Digitized by tine Internet Arciiive
in 2010 with funding from
University of Toronto
http://www.arcliive.org/details/transactionsof31amer
rHE AMEBlCAN SOCIETY OF MECHANICAL ENGINEERS
THE AMERICAN SOCIETY OF
ffi
AlKCfi WTOAT. FNOINFFI
\\\ \\S\rTI()\S
T^ATT7 M
UFFTf\'<,
OFIfi
T*.
^
<f.^f /^^. t
THE^ AMERICAN SOCIETY OF
ME( HANK AL ENGINEERS
TRANSACTIONS
VOLUME M
WASHINGTON MEETING
NEW YORK MEETING
1909
NEW YORK
PUBLISHED BY THE SOCIETY
29 West 39th Street
1910
I
Copyright 1910 by
THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS
OFFICERS
THE AMERICAN SOCIETY OF MECHANICAL
ENGINEERS
1909
FORMING THE STATUTORY COUNCIL
PRESIDENT
Jesse M. Smith New York.
VICE-PRESIDENTS
L. P. BuECKENRiDGE Urbana, 111.
Fred J. Miller Center Bridge, Pa.
Arthur West E. Pittsburg, Pa.
Terms expire at Annual Meeting of 1909
Geo. M. Bond Hartford, Conn.
R. C. Carpenter Ithaca, N. Y.
F. M. Whyte New York
Terms expire at Annual Meeting of 1910
PAST PRESIDENTS
Members of the Council for 1909
Ambrose Swasey Cleveland, O.
John R. Freeman Providence R.I.
Frederick W. Taylor Philadelphia, Pa.
F. R. Hutton New York
M. L. Holman St. Louis, Mo.
MANAGERS
G. M. Basford New York
A. J. Caldwell (Deceased) Newburg, N. Y.
A. L. Riker Bridgeport, Conn.
Terms expire at Annual Meeting of 1909
Wm. L. Abbott Chicago, 111.
Alex. C. Humphreys New York
Henry G. Stott New Rochelle, N. Y.
Terms expire at Annual Meeting of 1910
H. L. Gantt New York
I. E. MouLTROP Boston, Mass.
W. J. Sando Milwaukee, Wis.
Terms expire at Annual Meeting of 1911
TREASURER
WiLLiA-M H. Wiley New York
CHAIRMAN OF THE FINANCE COMMITTEE
Arthur M. Waitt New York
HONORARY SECRETARY
F. R. HuTTON New York
SECRETARY
Calvin W. Rice 29 West 39th Street, New York
V
PAST PRESIDENTS
Thurston, R. H 1880-1882 Died Oct. 25, 1903
Leavitt. E. D 1883 ^ Cambridge, Ma,ss.
Sweet, John E 1884 Syracuse, N. Y.
HoLLOWAY, T. F 188.5 Died Sept. 1, 18%
Sellers, Coleman 1886 Died Dec. 28. 1907
Babcock, George H 1887 Died Dec. 16, 1893
See, Horace 1888 Died Dec. 14. 1909
TowNE, Henry R 1889 New York.
Smith, Oberlin 1890 Bridgeton, N. J.
Hunt, Robert W 1891 Chicago, III.
LoRiNG, Charles H 1892 Died Feb. 5, 1907
CoxE, EcKLEY B 1892-1894 • Died May 13, 1895
Davis, E. F. C
Billings, Charles E
Fritz, John
Warner, Worcester R
Hunt, Charles Wallace .
Melville, George W . . . .
Morgan, Charles H
Wellman, S. T
.... 1894 Died Aug. 6. 1895
. . . .1895 Hartford, Conn.
1896 r BcThlehem. Pa.
1897 Cleveland, O.
.... 1898 New York.
.... 1899 Philadelphia, Pa.
.... 1900 Worcester, :\Ia.ss.
.... 1901 Cleveland, O.
Reynolds. Edwin 1902 Died Feb. 19, 1909
Dodge, James M 1903 Philadelphia, Pa.
FAST-FRESIDEXrS AND HONORARY COUNCILORS
1909
SwASEY, Ambrose 1904 Cleveland, O
Freeman, John R 1905 Providence, R. I.
Taylor, Fred. W 1906 Philadelphia. Pa.
Hutton, F. R 1907 New York.
HoLArAN, M. L 1908 St . Louis, Mo.
According to the Constitution, Article C 27, the five Past-Presidents who
last held the office shall be members of the Council, with all the rights, privi-
leges and duties of the otlier members of the Council.
EXECITIVE ( OMMITTKE OF IIIE ( OUN( IL
Jesse M. Smith, Cluiimian
Alex. C. Humphreys
¥. R. HuTTOiN
Fred J. Miller
F. M. Whyte
STANDING COMMITTEES
1909
FINANCE
ArthikM. Waitt (1), Chairmdii
Edward F. iSchxuck (2)
Waldo H. Marshall (5)
Geo. J. Roberts (3)
Robert M. Dixon (4)
HOUSE
Hknry S. Loud (1), Chairman Bernard V. Swenson (3)
William Cartkk Dickerman (2) Francls Blossom (4)
Edward Van Winkle (o)
LIBRARY
J(JHN W. LiEB, Jr. (4), Chairmnn
H. H. SUPLEE (1)
Chas. L. Clarke (5)
MEET IN (11^
Willis Fl. Hall (Ij, Chairiium
Wm. H. Bryan (2)
H. DE B. Parsons (5)
Ambrose Swasey (2)
Leonard W^aldo (3j
L. R. Pomeroy (3)
Charles K. Lucke (4)
MEMBERSHIP
Henry D. Hibbard (1), Chairman
Charles R. Richards (2)
Hosea Webster (5)
Francis H. Stillman (3)
Geor(;e J. FoRAN (4)
PUBLIC A TION
Arthur L. Williston (1), Chairman
D. S. Jacobus (2)
Geo. L Rockwood (5)
H. F. J. Porter (3)
H. W. Spangler (4)
RESEARCH
W. F. M. Goss (5), Chairman
Jas. Christie (1)
R. C. Carpenter (2)
R. H. Rice (3)
Chas. B. Dudley (4)
Note. — Numbers in parenthe.se.s indicate length of term in years that the member has yet to serve.
vii
SPECIAL COMMITTEES
1909
On a Standard Tonnage Basis for Refrigeration
D. S. Jacobus
A. P. Trautwein
John E. Sweet
E. F. Miller
On Society History
Chas. Wallace Hunt
On Constitution and By-Laws
Cnas. Wallace Hunt, Chairman
G. M. Basford
Jesse M. Smith
On Conservation of Natural Resources
Geo. F. Swain, Chairman
Charles Whiting Baker
g. t. voorhees
Philip DeC. Ball
H. H. Suplee
F. R. HUTTON
D. S. Jacobus
L. D. Burlingame
M. L. HOLMAN
Calvin W. Rice
On International Standard for Pipe Threads
E. M. Herr, Chairman
William J. BaLDWiN
On Thurston Memorial
Alex. C. Humphreys, Chairman
R. C. Carpenter
Fred J. Miller
Geo. M. Bond
Stanley G. Flagg, Jr.
Chas. Wallace Hunt
J. W. LiEB, Jr.
On Standards for Involute Gears
Wilfred Lewis, Chairman
Hugo Bilgram
D. S. Jacobus, Chairman
Edward T. Adams
George H. Barrus
Gaetano Lanza
On Power Tests
L. P. Breckenridge
William Kent
Charles E. Lucke
On Land and Building Fund
Fred J. Miller, Chairman
R. C. McKlNNEY
E. R. Fellows
C. R. Gabriel
Edward F. Miller
Arthur West
Albert C. Wood
James M. Dodge
On Student Branches
F. R. HuTTON, Honorary Secretary
viii
OFFICERS OF THE (iAS POWER SECTION
1909
CHAIRMAN
F. R. Low
SECRETARY
Geo. a. Orrok
GAS POWER EXECUTIVE COMMITTEE
F. H. Stillman, Chairman G. I. RocKWOod
F. R. HUTTON H. H. SUPLEE
R. H. Fernald
GAS POWER MEMBERSHIP COMMITTEE
Robert T. Lozier, Chairman D. B. Rushmore
Albert A. Gary A. F. Stillman
H. V. O. Goes G. M. S. Tait
A. E. Johnson George W. Whyte
F. S. King S. S. Wyer
GAS POWER MEETINGS COMMITTEE
Gecil p. Poole, Chairman E. S. McClelland
R. T. Kent C. T. Wilkinson G. W. Obert
GAS POWER LITERATURE COMMITTEE
C. H. Benjamin, Chairman L. S. Marks
H. R. Gobleigh T. M. Phetteplace
G. D. Gonlee G. J. Rathbun
R. S. DE Mitkiewicz W. Rautenstrauch
L. V. Goebbels S. a. Reeve
L. V. LuDY A. J. Wood A. L. Rice
GAS POWER INSTALLATIONS COMMITTEE
J. R. Bibbins, Chairynan A. Bement
L. B. Lent
GAS POWER PLANT OPERATIONS COMMITTEE
I. E. Mour.TROP, Chairman H. J. K. Freyn G. H. Parker
J. D. Andrew N. T. Harrington J. P. Sparrow
W. H. Blauvelt J. B. Klumpp A. B. Steen
V. Z. Garacristi G. L. Knight F. W. Walker
E. P. Goleman J. L. Lyon G. W. Whiting
G. J. Davidson D. T. MacLeod Paul Winsor
W. T. Donnelly V. E. McMullen T. H. Yawger
GAS POWER STANDARDIZATION COMMITTEE
G. E. Lucre, Chairman E. T. Adams
Arthur West James D. Andrew
J. R. Bibbins H. F. Smith
Louis G. Doelling
OFFICERS OF STUDENT BRANCHES
STUDENT BRANCH
Stevens Inst, of Tech..
Hoboken, N. J.
Cornell University.
Ithaca. N. Y.
Armour Inst, of Tech.,
Chicago, 111.
Iceland Stanford, Jr.
Universiry, Palo Alto.
Cal.
Polytechnic Institute,
Brooklyn, N. Y.
State Agri. College of
Oregon, Corvallis,
Ore.
Purdue University,
Lafayette, Ind.
Univ. of Kansas,
Lawrence, Kan.
New York Univ.,
New York
Univ. of Illinois,
L'rbana, 111.
Penna. State College,
State College, Pa.
Columbia University,
New York.
Mass. Inst, of Tech.,
Boston, Ma.ss.
LTniv. of Cincinnati,
Cincinnati, O.
Univ. of Wisconsin.
Madison, Wis.
.\UrHORIZED
HONORARY CH.\IR-
I
PRESIDENT
SECRETARY
BY CO UNCI I.
M\N
190S
Decern bei 4
Alex. C. Humphreys
H. II. Haynes
R. H. Upson
December 4
R. C. Carpenter
C. F. Hirshfeld
1909
March 9
C. F. Gebhardt
X. .1. Houghton
.M. C. Shedd
March 9
W. F. Duran.l
P. H. \'an Ktten
H. L. He.ss
March 9 W. D. Ennis J. S. Kerins
March 9 Thos. >L Cardner C.L.Knopf
March 9 L. ^". Ludy
March 9 P. F. Walker
November 9 C. E. Hougliton
November 9 W. F. M. (loss
November 9
Novemlier 9
November 9
Noveml)er 0
Novend)er 9
E. A. Kirk
H. S. Coleman
Harry Anderson
W. F. Colman
Fredk. A. Dewey
Percy Gianella
S. H. Graf
.1. R. Jackson
.John Ciarver
Andrew Hamilton
S. C. Wood
SUMMARY OF MEMBERSHIP
Dpcemhor31. 1909
United States
Alahama 19
Alaska 1
Arizona 5
Arkansas 2
California 74
Colorado 30
Connecticut 144
Delaware 18
District of Columbia 32
Georgia 19
Hawaii 3
Idaho 2
Illinois 241
Indiana o9
Iowa 9
Kansas 11
Kentucky 6
Louisiana 30
Maine 15
Maryland 33
-Massachusetts 339
Michigan 110
Minnesota 22
Mississippi 1
Missouri 64
^lontana 10
Address unknown
Nebraska 3
Nevada 5
New Hampshire lo
New Jersey 202
New Mexico 2
New York 1062
North Carolina 13
North Dakota 1
Ohio 277
Oklahoma 1
Oregon 11
Pennsylvania 4.59
1
3
69
3
13
15
9
11
28
Philippine Islands
Porto Rico
Rhode Island
South Carolina
Tennessee
Texas
Utah
Vermont
Virginia
Washington 15
West Virginia 8
Wisconsin 93
Wyoming 1
Total 3619
6
Foreign Countries
Africa 14
Australia S
Belgium 5
Canada 45
Central America 0
China 3
Cuba 4
England 45
Finland 1
France 10
Germany 8
Holland 1
Hungan.- 2
India
Italy
Japan
Mexico
New Zealand . .
Norway
Russia
.Scotland
South America.
Sweden
Switzerland. . . .
4
1
9
14
1
1
3
3
11
Total.
207
SUMMARY OF MEMBERSHIP
Bt Residence
December 31, 1909
Membership in United States 3619
Foreign membership ... 207
Address unknown 6
Total 3832
By Grade
December 31, 1909
Honorary members 15
Members 2565
Associates 398
Juniors 854
Total membership (including life members) 3832
MEMBERSHIP OF GAS POWER SECTION
Alabama.
United States
.... 4 Missouri .
California 7
Connecticut 5
District of Columbia 5
Delaware 1
Georgia 1
Illinois 22
Indiana 9
Kansas 2
Maryland 1
Massachusetts 25
Michigan 13
Minnesota 5
Nebraska 2
New Jersey 14
New York 141
Ohio 29
Pennsylvania 25
Rhode Island.
Vermont
Virginia
Washington. .
Wisconsin
7
1
1
2
17
Total 347
Foreign Countries
Belgium. .
Canada. .
Germany
Mexico
Switzerland.
Total.
GAS POWER SECTION
By Residence
Membership in United States 347
Membership in foreign countries 6
Total membership 353
By Grades
Members of the Society 222
Affiliates 131
Total 353
STUDENT BRANCHES
Armour Institute of Technology 16
Brooklyn Polytechnic Institute 22
Columbia University 4
Cornell Universitj^ 123
Leland Stanford, Jr., University 13
Massachusetts Institute of Technologj^ 1
Pennsylvania State College 35
Purdue University 3
State Agricultural College of Oregon 10
Stevens Institute of Technology 55
University of Cincinnati 23
University of Illinois 44
University of Kansas 7
University of Wisconsin 26
Total 382
ATTENDANCE AT MEETINGS, 1909
The following figures show the attendance at the several meet-
ings of the Society daring 1909:
Jjiuuary 12 New York 168
February 23 New York 133
March 9 New York 251
March 24 New York 625
April 13 New York 307
May 4-7 Washixgtox, D. C. Sprinc; Meeting .... Members. . .276
Guests 333 609
October 12 New York 192
November 9 New York 161
December 7-10 New York. Axxual Meeting Members. . .628
Guests 435 1063
CONTKNTS OF VOLUME 31
Washington, New York and Monthly Meetings, 1909
Page
Biography of Jesse M. Sinitli 3
No. 1229 Monthly Meetings, January to June; Washington Meeting. 5
Xo. 1230 Carl G. Barth The Transmission of Power by
Leather Belting 29
Xo. 1231 F. M. Whyte Safety Valves for Locomotives 105
Xo. 1232 P. G. Darling Safety Valve Capacity 109
Xo. 1233 Safety Valve Discussion 129
Xo. 1234 Ellis C. Soper A Unique Belt Conveyor 151
Xo. 1235 C. Kemble Baldwin Automatic Feeders for Handling
Material in Bulk 161
Xo. 1236 W. H. Kenerson A New Transmission Dynamometer 171
N^o. 1237 A. Kingsbury Polishing Metals for Examination
with the Microscope 181
Xo. 1238 C. L. Straub Marine Producer Gas Power 185
Xo. 1239 C. W. Obert Operation of a Small Producer-
Gas Power Plant 209
Xo. 1240 T. ^I. Phetteplace Offsetting Cylinders in Single-
Acting Engines 223
Xo. 1241 Presentation of Portrait of George W. Melville 253
Xo. 1242 G. A. Orrok Small Steam Turbines 263
Xo. 1243 E. M. IvENs Compressed Air Pumping Systems of
Oil Wells 311
Xo. 1244 C. H. Peahody The Specific Volume of Saturated
Steam 333
Xo. 1245 R. C. H. Heck Some properties of Steam 345
No. 1246 H. V. Wille A New Departure in Flexible Stay-
bolts 359
Xo. 1247 Hudson-Fulton Celebration 373
X'o. 1248 Meetings, October to December; Annual Meeting 381
X"o. 1249 Annual Reports of Council and Committees 409
X'o. 1250 Jes.se M. Smith The Profession of Engineering 429
Xo. 1251 R. C. Carpenter High-Pressure Fire-Service Pumps
of Manhattan Borough 437
X'o. 1252 Gaetano Lanza 1 Stres.ses in Reinforced Concrete
L. S. Smith / Beams 511
X'o. 1253 Walter Rauten- Design of Curved Machine Mem-
STRAUCH ber.s under Eccentric Lord 559
Xo. 1254 C. M. Allen Tests on a Venturi Meter for Boiler
Feed 589
No. 1255 G. F. Gkhhardt The Bitot Tubd as a Steam Meter. . 601
CONTENTS
Page
No. 1256 F. H Sibley | Efficiency Tests of Steam Nozzles. . 617
T. S. Kemble J ^
No. 1257 C. C. Thomas An Electric Gas Meter 655
No. 1258 D. M. Myers Tan Bark as a Boiler Fuel 685
No. 1259 J. R. BiBBiNS Cooling Towers for Steam and Gas
Power Plants 725
JSo. 1260 W. P. Caine Governing Rolling Mill Engines. ... 783
No. 1261 F. W. Dean An Experience with Leaky Vertical
Fire Tube Boileis 799
No. 1262 F. W. Dean The Best Form of Longitudinal Joint
for Boilers 823
No. 1263 C.M.Garland \ Testing Suction Gas Producers
A. P. Kratz / with a Koerting Ejector 831
No. 1264 J. R. BiBBiNS Bituminous Gas Producers 877
No. 1265 Walter Ferris The Bucyrus Locomotive Pile Driver 905
No. 1256 Henry Hess .Line-Shaft Efficiency, Mechanical
and Economic 923
No. 1267 A. F. Nagle Pump Valves and Valve Areas 953
No. 1268 A. F. Nagle A Report on Cast-Iron Test Bars. . 977
No. 1269 I. N. HOLLIS 1 ^^ . r- . r -r.-^^-
_, .^ .,. 1 Symposium on Cast-iron Jittmgs
E. F. Miller r ^ a i 4- .i o* nQo
. „ ^ tor Superheated Steam 989
No. 1270 Necrology 1039
No. 1271 Index 1059
TRANSACTIONS
OF
TPIE AMERICAN SOCIETY OF
MECHANICAL ENGINEERS
VOLUME 31-1909
T
HIS volume contains the papers and proceedings of The Ameri-
can Society of Mechanical Engineers for the year 1909,
covering the thirtieth year of the Society's history.
The newly-elected President, Jesse M. Smith of New York, was
introduced, as is customary, at the Annual Meeting of the Society
in December 1908, an account of which appeared in Volume 30 of
Transactions, covering the proceedings of 1908.
The annual report of the Council, presented at the annual meet-
ing of 1909, gives a record of the work of the year, and follows an
account of that meeting.
JESSE MEKKICK SMITH
Jesse M. Smith, President of the Society for 1909, was born
in Newark, Ohio, in 1848. He moved to Detroit, Michigan, with his
father's family in 1862. In 1865 he entered Rensselaer Polytechnic
Institute, Troy, New York, remaining there three years. The follow-
ing year he spent traveling in Europe, and entered L'Ecole Centrale
des Artset Manufactures, Paris, France, receiving after three years
of study the degree of M. E, in 1872. During his vacations, he trav-
eled among the manufacturing plants of France, Germany and Bel-
gium, and attended lectures in the Polytechnic Institute in Berlin.
After graduation in Paris he traveled three months among the iron
and machine works of England,
He began the practice of engineering in 1873, designing and super-
intending the erection of blast furnaces for smelting iron from native
ores with raw bituminous coal in the Hocking Valley, Ohio. He
made surveys of coal mines, opened mines and built coal handling
machinery for them. He surveyed and constructed railroads from
mines to furnaces.
Upon the death of his father in 1880, Mr. Smith returned to Detroit
and opened an oflEice as Consulting Engineer. He designed and con-
structed a high-speed center-crank steam engine with shaft governor,
containing the feature of the modern inertia weight governor, and
put it in operation driving a Brush dynamo producing 40 arc lights
in 1883, In 1890 he presented a paper before the Society on this
governor.
He represented the United States Electric Lighting Company in
Ohio and Michigan from 1884 to 1886, during which time he erected a
number of the early incandescent electric light plants, including
one of 1000 Ughts in the Stillman Hotel, Cleveland, Ohio, which was
the first hotel lighted exclusively and continuously by electricity
from its own plant. He returned to the work of consulting engineer
in 1886 and continued in it until 1898. During this time he designed
and erected several power plants and several plants for electric Hght-
ing and electric railways; also apparatus for steam heating with
exhaust steam in several large manufacturing plants.
3
4 JESSE MERRICK SMITH
He began in 1883 to be called as an expert witness in the U. S.
courts in patent litigation. This practice gradually grew and dis-
placed the work of consulting engineer until 1898, when he moved
to New York to continue his practice as expert in patent cases
exclusively.
Among the notable cases in which he has acted as expert are: steam
injectors under the Hancock Inspirator patents; cylinder lubricators
for locomotives; roller mills and middlings purifiers for flour manufac-
ture; cyclone dust collectors; quick action air brakes under Westing-
house patents; pneumatic tires for automobiles; automobiles under
the Selden patent; induction electric motors under Tesla patents;
pressure filters; incandescent electric lamps; steam heating apparatus;
typewriters; armored concrete construction; the cabulagi-aph, etc.
He became a member of the Society in 1883 and was a member of
the Council as Manager, 1891 to 1894, and Vice-President 1894 to 1896
and 1899 to 1901.
He is a charter member of the American Institute of Electrical
Engineers; La Soci6t^ des Ingenieurs Civils de France; 1' Association
des Anciens fileves de I'Ecole Centrale des Arts et Manufactures;
The Detroit Engineering Society; the Society for the Advancement
of Science; the National Geographical Society; the Engineers Club;
the Machinery Club and the Ohio Society of New York.
No. 1229
MEETINGS rJxVNUARV-.JUNE
NEW YORK MEETING, JANUARY 12
The first meeting of the Society for the year 1909 was held in the
Engineering Societies Building, on the evening of January 12, when a
paper on The Transmission of Power by Leather Belting was given by
Carl G. Barth.
This paper, with the discussion, constitutes one of the most compre-
hensive presentations of the subject of belting that has been given
before the Society. Mr. Barth has deduced a theory of belting based
on the well-known experiments of Lewis and Bancroft and other engi-
neers who have investigated different factors of the belting problem.
Moreover, in systematizing manufacturing plants and especially
machine shops, where the scientific operation of machine tools is
involved, a careful study was made of the whole belting problem,
resulting in additional material for his paper.
Following the discussion pertaining strictly to the subject-matter of
the paper, there was a general discussion upon the transmission of power
by electricity and by rope, and by the modern types of chains used for
power transmission. Written discussions were submitted by the fol-
lowing: A. F. Nagle, Prof. Wm. W. Bu-d, Prof. 0. H. Benjamin, H. K.
Hathaway, Prof. L. P. Breckenridge, and Prof. W. S. Aldrich. Oral
discussions were given by Henry R. Towne, Wilfred Lewis, W. D.
Hamerstadt, Fred. W. Taylor, Charles Robbins, Geo. N. Van Derhoef,
Walter C. Allen, Dwight V. Merrick, Fred. A. Waldron, S. B. FHnt
and A. A. Gary.
NEW YORK MEETING, FEBRUARY 23
A meeting was held in the Engineering Societies Builduig on Tues-
day evening, February 23, the subject for discussion being Safety
Valves. The meeting was opened by Frederic M. Whyte, general me-
chanical engineer of the New York Central Lines, with a paper upon
Safety Valves, giving special attention to locomotive practice. He
6 SOCIETY AFFAIRS
was followed by Philip G. Darling, mechanical engineer with Man-
ning, Maxwell & Moore, with a paper on Safety-Valve Capacity.
The papers and discussion, which covered locomotive, marine and
stationary practice, as well as the use of safety valves on heating
boilers, brought together such late data as were available and em-
phasized the need of further information for the purpose of establish-
ing a more rational and uniform practice.
The papers were discussed by: L. D. Lovekin, A. C. Ashton, A. B.
Carhart, E. A. May, H. O. Pond, F. J. Cole, Dr. Chas. E. Lucke, Jesse
M. Smith, G. P. Robinson, Wm. H. Boehm, H. C. McCarty, M. W.
Sewall, Geo. I. Rockwood, A. A. Gary, Dr. A. D. Risteen, F. L. Du-
Bosque, N. B. Payne, Frank Creelman. The discussion was ad-
journed to the Spring meeting.
NEW YORK MEETING, MARCH 9
A particularly interesting occasion was the lecture on Modern
Physics, given by Dr. William Hallock, Professor of Physics, Colum-
bia University, on Tuesday evening, March 9.
The lecture included a review of discoveries introductory to the
X-ray, radio-activity and allied phenomena; experimental demon-
stration of different forms of radiation, including heat; development
of the essential identity of radiant light, heat and Hertz waves,
together with the evidence of the electro-magnetic nature of light
radiations; differentiations between these forms of radiation and
those of so-called radio-active material, followed by the bearing of
the facts developed by radio-activity upon the possible genesis of
the chemical elements; the kinetic theory of gases and its relation to
the modern theory of solutions; the moving ion as the determining
factor ijn electrical conduction; the distinction between the chemical
and the physical ion; the atom and the relation of its structure to the
phenomena of radiation and absorption; the principle of relativity
and its relation to the structure of the atom and the electron; the
universal application of the /orce, mass, time theory to molecular and
cosmic phenomena.
JOINT MEETING ON CONSERVATION
A meeting of the national engineering societies on the conserva-
tion of our natural resources was held in the Engineering Societies
Building on the evening of March 24. Onward Bates, President
SOCIETY AFFAIRS 7
Am. Soc. C. E., who was expected to preside, was unable to
attend, and Dr. James Douglas, Past-President Am. Inst. M. E.,
acted as chairman. At the opening of the meeting, the Chairman
announced a congratulatory telegram from President Taft, which
was read by John Hays Hammond, President Am. Inst, M, E.
In his opening remarks Dr. Douglas said that in a great movement
of this kind there could be no dividing line between engineers in dif-
ferent branches of the profession. The great inventions like that of
the Bessemer process had required a combination of the skill of
engineers who had specialized in different fields. He said that in
looking back we must be struck with the advance made in the reduc-
tion of waste in the use of natural supplies, especially in saving coal,
both in mining it and in using it in metallurgical work.
The first address was upon The Conservation of Water, by John
R. Freeman, Past-President Am. Soc. M. E., Mem. Am. Soc. C.
E., consulting engineer of the Department of Additional Water
Supply for the City of New York. He spoke of the relation of
stream flow to lumbering, emphasizing the importance of accurate
stream measurements in order to obtain precise knowledge of the
effect of forests and of the value of water powers. Interesting
figures were given, comparing the efficiency of turbines of the old
days with those of the present time. Other phases of the conserva-
tion of water, such as the purity of the water courses, navigation,
irrigation, etc., were considered. He recommended the collecting of
facts by the different States, regarding the notable opportunities for
power development within their borders, and the making of care-
ful surveys, thus placing reliable information at the disposal of those
inclined to take advantage of such natural opportunities for power.
The address of Dr. R. W. Raymond, Secretary Am. Inst. M. E.
was upon Conservation by Legislation. He defined true conservation
as the diminution not of use but of waste. The best method for the
prevention of waste is by the progressive education of the people,
rather than by legislation. He urged that government information
pertaining to natural resources and their conservation should be
collected with care and not hurried, and stated without bias or argu-
ment in favor of any measure or policy. Hasty and ill-considered
legislation, especially if advocated by selfish interests, is a peril. He
dealt with specific examples of such legislation and urged that the
work of the departments of the Federal government should be care-
fully planned in advance instead of expanding without a definitely
arranged plan.
8 SOCIETY AFFAIRS
Charles Whiting Baker, Mem. Am. Soc. M. E., Editor of Engi-
neering News, spoke on The Waste of our Natural Resources by
Fire. He gave new statistics upon the fire laws in the United States,
with the striking illustration that we are burning every year in
this country a street of buildings a thousand miles long that would
reach from New York to Chicago. That this destruction is not
necessary is proved by the experience of European countries where
the per capita fire loss is in most cases only a few cents annually,
while in this country it is $2.50. Referring to the destruction by
forest fires, he said that effective laws for the protection of forests
must be enacted before capital will be invested in the development
or preservation of timber lands.
The last address was by Lewis B. Stillwell, Mem, A. I. E. E., con-
sulting electrical engineer, upon Electricity and the Conservation
of Energy. He illustrated by interesting figures the function of
electricity in the conservation of power resources, showing results
accomplished in three typical cases, namely, the plants of the
Niagara Falls Power Co., the Northeast Coast Power System at
Newcastle-upon-Tyne, and the plants of the Interborough Rapid
Transit Co., New York. The Niagara plant showed the possibility
in water power development and the Northeast Coast plant the
economy resulting from the substitution of large steam-driven units
for small steam plants, widely distributed. In the case of the Inter-
borough Company, comparisons were made of the cost under the
present system of electrical distribution and that which would
have obtained if locomotives had been used instead.
ST. LOUIS MEETING, APRIL 10
A meeting of members of the Society residing in St. Louis and vicin-
ity was called by Wm. H. Bryan, member of the Meetings Committee,
and held on Saturday evening, April 10. Prof. E. H. Ohle acted as
secretary and about twenty engineers were present. This was the
first monthly meeting of the Society to be held outside of New York
City. All present expressed themselves in favor of local meetings and
it was voted that a committee of three, composed of the chairman and
two others appointed by him, should be formed to lay out a plan of
organization and to report in sixty days.
The following topics were discussed: a local organization with occa-
sional professional and social meetings; 6 increase in membership; c
contributions to The Journal; d making up a party to attend the
SOCIETY AFFAIRS 9
Spring Meeting at Washington, May 4-7; c extending an invitation to
the Society to meet in St. Louis at some future time; /other means of
promoting the Society's welfare, not only locally, but generally.
JOHN FRITZ MEDAL AWARD
The John Fritz Medal, the only medal which the four National
Engineering Societies confer, was presented to Charles T. Porter,
Hon. Mem, Am. Soc. M. E., on Tuesday evening, April 13. The pre-
sentation took place in the auditorium of the Engineering Societies
Building, before distinguished invited guests and an audience
representing the entire engineering profession. The medal was
conferred upon Mr. Porter for his work in advancing the knowledge
of steam engineering and for improvements in engine construction.
Addresses were made by Dean W. F. M. Goss of the University of
Illinois, upon The Debt of Modern Civilization to the Steam Engine
as a Source of Power; by Prof. F. R. Hutton of Columbia University,
Honorary Secretary Am. Soc. M, E,, on The Debt of the Modern
Steam Engine to Charles T. Porter; by Robert W. Hunt of Chicago, on
The Debt of the Era of Steel to the High Speed Steam Engine; by
Frank J. Sprague of New York, on The Debt of the Era of Electricity
to the High-Speed Steam Engine.
Henry R. Towne, Past-President Am. Soc. M. E., and Chairman
of the Board of Award of the John Fritz Medal for 1909-1910, presided
at the meeting, and in his opening remarks spoke briefly of the
origin and history of the medal, introducing Dean W. F. M. Goss of
the University of Illinois.
At the close of Professor Goss's address, Mr. Towne in a short intro-
ductory speech recalled that Mr. Porter was the third person and the
first American to whom was accorded the distinction of Honorary
Membership in The American Society of Mechanical Engineers. On
account of this relation, Mr. Porter was introduced by Jesse M.
Smith, President of the Society, who said, by way of introduction:
The John Fritz Medal, estabUshed in 1902 by the American engineering pro-
fession as a meed of recognition for 'notable scientific or industrial achievement,'
was awarded in the year 1908 by a board representing the four National Engineer-
ing Societies, to a distinguished mechanical engineer for 'his work in advancing the
knowledge of steam-engineering and for improvements in engine construction.'
I present to you, and to this company, the engineer to whom this high distinction
has been granted.
10 SOCIETY AFFAIRS
He is honored because he saw the possibiUties of the high-speed steam engine;
because his mechanical genius in design made those possibilities real; and because
he recognized the necessity for, and then applied, the very best mechanical con-
struction to the realization of his ideals.
He then introduced into the development of the power plant an idea and an
influence so revolutionary as to make an epoch in the history of the art of engine
building; and which has been as world-wide in its effects as the use of the recipro-
cating engine.
Many of the present generation of engineers have inherited, without effort and
often without knowledge of their origin, the results which cost him many years of
painstaking study and experiment to establish.
That he may receive the John Fritz Medal awarded to him, I now have the
honor to present Charles Talbot Porter.
Mr. E. Gybbon Spilsbury, Chairman of the Board of 1908, by
which the award was made, said in presenting the medal to Mr.
Porter:
Under instructions from the Board of Award of the John Fritz Medal, it is my
privilege and pleasure to inform you that for your work in advancing the knowledge
of steam engineering and for improvements in engine construction, you have been
chosen as the worthy recipient of the medal for the year 1908-1909.
This medal was instituted in 1902 to commemorate the 80th anniversary of the
successful and honored career of our beloved colleague John Fritz, and its award by
a committee selected from the membership of the four great engineering societies
of the United States is the highest honor which the engineering profession can
V onfer on any of its members.
Charles Talbot Porter,in the presence of this distinguished company, I now present
you this medal, together with an engraved certificate of the award, and confer
upon you all the rights and honors and the distinction which attach to this emblem.
May you live long and happily to enjoy the appreciation which is your due at the
hands of those you have so benefited by your work.
After the presentation, Mr. Towne read the following telegram
from John Fritz:
With all my heart regret my inability to be with my dear friends and associ-
ates this evening. I cannot be with you in person, but I will be with you in
spirit. Please convey to my dear friend Porter my sincere congratulations and
best wishes.
Congratulatory cablegrams were also read from Wm. H. Maw, Editor
of London Engineering , from the sons and grandsons of Wm. A. Hoyle,
with whom Mr. Porter was associated in his early work, the Iron and
Steel Institute of Great Britain, the Institution of Mechanical Engi-
neers of Great Britain, E. D. Leavitt, Past-President Am.Soc.M.E.,
an early associate of Mr. Porter, and many others.
The addresses of the evening followed.
SOCIETY AFFAIRS 11
ADDRESS OF DEAN W. F. M. QOSS
Dean W. F. M. Goss spoke of the debt of modern civilization to
the steam engine. Dreams of the possibilities of steam belong to
the days of Addison, Steele, Swift and Defoe; days when there were bril-
liant men of letters, triumphs in architecture, achievements on the
battlefield, but when there were no means for performing industrial
work. There were no large factories in England because there was
no way by which their machinery could be driven. Mines were
abandoned because they were flooded with water; women and girls
were toiling in the mines amid suffering and degradation. The move-
ment of merchandise by land was laborious and traveling by sea slow
and dangerous.
Into the midst of such conditions came the steam engine. It freed
the mines of England from water, revived dormant industries, intro-
duced new systems of manufacture, supplied power, water and effective
means of sanitation to cities, and was later supplemented in all these
respects by electricity for lighting, power and transportation.
Steam usurped the place of wind in the propulsion of ships, and through
the agency of the locomotive has carried civilization to the farthest
ends of the earth. These achievements are direct contributions to
the upbuilding of civilization, the key-note of which is service. The
dwellers on the earth are beginning to see that if one nation suffers
severely, all are likely to suffer in some degree, and they are learning
sympathy for their fellow-men.
ADDRESS OF PROF. F. R. BUTTON
It was assigned to Prof. F. R. Hutton to speak in detail of the debt
of the reciprocating steam engine to the pioneer work of Mr. Porter.
It owes to him the first vision of the advantages that come from
making the crank shaft turn at a high rate of revolution, whereby
the weight of the motor per horsepower is reduced. From this
seed-thought has sprung the modern design of the motor for the
self-propelled vehicle and for the aeroplane. The high speed in-
volved the solution of difficult problems, owing to the necessity for
starting and stopping heavy parts of the mechanism in each revo-
lution. To Mr. Porter we owe the recognition of these problems.
Perhaps the most important debt of all is the requirement that the
standard of mechanical construction in the high-speed engine must be
of the highest type. We owe to Mr. Porter many manufacturing
details which now are commonplaces of modern practice.
12 SOCIETY AFFAIRS
Mr. Porter created a form of steam-engine condenser to be attached
directly to the engine and operated at a much higher rate of speed
than that at which the ordinary pump could be used; and finally,
invented a sensitive steam engine governor in two forms.
The address closed with a tribute to Prof. Chas. B. Richards, associ-
ated with Mr. Porter's early work of designing, and John F. Allen,
who had conceived many details of the first high-speed engine which
Mr. Porter combined into a harmonious whole.
ADDRESS OF ROBERT W. HUNT
Robert W. Hunt said it was scarcely conceivable that one could
have witnessed in a single lifetime the remarkable development in
the steel industry which he had observed since the birth of the Besse-
mer processes. These accomplishments were made practically pos-
sible by the discovery of a more rapid power. The early processes
were deliberate because man was habituated to slow movements.
The first power came from the slow-turning water wheel; later from
the slow-speed steam engine. Faster movements were obtained
through gears and belts. Among the first engineers to attach the
rolling mill engine direct to its train of rolls were John and George
Fritz, but the speed of the stroke of their engine was limited.
Charles T.Porter was the first to give to the rolling-mill a controllable
direct-connected economical high-speed engine.
Mr. Hunt referred to two engines in a rolling-mill plant in Troy,
N. Y.,in 1876. One set of rolls was driven by a walking-beam low-pres-
sure engine, taken from the 'steamboat Swallow, a Hudson River
boat, and the other set was driven by Porter-Allen engines. The .
contrast between the steamboat engine with a slow speed of 35 or
40 r.p.m., and Mr. Porter's little engines, humming away at high
speed, and accomplishing much greater results, was an instructive
sight.
ADDRESS OF FRANK J. SPRAGUE
Frank J. Sprague recalled that in 1867, at the French exhibition,
Charles T. Porter installed two Porter- Allen engines, the only high-
speed engines exhibited, to drive generators for supplying current for
lighthouse apparatus. While these engines were not directly coupled,
it is a curious fact that the piston speeds and revolutions were what
is common today in isolated direct-coupled plants. In the dozen years
following, Mr. Porter built many engines with certain common char-
acteristics, high piston speed and revolutions, solid engine bed and
SOCIETY AFFAIRS 13
babbitted bearings, but there was no direct coupliug to dynamos
until 18S0, when Mr. Porter installed a high-speed engine for Mr.
Edison in his laboratory at Menlo Park. Shortly after this Mr. Porter
was invited to construct for the Edison Station at Pearl Street, New
York, the first of a series of engines for so-called steam dynamos, each
independently driven by a direct -coupled engine.
Mr. Sprague likened the relations of electricity and the high-speed
engine, not to debtor and creditor, but rather to a close partnership,
an industrial marriage, one of the most important in the engineering
world, that of the prime mover and the electric generator. Here were
two machines, destined to be joined together, economizing space,
increasing economy, augmenting capacity, reducing investments,
increasing dividends. Primarily and largely due to Mr. Porter, the
high speed possibilities of the engine were commercially demonstrated.
BOSTON MEETING, APRIL 16
A meeting was held in Boston in the auditorium of the Edison
Building, on April 16, 1909, to discuss the advisability of holding meet-
ings of the Society in that city. Irving E. Moultrop, Manager of the
Society, was elected temporary chairman and Ralph E. Curtis tem-
porary secretary. About 160 members and guests were present,
including the President and the Secretary of the Society. There was
a general discussion in which the following participated: Irving E.
Moultrop, Henry Bartlett, Henry F. Bryant, Vice-President of the
Boston Society of Engineers, James D. Andrew, Fred R. Low, Paul
Winsor, Prof. W. W. Bird, Prof. Geo. F. Swain, Prof. L. S. Marks,
Prof. D. C. Jackson, Prof. Gardner C. Anthony, Francis W. Dean,
E. G. Bailey, Prof. C. G. Lanza.
During the discussion there were brief addresses by the President
and the Secretary. President Smith said that it is the desire of the
officers of the Society that these meetings be as free and open as is
consistent with the traditions and the high professional standards
which the Society has maintained during its thirty years of experi-
ence. He emphasized the fact that such meetings are meetings of the
Society as a body rather than of local sections or branches. Papers
presented would be published in The Journal when accepted, making
it possible to discuss them in all the cities where meetings are held.
He emphasized the Society's friendly spirit of cooperation with other
engineering societies and the particular esteem in which engineers and
members held the Boston Society of Civil Engineers.
14 SOCIETY AFFAIRS
Secretary Rice said it had become evident that two conventions a
year are not sufficient for a national society and that the holding of
meetings more frequently in one place does not create a national
spirit. He said the question before the meeting was, how the
engineering profession of Boston and vicinity can best get together
for the common and individual good; and stated that The American
Society of Mechanical Engineers desires to do what will best serve tlie
profession. He expressed the hope that that would be accomplished
by bringing together the various organizations in a common head-
quarters rather than by the formation of a new organization. Co-
operation and coordination, he declared, should be the motto of the
profession.
ST. LOUIS MEETING, MAY 15
A meeting was held at the Missouri Athletic Club, St. Louis, on
May 15, to discuss further the question of holding meetings of the
Society in St. Louis. William H. Bryan, member of the Meetings
Committee, presided, and Prof. E. L. Ohle acted as Secretary.
The report of the committee on organization, recommending that
the Society cooperate with the Engineers Club of St. Louis in the
matter of meetings and publication, was presented and discussed by
the following: M. L. Holman, Past-President Am.Soc. M.E., R. H. Tait,
Wm. H, Bixby, Thomas Appleton, Professor Westcott, F. L. Jefferies,
W. M. Armstrong, Prof. H. Wade Hibbard, J. A. Laird, Victor Hugo,
E. A. Fessenden. In opening the discussion Mr. Holman said that
the question of enlarging the sphere of usefulness of national engineer-
ing societies without interfering with local organizations is one which
has been given much thought by the engineers of the country. It
has taken years to bring the St. Louis Engineers Club, an earlier
organization than The American Society of Mechanical Engineers, up
to its present standing and it could not afford to take steps which
would interfere with its usefulness or impede its growth and impor-
tance. This movement, however, was not intended to antagonize
local clubs but was for the purpose of bringing more engineers into
the societies, both local and national, thus benefiting both organiza-
tions.
Following the discussion, the report was unanimously adopted and
the meeting concluded with an interesting running account, given by
Professor F. H. Vose, of the papers and discussion presented at the
Washington meeting.
SOCIETY AFFAIRS 15
BOSTOiN MEETING, JUNE 11
In accordance with the plans of the prelimhiary meeting of May
15, a professional meeting was held on June 11, and the paper on
Small Steam Turbines given by George A. Orrok at the Washington
meeting was presented for further discussion.
Prof. Ira N. HoUis, who presided, first outlined the work proposed for
the meetings of the Society in Boston, saying that the committee was
planning a number of meetings to be held during the fall and winter.
The subjects of several unusually timely papers promised for these
meetings were announced. As indicating the large number of engi-
neers in the vicinity who might attend, it was stated that notices had
been sent to 950, of which number 340 were members of the Society.
The meeting was addressed briefly by Secretary Rice, after which
Mr. Orrok's paper was read by Prof. E. F. Miller and the following
joined in the discussion: Dr. L. C. Loewenstein, J. A. London, Chas.
B. Rearick, F. B. Dowst, Chas. B. Edwards, V. F. Holmes, J. S. Schu-
maker, Prof. C. A. Read, Prof. I. N. Hollis, Prof. E. F. Miller, John T.
Hawkins, R. H. Rice, Chas. H. Manning, C. P. Crissey, Chas. B.
Burleigh.
IQ SOCIETY AFFAIRS
SPRING MEETING, WASHINGTON, D. C.
The 58th meeting of the Society was held in Washington, D. C, at
the New Willard Hotel, May 4-7. The total registration was 609, of
which 276 were members. Fewer professional sessions than usual were
arranged by the Meetings Committee in order that visiting members
and their guests might avail themselves of opportunities to see places
of interest at the national capital.
PROGRAM
OPENING SESSION
Tuesday Evening, May 4, «' ^-^^ o'clock
Reception of the members by the Washington Society of Engi-
neers and the local members at the New Willard Hotel. Music by
the Marine Band.
Address of welcome by Hon. Henry B. F. Macfarland, President
of the Board of District Commissioners.
Response by Jesse M. Smith, President of the Society.
SECOND SESSION
Wednesday Morning, May 5
Business Meeting: Reports of committees, tellers of election; new
business.
A Unique Belt Conveyor, Ellis C. Soper.
Discussed by T. A. Bennett, Harrington Emerson, Fred J.
Miller.
Automatic Feeders for Handling Material in Bulk, C. Kemble
Baldwin.
Discussed by T. A. Bennett.
A New Transmission Dynamometer, Prof. Wm. H. Kenerson.
Discussed by A. F. Masury.
Polishing Metals for Examination with the Microscope,
Albert Kingsbury.
SOCIETY AFFAIRS 17
Wednesday Afternoon
Special exhibition drill by troops at Fort Myer.
Wednesday Evening
Illustrated lecture by Arthur P. Davis, Chief Engineer of theU. S.
Reclamation Service, on Home-Making in the Arid Regions.
THIRD SESSION
Thursday Morning, May G
GAS power session
Report of the Standardization Committee.
Marine Producer Gas Power, C. L. Straub.
Discussed by Geo. Dinkel, Henry Penton, I. E. Moultrop, H.
M. Wilson, E. T. Adams.
I Operation of a Small Producer Gas Power Plant, C.W. Obert
Discussed by J. A. Holmes, J. H. Norris, W. A. Bole.
A Method of Improving the Efficiency of Gas Engines, Thos.
E. Butterfield.
Discussed by A. M. Greene, Jr., W. 0. Barnes.
Offsetting Cylinders in Single-Acting Engines, Prof. T. M.
Phetteplace.
Discussed by W. H. Herschel, J. H. Norris.
Thursday Afternoon
Reception of members by William H. Taft, President of the United
States.
Thursday Evening
Address, The Engineer in the Navy, by Rear-Admiral George W.
Melville, Ret., Past-President Am. Soc. M. E.
Address, Rear-Admiral Melville's Service to the Engineer-
ing Profession and to the Nation, and presentation to the National
Museum of a portrait of Rear-Admiral Melville; by Walter M. McFar-
land of Pittsburg, Pa. Acceptance of the portrait by Dr. C. D. Walcott,
Secretary of the Smithsonian Institution, representing the Nation.
18 SOCIETY AFFAIRS
FOURTH SESSION
Friday Morning, May 7
PROFESSIONAL SESSION
Small Steam Turbines, George A. Orrok.
Discussed by W. D. Forbes, R. H. Rice, Prof. R. C. Carpenter,
H. Y. Haden, F. D. Herbert, W. E. Snyder, W. T. Don-
nelly, F. H. Ball, C. A. Howard. The discussion was
continued at the Boston meeting, June 11.
Oil Well Tests, Edmund M. Ivens.
Discussed by F. A. Halsey, S. A. Moss, J. E. Callan.
Safety-Valve discussion, continued from the ebruary meeting in
New York: F. L. Pryor, E. F. Miller, G. H. Musgrave, A. B. Car-
hart, S. B. Paine, M. W. Sewall, A. C. Ashton, A. F. Nagle, J. J. Aull,
A. J. Hewlings.
Specific Volume of Saturated Steam, Prof. C. H. Peabody.
Discussed by Prof. W. D. Ennis.
Some Properties of Steam, Prof. R. C. H. Heck.
Discussed by S. A. Moss, G. A. Goodenough.
A New Departure in Flexible Staybolts, H. V. Wille.
Discussed by Wm. Elmer, W. E. Hall, Alfred Lovell, F. J. Cole.
Friday Afternoon
Trip by boat to Mt. Vernon.
LOCAL COMMIITEE
Walter A. McFakland, Chairman
Gtjstav Ayrbs Hervey S. Knioht
Albert H. Buckler Walter R. Metz
Chari^s Eli Burgoon George L. Morton
Howard A. Coombs Harold P. Norton
James B. Dillard Willard L. Pollard
William A. E. Doying John E. Powell
Charles E. Foster Alfred H. Raynal
H. A. GiLLis William B. Ridgely
James Hamilton W. E. Schoenborn
Frederick E.Healy George R. Simpson
Herman Hollerith Charles F. Sponsler
J. A. Holmes Lucien N. Sullivan
Arthxir E. Johnson Wiluam B. Upton
Frank B. King Charles V. C. Wheeler
Earl Wheeler
SOCIETY AFFAIRS 19
Committee of the Washington Society of Engineers
W. A. McFarland, Mem. Am. Soc. M. E., Chairman
A. E. Johnson, Mem. Am. Soc. M. E.
A. H. Raynal, Mem. Am. Soc. M. E.
W. E. Schoenborn, Mem. Am. Soc. M. E.
W. B. Upton, Mem. Am. Soc. M. E.
H. W. Fuller, Mem. Am. Inst. E. E.
John C. Hoyt, Mem. Am. Soc. C. E., Secretary Washington Soc. Engrs.
D, S. Carll, Mem. Am. Soc. C. E., President Washington Soc. Engrs.
Chairman of the Ladies' Committee, Mrs. James Loring Lusk
ACCOUNT OF THE MEETING
The Convention opened on Tuesday evening with a reception in
the large assembly hall of the New Willard, followed by dancing,
with music by the Marine Band. The reception was largely attended
and the occasion was a brilliant one. As the guests arrived they were
received by the President and Mrs. Smith, Mrs. W. L. Marshall, Mrs.
Charles D. Walcott, and Mrs. F. H. Newell.
D. S. Carll, President of the Washington Society of Engineers,
called the assembly to order at 9 o'clock, and extended a hearty
welcome to the Society on behalf of its local members and of the
Washington Society of Engineers. He then introduced Hon. Henry
B. F. Macfarland, President of the Board of Commissioners of the
District of Columbia.
Speaking on behalf of these same bodies and of the District of
Columbia, Mr. Macfarland referred especially to the work of engineers
in the city of Washington, and said in part: There is a particularly
warm welcome for the Society in the national capital, since engineers
more than the men of any other profession have made it what it is.
George Washington, in the year of the birth of the Constitution,
conceived the idea of a magnificent capital, then ridiculously out of
proportion to the youth, weakness and poverty of the new nation.
L' Enfant and Ellicott in the beginning, and a long line of able and
brilliant engineers since then, chiefly of the United States Army, have
rendered important service in carrying out his plans. The past nine
years, the great municipal building period of the city, have been
occupied with such engineering feats as the installation of the filtra-
tion plant, the sewage disposal system, the new pumping system,
the District government railway terminal work, the District govern-
ment l)uilding on Pennsylvania Avenue and its approaches, the
Connecticut Avenue bridge, and others of a similar character. Wash-
ington appreciates engineers.
20 SOCIETY AFFAIRS
President Smith in responding for the Society extended the thanks
of the members for this cordial welcome and their appreciation of
the interesting program prepared for their pleasure and entertain-
ment by the committees of the Washington Society of Engineers
and of the local membeiis.
Business Meeting Wednesday Morning, May 5
The report of the tellers of election was received and there being no
objection the President declared the names presented duly elected
to membership in the Society. The Ust follows this report.
Mr. Smith in behalf of the Membership Committee presented the
following proposed amendments:
C 10 An Associate shall be 30 years of age or over. He must have been so
connected with some branch of engineering, or science, or the arts, or indus-
tries, that the Council will consider him qualified to cooperate with engineers
in the advancement of professional knowledge. He need not be an engineer.
The committee recommends the following to be added at the end
of C 11 of the constitution.
A person who is over 30 years of age cannot enter the Society as a Junior.
The report of the Membersliip Committee pubUshed in Transactions,
Vol. 30, p. 550, gave in full the reasons for desiring the change.
The proposed amendments were discussed, and in accordance with
the rules governing the amendments to the constitution, were re-
ferred to the annual meeting for final action.
Prof. Ira H. Woolson, who was a member of the Membership Com-
mittee for five years, heartily commended the proposed change and
hoped it would become a part of the constitution.
Prof. F. R. Hutton proposed an amendment to C 45, adding
"Public Relations Committee" after "House Committee."
In view of the fact that it has been brought to the notice of the
Society that a movement is under consideration to increase and
improve the facilities for the work of the United States Patent Office,
Prof. F. R. Hutton introduced the following resolution:
Resolved, That this Society in convention assembled requests the
Council of the Society to consider the desirabihty of taking some
SOCIETY AFFAIRS 21
action in furtherance of the movement to increase the Patent Office
facilities, and, if deemed advisable, that they request the individual
members to take steps to urge their influence to this end upon their
Senators and Representatives.
The resolution was voted by the meeting.
Professional Session, Wednesday Morning
Four papers were presented at tliis morning session, two of which
related to the conveying of materials. The first was upon A Unique
Belt Conveyor, by Ellis C. Soper, of Detroit, Mich., and described an
installation consistiitg of a conveyor one-quarter mile long, so
located on an incline tliat less power is required to operate it empt}-
than when loaded. Datajupon performance were given. The
second was upon Automatic Feeders for Handling Material in Bulk
by C. Kemble Baldwin, of Chicago, 111. This contained outline
drawings and descriptive matter upon different designs of feeders,
to enable the engineer to select the type best suited to his needs.
The third paper was upon A New Transmission Dynamometer,
by Prof. Wm. H. Kenerson of Providence, R. I. This is made in the
form of a shaft coupling. The apparatus contains an oil chamber,
one side of which is a diaphragm, and is so arranged that pressure
is brought against this diaphragm directly proportional to the amount
of power transmitted. A gage or other registering apparatus is con-
nected with the oil chamber by a small tube which indicates the press-
ure and the water power transmitted.
The last paper was upon Polishing Metals for Examination with
the Microscope by Albert Kingsbury, Pittsburg, Pa., in which he
described the use of a polishing machine carrying discs faced with
common paraffin and charged with wet abrasives. This produces
excellent surfaces on all the harder metals and alloys, but has not
proved serviceable upon the soft metals, such as lead.
Wednesday Evening Lecture
On Wednesday evening Frederick H. Newell, director of the U. S,
Reclamation Service, was expected to lecture on Home Making in
the Arid Regions. As he could not be present a lecture on this sub-
ject was given instead by Arthur P. Davis, Chief Engineer of the
Reclamation Service.
The United States Reclamation Service in its seven years of exist-
ence has undertaken 26 projects situated in 16 different states and
22 SOCIETY AFFAIRS
territories of the West. It has invested in construction about S40,-
000,000. Nineteen projects have been brought to a point where
some land is now under irrigation. Water is ready for delivery to
about half a milUon acres. An average of about 10,000 laborers are
employed on this work, and over 55,000,000 cu. yd. of rock and
earth have been excavated. Over 2000 miles of canals have been
built and 56 tunnels have been bored, which have a total length of
over 13 miles.
Twelve large earthern dams and one high masonry dam have been
completed, and two other masonry dam? which will rank among the
highest dams in the world are in an advanced s^age of construction.
Many of these projects are in remote localities into which roads had
to be built, some of which were carved in precipitous rock, or tun-
neled through mountains. In the aggregate 342 miles of roads
and 793 bridges have been constructed.
In some localities, especially on the Pacific slope, the mild climate,
and the nearly perpetual sunshine, produce remarkable results in
the growth of fruits, which for color, flavor and physical perfection
cannot be equaled in a more humid climate. The chemical force
in sunshine and a perfectly regulated water supply are also evident
in the yields of vegetables and forage crops.
The lecture was illustrated by many beautiful slides.
Gas Power Section
At this session, F. R. Low, Chairman of the Gas Power Section,
presided, and Geo. A. Orrok acted as Secretary. Previous to the
reading of the professional papers reports were received from the
committees.
Membership Committee: The report showed a total member-
ship of 302, of which 177 were members of The American Society
of Mechanical Engineers and 125 were affiliates. The Membership
Committee is thoroughly organized with representatives in differ-
ent cities.
Literature Committee: Prof. C. H. Benjamin gave a verbal
report of this committee stating that the committee is organized
for work and had laid out a tentative program. It was hoped to
index the books on the subject of gas power and articles in periodi-
cals dealing with gas power and aUied subjects; also to present reviews
of new books and abstracts of important articles. There would be
two fields for work: one, a permanent one, and the other in the line of
SOCIETY AFFAIRS 23
current work relating to popular reviews and abstracts for the benefit
of members.
Plant Operations Committee: A verbal report offered by living
E. Moultrop reported progress and stated that standard forms for
obtaining operating data on gas power plants were in preparation.
The committee has a large membership and is widely scattered so
that it had been impossible to arrange a meeting, but the work had
been advanced as far as possible by correspondence.
Mr. C. L. Straub presented a report on gas-producer development
abroad, an abstract of which appears as part of his paper on
Marine Producer Gas Power, included in this volume.
Mr. Orrok stated with reference to the work of committees that it
is conducted with the idea that as the Gas Power Section has been
formed while the art is young it will be possible to place a record of
its development on file at the headquarters of the Society. Such
data in connection with the large library will place at the disposal
of anyone interested in the industry the available information upon
the subject of gas power.
Following the presentation of the reports came the professional
papers, the first of which was on Marine Producer Gas Power, by
C. L. Straub of New York. This paper explained the conditions
opposing the earlier adoption of producer gas power for marine service
and gave a summary of marine gas power plants in operation at
present. It compared the updraft and downdraft of producer gas
apparatus and contained comparative drawings of the steam equip-
ment and two types of producer gas equipment for a 306-ft. boat.
The next paper was upon The Operation of a Small Producer Gas
Power Plant, by C. W. Obert of New York. It presented a general
description of a producer gas power plant in the Westchester market
building of Swift & Company, Bronx Borough, New York. The
author outlined the operating and maintenance system developed
for keeping producers and engines in proper condition for continuous
operation.
A paper was presented upon A Method of Improving the Efficiency
of Gas Engines, by Thomas E. Butterfield of Philadelphia, Pa. It
related to the securing of higher efficiency by reducing the clearance
and increasing the compression and referred especially to a method
of diluting with an inert gas the charge drawn in during the suction
stroke of an Otto cycle engine. By this means premature ignition
and other troubles incident to high compression are avoided.
The last paper of the session was upon Offsetting Cyhnders in
Single-Acting Engines by Prof. T. M. Phetteplace of Providence,
24 SOCIETY AFFAIRS
R. I. It gave the results of an investigation of this subject in which
the various factors entering into tlie problem were taken into account.
Presentation of Portrait of Rear-Admiral Melville
On Thursday evening, a portrait of Rear-Admiral Geo. W. Melville
was presented to the National Gallery at a ceremony held in the audi-
torium of the New Willard. President Smith presided over the large
audience which assembled and an address was made by Rear-Admiral
Melville on The Engineer in the Navy. Mr. Walter M. McFarland
of Pittsburg, Pa., gave an appreciation of Melville as a man and of his
work for the nation and profession. The portrait was accepted for the
Nation by Dr. Chas. D. Walcott, Secretary of the Smithsonian Insti-
tution.
At the conclusion of the ceremony, President Smith asked that
Mr. Sigismond de Ivanowski, the Russian artist who had produced
so admirable a likeness of Melville, be escorted to the platform.
Mr. de Ivanowski told briefly and simply of his attempt to portray
the strong characteristics of his subject and displayed evident
pleasure that his efforts were so warmly appreciated.
Abstracts of the addresses are published with the professional
papers in this volume.
Professional Session, Friday Morning
Five papers, and a continuation of the Safety Valve discussion
given at the February meeting in New York, were scheduled for this
session. The first paper was upon Small Steam Turbines by Geo.
A. Orrok of New York. The various types of turbines now on the
market were illustrated and described and a number of steam con-
sumption curves were given to demonstrate the economy that might
be expected from machines of this type.
A paper on Compressed Air Pumping Systems of Oil Wells by
Edmund M. Ivens of New Orleans, La., was read, in which a descrip-
tion was given of compressed air plants at Evangeline, La., oil fields,
and the results of tests upon these plants with different types of
apparatus.
This was followed by the Safety Valve discussion, continued from
the February meeting in New York.
Two papers were next presented upon the properties of steam: one
by Prof. C. H. Peabody, of Boston, Mass., on Specific Volume of
SOCIETY AFFAIRS 25
Saturated Steam, and the other upon Some Properties of Steam by
Prof. R. C. H. Heck of New Brunswick, N. J. The former reviewed
the results of experiments which might form the basis of a computa-
tion of specific volumes at various temperatures and compared the
computed results with experimental determinations of the same quanti-
ties. The latter paper summarized the important work of Holborn
and Henning and compared the results of other investigators. These
two papers constituted another step ahead in the work that is now
being accomplished toward securing accurate information upon the
properties of both saturated and superheated steam.
The last paper was by H. V. Wille, Philadelphia, Pa., on A New
Departure in Flexible Staybolts. This paper proposed the employ-
ment of tempered spring steel in the manufacture of the stems of
staybolts, the ends being of soft steel so as to permit riveting over in
the boiler.
This session closed with a unanimous resolution extending the thanks
of the Society to those who had afforded such abundant entertain-
ment to their visitors.
Entertainment
During the convention an information bureau was conducted
at the Society headquarters by Chairman Walter A. McFarland of
the Local Committee, where the various excursions were organized.
These included trips not only to the public buildings, but to govern-
ment institutions and other points of technical interest, among which
were the Bureau of Standards, the station of the Potomac Electric
Power Co., the Union Railway Terminal, the Naval Gun Factory,
the District pumping stations, etc.
At the ladies' headquarters in the New Willard, tea was served each
day from four to six o'clock and the visiting ladies as well as many
members of the Society, accepted the hospitality extended by the
ladies at this time. Sight-seeing automobile trips for the ladies were
also arranged on each day, which were largely patronized and greatly
enjoyed.
On Wednesday afternoon alarge number took the trip to Fort Myer
to see the exhibition drill, and the evolutions performed by the
several troops, the unusual skill of both riders and drivers and the
thoroughly trained horses, called forth round after round of applause.
Two battalions of artillery with guns and caissons went through
evolutions of great complexity, and two troops of cavalry through
26 SOCIETY AFFAIRS
various formations, apparently equally difficult, while a troop of bare-
back riders captivated the audience by their horsemanship.
On Thursday afternoon the reception of members and guests by
President Taft in the East room of the White House was very
generally attended.
On Fridaj'- afternoon following the professional session, many went
by boat to Mt. Vernon to visit the estate and home of Washington,
and a wreath from the Society was placed on his grave.
A trip to Fort Myer was also made and the dirigible balloon of
the Signal Corps of the United States Army inspected.
ELECTIONS TO MEMBERSHIP
The following were declared elected to membership in the Society
upon the ballot of May 5, 1909, and their election reported at the
Washington Meeting:
Ahlquist, H., Syracuse, N. Y. Mackenzie, Edmund, Brooklyn, N. Y.
Babbitt, Edward F., Columbus, O. Mayall, E. L., Racine, Wis.
Behrend, Bernard A., Milwaukee, Wis. Morat, J., Yonkers, N. Y.
Billings, A. W. K., Havana, Cuba. Peck, Eugene Colfax, Cleveland, O.
Blaisdell, Benjamin H., Manila, P. I. Plunkett, Charles T., Adams, Mass.
Bloemeke, R. B., New York. Puchta, Edward, Chicago, 111.
Borde, G. U., New Orleans, La. Richardson, L. S., Rosebank, S. I., N. Y.
Bruckner, R. E., New York. Righter, Addison A., Chicago, 111.
Burt, Clayton R., Rockford, 111. Riley, Joseph C, Boston, Mass.
Bushnell, Douglas Stewart, New York. Roberts, Alvin L., Milwaukee, Wis.
Crockard, Frank H., Birmingham, Ala. See, Alonzo B., New York.
Davis, Alfred C, E. Liverpool, O. Shaw, James C, Kobe, Japan.
Duncan , Albert Greene, Boston, Mass. Sheridan, Richard B., Cleveland, O.
Ennis, J. B., Paterson, N. J. Shouvlin, Patrick J., Springfield, O.
Fletcher, E. LesUe, Bridgeport, Conn. Smith, Wm. W., Mexico, D. F., Mexico.
Funk, Nelson E., New York. Stacks, H. Roy, Philadelpliia, Pa.
Garvin, George K., New York. Still, F. R., Detroit, Mich.
Hazelton, W. S., Detroit, Mich. Svensson, J. Alfred, Brooklyn, N. Y.
Hem, H. O., Kansas City, Mo. Thomas, Horace T., E. Lansing, Mich.
Hogue, Oliver DriscoU, Boston, Mass. Thompson, Sanford Eleazer, Newton
Hunter, John A., Pittsburgh, Pa. Highlands, Mass.
Jewett, A. C, Orono, Me. Tiplady, John T., Cleveland, O.
Johnstone, F. W., Mexico City, Mexico.Tobin, R. P., Boston, Mass.
Jones, Walter J., New York. Trefts, John C, Buffalo, N. Y.
Kellogg, Harry F., Chicago, 111. Vail, Jesse A., Beloit, Wis.
Knight, Alfred H., Ann Arbor, Mich. Waters, W. L., Pittsburgh, Pa.
Kranz, William George, Sharon, Pa. Wells, Robert G., Kalimati, India.
McGuire, Charles H., Denver, Colo. Wills, C. Harold, Detroit, Mich.
SOCIETY AFFAIRS
27
PROMOTION TO MEMBER
Bibbins, James R., E. Pittsburgh, Pa.
Brown, J. Rowland, Mansfield, O.
Castanedo, Walter, New Orleans, La.
Chatard, William M., Baltimore, Md.
Cooke, Morris L., Germantown, Pa.
Dale, Orton G., New York.
Grover, Marcus A., Birmingham, Ala.
Hawks, Arthurs., Bethlehem, Pa.
Heisler, F. W., St. Marys, O.
Hunter, James F., New York.
Keith, Thomas Marshall, New York.
Kilgour, Dwight Foster, Boston, Mass.
Lea, Henry I., Chicago, 111.
Pomeroy, L. R., New York.
Robinson, G. P., Albany, N. Y.
Roe, Mark W., Akron, O.
Shiebler, M., New York.
Swan, John J., Cynwyd, Pa.
Whitted, Thomas B., Charlotte, N. C.
ASSOCIATES
Blanchard, A. S., E. Orange, N. J.
Bryce J. Wares, New York.
Carpenter, A. O., Franklin, Pa.
Castle, S. N., London, England.
Clancy, George W. A., Readville, Mass.
Clark, Frank S., Cincinnati, O.
Fuller, Ray W., Scranton, Pa.
Hart, Robert W., Winchester, Mass.
Howell, Frank Scott, New York.
Koenig, Adolph G., New York.
Pellissier, G. E., New York.
Sanguinetti, Philip C, New York.
Shields, George Rex, New York.
Vincent, Arthur S., Brooklyn, N. Y.
Willson, Ernest M., Charles City, la.
PROMOTION TO ASSOCIATE
Brooks, Paul R., Peabody, Mass.
Davoud, V. Y., Provo, Utah.
Dillard, James B., Washington, D. C.
Idell, Percy C, Hoboken, N. J.
Marshall, W. C, New Haven Conn.
Aldrich, Chester S., Boston, Mass.
Baendcr, F. G., Iowa City, la.
Bailey, H. Morrell, Johnstown, Pa.
Bateman, George W., Claremont, N. H
Bleyer, Charles F., Milwaukee, Wis.
Bond, Francis M., Washington, D. C.
Brown, Harry W., Allston, Mass.
Daugherty, Frank, Philadelphia, Pa.
Duncombe, Frederic H., New York.
I-essenden. C. H , Ann Arbor, Mich.
* afford, B. T., Lebanon, Ind.
Hamilton, Chester B., Toronto, Can.
Hildenbrand, Harry, Houston, Texas.
Home, Harold Field, Mohegan, N. Y.
Hudson, Robert A., San Francisco, Cal.
Jenks, Glen Fay, Philadelphia, Pa.
Kessler, Armin G., Ithaca, N. Y.
Lawrence, Gerald Peirce, Columbus, O.
Lawrence, S. E., Galveston, Texas.
Leahy, Frank E., Clairton, Pa.
Lee, Ralph A., Brooklyn, N. Y.
McGlone, R. G., Galveston, Texas.
Mack, George J., Albany, N. Y.
■ Meyer, C. Louis, New York.
Minck, Peter, Town of Union, N. J.
Newcomb, Robert S., New York.
Nicholl, John S., Yokohama, Japan.
Olmsted, George C, Milan, O.
Otto, Henry S., New York.
Phelps, Charles C, New York.
Pinner, Seymour W., Ann Arbor, Mich.
PuUs, W. Eugene, Saylesville, R. I.
Rupp, M. E., Culebra, C. Z., C. A.
Scheel, H. Van Riper, Passaic, N. J.
Searle, Wilbur C, Worcester, Mass.
Shenberger, G. H., Lansford, Pa.
Simpson, William K., New York.
Singer, Sidney C, Syracuse, N. Y.
Slee, Norman S., Barberton, O.
Smith, CD., Pittsburgh, Pa.
28
SOClteTY AFFAIRS
Juniors — Continued
Smith, Edward S., RoUa, Mo.
Stanton, Alden D., Brooklyn, N. Y.
Stanton, F. A. O'C, Hoboken, N. J.
Taylor, J. W., Massillon, O.
Thomas, Fred H., Mt. Vernon, O.
Thompson, Edward C, Boston, Mass.
Thurston, Edward D., Jr., New York.
Tuttle, I. E., Brooklyn, N. Y.
Wegg, David Spencer, Jr., Provo, Utah.
Whitcomb, L. A., Brooklyn, N. Y.
Whiting, R. A., New York.
Wiley, J. M., New York.
Woolley, Harold O., Dansville, N. Y.
Wyman, A. H., Milwaukee, Wis.
By direction of the Council announcement is also made of the elec-
tion of H. K. Hathaway of Philadelphia, Pa., elected on the ballot of July
25, 1907, as an Associate Member, but announcement of whose election has
not previously been made.
No. 1230
THE TRANSMISSION OF POWER BY LEATHER
BELTING
CONCLUSIONS BASED PRINCIPALLY ON THE EXPERIMENTS OF
LEWIS AND BANCROFT
By Caul G. Bauth, Philadelphia, Pa.
Member of the Society
In his paper, Slide Rules in the Machine Shop as a Part of the Taylor
System of Management, read December 1903, the writer referred to an
improved theory and new formulae developed by him for the pulling
power of belting, wliich had been applied in connection with the slide
rules described. He also stated his expectation of presenting his
theory and conclusions to the Society at some future time.
2 These conclusions have since been successfully applied in prac-
tice by the extensive daily use of these slide rules in task-setting for
machine operations, and the present paper was prepared with the
general view of submitting this theory to the Society for the criticism
and consideration of members who are interested in this subject, and
with the special view of supplementing Mr. Taylor's paper, On the
Art of Cutting Metals. All the experimental and mathematical data
for the slide rules were presented in his paper, excepting the data upon
the pulling power of belting, an important element in these slide rules
when applied to belt-driven machines.
3 The theory to be presented is only an additional attempt, and it
is hoped a fairly successful one, to do something along lines repeatedly
touched by various other investigators among the members of the
Society, and the writer is glad to acknowledge his indebtedness to
nearly all of these, as his work has principally consisted in taking
advantage either of carefully conducted experiments recorded by
them, or of suggestions of various kinds that have stimulated his
interest and set him thinking.
Prosonted at the New York monthly meeting (January 1909) of The
t^MEKicAN Society of Mechanical Engineers.
30 TRANSMISSION OF POWER BY LEATHER BELTING
4 Most notable is the paper presented by Mr. Wilfred Lewis at the
Chicago meeting in 1886, in which he recorded a series of experiments
conducted by himself in the spring of 1885, under the direction of
Mr. J. Sellers Bancroft and at the works of William Sellers & Com-
pany, Philadelphia. In his paper was shown for the first time the
fallacy of the then universal and still common assumption, that the
sum of the two tensions in a belt is constant for all loads. It was also
shown that the coefficient of friction between a belt and its pulley is
considerably higher than was commonly assumed for ordinary work-
ing conditions, and that this coefficient varies greatly with the velocity
of sHp, a fact that has also been pointed out by other investigators.
5 Mr. Lewis did not, however, even attempt to develop either
empirical or rational mathematical formulae to represent the facts that
he established, though his experiments, as will subsequently appear,
contained all that was necessary for a complete mathematical exposi-
tion of the subject.
6 An attempt at an empirical mathematical exposition of the
relations between the two tensions in a belt in accordance with the
facts established in Mr. Lewis' paper, was later made by Prof. Wm.
S. Aldrich in a paper read at the New York meeting in 1898. Mr.
Aldrich made an original layout of a great number of Mr. Lewis'
experiments in a manner that seemed to the writer to indicate an
excellent way to investigate the subject, and which resulted in a dis-
cussion of Mr. Aldrich's paper, the substance of which has formed the
basis for all of the writer's subsequent work on the subject. ':
7 But while the writer's complete theory could have been worked
out without it, its practical application to the running of belt-driven
machine tools could never have been made in the present satisfactory
manner without the facts made known by Mr. Taylor in his paper,
Notes on Belting, read at the New York meeting in 1893.
8 In this paper Mr. Taylor showed the economy of running belts
under much lower tensions than those commonly used, and that the
ultimate strength of a belt, or rather of the joint in a belt, does not
form a proper basis for the working tension of a belt, since a belt will
not long retain a tension that is even a small fraction only of its ulti-
mate strength (see Fig. 4). However, Mr. Taylor's facts and figures
were derived from comparatively slow-running belts, and he gave
nothing that could be directly applied to the higher and more eco-
nomical belt speeds. The modification and extension of Mr. Taylor's
ideas to include high speed belts have therefore been part of the
writer's personal work also.
TRANSMISSION OF POWER B\ LEATHER BELTING 31
9 A summary of the writer's work on this subject follows;
a To establish a mathematical formula for the relation be-
tween the tension in a belt and the stretch due to this
tension, based on experiments made at different times by
Mr. Wilfred Lewis, Prof. W. W. Bird and himself. See
Fig. 1, 2 and 3 and Par. 1-18 of the Appendix.
b By means of the knowledge of the elastic properties of
leather belting expressed by this formula to develop a
formula for the relations between the tensions in the two
strands of a belt transmitting power, which formula takes
account of the influence of the sag in a horizontal belt, and
agrees substantially with the results of the experiments
made by Mr. Lewis, when plotted in the manner first done
by Professor Aldrich. See Fig. 6 and Par, 19-38 of the
Appendix.
c To establish a formula to express the relation between the
coefficient of friction between a belt and a cast iron pulley,
and the velocity with which the belt slips or slides over the
pulley, as revealed by plotting the results likewise obtained
by Mr. Lewis. See Fig. 7 and Par. 48 of the Appendix.
d The construction of a diagram embodying the formula
expressing the relation between the two tensions in a
belt, the well known formula for the loss in effective ten-
sion due to centrifugal force and the likewise well known
formula for the ratio between the effective parts of the
two tensions, as determined by the coefficient of friction
and the arc of contact of the belt on its pulleys. These
formulae are so correlated on the diagram that problems
dealing with the contained variables may be solved
graphically, while a direct algebraic solution is possible
only for a vertical belt, or what is the same thing, by
neglecting the effect of the sag of a horizontal belt.
See Plate 1 and Par. 11-24.
e Also, by means of the better knowledge gained of the elastic
properties of leather belting, to develop a formula for the
creep of a belt on its pulleys due to the difference in the
tensions in the two strands, along the lines outlined by
Professor Bird in his paper on Belt Creep, read at the
Scranton meeting in 1 905 . See Par. 41-44 of the Appendix.
32 TRANSMISSION OF POWER BY LEATHER BELTING
/ The construction of diagrams showing the pulling-power
and other relations of the two tensions of a belt of 1 sq.
in, cross section and 180 deg. arc of contact at different
speeds, under certain conditions and assumptions recom-
mended by the writer. See Fig. 1, 2 and 3, and Par. 38-52.
Also a modification of these diagrams for extended prac-
tical use, on which may be read off: (1) The pulling
power of a belt of any width and thickness and any arc
of contact between 140 and 180 deg.; (2) The initial
tensions below which the belt must not be allowed to fall
in order to confine the slip and the consequent loss of
efficiency of transmission within certain limits; (3) The
initial tension to which it is recommended that the belt be
re-tightened after falling to this minimum limit. See
Plate 2 and Par. 53-66.
g Finally, the construction of a slide rule serving the same
purpose as the diagram just mentioned, but which is
much handier than the diagram. See Fig. 5.
10 With these statements the explanation of the diagram Plate
1 will now be taken up.
DESCRIPTION AND USE OF THE DIAGRAM PLATE 1
1 1 Taking the extreme left-hand bottom corner point as the origin
distances along the bottom line represent the variable tension in the
tight strand or side of a belt in te'rms of the initial tension, while
vertical distances measured to any of the bottom group of curves in
the middle field of the diagram represent the corresponding ten-
sion in the slack side of the belt, also in terms of the initial tension.
12 The particular curve against which to read off a certain ten-
sion depends on the center distance of the pulleys of the belt in con-
nection with its initial tension per square inch, and is found by con-
sulting the small diagram directly to the right of this group of curves,
in the following manner:
13 Read off the center distance c along the extreme right side of
this diagram, then follow along the diagonal to the left from this
reading of c until it intersects the vertical line that extends up from
the reading of the initial tension t^ on the base of the diagram.
14 From this point of intersection of c and t^ go horizontally
to the left to the reading of the corresponding value of the ratio
TRANSMISSION OF POWER BY LEATHER BELTING 33
—275 , which leads directly to the proper curve in the bottom group
*•
of curves in the middle section of the diagram.
15 Against this curve there can now be read off any simultaneous
tensions in the two strands of the belt corresponding to these par-
ticular values of c and ^o of the belt under consideration.
16 Having noted this curve, and turning to the extreme right hand
section of the diagram, the ratio of the effective tensions in the two
sides of the belt corresponding to the particular coefficient of friction
4>, and the particular arc of contact a which we wish to count on,
may be determined.
17 To this end we read off the arc of contact in degrees on the
extreme right-hand side of this section of the diagram, follow this
reading horizontally to the left until it intersects that Une radiating
from the bottom left corner of this section which is marked with the
value assumed for 9S at its termination in the extreme top line of the
section, and then from this point of intersection go vertically up or
down as the case may be, until we meet the single curve drawn in this
section of the diagram. From this point in the curve we now go
horizontally to the extreme left side of this section of the diagram
and there read off the value of the ratio of the effective tensions,
which is
t, 1
for a belt running so slowly that the centrifugal force has no percept-
ible influence, and equal to
L -L 1
when the centrifugal force reduces the total tensions to the effective
tensions fj ~ ^c ^^^ h ~ h respectively.
1
18 From the point representing we now draw a line to the
extreme left bottom corner point of the whole diagram.
19 Any two simultaneous coordinates to this slant line counted
from its extremity in the base Une of the dl&gri^in, will then also be
1
in the ratio
34 TRANSMISSION OF POWER BY LEATHER BELTING
20 Passing now to the extreme slanting left side of the diagram,
we read off the velocity V of the belt, follow this reading horizontally
to the right until we intersect that radiating line from the extreme left
bottom corner point of the diagram which is marked with the initial
tension t^ of the belt per square inch, where it terminates, either in
the extreme top line of this section of the diagram, or against its
extreme right side.
21 From this point of intersection we now go down vertically until
we reach the long 45-deg. diagonal of the diagram, on which we then
read off the ratio — , or the loss in effective tension in terms of the
initial tension, due to the centrifugal force in the belt.
22 From this point on the long 45-deg. diagonal of the diagram
we now finally draw a line parallel to the line previously drawn to
1
represent the ratio - ^ ^ and extend it to intersect the curve first of all
e
determined to represent all possible relations between the two ten-
sions in the belt.
23 The coordinates of this point then give the two tensions in the
belt in terms of its initial tension. By extending the ordinate up to
intersect that curve in the middle group of curves which is marked
with the same value of 25 ^^ ^ts terminal in the right side of this
middle section of the diagram as the curve dealt with in the bottom
group of curves, we read off the difference of the two tensions; that is,
the effective pull of the belt, in terms of the initial tension.
24 By likewise extending the ordinate all the way up to meet that
curve in the top group of curves which is marked the same as the other
two curves, we may also read off the sum of the two tensions, in terms
of the initial tension.
25 Example. A belt of the pulley center-distance c = 200 in.
and of 2^ sq. in. cross section, has an initial tension T^, = 175 lb.
and runs at a velocity V = 2000 ft. per minute. The arc of contact
of the belt on each pulley may be taken as 180 deg., and the coefficient
of friction <f) as 0.5. What will be the centrifugal tension T^ = 2.5
t^ in the belt, what the tension T^ = 2.5 t^ in the tight side, and what
the tension T^ = 2.5 t^ in the slack side? Also, what will be the effec-
tive pull P = T^ - T2 and what the sum T^ + T^ of the two tensions?
26 Solution. The steps have been indicated on the diagram
by little circles around the points on the several scales of variables
r—
FOLDEK No. 1
TKANSACTIONS THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS VOLUME HI
TRANSMISSION OK POWER BY LEATHER BELTIXG
0^<, .1 .2 J .4 05 £ .7 .8 3 1.0 .1 .2 .3 A L5 .6 .7 .8 .9 2.0 .1 .2 .3 A 25 £ .7 .8 .9 30 " ,^^J°J'° c , +
( TENSION IN TIGHT SIDE or BELT ,n TERMS or INITIAL TENSI0N-8,-^"l INITIAL TENSION PtRSo.l.-t.
^-LOSS IN TENSION due TO CENTRIFUGAL FORCE ^ ^
IN TERMS OF INITIAL TENSION-^
PJ'Atk 1 Di.iORAM FOR THu GRAPHICAL Solution of Formhi.ak for Horizontal Belts
EXPONENT yot
TRANSMISSION OF POWER BY LEATHER BELTING 35
that correspond to the values of these variables in the example.
Thus, in the small diagram at the bottom of the chart, between the
other two diagrams, circles have been drawn around the points indi-
cating the unit initial tension
t, - '-'' = 70
2.5
and the center distance 200, giving the resulting value
c2 200^ , . , ,
1, approximately.
l2-5 702-5
"0
which determines the particular curve in each of the three groups ol
curves in the middle section of the diagram which apply to the belt
under consideration.
27 Also, in the right-hand section of the diagram circles have been
drawn indicating the coefficient of friction ^ = 0.5, and the arc of con-
tact 180 deg., giving the resulting value of the ratio between the effec-
1
tive tensions -^-^ = 0.208.
1
28 From the point ^^ = 0.208 has also been drawn a line to the
extreme left bottom corner of the whole diagram, the ratio of any
two coordinates to this line thus being 0.208 also.
29 Again, in the left-hand section of the diagram circles are drawn
about points indicating a velocity V = 2000 ft. and the initial ten-
sion t^ = 70, leading to a resultant value of
t T
'« = ^« = 0.202
to T,
This means that 0.202 X 175 = 35.35 of the total 175 lb. of initial
tension in the belt, is made ineffective by the centrifugal force due to
V = 2000 ft.
30 From the point — = 0.202 has also been drawn a line parallel
to the line expressing the ratio
\ = 0.208
36 TRANSMISSION OF POWER BY LEATHER BELTING
SO that the inclination of this line expresses the same ratio between
the effective belt tensions; or
r, - r. ^'-^_ 0.208
31 The point of intersection of this line with the curve previously
found to express all possible relations between the working tensions
in the two strands, has also been encircled, and a vertical line has
been drawn through this point upwards until it intersects that curve
in the top group of curves which is marked 2.5 = 1> ^'^'^ which
accordingly expresses all possible values of the sum of the two ten-
sions in the belt under consideration.
32 The intersection of this vertical line with that curve in the
middle group which is likewise marked —275 = 1, and which accord-
ingly expresses all possible values of the difference of the two tensions,
has also been encircled.
33 Taking the readings of the point encircled in the bottom curve,
we find
d^ = ^A = 1.81, and d^ = ^-^ = 0.535
K T,
We therefore get
T, = 1.81 X 175 = 316.75 lb. and
T^ = 0.535 X 175 = 93.63 lb.
34 From this we again get
P = T,-T^ = 316.75 - 93.63 = 223.12 lb.
as the effective pull of the belt. Also, T^ + T^ = 316.75 + 93.63
■= 410.38 lb. as the sum of the tensions, as against 175 X 2 = 350
lb., the initial sum.
35 But usually we would not be interested in the separate values
of the tensions, and then we would read off directly by the encircled
point in the middle group of curves,
P^L = ^l^I^ = 1.276
f T T
'ft in i n
TRANSMISSION OF POWER BY LEATHER BELTING 37
which gives P = 1.275 X 175 = 223.12 lb., the same answer as
above.
:^S6 If also interested in the sum of the tensions, we would read this
off directly by the point encircled in the top group of curves,
^.V+ ^2 = ^i-L^2 =,2.345
which gives Ti + T, = 2.345 X 175= 410. 38 lb., the same answer
as above.
37 The solution of problems involving long horizontal belts is thus
readily enough effected by means of this diagram, but a little considera-
tion will also make it evident that the difference in results obtained by
taking the length of a belt into account and by neglecting the same
is but slight, except for greater lengths and lower initial tensions
than are usually employed in practice. Ordinarily, therefore, we
would use merely the very bottom curves in the bottom and top
groups, and the very top curve in the middle group of curves in the
middle section of the diagram, which curves are all marked
^^ = 0
/ 2.6
and thus make no difference between horizontal and vertical belts
except for exceedingly long belts.
DESCRIPTION OF DIAGRAMS, FIG. 1, 2, and 3
38 We will now take up the consideration of diagrams Fig. 1, 2,
and 3, which form] the basis of the large diagram Plate 2. These
diagrams are worked out theoretically for vertical belts only, but may
be applied without hesitation to horizontal belts of the lengths usually
met with in practice.
39 In his paper Notes on Belting, Mr. Taylor showed, as already
mentioned, the economy of running belts under much lower tensions
than those commonly figured on in proportioning a belt to do a cer-
tain amount of work.
40 He also divided the belts with which he dealt into two classes :
those whose dimensions he could not well increase over what he found
in use, such, for instance, as the cone belts on lathes and other machines
provided with a cone pulley in a limited space; and those he could
38
TRANSMISSION OP TOWER BY LEATHER BELTING
20
300
90
80
70
60
250
40
30
20
10
200
QQ
u- 90
60
to
I
CO
to 150
o
o 30
20
100
-2L 90
CD „
Q- 80
70
60
50
40
30
20
y
^
^
y^
^
^
k
[^
\^"^
>vi .
\^^
""
V
s\J^^JL
^
TENSION
1 1
IN TIGHT SIDE +
III.
iTENSION IN SLACK SIDE=1
„*tt
=240
— ^
■^
_j-f
-^
^'
'zpiOfi/ ...
r
"^
\.,
/
f^
f:s
'H
^ ' ' '
-'
f
^'
s^
^^-^
JV- -
.-'
■V
k
t-m
\,
\
V
^
\
N
^
/
^'
i^''.
0,,
\
— -
.^
/
/
%,
fe
\
\
-^
//
f
i
^*U
^^
K
/
y
h-
INI
HALT
tNS^
^ ^
\
\
//
/
/
^^-^
/ ,'
\
\
\
/
/
44
>^
\
y,
/
X
X
/
K,
Y
,
N<
/
\
A
V
'>^
.^^
y
\
/
V
\
\
US^
^^%
r^
^
/
\
\
\
X
^ "ft
»
/
\,
V\
ktc
A
y
\
s\
\
,v^,'^
^
\
\\
---
-<^^>"
r
^^
\5^
:i:
c<^T
y^
/
y
^
-^
\
\
_^
^^
^^
\
1000
2000
.„„„ 3000 4000 5000 6000
VELOCITY OF BELT iN FEET PER M1NUTE=V
Fig. 1 Diagram showing the Relations of Pulung Power to Tensions
AT All Speeds, for a Machine Belt of 1 sq. in. Section. (See opposite page
AND Par. 9/).
TRANSMISSION OF POWER BY LEATHER BELTING
39
rl60
1000 2000 3000 4000 5000 5557
VELOCITY OF BELT in FEET PER MINUTE=V
Fig. 2 Diagram Showing the Relations of Pulung Power to Tensions
AT All Speeds for a Countershaft Belt of 1 sq. in. Section
140
Fig. 1 is plotted for: Arc of contact a=180 deg., coefficient of friction (J) = 0.54— . - , -
(See Par. 51); and the sum of the tension in the tight side of the belt and one-half the tension
in the slack side, A = 240 lb. for all velocities. (See Par. 44.) The dotted curve marked tm
gives the initial tension which for p, the same as that figured for A = 240 lb., corresponds to
A = 320 lb. This tm, the average value of which is about 185 lb. between practical limits of
velocities, is the maximum initial tension, to which a machine belt is retightened whenever the
tension falls to the minimum initial tension <„.
Fio. 2 is plotted for the same data as Fig. 1, except that here A = 160 lb. for the full-diawn
curves. (See Par. 44.) The dotted curve marked tm gives the initial tension which, fo'- p the
same as that figured for A = 160 lb., corresponds to A = 240 lb. This tm, the average value of
which ia about 142 lb. between practical limits of velocities, is the maximum initial tension to
which a countershaft belt Ls retightened whenever the tension falls to the minimum initial ten-
sion <„.
40 TRANSMISSION OF POWER BY LEATHER BELTING
readily increase by providing larger pulleys, such as the belts leading
from line-shafts to the counter-shafts of machines.
41 For belts in the first class he adopted higher tensions than for
those in the second class. He also devised a set of belt-clamps pro-
vided with spring balances, by means of which he could make sure
that a belt was put up under a specified initial tension, and could also
ascertain its fall in tension at any time desired.
42 All this seems to be the first effort made by an engineer to pay
any systematic attention to the belting in a shop, which even today
is usually left entirely to the rule-of-thumb method of the machinist or
millwright.
43 The reason why Mr. Taylor had adopted, and accordingly
recommended, lower belt-tensions than usually counted on in propor-
tioning a belt to do a certain amount of work, was that a belt quickly
loses its tension if it exceeds a certain amount, and thus in order to
maintain such a tension, approximately, requires frequent retighten-
ing, which is a source of too much expense and leads to a rapid destruc-
tion of the belt. See Fig. 4 and Par. 8.
44 Taking Mr. Taylor's data as a starting point, the writer has
finally adopted the rule, as a basis for his use of belts on belt-driven
machines, that for the driving belt of a machine the minimum ini-
tial tension must be such that when the belt is doing the maximum
amount of work intended, the sum of the tension on the tight side of the
belt and one-half the tension on the slack side will equal 240 lb. per
square inch of cross-section for all belt speeds; and that for a belt driving
a countershaft, or any other belt inconvenient to get at for retighten-
ing or more readily made of liberal dimensions, this sum will equal
160 lb.
45 Further, the maximum initial tension, that is, the initial ten-
sion under which a belt is to be put up in the first place, and to which
it is to be retightened as often as it drops to the minimum, must be
such that the sum defined above is 320 lb. for a machine belt, and S40
lb. for a counter-shaft belt or a belt similarly circumstanced.
46 The reason for adopting a uniform sum of the tension in
the tight side and one-half the tension in the slackside, as mentioned
above, instead of either a uniform initial tension, or a uniform maxi-
mum tension alone, is that the aim has been to get as uniform periods
as possible for the retightening of belts at all speeds.
47 But evidently, while the maximum tension in a belt must be
the greatest factor in determining the rapidity with which the belt
will lose its tension as a whole, the accompanying tension in the slack
TRANSMISSION OF POWER BY LEATHER BELTING 41
strand or side must also have some influence, though not proportion-
ally to the same extent; and hence, the idea occurred to the writer of
taking it into account in the manner and to the extent stated.
48 On the diagram Fig. 1, various formulae have been plotted
for 240 lb. as the constant sum at all speeds of the tight tension and
one-half the slack tension per square inch cross section of belt; for a
coefficient of friction that varies with the velocity according to For-
mula 13 in the Appendix; and an arc of contact of 180 deg. The rela-
tions of the various tensions in the belt for all speeds may there be
studied to great advantage. It will thus be seen that the centrifugal
tension completely balances the initial tension at a belt speed of
6805 ft. per minute.
49 On the diagram in Fig. 2 some of these formulae have likewise
been plotted, with the lesser value of 160 lb. as the constant sum of
the tight tension and one-half the slack tension, but for the same
values of the coefficient of friction and the arc of contact. Here the
centrifugal tension balances the initial tension at the speed of 5557
ft. per minute.
50 The diagram, Fig. 1, represents the writer's practice in connec-
tion with machine belts; that in Fig. 2 his practice in connection
with counter-shaft belts (see Par. 44). Both diagrams were used
as the basis for the construction of Plate 2, and for the slide rule illus-
trated in Fig. 5.
51 In the diagram Fig. 3 are given the horse power outputs per
square inch of section of belts running under the conditions imposed
in the diagrams Fig. 1 and 2.
52 It will be seen that the maximum output is 13.8 h.p. per square
inch of a belt under the conditions imposed in Fig. 1, and that this
is for a speed of about 4000 ft. (more exactly 4032 ft.); and that for
the conditions imposed in Fig. 2, the maximum horse power is 7.46 per
square inch, and that this is for a speed of about 3250 ft. (more exactly
3247 ft.).
DESCRIPTION OF DIAGRAM PLATE 2
53 In the diagram Plate 2, the data given on the diagrams
Fig. 1, 2, and 3, for a belt of one square inch of section, and an arc
of contact of 180 deg., have been so modified that almost any problem
relating to belting of any size and any arc of contact can be solved.
54 This will best be illustrated by the following two examples:
55 Example 1. The maximum cone step on the counter-shaft of
42
TRANSMISSION OF POWER BY LEATHER BELTING
a lathe is 22 in. in diameter and wide enough to carry a 3 in. double
belt. The speed of the shaft is to be 300 r.p.m. Assuming the thick-
ness of a 3 in. double belt to be fV in., and the arc of contact of the
belt to be 170 deg.: (a) What pull can the belt be] counted on to
exert, and what horse power will it transmit with ^this pull? (6)
Under what initial tension will the belt first be put up, and retightened
MACHINE
BELT
t+it|=240
COUNTER
SHAFT BELT
t+5tfl60
0 1000 2000 3000 4000
VELOCITY OF BELT in FEET PER MINUTE=V
Fig. 3 Horsepower Output Corresponding to Belt Pulls in Fig. 1 and 2
from time to time? (c) And what minimum initial tension must it
not be allowed to fall below to insure the above-determined pull
without undue slip?
56 Solution. To get the answer to question (a), we first turn to
the small bottom portion of the diagram Plate 2, and on its right hand
side note the point reading 300 r.p.m. From this we pass horizon-
tally to the left until we intersect the vertical line from the point
TRANSMISSION OP POWER BY LEATHER BELTING
43
reading 22 in. on the scale of pulley diameters at the bottom line of
the diagram. From the point of intersection we follow the diagonal
KO
ISO
140
130
\
i
120
\\
■•Vs
110
1 '\
\
1
>>
1
\
■^
in
100
^^ \
<-*J
A!'.
'';— .
\
'■
^ 90
■ ^ —
^
. 1 \
T
rm
-1-II
5 80
Q.
g 70
o
z
2 50
to
LU
•-40
30
20
10
Z 9 16 23 30
NOVEMBER
1900
7 14 21
DECEMBER
II 18
JANUARY
8 15 22
FEBRUARY
I 8 15 32
MARCH
1901
Fig. 4 Experiments Made on the Fall in Tension in Two Belts 5f in. Wide
BY ^i in.Thick Driving a Large Rotary Planer at the
Works of the Bethlehem Steel Company
(See Par, 8 and Par. 4.3)
The peculiarly high tensions measured on four days, during the latter part of
February 1901, were probably due to something sticking in the belt scales used.
line upwards to the bottom line of the main portion of the diagram -
and there read the velocity of the belt to be about 1700 ft. per min.
44
TRANSMISSION OF POWER BY LEATHER BELTING
H
W
H
a
o
o
02
O pM
B
o
>
o
a
a
«
o
w
o
02
Wur.i:i. .1 \.i\i: v; ■ .^ >"/•
FOLDER No. 2
.Si
TRANSACTIONS THE AMERICAN SOCIETY OK MECHANICAL ENGINEERS VOLUME ;il
TRANSMISSION 01' I'OVVKR fiY LEATHER BELTING
VELOCITY foR INITIAL
TENSION IN BELT
MAXIMUM. MINIMUM.
DIAMETER OF PULLEV
Plate 2 Geneual Belting Diagram iNconponATiNQ the Author's Pkactice
TRANSMISSION OF POWER BY LEATHER BELTING 45
57 We now note the point that corresponds to this belt speed of
1700 ft. per min. in that scale on this same bottom line of the main
diagram which is marked " Velocity for Pull of Machine Belt" and
interpolate a vertical line extending upwards from this point. Then
leaving this for the time being, we turn to the extreme left hand
portion of the diagram and there note the point corresponding
to the belt thickness ^ in. on the vertical scale to the extreme left,
and also the point marked 3 in, on the scale of belt widths at the
bottom of the diagram. Following these two points, respectively,
horizontally to the right and vertically upwards, until they intersect,
we now follow the diagonal from this point of intersection until it
terminates against the vertical line marked 170 deg. at the top of the
diagram, in the field marked "Arc of Contact," and then continue
horizontally until we intersect the interpolated vertical line for the
belt speed 1700 ft. already noted.
58 From the point of intersection we follow the diagonal until we
meet the vertical scale of pounds, on which we now read the belt pull
to be 140 lb.; and continuing horizontally until we meet the vertical
line extending upwards from the point corresponding to the belt speed
originally found on the scale of belt speeds in this section of the dia-
gram, and from this line diagonally to the vertical scale of horse power,
we read ofT the horse power transmitted to be 7.2. All of these move-
ments are indicated on the diagram by little circles around the vari-
ous points of intersection.
59 To get the answer to question (6), we proceed exactly as before,
with the width and thickness of the belt, except that we follow the
diagonal across the portion of the diagram headed " Arc of Contact"
until we meet the border line for 180 deg. From here on we proceed
horizontally until we meet the vertical line that corresponds to the
belt speed 1700 ft. in the field marked "Velocity for Maximum
Initial Tension for Machine Belt." From the point of intersection on
this vertical line we then pass diagonally to the scale of pounds, and
there read the maximum initial tension to be 168 lb. Those move-
ments for this solution on the diagram that differ from those for the
answer to question (a), are indicated by little filled-in circles around
the various points of intersection noted.
60 For the answer to question (c), we proceed in every respect as
we did for question (b), except that we of course start from the
point corresponding to the belt speed 1700 ft. in that field of the scale
on the top line of the diagram which is marked " Velocity for Mini-
mum Initial Tension for Machine Belt." The answer read off on
46 TRANSMISSION OF POWER BY LEATHER BELTING
the vertical scale of pounds is 113 lb. The movements for this
solution on the diagram that differ from those for the answer to ques-
tion (6), are indicated by little dotted circles around the points of
intersection.
61 Example 2. The counter-shaft in Example 1 is to be driven
by a belt to run at a speed of 3000 ft. per min. (a) What diameter
of pulley is required to give this belt speed? (6) What pull must the
belt transmit? (c) What width of double belt must be used? {d)
And what will be the initial tension under which the belt must be put
up, and to which it must be again retightened after falling to the
minimum? (e) What will be its minimum tension?
62 Solution, (a) As the counter-shaft is to make 300 r.p.m. and
the belt is to run at 3000 ft. per min., we turn to the small diagram
at the right hand bottom corner of the main diagram, proceed hori-
zontally to the left from the point marked 300 on the scale of revolu-
tions on the right, until we meet the diagonal line from the point
marked 3000 on the horizontal scale of velocities. From the point
of intersection we then go vertically down to the scale of pulley
diameters, and there read off 38 in. as the nearest even diameter.
63 (6) To get the pull of the belt we remember that the cone
belt was found in Example 1 to transmit 7.2 h.p. We therefore note
that point on the vertical scale of horse powers at the extreme right
of the main diagram which corresponds to 7.2, and then follow the diag-
onal from this point towards the left, until we meet the vertical
line extending up from the point marked 3000 on the scale of velocities
on the bottom line of this portion of the diagram. From this point
of intersection we continue horizontally to the left to the vertical
scale of pounds, on which we then read off the pull 80 lb.
64 (c) From the point corresponding to this 80 lb. we now
continue diagonally to the left until we meet the vertical line extend-
ing up from the point corresponding to the belt speed 3000 on the
scale marked '' Velocity for Pull of Counter-Shaft Belt" at the bottom
of this central portion of the main diagram. From this point we
continue horizontally to the vertical line corresponding to the arc of
contact, which, not being given, we will assume as 160 deg., and then
again diagonally in the extreme left hand section of the diagram.
Any simultaneous readings of width and thickness from points in the
diagonal along which we are now moving, will then give a proper belt,
and assuming as in Example 1 a thickness of ^ in., we find the width
to be 3^ in.
65 {d) To find maximum initial tension for this belt, we proceed
TRANSMISSION OF POWER BY LEATHER BELTING
47
Ph S
48 TRANSMISSION OF POWER BY LEATHER BELTING
exactly as in Example 1, except that we use the scale marked " Veloc-
ity for Maximum Initial Tension for Counter-Shaft Belt" at the top
of the middle section of the diagram, and then read this off on the
scale of pounds as 157 lb.
66 (e) Similarly, we find the minimum initial tension to be 97.5 lb.
MEANS OF SECURING AND MAINTAINING DEFINITE TENSIONS IN BELTS
67 In his paper. Notes on Belting, Mr. Taylor referred to belt-
clamps provided with spring balances for weighing the tension in
a belt. In the case of endless belts these scales are put directly on a
belt in its final position over its pulleys, while in the case of a belt
with wire lacing, this is cut to length under the required tension on
the specially designed belt bench illustrated in Fig. 6. As will be seen,
this bench is provided with a pair of pulleys which can be so adjusted
that a tape-line will measure the same around these pulleys as over
the pulleys on which the belt is to run. A belt cut and laced to give
a certain tension when the bench pulleys have been properly adjusted,
will then be of a length to assume the same tension over its own
pulleys.
68 This indirect way of securing a desired tension in a belt was
first suggested by our fellow member, Mr. Gullow Gulowsen, who also
made the drawings from which the first bench and the first improved
belt scale were made by the Bethlehem Steel Company in the year
1900.
APPENDIX
ELASTIC PROPERTIES OF BELTING
The only experiment recorded by Mr. Lewis to establish the elastic properties
of leath. r belting is the following:
2 " A piece of leather belting 1 sq. in. in section and 92 in. long, was found by
experiment to elongate \ in. when the load was increased from 100 to 150 lb., and
only ^ in. when the load was increased from 450 to 500 lb. The total elongation
from 50 to 500 lb. was 1^ in., but this would vary with the time of suspension,
and the measurements here given were taken as soon as possible after applying
the loads. "
i
•^
H^^ —
_„,*--'
"
94"
' — '
... —
^..--■^
_^^^
i
-^
^-'^^
MS
y
^^
iti9?»)
y
'-+♦
93°
y
y
1
^
/
i
/
/
/
no-
100
200 300 400
TENSION IN POUNDS per SQ.lN.= t
500
Fig. 1 Plot of Experiments on a Piece of Bei/ting 1 sq. in. in Section
AND 92 in. long. Test Made by Wilfred Lewis at the Works of
Wm. Sellers & Co., Philadelphia, in 1885
3 Theae data have been plotted in Fig. 1, in which they are remarkably well
represented by the formula
Lt = 92 ( 1 + ^^
830
in which L^ is the elongated length of the belt, and t the load or tension per square
inch.
The full development of the mathematical formulae of this paper, with some other related
mattpr, is given in an unpublished supplement to the Appendix, which is on file in tke
Library of the Society for the use of any who wish to verify the mathematical work.
50
TRANSMISSION OF POWER BY LEATHER BELTING
4 On the strength of this formula the writer originally established the theorem
that the sum of the square roots of the tensions in a belt is constant for all loads, when
no attention is paid to the weight of the belt.
5 However, he soon realized that it would not be safe to build a theory on a
single experiment of this nature; and hence, in July 1901, while in the employ of
the Bethlehem Steel Company, he undertook a series of similar experiments, and
obtained permission of William Sellers & Company to use their emery testing
machine for that purpose with the assistance of their shop engineer, Mr. Leonard
Backstrom.
G Nine pieces of belting were tested in all. The results upon one of those
pieces are shown in Fig. 2, which is tj'pical of all of them. Similar diagrams
representing the other tests are filed with the unpublished supplement to the
Appendix. In all cases the tests were made as rapidly as the loads could be
adjusted and the extensometer readings taken.l
Fig. 2
100 200
TENSION IN POUNDS per SQ.lN.=t
Plot of Experiments on a Piece of Slightly used Double Belting
3rG IN. Wide by | in. Thick
Test made by the writer at the works of Wm. Sellers & Co., Philadelphia, in 1901.
The Supplement contains eight similar plots.
7 Each piece was several times subjected to a complete cycle of loads between
two extremes. During the first few cycles the belts invariably showed different
results, but always gave, eventually, practically the same readings for a number of
cycles in succession, and these are the readings plotted in the figures.
8 The small filled-in circles represent the readings for an increasing load and
the small open circles those obtained for a decreasing load. It is rather astonish-
ing how much lag is shown by every belt. Unquestionably this has some influ-
ence on the law of change of tension in a belt, from its minimum to its maximum,
along its contact with a pulley. This matter has been given some consideration
in the Suppl ;ment.
9 On account of this lag, apparently it would have been desirable to subject
some of these belts to a series of smaller cycles, each between adjacent limits of
the load. The best the writer could do with the results obtained was to average
TRANSMISSION OF POWER BY LEATHER BELTING
51
the loop of each cycle by means of a parabolic curve, and thus obtain a value for
the constant E for each belt on the supposition that the formula
Lt = L 1 +
E
[11
ia approximat<>ly correct. In the various formulae given, however, L is not the
original 15 in. of length to which the extensometer was originally adjusted for each
belt, but an ideal length only, for the estimation of the relations between the ten-
sion and the stretch for values never approaching close to zero.
10 But the best experiments for ascertaining the relations between the tension
and the stretch in belts are unquestionably those by Prof. W. W. Bird, published
in his paper on Belt Creep, read at the Scranton Meeting in 1905.
1 ! y-
-jjfj-
A^
1 , ■ J ' j
\r~ — h —
1 Pi '
r^ i I '
■fill
, /'' ■i-^^\
\ i '
Lt=29A.3?
1^^)
1
1 \^f'
fs.n .■ ' f*r/:
^w
1 I t
w/ 1
/
/
\J\ !
,
» 1 ' >
Lt=^9j:52(K^)
r- . ■ ,
1 ! !
Ml!
y 1 1
r ' '
/
295'
0 100 200 300
TENSION IN POUNDS per SQ.lN.=t
Fig. 3 Plot of Experiments on a Single 4-in., Endless, Running Belt
Test made by Prof. Wm. W. Bird at the Worcester Polytechnic Institute.
The Supplement contains a similar plot on a single 6 in. laced belt. See Professor
Bird's paper on Belt Creep in Volume 26 of the Transactions.
11 These have been replotted by the writer after making some slight correc-
tions in the lengths given by Professor Bird, allowing for the influence of the
sag in the belts and in the tensions given, by the addition of the estimated cen-
trifugal tension, which was not measured by Professor Bird. The centrifugal
tension was estimated after obtaining from Professor Bird the information that
the belts were run at a speed of about 1000 ft. per minute. One of the diagrams
is shown in Fig. 3 and another has been filed with the Supplement.
12 The plots made by the writer differ further from those in the original paper
in that he laid off the tension in pounds per square inch of section of the belts.
13 It will be noted that the results have been approximated both by a dotted
line representing a special form of the broadly general formula
Lt = L 14-
[2]
52 TRANSMISSION OF POWER BY LEATHER BELTING
and by a full line representing this same formula with the special value } for n;
or in other words, Formula 1. For regularity of results, these experiments are
remarkable.
14 While the dotted curves with their more complicated formulae represent the
experiments more closely, the full curves with their simpler formulae also cover
the results so well that they may be looked upon as an excellent justification for
the assumption previously made by the writer on the strength of the experiments
made by Lewis and himself, namely, that within the limits of ordinary working
tensions of a belt, the difference between the lengths of a belt at different tensions is
proportional to the difference between the square roots of those tensions.
15 This proportion is implied in the general Formula 1, when by L we imply, not
necessarily the slack length of a belt, but an ideal slack length on the basis of
which the formula gives reliable results between ordinary working limits of t.
16 Taking the average of the values of E in all twelve sets of experiments we
get 895. Leaving out two experiments, one with a value of E exceeding 1000 and
another for which E was less than 800, and taking the ten remaining experiments
with values of E between 800 and 1000, we get 890; while if we take the average
of only the two experiments by Professor Bird we get only 861.
17 As will be seen hereafter, the writer has adopted 864 as an average working
value, because this figure, combined with certain other constants, results in the
simple final constant coefficient 0.04'^in the right member of Equation 5. For
an average practical working formula on which to build an improved theory for
the transmission of power by leather belting, we thus have
in which Lt equals the length of a belt under the unit tension t when its slack
length is L.
18 However, it will appear later on, that the value 864 adopted for E has
significance only in the formulae developed for long horizontal belts, as £/ disappears
in these formulae when the weight of the belt is neglected.
LAW OF VARIATION IN THE TWO TENSIONS OF A LONG HORIZONTAL
BELT
19 In developing an expression to represent the law of variation in the two
tensions of a long horizontal belt, the free strands of the belt only are considered,
and then, for the sake of argument, are assumed to be attached to the ends of
two double levers fulcrumed in the middle, as shown in Fig. 4 and 5.
20 That the parts of the belt in contact with the two pulleys remain at prac-
tically constant length independent of any variation in the tensions of the two
strands, and thus have no material influence on this variation, will be shown in the
Supplement.
21 In Fig. 4 the levers are parallel to each other, and the two strands of belt-
ing whose normal slack lengths I are supposed to be equal, must form equal cate-
naries under equal unit tensions <,,, corresponding to the equal initial tensions in
the strands of a belt continuous over its two pulleys.
TRANSMISSION OF POWER BY LEATHER BELTING
53
22 In Fig. 5 the levers have been moved through equal angles in opposite
directions, thus tightening the bottom strand to the unit tension <,, and slackening
the top strand to the unit tension L, with corresponding changes in the respec-
tive catenaries.
23 Under this arrangement the sum of the chords of the catenaries remains
constant under all variations of the tensions, a condition that corresponds to that
of an actual working belt over its two pulleys, which remain at a constant distance
apart under all conditions of tension in the belt
24 In considering the problem, the customary approximations in dealing with
catenaries were made, in connection with Formula 1 for the elastic properties of
leather belting, and then the following general formula developed for the relations
between the unit working tensions <, and t^ in a belt, as dependent on the initial
unit tension t^, the weight W per square inch of cross section of each free strand,
and the elasticity constant E in Formula 1.
W^E I 1 1
[4]
FIG. 5
Fig. 4 and 5 Representation of Change in Tensions and Sags in Strands
OF A Horizontal Belt
25 The average weight of a cubic inch of leather belting being about -^-^ lb.
I
we may write W = -, in which I is the center distance in inches between the two
pulleys of a horizontal drive; and by also assuming 864 as the average value of E,
as already done in Formula 3, we may write more specifically
Vf. + %/<,-= 2 V<„ + 0.04? ( i + ^^^ - ^^^
[5]
54 TRANSMISSION OF POWER BY LEATHER BELTING
26 P'or very short belts the last terms in Equations 4 and 5, which are only
tentatively solvable equations, become very small as compared with the term
^\/tg, and by neglecting it entirely we write for vertical belts as well as for short
horizontal belts, or even approximately for all belts:
V/; + Vi, = 2 VFo [61
This formula may be enunciated as a new theorem of the relations of the ten-
sion in a belt, thus: Under any variation of the effective full of a belt, the sum
of the square roots of the tensions in the two strands remain constant, as against
the old fallacious supposition that the sum of these tensions remains constant.
27 However, without entirely neglecting the last term in Formulae 4 and 5
above, these may be made solvable with respect to <,, by first approximating the
value of this term by the substitution in it alone of an average relation of the
tensions t, and ij- As will appear from a study of the diagram Plate 1 in the
body of the paper, such an average relation of the tensions is
t^
t = -
h
28 Substituting this in the last term of Equation 5 and then solving this with
respect to f, we get
I, =
2v/?,-V<; + 0.04P(;^ + i-|
m
which gives practically identical results with the original Equation 5 for such val-
ues of <o and I as fall within ordinary practice.
29 If we express the tensions ti and t^ in terms of the initial tension t„ by writing
^ , '2
5, =• — and 5, = — , Formula 7 reduces to
to ' t,
2 - V^, + 0.04 ^ 2- / 5^2 ^ ^^ _ 2 j • [8]
in which form it is under certain circumstances more readily applied.
TESTING FORMULA 8 BY THE RESULTS OF LEWIS' EXPERIMENTS
ON HORIZONTAL BELTS
30 In Fig. 6 the two tensions simultaneously obtained by Mr. Lewis in one
series of his experiments have been plotted in terms of the initial tensions of the
belt, the tensions in the tight side of the belt being laid off horizontally and the
tensions in the slack side vertically, in the same manner as is done on the diagram
Plate 1, in the body of the paper. Similar diagrams, representing five additional
series of experiments made by Lewis, are filed with the Supplement to the Appen-
dix.
31 However, as the apparatus used by Mr. Lewis did not measure the cen-
trifugal tension in his belts, and as he had no occasion to calculate values of thig
TRANSMISSION OF I'OWER BY LEATHER BELTING
55
quantity, these have been calculated for the present purpose, and added to the
effective tensions tabulated by Mr. Lewis.
32 Each experiment is represented by one of the small filled-in circles, and
is numbered the same as in the tables from which the experiments were taken from
Mr. Lewis' paper.
33 Unfortunately, but very naturally, the initial tension did not remain con-
stant throughout a set of experiments, and in plotting the tensions it was there-
fore necessary to estimate for each individual experiment where the initial tension
was between the values measured at the beginning and at the end of each set of
experiments. This was done by assuming that the initial tension measured at the
beginning of a set of experiments held good for the first experiment, and that the
initial tension measured at the end of a set of experiments held good for the last
experiment, and that there was an equal drop for each experiment.
34 That the initial tension was not constant during each set of experiments
is the reason why the actual tensions obtained were not plotted, but instead their
ratios 5, and d^ to their respective initial tensions.
35 In each figure. Equation 8 is also given with the center distance of the
pulleys for each particular belt introduced as an approximate value of /.
3 A H S .1 .S 3 20 i .2 :i A 2i £ J J3 S 3jO J .2 J /» 35 4
tension in tight side of belt m terms or initial tension=§=|^=|^
Fig. 6 Plot op Experiments by Wilfred Lewis
Horizontal Double Belt 2i in. wide, -j^ in. thick and 32 ft. long. 20 in. Pulleys.
Average Value of Initial Tension t„ = 70 lb. per sq. in. (The Sup-
plement contains five similar plots.)
36 By the introduction of the value obtained for ^g in each experiment plotted,
a corresponding value was calculated for 5, by the formula mentioned, and this
value also plotted, and then a curve was drawn to cover the points thus calcu-
lated. The points themselves are indicated by the little circles drawn around
them.
37 The close coincidence between the curves representing Formula 8 and the
experimental results, is certainly all that can be desired in the way of an experi-
mental verification of the soundness of Formulae 7 and 8.
38 The other, lower curve drawn on each diagram represents the relation be-
tween the tensions in a belt when the influence of its weight is neglected, as given
by Equation 6, which is also given on each diagram, in the form
§0.5 4. 5^06 =, 2 [9]
56 TRANSMISSION OF POWER BY LEATHER BELTING
BELT CREEP AND ITS INFLUENCE ON THE COEFFICIENT OF FRICTION
BETWEEN A BELT AND ITS PULLEY
39 In the paper on his experiments, Mr. Lewis drew the conclusion that the
friction between a belt and its pulley varies greatly with its velocity of slip, so
that the greater the slip the greater the friction. But as he did not make a sub-
stantial study of the elastic properties of leather, upon which the phenomenon
of belt creep depends, he had no means of distinguishing in his experiments be-
tween the necessary slip due to the creep of the belt, and the amount that was
slip pure and simple of the belt as a whole.
40 In most of his experiments the belt did an amount of work that called for
much greater friction between the belt and its pulley than that corresponding
to the creep of the belt alone, and this resulted in additional or true slip that
produced the friction needed to make the belt exert the pull called for.
41 By means of Formula 3 it was possible also to derive a formula that gives
a good idea of the actual creep of a belt in terms of the tensions in its two strands,
which was the object of Professor Bird's paper on Belt Creep, from which the
experiments plotted in Fig. 3 of his paper were taken.
42 This formula, the mathematical development of which is given in the Sup-
plement, is
2 864 + \/t
in which
V = actual average velocity of the creep of the belt on each of its two
pulleys.
F, = velocity of the tight strand of the belt, which is the same as the cir-
cumferential velocity of the driving pulley,
y, = velocity of the slack strand of the belt, which is the same as the cir-
cumferential velocity of the driven pulley.
43 The total creep of the belt on both pulleys together expressed in per cent
of F, is then
^64 + Vt^
44 It must be borne in mind, however, that Formulae 10 and 11 take account
of creep only, and have nothing to do with any additional slip due to a sliding of
the belt as a whole over its pulleys, though the expression
t, = ^ (F, - F,)
taken by itself always represents the total sum of the average creep of the belt
and the additional sliding of the belt as a whole, over each of its pulleys, when
such additional sliding does take place.
45 Considering the matter in this light Mr. Lewis calculated and tabulated
the velocity of shding from the observed loss in speed between the pulleys in his
various experiments.
46 In Tables 1 and 2 appear some of the data thus tabulated by Mr. Lewis.
However, instead of tabulating merely the effective tensions measured by him,
the centrifugal tensions have here been figured and allowed for, and then the total
TRANSMISSION OF POWER BY LEATHER BELTING 57
tensions thereby obtained subsequently converted into tensions per square inch
of cross-section. A column has also been added giving the percentage of aver-
age slip due to belt creep alone, as figured by Formula 11.
47 It will be seen that in most of the experiments the velocity of sliding greatly
exceeded the average due to the elastic creep alone, and that thus the belt as a
whole slid over the pulleys in addition to the elastic creeping, thus showing that
the friction corresponding to this creep alone was not enough to produce the pull
the belt was called on to perform.
4S The relation between the total average velocity of sliding of the belt on
each pulley, and the corresponding coefficient of friction calculated by Mr. Lewis
by the formula
Ratio of Effective Tensions = c^*^
and also copied in Tables 1 and 2 of this paper, is plotted in the diagram Fig. 7
and in a similar diagram of the Supplement. On these diagrams is also shown
a curve representing the equation
[12]
in which 4> is the coefficient of friction and v the total average sliding velocity of
the belt in feet per minute. These results were obtained from belts that had been
in active service, and tested without the application of any belt dressing.
49 As a somewhat conservative average the curve is seen to cover the results
obtained with the belts in a normal condition in a highly satisfactory manner.
50 The question now arises, What coefficient of friction ought to be assumed
in calculating the pulUng power of a belt at any given speed? In view of the fore-
going it does not seem right to assume an average coefficient for all belt speeds
Nor would it be right to base it on an average total sliding velocity of a belt cor-
responding to a fixed percentage of the belt speed, for even a very low percentage
would mean a very high sliding velocity in the case of a high-speed belt, while
a high percentage would mean only a moderate sliding velocity in the case of a
slow-speed belt, and it would seem that the speed with which a belt slides over
its pulley would principally determine the life of a belt that meets with no accident.
51 After considerable study over the subject, the writer has assumed a vari-
able coefficient of friction expressed by the empirical formula
140 [13]
^ = 0.54 -
^ 500 -H F
in which V is the velocity of the belt in feet per minute
52 Equating Formulae 12 and 13 we get
_ 160-1- 0.88 7
*~ 85-1- 0.03 7
as the velocity of sliding on each pulley in terms of the velocity of the belt itself
53 As the percentage of slip between the circumferential speeds of the two
pulleys of a belt is twice the percentage of the average total velocity of sliding t;
of the belt over each pulley, we may now write
200 V 200 160-!- 0.88 7
X = = . [14]
7 7 85 + 0.03 7
58
TRANSMISSION OF POWER BY LEATHER BELTING
a, = NOIlDlbJ JOiN3l3UJ303
TRANSMISSION OF POWER BY LEATHER BELTING 59
as an expression for the percentage of slip corresponding to the coefficient of
friction expressed by Formula 1.'].
54 In Table 3 are listed simultaneous values of (/> and x as expressed by For-
mulao 13 and 14 and also, for the sake of comparison, for ^ figured by the formula
400
^ 800 + F
V
which is the value of 6 by Formula 12, for r = — , that is. for a uniform slip of
V 0' '200
one per cent at all belt speeds.
55 A study of Fig. 7 and the similar figures in the Supplement will show
that the variation in the coefficient of friction with the initial tension of the belt
is so conflicting as to make it best to leave this out of consideration entirely, and
so adhere to the customary assumption that the coefficient of friction is inde-
pendent of the intensity of the pressure. Therefore, as our whole theory of the
variation of the coefficient with the belt speed rests entirely on the formula
Ratio of Effective Tensions = «"^"
t his will be used unhesitatingly, in spite of the fact that its absolute validity will
be disputed in the Supplement for a number of reasons.
EFFECT OF CENTRIFUGAL FORCE IN A BELT
56 While the effect of the centrifugal force in a fast-running belt seems to have
been fully understood by all who have previously treated the subject before this
Society, it is still but imperfectly understood by many engineers and mechanics;
and has even been treated in a wrong way by certain text-book writers who have
followed the work on belting by the late Professor Ruleaux. It has therefore
seemed desirable, in the Supplement to the Appendix, to go into details on this
part of the subject.
57 Subsequently, the following general formula was developed for the loss in
effective tension in a belt, due to its centrifugal force:
w
ic = V^
300 g
in which
Ic = loss in effective tension per square inch of cross-section of belt.
V = velocity of belt in feet per minute.
w = weight of one cubic inch of belting in pounds.
g = acceleration of gravity in feet per second.
Substituting w = -Jjj, as in Par. 25, and g = 32-J^, we have more specifically
^c = 0.000003454 V^ [15]
which is substantially the same formula as that given in Mr. Nagle's paper, For-
mula for the Horse-Power of Leather Belts, read at the Hartford meeting in 1881 .
60
TRANSMISSION OF POWER BY LEATHER BELTING
FORMULA FOR PULLING POWER OF A HORIZONTAL BELT IN TERMS
OF ITS INITIAL TENSION
58 For a horizontal belt on pulleys of the center distance c and one square
inch of cross-section we now have, by substituting c for I in Formula 7,
t. =
2V,.-V^ + 0.04cM;i + i-i
i, - tc
Ratio of effective tensions = = e'P'^
U - tc
to = 0.000003454 V^
= 0.54 -
140
500 -F
[16]
[17]
[15]
[1.3]
per square inch
of cross-sec-
tion of belt.
in which formulae
c = center-distance of pulleys, in inches.
/, = initial tension.
ti = tension in tight strand or side.
<2 = tension in slack strand or side.
tc= centrifugal tension; or, more correctly, loss in
effective tension due to centrifugal force,
p — effective pull.
« = basis of Naperian system of logarithms, 2.71828.
<i> = coefficient of friction between belt and pulley.
a — the lesser arc of contact of belt on pulleya, in radians = — X arc in
180
degrees.
V = velocity of belt in feet per minute.
59 However, an attempt to combine these five equations algebraically to
obtain an expression of p in terms of t^, V, and a, leads to an equation solvable by
trial only, and for this reason the diagram Plate 1, the use of which was explained
in Par. 11 to 24 in the body of the paper, was constructed to effect the solution
graphically.
60 For the construction of this diagram Equation 5 was used after substitut-
ing c for I as above,
^1
<.
, and d.
as in deriving Formula 8 from 7. It thus became
c'' / 1 1
V5, + \/8,-2 = 0.04-- ( -„ + _ - 2
[18]
TRANSMISSION OF POWER BY LEATHER BELTING 61
and was in that form solved tentatively to obtain a aeries of points in a series of
C"
curves, each representing a certain value of the factor ^ ^. Onthe diagram these
y"'
curves form the bottom field of curves in the middle section.
61 The equation was in each case first solved to obtain an approximate value
only of 0,, m terms of an assumed value of ^2, by resorting to the approximation
J,- for in the right-hand member of the equation, as shown by Formula 8.
"'■' . 1
62 Then by substituting this approximate value of S^ m the term ^ ^ in the
right-hand member of Equation 18 this was again solved for a still closer value
of 5j.
63 For the lesser values of , this closer value of d^ differed mappreciably
from the first approximation, while for the greatest values plotted on the diagram,
the equation was solved twice to get the values actually plotted.
64 However, these greatest values of never occur in the practical use of
belting, and hence the very construction of the diagrams under consideration
proved the validity of the approximations d^^ for — , which is equivalent to the
approximation-^ for— resorted to in Par. 27 in modifying Equation 5 to Equii-
tion 7.
FORMULAE FOR PULLING POWER OF VERTICAL BELTS IN TERMS OF
INITIAL TENSION
65 For a vertical belt the relation between the tensions of a belt is expressed
by the simple Equation 6, and this can readily be combined with the rest of the
equations listed in Par. 58 (in the manner done in the Supplement to the Appen-
dix), which leads to the following formula:
,/e^«+l I 4e^^« 0. 000003454 V \ ,
"-■'(eT^rm-V „,«_„.+ 1 )'' ['»!
66 This formula is simple enough, though a great improvement over the
one derived on the erroneous supposition that the sum of the tensions is constant
for all loads.
67 By means of this formula the pulling-power of a belt can easily be deter-
mined in terms of its initial tension. However, for a uniform unit initial tension
for all speeds, the unit tension in the tight side would vary so much that belts
ruiming at different speeds but tightened to a uniform maximum unit initial ten-
sion, and allowed to run until this had dropped to a uniform minimum tension,
would require re-tightening at greatly different periods.
68 As already pointed out the writer has arrived at the conclusion that the
periods at which belts ruiming at different speeds will have to be re-tightened,
will b'i nearly constant if they are all made to do their work at such initial tensions
as under full load will result in the same sum of the tension in the tight side and
62 TRANSMISSION OF POWER BY LEATHER BELTING
one-half the tension in the slack side of the belt, at the two extremes of the initial
tensions, just before and after retightening.
PULLING POWER OF BELTS IN TERMS OF A CONSTANT SUM OF THE
TIGHT TENSION AND ONE-HALF THE SLACK TENSION, AT ALL SPEEDS
Vertical Belts
69 This condition is expressed by the equation
ii + i to = A = a constant
Combining this with Equation 17 (in a manner shown in the Supplement) we get
(e^a _i) (2A - 0.00001036 V^)
p=. ^ [20]
2e"^« + 1
<. = ^^ r22i
and
_ 4 A -p + V(4A-py-9p^
"12
These formulae are the ones plotted in the diagrams Fig. 1 and 2 in the body
of the paper, for A = 240 and 160 lb., respectively.
70 By a similar treatment (as shown in the Supplement) we are also able to
get an expression for the initial tension in a horizontal belt, which gives results of a
high degree of accuracy. This expression is
[24]
which is evaluated by first determining p by Formula 20, and subsequently f, ,
and <2 by P'ormulae 21 and 22, as for a vertical belt, paragraph 09.
71 However, while this formula is of great theoretical interest, it is hardly of
much practical vakie; as the initial tension determined by it will differ but little
from that determined for a vertical belt by Formula 23, except for belts of extra-
ordinary lengths.
72 One very interesting general conclusion may now be drawn from Formula
24; namely, that while actually doing work two horizontal belts of unequal lengths
may be under precisely the same tensions, but this being the case, when idle the
longer belt will be under a slightly lower initial tension.
73 It appears, however, that the popular notion that horizontal belts drive a
great deal more than vertical belts, is not well founded
TABLE 1
EXPEKIMES'TS BY WlLFRED LeWIS, AT THE WoRKS OF Wm. SeLLEEIS & Co., PhILADELPBIA,
1S85, ON' Single Belt 5i in'. Wide by /t in. Thick and in Ordinary Working Con-
dition Without Belt Dressing. Belt Speed = 800 ft. per Minote. These Experi-
ments ARE Plotted i.n Fig. 7 of the Supplement, which see. See also his Paper,
No. 198, Vol. 2 op Transactions, Table 1
.So .So
>>■£ »
& H
Q 0)
o -a
N - I'
2 "o
I w ~
,^ II
•lira
o.
o o »-
■S '^ II
Cm
^■^ I I I ^ ^ ^
0 0~-~ Ohi3.2o
C art!
•s- £• II :«
u a> a
i 5 c2. V ■"
o
60
61
62
63
65
81.6 lb.
persquareinch
125.33
131.42
142.00
152.41
179.92
58.67
0.5
2.0
0.251
0.41
46:58
0.9
3.6
0.336
0.53
42.00
1.7
6.8
0.407
0.62
35 . 75
3.0
12.0
0.490
0.73
29.92
12.0
48.0
0.610
0.91
66 I '
177.42 1
77.42 t
1
0.5 1
2.0
0.270
0.52
68
198.25
64.92
0.8
3.2
0.365
0.69
69 127.5 1b.
208.77
58.67
1.0
4.0
0.418
0.77
70 persquareinch
219.08
50.75
1.7
6.8
0.472
0.87
71
229.50
46.17
2.6
10.4
0.545
0.95
72
244.08
44.08
3.8
15.2
0.569
1.02
73
256 . 58
39.92
3.5
22.0
0.623
1.10
74
252.42
35.75
8.6
34.4
0.677
1.13
"1 J
283.66
33.67
1
15.2 ]
60.8
0.719
1.25
TABLE 2
Experiments bt Wilfred Lewis, at the Works of Wm. Sellers & Co., Philadelphia,
1885, ON Double Belt 2} in. Wide bt jj in. Thick and in Okdinart Working Con-
dition Without Belt Dressing. Belt Speed = 800 ft. pee Minute. These Experi-
ments ARE Plotted in Fig. 19 of the Supplement, which see. See also his Paper,
No. 198, Vol. 2 of Transactions, Table 2
a o
-a ^
m
II
<o II
"I M m
t3
_ a
c 5
fc. g
§ 2 ^° ^«
p. o,
Ir: 4> ■
•30
= o ■
o 00
105
104.9
47.5
0.3 !
1.2 .
0.263
0.38
106
73.5 lb.
123.4
37.5
0.8
3.2
0.395
0.57
107 ,
persquareinch
146.0
32.6
1.7
6.8
0.511
0.73
108 !
1
1
171.5
29.5
4.3
17.2 1
0.600
0.87
121
124
125
126
127
128
131
133
134
135
283.0 lb.
per square inch
343.5 lb.
per square inch
403.0
450.0
465.0
482.2
497.5
511.3
557.0
589.5
603.0
618.0
175.9
137.5
124.8
113.5
99.2
227.0
187.2
162.4
148.2
134.0
0.7
1.5
2.3
3.7
10.1
0.5
1.1
1.8
2.7
5.1
2.8
6.0
9.2
14.8
40.4
2.0
4.4
10.8
20.4
0.267
0.387
0.424
0.469
0.523
0.261
0 . 3,50
0.414
0.450
0.490
0.77
1.07
1.17
1.28
1.39
0.85
0.99
1.30
1.39
1.49
64
TRANSMISSION OF POWER BY LEATHER BELTING
TABLE 3
Relations Between Coefficient of Fbiction, Velocity of Sliding and Belt Speed
See Par. 54
V = Velocity in
feet per minute
c
00 M
X 00 o
o , o 6
§ ^+ +
II O 00
Velocity of Slid-
ing V =
160 + 0.88F
o
d
+
00
n It.
§ § •«
o Sh us
o d
Velocity of Slid-
ing; at 1 per cent
slip
" =200
II ^
-a- »g +
"o "o (N -f-'* O
a c 'I' »
■%-c 1 '
5H ■- to "=-
gfc, d °
o B
0
8
1.8S
0.260
0.00
0.100
50
9.432
2.36
0.285
0.25
0.129
100
5.636
2.82
0.307
0.50
0.156
200
3.690
3.69
0.340
1.00
0.200
300
3.010
4.51
0.365
1.50
0.236
400
2.640
5.28
0.3S4
2.00
0.267
500
2.400
6.00
0.400
2.50
0.292
600
2.227
6.68
0.413
3.00
0.314
700
2.090
7.32
0.423
-3.50
0.333
800
1.983
7.93
0.432
4.00
0.350
900
1.889
8.50
0.440
4.50
0.365
1000
1.808
9.04
0.446
5.00
0.378
1200
1.675
10.05
0.458
6.00
0.400
1400
1.566
10.96
0.466
7.00
0.418
1600
1.474
11.79
0.473
8.00
0.433
1800
1.394
12.55
0.479
9.00
0.446
2000
1.325
13.25
0.484
10.00
0.457
2500
1.180
14.75
0.493
12.50
0.479
3000
1.067
16.00
0.500
15.00
0.495
3500
0.974
17.05
0.505
17.50
0.507
4000
0.898
17.95
0.509
20.00
0.517
4500
0.832
18.72
0.512
22.50
0.525
5000
0.768
19.40
0.514
25.00
0.531
5500
0.727
20.00
0.517
27.50
0.536
6000
0.684
20.53
0.519
30.00
0.541
6500
0.646
21.00
0.520
32.50
0.545
DISCUSSION
Henry R. Towne. The earliest investigation of this subject
was by General Morin, of the Conservatoire des Arts et Metiers,
who gave, in a volume published, I think, about 1850, the results of
his experiments to determine the coefficient of friction of belts on
pulleys, and algebraic formulae to express the power transmitted under
varying conditions. For many years these formulae were accepted
TRANSMISSION OF POWER BY LEATHER BELTING 65
universally. General Morin's experiments were made under labora-
tory conditions.
2 In 1867 I made a series of experiments to determine, under
conditions approximating those of actual use, the coefficient of
friction and also the tensional strength of commercial belting. These
experiments, and a discussion by the late Robert Briggs on the
mathematical conditions involved in the problem, were pubhshed in
the Journal of the Franklin Institute in 1868. Under the title of
the Briggs and Towne Experiments, the conclusions thus reached
were quoted and accepted for many years, by Professor Ranldne,
Professor Reuleaux, Professor Unwin, and many other technical
writers. A. F. Nagle, in a valuable paper contributed to the Trans-
actions of the Society in 1881 (Vol. 2, p. 91), accepted the results
of the Towne experiments as the basis for his discussion of the
mathematical problems involved.
3 The Transactions for 1886 (Vol. 7) contained two important
contributions to the literature on this subject. One of these is a
paper by Professor Lanza (p. 347), which first prominently calls
attention to the importance of syeed of slip as a factor in the trans-
mission of power by belting. The other is a paper by Wilfred
Lewis (p. 549) giving the results of a long and elaborate series of
experiments in the shops of WiUiam Sellers & Co., and demonstrat-
ing, among other things, that the proposition first enunciated by
General Morin, and accepted unquestioningly by all subsequent
authorities, namely, that the sum of the tensions is constant (T^+T^),
does not hold true in all cases, and is therefore erroneous.
4 The Transactions for 1894 contains another most valuable paper,
by Fred. W. Taylor (p. 204), giving the results of his large expe-
rience covering many years in the use and observation of belting
under the conditions of actual practice. Many new and important
deductions based on the investigations of Mr. Taylor are availed
of by Mr. Barth in the conclusions and recommendations contained
in his paper. One of the most important facts demonstrated by
Mr. Taylor relates to the value of increased thickness of belts, and
the resulting expediency of a larger and more general use of double
belts. He was also the first to demonstrate and set forth clearly
the economic gain to be derived from the scientific care of belting.
5 Finally, Mr. Barth, avaiHng himself, as he has stated, of the
work of his predecessors, especially that of Mr. Lewis and Mr.
Taylor, has completed, for the present at least, the study of this
problem, which has thus extended over some sixty yearSj giving us an
66 TRANSMISSION OF POWER BY LEATHER BELTING
elaborate and apparently a conclusive demonstration of the sound-
ness of the mathematical conclusions finally reached, furnishing work-
ing formulae for practical use, and presenting a most ingenious appli-
cation of the slide rule to the problems involved in the practical use
of leather belting.
6 The Society is to be congratulated on including in its roster of
membership the names of all those since General Morin who have
taken the lead in ascertaining the facts and in determining therefrom
the rules which govern the application of leather belting to industrial
uses.
7 Mr. Earth's system has now been in use for about two years in
the works of the Yale & Towne Mfg. Co., Stamford, Conn., where it has
accomplished a substantial increase in economy and efficiency.
Wilfred Lewis. I am clearly of the opinion that Mr. Barth has
discovered and formulated principles of the greatest practical value
in the solution of the problems of the transmission of power by
leather belting.
2 It is difficult in a paper of this kind to separate the practical
from the theoretical without discarding the most valuable part of
the undertaking. The laborious work done by the author in order to
reach his conclusions, and recorded in the appendix to this paper,
is really the basis of the superstructure reared by him and gives
the reader some idea of the immense amount of patient research and
good sound reasoning employed in building up a complete analysis
of the subject.
3 Mr. Barth is the first, I believe, to analyze the peculiar elastic
properties of leather, and to demonstrate in a convincing way the
effects of these properties in the use of belting under varied con-
ditions. His analysis of the combined effects of elasticity and sag
is very original and ingenious, and even aside from the results obtained
his methods cannot fail to interest investigators in other fields of
research. Difficult and complex problems have been solved by
making certain assumptions and approximations that are quite allow-
able as the means to an end, and it is in these short cuts from the
intricate and unwieldly to the simple and practical that he has dis-
played such remarkable ingenuity. At the same time, for those not
enough interested in every step to care to follow a mass of mathe-
matical formulae, Mr. Barth has presented his conclusions in a form
available for immediate use.
TRANSMISSION OF POWER BY LEATHER BELTING 67
4 Popular impressions, even though well founded, are often
exaggerated beyond reasonable bounds, and while it is true that
horizontal belts of considerable length are preferable in the trans-
mission of power to vertical or shorter ones, it will be a surprise, I
believe, to engineers, that there really is so little advantage in a
long horizontal belt over any length of belt in any position. All
this results from the exposure of the fallacy that the sum of the
tensions is constant, a belief exploded 23 years ago, although the
far-reaching effect of the exposure on the transmission of power by
belting has never before been so clearly expounded.
5 The author's treatment, also in the unpublished supplement
to the appendix, of the effect of variations in pulley diameter upon
the transmission of power, I believe to be absolutely original, and
his conclusion that a belt will slip on a driven pulley before it will
slip on a driver of the same diameter indicates a subtlety of analysis
rarely displayed in our proceedings, and is a fair index of the pains-
taking care with which the whole paper has been written. Although
not perhaps of very great practical importance, as a new discovery,
the analysis might well be included in the appendix to the paper,
rather than in the unpublished supplement to the appendix.
W. D. Hamerstadt. The writer has been somewhat closely
associated -uith work on pulley and belt drives, and recently has had
occasion to compare the results of some carefully conducted experi-
ments with the results which might be expected from the use of
formulae as proposed in Mr. Earth's paper. Considering the many
variable factors, these comparisons are remarkably favorable, and
for average conditions of operation, the relationships which have been
established would appear to hold quite true.
2 One almost vital point of consideration in the actual design
of belt drives seems to have been touched upon but lightly, however,
and then in a manner which, as the author himself has stated, leaves
some room for discussion — namely, values of the coefficient of fric-
tion to be used in the formulae given, under varying conditions of
service. While the value of the coefficient of friction will not affect
the theory of belt transmission as given, it will seriously affect the
size of drive required to do a given work, and having now a good
theoretical basis for work, and being assisted by the observations of
others, additional experimental work might well be done for the deter-
mination of such values, using as nearly as possible good average
leather belting and operating under actual conditions of service.
68
TRANSMISSION OF POWER BY LEATHER BELTING
3 Based partly on the conclusions of Professor Lanza or of Mr.
Wilfred Lewis, as given in early papers before the Society, and partly
on the very mechanical reasonableness of the thing as he puts it,
Mr. Barth assumes that, given a belt and pulley, the value of the
coefficient of friction to be used in any case will be determined to a
0 2 4 6 8 10 12 14 16 IS 20 24 28 3Z 36 40
Fig. 1 Relation Between Coefficient of Fkiction and Velocity of Slip
.0
.9
— EFf
ECTl
\Z RAl
tGE —
s^
=a
ss
.U
.7
6
.5
.4
.3
2
^
r^
S^
■■=■
=^
^
"S^
fc=
-^
^
^
^
>^
n
^
y^
^^
::::;
F-«
==
=
^
^
^
A
^
^
'
y
Fig
0 1^ 2
2 Relation Between Coefficient
3 4 5
OF Friction and Percentage of Slip
great extent by the velocity with which the belt sUdes on its pulley.
Taking then a curve representing average relations between these
two factors for any convenient speed of belt, values are at once avail-
able for the coefficient of friction for any speed of belt and any condi-
tion of sHp desired.
4 That there exists some ground for such reasoning cannot be
TRANSMISSION OF POWER BY LEATHER BELTING
69
1
"' c
V
c
c
0 o
>
01
el
4)
•a
^
=5 a,
"S
c
a
«
S
E
'S
j
0) Q^
a
o
O
2
1 (0
s:
^ 0
2
>
0)
e
(0
"a
e
o
t
T3 4)
3
c
o
42
H
c •«
C3
<
c
0
a
r3
c9
i tf
:i
.£• '•
6^
1
1
^ m
§ 1"
v
3
2
1
_ 3
+
O C3
2
n
Ex
H
a 1
o
O
o
•*
S
><
.2 o
N
C4
«
C4
U
Q ^
1 C}
1 a,
1-J
'^
ri
-*'
1
3
d
o
o
O
d
d
•o
c
c
c
c
T3
e!
•a
v
d
oi
c
o
C
ei
0
0
C
1
§ 1
01
-a
3
;
O
- <1>
T
c
'O o -g
"b
>•
5
2;
c
C
a
Q
Q
-ts
H
;
-C
-15
"C
P9
^1
K
><
X
X
■*
ir
cs
lO
M
"O
•2
<D
«>
c
_a
3
V
1
■a
d
"3)
A a
Ji 3
"t
1
3
^
1
?
h]
Is o
c
o
<d
o
O Q
a
Q
o
c
»
o
o
«
o
o
H H .
c
•^
t^
J t- Z
o
c
o c
(N
c^
■^ If
00 ^
i
o
B.
Ii4
o
1"
CO
c-^
<N
&
~f'
&
m
o
1
m
H
H
■*
(N
: H ^
0.
o
_2
01
: t-' J
d
"3
3
3
6
o
03
03
,*
• -t; c
.
i^
o
»-<
S
H
H
H
•g
c
1
i
-1
ii
1
1^
0)
: H^
t3
o
-2^
1"
>
T3 «:
"3.
a
>
8
a
O
g^
IS
>
3
u
o
o
m
H
g
<
«
o
Q
w
b
D
c
J
a -
0>
-o
^
«.-i
0
o
i!i
^
s
o .S
70 TRANSMISSION OF POWER BY LEATHER BELTING
denied, but a brief comparison of results actually obtained from
tests performed from time to time, on belts operating at widely
different speeds of service, leads one seriously to question its applica-
tion to practice. Such comparisons rather lead one to expect more
nearly correct results when drives are designed on the basis of relative
slip between belt and pulley.
5 Fig. 1 and Fig. 2 show a series of curves representing, for a
number of different tests, the relation existing, first between values of
the coefficient of friction and velocity of slip, and second, between
values of the coefficient of friction and percentage of slip. Informa-
tion regarding the data from which these curves were plotted is given
in the table. Letters designate corresponding tests in either set
of curves.
6 Referring to Fig. 1, it will be noted that for each different speed
of belt there appears to exist a clear and well defined relation between
values of the coefficient of friction and velocity of slip; at lower
velocities of slip more especially, the value of the coefficient appearing
to be higher for slow-speed belts and lower for high-speed belts.
Obviously then, any curve representing a relation between values of
the coefficient of friction and velocity of slip holds true only for that
speed of belt for which it is plotted and cannot be used indiscrim-
inately for all speeds of belts. The effective range of velocities of
slip which would be used in the design of belt drives would probably
be from 0 to 25 ft. per minute, as indicated in Fig. 1, and the error
which might be incurred then in using either of the extreme outside
curves shown (even though they do not represent maximum possible
range of speeds of belts) would vary from about 70 per cent to values
almost infinitely large. The curves B^ and C^ indicate the relation
that would presumably have held true between values of the coeffi-
cient of friction and velocity of slip for belts B and C had those belts
been such as to have shown a maximum value of the coefficient equal
to that of belts A, E or F.
7 From Fig. 2 it will be noted that for any given speed of belt
the same general relation between values of the coefficient of friction
and per cent of slip appears to hold true, and belt drives designed on
such a basis might then reasonably be expected to give anticipated
results at all speeds of operation. When curves B^ and C^ are plotted
to represent higher values of the coefficient of friction for belts B
and C, as in Fig. 1, the similarity in form of these curves is remarkable
— the more so as they represent tests performed in some cases over
twenty years apart.
TRANSMISSION OF POWEK BY LEATHER BELTING 71
8 As an example of the results to be expected when drives are
designed on the basis of velocity of slip of the belt, as proposed
in the paper, let two extreme conditions of service be taken: one a
drive operating at a speed of 400 ft. per minute and at a slip of
about 2^ per cent, the other a belt operating at a speed of 5000 ft.
per minute and at a slip of about 1 per cent. The slow-speed belt
would then have a velocity of slip of 5 ft. per minute on each pulley,
the high-speed belt of 25 ft. per minute. Referring to Fig. 7 in Mr.
Earth's Appendix, it will be found that for such velocities of slip
the values of the coefficient of friction to be used should be respect-
ively 0.38 and 0.53; but the maximum value of the coefficient of fric-
tion at even the highest velocities of slip of 60 ft. per minute, as shown,
is only about 0.57 and it appears then that the overload capacities
of all drives is Umited to about that value. This amounts, in the
case of the slow and high-speed belts given, to about 50 and 7^ per
cent respectively.
9 It is safe to say, however, that fully 80 to 90 per cent of all
high-speed drives are used in connection with electrical machinery,
and for such work drives must have an overload capacity of at least 50
per cent of their rated capacity. It would be necessary then, for
such practice, that belts be originally designed for correspondingly
lower velocity of slip, amounting in this case to a velocity of about 5
ft. per minute. The high-speed belt, noted above, at ordinary con-
ditions of service would then operate at a relative total slip between
pulley and belt of only one-fifth of one per cent.
10 While the author has had occasion to observe a large number
of successful high-speed drives on electrical machines, just the con-
dition of slip here indicated has never been noted, but in almost every
case the overload capacity of the drives has been noted as a function
of the relative slip between belt and pulley.
11 When it is further considered that, generally speaking, the
point at which the belt will leave its pulley is a function of the per
cent of slip and is taken independent of the speed of belt, it certainly
appears that the relative slip cannot but play an important factor
in determining values of the coefficient of friction to be used for such
drives.
Fred. W. Taylor.* The belt is one of the oldest and most
commonplace of the elements used in shop practice, so that engineers
designing new establishments or remodeling old ones, who wish to
be up-to-date, naturally incline toward the use of the electric drive
' Discussion abstracted.
72 TRA.NSMISSION OF POWER BY LEATHER BELTING
rather than the belt. There is no doubt, however, that this has led
to the use of the electric drive in many instances where the belt would
be far more economical and satisfactory in almost every way.
2 In the average machine shop, for instance, the writer is pre-
pared to say that for more than half of the machines the belt drive
can still be used with greater economy and with more satisfactory
results than the electric drive; only on the assumption, however, that
the belting is systematically cared for. The most serious objection
to the belt drive as generally used is the loss of time due to interrup-
tion to manufacture when retightening and repairing, and to the
loss of driving power and consequent falling off in output, when the
belt is allowed to run too slack. Belts can be tightened and repaired
at regular intervals after working hours, however, with the use of
spring-balance belt-clamps to get the right tension, causing thus
practically no interruption to manufacture.
3 As will be explained later, it has been shown by an accurate
record kept through a long term of years, that in the average machine
shop the average cost per belt per year is $2.25. This includes the
original cost of the belt, plus all labor and materials used in main-
taining, repairing and cleaning it throughout its life. No similar
statistics for the maintenance and renewal of the motor drive seem
to be available, but I think no one will contend that the latter can in
any way approach this economy.
4 In a great number of cases the electric drive should be used in
the machine shop, but in the writer's judgment the burden of proof
still rests on the motor drive to show in each case that the economy
in delivery and removal of work more than makes up for the extra
cost of installation and maintenance, and for the delays incident to
repairs, blowing out fuses, etc. In large machines economy lies on
the side of the motor drive in many instances, but with almost all
small machines the belt drive should still be used. In view of these
facts, the belt drive is hardly a back number.'T In fact, the manager of
one pulley manufactory told me recently that even during the dull
times his company had been selling from twelve to fifteen thousand
pulleys per month.
5 Under the rules still in common use, a large proportion of belt
drives are badly designed, and belts are used under heavier tensions
than they^should be for all-round economy.P^All who have experi-
mented with belting or who have been interested in the mathe-
matics of belting, will be filled with admiration at the remarkable
analysis which Mr. Barth has made of this diflacult problem. Even
TRANSMISSION OF POWER BY LEATHER BELTING 73
Mr. Lewis, whose experiments and scientific conclusions have properly-
been s;iven first place among writings on this subject, tells us in his
paper that life is too short to attempt a complete mathematical solu-
tion of the problems involved. Yet this is precisely the task which
has been accomplished by Mr. Barth.
6 The experiments of Messrs. Briggs and Towne and those of
Messrs. Bancroft and Lewis will remain for many years as classic
monuments in the development of our scientific knowledge of belt-
ing laws; but Mr. Barth's remarkable analysis of the work of former
experimenters, supplemented by his accurate though less voluminous
experiments on the elastic properties of belting and on the rate and
extent of the fall in tension of belts, has rendered his conclusions as
to economical speeds and the proper sizes of belts more reliable than
those of any previous writer. His final recommendations should be
accepted, therefore, rather than those in the papers of Messrs. Towne,
Lewis, or the writer. ^
7 It may be of interest to know how the figure of $2.25, quoted
earlier in the paper as the cost per belt per year, was found.
8 In the new machine shop of the Midvale Steel Company,
beginning in the year 1884, the writer experimented^ with all of the
belts in the shop, in practical use; and upon the comparative values
of the four leading types of leather belting then in common use.
This experiment lasted nine years with belting running night and day
(equivalent to eighteen years running ten hours per day). Exact
records were kept of all items affecting the life and economical use
of belting, and at the end of the experiment, among other items, it
was found that the average belt cost (under the ordinary belt rules
then in use, as, for example, those used on the cone pulleys of the
various machines; and on the ten hour basis) $3.34 per belt per year
for the first cost plus all labor and materials used in maintenance
and repairs. These are double belts, averaging 29 ft. long by 3.8
in. wide.
9 These belts were run under too high ftension lor economy,
however. They lasted on an average 14 years (ten hours per day).
The remaining belts in the shop, which proved more economical,
lasted on an average not far from 28 years (ten hours per day), and
cost per year per belt less than $2.50 for first cost and maintenance,
* These experiments are described in a paper entitled Notes on Belting, pre-
sented before the Society December 1893, and forming part of Volume 15 of
the Transactions.
74 TPIANSMISSION OF POWER BY LEATHER BELTING
etc. And this although they were materially larger than the cone
belts, averaging 50 ft. long by 4.84 in. wide. The machines in this
shop averaged much larger than in the average shop, and an investi-
gation has led me to the conclusion that in the average shop the aver-
age belt would be about equal to a 3-in. double belt, 20 ft. long. The
first cost plus the maintenance of this belt would not be greater than
$2.25 per belt per year.
10 The care of belting should be entirely taken out of the hands
of the men who are running the various belt-driven machines, and
belts should be systematically retightened at regular intervals, with
belt-clamps fitted with spring-balances, each belt having the tight-
ening strain carefully figured in advance. Belting should also be
cleaned at regular intervals, and should be softened with the small
amount of belt-dressing which is needed to keep it in perfect con-
dition. A laborer can be quickly trained to tighten and care for all
the belts in the shop during the noon hours and on Saturday after-
noons and at other times when the shop is not running.
11 Two elements of great importance in Mr. Earth's paper are
The Influence of Pulley Diameters on the Sum of the Tensions of the
Belt and a condensation of the discussion of the formula. Ratio of
Effective Tensions e 4>°^. Not only has this discussion a great the-
oretical interest, but the conclusions have a distinct practical
value.
Charles Robbins.^ In applying motors in textile mills where the
belt has been in use for years and the proposition is essentially that
of constant and uniform speed we discovered that the capacity of the
spinning frames was largely increased. This led us to make tests upon
the loss of speeds, or slip of belts and their lack of uniform operation.
The net result in using the induction motor instead of the belt is an
increased production of at least seven per cent, and in some instances
even ten per cent. Probably some of this increase is due largely to
the fact that the belting systems tested were not designed in accord
with Mr. Barth's system; but I believe that a system of belts will never
approach the uniform and constant speed of an induction motor.
2 The question of efficiency may be classified as (1) the primary
efficiency from the engine shaft to the shaft of the driven machine;
'Charles Robbins, Westinghouse Electric and Manufacturius Company, East
Pittsbiire. Pa. Discussion abstracted.
TRANSMISSION OF POW^R BY LEATHER BELTING 75
(2) the economies which result from the use of tlie electric motor
drive. These secondary economies, which are undoubtedly the most
important, will vary with the class of industry to which the electric
motor is applied. It is greatest in those industries where the load-
time factor of the installation is lowest and where the inherent charac-
teristics of the electric motor are of greatest value. These charac-
teristics are as follows:
a Ability to adjust the speed according to the demands of the
work.
b Absolute certainty of a uniform and constant speed.
3 While these two characteristics may seem to be opposed, they
are important factors in the increase of production of different types
of macliines. As widely separate examples: for a machine tool the
readiness with which the speed of a motor may be varied to the right
quantity for the work required contributes to its increase of production;
on the other hand textile mill service requires an absolutely constant
and uniform speed, which is obtained from the induction motor.
4 In determining the value of an electric motor drive the essential
point is always the secondary, or accruing economies from its use,
rather than the primary economy, although when the primary is added
to the secondary the net result will be extremely satisfactory.
Geo. N. Van Derhoef.^ The author's plan of proportioning
belts so that the slack will be taken up at approximately regular
intervals of time, regardless of speed or power transmitted, is excel-
lent from a theoretical point of view. He is obliged, however, to
divide belts into two classes — machine belts and countershaft belts,
under different initial tensions, and therefore with different periods
between adjustments. I think it will be absolutely necessary to
provide more classes. In some cases first cost is of greater impor-
tance, and in other cases the expense or inconvenience of taking up
belts is the main consideration. With a large belt, running night
and day, the stopping of the drive to take up the belt is a serious
matter. In the case of many drives, however, this is a matter of
small moment.
2 I have had considerable experience with large quarter-twist
belts, running from 12 to 20 in, in width, for connecting horizontal
and vertical shafts, and have seen results that appear incredible in
view of much of the theoretical data published on belting. These
belts were under high unit-tension, and always subjected to reverse
' Discussion abstracted.
76 TRANSMISSION OP POWER BY LEATHER BELTING
bending over deflecting idlers. Probably one reason for the success
of belts of this kind is the automatic regulation, within limits, of the
slack-side tension, due to the belt worldng up and down across the
face of the pulley on the vertical shaft. As far as I have observed,
belt drives of this kind, when properly designed and erected, have been
as satisfactory as horizontal belts with about the same distance
between centers.
3 Possibly the larger unit-stresses frequently used necessitate a
slight actual slipping of the belt on the pulleys, with some correspond-
ing increase in the coefficient of friction. This should not necessarily
be 'regarded as poor practice, but simply as a factor to be weighed
against savings in first cost, friction losses, etc. There seems to be
no fundamentaJ objection to slipping within certain limits, pro-
vided such slip is a constant quantity. All belts are continually
sliding, to some extent, on the surface of the pulleys, due to the
theoretical creep caused by the elasticity of the belt. A little more
would not necessarily be serious. The surface of a well finished
leather belt is such that sliding on a polished iron pulley will not cause
much harm provided the heat generated by the slip is dissipated with
sufficient rapidity to prevent the temperature of the belt surface from
rising too high. This, of course, involves a loss of energy, as do very
large belts under low tensions, and the crowning of pulleys. The
writer desires to emphasize that due consideration should be given
to all the factors involved.
4 Spring belt-clamps should be used wherever practicable, and
ought not to be very expensive if manufactured in reasonable quanti-
ties. In the majority of cases, however, we shall have to be satisfied
with figuring belts properly, and leave the actual initial tension to fate.
5 The idea that the maximum working-stress of a belt should
not be determined by its ultimate strength is, I believe, correct.
This becomes more apparent in studying transmission ropes. It is
a well-known fact that the maximum unit-stress for a manila trans-
mission rope should be of such amount that the side-pressure between
the lubricated fibers of the rope will not cause abrasion when the
ropes bend over the sheaves, and the fibers slide on one another.
Probably some such internal action takes place in the case of leather
belts. In transmission ropes the ultimate strength bears a greater
ratio to the proper maximum working stress than is the case with
leather belts. Manila rope is therefore a very safe transmitting band.
6 The constant lengthening of belts in service has its counterpart
in ropes. Where a rope is simply carried around two sheaves, as
TRANSMISSION OF POWER BY LEATHER BELTING 77
in the separate rope system, the general equation of the rope is
without question similar to that which the author has shown to be
true of leather belts.
7 The continuous system of rope transmission, with its automatic
tension carriage, has the slack-side tension maintained at a minimum.
This is one of the fundamental reasons why the continuous system
can transmit the same amount of power at the same rope speed and
with the same rope life, with less rope than is possible with the sepa-
rate wrap system. A few years ago the continuous system was looked
upon by most engineers with considerable scepticism; its enormous
development in the last quarter of a century is due simply to its
basis on absolutely sound mathematical principles.
Walter C. Allen. My contribution to the discussion will
relate to the practical results obtained from the installation of an
improved method of caring for belting, rather than to the technical
phases of the question. In this connection a brief description of
the working out of the improved system in the works of the Yale &
Towne Mfg. Co. may prove interesting.
2 The problem of transmitting large amounts of power by means
of belting is not a serious one with us, as our power is for the most
part transmitted electrically; each room is provided with one or
more motors, and the power is distributed from them through line
and countershafts to the machines. The great majority of our belts
are small, and many of them run at high speeds. Altogether we have
about 4800 belts, so that their proper maintenance is an important
and somewhat difficult problem.
3 Early in 1905, at Mr. Barth's suggestion we undertook an inves-
igation of our belting and the methods employed in its upkeep,
as a result of which we decided to adopt a system of caring for belting
recommended by Messrs. Taylor and Barth. For the sake of brevity
I have divided my notes into comparative statements, of the con-
ditions before and after the adoption of the new method as'affecting
each element of this important subject. It may seem that the con-
ditions existing before the installation of the new plan were dis-
tinctly bad, but I venture to say that they were as good as those in
many manufacturing establishments at the present time, if not
better. The improved conditions, however, are so infinitely superior
to the old that by comparison the latter appear extremely anti-
quated and crude.
4 Tensions. Under the old plan we had no means of knowing
78 TRANSMISSION OF POWER BY LEATHER BELTING
with any accuracy the tension of a belt. It was left to the individual
judgment and experience of those doing the repairing, so that inevit-
ably the tensions of the belts varied in proportion to the variation
of judgment of the repair men.
5 The first step in the reorganization was the building of a belt
bench and the provision of tension scales such as are shown in Fig. 6.
These are used now altogether for the determination of tensions.
6 Records. Under the old regime we had no records whatever
of our belts.
7 Under the new plan we have a record of each belt showing its
location; its type, i. e., whether open or crossed, countershaft or
machine belt; kind of leather; thickness, width and length. These
records also show for each belt the dates of inspection.
8 Organization. Under the old plan our millwrights cared for
the heavy belts, but the repairing was done only when the belt gave
way, or stretched so that it failed to transmit the necessary power.
The small machine belts were cared for by the individual macliine
operators, rhany of whom knew absolutely nothing about belting,
and in some cases our investigations showed that ignorant operators
had attempted to tighten a belt by cutting out a piece, and, finding
that they had cut out so much that the belt would not go over the
pulleys, were then compelled to cut out still more and set in a piece
in order to make the belt long enough to do the work. In these
cases also the belts were not repaired until they actually gave out
through breakage or failed to give the necessary pull.
9 Under the new plan a gang of four men do absolutely nothing
else but inspect belting and attend to the repairs and retightening.
A belt room has been provided in which is an annunciator, and a
series of push buttons are arranged at the telephone central, so that
in case of an accident to a belt the foreman or gang boss can call the
belt man easily. In a plant as large as ours the annunciator results
in a great saving of time.
10 A tickler system was installed by means of which <;he belt
gang are notified regarding the belts to be inspected each day.
After the inspections are made these tickler cards are returned to
the office where the proper records are made and the ticklers put
back for the next inspection.
11 These belt men take their lunch hour from 11 to 12 o'clock,
working during the noon hour, and are thereby enabled to repair
many belts which could not be repaired when the works are running,
without loss of time to other employees.
TRANSMISSION OF POWER BY LEATHER BELTING 79
12 Fastening. Under the old plan there was no fixed rule regard-
ing the fastening, rawhide lacing and belt hooks being used indis-
criminately. Under the present plan Jackson wire lacing, put into
the belts by means of a machine, is universally used. For continuous
belts, under the old plan we used a kind of glue which took from three
to ten hours to set satisfactorily. Under the present plan we are
using a special glue which will set hard in thirty minutes. This also
results in a saving of time in the case of an accident to continuous
belts.
13 Belt Dressing. Under the old plan comparatively little belt-
dressing was used, but in many cases rosin was used through ignorance
of the fact that it causes the belting to deteriorate rapidly. We
now use entirely Plomo belt-dressing, which is extremely useful and
tends to prolong rather than to shorten the life of the belt.
14 Reclamation of Belts. Under the old plan no reclama-
tion was attempted, but at the present time we reclaim a consider-
able amount of belting each year. Belting damaged on the edges
is cut down and used for narrower belts, short pieces are scarfed
and glued together and the oil is taken out of oily belting and the
belts used over again.
15 Kind of Belting. Several kinds of belting were used under
the old plan, but we have gradually standardized our belting until
at the present time practically nothing but a high grade of oak-
tanned belting is used.
16 Cost of Up-Keep. Of course there was no method of determin-
ing the cost of maintenance under the old plan. Our records show
that during the year 1906 the labor-cost of maintaining our belting
system was 96 cents per belt. During 1907 it was 73 cents and dur-
ing 1908, 45 cents. This decrease has of course been due to the
increased efficiency of the men doing the work and to the fact that
experience has indicated where inspection periods could be lengthened
out, and also to the fact that the belting is now in such condition that
expensive breakdowns seldom occur.
17 The foregoing statements describe briefly the various features
of the old and the new plans; a summary of the advantages of the
new plan follows:
a Decreased cost of belting. The cost for the year 1907 was
only about 60 per cent of that for 1906, despite the fact
that we installed more new machinery in 1907 than in
1906.
80 TRANSMISSION OF POWER BY LEATHER BELTING
6 Increased efficiency of machines, due to the fact that the
tensions are maintained much more uniformly than for-
merly.
c Continuous production by both men and machines, due to
decreased interference due to belt-breakdowns.
d Uniform type of belt lacing, decreasing danger to employees.
e Decreased cost of maintenance.
f Under the present plan the cost of maintenance appears as
a separate item where it can be watched and compared
with that of previous periods to determine the relative
economies, while under the old plan the figures were
combined with a mass of others so as to make it impos-
sible to determine how much it had cost.
CONCLUSION
18 When we first commenced to install the new system we had
all sorts of trouble as is generally the case with any new thing.
The plan was opposed by foremen, gang bosses and workmen, each
of whom had an idea that the new tensions were entirely wrong, and
that the machines would never do the work properly, unless they
could adjust the belting according to their individual ideas. One
of the best evidences of the value of the present plan is that this
antagonism has entirely disappeared, and what was at first con-
sidered by many an interference and a hindrance is now accepted as
a help and is believed to be entirely satisfactory by those competent
to hold an opinion.
Mr. Taylor. The original experiments at the Midvale Steel
Works were started in 1884; 17 years later, when all^the machinery
in that shop was taken out, one of the belts, which was of the type
of those run under proper rules, that is, approximately the low tension
suggested by Mr. Barth, had run all that time night and day under
heavy tension. During this time it had required tightening only nine
times, and at the end of the equivalent of 34 years of ten-hour service
that belt came off its pulleys and was immediately put to work on
another machine, in good condition. This instance of the life of a
belt properly taken care of and properly tightened will be a sur-
prise to the man accustomed to see a belt go out of use in from two
to five years. This statement has just been determined.
TRANSMISSION OF POWER BY LEATHER BELTING 81
D WIGHT V. Merrick.^ As I am interested in chain drives, I
will draw attention to some experiments made by Hans Renold,
Ltd., of Manchester, England, and embodied in a pamphlet issued
May 1908, comparing the relative efficiency of chain and belt drives
on automatic machines. Mr. Renold claims that with the chain
drive the output was increased 20 per cent, fewer drills and parting
tools were used, and a better finish was obtained on the work. He
says: " The tool did its work unflinchingly at every part of the revolu-
tion of the spindle — no more and no less." He further states that the
wear and tear on the spindle and countershaft bearing was consider-
ably reduced. These statements were so striking that the Link-
Belt Company, with which I am associated, decided to make further
tests. In one of these which I was detailed to make I maintained a
constant feed and speed and used the same tools with each drive,
and in all cases the tool was used until it became necessary to re-grind,
the object being to cut off as many pieces or drill as many holes as
possible before this condition was reached. The tool when chain-
driven did considerably more work than when belt-driven. I quote
from my report as follows:
2 These tests were made on a 3 in. by 36 in. Jones and Lamson
turret lathe, with "blue chip steel" cutting-off tool {% in. wide, cutting
off cold-rolled shafting 2^ in. diameter, feed 0.012 in. per revolution.
3 Care was taken in forging, treating and grinding the several
tools used, to insure uniformity in their cutting qualities; but to
obviate the possibility of the results being affected by the cutting
qualities of the different tools, each tool was used with both drives.
4 One of the tools cut off 16 pieces when chain-driven before it
became necessary to re-grind, and only 9 pieces when belt-driven.
The cutting speed in both cases was 94 ft. per minute, feed 0.012 in.
per revolution, and another tool cut off 8 pieces when chain-driven
against 5 when belt-driven. In this latter case the cutting speed was
130 ft. per minute, feed 0.012 in. for chain and belt.
5 As the cutting periods in the above test were so short, two more
series of tests were made with longer continuous periods. These tests
were made on a drill press \vith new | in. carbon steel drills in a soft
cast-iron block, 3 in, thick. The same drill was used on both drives,
and was carefully and uniformly ground for each test.
6 One of the drills when belt-driven drilled 31 holes before it
became necessary to re-grind, but when chain-driven the same drill
drilled 57 holes, the cutting speed in both cases was 62 ft. per minute,
1 Dwight V. Merrick, Link-Belt Mfg. Co., Nicetown, Philadelphia, Pa.
82
TRANSMISSION OF POWER BY LEATHER BELTING
TABLE 1
Results of Experiments om a 3-in. by 36-in. Jones & Lamson Turret Lathe 0.012 in
Feed per Revolution.
^
i
1
M 0
a o
Kind of
No. OF
Metal Cut
Cutting Speed ,
Condition
Time
R.p.M. of
o
Drive
Pieces
BY Tool
Ft. per Minute
op Tool
Minutes
Spindle
Inches
Beit
6i
8.125
94
Rmned
5.91
143
2
Ch^n
16
20.
94
Ruined
14.70
143
Belt
9
11.25
94
Ruined
7.15
143
Chain
7i
9.843
128
Ruined
4.81
196
5
Belt
4i
6.093
134
Ruined
2.72
203
Belt
1
1.25
134
Good
0.51
203
4
Belt
i
0.312
151
Ruined
0.12
231
Chain
1
1.25
126
Good
0.54
193
Chain
i
0.937
151
Ruined
0.39
231
Belt
1
1.25
129
Good
0.53
197
1
Belt
i
0.625
146
Ruined
0.15
223
Chain
1
1.25
129
Good
0.53
197
3
Chain
1
1.25
149
Fair
0.44
228
Chain
f
0.468
195
Ruined
0.15
299
Note: A higher cutting speed was obtained by the chain drive.
TABLE 2
REfsuLTS of Experiments on a Drill Press with New J-in. Diameter Carbon Steel
Drills, 0.018 in. Feed per Revolution, in a Soft Cast- Iron Block, 3-in. Thick
>^
Condition
o
Kind of
Drive
Holes
Drilled
Number
Metal Cut
BT Drill
Inches
Cutting Speed
R.P.M.
OF Drill
AFTER
Drilling
Holes
Time
Minutes
R.p.M. or
Spindlb
1
Belt
31
93
62.2
Started to
18.91
273
1
Chain
57
171
60.5
nun
Comer
rounded,
needed
grinding,
starting to
ruin
35.91
264
2
Chain
37
111
62.2
Starting to
22.57
273
2
Belt
20
60
62.2
ruin
Starting to
ruin
12.20
273
Note: A great many more holes were drilled by the chain drive.
TRANSMISSION OF POWER BY LEATHER BELTING
83
TABLE 3
Results of Experiments on the same Drill Press as in Table 2 with New 1;]-in. Diam_
ETER Carbon Steel Drills, 0.018 in . Feed per Revolption, in a vest hard Cast.
Iron Block, 3 in. Thick
1^
0
Kind of
Drive
Holes
Drilled
Ndmbicr
Metal Cot
BY Drill
Inches
Cdttinq Speed
R.P.M.
Condition
OF Drill
AFTER
Drilling
Time
Minutes
R.P.M. OF
Spindle
Holes
3
Chain
17
51
28.0
Starting to
19.84
148.8
3
Belt
14
42
28.4
ruin
15.40
151.4
4
Chain
17
49i
28.0
Started to
run on 17 th
hole 1^ in.
deep
18.48
148.8
4
Belt
13
39
28.2
Started to
niin
14.44
150.0
Note: More holes were drilled by the chain drive, but the percentage of gain was not
anywhere near as great as in Table 2.
feed 0.018 in., and another drill at the same speed and feed drilled
37 holes when chain-driven, against 20 when belt-driven.
7 The other series of drill tests was made on the same drill
press, with ff in. carbon steel drills on a very hard cast-iron block,
3 in. thick. One of the drills when chain-driven drilled 17 holes
before it became necessary to re-grind, against 14 holes when belt-
driven; the cutting speed in both cases was 28 ft. per minute, feed
0.018 in. per revolution, and another drill did 17 holes, chain-driven,
against 13 belt-driven, same feed and speed as above.
8 The results were so gratifying that further tests are being made
on four similar automatic machines, at our plant in Indianapolis,
two fitted with belt drives and two with chain drives. The same feeds
and speeds will be maintained with each drive throughout the series
of tests, but a variety of tests will be made to establish the maximum
efficiency of both belt and chain drives, to the best of our ability.
The results will all be tabulated and published in a pamphlet in the
near future.
9 The accompanying tables contain the tabulated results of my
experiments.
F. A. Waldron, After listening to this paper, one naturally
asks the question, What is its commercial value? Mr. Allen has
answered this very well, but I will give a little of my own experience
with the system.
84 TRANSMISSION OP POWER BY LEATHER BELTING
2 At the plant of the Yale & Towne Company, most of the
responsibility for the condition of belts, prior to the author's going
there, was placed with me and I am willing to take any criticisms.
I became an prdent advocate of Mr. Earth's work on belts, however,
particularly because of the practical results obtained.
3 After leaving the Yale & Towne Company, I had occasion to
purchase a Barth bench and spring-balance and apply the elements
of the system without spending a large amount in replacing counter-
shafts. I established the system of varying tensions on different
machines. A light countershaft would not stand as heavy tension
on the belt as the author originally prescribed. Tensions on belting,
lengths, taking up, etc., were recorded. A record of complaints
received in the millwright department for a specified number of coun-
tershafts and machine belts had been kept, and for ten days or two
weeks before installation something like 150 complaints came in.
After complete installation of the Barth bench and scales and the
Barth system, the complaints dropped to 80 for two weeks, and six
weeks later to 35, showing the commercial results of systematic care
of belts.
4 Belts as low as 1^ in, wide, and some heavy double belts three
to four inches wide, were the limits on size.
5 This system was installed almost at the cost of my reputation,
and on leaving that concern I supposed that the belt bench and
bench-scales would be relegated to the scrap heap. Having an
opportunity to put in a belt-bench and scales elsewhere, however,
I wrote the firm asking if they did not want to sell the bench and
scale and they said "no."
A. A. Gary. I was much interested in Mr. Allen's remarks
concerning the employment of Mr. Barth's system and formulae for
the selection and proper application of belts to drive the numerous
machines at the Yale & Towne plant, but explanation of one essential
point is needed to show how this can be practically accomplished.
2 As I understand, one important factor required in the formula
used to determine the proper initial tension to which each belt must
be subjected when put in place, is the horse-power to be transmitted
by that belt. It has been stated that 4000 belts are used in this
plant, operating perhaps one-half that number of machines. I would
like to know the method employed to determine the power require-
ments of each of these macliines so as to obtain the required value of
this factor when the formulae are used.
Q TRANSMISSION OF POWER BY LEATHER BELTING 85
3 If we merely guess at the power required, we depart from the
exact scientific method of determining information in our belt prob-
lems and recede toward the "rule of thumb" method, as a formula
is no more exact than is the value of the most uncertain quantity
employed in its solution. If Mr. Barth can give us any "short cut"
method for determining the power required by machines to be driven
by belts, he will funish information that will give his formulae a very
practical value.
A. F. Nagle. This paper does not pretend to present new
facts, but sets forth, in mathematical formulae and diagrams, data
obtained by, .Messrs. Lewis, Taylor and others. It also diagrams
some simple arithmetical computations. As a work of mathematical
study and diagrammatic representation, the paper is admirable, but as
a practical aid to a busy engineer, it seems to me too compUcated.
The only part which holds my attention is the diagram in Fig. 3..
giving the horse power of belts at different velocities, and of two types
spoken of here as countershaft and as main di'ive belts, but more
commonl}^ designated as "single and double thickness." The reason
for this distinction is that while the stress in the net solid body of the
leather is taken to be the same in each case, in "single" belts the
joint is a butt joint and is laced. This cuts away more of the belt
than where the belt is of double thickness, lapped and cemented, or
riveted; the difference being in the character of the joint rather than
in the strength of the belt.
2 For comparison then, we can take Mr. Earth's estimate of the
relative strength of these belts, as 160 to 240 or 1.0 to 1.50. Mr.
Towne found these to be as 1.0 to 1.82, and in my early studies, I was
inclined to adopt this ratio; later, however, I have used the ratio of
275 to 400 or 1.0 to 1.45.
3 The belt problem is very far from being one of pure mathematics.
As in most engineering problems, there is about 5 per cent of scientific
knowledge involved, and fully 95 per cent of good judgment based
upon experience. We rarely know the exact power to be transmitted
except in the case of prime movers. The arc of contact, the velocity,
and the stress we are willing to put upon the leather, are all easily
determined, but we cannot decide upon the coefficient of friction by
formula. A new leather belt upon an iron pulley may not have a
coefficient of friction of 25 per cent, while the same belt, well worn and
well groomed, will give 65 per cent in a clean, dry room ; put the same
belt in a wet place, hke a tannery, or a dusty place, Uke a stone-
crushing plant, and we have an entirely different coeflacient.
86 TEANSMISSION OF POWER BY LEATHER BELIING
4 It seems to me that the designing engineer, even though he
understands the mathematics of the belt problem, if ignorant or
unappreciative of the practical conditions under which the belt works,
will be liable to make a mistake. On the other hand the engineer
famiUar with the conditions, but ignorant of the mathematics involved,
is also liable to error in his conclusions. A cautious man will endea-
vor to err on the safe side, feeHng no doubt as did our venerable ex-
President John Fritz, who when remonstrated with for making
some machinery needlessly strong, replied, " If I do, nobody will ever
find it out."
5 On general principles, it is of course desirable to work belts,
like other members of a machine, with large coefficients of safety, but
engineering in its last analysis is a question of finance and we must
"hew as close to the line" as possible. Mr. Towne found the ulti-
mate strength of laced belts to be 200 lb. per inch width {-h in.) thick,
and used ^ of this, or 66§ lb. as a safe working stress. Mr. Towne also
found a coefficient of friction of 42 per cent to be safe. The general
practice of the day has been quite close to these factors, but if I under-
stand his diagrams correctly, Mr. Barth has departed far from them.
6 In 1881 I read before the Society a paper giving for the first
time, I believe, a belt formula which took cognizance of the effect of
centrifugal force. The data used therein were based principally
upon Mr. Towne's experiments. The results obtained were well
within the safe limits of previous practice for low speeds, but at high
speeds my formula showed the deviation. Common formulae gave
results (see Kent, Mech. Eng. Pocket Book, p. 879) as follows:
For single belt 1 in. wide,
Velocity (1) (2) (3) (4) Nagle
600 ft. per min. 1.09 h.p. 0.55 h.p. 0.60 h.p. 0.82 h.p. 0.73 h.p.
Barth gives only 0.40 h.p.
For double belt,
Common Formula Nagle Barth
1.17 h p. . 1.24 h.p. 0.68 h.p.
7 For the purpose of giving a clear conception of Mr. Barth's
deviation from the others, I repeat my formula here:
C.V.tw{S - 0.012 V^)
h.p. =
550
C is a constant expressing the adhesion of the belt upon the pulley
under a unit of stress of belt. Its value is expressed by the equation
TRANSMISSION OF POWER BY LEATHER BELTING
87
C = 1 — 10 000''58/a ^vi^gre a is the arc of contact and/ the coefficient
of friction. The other quantities are as follows:
V = velocity in feet per second.
5 = stress upon leather per square inch, which I have taken
at 275 lb. for laced and 400 lb. for riveted belts.
t and w are the thickness and width respectively in inches.
550 ft. lb. = horse power per sec.
14
13
12
11
10
9
1
4
^
,»'
—
/^
^ ^
*sl^
—
^
;^
^
I'
^
r
5
^
^
00*2
oo
'^
^
^
y
- — ' 1
Nagle
4
__-5
s^"^
/
t^
:^
-"^ — -^Bar
tb-(D
>able
3
^
-
^
^
^^
^^
>^
Si
A
s^ ^^
1
>*^li5*^^^^
__J_ — ^ ■
J
^fn^
1
10
20
30
40 50 60
70
80
90
100
600
1200
1800
Feet per Second
2100 3000 3600
Feet per Minute
4200
4800
5400
GOOO
Fig. 1 Comparison op Different Belt Formula, Based Upon Belts 1
In. Wide and ^ In. Thfck for Single and \ In. Thick
FOR Double Belts
8 To illustrate the solving of this equation, let a = 180 deg. an
/ = 0.40, then
180 X 0.40 X 0.00758 = 0.54576
10-0.54576 _ iQg iQ y^ 0.54576 = 1 X 0.54576
0.54576 is a logarithm of which 3.513 is the number. This being a
minus — coefficient, we must take its reciprocal or 0.284; subtracting
this from 1, we get 0.716. The result could have been obtained by
subtracting the log 0.54576 from 1, giving 1.45424, and this gives
0.2846 as its number direct.
88 TRANSMISSION OF POWER BY LEATHER BELTING
9 In Kent's Mechanical Engineering Pocket Book, p. 878, tables
are given based on this formula, which facilitate its use. I wish to
call attention to the wide divergence of Mr. Earth's conclusions from
those commonly used as well as from my own, as plotted in Fig. 1
herewith. I have reduced his figures to the same thickness as mine,
namely -h in. for single and \ for double.
10 This work has been done somewhat hastily and I hope the
author will check it at least so far as relates so the interpretation of
his diagram. If my work is correct, I am puzzled to understand why
his tables of belt horse power differ so much from mine.
Prof. Wm. W. Bird. I feel very much pleased and highly compli-
mented to see the results of Mr. Barth's mathematical analysis of
ray experiments on belt creep. On a few points, however, I am still
in doubt in regard to his general conclusion. I believe:
a That the elasticity of a belt varies with the velocity and
that at very slow speeds the sum of the tensions would
remain constant, while at high speeds, if it were not for the
centrifugal force, the sum would increase] practically
with the load. If this is true, a belt will improve with
higher speeds and will not reach a maximum at 4000 ft.
per minute as shown in Fig. 3.
h That in determining the size of a belt for a given load, the
diameter of the smaller pulley should be considered. A
belt will do relatively less on a small pulley than on a large
one, other conditions being the same.
c That the crowning of the pulleys should be considered as it
affects the life of the belt.
d That the carrying capacity of a belt should not be given in
terms of square inches of cross section, as a double belt
with one square inch of cross section will not transmit as
much as a single belt of the same cross section.
2 I have recently made some tests on compound or rider belts
and have been somewhat surprised at the relative movements of the
main and rider belts with various pulley ratios, from which I have
concluded that if these belts were glued together, as they would be in
a double belt, numerous internal stresses and strains must be set up
when the belt passes over a pulley, especially over a small one.
3 The crown is a very serious matter on a small pulley, as the
following figures will show. Take a 4 in. pulley with a crown of 0.2
in. in the diameter, if the belt wraps 180 deg., the length in the
TRANSMISSION OF POWER BY LEATHER BELTING 89
middle of the belt will be 0.31 in. greater than on the sides; this means
a stretch of 0.05 in. per inch or 1000 lb. stress on middle fiber, taking
modulus of elasticity as 20,000 lb. The belt must slip or be ruined,
for this stress does not include load or initial tension and is in itself
enough to stretch the belt beyond the elastic limit. The slipping
necessary to adjust this stress must influence the friction and slipping
of the belt as a whole.
4 I would like to have Mr. Barth answer a question which I have
been asked a great many times, — why does the sum of the tensions in a
belt increase with the load? I would also like to have him calculate
with his slide-rule the size of 'a belt for [the following conditions:
20-h.p. motor, 6-ih. pulley, 1200 r.p.m., to drive a shaft 12 ft. away
at 200. Would the same belt last as long if the drive were reversed,
that is, a shaft running at 200 r.p.m. driving a generator with 6 in.
pulley at 1200? I would like also to ask Mr. Barth or any engineer
present whether he would use the same size belt with the 6 in. pulley
as a driver as with it as a driven, and with the same size of belt; and
whether in this case it would last longer, other conditions being equal.
5 Anyone who has undertaken an investigation of the belt prob-
lem knows that it is almost impossible to keep conditions constant —
humidity, oil in the belt, surface of pulley, etc., seem to change with-
out notice and complicate the work.
6 I wish to congratulate Mr. Barth on his efforts to advance the
theory of the transmission of power by leather belting, and to agree
with Mr. Lewis in the conclusion of his paper presented in 1886,
"That there is still need of more light on the subject."
Prof. C. H. Benjamin. I have been asked to contribute to the
discussion of Mr. Barth's paper; technically, I am afraid I can not
criticise it or add to it, for it leaves but little more to be said. Sen-
timentally, I can not but regret the gradual disappearance of our
terra incognita, both geographical and mechanical. Time was when
large areas on the map bore the encouraging legend "Unexplored
Wilderness" or "Great American Desert" and left room for the free
play of the imagination. Today you miss those fascinating areas
and are tied down to realities.
2 Not many years ago, the grinding of a lathe tool was an interest-
ing experiment, attended with much uncertainty, and the matter
of feeds and speeds offered an alluring field for investigation. Mr.
Taylor has spoiled all that for us and now our imagination is worked
by slide-rule.
90 TRANSMISSION OF POWER BY LEATHER BELTING
3 Time was when the possibilities of belting were vague in outline
and when coefficient of friction, slack tension and belt creep were
rather shadowy phantoms. It was pleasant then to speculate on
what the belt would do and how long it would do it and the man with
the longest memory had the advantage. But now, Mr. Lewis, Mr.
Bird and Mr. Barth have taken all the romance out of it and another
illusion succumbs to the deadly aim of the slide-rule.
4 Perhaps I take a malicious pleasure in noting that one or more
factors of the problem are still out of harness and a trifle intangible.
Our old friend, the coefficient of friction, is in hiding under the belt
sporting with those other elusive fairies, modulus of elasticity and
belt-creep. After all, what does it matter? Aside from the interesting
theoretical questions involved, what we need to know is, first, how
wide a belt to use at a certain speed to transmit a certain power, —
Mr. Taylor has answered this question. Second, how tight to lace
or cement that belt that it may do the work for a reasonable time
without relacing, — Mr. Barth has told us that.
5 I began experimenting on belts 25 years ago and have been at it
more or less since. With a fixed pulley and a slipping belt, I found no
difficulty in proving <f> = 0.42 after Rankine, but when I built a
belt machine and tested belts under running conditions, 4> lost all
its constancy and might as well have been called x. Working back-
wards from the measured tension and using the old formula, I found
^ to vary with the load, the speed, the kind of pulley, the age of the
belt, the weather and the dominant political party — in fine, to be so
mysterious and intangible a quantity as to be useless for practical
purposes.
6 The sum of the tensions also varied in a manner that did not
admit of rational explanation. And right here let me say that the
reasons for Mr. Barth's assumption of constancy for (t^ + ^ Q are
hardly clear to me. Why call that constant which is not constant?
Why call anything constant except as it is shown to be so by measure-
ment? This is not said in criticism but for the sake of information.
7 There is one aspect of the paper that deserves special attention
and that is the recognition of the fact that a belt is an elastic connector
with a variable length and variable tensions. Most writers on the
subject have treated belting as if it were a non-extensible element
which could be exactly represented on paper and whose behavior
was capable of exact mathematical analysis. A belt in action is
almost like a thing alive, squirming, lengthening, shortening, its
tension changing back and forth with a variable modulus of elasticity
TRANSMISSION OF* POWER BY LEATHER BELTING 91
and a lag in its changes due to its contact with the pulley and the
short time intei-vals. A belt must be tested to be appreciated and
theory must wait upon experiment.
8 I fully appreciate the value of Mr. Earth's analysis and can see
that his methods will result in marked economies in establishments
where many large belts are used and where conditions are pre-deter-
mined. In the smaller shop, where conditions vary, and in isolated
cases with differing sorts of pulleys, differing kinds of belt, new and
old, I feel that each case will have to be settled on its own merits.
Until more experiments are recorded, the average machine-designer
or millwright will have to be guided largely by his own judgment and
experience in determining the width and tension of each belt. Let
us have more experiments.
H. K. Hathaway.^ To the scientist and machine-tool designer
the value of Mr. Earth's paper will unquestionably be immediately
apparent, but the writer feels that the paper does not represent with
sufficient clearness features of the problem that are of inestimable
value to the engineer concerned with running a shop. Assuming
that the designer takes care of the sizes of belts required, and the
speed at which they should be run, in accordance with the conclusions
of such eminent authorities as Mr. Taylor, Mr. Earth, and Mr. Lewis,
a great deal is lost unless the shop-man properly cares for the mainte-
nance of such belts. As Mr. Earth has pointed out, the care and
maintenance of belting in the great majority of shops is done by rule
of thumb, and left entirely to the judgment of the shop millwright or
the workman operating the machine.
2 The efficiency of a belt-driven machine largely depends upon the
tension of the belts being properly maintained at a point above the
minimum initial tension at which they will transmit the power
required. This point Mr. Earth has only slightly touched on, whereas
the writer feels that this subject should have occupied a section fully
as large as the body of the paper presented.
3 If a machine stands idle during working hours while the belt
is being repaired or tightened it produces nothing during that time,
and there is a distinct loss to the manufacturer. If a machine stands
idle for one-half hour out of ten hours working time there is a loss
of 5 per cent in the output of that machine, and if in a shop having
100 machines, 10 machines out of the 100 lose one-half hour each day
I H. K. Hathaway, Tabor Mfg. Co.. Philadelphia, Pa.
92 TRANSMISSION OF POWER BY LEATHER BELTING
on account of repairs to belts it amounts to a loss of 0.5 per cent
on the total output of the shop. This feature, however, is probably
not so bad as the loss in output due to the machine belts being run so
loose that they cannot begin to take the feeds, speeds, and depths of
cut for which the machines are designed and that the tools will stand.
4 The writer has had considerable experience with the system of
maintenance of belting mentioned in Par. 67 of Mr. Earth's paper, and
will describe it briefly.
5 Almost every foreman or superintendent, in attempting to
bring up the speeds of his machines to something like what he knows
to be possible, has found that such attempts usually result in the
belt's slipping or breaking, or the lacing giving out, and knows that
where the care of belts is left to ^the man on the machine, only in a
very few cases can the belts be depended upon to do the maximum
amount of work. If, therefore, the maximum feed, speed and depth
of cut are to be prescribed and used, as is done by the aid of Mr.
Earth's shde-rules under the Taylor system, it is essential that belts
of the best quality and of the proper proportions be used, and that
they be kept in first-class condition and at the proper tension, so that
they can be reUed upon to give the pull required. It is also necessary
that all repairing, tightening, and inspection of belts be done outside
of working hours that there may be no loss of output from interrup-
tion to manufacture. In order to accomplish these objects the follow-
ing system has been evolved.
6 A record is kept for each belt in the shop on the form shown
as Fig. 1, on which are given all standard data for each belt in ques-
tion.
7 When a new belt is to be put on, or an old belt to be inspected
or tightened, the special belt fixer's bench developed by Mr. Gulowsen
is used, together with the belt-tension scales referred to by Mr.
Earth. With this apparatus it is possible for one man to remove,
tighten and replace almost any belt in from six to eighteen minutes.
In putting on a new belt, or tightening an old one, the drums or
pulleys on the belt bench are set by means of a steel tape to corre-
spond with the distance over the actual pulleys, as previously deter-
mined, and shown on the belting record as "Length over Pulleys."
A roll of belting, of the proper width and thickness, is next placed in
the open drum, and passed through one pair of clamps of the belt
scales around the drums or pulleys and through the other pair of
clamps of the belt scales. The clamps are then tightened on the
belt and the belt drawn up by means of the screws until the spring
TRANSMISSION OF POWER BY LEATHER BELTING
93
a
0
a
o
: . S
o
<
a
z
o
z z
: 0
• I <r
a
* H § 0 M
H
" z s ;:; 0 M
O
1
■< S P 0
a
n
SH go^
: §
rt
a
1 .<o
>
i >
^
■ 1
z
•< a
a s
« •<
1 c
I
c
c
1
c
s
1
■3 ^
S c
0 ^
0 E-
ii ^
MARKS
Done
ONDI-
r, ETC.
a w 0 :a
P^ S - "
h
i^«z
« § g 0
Z
g
Pi
B.
0 3 0 s
Date
Dressing
PUT ON
During
Service
J)
'
g
c
Date
TO BE
Cleaned
AND
Greased
Again
n
"i
PM
>
9
c
1
(V
C
Q
C
g
5
•
5
1
c
o
2
c
1
&
0
a
i
"5
C
*«
Date
leaned
AND
REASED
h^l
H
c
a
K
«
0 0
sis g S
g S g a S <
;; z 2 « 5 0
Q«« H<
" g '^ .
D a H
'! S H
to
a
l| Z U Q
!S "3
5 ,-5 «
5 S
i «° W z
is -a
< M
|i "^ ^ 3
^ « 5
E-i
!l 2 oPq
a 0
M
' §^i
£•1
a '^ *
0 s
£
c
0
1
■3
en w B,
;: i
1
§
0
•3
3
Pi
"5
H
g;
o
a
c3
■3
*
J
s
s
Q
Q
3
a
94 TRANSMISSION OF POWER BY LEATHER BELTING
balances between the two pairs of clamps record the tension required,
after which the belt is cut off so that the two ends will come together,
and the belt is laced on a belt-lacing machine and put on its pulleys.
8 A memorandum, which also serves as the belt fixer's order and
time card, giving him all necessary instruction, is then placed in what
is called the "tickler," a portfolio having a compartment for each
day of the year, under the date on which the belt will probably
require re-tightening, and on that day it will be removed from the
tickler, together with the memoranda for any other belts requiring
attention, and sent to the belt fixer for attention during the noon
hour and after quitting-time.
9 These belts are then removed from their pulleys, taken to the
belt bench and tested to ascertain whether they require tightening;
if the tension is found to have fallen to approximately the minimum,
they are drawn up to the maximum tension as previously described,
a piece is cut from one end, the belt is re-laced]^and put back in
place and these facts are noted on the belt fixer's memorandum,
which is then returned to the planning department, and entered on the
belt record; and a new memorandum placed in the tickler under the
date on which the belt will again require attention. Notices for
scraping, cleaning and greasing the belts at proper intervals are also
placed in the tickler.
10 The length of time a belt will run before the tension will
fall to the minimum at which it will pull all that is required, has been
determined from experiments, and a belt seldom requires attention
before the time set for re-tightening; when this does occur, however,
a belt-dressing which does not injure the belt, but which will enable
it to pull properly until noon or the end of the day, is applied, and the
memorandum is removed from the tickler and another placed under
its next date for re-tightening.
11 The system described accomplishes four things of vital impor-
tance to economical production:
a Freedom from interruption to production from having
to repair belts during working hours, by having all belts
systematically inspected and all breakdown and shppage
anticipated and prevented before they occur.
h Possibility of using the maximum feeds, speeds and depths
of cuts at all times.
c Increase in life of the belt owing to all belts being of the
proper dimensions and properly laced and spliced and
run at the proper tension.
d Reduction of cost of maintenance to a minimum.
TRANSMISSION OP POWER BY LEATHER BELTING
95
12 Mr. Earth's belting slide-rule is used in determining the
dimensions of the belts, the maximum and minimum tensions. The
writer can speak from experience of the great value of the belting
slide-rule in solving the belting problems that confront the shop
engineer, and while the mathematical features of Mr. Earth's paper
Oui
In
Order Number
D L
Departi
Day Ra
nent
te
Man's Time
Max. Tension
Min. Tension
Belt Symbol
Cleaned and Grea
sed
Dressed While in
Use
It
Dressing Used
Length Put In
Amount Taken 0\
Length of Splice.
Cemftnt TTspd
Tension in Lbs.
Indicated by
' Before
After. .
Tightening
Each Spring Balance^ >
Workman's Name
Entered in
Pay
Sheet
Cost
Sheet
Belt
Record
T-. TT,^ time
UAT WORK
note
Fig. 2 Belt Fixer's Order and Time Card
are unquestionably interesting to many, the writer feels that, like
himself, many will be glad to accept Mr. Earth's figures without
question provided they can have the slide-rule.
13 It is a fact that in the average shop very few belts become unfit
for use through legitimate wear, but rather through accidents or
improper care. Where the care of the belts is left to the workman,
the belts are usually far too loose, and when a belt slips it is less
96 TRANSMISSION OF POWER BY LEATHER BELTING
trouble for the workman to reduce his speed, feed, or depth of cut,
or as a last resort to use rosin to make the belt pull. This use of rosin
will ruin any belt in a very short time.
14 Very few machinists or even foremen know how to tighten or
lace a belt properly, the amount to be taken out being usually
guessed at, and a great deal of time is lost through the machine's
standing idle while the cutting and trying is going on. The writer
has seen a good machinist run a cone belt, which he had made too
tight, on "cross cones," i.e., on steps not in line with each other,
with the result that it twisted itself up like a corkscrew and was
practically ruined.
15 Another cause of premature ruin of belts is improper lacing, the
ends not being cut square and the lacing on one side stretching more
than the other, causing the belt to run crooked.
16 Cemented splices, when properly made, give the best results.
Machine lacing, using a spiral wire lacing, while not so good as a
cemented splice, is very satisfactory, however, and more convenient,
and takes less time for putting on and taldng off belts for the purpose
of testing and tightening on the belt bench. A belt joined by a
cemented splice must be tested and spliced in position, which is not
so convenient as on the belt bench, especially in the case of over-
head belts. Even where cemented splices are used the belt bench is
convenient for cutting new belts, or re-tightening to a length giving
the proper tension, and for repairs. Only one wire joint is used
in any belt, splices being made if a section becomes damaged so that
a new piece must be set in. The average belt if cared for under this
system will last from six to eight years.
17 The tension on a new belt falls very rapidly, and our present
practice is to tighten it after 24 hours, then in 48 hours, then in one
week, then in two weeks, and so on doubling the length of intervals
until it gets to three months; from this point we must ascertain by
trial for each belt how much greater the intervals may be. This of
course depends upon the severity of service the belt is called upon
to perform as well as the quality of the belt.
Prof. Wm. S. Aldrich. In the first place, the academic discus-
sion of the constancy of the sum of the belt tensions under all loads
is finally set at rest. Now that we really know what is what, by the
invaluable series of experiments referred to, the wonder is that this
fallacy of the constancy of the sum of the belt tensions is so per-
sistent.
TRANSMISSION OF POWER BY LEATHER HELTING 97
2 It was doubt of this position that led the writer to analyze
for himself the experiments on belting then available, those of
Wilfred Lewis and J. S. Bancroft, undertaken for Wm. Sellei-s & Co.,
and of Professor Lanza, of the Massachusetts Institute of Technology.
Both of these were recorded in papers read before the Society, and
published in Vol. 7 of the Transactions. It is remarkable that these
classic experiments have been before the world thus long, and vet so
little studied and respected, and, as far as the writer is awar(>, have
not been superseded by experiments in their special field with more
modern apparatus. Until they are superseded, Mr. Barth's con-
clusions must stand, a remarkable instance of the deductive reasoning
by which it would seem that engineering progress must be made.
3 On the other hand, Mr. Barth has built up, in characteristic
fashion, from theoretical considerations more or less influenced by a
knowledge of the phenomena of belt-transmission, combined with the
physical properties of belting, certain new and helpful relations that
must govern in the future. Such is his " new theorem of the relations
of the tension in a belt," that " under any variation of the effective
pull of a belt, the sum of the square roots of the tensions in the two
strands remains constant, as against the old fallacious supposition
that the sum of these tensions remains constant." (Appendix, Par.
26). Therefore,
V7\+ VT^=2VT^ (1)
4 Now, if we can obtain a similar relation for the difference of the
square roots of the tensions; then we shall have at once, by the usual
formula for the product of the sum and difference of two quantities,
the difference of their squares; that is, in this case, the difference of
the squares of the square roots of the tensions, which is the difference
of the tensions, or the pulling power sought.
5 This much needed "difference of square roots of tensions" has
been indicated by Mr. Barth (Appendix Par. 14), "on the strength
of the experiments made by Mr. Lewis and himself, namely, that
within the limits of ordinary working tensions of a belt, the differ-
ence between the lengths of a belt at different tensions is proportional
to the difference between the square roots of those tensions." We
thus have,
L,-L,= K {VT, - V¥,) (2)
in which K is a constant, dependent upon the material of the belt,
and determined by experiment on the belt.
98 DISCUSSION
6 Combining with Equation 1, we have, as abeady indicated
T,-T, = 2 Vt, (L, -L,) ^ (3)
It seems to the writer that this might possibly be a helpful deduc-
tion, though it may be without much practical appUcation; so that
knowing the initial unit tension T^ and the lengths of the belt under
the tensions T^ and Tj* together with the constant K, its pulling
power {Ti — T^) is known. It seems, therefore, necessary to know
the difference in the lengths of the belt, due to differences in the
belt tensions, that is, to the different driving powers under which it
is expected to operate the belt, or in other words, to calibrate the
belt-performance for this use.
7 It may be remarked, in passing, that the constant K is to be
found from the experiments of Mr. Lewis, as analyzed by Mr. Barth
(Appendix, Equation 3),
^. = ^(1 + 864' W
in which L^ equals the length of belt under the unit tension t when its
slack length is L. From this, by analogy with the above Equation
2, we have,
«-8^ (^>
8 It will no doubt appear that the writer is still inclined to let
the arc of contact and the coefficient of friction of belts take care of
themselves, notwithstanding the keen discussion that has centered
about the fourth conclusion in his paper, referred to by Mr. Barth;
namely,
" (4) The ratio of the tensions of a belt-transmitting power
cannot be calculated with any degree of accuracy by
means of the well-known belt formula:
T <f>a
7f=e (6)
involving the arc of contact a and the coefficient of fric
tion 4>:'
9 This relation is no doubt a guide and a help, indicating the
way the ratios of belt tensions are most likely to be involved. But
TRANSMISSION OF POWER BY LEATHER BELTING 99
it certainly requires a radical modification to adapt it to any reliable
use in predetermining the ratio of tensions for lacing up belts for
given pulling power. Mr. Barth has wrought out these modifica-
tions with excellent results, judged by the adaptabiUty of his slide
rules, and the closeness of approximation to actual conditions (within
the limits assigned) of the assumptions upon which they are based;
namely (Par. 44), "that for the driving belt of a machine the mini-
mum initial tension must be such that when the belt is doing the maxi-
mum amount of work intended, the sum of the tension on the tight
side of the belt and one-half the tension on the slack side will equal 240 lb.
-per square inch of cross section for all belt speeds; and that for a belt
driving a countershaft, or any other belt inconvenient to get at
for re-tightening or more readily made of liberal dimensions, this
sum will equal 160 lb."
10 Here, then, is a definite acceptance of things as they are, and
a straightforward assumption involving additive relations of belt
tensions of leather belting, as it is made and used, and conformable to
experience, rather than their ratios agreeable to theoretical formula,
involving coefficient of friction and arc of contact. This latter rela-
tion (6) is as elusive as the traction-coefficient in railroad work; and
engineers probably will have their own opinions about each until
some genius can predetermine what coefficients of friction are to be
expected in every instance, and so properly introduce the friction
for dynamic conditions into a formula based entirely upon a con-
sideration of statical relations.
The Author. In reading the unexpectedly numerous discussions
of this paper, the author is pleased to note the general appreciation
of it as a contribution to the literature of its kind, but regrets the
assumption by two or three of the discussors that he considers the
paper final in its application of the theories developed. All that is
claimed is that he has taken practical advantage of the experimental
data at his disposal, and has taken the pains to do mathematical jus-
tice to them, deriving therefrom excellent results in the scientific
running of machine tools whose belts have been tightened and
worked according to the rules thus established.
2 While the author feels guilty, therefore, of narrowing to a con-
siderable extent the scope within which Professor Benjamin's imagi-
nation may still run rampant, so far as the behavior of a leather belt
goes, he fully agrees that further experiments are needed in order to
determine the coefficient of friction under all the variable conditions
100 DISCUSSION
under which belts are called on to drive; and yet more, in order to
settle conclusively whether the coefficient of friction is a function of
the velocity of slip, as he has assumed, or of the percentage of slip,
as indicated by Mr. Hamerstadt's study of numerous experiments at
different belt speeds, though the latter seems contrary to the mechan-
ical principles involved in the phenomenon of slip.
3 For the special benefit of Professor Bird, the writer will even
say, that while all he knows about belting could probably be reduced
to a pamphlet three times the size of his paper, a good-sized volume
would probably be required to hold all he does not know but would
like to know about belting, and a small library would be required
to record all he does not care a straw to know about the subject.
4 But while the writer agrees with Mr. Hamerstadt as to the
desirability of further experiments and will look forward to these
with the keenest interest, he does not see the force of his argument
about the necessary overload capacity of a high-speed belt, on a
motor with an overload capacity; surely we need only make the belt
b'g enough to take care of the overload as a normal load, and be
satisfied to have it unnecessarily large for the rated capacity of the
motor; just as a bridge intended for a light normal load must still be
made strong enough for any anticipated occasional extra load.
Trouble arises only when we do not know how to design a bridge
properly, or v.hen we get an occasional extra load which we have had
no reason to anticipate.
5 Though the writer had not expected to be forced to express
himself on the question of belt-drives versus electric motor drives, he
will say, in view of Mr. Robbins' remarks, that he believes that during
the past decade hundreds of thousands of dollars, if not millions, have
been more than wasted by the substitution of motor-drives for belt
drives. Such a change has often been advantageous, of course, and
is occasionally recommended to his clients by the writer: the trouble
has been that the enormous investments of electrical manufacturing
establishments have forced the electrical salesman more than any
other to create a demand for his product, so that not only has he
allowed his enthusiasm to run away with him, but he frequently has
recommended his product against his own biased judgment; persuad-
ing the incompetent shop manager or superintendent to accept his
product as a remedy for a small output that is in reality due to a
complication of causes that could be cured only by the application of
a number of remedial measures.
6 The writer believes, however, that a reaction against this indis-
TRANSMISSION OF POWER BY LEATHER BELTING 101
criminate electrification of machine shops has already set in. aside from
the influence of the industrial depression, and that the electric drive
will be installed, in the near future, only when conditions make it
unquestionably more advantageous than the belt-drive.
7 As touched upon by Mr. Van Derhoef, the elastic properties of
transmission-rope are probably similar to those of leather belts, and
it seems to be in order for someone to ascertain them by the neces-
sary experiments, and subsequently to apply this knowledge by use
of the writer's methods.
8 The writer values Mr. Allen's statements of the advantages
derived from the adoption of the Taylor system, as introduced by the
author in the Yale & Towne Mfg. Co. 's plant, where 4800 belts are thus
taken care of. Mr. Allen, in conjunction with Messrs. Tajdor, Hath-
away and Waldron, has thus supplemented the scant attention paid
in the paper to the aspect of the subject most practically important.
The reason for this omission is that in his work with belting the
writer has derived by far the greatest personal satisfaction from the
solution of the mathematical problems involved, and he has been
unable to eliminate entirely the personal interest.
9 It is not possible to answer here Mr. Gary's question as to how
to estimate the horse-power required to drive each machine in a
large plant, but the writer will be pleased to give him, and anybody
else interested enough to pay a visit to Philadelphia, an idea of how it
is done, by means of slide-rules especially constructed for the purpose.
10 As a further answer to Professor Bird's various statements
and questions, the writer will only say that on a more careful reading
of the paper, as well as the Appendix and Supplement, he will find
most of them answered. For instance, the most valuable mathemat-
ical developments in the Appendix and Supplement answer the
question why the sum of the tensions of a belt increases with the
load; and study of this will help him to formulate for himself an
answer to his non-mathematical questioners. As to the effect of
crownings a mall pulley, the writer heartily agrees with Professor
Bird in a general way, though surprised to note with what confi-
dence the latter estimates the difference between the tensions in the
middle and edge-fibers of a belt running over such a pulley.
11. The author is sorry that the considerable trouble to which
Mr. Nagle has gone to make comparisons with his own earlier formulae
for the horse-power transmitted by leather belting, is based on a mis-
understanding of the fundamental basis of the author's work.
12 As stated in the paper, the author bases his figures on a certain
102 DISCUSSION
tension per square inch of belt, independent both of the strength of
the belt itself and its thickness, and of the strength of the lacing,
except that the latter must be in excess of the maximum tension
brought to bear on a belt while delivering power. The author, there-
fore, makes no distinction between a single and a double belt, but
merely considers the tension per square inch of section, as it has not
been definitely proven that the coefficient of friction depends materi-
ally on the area of the surface presented by the belt against the pulley.
13 As Mr. Nagle somehow has assumed that the two horse-power
curves in Fig. 3 are meant respectively for a single and a double belt,
whereas they stand for something totally different, it unfortunately
follows that the comparisons made by him of his own and the author's
ideas as to what power a belt should be counted on to transmit, have
completely miscarried.
14 Mr. Nagle says that we cannot decide upon the coefficient of
friction by formula. This is unquestionably so, but it is also true
that, having roughly decided, by one means or another, what we wish
to count on as the coefficient of friction at any one velocity of a belt,
we may to great advantage make an empirical formula to represent a
perfectly-graded change of this coefficient with the velocity; and only
by so doing can we effect a mathematical solution of the belt problem
that is an improvement on the unquestionably wrong assumption of a
coefficient independent of the velocity of the belt, such as 0.42, origin-
ally recommended by Mr. Towne, or 0.28, recommended by the late
Professor Ruleaux.
15 The author fully agrees with Mr. Nagle that " a new belt on an
iron pulley may not have a coefficient of friction of as much as 0.25,
while the same belt, well worn and well groomed, will give 0.65 in a
clean, dry room;" and, more than that, knows that this elusive quan-
tity will vary all the way from almost 0 to 1.50. However, just
because the author is a practical and practicing engineer, though very
fond of a little pure mathematics in the handling of practical engineer-
ing problems, he has adopted something as a standard, this something
being a variable lying happily between the great extremes, instead of
being merely a single average between the extreme values.
16 The author is not at all disappointed because a perfectly new
belt will not give the output required, at its minimum tension, without
the resort to a temporary application of some good adhesion-produc-
ing belt-dressing; nor on the other hand, when a "well worn and
groomed" belt at times is capable of giving the output required, at a
little less initial tension than the one he aims at maintaining by the
TRANSMISSION OF POWER BY LEATHER BELTING 103
means more fully described in the discussions submitted by Mr. Allen
and Mr. Hathaway.
17 Mr. Nagle also remarks that we rarely know the exact power to
be transmitted except in the case of prime movers, which no doubt
is true, so far as the work of most engineers is concerned. However,
in the author's practice at least, the maximum output of every belt
put up on any machine is known; simply because he personally sets
the limit, and has means of seeing that the same is never exceeded.
18 Mr. Nagle refers to his paper read in 1881 as the first one to
recognize the effect of centrifugal force in a belt. A correct formula,
however, for the loss of effective tension in a belt, due to its centrifugal
force, was given by Weisbach, at a much earlier date. This fact does
not detract, of course, from the value of Mr. Nagle's paper, in which,
probably for the first time, this matter was presented in a manner that
made it readily available to the busy, practical engineer.
19 As regards Mr. Nagle's suggestion that the data have not been
presented in a sufficiently handy form for the busy engineer, the
author believes Mr. Nagle has failed to appreciate the slide-rule illus-
trated in Fig. 5, which contains these data in a form which for handi-
ness leaves tables and diagrams far behind, while he at the same
time is not ready to admit that there is anything the matter with the
various diagrams that give the same data.
SAFETY VALVES
No. 1231
SAFETY VALVES FOR LOCOMOTIVES
By Frederic M. Whyi , New York
Member of the Society
It is the purpose of this paper to present some ideas about safety
valves for steam boilers and particularly for locomotive boilers.
2 Just how the capacity of the first valve used on a steam boiler
was determined, or what relation this capacity had to the generat-
ing capacity of the boiler, may be recorded somewhere in history, but
it is doubtful if either fact was recorded or even determined. So far
as locomotive work is concerned, the same ignorance prevails todav;
but there is good promise that tliis ignorance will soon be dispelled.
In marine work certain formulae have been devised for calculating
the sizes of safety valves, and these formulae have been accepted,
more or less blindly, it is thought.
3 The general practice in locomotive work has been to determine
in an " offhand" way the size and number of safety valves to be used,
and former practice has guided the determination entirely. If a
larger boiler is to be used the valve capacity may not be increased,
depending upon the judgment of the person whose duty it is to deter-
mine the capacity. Again, the capacity has been indicated in an
indifferent manner, being expressed as a "size," meaning the diam-
eter of something more or less uncertain; while the other dimension,
the lift, which is necessary to give an indication of the capacity, is
entirely ignored.
4 But to know the exact capacity of the available valves is not
sufficient; it is quite as important to know how much steam is to be
released and in what length of time it should be released. It will
be comparatively easy to determine the capacity of safety valves, if
indeed the elaborate tests which have already been made — data
Presented at the New York monthly meeting (February 1909) of The
American Society of Mechanical Engineers.
106 SAFETY VALVES FOR LOCOMOTIVES
from which it is hoped may be presented in the discussion^ — have
not already solved part of the problem; more difficult will be that part
of it which is concerned with the quantity of steam to be released and
the rate of the release. The subject is of mutual interest to the valve
manufacturer and the user, — the design of the valve for capacity and
wear to be worked out by the manufacturer; and the capacity which
is to be used, both in volume and in number of valves, and the rate of
release, to be determined by the user with the assistance of the manu-
facturer.
ESSENTIALS OF A SAFETY VALVE ON A LOCOMOTIVE
5 The design of the valve will include the diameter of the con-
trolling opening and the passages leading to it from the steam volume,
as well as those leading from it to the atmosphere, the shape and
material of the seat, the amount of lift of the valve, and the shape
and material of the valve face, the spring and its protection, the
adjustment, the muffler, if one is to be used, and the action of the
valve in lifting and in seating.
6 It will not be necessary to discuss the diameter of the control-
ling opening, and of the passages to and from it, in view of the sugges-
tion here made that instead of indicating the capacity of a valve in
a very rough way by the diameter of some opening, the capacity be
expressed in pounds of steam at certain pressures. The shape of the
seat and of the valve face may or may not be of importance; but this
will be referred to again. The material in the seat and face will
naturally be that which will best withstand the effects of the flow of
steam over them, and the possible pound of the valve when seating.
7 The reliability of the spring and the effect of heat upon it are
very important points. Adjustments should be readily made, but
on the other hand to get out of adjustment should be practically
impossible. The capacity of the muffler need not be questioned,
except in extreme designs, but the indicated capacity should be that
of the valve complete, with or without muffler, according to the in-
tended use of the valve; then it is important only that it deaden suffl-
ciently the noise of the escaping steam.
8 The action of the valve in lifting and in seating, the desirability
of a forewarning that the maximum pressure is about reached,
and the operating conditions which bear upon this question of fore-
' These data are given in the paper "Safety Valve Capacity" which follows.
They were presented as a discussion'and afterwards published in The Journal as
a paper under the above title. — Editor.
SAFETY VALVES FOR LOCOMOTIVES 107
warning, are correlative. With any Idnd of steam-generating plant
it ought to be quite sufficient if those immediately responsible for the
quantity produced, and for its use, know what is available; in station-
ary and in marine work this is generally true, and steam gages can be
placed within view of those who should know what the pressure is at
any time. Unfortunately in locomotive work, however, it has become
perhaps desirable that others than those within view of the gage in
the cab know something about the steam pressure, and inasmuch as
the fireman is wilUng, and sometimes anxious, that they should know,
he takes the only means at hand to inform them when he thinks that
the results of his labors are good, and "fires against the pop" so that
everybody within hearing or sight of the valve knows by the escaping
steam that the fireman is doing his duty. If when a train is ascend-
ing a grade the conductor at the rear sees steam escaping from the
valve he knows the train will get up the grade; on hard grades he
will watch for the only indication which can be given him, and the
fireman tries to present this indicator, the escape from the valve,
the "white feather."
9 Numerous similar examples might be mentioned, but assum-
ing that such an indicator of steaming conditions has grown to be a
necessity, undesirable as it may be, how can it be produced at the
least expense? Surely not with a valve from 2^ to 4 in. in diameter
and open to its full capacity. Two devices, at least, are available
to give the indication at a lower cost than the full open valve: the
"simmering" valve, which will open slightly for two or three pounds
about the normal maximum, then open full, and just reversing this
order in seating; the other, the small pilot valve, which will open at two
or three pounds pressure below the working valve. The first method
will have some bearing on the kind of metal to be used in the valve
seat and valve face and possibly upon the shape of the exterior edge
of the valve and the opposing surface of the seat. The second
method means the addition of the small valve, an additional cost for
which there will be no need if the first method can be developed
successfully.
RELATION OP VALVE CAPACITY TO STEAM-GENERATING CAPACITY
10 There remains for consideration the relation of valve capacity
to steam-generating capacity, and the unit capacity of the valves
which will make up the total valve capacity. The fact that in loco-
motive work the total valve capacity has not been as great as the
maximum steam-generating capacity should be ample proof that
108 SAFETY VALVES FOR LOCOMOTIVES
such valve capacity is not necessary. The reason for this is,
of course, that on account of using the exhaust steam for producing
the forced draft, when the demand for steam from the boiler is reduced
or entirely cut off, the forced draft is automatically reduced or cut off,
and the generating capacity is reduced so that it is not necessary that
the safety valves release the full generating capacity. Probably it
will be largely a question of opinion what per cent of the total generat-
ing capacity the valve ought to have, although it is possible that as
attention is centered upon this question some more or less positive
solution of it may result.
11 Having fixed upon the per cent of the generating^capacity to
be provided for in the valves it will be necessary to determine the
desirable unit capacity of the valves. Some States require that each
locomotive boiler have at least two valves. Starting with this condi-
tion, consideration of the maintenance of the valves indicates that
they should be duplicates and therefore that each have a capacity
equal to one-half the required generating capacity. If a number of
boilers of different capacities are to be considered, then the smaller
ones would probably be provided with the same valves as the larger
ones for the purpose of duplication. There are some large boilers
for which three valves might be necessary, because the necessary
capacity in two units might make the valves abnormally large for
construction purposes. Also it is worth while to consider whether
undesirable results would come about from opening almost instan-
taneously an escape of steam from the boiler to the atmosphere. No
suggestions are offered on this, but it is hoped that something bear-
ing on the question may be developed in the discussion.
12 It is suggested that instead of setting the several valves on
a boiler at different pressures, all the valves be set at the same
pressure, with the idea that each of them will operate frequently
enough to keep all in working condition, rather than run the risk
of one valve being found out of condition when it is required for
action.
13 It is probable that some time it will be found desirable to
consider the minimum distance above the steam releasing surface of
the water at which the safety valve seat may be placed.
No. 1232
SAFETY VALVE CAPACITY
By Philip G. Dakung, New York
Associate Member of the Society
The function of a safety valve is to prevent the pressure in the
boiler to which it is applied from rising above a definite point, to do
this automatically and under the most severe conditions which can
arise in service. For this, the valve or valves must have a reliev-
ing capacity at least equal to the boiler evaporation under these
conditions. If it has not this capacity, the boiler pressure will continue
to rise, although the valve is blowing, with a strain to the boiler
and danger of explosion consequent to over-pressure. Thus, with
the exception of the requisite mechanical reliability, the factor in a
safety valve bearing the most vital relation to its real safety is its
capacity.
2 It is the purpose of this paper to show an apparatus and method
employed to determine safety valve lifts, giving the results of tests
made with this apparatus upon different valves; to analyze a few of
the existing rules or statutes governing valve size; and to propose a
rule, giving the results of a series of direct capacity tests upon which
it is based; to indicate its application to special requirements; and
finally its general bearing upon valve specifications.
3 Two factors in a safety valve geometrically determine the area
of discharge and hence the relieving capacity: — the diameter of the
inlet opening at the seat and the valve lift. The former is the
nominal valve size, the latter is the amount the valve disc lifts verti-
cally from the seat when in action. In calculating the sizes of valves
to be placed on boilers, rules which do not include a term for this valve
lift, or an equivalent, such as a term for the effective area of discharge,
assurne in their derivation a lift for each size of valve. Nearly all
existing rules and formulae are of this kind, which rate all valves of
a given nominal size as of the same capacity.
4 To find what lifts standard make valves actually have in prac-
Presented at the New York monthly meeting (February 1909) of The
American Society of Mechanical Engineers.
110
SAFETY VALVE CAPACITY
Fixed Center Shaft
Driven by a siiuiU Motor
Comicction Tapped iiito
dilt'erent Places in Valve
Case Exhaust Pipe etc.
to determine Back Pressure
Connection to Boiler
ConiMctetl to Boiler
as in.SerN-ice
Fig. 1 Safety Valve Lift-Recording Apparatus
SAFETY VALVE CAPACITY 111
tice, and thus test the truth or error of this assumption that they are
approximately the "same for the 'same size of valve, an apparatus has
been devised and tests conducted upon different makes of valves.
With this apparatus not only can the valve Hft be read at any moment
to thousandths of an inch, but an exact permanent record of the lift
during the blowing of the valve is obtained, somewhat similar to a
steam engine indicator card in appearance and of a quite similar use
and value in analyzing^the action of the valve.
5 As appears in Fig. 1 the valve under test is mounted upon the
boiler in the regular manner, and a small rod is tapped into the top
end of its spindle, which rod connects the lifting parts of the valve
directly with a circular micrometer gage, the reading hand of which
indicates the lift upon a large circular scale or dial. The rod through
this gage case is solid, maintaining a direct connection to the pencil
movement of the recording gage above. This gage is a modified
Edson recording gage with a multiplication in the pencil movement
of about 8 to 1, and with the chart drum driven by an electric motor
of different speeds, giving a horizontal time element to the record.
The steam pressures are noted and read from a large test gage gradu-
ated in pounds per square inch, and an electric spark device makes
it possible to spot the chart at any moment, which is done as the
different pound pressures during the blowing of the valve are reached.
The actual lift equivalents of the pencil heights upon the chart are
carefully calibrated so that the record may be accurately measured
to thousandths of an inch.
6 In testing the motor driving, the paper drum is started and
the pressure in the boiler raised. The valve, being mounted directly
upon the boiler, then pops, blows down and closes under the exact
conditions of service, the pencil recording on the chart the history
of its action.
7 With this apparatus, investigations and tests were started upon
seven different makes of 4-in. stationary safety valves, followed by
similar tests upon nine makes of muffler locomotive valves, six of
which were 3^ in., all of the valves being designed for and tested at
200 lb. The stationary valve tests were made upon a 94-h.p. water-
tube boiler made by the Babcock & Wilcox Co. (See Fig. 2.) The
locomotive valve tests were made upon locomotive No. 900 of the
Illinois Central R. R., the valve being mounted directly upon the
top of the main steam dome. (See Figs. 4 and 5.) This locomo-
tive is a consoUdation type, having 50 sq. ft. of grate area and
2953 sq. ft. of heating surface. Although a great amount of addi-
112
SAFETY VALVE CAPACITY
Fig. 2 Valve-Lift Apparatus as Used with^the Stationary Tdst Boiler
AT Bridgeport, Conn.
SAFETY VALVE CAPACITY
113
tional experimenting has been done, only the results of the above tests
will be quoted here. These lift records show (with the exception of
a small preliminary simmer, which some of the valves have) an abrupt
opening to full lift and an almost equally abrupt closing when a certain
lower lift is reached. Both the opening and closing lifts are signifi-
cant of the action of the valves. (See Fig. 3.)
8 The results of the 4-in. iron body stationary valve tests sum-
marized are as follows: of the seven valves the average lift at open-
ing was 0.079 in. and at closing 0.044 in., or excluding the valve with
the highest lifts, the averages were 0.07 in. at opening and 0.037 in.
at closing. The valve with the lowest Ufts had 0.031 in. at opening
and 0.017 in. at closing, while that with the highest had 0.137 in.
and 0.088 in. Expressing the effective steam-discharge areas of the
r::
Fig. 3 Typical Valve-Lift Diagrams
valves taken at their opening'Jlifts as percentages of the largest
obtained, the smallest had 31.4 per cent, the next larger 40.8 per cent,
and the next 46.6 per cent. Of the six 3^-1^. muffler locomotive
valves the summarized hfts are as follows: average of the six valves,
0.074 in. at opening and 0.043 in. at closing. Average, excluding
the highest, 0.061 in. at opening and 0.031 in. at closing. The lowest
Uft valve had 0.04 in. opening and 0.023 in. closing; the highest,
0.140 in. opening and 0.102 in. closing. As percentages of the largest
effective steam-discharge area, the smallest was 36.4 per cent, the
next larger 39.8 per cent, and the next 46.4 per cent. In both the
114
SAFETY VALVE CAPACITY
stationary and locomotive tests, the lowest lift valve was flat-seated,
which is allowed for in the above discharge area percentages.
9 The great variation — 300 per cent — in the lifts of these stand-
ard valves of the same size is startling and its real significance is
apparent when it is realized that under existing official safety valve
rules these valves, some of them with less than one-third the lift and
capacity of others, receive the same rating and are listed as of equal
relieving value. Three of these existing rules are given as an illus-
tration of their nature: the United States Supervising Inspectors
Fig. 4 Valve Lift Apparatus as Erected for Locomotive Testing at
BuRNSiDE, III,
Rule, that of the Board of Boiler Rules of Massachusetts, and the
Boiler Inspection Rule of Philadelphia.
BULE OF UNITED STATES BOARD OP SUPERVISING INSPECTORS
W
A - 0.2074 X^=r
A "» area of safety valve in square inches per square foot of grate surface,
ly = pounds of water evaporation per square foot of grate per hour.
P = boiler pressure (absolute).
SAFETY VALVE CAPACITY
115
10 In 1875 a special committee was appointed by the United
States Board of Supervising Inspectors to conduct experiments upon
safety valves at the Washington Navy Yard. Although the pres-
FiG. 3 Detail of Lift Apparatus at Burnside, III., Showing Locomotive
Valve
sures used in these experiments (30 and 70 lb. per square inch) were
too low to make the results of much value today, one of the conclu-
sions reported is significant:
116 SAFETY VALVE CAPACITY
a That the diameter of a safety valve is not an infallible test
of its efficiency.
b That the lift which can be obtained in a safety valve,
other conditions being equal, is a test of its efficiency.
1 1 The present rule of the board, as given above, formulated by
L. D. Lovekin, Chief Engineer of the New York Shipbuilding
Co., was adopted in 1904. Its derivation assumes practically a 45-
deg. seat and a valve lift of 3^ of the nominal valve diameter. The
discharge area in this rule is obtained by multiplying the valve lift
D
— by the valve circumference (tt X D) and taking but 75 per cent of
the result to allow for the added restriction of a 45-deg. over a flat
seat. The 75 per cent equals approximately the sine of 45 deg. or
0.707. This value for the discharge area, i.e.,! 0.75 X tt X ^ ), issubsti-
P
tuted directly into Napier's formula for the flow of steam, w= a ^-.
Thus in the valves to which this rule is applied the following lifts
are assumed to.exist: 1-in. valve, 0.03 in.; 2-in. valve, 0.06 in.; 3-in.
valve, 0.09 in.; 4-in. valve, 0.13 in.; 5-in. valve, 0.16 in.; 6-in. valve,
0.19 in.
\^2 Referring back to the valve lifts as given in Par. 8, it is seen
that the highest lift agrees very closely with the lift assumed for 4-in.
valves in the rule of the Board of Supervising Inspectors. If the
lifts of valves of different design were more uniformly of this value,
or if the rule expressly stipulated either that the Uft of 3V of the valve
diameter actually be obtained in valves qualifying under it; or, if not,
that an equivalent discharge area be obtained by the use of larger
valves, the rule would apply satisfactorily However, the lowest
lift valve actually has but 25% of the lift assumed for the 4-in. valves
in the rule; the next larger less than 50%; while the average lift of
these valves, excluding only the highest, is but 50% of the assumed
lift in the rule for 4-in. valves.
MASSACHUSETTS RULE
, IOX70 ,,
A ^Xll
A = area of safety valve in square inches per sq. ft. of grate surface.
w = pounds of water evaporation per square foot of grate surface per
second.
P = boiler pressure (absolute) *t which valve is set to Wow.
SAFETY VALVE CAPACITY 117
13 One of the most recently issued rules is that contained in the
pamphlet of the new Massachusetts Board of Boiler Rules, dated
March 24, 1908. This rule is merely the United States rule given
above with a 3.2 per cent larger constant and hence requiring a valve
larger by that amount. The evaporation term is expressed in pounds
per second instead of per hour and two constants are given instead of
one, but when reduced to the form of the United States rule it gives
W
A = 0.214 X p. Figuring this back as was done above with the
United States rule, and taking the 75 per cent of the fiat seat area
as there done, this rule assumes a valve lift of jU of the valve diameter
instead of the 3^ of the United States rule. This changing of the
assumed lift from 3^ to ^ of the valve diameter being the only dif-
ference between the two rules, the inadequacy of the United States
rule just referred to applies to this more recent rule of the Massa-
chusetts Board.
PHILADELPHIA RULE
22.5 G
a =
p + 8.62
a = total area of safety valve or valves in square inches.
G "= grate area in square feet.
p = boiler pressure (gage) .
14 The Philadelphia rule now in use came from France in 1868,
where it was the official rule at that time, having been adopted and
recommended to the city of Philadelphia by a specially appointed
committee of the Franklin Institute, although this committee frankly
acknowledged in its report that it "had not found the reasoning upon
wliich the rule had been based," The area a of this rule is the
effective valve opening, or as stated in the Philadelphia ordinance
of July 13, 1868, " the least sectional area for the discharge of steam. "
Hence if this rule were to be appUed as its derivation from the French
requires, the lift of the valve must be known and considered when-
ever it is used. However, the example of its application given in
the ordinance as well as that given in the original report of the
Franklin institute committee, which recommended it, shows the area
a applied to the nominal valve opening. In the light of its derivation,
this method of using it takes as the effective discharge area the
valve opening itself, the error of which is very great. Such use, as
specifically stated in the report of the committee above referred to,
118 SAFETY VALVE CAPACITY
assumes a valve lift at least i of the valve diameter, i.e., the practi-
cally impossible lift of 1 in. in a 4-in. valve. Nevertheless, this is
exactly the method of use indicated in the text of the ordinance.
15 The principal defect of these rules in the light of the preceding
tests is that they assume that valves of the same nominal size have
the same capacity and they rate them the same without distinction,
in spite of the fact that in actual practice some have but J of the capac-
ity of others. There are other defects, as have been shown, such as
varying the assumed lift as the valve diameter, while in reality with
a given design the lifts are more nearly the same in the different sizes,
not varying nearly as rapidly as the diameters. And further than
this, the lifts assumed for the larger valves are nearly double the
actual average obtained in practice.
16 The direct conclusion is this, that existing rules and statutes
are not safe to follow. Some of these rules in use were formulated
before, and have not been modified since, spring safety valves were
invented, and at a time when 120 lb. was considered high pressure.
None of these rules takes account of the different lifts which exist in
the different makes of valves of the same nominal size, and they
thus rate exactly aUke valves which actually vary in lift and reUev-
ing capacity over 300 per cent. It would therefore seem the duty
of all who are responsible for steam installation and operation to
leave the determination of safety valve size and selection no longer
to such statutes as may happen to exist in their territory, but to
investigate for themselves.
17 The elements of a better rule for determining safety valve size
exist in Napier's formula for the flow of steam, combined with the
actual discharge area of the valve as determined by its lift. In
Steam Boilers, by Peabody & Miller, this method of determining
the discharge of a safety valve is used. The uncertainty of the
coefficient of flow, that is, of the constant to be used in Napier's
formula when applied to the irregular steam discharge passages of
safety valves, has probably been largely responsible for the fact that
this method of obtaining valve capacities has not been more
generally used. To determine what this constant or coefficient of
flow is, and how it is affected by variations in valve design and
adjustment, an extended series of tests has recently been conducted
by the writer at the Stirling Department of the Babcock & Wilcox Co.,
at Barberton, Ohio.
18 A 373-h.p. class K, No. 20 Stirling boiler, fired with a Stirhng
chain grate, with a total grate area of 101 sq. ft., was used. This
SAFETY VALVE CAPACITY
119
boiler contained a U-type superheater designed for a superheat of
50 deg. fahr. The water feed to this boiler was measured in caHbrated
tanks and pumped (steam for the pump being furnished from another
boiler) through a pipe line which had been blanked wherever it was
impossible, with stop valves and intermediate open drips, to insure
against^ any leakage. The entire steam discharge from the boiler
was through the valve being tested, all other steam connections from
the boiler being either blanked or closed with stop valves beyond which
were placed open drip connections to indicate any leakage. A constant
H^.
Fig. 6 Akrangement of Valve with Micrometer Spindle Used in the
Direct-Capacity Testing at Barberton, Ohio
watch was kept throughout the testing upon all points of the feed
and steam lines, to insure that all water measured in the calibrated
tanks was passing through the tested valves without intermediate
loss.
19 The valves tested consisted of 3-in., 3^-in. and 4-in. iron
stationary valves, and l^in., 3-in. and 3^-in. locomotive valves,
the latter with and without mufflers. These six valves were all
previously tested and adjusted on steam. Without changing the
position of the valve disc and ring, the springs of these valves were
120 SAFETY VALVE CAPACITY
then removed and solid spindles, threaded (with a 10-pitch thread)
through the valve casing above, inserted. Upon the top end of these
spindles, wheels graduated with 100 divisions were placed. Fig. 6
shows the arrangement used with the locomotive valves, the spindle
and graduated wheel being similar to that used with the stationary
valves. By this means the valve lift to thousandths of an inch was
definitely set for each test and the necessity for constant valve lift
readings, with that source of error, eliminated.
20 In conducting the tests three hours' duration was selected as
the minimum time for satisfactory results. Pressure and tempera-
ture readings were taken every three minutes, water readings every
half hour. A man stationed at the water glass regulated the feed
to the boiler to maintain the same level in the boiler during the test;
other men were stationed, one at the water tanks, one firing and one
taldng the pressure and temperature readings. Pressure readings
were taken from two test gages connected about 4 in. below the
valve inlet, the gages being calibrated both before and after the series
of tests was run and corrections applied. In all 29 tests were run,
of which 15 were 3 hours long, 4 were 2^ hours, 3 were 2 hours, and
7 of less duration.
21 Tests numbered 1 to 5 were preliminary runs of but one hour
or less duration^ apiece, and records of them are thus omitted in
Table 1, which gives the lifts, discharge areas, average pressure and
superheat, and the steam discharge in pounds per hour of each
of the other tests. The discharge areas in the valves with 45-deg.
seats are given by the following formula which is easily derived
geometrically :
a= 2.22 X D X I + 1.11 X J?
where
a = effective area in square inches
D= valve diameter in inches
^= valve Hft in inches
In tests 8 and 23, where the width of valve seat was 0.225 in. and
0.185 in. respectively, and the valve was thus slightly above the
depth of the valve seat, the area was figured for this condition.
22 As previously stated, the appHcation of these results is in
fixing a constant for the flow of Napier's formula as applied to
p
sarety valves. The formula is w =a - in which w equals pounds
discharged per second, P equals the absolute steam pressure behind
SAFETY VALVE CAPACITY 121
the orifice or under the valve and a equals the effective discharge
opening in square inches. This may be stated as E = C X a X P',
in which E equals the pounds steam discharge per hour and C equals
a constant. The values of E, a and P being given for the above
tests, C is directly obtainable.
23 Figuring and plotting the values of this constant indicate the
following conclusions:
a Increasing or altering the steam pressure from approxi-
mately 50 to 150 lb. per square inch (tests 14 and 10)
does not affect the constant, this merely checking the
applicability of Napier's formula in that respect.
b Radically changing the shape of the valve disc outside
of the seat at the huddling or throttling chamber, so-called,
does not affect the constant or discharge. In test 15
the valve had a downward projecting lip (as in Fig. 1),
deflecting the steam flow through nearly 90 deg., yet the
discharge was practically the same as in tests 10 and 14,
where the lip was cut entirely away (as in Fig. 6), giving
a comparatively unobstructed flow to the discharging
steam.
c Moving the valve adjusting ring through much 'more than
its complete adjustment range does not affect the con-
stant or discharge. (Tests 16 and 17.)
d The addition of the muflBer to a locomotive valve does
not materially alter the constant or discharge. There is
but 2 per cent difference between tests 10 and 13.
e Disregarding the rather unsatisfactory IJ-in. and 3-in.
locomotive valve tests, the different sizes of valves
tested show a variation in the constant of about 4 per
cent when plotted to given lifts.
/There is a slight uniform decrease of the constant when
increasing the valve lifts.
24 The variations indicated in the last two conditions are not large
enough, however, to impair materially the value of a single constant
obtained by averaging the constants of all the 24 tests given. The
selection of such a constant is obviously in accord with the other four
conditions mentioned. This average constant is 47.5, giving as the
formula E = 47.5 X a X P. Its theoretical value for the standard
orifice of Napier's formula is 51.4, of which the above is 92^ per cent.
25 To make this formula more generally serviceable, it should
be expressed in terms of the valve diameter and lift, and can be still
122 SAFETY VALVE CAPACITY
further simplified in its application by expressing the term E (steam
discharged or boiler evaporation per hour) in terms of the boiler heat-
ing surface or grate area. For the almost universal 45-deg. seat the
effective discharge area is, . with a slight approximation, L X sine
45 X TT X I), in which L equals the valve lift vertically in inches and
D the valve diameter in inches. Substituting this in the above
formula gives E = 47.5 X Lx sine 45 X tt X Z) X P, or £/ = 105.5
X L X D X P.
26 The slight mathematical approximation referred to consists
in multiplying the {L X sine 45) by {n X D) instead of by the exact
value {k X D -\- ^L) . To find directly the effect of this approxima-
tion upon the above constant, the values for E, L, D and P from the
tests have been substituted into the above formula and the average
constant re-determined, which is 108.1. The average lift of all the
tests is 0.111 in. Plotting the constants obtained from the above
formula in each test, as ordinates, to valve Hfts, as abscissae, obtain-
ing thus the slight inclination referred to in Par. 23 /, and plotting
a fine with this inclination through the above obtained average con-
stant 108.1, taken at the 0.111-in. average lift, gives a line which at
a maximum lift of say 0.14 in. gives a constant of just 105. At
lower lifts this is slightly larger. Hence 105 would seem to be the
conservative figure to adopt, as a constant in this formula for general
use, giving
E= 105 XLXDXP
This transposed for D gives:
pi
D = 0.0095 X i^^^p
Note that the nominal valve area does not enter into the use of this
formula and that if a value of 12, for instance, is obtained for D it
will call for two 6-in. or three 4-in. valves. For flat seats these con-
stants become 149 and 0.0067 respectively.
27 The fact that these tests were run with some superheat (an
average of 37.2 deg. fahr.) while the majority of valves are used
with saturated steam, would, if any material difference exists, place
the above constants on the safe side. The capacities of the stationary
and locomotive valves, the lift test results of which are summarized
in Par. 8, have been figured from this formula, taking the valve lifts
at opening, and in pounds of steam per hour are as follows:
SAFETY VALVE CAPACITY 123
Of the seven 4-in. iron body stationary valves, the average
capacity at 200 lb. pressure is 7370 lb. per hour, the
smallest capacity valve (figured for a flat seat) has a
capacity of 3960 lb., the largest 12,400 lb., and of the
six 3^-in. muffler locomotive valves at 200 lb. pressure,
the average capacity is 6060 lb. per hour, the smallest
4020 lb., the largest 11,050 lb.
28 To make the use of the rule more direct, where the evaporation
of the boiler is only indirectly known, it may be expressed in terms
of the boiler heating surface or grate area. This modification con-
sists merely in substituting for the term E (pounds of total evapora-
tion) a term H (square feet of total heating surface) multiphed by
the pounds of water per square foot of heating surface which the
boiler will evaporate. Evidently the value of these modified forms
of the formula depends upon the proper selection of average boiler
evaporation figures for different types of boilers and also upon the
possibility of so grouping these boiler types that average figures
can be thus selected. This modified form of the formula is
D=CX ^
L X P
in which H equals the total boiler heating surface in square feet and C
equals a constant.
29 Values of the constant for different types of boilers and of
service have been selected. These constants are susceptible of
course to endless discussion among manufacturers, and it is undoubt-
edly more satisfactory, where any question arises, to use the form
containing term E itself. Nevertheless the form containing the
term H is more direct in its application and it is beheved that the
values given below for the constant will prove serviceable. In apply-
ing the formula in this form rather than the original one, containing
the evaporation term E, it should be remembered that these con-
stants are based upon average proportions and hence should not be
used for boilers in which any abnormal proportions or relations
between grate area, heating surface, etc., exist.
30 For cylindrical multi-tubular, vertical and water-tube station-
ary boilers a constant of 0.068 is suggested. This is based upon an
average evaporation of 3^ lb. of water per square foot of heating
surface per hour, with an overload capacity of 100 per cent, giving
7 lb. per square foot of heating surface, the figure used in obtaining
the above constant.
124
SAFETY VALVE CAPACITY
02 c^
a
m
H
O
><
.
H
<
w
(U
n
<
a
n
<
H
«
>
•
<
d^
I>
X
'^
J
g
^
<5
4
r/)
M
u
f-i
0
W
n
hJ
■<
m
«
<j
n
H
-
•o
(U
A
a>
t~
•-; "p
£
B
0]
o
>
■3
>
_>
<B
>
"O *hn
<
a
H
i
IN
CO
£
p
3
£
£
p
>
o
S
o
o
o
H-1
0
e
0
0
M
'^
a
si
0
"3
>
"o
<o
0.
>>
H
3
m
£^
"o
c
.2
CO c
£ -B
p •'~
m 0
£ Oh
O ci d
ci
u
c9
S
a
O
"5
3
(D « 0
lis
0
<u
a.
lU
eu to
iz; « P5
_i9
• —
"—
• —
— -
s_
H
0
^.^
^0
^1
a
50
"5
■*
o
CO
o
c!>
N
lO
o
o
o
O)
00
CO
o>
a>
la
•0
CO
<N
t^
■*
o
■*
C4
0
0
0
CO
CO
Q
d-
d
o
«o
00
N
w
M
d
""■
d
d
l-H
"^
""■
"^
■^
« 2
»• «
"o
g
o
^
o
o
o
o
o
0
0
0
0
0
a «
S p
(N
(M
o:
t^
00
0
CM
0
t^
0
;; •<
H n
O
<N
CO
t^
M<
CO
CO
t^
a>
Q gd^i^
3
ra
3
00
t^
00
•*
CO
00
CO
00
00
00
a,
■<
fcl
(O
^
(N
O
o
e
(N
^
CO
0
CO
„
Ci
lO
OJ
r4
d
00
^
06
r^
d
<N
d
P
n
■*
■*
•*
■*
CO
CO
■*
CO
10
CO
11
iO
02
P
P
hi
a'
t>.
■*.
t^
■^
t^
lo
CO
CO
<N
■*
»o
0
oo
t—t
U5
lO
d
CO
■N
d
CO
CJ
CO
00
(N
S
5
o*
U5
rt<
CO
T»t
■*
■o
iO
T)<
10
TT
CO
'!<
<£
^
02
■^
'^
'^
"^
"^
'^
"^
"^
a
m
>o
U3
10
^5
v
05
OS
o
"i"
o
o
lO
0
0
0
0
0
j3
«o
CO
00
o
•*
o
CO
■<J<
Tl<
•*
■*
u
o
d
^ h3
a
>
d
d
d
d
d
d
d
d
d
d
d
d
i«^
1
—
^-
.s"
13
s
h^
a
1^
H
9
o
1
a
5
a
g
p
,
'""i
3
i
^
,
,
CI
I
J3
0 0
5
<<
■<
-
M «
<!
bi
u
p
"r?
c rt
C3
(d
.
3
•<
^1
1-
1
a
^
Ul
^
4J
i
H
£?
•a
g
tf
02
b
•<
lU
<u
^
a
d
>
B
s
s
"
3
B
0
^
H
H
^
■43
O
M
u
■^
s
H
a
.>
■5
02
o
S
^
-^
=
=
3
o
S
=
'
5
m
0.
0
=
Q
(i.'
o
u
o
Hi
c^
u
fe
^
(4
W
a)
J3
tf
4
i
a S
.S
4>
GQ
^
CO
02
•i
o t^
2
a
3
Hn
-♦>
D
o
CO
M
CO
CO
Ol
CO CO
CO
N
M
CO
CO
Q
a
«
H H
9 n
fi a
«o
t^
00
a>
o
■-^
c^
CO
■«)"
10
CO
t^
H p
!z:
SAFETY VALVE CAPACITY
125
..
^
5
0
8
_
fc"
■5
B
si
a>
o
1
§
0-
>
S
■3
0 0 "
0
09
o
t>
0 0
^
'
■^2; ^
M 00
0
0 CO
0
CO
0 CO
CO rl<
t^ CO
tf>
« OS CD
00
Tjl
t~ >-l
'^ e<3
<N O
"5
lO r»< CO
c^
00
t^ t^
00 0
Tl< <N
(N
(N >0 --i
00
00
CO ■^
00 CO
d d
d
d d rt
d
d
d d
d d
U3 O
IC
10 0 0
0
0
0 "5
0 0
rH lO
(N
t^ (N CO
0
t-
CO 0
CO 10
lO >o
0
CT> CO CO
CO
CO
-H 1^
0 03
C^ •-I
(N
-1 Tl« 00
CO
CO
10 CO
(^ ■<<<
— Q>
V
0 CO CO
t^
■*
CO us
1^ o>
a
c • •
0
0 e<i CO
00
00
ui d
06 co'
jz;
^^^
Ti<
•*
•* IN
•«< ■*
« N
M
t^ 00 0
CO
^
00 •*
•* CO
d -^
d
t^ CO 03
^
d
<N IN
00' d
T»< lO
f-i
■* ■* CO
T»l
•* T(l
CO CO
J^ o
10
10 0 0
>o
0
0 0
0 0
O CO
r^
r^ 1^ ■*
0
CO
0 r^
CO 01
rt o
0
0 0 -H
1 0
"". R
d d
d
ado
d
d
d d
d d
bi
•0
u
0)
V
V
S
a
m
a
S
s
S
s
sx
H
ja
'%■
«
'$
■3 '
<
5^5
'
'4'
<
U
eS
■l
cS
a
3
1
g"
a
o
tf
tf
>.
CO
b
>.
a>
oj
03
03
.2
.2
D
CQ
ca
H
>
O ,
a"
o
o
o
0
13
>
5
; t. ,
=
2
5 3
'i^
fa
fe
0
Hi
tf
«
0
a
■i
_a
a
1-1
«
CO
CO
•«•
-» HM
N -4
w
N -H CO
CO
CO
CO Cvl
CO CO
1
00 a>
0
^ IN CO
■*
>o
CO r-
00 a>
Oi
Ol (N (N
(N
(N
(N <N
N (N
1
a -- -3
n
Q
X
^
X
^
w
1^
0)
-5
"3
+
+
>
"W
X
J3
X
Q
—1
XI
^^
X
IN
C^_
0.
0
+
II
^
X
e
a
03
Q
j3
X
2: a
O >" rt
.2 03
II 0
126 SAFETY VALVE CAPACITY
31 For water tube marine and Scotch marine boilers, the sug-
gested constant is 0.095. This is based upon an overload or maxi-
mum evaporation of 10 lb, of water per square foot of heating surface
per hour.
32 For locomotives the constant 0.055 was determined experi-
mentally as explained below. Special conditions to be considered
in locomotive practice separate it from regular stationary and marine
work. In the first place the maximum evaporation of a locomotive
is possible only with the maximum draft obtained when the cylinders
are exhausting up the stack, at which time the throttle is necessarily
open. The throttle, being open, is drawing some of the steam and
therefore the safety valves on a locomotive can never receive the
full maximum evaporation of the boiler. Just what per cent of this
maximum evaporation the valve must be able to relieve under the
most severe conditions can only be determined experimentally.
Evidently the most severe conditions obtain when an engineman
after a long, hard, up-hill haul with a full glass of water and full
pressure, reaching the top of the hill, suddenly shuts off his throttle
and injectors. The work on the hill has brought the engine steam-
ing to its maximum and the sudden closing of throttle and injectors
forces all the steam through the safetj^ valves. Of course the
minute the throttle is closed the steaming quicldy falls off and it is at
just that moment that the most severe test upon the valves com.es.
33 A large number of service tests have been conducted to deter-
mine this constant. The size of valves u^Don a locomotive has been
increased or decreased until one valve would just handle the maxi-
mum steam generation, and the locomotive heating surface being
known the formula was figured back to obtain the constant. Other
special conditions were considered, such as the liability in locomotive
practice to a not infrequent occurrence of the most severe conditions;
the exceptionally severe service which locomotive safety valves
receive; and the consequent advisability of jDroviding a substantial
excess valve capacity on locomotives.
34 As to the method of applying the proposed safety valve capac-
ity rule in practice, manufacturers could be asked to specify the
capacities of their valves, stamping it upon them as the opening and
closing pressures are now done. This would necessitate no extra
work further than the time required in the stamping, because for
valves of the same size and design, giving practically the same lift,
this would have to be determined but once, which of itself is but a
moment's work with a small portable lift gage which is now manu-
SAFETY VALVE CAPACITY ]27
factured. The specifying of safety valves by a designing engineer
could then be as definite a problem as is that of other pieces of
apparatus. Whatever views are held as to the advantages of high
or low Ufts, there can be no question, it would seem, as to the advan-
tage of knowing what this lift actually is, as would be shown in this
specifying by manufacturers of the capacity of their valves. Further,
as to the feasibility of adopting such a rule (which incorporates the
valve lift) in statutes governing valve sizes — this would involve the
granting and obtaining by manufacturers of a legal rating for their
valve designs based upon their demonstrated lifts.
35 This paper has dealt with but one phase of the subject of safety
valves in order that its conclusions might be drawn more clearly.
The apparatus and tests shown indicate that the lifts and capacities
of different make valves of the same size and for the same conditions
vary as much as 300 per cent, and that there is therefore the ha-
bility of large error in specifying valves in accordance with existing
rules and statutes, because these rules as shown rate all valves of
one size as of the same capacit}', irrespective of this variation.
A simple rule is given, based upon an extended series of direct capacity
tests, which avoids this error by incorporating a term for the valve
lift. Finally, the method and advantage of applying this rule in
practice have been briefly indicated.
No. 1233
DISCUSSION UPON [SAFETY VALVES
The subject of safety valves was extensively discussed at the New York
monthly meeting, February 1909, when the preceding papers by Frederic M.
VVhyte and Philip G. Darling were presented. The discussion was continued
at the Spring Meeting at Washington, May 1909. Full reports of both dis-
cussions were published in The Journal for April and June 1909. In what
follows, synopses only are given of the most important engineering data
presented upon the proportions of safety valves.
Luther D. Lovekin Miscussed the rules for marine work adopted
by the Board of Supervising Inspectors. In 1903 the regulations con-
cerning safety valves prescribed by the United States Board of Steam-
boat Inspectors were investigated by Mr. Lovekin, for the purpose
of formulating new rules. The rule in use was based on grate sur-
face without regard to the amount of coal burned in a given time.
The rules finally prepared by Mr. Lovekin and adopted by the Board
of Supervising Inspectors are based on evaporation and on Napier's
formula for the flow of steam. The formula for the required area of
discharge for a valve is derived as follows:
Let
P = absolute pressure
W = weight discharged per hour in pounds
A = area valve opening in square inches
d = diameter of valve in inches
a = area of valve of diameter d
From Napier's rule
360 A P
w = -^~
For safety valve practice allow 75 per cent of this and restrict the
lift of valve to ^V diameter. Then
270 P ndr
W = — y— X ^ = 4.821 Pa
W
a = 0.2074 -^
^Chief Engineer, New York Shipbuilding Co., Camden, N. J.
130 DISCUSSION
If W represents the weight of water in pounds evaporated per square
foot of grate surface per hour the above formula will give the area
of valve required per square foot of grate surface.
2 A table of safety valve sizes was prepared by the aid of this
formula, but the board failed to state in their rules that the sizes
were based on a lift of 1/32 of the diameter. In commenting upon the
above formula, Mr. Lovekin brought out the following points :
3 The clear area between the valve and its seat (due to having a
lift equal to 1/32 of its diameter) is only about 1/11 of the area cor-
responding to the nominal diameter found by the formula. There-
fore it would seein that the inlet from the boiler to the safety valve
need be equal in area only to the free area between the safety valve and
its seat. This would reduce the opening in the boiler to about 1/11 of
the area used at the present time. Experiments have shown, how-
ever, that a free entrance from the boiler to the safety valve is abso-
lutely necessary to prevent chattering. Just what this relation is has
not been determined. It would depend entirely on the length of the
nozzle or pipe connecting the safety valve to the boiler. In most
cases, safety valves are bolted either directly to the boiler or to a
casting \shich is bolted directly to the boiler, so there would be very
little gain in reducing the inlet nozzle to a safety valve. If safety
valves are connected to the outlets of dry pipes to boilers, it is advis-
able to have at least 25 per cent excess area through the slots in order
to prevent excessive pressures in the boilers.
4 Some rules insist on an outlet area equivalent to the full bore
of the safety valve, which seems inconsistent if we have only 1/11 of
the area for the steam to pass through at the valve seat. An outlet
from a safety valve equal to 1/2 the nominal area of the valve would
no doubt suffice in all cases. Most of the United States battleships
are equipped in this manner. While the United States cruiser Tennes-
see was on trial the main engines were stopped suddenly. All the
the safety valves responded instantly and though the steam pressure
went up to 10 lb. above popping point, no trouble was experienced,
proving that a combined area of outlet pipes equal to 1/2 the area of
the safety valves was sufficient.
Albert C. Ashton said that while the tests show what pop safety
valves would accomplish under certain favorable conditions, they
have not clearly demonstrated that high-lift valves so made are a
success on all applications. They certainly have shown many failures
in locomotive service during the past year and must still be classed
as an experiment.
SAFETY VALVES 131
2 Safety valves should never give such a large and sudden relief
as to affect the water level in a boiler, neither should they close so
suddenly as to cause a shock to the boiler by the quick stoppage of the
flow of steam. High-lift valves which do this are not as practical
as lower-lift valves which give a somewhat slower and easier
relief.
3 The tests which Mr. Darling has explained show an average
lift of 1/8 in. for the high-lift valves which is about double the lift of
standard valves. Such high lift seems to be excessive, although there
may be some virtue in making valves with a little higher lift than the
common standard of 1/16 in.
A. B. Carhart, speaking on the proper rating of safety valves and
their relation to boiler capacity, said that the limit of diameter of
valves for stationarj' boilers should be 5 in., and for locomotives 3i
in.; common practice is in accord with this. Valves of 1 sq.in. dis-
charge area are the largest advisable for locomotives. A total dis-
charge area of 2 sq. in. for locomotives having 35 sq. ft. grate area,
and of 3 sq. in. for the largest ones having 50 sq. ft. grate area, has
been demonstrated to be amply sufficient. The capacity might be
divided as follows: (a) muffled valve with close adjustment; (6)
reserve valve regulated for reasonably greater discharge; and (c) an
emergency valve as the ultimate protection against explosion, the
other two simply to limit the working pressure under ordinary con-
ditions. Valves generally discharge more steam than engineers are
availing to permit. The strain on the boiler is dangerous when the
opening is too large and sudden; and if water is lifted, it chokes the
relief through the safety valve and endangers the cylinders.
2 A smaller valve with high lift is not the equivalent of a valve
of larger seat diameter and less lift wdth the same discharge area;
the smaller valve gives a smaller percentage of steam discharge, there
is greater danger of sticking of the guide wings in opening and more
trouble from pounding of the seat and leaking, and the outlet area
becomes too large in proportion to the inlet, causing chattering and
ineffective relief to the boiler, besides requiring an undesirable exces-
sive spring compression. The lift should not exceed 0.08 in. for loco-
motive valves, and 0.10 in. for stationary valves used at lower
pressures; prudence and economy would reduce rather than increase
this hmit. Many valves of high lift have been produced in past
years, but all have been withdrawn because of rational objections
developed in their use.
132 DISCUSSION
3 Every valve has a wide range of lift, which can often be varied
from 0.04 in. to 0.10 in. by simple adjustment, and to still greater
limits by a change of springs. In valves as commonly made, limited
lift is a matter of preference, not of necessity; and such valves are
giving entire satisfaction in service, with demonstrated safety under
all conditions, under the present rules and ratings.
4 All internal work that must be extracted from the escaping
steam, to accomplish high lift of the disc, reduces the velocity and
efficiency of the relief and requires an undue throttling of the outlet,
strangling the discharge instead of relieving the boiler. This con-
dition is described in the early patent to Richardson, of January 19,
1869, showing the over-lapping regulating ring.
5 Napier's formula was used as the basis for calculating safety-
valve areas as long ago as the tests made by the United States Board
of Supervising Inspectors of Steam Vessels in 1875; and reports
of tests of safety valves made at the Massachusetts Institute of Tech-
nology can be found in Peabody and Miller's text book on Steam
Boilers printed more than a dozen years ago, showing lifts of 0.07 in.
and 0.08 in., with an efficiency of 95 per cent of the calculation by
Napier's formula as there applied.
E. A. Ma\* spoke on the proper method of rating safety valves for
low-pressure boilers. A safety valve on a low-pressure boiler is rarely
called upon to exhaust all the steam-generating capacity, due to several
conditions :
a In the majority of heating plants, the full amount of radia-
tion is almost always in service, caring for a large percent-
age of the steam generated, and even when the radiation is
nearly all cut out there is still circulation through the
piping.
b Practically every steam boiler used in low-pressure work,
which rarely calls for gage pressure in excess of 2 lb., has
its damper regulator which, when properly set, checks com-
bustion when 2-lb. pressure is reached.
c Chimney conditions in the majority of heating plants make it
almost impossible to drive the boiler to its maximum
steam-generating capacity, i.e., the maximum capacity
obtainable with every condition exactly right.
^Mechanical Engineer, American Radiator Company, Chicago.
SAFETY VALVES 133
2 In practically all house installations at least two of these condi-
tions exist, and in a majority all three, so that we would have to select
a valve out of all proportion to actual requirements in order to
exhaust all the steam which might be generated by the boiler under
its full steam generating capacity under ideal conditions.
3 This brings us to a consideration of maximum capacity and
how it is established : whether (a) by the heating surface of the boiler
alone; (b) by the grate surface; (c) by the fuel-carrying capacity;
(d) by the rate of combustion; or (e) by all combined. Scarcely any
two manufacturers of low-pressure house-heating boilers agree in
this particular. One may rate solely on the area of heating surface,
another on the grate surface, and still another on the amount of fuel
the grate will carry, but the writer believes that none of those factors
should be considered alone.
4 In view of the wide variation in methods employed by manu-
facturers in ratiiigs of boilers, as well as in the rules employed by users
of safety valves, it would be difficult to select a proper size valve based
on grate dimensions only. If valve manufacturers would indicate, in
addition to the size of the valve, its capacity at different adjustments
for exhausting steam, it would help materially. Valves could in fact
be designed and sold on their exhaust capacity without regard to
specific size, i. e.; owing to variation in design, one valve might have
a larger diameter with a lesser lift than another, while their capacity
for exhaust might be identical.
5 The simplicity of this method will be appreciated by anyone
considering the rules and formulae in effect at present. If the law
specified, however, that for a certain evaporative power or rating of
boiler a certain exhaust capacity should be maintained in the valve
each manufacturer could determine for himself the proper valve to
use.
F. J. Cole quoted from a letter of an Enghsh locomotive builder
stating that the "Ramsbottom" duplex safety valve is almost uni-
versally adopted there. It was introduced in 1858 and made 3 in.
in diameter. Notwithstanding that boilers have since nearly
doubled in capacity and pressures have been increased 50 per cent,
this size is still used, which shows the disregard of proportion of
safety valve to any other part of a locomotive boiler.
2 It is probable that general foreign practice for locomotive
safety valves is systematized no more than in England or America.
On account of the peculiar conditions governing the draft of loco-
134 DISCUSSION
motives the same necessity does not exist for safety valve regula-
tion as in the case of marine or stationaiy boilers, the action of
the exhaust automatically talcing care, in large measure, of the
generation of steam.
3 Mr. Cole stated that he is in favor of a thorough investigation
looking towards the formulating of definite and authoritative rules
for the application of safety valves to locomotives, and invited atten-
tion to the following suggestions for their preparation :
a The diameter, number and kind of safety valves to be based
on their capacity for discharging pounds of steam per
second at different pressures.
b The maximum amount of steam which the safety valves may
be required to discharge when the throttle is suddenly
closed after the fires have been urged to their maximum
rate, to be based on the square feet of equated heating
surface, so that the relative values for evaporation for
various kinds of heating surface, whether of firebox, water
tubes for supporting arch brick, long and short boiler
tubes between the limits of 10 and 21 ft. in length, and
values for different spacing of boiler tubes, will be taken
into consideration. Or, what would be simpler, some ap-
proximation of average value of heating surface, equated
to account for difference in length and spacing of tubes;
the firebox heating surface in this case to be considered as
a certain percentage of the whole for all sizes of locomotives.
4 A great diversity of practice exists in the spacing of flues in
locomotive boilers. The variation in length ranges in common prac-
tice from 10 ft. to 21 ft. These two conditions make the use of
unequated heating surface somewhat unreliable as an absolute guide
for the amount of water evaporated. It is evident in two boilers hav-
ing the same diameter and the same length between flue sheets that
one will contain a much larger amount of heating surface if the flues
are spaced 11/16 in. apart than the other if they were spaced 1 in.,
and both these figures are within the limits of what is accepted as
good practice. Furthermore the heating surface of flues of the same
diameter and, saj', 11 ft. long, will be much more effective per square
foot than if the flues were 21 ft. long. Firebox heating surface is, of
course, very much more efficient than tube heating surface, and the
water-tube heating surface for supporting firebricks is more efficient
than the ordinary boiler tubes.
SAFETY VALVES
135
5 Tests show that the evaporation of boilers is somewhat inde-
pendent of the tube spacing, and probably is more in direct relation
to the cubical contents, as it is a matter of common knowledge that
the steaming capacity of boilers does not vary in direct proportion to
the amount of heating surface if a great variation exists in the spacing
of the tubes.
6 The evaporation per square foot of heating surface in locomo-
tives is a variable quantity, ranging from 6 lb. or even less to 15 or 16
lb. per square foot of heating surface per hour. On the authority of
Professor Goss, from Purdue University tests, it may be stated that
the evaporation in a very general way, and the draft produced by the
blower and exhaust in terms of inches of water, will be approximately
as follows:
1-in. draft will evaporate 3.0 lb. per foot of heating surface per hour
2-in. "
6.0 "
3-in. "
8.2 "
4-in. "
10.0 "
5-in. '
11.4 "
6-iii. "
12.G "
7-in. '
14.0 "
8-in. '
15.0 "
Dr. Chas. E. Lucke stated that another element in the safety valve
question, of minor importance, perhaps, is the time element. He
had experimented for many years with rapidly rising and rapidly
falling pressm-es, and believed that increase in pressure in a chamber
may go momentarily far beyond what a safety valve is set for. Be-
cause this excess is only momentary and measured in fractions of
seconds, it should not be considered of no consequence ; it is indeed
of far more consequence, as a suddenly applied load cannot be resisted
by the metal under stress as well as a steady load.
2 If then by any romote series of circumstances the pressure in the
boiler suddenly rises, as it may, the time element will enter in, the
pressure will go higher than the safety valve is set for, before the valve
opens, and will suddenly stress the entire structure. This subject
should be stu died eperimentallj^, with the others involving the steady
rate of steam discharge, to discover if it is of any consequence in practi-
cal safety valve work. Although Dr. Lucke had never seen the pres-
sure rise in a steam boiler in this way, he believed it could so rise, pro-
ducing the effects described.
136 DISCUSSION
Jesse M. Smith thought Dr. Lucke had touched on a point which
needed investigation. Another point along the same line is the danger
of having a safety valve too large, particularly if it be of the "pop"
kind. If a boiler be stored with water at a temperature corresponding
to 150-lb. pressure, and that pressure be suddenly reduced, a portion
of the water will instantly flash into steam and the boiler may be
greatly strained and may explode. There is danger from having
a safety valve too large as well as from having it too small.
2 Those who have had to do with the investigation of boiler
explosions, and particularly those being questioned with regard to
them in the courts, will realize the necessity for rules based upon
scientific investigation and reason, instead of rules having no special
reason for their existence except that they have existed for hundreds of
years or more.
Garland P. Robinson stated that the commission with which he
is connected had collected reliable data on about 7500 locomotive
boilers, and he had recently calculated the valve capacity of 1000 of
these boilers for the purpose of finding the average practice for safety-
valve equipment. The greatest variations were noted. For in-
stance, boilers using 180-lb. pressure with valves of re-in. lift have
two 3-in. valves to take care of the evaporation from 1750 to 3350
sq. ft. heating surface. It was also found that two 2^-in. valves were
used to take care of 900 to 1900 sq. ft. of heating surface. These
cases represent whole classes and not individual boilers. Therefore
it would appear that no rule has been followed to determine the size
of valve required.
2 The heating surface, all things considered, is the best unit of
measurement for determining the size of safety valves for locomotive
boilers. In his opinion a formula based on the heating surface and
providing for 50 per cent of the maximum evaporation of the boiler,
will give satisfactory results for locomotives. A formula for size
of safety valves for locomotive boilers can be derived in the manner
shown in Mr. Darling's paper on Safety Valve Capacity.
3 For locomotive valves with 45-deg. valve-seats, Mr. Robinson
used the formula
^ heating surface
and for locomotive valves with flat valve-seats, the formula
^ heating surface
^ = "°^^ LXP
SAFETY VALVES 137
4 He had checked 1000 boilers and found the average constant to
be 0.0441 for present practice. Included in the 1000 boilers, however,
are a number which evidently have valves of insufficient size, as the
constant in their case is only 0.024. Eliminating this class of boilers,
the constant for average practice is about 0.05, as given in the formula,
H. C. McCakty^ said that his experience had clearly proved that
safety valves with U7iusual discharge, resulting from increased lift of
valve, cause a violent disturbance in the water level, especially on the
large modern locomotive boilers, and in proportion to this disturbance
is the volume of water passing the throttle valve, and hence to the
steam chest and cylinders, increased. Railroads will be relieved of
many expensive repairs by reversing these conditions, and thus pro-
duce the driest steam possible for the engine. To this end, the
throttle valve, as is well known, is located at the highest possible
point in the boiler. Further to secure greater locomotive efficiency
in this direction, the safety valve should be at as high a point on the
boiler as clearance limits permit, and with an independent short con-
nection of ample dimensions to the boiler.
2 Mr. McCarty said his further observations have been that the
location of the valve on a boiler has much to do with the normal crest
of the water. Air-brake shocks in train and similar effects, in conjunc-
tion with high-lift valves, have been a frequent cause of locomotive
failures through the combination of undesirable conditions, all of
which cause a greater volume of water to pass through the throttle
valve and safety valve.
3 In the speaker's experience in locomotive service he had never
had even a suggestion of the necessity or the advisability of increas-
ing the lift of the valve; on the contrary, the reverse condition, from
a service standpoint, presents itself. The possible limited economy in
first cost of a slightly smaller valve having increased lift, to accom-
phsh increased discharge, compared with the next larger size valve
with normal lift, is deceptive, as the short life and expensive mainte-
nance of the high-lift valve make it not only an expensive burden to
the railroads, but an unreliable device.
M. W. Sewall said that if the evaporative capacity of the boiler and
the delivery capacity of the safety valve were adapted to each other,
no difficulty need arise from the use of high-lift valves. The areas
'President Coale Muifler and Safety Valve Co., Baltimore, Md.
138 DISCUSSION
of approach to the valve and discharge from it should then be such
as practice has already shown to be essential. The usual diameters
have been mentioned in the discussion as if the}- could not be changed,
and the high and low lifts have been spoken of as related to those
diameters. As the high-lift valve has a greatly increased discharge
capacity, however, it should be reduced in diameter and an entirely
new adaptation of diameters to " pounds of steam discharged "
should be made. A manufacturer could then adopt any desired com-
bination of diameter and lift and the valves would be rated on the
pounds of steam delivered per second.
A. A. Caki:, in discussing safetj'-valve springs, said that the ratio
between the pitch diameter of the spring and the diameter of the wire
composing it should not be less than 5 to 1, but 7 to 1 is a better mini-
mum proportion. A pop- valve spring should not be wound to a
smaller proportion than 7 to 1, and with such a spring coiled to a
smaller ratio he had found a considerable breakage resulting.
2 One matter deserving careful attention in the design of pop-
valve springs is the shape of the section of wire used. Unquestion-
ably, the best and safest wire for springs is that of round section. The
principal stress occurring in the wire of a helical spring is that of
torsion, and in a wire of square section the greatest fiber stress occurs
at the corners of the square, which are the most distantly removed
from the center of the section.
3 The only advantage gained by the use of square-wire springs is
a slight reduction of the space required for a spring having the same
resistance to compression.
4 The most durable of all is the helical spring designed to resist
extension, known as an extension spring. When this spring is prop-
erly applied, the load is carried directly along the line of the spring's
axis, thus doing away with the "buckling" which so frequently im-
poses harmful bending strains (in addition to the torsional strain)
in the wire composing compression springs. The use of compression
springs for pop valves has become almost universal, but there is
no reason why extension springs of good design cannot be used for
this purpose.
F. L. DuBosQUE thought that the formula in the U. S. marine
laws has the serious defect that some of its factors are left to the
opinion of any one of a great number of persons concerned in its use.
The factor W is made up of two quantities, the calorific value of the
SAFETY VALVES 139
fuel and the amount burned per square foot of grate surface, and the
\ahio of these factors can with reasonable judgment be varied so as
to vary the size of the safety valve at least 50 j)er cent. It is now im-
possible for a tlesigner to specify the size of a safety valve on a marine
boiler without first obtaining from the United States inspector his
opinion on the value of these two factors, notwithstanding the fact
that the inspector who is compelled to decide this question cannot
i:)ossibly have as much information to assist him as the designer.
2 This new formula, therefore, has not in any way improved the
Rules of the Steamboat Inspection Service and, as pointed out above,
has added only a complication. As to results produced by it, it is
easy to see that by selecting proper proportions for the two factors
that make up W, — and these factors both may be within reasonable
limits, — the same result will be obtained as by the old formula. The
old formula at least gave the designer a certain basis to work on, and
if he was designing his work with the proper regard for safety he had
the privilege of deviating from the formula if he felt it did not pro-
vide a valve of large enough size.
3 This new formula is also similar for cylindrical and water-tube
boilers. Practical operation shows that a safety valve on water-tube
boilers should be much smaller than on cylindrical boilers of equal
evaporative power. A sudden release of steam pressure in a water-
tube boiler with its limited water-line area causes more damage by
lifting the w^ater within the l:)oilef than can be caused by a moderate
increase in steam pressure.
L. D. LovEKiN, in replying to the remarks by Mr. DuBosque, said
that he was not aware of the trouble he had caused marine engineers,
and still further, he saw no reason for such trouble. He had dis-
cussed the matter fully wnth a number of engineers and showed them
the new formula which he proposed submitting to the Board, and all
agreed that his formula was based upon common sense.
2 Any safety valve based on one square inch of opening for three
square feet of grate area for a Scotch boiler, and one square inch of
opening for six square feet of grate area for a water-tube boiler, is
absurd, and yet this was the formula used by the United States
Inspectors for many years.
3 Xelson Foley, of England, states that safety valves may be
made capable of hfting, say | of their diameter; that a high lift is
useless and may be an evil if anything gives way; that the waste steam
pipe, when not under the Board of Trade rules, may be equal in area
140 DISCUSSION
to the opening with the lift just mentioned, i. e., the area of the waste
steam pipe would be one-half the gross cross-sectional area of the
valve.
4 Our United States Navy Steam Engineering Department, with
all their experience in connection with boilers, have agreed with sev-
eral prominent authorities abroad on a lift of | the diameter of the
valve. It does not follow, however, because a valve has provision
for a lift equal to | of its diameter, that it ever lifts this amount. It
is simply a provision in case the valve is required to be lifted by the
safety-valve hand-operating gear usually provided on all ships.
5 The area of waste steam pipe on all our recent naval vessels is
made | the gross cross-sectional area of the valve, which accords
with the statements of Mr. Foley.
6 It is a coincidence that while the present rule might give an ex-
cessive lift on sizes above 4|-in. diameter, it averages up closely to the
sizes recommended by many manufacturers for valves below 4|-in.
diameter. The rate of evaporation of 180 lb. in the present rule
almost coincides throughout with the Board of Trade formula for
safety valves under natural draft.
7 It appears therefore that the Board of Trade thought it wise to
keep all boilers under natural draft at the same rate of evaporation,
i. e., all boilers worked under natural draft are assumed to be capable
of evaporating 180 lb. of water per square foot of grate surface, which
seems to be a safe maximum rate for any marine boiler under natural
draft.
8 When forced draft is used, under the Board of Trade regulations,
the area of the safety valve must not be less than that found by the fol-
lowing formula.
(estimated consumption of coal\
per square foot of grate, in I = area of valves required,
pounds per hour -=-20 /
A equals the nominal area of the valve, based on its diameter, as found
from the table of safety valve areas under the Board regulations.
9 The results of experiments on safety-valve lips illustrated in
the figures may be of interest to the members of the Society. These
experiments were made by Nelson Foley to determine the effect of
adjusting the lip on safety valves.
Nathan B. Payne believed the most important point brought out
by the discussion was that there is no proper standard of measurement
SAFETY VALVES
141
rJK-2 Fig.3
Fig. 1 Results of Experiments on Safety-Valve Lips
VALVES ROSE AT 81 LB. AND LIFTED ABOUT ^ IN. RATIO OF VALVE AREA TO
GRATE AREA § SQ. IN. TO 1 SQ. FT. FIG. 1 VALVE CLO.SED IN J MIN. AT 79 LB.,
AND VIBRATED CONSIDERABLY. FIG. 2 BLEW STEADILY, WITHOUT CLOSING.
FIG. 3 CLOSED IN 1 MIN. AT 80 LB. PRESSURE DROPPED STEADILY. FIG. 4
SAME AS IN FIG. 3. CLOSED IN f MIN. AT 79^ LB. LESS VIBRATION THAN
IN FIG. 1.
142 DISCUSSION
for the safety valve's capacity. Whether a high-lift or a low-lift
valve is selected, what we must have is some way of measuring the
relieving capacity. When we buy a 4-in. valve, for instance, we want
to know whether that valve has relieving capacity for a 100-h.p. or
a 200-h.p. boiler, or what size it is suited for.
2 We have been thinking with regard to the relieving capacity
of safety valves that we need consider only one dimension, but it is
absolutely impossible to determine the amount of relieving capacity
in a given time without knowing the lift of the valve off the seat,
so as to get the effective area of opening. The question for the
user to decide is how much relief he can get from a given make and
size of valve. If one maker offers a safety valve having :^-in. lift,
and another offers a i-in. lift, each should state how many pounds
of steam per hour his valve will relieve.
H. 0. Pond said the question of high-lift and low-lift valves seemed
to be one simply of capacity. If the low-lift valve will deliver a cer-
tain number of pounds of steam at a given pressure and temperature,
and its capacity under these conditions is known, this is the principal
thing required. The same test applies in the case of the high-lift
valve, the essential point, however, being to know how much steam
the valve will discharge. Undoubtedly a high-lift valve will give
more capacity than a low-lift valve having the same diameter of
opening. This being so, we could use a smaller valve of the high-
lift type, which would be an advantage in many ways.
F. L. Pryor summarized the results obtained from tests made to
obtain the blowing-off pressure of safety valves when tested with
water and with steam.
2 A standard 4-in. pop safety valve, set for 125 lb., was mounted
on a 4-in. pipe and so connected that either steam pressure or water
pressure could be admitted to the valve. One set of tests was made
over a period of 15 days, the test of one day being with steam and the
following day with water. In a second series of tests, the valve was
tested at three different settings on the same day, viz. 104, 131 and
159 lb. The third series of tests was made with the valve at a num-
ber of different settings, from 105 to 165 lb., one measurement being
made directly after the other and no precaution taken to insure that
the valve had returned to its normal temperature after the preceding
test, except that before operating with water pressure a considerable
amount of water was flushed through the valve.
SAFETY VALUES 143
3 The results obtained in all the tests were in practical agreement,
antl indicated that the l)lo wing-off pressure with steam and with water
did not differ to any great extent, although the pressure to blow off
with water was higher than with steam.
4 In the case when the valve was allowed to cool 24 hours, the
water pressure required to open it was about 3| lb. higher than the
steam pressure. In the tests where the valve was cooled thoroughly
with water, the pressure with water was about 3 lb. higher than with
steam. In the rapid change test the water pres'sure amounted to
about 2.6 lb. more than the steam pressure.
5 In all tests the steam and water pressure record was that at
which the valve was in full operation. In the case of the steam pres-
sure test there were two distinct points below full open pressure which
could also have been noted: when the valve began to leak, which
occurred about 2 lb. below the final blowing-off pressure, and when
the rate of flow suddenly increased, which was about 1 lb. below
maximum.
A. B. Cakhart, speaking on the proportions of safety-valve parts,
said that the specifications which require valve seats to be made of
non-corrosive metal, and the rules which compel every valve to be
tried and lifted by the lever every day, aim to overcome the ever-
present danger that the valve may stick upon its seat and fail to open
at the critical moment. But the greatest cause of the sticking of the
valve, when it does occur, is not corrosion of the seat face, but the
binding friction of the disc-guides against the side of the well or
throat of the valve. This cocking or binding effect can be decreased
by any modification of design which will reduce the diameter of the
cylindrical guide, or which will bring the guiding surface close to the
plane of the seat, both of which will reduce the moment of the friction
or cocking stress.
2 Any device which reduces the lift of the disc and the spring
movement to the least possible amount will also reduce the eccentric
spring action and its effect, and any valve design which contemplates
an unnecessarily large lift or compression disadvantageously magnifies
this effect.
3 An early and still common form of safety valve has the seat
opening beveled at an angle of 45 deg. The effective steam passage
is therefore measured by the sine of 45 deg., which is approximately
only 0.7 of the actual compression of the spring when the valve
opens, so that the spring must necessarily compress about Ih times
144 DISCUSSION
the effective lift. Even this does not always afford a free passage for
the steam tojthe air where there is vertical overlap of the regulating
ring against the lip of the disc in order to increase the lift against the
greater pressure of the shortening spring.
Edw. F. Miller presented the following method of obtaining the
valve discharge area based on the rate of fuel consumption. The
weight of steam flowing through an orifice with a slightly rounded
entrance may be figured quite accurately by Napier's formula (some-
times called Rankine's formula). Its accuracy for commercially
dry steam has been shown by tests made under pressures varying
from 30 to 150 lb. There are a number of papers on this subject in
the earlier volumes of the Transactions. According to the formula
the weight of steam discharged per second through an orifice with
FP
slightly rounded entrance is — where F is the area of the orifice in
square inches and P is the pressure in pounds absolute on one square
inch.
2 The discharge per second through an orifice with a sharp edge
at the entrance, as would be the case in a safety valve, has been found
from actual tests on valves to be 0.95, the amount figured from this
formula. The opening needed in a safety valve may be figured as
follows:
G = grate area.
U = rate of coal consumption per square foot of grate per
hour.
9 = probable evaporation per pound of coal under actual
conditions.
G X iil X 9
= weight of steam made per second.
3600
Equate this to the preceding expression and solve for F:
GXRX9 ^ FXP
3600 ' 70
p^GxBX9X70
3600 X P X 0.95
3 The area of the opening through a safety valve is equal to
the inner circumference of the seat times the effective lift. For a valve
with the seat at an angle the effective lift is equal to the lift multi-
plied by the cosine of the angle the seat makes with a horizontal.
SAFETY VALVES 145
4 For a 45-deg. angle the effective lift is 0.707 X lift. Calling
D the inner diameter of the valve, the opening is
~ X D X lift X 0.707
Substituting this for F,
.i)X lift X 0.707= ^X^><^X^O
3600 X P X 0.95
If the lift of the valve is -j^,, in.,
^^ GX/2X9X70 _ GR
3600 X P X 0.95 X ^ X 0.707 X 0.1 ~P X 1.206
5 If the lift is 0.05 instead of 0.10, then the valve diameter D
is doubled. Doubhng the pressure will make the same valve with
the same lift take care of double the weight of steam. For illustration :
Grate area = 2'".
Coal consumption = 18 lb. per square foot per hour.
Pressure = 120 lb. absolute.
^^ ^X18_3^
120 X 1.206
Pressure, 150 lb. absolute.
Grate area, 50 sq. ft.
Coal consumption, 25 lb. per square foot per hour.
50X25
150 X 1.206
= Z)=6.9in.
A valve as large as this would be replaced by two of equivalent capacity.
The circumference = 3.14 X 6.9
6.9
Two smaller valves of diameter '- = 3.45 will give the same cir-
cumference and the same discharge with the same lift.
George H. Musgrave^ The function of the safety valve is two-
fold: (a) it gives notice of the highest pressure permissible; (6) it
gives the alarm that more water or less fuel is needed. He had been
told by engineers in the marine service, that through the use of safety
valves with excessive lift and quick discharge, their engines had been
plugged by taking over water. He had known of numerous occasions
in locomotive service where there have been very disastrous results.
'General Sales Agent, Star Brass Mfg. Co., Boston, Mass.
146 DISCUSSION
2 If the same principle is to be introduced in high-lift locomotive
safety valves that is now used in injectors to raise water, what is to
prevent the syphoning of the water to the throttle valve, and its
flowing through the dry pipe and into the cylinders? From his long
experience, originally in locomotive service, afterwards in marine and
stationary service and at the present time, on safety valves for all
uses, he would suggest that the medium-discharge type is the safest
and most satisfactory valve to use. Any valve that will materially
disturb the water level and have a tendency to raise it is dangerous.
M. W. Sew ALL, speaking on exact regulations for valve propor-
tions, said that the public authorities and insurance companies
should establish means of regulation in regard to the following :
a Flange diam.eters for various rates of discharge.
b Requirements as to minimum discharge of pounds of steam
per second within given ranges of pressure.
c Requirements as to non-corrosive seats or other operating
parts, strength of parts, means of operation by hand, and
security against being put out of adjustment by ill-dis-
posed persons.
Albert C. Ashton was opposed to high-lift valves, since they open
and close so suddenly as to injure the boiler and its connected fittings,
as well as the valve itself.
2 He thought that any revised rule for the size of pop safety valves
should not prescribe a capacity of relief that could be obtained only
with a high-lift valve, as suggested by Mr. Darling and Mr. Lovekin,
who approve of a lift equal to J5 of the diameter of the valve.
A. F. Nagle presented the following table on the size of safety
valves for boilers of a given power, based upon the following data:
a A boiler horsepower is the term used to express the evapora-
tion of 34.50 lb. of water per hr. from and at 212 deg. Fahr.
b A spring safety valve can and should be depended upon to
lift M of its diameter.
c The flow of steam follows closely Napier's formula, reduced
to 92| per cent by Mr. Darling's experiments (Par. 24).
SAFETY VALVES
147
r/ The fornmla usod in the computation is h.p. = 0.0951 D'-P,
transformed from Mr. Darling's formula, where
h.p. = boiler horsepower.
D = diameter of valve in inches.
P = absolute steam pressure.
2 In using this table, allowance must be made for what is likely
to be the maximum horsepower of the boiler and not its normal rating.
Fifty per cent overload is not unusual, and double the rating, while
not impossible, is not liable to pass through the safety valve.
HORSEPOWER OF BOILERS AND SIZE OF SAFETY VALVES
Steam
Safety Valves
Pressure
Pounds
2 In.
2Jln.
Sin.
3J in.
4 in.
4iln.
100
44
68
98
134
175
221
125
53
83
120 163
213
269
150
63
98
141 I 192
251
318
175
72
113
162 221
289
366
200
82
128
184
250
327
414
225
91
142
205
280
365
462
250
100
167
227
309
403
510
Note. — Roughly every 4 lb. of coal burned per hour represents one boiler h.p.
Jerome J. Aull thought that the proposed rule should include a
term for a fixed lift rather than a variable one, for the reason that with
the latter would result a hopeless confusion of safety-valve openings
in boilers of the same size. Thus under Mr. Darling's rule a boiler
of a certain size might be provided with a safety-valve connection
varying from 2| in. to 4 in. in diameter, depending upon the make of
valve specified. It would be far more convenient and satisfactory
to standardize safety-valve connections so that any valve having the
required capacity could be used. To do this it would be necessary
that the valves themselves be standardized within certain set limits
and this could be done only by a body of disinterested engineers,
properly authorized to investigate the subject.
2 Mr. Aull condemned high lift as it made the seats and spring-
bearings subject to a severe pounding action; there is more danger
of chattering; close adjustment is not possible; there is danger of
lifting water; and th e boiler seams are sometimes strained to the open-
ing point.
148 DISCUSSION
Philip G. Darling, in closing, dealt with the many different values
at present being advocated for safety-valve lifts. Recent articles
place this maximum limit variously at 0.05 in., 0.06 in., 0.08 in., 0.09
in., and 0.14 in., for the same size valves.
2 It is well known by those in t!ouch with foreign manufacturers
that valve lifts, spring compressions and other valve elements which
are radically different from what has been the general practice in
this country, are being used successfully, and in some places univer-
sally.
3 Two cases will illustrate this. The springs on 3|-in. triplex
valves of the Thornycroft design, used widely in English marine
practice, are not only of the exposed type, but have, when set for a
designed pressure of 250 lb., a compression of 4 in. These are regular
safety valves of the same principles as our own duplex valves. To
those who would condemn a compression of § in. to | in. as radically
high and unsafe this instance should be suggestive and help to broaden
their conceptions of the possibilities of safety-valve design. Again,
in London Engineering, February 26, 1909, reprinted in Power, March
30, 1909, J. H. Gibson tells of exceptionally good results obtained in
a valve having 0.21-in. lift. He says: "We think we are justified
in the assumption .... that anything tending to reduce the
size of these important fittings (safety valves), which have been
growing to abnormal proportions of late, is a step in the right direc-
tion. "
4 High lift is not synonj^mous with excess safety-valve capacity.
A boiler's evaporation absolutely determines the necessary safety-
valve capacity. In a given boiler the pounds of steam per hour which
the valve should be able to relieve can be definitely figured and all
that is further needed, in making the correct valve specification, is the
capacity of the safety valves.
5 It is not a question of lift for itself, but of requisite relieving
capacity, and if this is obtained with a 3-in. instead of a 4-in. or 4|-in.
valve there is a positive, real advantage, not only in original cost but
in the maintenance and better action of comparatively small rather
than large valves.
6 It is thus not a uniformity in the lifts of different valves which
the engineering public should demand, but rather the practice of
stating relieving capacities, based on the actual lifts existing in the
valves themselves. If the capacities were stamped upon the valves,
as already done by one maker, it would give a rational basis for use
in the application of safety valves to boilers.
SAFETY VALVES
149
7 It has been objected that capacities thus published could not
be verified without actual capacity runs, such as the Barberton tests
recorded in the paper, on the ground that in some valves the effective
area of discharge at the seat, upon which the formula is based, is not
the smallest discharge area; or even if it is, that there is a material
throttling or holding back of the steam flow. Valves containing
the original Richardson adjusting ring have been cited as designs in
which this choking occurs.
8 In order to secure information upon this matter prior to con-
ducting the direct capacity tests at Barberton referred to in the paper,
the effective discharge areas at the seat and at the most contracted
passage between the lip and adjusting ring were figured and plotted
for the different valves tested at different lifts. Further, a 3^-in.
valve was constructed having this Richardson ring and projecting
disc lip design, and for the same valve another disc and ring in which
the projecting Hp was cut entirely away. In the former the discharg-
ing steam was deflected through practically 90 deg., and in the latter
the steam had a free straightaway passage. These two designs were
radically different and fairly represented the extremes of what on the
one hand seemed to be a choked or impeded steam discharge passage
and on the other a free open one.
9 The most effective discharge areas of the two taken at the seat
and at the most contracted part of the passage between the lip
and rings are given in the table in square inches for different lifts.
TABLE 1 EFFECTIVE DISCHARGE AREAS
Valve with Projecting Lip and Richardson Ring
Valve without the Lip
Most Contracted
Most Contracted
Lift
At Seat
Point Beyond
Seat
At Seat
Point Beyond
Seat
0.02
0.16
1.20
0.16
2.01
0.06
0.47
1.40
0.47
2,14
0.10
0.79
1.76
0,79
2,29
O.U
1.11
2.27
1.11
' 2.43
These areas, taken with Napier's formula, give a method of figuring
the theoretical pressure existing in the "throttling chamber " under the
disc lip; that pressure being to the boiler pressure as the effective dis-
charge area at the seat is to the most contracted area between the lip
and ring beyond. The highest pressure thus indicated in the throt-
tling chamber is less than 50 per cent of the corresponding boiler pres-
150 DISCUSSION
sure. This pressure in the "throttHng chamber" being the discharge
pressure of steam passing over the valve seat, and the full flow of
Napier's formula being practically unaffected by any discharge pres-
sure less than 60 per cent of the original or boiler pressure, the theoreti-
cal conclusion is that the discharge from neither of these valves would
be affected by the disc design or discharge areas outside of the valve
seat.
10 In replying to the references made to disastrous results to boilers
such as the opening up of seams and fittings due to the sudden release
or cutting off of steam by the safety valve, Mr. Darling discussed
the sudden change of pressure due to opening and closing of throttle
valves and blow-off valves, concluding that the shock to the boiler
from this source would far exceed that due to the closing of the safety
valve.
11 The larger the safety valve compared with the boiler the
greater the shock to the boiler due to its action, if such shock exists.
A 5-in. valve mounted directly upon a94-h.p. test boiler would increase
or accentuate this tendency to strain over say a 3i-in. valve on an
800-b.h.p. locomotive surely 12 or 13 times. Yet with a most sensi-
tive boiler pressure test gage graduated to pounds and mounted upon
this 94-h.p. test boiler, absolutely no recoil of the gage hand upward
either at the opening or closing of a 5-in. valve is perceptible. It
would seem that some increase of pressure such as would be indicated
upon the gage would be positively necessary to transmit a strain to
the boiler.
12 Two cases had recently come to his notice in which loco-
motive safety valves had loosened from their spud connections
and had blown off while the boiler was under its full steam pressure.
One was a 3i-in and the other a 4-in. valve, which therefore opened
areas of 9.6 and 12.6 sq. in., respectively, while the maximum cor-
responding safety-valve discharge area could be but a little over
one square inch. Yet no damage to the boilers was experienced.
The blowing off of 2-in. locomotive whistle connections had been
cited as a not infrequent occurrence. The steam-relief in such acci-
dents is of course more sudden than with a safety valve, and the
fact that this opening of ten to twelve times the maximum discharge
area of the corresponding safety valves results in no further incon-
venience than the replacing of the fittings raises some question as
to the actual disaster impending in the use of valves having a dis-
charge area of but 1 sq. in.
No. 1234
A uniqup: belt conveyor
I'v E. (.'. SopER, Detroit, Mich.
Member of the Society
It is quite possible that a description of a belt conveyor a quarter
of a mile long, and requiring more power to operate empty than
loaded, will be interesting to some of the members and since its
installation and operation are at variance from the prescribed rules
of conveyor design, we beg to submit the following:
2 The belt conveyor was built during the summer of 1908 in one
of the large portland cement plants of the South. It consists of a
24-in. 8-ply canvas belt in two sections, one section about 1000 ft.
between centers, and the other with 1100 ft. between centers, its
function being to convey the shale used in the manufacture of the
cement, from the shale quarry to the plant. The shale deposit is
located on a mountain about 247 ft. above the shale storage tanks,
as shown in profile, Fig. 1. The two sections intersect at an angle
of 140 deg. 40 min., so that the blasting from the limestone quarry
does not interfere with the operation of the belt. The belt conveys
the shale around the limestone quarry, as shown in plan. Fig. 1.
3 The belt is flat and carried by rollers, the top row having 4 ft.
between centers and the return idlers 12 ft. between centers. Guide
rollers are placed with about 40 ft. between centers along both upper
and lower belts. (See Fig. 2.) The majority of manufacturers of
belt conveyors recommend the maximum length between centers of
a single belt to be about 700 ft. to 800 ft.
4 Referring to Fig. 1, the belt conveys the material down-hill,
and to this fact is due the apparently parodoxical results in power
required to operate, shown in Tables 1 and 2.
5 Because of the extreme length of the belt, and the fact that
there is no roof or other covering, it was necessary to install some
system for taking care of the expansion and contraction, in addition
Presented at the Spring Meeting, Washington, May 1909, of The American
Society of Mechanical Engineers.
152
A UNIQUE BELT CONVEYOR
to the ordinary stretch of the belt, which is taken up in the majority
of installations by 24-in., 36-in."^or 48-in. takeups, according to
length of belt. A set of 36-in. takeups, (Fig. 3) was installed at
the upper end of each of these belts to maintain alignment and equal
Shale Quarry
400
^'^1.446.65
300
^
El.268^38^^^
200
El.227.73
1
"ElTlolT
10 12 14
ELEVATION
PLAN
Fig. 1 Profile Showing Elevation and Plan of Conveyors
tension on each edge of the belt. The system installed acts as a
tension carriage and makes it less often necessary to cut out the
-4'o—
Tojp or Carrying Belt \
Lower Belt-
FORWARD IDLERS
-36'^
RETURN IDLERS
Fig. 2 Details op Forward and Return Idlers
slack in the belt , and in cool and wet weather the belt adjusts itself, the
increased tension due to contraction raising the weight in the tower.
A 10-h.p. motor drives each section. The lower section has a 6-ft.
drop and requires approximately 5. Ih.p. to operate empty and 5. Ih.p.
A UNIQUE BELT CONVEYOR
153
«
o
s
O
O
154
A UNIQUE BELT CONVEYOR
Fig. 4 View of Discharge from Fig. 5 Looking Down onFirst
Upper to Lower Section
or Lower Section
Fig. 6 "^Side View of Upper Section
A UNIQUE BELT CONVEYOR
155
to carry a load of 1200 lb., as shoveled by ten men. (See tests which
follow.) The discharge from the upper to the loAver sections through
a chute is shown in Fig. 4. There is no spilling of material at any
point of the travel, and pieces of shale a cubic foot in size are carried.
The upper section is driven, contrary to practice, at the upper end,
the pull being on the lower or slack side of the belt, but in this case,
due to the pull of gravity on the top side, the belt was found to work
better with the pull on the lower side.
Fig. 7 General View Showing Both Belts
6 The several halftones give views of the belt taken from different
points. In clearing a way through the woods, the poles obtained were
utilized for trestling and the planking was obtained from the scrap
pile of concrete-form lumber.
7 Fig. 6 is a side view of the lower end of the upper section, show-
ing the two depressions in the belt, and though these depressions do
not conform closely to the prescribed radius of 300 ft., there is no
lifting of the belt from the carrying idlers.
8 Power tests were made on the two sections after the belt had
been operating a few days, with the following results; the speed of
156 A UNIQUE BELT CONVEYOR
the belt of the lower section, which has a grade of 2.4 per cent for 665
ft., or 0.024 X 665 = 16 ft., was 146 ft. per min.; the belt was driven
by a 10-h.p. direct-current Westinghouse motor, and was loaded 2.2
lb. per ft. for a distance of 550 ft., or 1210 lb.; this load fell 16 ft. in
5 min. Then
= 3520 ft.-lb. of work exerted bv load
5
or,
= — h.p. (approx.) helping to pull the belt.
33,000 11 ^ ^^ ^ ^ ^
When the belt was loaded as above, a test of the motor showed that
16 amperes, 239 volts, or 5.1 h.p., were required. There was no
appreciable difference in the ammeter and voltmeter readings, when
belt was empty or loaded, as in test.
9 When the belts were installed, after trying them out and ascer-
taining how easily they could be operated, a sprocket was placed on
the tail-shaft of the lower section and also one on the head-shaft of
the upper section, and the two sprockets were connected by a vertical
quarter-twist chain. The idea was to drive both belts by a 10-h.p.
motor at the head of the lower belt section, after all shafts had
become well seated in the bearings and the stiffness had disappeared
from the belt and it was in good operating condition. This was also
necessary in order to take up the slack in the upper section when
starting, and the speeds were such that the top side of the belt ran 3 ft.
per min. faster than the lower side. The results of a series of tests are
given in Tables 1 and 2.
TABLE 1 POWER TESTS OF BELTS UNDER CONDITIONS NOTED IN TEXT
Time
Volts
Amperes
Watts
H.p.
Notes
(A.M.)
9:50
10:08
207
210
12
12
2484
2520
3.3
3.3
/ Belts chained together
\ Eight men loading
10:09
208
14
2912
3.9
/Connecting chain off, 10-h.p.
10:11
210
14
2940
3.9
1 motor only
10:15
10:20
10:35
200
200
200
14
15
16
2800
3000
3200
3.7]
4.0[
4.2J
Gradual increase in electrical load
due to decrease in shale load
Note: Low voltage due to very small mains and long distance (2500 ft.).
A UNIQUE BELT CONYEVOR
167
TABLE 2 SECOND SERIES OF TESTS
Time
(P.M.)
Volts Amperes
Watts h.h.
i
Notes
2:00
2:15
2:25
2:35
3:45
3:50
194
180
182
186
195
185
16
16
18
18
14
19
3104
2886
3275
3348
2730
3515
4.1
3.8]
4.4 [
4.4J
3.6
4.7
Empty. Connected to lower belt by
chain
All readings at motor and not in-
cluding line loss
Loaded by seven men
Loaded as before, but with connect-
ing chain off. 10-h.p. motor only
Note : Readings taken on motor at upper end of upper belt-section.
Initial and Operating Costs
10 Tables are given herewith upon the first cost of the equipment
(Table 3) and the cost of operation and maintenance (Table 4).
Table 4 is based upon a capacity of 200 tons conveyed in ten hours.
Inasmuch as the capacity is directly proportional to the speed, if it
was desired to increase the capacity of the conveyor, it would only
be necessary to increase the travel of the belt per minute, and from
experience, it is quite possible that by doubling the load the power
required to operate would be reduced 50 per cent.
11 The operation costs given in Table 4 are taken from actual
practice. Doubling the capacity per day and assuming above costs
to be approximately the same -reduces the actual cost of conveying
to S0.0038 per ton. Interest and depreciation, $0.0063, or a total
of $0.0101.
TABLE 3 COST PER FOOT OF COMPLETED BELTS INCLUDING ELECTRICAL
MOTORS. TRESTLING. ETC.
Uatebials
Total Cost
Cost pbb Ft.
Lumber
S 496.34
5361.52
1435.77
637.11
193.20
962.20
$0,238
Belt
2.58
Castings
0.69
Electrical equipment, including two 10-h.p. motors
Miscellaneous: nails, bolts, screws, iron, etc
Labor
0.316
0.093
0.46
S9106.16
$4.37
Note: Length of first section, center to center, 998 ft.; second section, 1082 ft.; total,
2080 ft.; takeup, 15 ft.
Cost of castings includes machine work, etc.
158 A UNIQUE BELT CONVEYOR
12 Regarding the operation of the belt: after the stiffness had
disappeared there was very httle slipping at the head or drive pulleys,
and there was sufficient lubrication in the shale itself to form a water-
proof covering about J-in. thick on the belt, thereby protecting it not
only from wear but from the action of the elements, and proving a
very good dressing to keep the belt pliable. Because of the slow
speed, etc., there are very few repairs necessary to the belt, and in
this instance, being coated as described above, the belt should last
several years.
TABLE 4 COST TO OPERATE AND MAINTAIN BELT CONVEYOR
PER 10 HR. pgjj ^Q^
DAY
Power
lOh.p. at$0.004perh.p.-hr $0.40 $0,002
Labor
Boy oiling, etc $0 . 75
Taking up slack once in 7 days, 2 men, 3 hr. at
$0.20perhr 0.171 0.92 0.0046
Supplies
Belt Lacing 0 . 10
Waste, Resin, etc 0.10 0.20 0.001
Total $1.52 $0.0076
Oil (no charge, using waste oil from large crushers) .
Interest, etc.
Interest, Depreciation, Renewals, 10 per cent on
investment of $9200 2.52 0.0126
Grand Total $4.04 $0.0202
DISCUSSION
T. A. Bennett. Mr. Soper's paper, while giving a practical descrip-
tion of a certain installation, hardly seems to warrant the word
" unique." There are many conveyors in regular practice just as
long — conveyors which run downhill — in fact, conveyors that need
a brake. As for size, there is a 36-in. movable belt conveyor in New
York over a thousand feet long, used for filling in the refuse from the
A UNIQUE BELT CONVEYOR 159
city on Hiker's Island. The driving arrangement mentioned, from
the receiving end of the conveyor, is also common practice.
2 Regarding the maximum length of conveyors, with a flat belt,
as in tliis instance, the limit would be merely the cost of installation as
practically any tensile strength desired can be obtained by increasing
the number of plies of the belt. With a troughed belt the hmit of
length would be the tensile strength of the thickest belt that would
conform to the trough of the idlers. This limit approaches somewhat
the Umits the author mentions, although such belts have been put in,
as above, for lengths of a thousand feet or more.
3 The take-up has been in use for over five years in belt-conveyor
practice. There is one installation at Bilbao, Spain, handling iron
ore, which runs down an incline of 13 deg. and needs a brake, and has
a counterweighted take-up working in a vertical plane. The take-up
and drive are located on the return belt near the foot of the incHne.
4 The tonnage of the conveyor is so small that the cost of mainte-
nance per ton is also misleading. The wear of a belt is occasioned by
the material coming in contact with it when dehvered to it. A
narrow stream of material permits each particle to come in contact
with a small proportion of the total width, whereas a wide stream
utilizes the full width of the belt and furthermore carries a large part
of the material on top of the belt without ever touching it. I beUeve
the capacity of this conveyor is something like 20 tons per hour,
whereas such a conveyor should easily handle 200 tons per hour.
Therefore the average cost for maintenance of the belt per ton carried
is high.
Harrington Emerson. The last words in this paper are "the
belt should last several years." In the last Hne of Table 4, it is
stated that interest, depreciation and renewals amount to 10 per cent
on an investment of $9200. Now, if the belt is to last only a few
years, 10 per cent is not sufficient to cover interest, depreciation and
renewal. Assuming that the belt lasts four years, the depreciation
account alone would be S2300. That would increase the cost per
ton from $0.02 to about $0,035.
Fred J. Miller. It might be well to consider that there are other
things that constitute part of this plant as well as the belt. As I
understand, it is stated that the belt may last only a few years, but
the rest of the plant may last enough longer than ten years to make
10 per cent a fair total charge for depreciation.
160 DISCUSSION
The Author.* The belt conveyor in question has now been in
operation about eighteen months, in which time less than $3 in
repairs has been expended. The belt itself shows little wear and
should last ten years. Of course the driving and carrying mechan-
isms will last indefinitely under ordinary conditions, as there is little
wearing of the working parts, due to the slow speed of the belt.
2 The installation, as stated previously, though ample for a
capacity of 200 tons per hour, is required to carry not over 20 tons
per hour, and certainly the cost per ton for maintenance and other
charges is out of proportion to what it would be were the belt carry-
ing an5rthing like full load.
3 The motor at the receiving end of the upper belt has been taken
out and the belt has been driven for several months by the "lower"
10-h.p. motor, the "upper" belt being driven by the "quarter-twist"
chain mentioned in the paper. At the time of writing the paper,
this was to our knowledge, the longest single-driven belt conveyor
(about 2150 ft.) in operation.
4 The writer has since learned of a slightly longer belt carry-
ing grain. Its speed is 1800 ft. per min., and hence a much greater
power is necessary. As to the power required to operate it, it
is reasonable to assume that if it takes about 6 h.p. to operate
the belt when empty, and 3 h.p. when loaded by 20 men, the
belt will practically run itself when loaded by 40 men, and will
require a band-brake when loading is increased above this number.
^ Abstracted.
No. 1235
AUTOMATIC FEEDERS FOR HANDLING
MATERIAL IN BULK
By C. Kbmble Baldwin, Chicago, III.
Member of the Society
In the writer's paper on the Belt Conveyor, read before the Society
in June 1908, mention was made of the advisability of using some
type of automatic feeder when feeding a conveyor from bulk, for
example, from a storage bin. This brief mention of the automatic
feeder brought so many inquiries for information ^regarding feeders
for various materials that this paper has been prepared in order to
present a brief description of the various types now in use. The
cuts are not intended to show the construction, but to illustrate the
principle involved, so that they may be compared.
2 Careful study of this subject reveals the fact that a particu-
lar type of feeder has been developed in a certain industry, or local-
ity, and is little used elsewhere. The types illustrated and described
below are only those which have come under the writer's personal
observation in many processes and locations within the past fifteen
years. There may, therefore, be many other types.
3 When handling dry, free-flowing material such as grain, from
a storage bin to a conveyor, a feeder is not required, as a simple gate
may be set to give the desired opening, thus allowing the proper
quantity to flow from the bin. Should the material be of varying
size, such as mine-run coal, a simple gate is not satisfactory unless
constantly attended; even then it is impossible to get the same con-
stant, regular feed that a properly designed feeder gives. If the gate
is raised high enough to allow a large lump to pass, there usually
results a rush of fine material, which floods the conveyor before the
gate can be closed. The automatic feeder, therefore, not only saves
the expense of an attendant, but insures a constant and regular feed,
irrespective of the size of the material.
Presented at the Spring Meeting, Washington, May 1909, of The American
SociETT OF Mechanical Engineers.
162 AUTOMATIC FEEDERS FOR HANDLING MATERIAL
*-UrtDERCUT G>\T£
Fig. 1 Undercut-Gate Feeder
Fig. 2 Lifting-Gate Feeder
SHftFT FOR GtftR OR SPRQCKE
Fig. 3 Screw-Conveyor Feeder
AUTOMATIC FEEDERS FOR HANDLING MATERIAL 163
4 Fig. 1 shows the undercut-gate feeder, with a body either of
cast-iron or steel plate. Pivoted near the top is the undercut gate —
which is swung back and forth by a connecting rod from crank or
eccentric. This type of feeder is best adapted to fine-sized, free-flow-
ing material. Material containing lumps is likely to bridge. As the
feed is intermittent, the feeder is generally used in connection with
chain or bucket conveyors, the strokes being timed to feed material
between the flights, or into the buckets. The capacity may be
changed only by changing the length or the number of strokes. As
the length of stroke is more easily changed, it is preferable to use a
crank with an adjustable throw rather than an eccentric. P^lliptic
ger.rs are sometimes used to give a quick return, but in practice
this quick return has not been found of sufficient value to justify
the t-pi'cial and more expensive gears.
5 The lifting-gate feeder, shown in Fig. 2, also gives an inter-
mittent feed and is therefore used only with a chain or bucket con-
veyor. The chute is hinged, so that when down, the material will flow
out of the hopper, but when raised above the angle of flow of the
material, the discharge is stopped. The moving of the chute may be
accomplished by a connecting rod receiving motion from either crank
or eccentric. This feeder will handle material regardless of size, but
it must be free-flowing material, so that it will move by gravity when
the chute is lowered to the angle of flow. The capacity may be
adjusted by varying the number of strokes, also, in a measure, by
increasing the length of the stroke, thus increasingthe maximum angle
of the chute and causing the material to flow more quickly.
6 The screw-conveyor feeder, illustrated in Fig. 3, will deliver
a constant stream of material, but in this case also it must be of such
a nature that it will flow by gravity to the screw. The capacity can
be changed only by altering the speed of the screw shaft. This type
of feeder has a large field in the handling of pulverized material, such
as coal, cement, etc.
7 The roll feeder, shown in Fig. 4, is extensively used in the mineral
industries for handling both large and small materials. The roll is
so located under the hopper that the material will not flow when the
roll is stationary, but when rotated it will carry the material forward.
The capacity is determined by the speed and width of the roll, and
the thickness of the stream, as fixed by the adjustable gate.
8 The roll feeder has been successfully used in handling iron-ore,
coke and stone from the bins to the weigh cars for furnace changing.
Edison used this type for feeding ore and stone from bins to crush-
ing-rolls. The disadvantage is the head-room required, owing to
164
AUTOMATIC FEEDERS FOR HANDLING MATERIAL
/
♦
'
ROLL-> =
E
rsc-i
j^--**--jr
=
.,--^^— -^-^-v-^
Fig. 4 'Roll Feeder
the large roll necessary to satisfactory operation. For handling mine-
run material, the ^oll should be 6 ft. to 8 ft. in diameter and in many
cases it is not possible to obtain this space.
9 The rotary-paddle feeder, Fig. 5, acts not only as a feeder, but
as a measuring device. It is used for fine material which flows
readily from the blades. The capacity is fixed by the speed of the
paddle shaft.
10 The revolving-plate feeder, shown in Fig. 6, is used mostly
for feeding stamp-mills. The inclined plate driven by gears, placed
either above (as shown) or below, moves the material out of the hopper
Fio. 5 Rotatino-Paddle Feeder
AUTOMATIC FEEDERS FOR HANDLING MATERIAL
165
RtVOLVING Pl-BTt
Fig. 6 Revolving-Plate Feeder
where it is scraped off by the skirt-board. When the skirt-board is
made adjustable, sticky material may be handled by this feeder
because the curved plate will scrape the material off the revolving
disc and into the chute. The capacity is fixed by the speed of the
plate and the location of the adjustable gate.
11 Fig. 7 illustrates the apron-conveyor feeder used for handling
material of all sizes. The conveyor may be of any of the various
types of apron flights, depending on the nature of the material handled.
The chain should be provided with rollers or wheels traveling on
CUWVCO ftPBOM FLICHT
Fig. 7 Apron-Conveyor Feeder
166 AUTOMATIC FEEDERS FOR HANDLING MATERIAL
track to prevent the apron from sagging. The capacity is fixed by
the speed of the apron and the position of the adjustable gate.
12 The disadvantage of this type is the inherent disadvantage
of the apron conveyor. If the flights become bent or buckled,
the material leaks through or catches between them. It has an
advantage over other feeders in that it may be used to carry the mate-
rial a greater distance.
13 A rubber or canvas belt may be used in place of the apron, in
which case the belt is supported by idlers placed close together.
14 The swinging-plate feeder, shown in Fig. 8, is used for handling
coal and such material of all sizes. It consists of two castings pivoted
at their tops and swung alternately so as to move the material forward
on the bottom plate. The plates are moved by connecting-rods
\ COHMtCTIM
Fig. 8 Swinging-Plate Feeder
from a crank or eccentric through a rocker shaft. The capacity is
fixed by the length and the number of strokes, but as it is limited to
the amount of material displaced by the plates, a wide range is not
possible.
15 The disadvantages are the lack of adjustability and the
tendency of the material to pack. It will also be noted that the
feeder is not self-cleaning, so that the bottom plate always contains
material which is very liable to freeze in winter.
16 The plunger feeder, illustrated in Fig. 9, is similar in operation
to the swinging-plate feeder in pushing the material along the bottom
plate. The plunger may be built either in one or two parts, moving
ahead alternately and driven through a rocker shaft, as in the case
of the one previously described. The capacity is fixed by the number
and length of the strokes and the location of the adjusting gate.
AUTOMATIC FEEDERS FOR HANDLING MATERIAL
167
This type has the same disadvantages as the swinging plate feeder,
the most serious being that it is not self-cleaning.
17 Fig. 10 shows the reciprocating-plate feeder, consisting of a
plate mounted on four wheels forming the bottom of the hopper.
When the plate is moved forward, it carries the material with it, and
when it is moved back the plate is withdraAvn from under the material,
allowing it to fall into the chute. The plate is moved by a connect-
ing rod from crank or eccenti'ic. The capacity is determined by the
length and number of strokes and the location of the gate. The
disadvantages are the lack of adjustment and the inability to clear
the feeder of material.
18 The shaking feeder. Fig. 1 1, consists of the shaker-pan located
under the opening in the bottom of the hopper at such an angle that
the material will not flow when the pan is stationary. When given
HOPPER
./
nOJUSTflBLE GATE
DISC CHW1K I RBIL3 J^
SKIRT B0ffRD5
CONWECTinG ROD
Fig. 9 Plunger Fekder
a reciprocating motion by the crank and connecting-rod, the material
is moved forward on the pan. The front end of the pan is carried by
a pair of flanged wheels; the back end is suspended by two hanger-
rods, each being provided with a turn-buckle so that the angle of the
pan may be varied. The crank having an adjustable length of stroke,
there are three variables, viz: number of strokes, length of stroke;
and inclination of the pan. As the number of strokes is difficult to
change, and the others easily changed, the feeders are usually de-
signed for about 75 strokes per niin., a number determined by
experiment. The angle of the pan is fixed by the capacity desired
and the nature of the material handled. For coal, stone, ore, etc.,
8 deg. to 10 deg. is sufficient, while clay and other sticlcy substances
require from 15 deg. to 20 deg. The length of stroke varies from 4 in.
to 12 in., so that a large range is possible.
19 A feeder designed to handle 400 tons per hr. of mine-run coal
168
AUTOMATIC FEEDERS FOR HANDLING MATERIAL
Fig. 10 Reciprocating-Plate Feeder
was changed in five minutes to deliver 30 tons per hr., by shortening
the length of stroke and lowering the pan until nearly horizontal.
20 Not only has this feeder the widest possible range in capacity,
but it is self-cleaning, a very important feature. From the cut it
will be noted that the pan is placed under the opening and the
material rests directly on the pan, so that when the pan is moved the
material in the hopper is moved, which prevents the material from
bridging.
21 The shaking feeder has none of the disadvantages of the
other types for general use, and possesses many advantages which the
others lack. Owing to its great flexibility it is more easily standardized
Fig. 11 Shaking Febdeb
AUTOMATIC FEEDERS FOR HANDLING MATERIAL 169
and will successfully handle practically any material, regardless of
size or condition. If desired the bottom plate may be perforated to
screen out the fine material, thus acting as both screen and feeder.
This is not possible with any of the other types.
22 The power required by all of the types is so small that it is
not an important consideration. The shaking feeder mentioned
above, which handled 400 tons of coal per hr., required but 3.5 h.p.
23 The preceding cuts and descriptions will give a general idea
of the different types and their possible uses, so that an engineer may
readily choose the best type for the work to be done. The point that
should be kept in mind is, that it is always advisable to gear the
feeder to the conveyor, crusher, or other machine which it feeds so
that they will both start and stop simultaneously.
DISCUSSION
T. A. Bennett. An automatic feeder is absolutely necessary in
some installations; in others it is demanded for economic reasons.
For example, run-of-mine coal, on account of constriction in the chute,
requires properly a 36-in. belt, although it can be handled on a 30-
in. belt. With an automatic feeder it is possible to use a 24-in. con-
veyor provided the capacity will permit. In handling damp sand,
a large chute opening is necessary and this usually requires at least
a 16-in. conveyor belt. With a feeder, however, this width can easily
be reduced to 12 in. A 12-in, conveyor has capacity to take care of
nearly every problem in handling damp sand.
2 The feeder is also economical in filling in the blank spaces on
the belt. The loading is usually intermittent, the belt being either
over-loaded or under-loaded intermittently; the feeder can be regu-
lated to give a uniform maximum loading, greatly increasing the
capacity. Intermittent loading also increases the wear on the belt.
As the only wear worth considering is that done by the material
coming in contact with the surface of the belt in the delivery of ma-
terial to it, the larger the load the less is the proportionate wear on
the belt.
3 A type of feeder has been developed similar to that in Fig. 8,
but doubled; that is, a hopper and either one or two swinging plates at
each end and pushing to an outlet at the center part of the skirt
boards. This type is now working very satisfactorily in two large
plants, the Hudson Company's power house at Jersey City, N. J.,
and the Illinois Steel Company at Joliet, III. The chief advantage
over any other is the saving of headroom. Where there is sufficient
headroom, the shaker feeder is correct practice and is in general use.
170 DISCUSSION
The Author. Mr. Bennett's discussion emphasizes three impor-
tant points with reference to the use of automatic feeders in connection
with belt conveyors:
a The installation of the feeder frequently permits the use of
narrower belts.
b Delivery of an even and continuous stream of material en-
ables the conveyor to operate at its maximum capacity.
c By loading the conveyor to its full capacity, a smaller pro-
portion of the load comes in contact with the belt, there-
by reducing the wear per ton carried on the belt at the
loading point.
2 The type of feeder mentioned by Mr. Bennett as similar to that
illustrated in Fig. 8 is a variation of that type, used where material
is taken from two hoppers; instead of tAvo swinging plates placed side
by side, there is a single plate under each hopper. These two plates
are connected by rods, so that when one plate is in the forward stroke
the other will be in the back stroke. It is therefore nothing more
than two single-plate feeders so connected that they operate together.
3 Mr. Bennett is mistaken regarding this type saving headroom
over any other type. It requires about the same headroom as the
plunger feeder (Fig. 9) and the reciprocating plate feeder (Fig. 10).
The great drawback of the swinging-plate feeder is that it is not self-
cleaning, so that if exposed in winter the material will freeze to the
bottom plate.
4 This type of feeder was originated by Mr. Lincoln Moss of New
York, and the two installations mentioned by Mr. Bennett were
designed by Mr. Moss under the writer's direction.
No. 1230
A NEW TRANSMISSION DYNAMOMETER
By Pkof. Wm. H. Kenerson, Providence, R. I.
Member of the Society
The author has received from time to time many requests for a
simple transmission dynamometer, and has himself often felt the
need of one which would be more generally applicable than those now
in use. These continued requests, together with the requirements
of a definite problem whose solution demanded a rigid transmission
dynamometer in the form of a coupling, led to the design and con-
struction of the instrument described below. The accompanying
illustrations show the construction of the dynamometer and its
method of application and use. In Fig. 2 and Fig. 4 the correspond-
ing parts of the dynamometer are given the same letters and are
referred to in the text.
2 The couplings A and B, each keyed to its respective shaft, are
held together loosely by the stud bolts C. The holes in the flange A
are larger than the studs C, so that these studs have no part in trans-
mitting power from one shaft to the other. The power is trans-
mitted from A to B through the agency of the latches L, four of
wliich are arranged around the circumference of the flange B. These
latches are mounted and are free to turn on the studs E. The two
fingers of the latches engage the studs F on the flange A. On the
ends of each latch are knife-edges parallel to the stud about which
the latch turns. For either direction of rotation of the flange A
the latches L, which are in effect double bell-crank levers, will exert
a pressure on the disc G, tending to force it axially along the hub of
the coupling B, and this pressure, it will be seen, is proportional to
the torque.
3 Between the end-thrust ball, or roller, bearings M M, is held
the stationary ring S, which is the weighing member. 0 is a thrust-
collar screwed on the hub of B, and P is its check nut, which is ordi-
Presented at the Spring Meeting, Washington, May 1909, of The American
Society of Mechanical Engineers.
172
A NEW TRANSMISSION DYNAMOMETER
narily pinned to the hub when in position. The stationary member
S, in the form of a ring surrounding the shaft, is prevented from
rotating by fastening to some fixed object the attached arm shown
in the view (Fig. 1) of the assembled instrument. In this ring is an
annular cavity covered by a thin, flexible copper diaphragm D,
against which the ball-race of one of the thrust bearings presses.
The edge of this ball-race is slightly chamfered to allow some motion
Fig. 1 Dynamometer for 2-in. Shaft, Weight 60 lb.
to the diaphragm. The cavity is filled with a fluid, such as oil, and
connected by means of a tube to a gage. The oil pressure measured
by the gage is proportional to the pressure between the thrust- bear-
ings, which in turn is proportional to the torque.
4 The instrument may be calibrated in the torsion-testing machine
or by means of a sensitive friction brake. Fig. 6 is an actual cali-
bration curve for a small instrument, obtained by hanging standard
A NEW TRANSMISSION DYNAMOMETER
173
weights at proper distances from the shaft on a horizontal lever
attached to the shaft, and reading the pressures indicated by the
gage for the various torques shown in the diagram. For ordinary
purposes, however, it is not necessary to calibrate the instrument by
actual trial, since computations of the oil pressures for the various
torques from the lengths of the lever-arms and diaphragm area
check very closely those thus obtained.
5 It will be seen that the weighing means is similar to that
employed in the Emery testing macliine, which is recognized as being
extremely accurate. It will be possible to employ the Emery flexible
Fig. 2 Dynamometer Shown in Section
steel knife-edges on the levers, if desired, but this has been found in
practice an unnecessary refinement.
6 The construction makes the coupling as nearly rigid as materials
will permit, the movement of the diaphragm being extremely small.
The only flow of oil through the copper connecting pipe is that suffi-
cient to alter the shape of the Bourdon tube, if that be the form of
gage employed. As soon as the normal position of the gage is reached
this flow ceases, hence there can be no fluid friction. It is possible
therefore, to use as long and as small a tube as desired, without intro-
ducing error. Where the gage is placed at a distance above or below
the coupling, correction should of course be made for the static head.
174
A NEW TRANSMISSION DYNAMOMETER
E^
A NEW TRANSMISSION DYNAMOMETER
17.5
7 Other means than the gage shown may be employed to measure
the fluid pressure. Where extreme accuracy is desired it will be well
to employ the weighing device used with the Emery testing machine.
The manograph has been used in this connection to measure varia-
tions in torque too rapid for indication by the ordinary gage. For
example, the variations in torque in a single revolution of the shaft
of a 3-cylinder gasolene engine have been recorded with its aid.
8 Where the rate of rotation of the shaft is variable and it is
Fig. 5 Dynamometer Placed between Flanges in Machine-Shop Drive
3-lNCH SHAFT. SPIRAL RUNNING TO THE WALL IS OIL PIPE TO GAGE
desired to indicate the horsepower direct, the combination of gage
and tachometer shown in Fig. 7 is employed. The hydraulic gage
is connected to the coupling described, its pointer therefore indicat-
ing torque. The pointer of the tachometer shows the number of
revolutions per minute. Being a function of the revolutions per
minute and the torque, the horsepower will be indicated by the inter-
section of the two pointers and suitable curves on the dial as shown.
Arrangements for recording or integrating the work done may also be
attached to the coupling.
176
A NEW TRANSMISSION DYNAMOMETER
50
■g ^
a
"i *^
u
K 35
'^-•30
f«
Sf 20
1
1
/
[/
/
(
/
(
y
/
/
S 15
p.
/
o 10
1/
5
/
0 250 500 750 1000 1250 1500 1750 2000 2250 2500
Torque, iiic'h-poumls
FiQ. 6 Calibration Curve for Transmission Dynamometer,
Fig. 7 Combination Pressure Gage and Tachometer Indicating Torque
Revolutions per Minute and Horse Power
A NEW TRANSMISSION DYNAMOMETER 177
9 A summary of some of the more important characteristics of
the instrument follows:
o The instrument is compact. The example shown in Fig.
3 and Fig. 4, which is designed to transmit 30 h.p, at 500
r.p.m., is about 5f in. in diameter and weighs about 25
lb. That shown in Fig. 5 driving a 3-in. shaft is about 13
in. in diameter and weighs about 160 lb.
b It is as rigid as an ordinary flange coupling.
c It may be made in the form of a coupling, and will then
occupy about the same space as the usual flange coupling,
or it may be made in the form of a quill on which a pulley
is mounted. This form may be made in halves for appli-
cation to a continuous shaft.
d It will indicate for either direction of rotation of the shaft.
e The torque may be read and recorded or the work inte-
grated at a considerable distance from the coupling.
/ The readings do not require correction for different speeds
of rotation. All parts containing oil are stationary, hence
are unaffected by variation in speed. Other parts are
likewise unaffected by centrifugal action.
g It may be made very sensitive and accurate. The construc-
tion lends itself very easily to variation of range of appli-
cation and to varying degrees of sensitiveness, since the
oil pressure, and hence the sensitiveness of the instrument,
depend upon the area of the diaphragm, the relative
lengths of the arms of the latches L, and the diameter of
flanges. Its accuracy is dependent mainly on the degree
of accuracy of the means employed to measure the fluid
pressure, of which a number of forms, other than the
usual pressure gage, are available.
h The only power absorbed is the small amount due to the
friction of the ball, or roller, bearings, and this can be
determined from the pull of the retaining arm. It is
unnecessary to make correction for this, however, since
the amount is so small as to be negligible.
i Since the only wearing parts are the ball, or roller, bearings,
which may be lightly loaded, the instrument should not
be deranged easily. Because of the very small volume of
oil contained in the weighing chamber, ordinary tempera-
ture changes do not affect the calibration. All parts con-
taining oil are stationary, hence all joints may be soldered
and leakage entirely prevented.
178 DISCUSSION
j With suitable material and ordinary workmanship, it is
believed that there is little likelihood of failure of any
part of the instrument. It is conceivable, however, that
the balls or rollers, although lightly loaded, might -crush;
the diaphragm might shear; or the stationary member,
although bearing only its own weight and lubricated,
might seize to the hub. Remote as are any of these
possibihties, should any or all of them occur, the worst
that coul'l happen would be the tearing-off of the oil pipe
and retaining arm, when the whole would revolve as a
solid coupling. In no case can the coupling fail to drive
the shaft because of its variation from the standard form,
since, in addition to the driving latches employed to
carry the load normally, the same number of connecting
bolts may be employed as in the ordinary coupling, which
will still hold the coupling together should the latches
fail. Since, however, these latches are farther from the
shaft, they should, if properly constructed, be less likely
to fail than the connecting bolts usually emploj^ed.
10 It is believed that uses for the instrument here described will
suggest themselves, and it is with the hope that the device will prove
of some interest to those who deal with the use and transmission of
power that the matter is presented to the Society.
DISCUSSION
A. F. Masury. I want to say a few words as to how the Kenerson
dynamometer may be applied to the betterment of design in motor
vehicles. In the first place, we must have exact data regarding the
effect of road irregularities, wind pressures, and the resistances set
up by grades and speed. These figures are absolutely necessary in
order to determine the best torque to apply on motor gearing and
equipment, such as tires, etc., to overcome the existing conditions in
each particular car.
2 At present we have two recourses: first, the figures procured
by Mr. S. F. Edge at the Brooklands track in England with his Napier
car. He first calibrated his motor and then made the test on the
track. These figures must necessarily include many errors. Second,
the dynamometer at the Automobile Club of America in New York.
Here again satisfaction is not entirely procurable as all the road con-
ditions are obtained artificially by attachments on the machine.
A NEW TRANSMISSION DYNAMOMETER 179
3 With the Kenerson dynamometer we can certainly get exact
readings while the machine is working on the road. These can even
be made graphic if desired. Manufacturers wll thus have available
means of getting information which should result in more perfect
design.
4 There is one thing more, the dynamometer of the Automobile
Club of America cost, I believe, in the vicinity of S15,000 by the time
it was completely installed, while the price of the Kenerson machine,
around $500, will make it possible for even a small manufacturer to get
his own reading.
The autlior desired to present no closure. — EoiToit.
No. 1237
POLISHING METALS FOR EXAMINATION WITH
THE MICROSCOPE
Bt Albert Kingsbuby, Pittsbubq, Pa.
Member of the Society
In 1902 the writer made experiments to find the most suitable
method of polishing samples of metals for microscopic examination
The polisliing of the surface is one of the most important as well as
most troublesome details of metallography, particularly when high
magnification is required.
2 At the outset, trials were made of all the methods of which
descriptions have been published. Some of those methods have been
successfully employed by various metallographists, as shown by
numerous reproductions of excellent micro-photographs in different
publications. Nevertheless the writer did not find any of these free
from objectionable features. The ideal method should produce a
fairly flat surface, free from excessive relief of the harder constituents,
rounded edges at flaws, or scratches and smearing of the metal. The
method should be simple, the materials employed readily available,
and the process as rapid as consistent with the first-named requisites.
None of the published methods embodied all these requisites, nor is
a perfect method likely to be found. However, the method finally
developed by the writer appears to him superior.
3 The preliminary trials were made with rotating discs covered
with various materials, including canvas, felt, silk, leather, chamois,
parchment, paper, wood, pitch, asphalt, resin, shellac, beeswax, etc.
The polishing powders included commercial abrasives, such as emery,
carborundum, tripoli, crocus and jewelers' rouge; also precipitates,
such as carbonates and sulphates of the alkaline earths. Attempts
were made to obtain fine finishing powders by the levigation process
from commercial abrasives. These abrasives were tried both wet and
dry and with various speeds of the discs. Hand polishing was also tried.
Presented at the Spring Meeting, Washington, May 1909, of The American
Society of Mechanical Engineers.
182 POLISHING METALS FOR EXAMINATION
It is needless to detail the objectionable features encountered, which
are probably familiar to all metallographists.
4 The method finally adopted was the result of two distinct dis-
coveries: {a), that ordinary paraffin Avax makes a good polishing bed;
(6), that excellent polishing powders of certain grades are commer-
cially available.
5 The paraffin is used as a facing for rotating discs of metal,
preferably brass, about 8 in in. diameter. The discs are grooved on
the flat face for anchoring the wax. To prepare the discs, they are
warmed to about 100 deg. cent., and laid flat, and the melted paraffin
is poured on them to a depth of about ^ in., a removable ring or band
retaining the melted wax. The whole is tli^n covered to exclude
dust and allowed to cool. After the wax has solidified it may be
dipped in water to hasten the hardening. Since the wax has very
little viscosity when melted, all hard foreign particles, which might
) )roduce scratches in the samples, settle out before the wax hardens,
the elimination being practically complete. No advantage in this
respect was gained by keeping the wax in a fluid condition on the disc
for several hours in an oven. After the hardening of the wax the
discs are placed on the spindle of the polishing machine and the face
of the wax is turned true and flat by a hand-tool.
6 In the writer's machine the spindle was horizontal and four
discs were used for abrasives of progressive fineness, two discs being
placed back to back at each end of the spindle. The disc used for the
final polishing should not be perforated and the wax should be con-
tinuous to the center of the disc, as that part is best for the finishing
touches to the sample. This latter disc should be at the right-hand
end of the spindle. The speed of rotation should be about 200 r.p.m.;
a higher speed throws off the polishing powder with the water used,
and a lower speed makes the work too slow. A stationary sheet-
metal strip about 3 in. wide bent over the discs serves as a screen.
7 The polishing powders, in the order used, were as follows: (a)
commercial flour of emery; (6) washed Naxos emery, 3/0 grade; (c)
washed Naxos emery, 7/0 grade; (d) soft optical rouge, light grade.
These were obtained from the George Zucker Co., New York, except
the first, which is available everywhere.
8 The emery powders were mixed to a paste with water in tall
glass jars provided with covers; the paste was apphed to the rotating
discs with small brushes as required, the brushes being kept in the
jars when not in use. The rouge was in cake form, best applied by
holding a small piece in the hand, wetting both the rouge and the wax,
and pressing the rouge lightly against the rotating surface.
POLISHING METALS FOR EXAMINATION 183
9 A small quantity of water is required throughout the polishing
process, but water cannot be used very freely without wasting the
powders. The water is best applied as required, from an ordinary
chemist's wash-bottle, held in the left hand while the right hand
manipulates the sample. No water pipes or drains are required for
the polishing machine. Distilled water may be used if available.
If tap water is used, it should be drawn into large jars provided with
covers and siphons, and allowed to stand a day or more before use,
in order that all gritty particles in suspension may be deposited.
The inner ends of the siphon tubes should be at least 3 in. above the
bottom of the jars.
10 The treatment of the samples is as follows: the samples are
first di-essed to shape and size' by any convenient method, the surface
to be polished made flat by an emery wheel or file, and the sharp
edges rounded to prevent cutting into the wax. The dimensions of
the samples should depend to some extent upon the coarseness of
structure. For normal iron and steels, and for much other work,
a f-in. cube is a convenient sample. Massive castings sometimes have
grains an inch or more in diameter, and correspondingly large samples
are rec^uired. The samples are held flat against the waxed discs,
which are kept well covered by the polishing paste, using successively
the flour of emery, the 3/0 emer}^, the 7/0 emery, and the rouge, on the
several discs. At each grinding with emery the sample should be
held without rotation and with a slow transverse motion across the
face of the disc until the grinding marks show over the entire surface.
The sample may then be given a quarter turn, so that the new marks
cross the old ones, and so on. The discs must be kept wet continu-
ally while grinding. With each grade of powder the grinding should
continue for some time after the marks of the last previous grade
have disappeared, especially with soft metals, since the scratches
cause a flow or disturbance of the metal to a minute depth below the
surface, and if this disturbed metal is not ground off, the deep effect
of the scratches becomes apparent on etching. In the final polishing
on the rouge disc, the sample should be continuously rotated; this is
most readily done by moving the sample nearly in a circle about the
center of the disc in an opposite direction from the rotation of the disc.
This keeps the direction of the grinding marks constantly changing,
and avoids grooving. The finishing should be done near the center
of the disc, the slower motion being most effective for very fine polish-
ing. After grinding with one grade of powder and before proceeding
to the next, the samples and the operator's hands should be thor-
184 POLISHING METALS FOR EXAMINATION
oughly washed; and the hands and the apparatus should be kept free
of dust or dirt, to secure a polish free from scratches.
11 The most important item to be noted by the beginner is the
liability of the paraffin to adhere to the samples when the grinding
is begun, particularly in the case of the rouge disc. When the sample
is first brought into contact with the disc, especially if the latter has
been freshly prepared, the paraffin nearly always smears over the
surface of the sample in a second or two, and if the sample is not
removed and cleaned at once the result is a roughened disc, requir-
ing re-turning with the hand-tool and re-application of the paste.
Therefore the sample should at first be touched very lightly to the
disc, and at once removed and wiped with the finger, or with a cloth.
If this is repeated several times, the surface of the sample will no
longer become coated with paraffin but can be ground continuously,
except when a fresh coating of paste is required by the disc. One
great advantage of the paraffin disc over discs covered with cloth or
felt, is that if the disc becomes roughened or cut, it can readily be
turned smooth and true again.
12 For cleaning the samples after polishing, the best material is
a stock of old linen or cotton cloth well-laundered and cut to 3-in.
squares. These small pieces are preferable to larger ones, since they
can be discarded for fresh ones after once using. The old cotton or
linen is also the best material for cleaning the lenses and mirrors of
the optical apparatus, being superior to chamois for this purpose.
13 The time required for polishing a sample varies somewhat
with the hardness. A single sample of normal steel, cast iron, or
wrought iron, may be finished in fifteen minutes; a set of five or six
such samples may be finished in an hour. Hardened steels require
a slightly longer time. The method has not thus far proved service-
able for very soft metals and alloys, particularly lead, owing to the
persistent adhesion of the paraffin to the surface of the sample.
The harder alloys polish well by this process. The finished surface
presents a minute relief of the harder constituents, but much less
than is produced by the use of felt or other very soft materials.
14 The paraffin beds are more durable than might be supposed;
on long standing at summer temperatures the surfaces become dis-
torted by the flow of the wax, but they can always readily be made
true by the turning tool. The harder paraffin (ceresin) offers no
advantages over ordinary paraffin, except that it flows less at sum-
mer temperatures. It is serviceable for use with the emery powders
but too hard for best results with the rouge.
No. 1238
MAKINE PRODUCER GAS POWER
A COMPARISON OF PRODUCER-GAS AND STEAM EQUIPMENTS
By C. L. Steaub,* New York
Non-Member
So much interest is exhibited both by the engineering profes-
sion and the general public in the application of producer gas power
to marine, commercial and naval service, that a brief summary of
recent progress in this field appears timely.
2 Any innovation which makes for improvement in present
practices, surely, though sometimes slowly, achieves its end. Pro-
ducer gas power, on impartial analysis, offers so many benefits to
marine service that it appears strange indeed that more rapid prog-
ress has not been made in its adoption. The delay appears to be
due to several causes.
3 The marine public, which since the days of the Clermont has
exclusively associated the term "motive power" with steam, has
every reason for demanding exact and conclusive evidence of the
superiority of gas power or any other power, before adopting it
in lieu of present methods. This evidence is only now slowly
coming forth. Many who have been credited with authority by the
engineering profession and others, either through ignorance or through
misinformation, have beset the way of marine gas power with
numberless imaginary obstacles, ridiculous in proportion to the real
difficulties, but sufficient nevertheless to instill some doubt of the
possibilities of the system into the minds of the waiting public.
4 Only recently has such progress been made in the development
of gas power for marine work, as to warrant its early adoption in
commercial service. Two years ago, less than 300 h.p. in the aggre-
gate was being developed by marine producer'gas power installations;
these were experimental in nature and were of the German Capitaine
' With the Loomis-Pettibone Co., New York.
Presented at the Spring Meeting, Washington, May 1909, of The American
Society op Mechanical Engineers.
186 MARINE PRODUCER GAS POWER
type. There are now installed and accepted 23 Capitaine marine
plants, aggregating 2035 h.p., a partial list of which follows:
a Emil Capitaine: Launch, 60 b.h.p.; 4-cylinder single-acting, 4-cycle
engine; boat 60 ft. long, 10 ft. beam, 4ft. draft; ran an average speed
of 10 miles for 10 hr. on 412 lb. of anthracite coal.
h Rex: Sea-going Swedish boat; 102 ft. long; 22 ft. beam, carries 350 tons
on 9-ft. draft; fitted with a 3-cylinder single-acting, 45-h.p. engine at
.300 r.p.m.
c Capitaine: Tow boat at. Genoa; length 47 ft., beam 12 ft., draft 7 ft.;
fitted with a 3-cylinder, single-acting, 4-cycle engine, 105 b.h.p. at
240 r.p.m.
d Duchess: Canal barge; length 71 ft., beam 7 ft. 1 in.; carries 20 tons
cargo on 42-in. draft; fitted with double-cylinder, single-acting, 4-cycle
engine of 25 b.h.p.
e Dusseldorf: Tug at Hamburg; fitted with a 4-cylinder, single-acting,
4-cycle engine, 60 b.h.p. at 240 r.p.m.
/ Isee: Tug, fitted with a 3-cylinder, single-acting, 4-cycle engine, 45
b.h.p., 300 r.p.m.
g Wilhclm: Combination freight and passenger Rhine boat, fitted with a
5-cylinder, single-acting engine, 175 b.h.p. at 240 r.p.m.
h Badenia: Rhine freight boat, fitted with a 2-cylinder, single-acting,
4-cycle engine of 30 b.h.p.
i Katrina: Canal freight boat, fitted with a 3-cylinder, single-acting,
4-cycle engine, 45 b.h.p.
i Marie: Canal freight boat; fitted with a 3-cylinder, single-acting,
4-cycle engine, 45 b.h.p.
k Hoffnung: Combination freight and passenger Rhine boat, fitted with
a 5-cylinder, single-acting, 4-cycle engine of 210 b.h.p.
I Amersie: Volga freight boat, fitted with a 4-cylinder, single-acting,
4-cycle engine of 60 b.h.p.
m No. 58: Canal freight boat, fitted with a 4-cylinder, single-acting,
4-cycle engine of 60 b.h.p.
5 In addition to the above there were a number of freight boats,
the dimensions and names of which we were unable to obtain, but
whose power plants varied in capacity from 30 to 175 h.p. each.
II H. M. S. Rattler: An old gun boat, 165 ft. long, 29 ft. beam, originally
fitted with a triple expansion engine. The gas engine is 5-cylinder,
single-acting, 4-cycle. Cylinders 20 in. diameter by 24 in. stroke,
developing 500 b.h.p. at 120 r.p.m. This engine is started by means
of a mixture of gas and air which is pumped into the cylinders at a
pressure of about 95 lb. per sq. in. This complete plant was designed
entirely in the Capitaine Works at Diisseldorf. The total weight
of the entire plant, including the donkey boiler for working the pumps
and auxiliaries, is 94 tons, as compared with 150 tons in the case of
the displaced steam engine. A consumption of 1525 lb. of coal was
made for a measured distance of 45 knots on an average speed of
MARINE PRODUCER GAS POWER 187
10^ knots per hr. The cost per mile for fuel with coal at 15s. Qd. per
ton is $0,064 U. S. currency. This boat made a maximum speed of
11.3 knots per hr. against a 1^ knot current at 110 r.p.m. of the
engine shaft.
6 All of the above plants by their design and construction are
restricted to operation on anthracite coal, coke or hard-burned char-
coal, and any plant so restricted by its design to one class of fuel is
seriousl}' limited in its scope of application. The development of a
simple marine gas-producer for use with any class of solid fuel is a
necessity, if the system is to be considered seriously by the marine
profession.
7 The writer is fortunate in having been associated with some
recent American developments both in stationary and marine gas-
power plants, a brief survey of a portion of which will enable us to
draw more clearly the comparison between a typical steam and a
possible gas installation.
8 There are in commercial operation in this country today, two
distinct types of stationary power gas-producers which are suited by
their design for operation on almost any class of solid fuel. They may,
by their systems of operation, be qualified as up-draft and down-
draft producers.
9 In the up-draft producer, the fuel is charged into the generator
through an air-tight mechanism at the top, while air and steam, or
air and products of combustion are admitted at the bottom of the
fuel bed, and passing upward, leave the generator at the top in contact
with the fresh fuel. Almost all of the hydrocarbons are unfixed
when leaving the generator with the hot gas, and are condensed later
in the gas coolers or scrubbers and gas mains, forming large amounts of
tar, which, if not removed to a minute degree, will positively prevent the
operation of the engine. The removal of tliis tar is troublesome and is
accomplished at a loss of power and efficiency. The fuel in the upper
zone of the bed in the up-draft producers cokes and cakes so seriously
as to require continuous poking of the fuel bed, either mechanically or
by hand. These features and others in this type of apparatus contrib-
ute to limit the rates of combustion per sq. ft. of grate to a relatively
low quantity. All things considered, therefore, this type of appara-
tus has not lent itself agreeably to modification for marine service.
10 In the down-draft type of apparatus, the fuel is charged by
hand through a large door at the top of the producer, which is nor-
mally in an open position, allowingthe operator unrestricted inspection
of the whole upper zone of the fuel bed. The hydrocarbons con-
188 MARINE PRODUCER GAS POWER
tained in the fuel are driven off in the upper zone, mixed with air and
almost completely burned, and the burnt products, passing downward
through the relatively deep bed of fuel, are decomposed and regener-
ated into carbon monoxid and hydrogen gases. All of the tar and the
lighter hydrocarbons are completely fixed in this process, and no tar
is found in condensation in any portion of the plant after cooling.
Coking or caking of the fuel bed is not detrimental, but on the other
hand assists in keeping the fire in the open porous condition, which is
desirable and necessary where high rates of combustion obtain. This
feature eliminates the poldng necessary in the up-draft apparatus.
The gas leaves the bottom of the producer through brick-lined connec-
tions, and a portion of the sensible heat is extracted in passing through
an economizer. The gas is then cooled and washed and passed
through an exhausting mechanism, whence it is delivered under
pressure to the engine.
1 1 This type of apparatus lends itself admirably to the high rate
of fuel combustion, which for the sake of economy in space and weight
is desirable in marine service. There are in actual commercial opera-
tion today, a number of plants of this type having an average fuel
consumption of over 40 lb. of good bituminous coal per sq. ft. of grate
per hr. These producers are sold on a rating of from 18 lb. to 20 lb. of
fuel per sq. ft. of grate per hr., which is almost 100 per cent greater
than the average rating of the up-draft type of producers.
12 Undoubtedly a better method of measuring the ability or
success of these two systems, is to make note of the number and capa-
city of plants of each type in actual operation on engine service. A
report of the committee on gas engines of the National Electric Light
Association, spring of 1908, showed that in gas-engine power plants,
of capacities of over 300 h.p. each, there were in operation 32 plants of
both types having a total capacity of 57,225 h.p. Of these, 4 plants
were of the up-draft type, having an aggregate capacity of 4050 h.p.,
and 28 plants were of the down-draft type, with an aggregate capacity
of 53,175 h.p. The latter contain the Loomis-Pettibone gas-generat-
ing apparatus, some of which has been in operation on engine service
for 13 years.
13 Three years have been devoted to the modification of these
stationary plants for marine service. The work involved a reduction
in the size and weight of the generators; complete revision of the
scrubbing, gas cleansing and exhausting mechanism; elimination of all
gas holders, storage receptacles, mixing chambers, etc.
14 The plant as modified to date has a light compact producer,
MARINE PRODUCER GAS POWER 189
which while retaining the same rate of combustion as the stationary
apparatus, has materially reduced dimensions and weight of the shells,
brick lining, fittings, etc. The economizer boilers which were used on
stationary work have been abandoned, and replaced with light air-
heating economizers. The gas coolers no longer contain any coke or
brokenmaterial, or wooden trays, and are built of very hght, non-corro-
sive sheet metal, and arranged for eitlier vertical or horizontal posi-
tions, the latter arrangement being convenient for space which would
be otherwise wasted in the vessel. The cooled and partially cleansed
gas is drawn through the above portion of the plant by a centrifugal
gas-cleaning exhauster, driven by direct-connected motor. The gas
passes directly from the exhauster under pressure, through an auto-
matic pressure-regulating valve, to the engine manifold.
15 That the plant is adaptable for marine service with regard to
space occupied and weight, may be seen from the following conserva-
tive estimate:
Plants of from 100 to 500 h.p. each occupy from 0.4 to 0.5 sq. ft.
per h.p., and weigh from 70 lb. to 90 lb. per h.p., including all
auxiliaries, piping, etc. ; plants of from 500 h.p. to 1000 h.p.
occupy from O.SOsq. ft. to 0.45sq. ft. per h.p., and weigh from
401b. to 70 lb. per h.p., including all auxiharies, piping, etc.
16 Undoubtedly the rational opportunity at the present time for
marine gas power lies in commercial service, in which regard the most
rapid advancement in America has been made in the freight, ore and
fuel carriers of the Great Lakes.
17 We have therefore taken for our example a ship built from
the designs of Messrs. Babcock & Penton within the last year.
For the sake of clearness, the views show only the machinery space;
all of the ladders, stairways and grates have been omitted from the
plans, and the piping is shown only on the gas installation. The
machinery installation proper is all there, however, and while the
parts eliminated are merely accessory, the contrast between the two
plants would be all the more striking were they included.
18 The boat is a modern lake freighter and represents the best
standard practice in this service. She is 306 ft. long over all, 45 ft.
beam and 24 ft. deep. Her present power equipment consists of a
single-screw, triple-expansion, three-crank condensing engine, 18-30-
50 by 36-in. stroke, which indicates 1050 h.p. at 90 to 95 r.p.m.
The vessel is fitted with a jet condenser and has independent
190 MARINE PRODUCER GAS POWER
steam-dri'ven reciprocating, bilge, vsanitary and feed pumps. The
complete engine room weight, including piping and alJ auxiharies, is,
in round figures, 182,000 lb.
19 The boiler room equipment consists of two single-ended Scotch
boilers 11 ft. 10 in. mean diameter, 11 ft. length over heads, operat-
ing on a working pressure of 180 lb. per sq. in. Each boiler is fitted
with two 42-in. corrugated furnaces and has tM'O hundred and forty-
four 2f-in. tubes. The grate surface is 36 1 sq. ft. and the heating
surface 1642 sq. ft. in each boiler.
20 The boilers are fitted with forced draft from a 66-in. steam-
driven fan. The air for the draft is taken from the stoke hole and the
fan is located in the engine room. The fan discharge passes through
air heaters in the up-take and thence through ducts to the under side
of the grates. The complete boiler-room weight, including water in
the boilers, but not fuel, is 170,000 lbs. These weights are actual
rather than mere estimates.
21 The coal bunker extends from the main deck to the tank top
and is arranged atliwartship. It has a capacity of 170 tons. The
bunker doors face the stokers on the stoke hole floor. The bunker is
6 ft. fore and aft at the stoke hole. The distance from the forward to
the after bulkhead in the boiler room is 24 ft. 0 in. The distance from
the forward to the after bulkhead in the engine room is 22 ft. 0 in.,
maldng a total over-all length for the plant, including bunkers, of
52 ft. 0 in.
22 The coal consumption on this vessel is from 1.80 lb. to 2 lb.
per i.h.p. lir. This coal is of approximately 13,500 B.t.u. per lb.
23 The problem of substitution of gas for steam, aside from the
design -of the construction of the gas producers or cylinders of the gas
engines, has been thoroughly worked out by Messrs. Babcock & Pen-
ton, of Cleveland. The illustrations show two different arrangements
of gas producers with the same engine. The proposed gas engine is a
four-cylinder, double-acting, reversing type, having cylinders 24
in. bore by 36 in. stroke, delivering 1000 b.li.p. at 100 r.p.m. The
reversing is accomplished by means of compressed air, which is used
to shift the cams from the head to the stern position. Compressed
air is admitted to the cylinders by timed cams in proper cycle. The
crank shaft of the engine is rigidly coupled to the tail shaft of the screw.
24 The illustrations show a column-framed engine. Since making
this layout, the design of the engine has been modified to meet all of
the present marine conditions now found in marine engine design on
the Lakes. In fact, with the exception of the condenser shown on the
MARINE PRODUCER GAS POWER 191
steam drawings, the gas-engine frame will be very similar to that of
the steam engine.
25 For the generation of current to drive the auxiliaries, there
will be installed a double-cylinder, double-acting gas engine, direct-
connected to a 50-lv^v. direct-current generator. All of the pumps
and auxiliaries will be motor-driven. A smaller direct-connected
unit operating on oil will be used for pumping, blowing fires, or
other service, when the gas plant is down. Allowing a distance of
4 ft. 3 in. between the forward bulkhead in the engine room and the
forward side of the flywheel, which distance is one foot greater than
that in the steam installation, we have an over-all distance between
forward and after bulkheads in the engine room of 19 ft. 6 in.
26 As previously stated, two arrangements of the producer room are
shown. The first, the four-generator plant, consists of four 6 ft. by
9 ft. generators, each fitted with independent economizers. The for-
ward pair and the aft pair are connected independently to two
horizontal gas scrubbers which are shown slung under the main deck
beams. The gas passes from these scrubbers to independent motor-
driven centrifugal gas-cleaning fans, whence it is delivered, either
through common connection to a purge or blow-off pipe which also
acts as a by-pass, or through two gas pressure regulator valves to the
air and gas mixing valve at the engine manifold. The 6 ft. generators
require only one cleaning door each. As a I'esult a single cleaning
space suffices for the four macliines, allowing them to be grouped with
reference to athwartship space, so as to give ample room on each side
of the vessel for coal bunkers. The total space occupied by the pro-
ducer plant is 21 ft. 10 in. athwartship, and 15 ft. between forward and
after bulkheads. The producer room weight, including generators,
economizers, piping, and scrubbers, complete, of the four-generator
set, is 110,000 lb. This weight is estimated, but has been carefully
checked and completely covers all the mechanism. In addition to the
above mechanism, there will be a heating boiler which is shown on the
main deck. This boiler will serve to furnish low-pressure steam for
heating the vessel and supplying hot water for washing down decks,
etc. This boiler, with water, will weigh about 8000 lb.
27 The two-generator producer plant, which will undoubtedly be
the one installed, will consist of two 8 ft. diameter by 9 ft. 6 in. generat-
ors, connected to indepenrlent air economizers and each fitted with
an independent horizontal scrubber, located athwartship under the
main deck beams. The gas outlet at the scrubbers will be connected
with a cross-over, so that either exhauster may operate either or both
192 MARINE PRODUCER GAS POWER
producer plants. The exhausters are installed in duplicate and are
connected with common purge or blow-off and common gas outlets
leading either through one pressure-regulator valve, or through a by-
pass direct to the air and gas mixing valves at the engine manifold.
28 On account of the fact that the 8-ft. generators require two
cleaning doors set at 120 deg. the double generator unit plant will
require the full athwartship space in the producer room. The approxi-
mate floor space occupied, therefore, will be ;30 Jft. athwartship and
15 ft. between forward, and^ aft ^bulkheads. The producer-room
weight, including generators, economizers, piping and scrubbers
complete for the two-generator set, is 82,000 lb. jThis weight is
estimated, but has been carefully checked and completely covers all
of the mechanism. As in the case of the four-generator plant, a
low-pressure boiler for heating service will be installed. In the two-
generator plant, however, this boiler will be located on the producer-
operating floor, so that one set of firemen may suffice for both.
29 The only guide we have for estimating the probable fuel con-
sumption for this service is found in the large number of stationary
producer gas power plants now in operation. Fortunately, in marine
service, the load factor will be uniformly much higher than that found
in any stationary service to which gas power is applied at the present
time. The builders of this apparatus are prepared to guarantee one
brake horse power per hr. on one lb. of good bituminous coal, averag-
ing 13,500 B.t.u. per lb.
30 Messrs. Babcock & Penton, the engineers who designed and
built the steam plant, and who have spent years on the problem of the
substitution of gas for steam, have suggested that the coal bunker,
which will be placed above the charging deck of the producer, should
have a capacity of about 80 tons of coal. These bunkers will run from
the charging deck to the deck-house and will have doors opening
closely adjacent to the charging doors of the generators, so that little
or no coal passing on the operating deck will be required.
31 In making the comparison shown in the table, it is unnecessary
to go into the cost of fuel, labor, hours of service, etc., as these ele-
ments vary with every class of service. In this particular proposition,
it will suffice to state that the engineers who have been working on
this substitution problem have conservatively figured that with the
saving in fuel and the increased cargo carried, the cost of the com-
plete plant will be saved in two years of operation.
32 While the gas plant here described has neither been constructed
nor ordered at this writing, its forthcoming will not be long delayed,
MARINE PRODUCER GAS POWER
193
TABLE I COMPARISON OF POWER PLANTS FOR GREAT LAKES
FREIGHT-CARRIER
Length over all 306 ft. 0 in. Displacement. . 6000 tons gross, 6600
Beam 45 ft. 0 in. tons net.
^«P*^ 24ft. 0. in. (.^^g^ 4200 tonsnet, 18 ft. draft
Speed, 12 statute miles per hr. on 900
i.h.p.
Steam Gas
engine room engine room
3-cylinder triple-expansion, condens- 4-cylinder, 4-cycle, double-acting, gas
ing, 18-30-50 by 36 in., 1050 i.h.p. at engine, 24 in. diam., by 36 in. stroke
90 to 95 r.p.m. 1000 b.h.p. at 100 r.p.m.
AuxiUaries steam-driven Auxiliaries motor-driven
Length between bulkheads, 22 ft. 0 in. Length between bulkheads, 19 ft. 6 in.
Engine room weights, including auxili- Engine room weights, 105,000 lb.
aries and piping, 182,000 lb.
BOILER ROOM
2 single-ended Scotch boilers fitted
with economizers, forced draught.
Length each boiler, overheads 11 ft.
Oin.
Mean diameter, each, 11 ft. 10 in.
Two 42-in. furnaces each
244 2|-in. tubes, each
Grate surface, each, 36.75 sq. ft.
Heating surface, each, 1642 sq. ft.
Boiler room weight, water in boilers,
no fuel, 170,000 lb.
Length boiler room 24 ft. 0 in.
Length boiler room, includes bunkers,
30 ft. 0 in.
Square feet boiler room, including
bunkers, 900
Square feet per h.p., 0.9
Bunker capacity, 340,000 lb.
Total weight, machinery and fuel, 692,-
000 lb.
Total length of machinery space includ-
ing bunkers, 52 ft. 0 in.
PRODUCER ROOM
Two down-draft gas producers and
auxiliaries
Diameter shell, each generator, 8 ft.O in.
Inside diameter, lining generator, 6 ft.
Sin.
Height shell, each generator, 9 ft. 6 in.
Grate surface, each generator, 30.67
sq. ft.
Producer room weights, no water, no
fuel, 82,000 lb.
Length producer room, includes bunk-
ers, 15 ft. 0 in.
Square feet producer room, 450
Square feet per h.p., 0.45
Square feet producer room with four
smaller generators, 330
Square feet per h.p., four generators,
0.33
Bunker capacity, 160,000 lb.
Total weight, machinery and fuel,
347,000 lb.
Total length of machinery space, 34 ft.
6 in.
Saving in weight, 355,000 lb.
Saving in fore-and-aft length, 17 ft. 6 in.
Saving in cubic space 17 ft. 6 in. by 32
ft. beam by 20 ft. high, 11,200 cu.ft.
194
MARINE PRODUCER GAS POWER
MARINE PRODUCER GAS POWER
195
^ai}=^.
'fj CO
S O
X o
^5 2
CO
- -A
O ^
t^
196
MARINE PRODUCER GAS POWER
MARINE PRODUCER GAS POWER
197
198
MARINE PRODUCER GAS POWER
MARINE PRODUCER GAS POWER
199
200 DISCUSSION
and this comparison, while somewhat premature, is made to present
the possibilities of marine producer gas power to those interested in
its future.
33 A marine bituminous gas plant, similar in construction and
operation to the one described, but of 300-h.p. capacity, has been in
commercial operation driving a six-cylinder, single-acting, reversing
marine gas engine for over a year. The results obtained give ample
security for the statements made in this paper, and point to the early
adoption of this type of prime mover for our marine commercial ser-
vice.
DISCUSSION
C. L, Straub. We have received reports from abroad of progress
in the marine gas-producer field, a summary of which I hope will
prove of interest.
2 In Holland teaming is practically unknown and local freight
trains are never run, the canals being used for moving freight from
one city to another or between different parts of the same city.
The canal barges range from 40 ft. in length, with an 8-ft. beam,
drawing 3 ft. of water, to 150 ft. in length, with a 20-ft. beam, drawing
6 ft. of water. The majority of the barges are hand-propelled, about
9 per cent have steam equipment, while about 6 per cent are pro-
pelled by gas engines. Of the last-named, a few — about | of 1 per
cent — use gas producers, the others using liquid fuel.
3 The gas-engine barges range in size from 40 ft. to 130 ft., the
engines from 10 h.p to 300 h.p., and the engine speed from 250 r.p.m.
for the larger to 400 r.p.m. for the smaller engines. The reason for
the small number of power boats is the great scarcity of fuel. Holland
is without coal mines or any natural growth of timber. Hence coal
is expensive and difficult to obtain, though wood is more plentiful as
large quantities of lumber are shipped in from Germany. Peat is
used a great deal, while compressed peat and some domestic and
imported briquettes are burned to some extent.
4 The gas-engine boats using liquid fuel are more handicapped by
fuel conditions than the steamboats are. The Standard Oil Co.
supplies fuel in a few large cities such as Amsterdam and boats of
large power can work around these fuel depots only. This limit to the
Note.— =-At the time of the presentation of his paper Mr. Straub also pre-
sented a report on gas producer development abroad, with special reference
to marine work. An abstract of this report is here given as a part of the
discussion on his paper. — Editor.
MARINE PRODUCER GAS POWER 201
radius of action prevents the extensive equipment of boats with gas
engines operating on liquid fuel.
5 The producers installed on the comparatively small number of
boats include principally modifications of standard German stationary
types. They are not successful and apparently cannot be made to
give continued good results with the fuels available. Producers that
will operate economically on peat, wood, briquettes or poor coal, and
require little attention, are in great demand. The engines are giving
satisfaction. Boat owners want gas-engine equipments but not for
use with liquid fuel. There is a great demand for gas-producer equip-
ment of capacities up to 300 h.p. for inland waterway traffic, and for
capacities up to 600 h.p. for general inland and foreign traffic.
Holland thus offers an attractive field, because success with these
small capacities will open the way for larger equipments.
6 A test run of two tug-boats was made from Hamburg to Kiel
and return, one boat having a steam equipment, and the other a gas-
producer equipment. The results are shown in Table 1. The
weather was rough and the speed maintained 8^ knots an hour.
The coal consumption for a period of 10 hours was as follows: Gastug,
530 lb. anthracite coal; Elfrieda, 1820 lb. bituminous coal; an econ-
omy in coal consumption of nearly 3.5 to 1 in favor of the gas-pro-
ducer equipment.
TABLE 1 COMPARISON OF STEAM AND GAS-PRODUCER EQUIPMENTS
Gas
Steam
Name
Length
Beam
Horsepower
Towing meter pull
Gastug
44 ft. 3 in.
10 ft. 6 in.
4 cyl., 70 b.h.p.
2140 lb.
Elfrieda
47 ft.
12 ft.
Triple-Exp., 75 b.h.p.
2020 lb.
7 Herr Korting has established a marine department at Kiel, and
has practically a monopoly of the government submarine work. He
is experimenting and preparing to take up larger work. The Ger-
man government has decided to adopt gas-engine boats for all war-
ship jaunches, pinnaces, videttes, portable torpedo boats, and the like.
8 The Niirnberg Company (Vereinigte Maschinenfabrik Augs-
burg und Maschinenburg Gesellschaft Niirnberg A. G.) have installed a
number of small Diesel oil engines for marine use. Both suction and
pressure producers are built by them. As to fuel, the company states
that anthracite, coke, charcoal and lignite briquettes are the most
202 DISCUSSION
suitable for producers. Efforts to gasify raw pit coal have not been
entirely successful, at least in medium and small-sized plants.
9 In Great Britain, Vicker Sons & Maxim, Ltd., have built 40,000
h.p. of marine high-speed gas engines. Plans have been drawn for an
oil-engine torpedo-boat destroyer of about 30 knots speed. With the
same dimensions and speed, the oil-engine destroyer saves enough
weight and space so that the armament may be increased from the
one 12-lb. and five 6-lb. rapid-fire guns of the steam destroyer to four
33-lb. and two 6-lb. rapid-fire guns. Moreover the quantity of
ammunition (number of rounds per gun) is the same for l30th boats,
although the guns and ammunition per round of the oil-engine
destroyer are much heavier than those of the steam-engine destroyer.
Sufficient fuel capacity is provided to allow a speed of 30 knots an
hour with a radius of action 6^ times greater than that of the steam-
driven destroyer. Also more space below decks allows superior
accommodations for the crew.
10 A marine gas-producer equipment built by the Beardmore
Company under the Capitaine patents was tried out with satisfactory
results. The capacity of the plant was 70 h.p., the engine having
four cylinders 8 J in. in diameter and of 11-in. stroke. The weight
of the whole was shghtly over 13,440 lb. The equipment was installed
in a launch 60 ft. long and of 10-ft. beam. A 10-mile run was made
in one hour without recharging the producer.
George Dinkel. I would like to ask the author if there are any
gas-producers working with the small grades of the steam sizes of
anthracite coal, such as No. 3, 2 and 1 buckwheat, especially as he
states in Par. 8 that there are in commercial operation in this country
today two distinct types of stationary power gas-producers suited by
their design for operation on almost any class of solid fuel. Where
are those producers being used on the same grade of steam sizes, and
what has been the result?
Henry Penton.^ As the members probably know, freight is
carried on the Great Lakes at a lower cost than anywhere else in the
world, and over 75 per cent of the merchant steam tonnage of the
United States is built on the Great Lakes. We are, of course, con-
stantly seeking methods of reducing carrying costs, and so far as the
ship is concerned the most important item of expense is that of power.
We have for some time been firm believers in gas power: power for
1 Henry Penton, Babcock & Penton, Cleveland, O.
MARINE PRODUCER GAS POWER 203
propulsion, however, is only one of the problems to be met; the prob-
lem of the application of gas power to the auxiliary service has given
us more concern. At present our ships are handled entirely by steam,
both in port and out, and we must have power available at all times.
2 We believed we could depend upon the combustion engineer to
perfect the producer- gas engine when the opportunity arrived; but
it has arrived and while we believe the producer to be satisfactory
we are not yet satisfied with the engine.
3 We first considered the installation of gas power in one of our
modern ships where the average horsepower (indicated) is in the
neighborhood of 2000, but subsequently decided to select a smaller
type, believing that we should practice creeping before walking.
While the ship to which the author has alluded represents the best
standard practice in design, she is not a representative lake steamer
in that she is relatively small. Her engines develop approximately
1000 i.h.p. Her carrying capacity is about 4080 tons on 18 ft.,
while our modern ships run to 12,000 and 13,000 tons on the same
draft. The fuel is necessarily bituminous coal, fairly uniform in heat
value and averaging about 13,500 B.t.u. Anthracite and coke are out
of the question both as to delivery and cost-
4 It should be noted that the fuel consumption given by the
author includes fuel used for all purposes aboard ship charged against
the actual indicated horsepower. This is the customary method of
stating the consumption and is used merely for purposes of compari-
son. The propelling engines actually do their work on an average of
1.5 lb. to 1.65 lb. per i.h.p. per hr., including the auxiliaries.
5 In Par, 25 mention is made of the installation of a 50-kw. direct-
current generator. I think, however, it will be necessary to use at
least two of this size, depending somewhat on the method adopted
for handling the windlass, which calls for the largest individual motor
of any of the apparatus. There must be no such thing as a generator
shutdown. As just stated, every function of the ship, including
propulsion, is now performed by steam, and power must be available
every instant from the time the ship goes into commission in the spring
until she is laid up the following winter. If we take out steam we
must provide something just as available in its place. While a great
part of the time the output will be small, we must be able to take care
of the maximum requirements.
6 It may be wondered why, as stated in Par. 26, the auxiliary
boiler is required for heat and for washing down decks. These ships
run until well into the winter and when the weather becomes severe
204 DISCUSSION
they ice up badly, and it is not uncommon to make port with 200 or
300 tons of ice aboard. The quickest method of clearing away is
with the hot-water hose. I hope we shall be able to use the exhaust
gases for generating steam at sea and thus operate the boiler without
the use of coal; but I do not know whether this is yet feasible.
7 With reference to figuring the elimination of cost in two years,
the operation is brought about in this way: some of our ships get in
more, some not so many, but the average is not far from 20 round
trips per year; and taking into consideration the reduction in weights,
which means additional revenue-producing cargo capacity; reduced
space, which in some trades is also additional cargo capacity; and
reduction in fuel, which is both reduced expense and additional cargo,
I have succeeded in convincing myself that we can get even in about
1^ seasons; but two seasons is perfectly satisfactory, and you can
readily see that the addition of one or more trips in the year wUl go a
long way toward the extinction of that cost.
Irving E. Moultrop. Examination of Table 1 gives some very
interesting information. It is rather surprising that the total weight
of the complete gas-power plant is so much less than that of the steam
plant. Of course, the steam plant has a number of auxiliaries which
the gas plant does not require, and these auxiliaries are quite heavy,
yet up to the present time, in stationary practice, the total weight of
a gas engine has been verj'- much in excess of that of a steam engine for
the same power. The two prime movers discussed in this paper oper-
ate at the same speed; one would naturally assume, therefore,
that the extra weight in the gas engine would go far toward making
up for the saving in weight due to the omitting of a number of steam
engine auxiliaries. As the total weight of the gas engine room machin-
ery is only about 60 per cent of that of the steam engine room machin-
ery, one naturally wonders if the factor of safety in the gas engine
design has not been reduced to save weight, or if this is not the case, if
some special weight-saving features have not been introduced in the
gas engine design, which might have been used to equal advantage in
the steam engine.
2 Comparing the producer room with the boiler room it is inter-
esting to note that the total grate area of the producers is only about
five-sixths of that of the steam boilers. Stationary practice has
shown that the best producer results are obtained at a very much
lower rate of combustion per square foot of grate than in good steam
boiler practice. It would be interesting to know how the engineer of
MARINE PRODUCER GAS POWER 206
the gas-power plant expects to obtain capacity out of its producers
when the full consumption per square foot of grate area in the pro-
ducer will exceed what is considered good practice on the grate of
steam boilers.
3 In comparing the total weight, machinery and fuel, in the gas-
power plant with the steam plant, and also the total space occupied,
it should be noted that in the gas plant the bunker capacity is less
than half that in the steam plant.
Herbert M. Wilson.^ Perhaps it would not be a breach of con-
fidence for Mr. Straub to tell us something concerning the gas producer
for the new non-magnetic vessel of the Carnegie Institute. This
vessel is being constructed with as little iron as possible, for use in
magnetic surveys; and I understand the gas producer and gas engine
were selected for auxiliary power because of the small weight of metal
required and the possibility of substituting bronze and other non-mag-
netic metal for iron and steel. Their vessel is about to be launched,
and perhaps Mr. Straub could tell something of the gas producer engine
plant which is under construction for actual operation.
E. T. Adams. Great changes in the weight required have come
about in the past few years. The first designs in any line of manufac-
ture uniformly carry unnecessary weight in the various parts and it is
sale to say that the gas engine of today, built with the same factor of
safety, would be 25 per cent lighter than any engine of the same horse-
power built three years ago.
2 This applies to engines in use for ordinary commercial purposes,
as electric lighting or power. In view of this fact the statements
of the author on this point are not at all surprising. The increasing
use of steel and the modification of design based on experience have
led to great reduction in weight of engines for special purposes, such
as are here specified.
The Author. With regard to the question of the speed of vessels
depending on the character of power equipment, taking a given hull
and a given power equipment, the craft will go at a certain speed, the
character of the equipment notwithstanding. The comparative
space required by steam and by gas equipments can be best shown
by an example given by Capt. A. B. Willits, U. S. N., in an article.
Gas vs. Steam for Marine Motive Power, printed in the United
States Naval Institute Proceedings, December 1908. Mr. Willits
* With U. S. Geological Survey, Washington, D. C.
206 DISCUSSION
states that the floor space occupied by the boilers in the New Hamp-
shire equals 0.33 sq. ft. per h.p., and that the weight per b.h.p. is
110 lb. The power of these boilers is rated at their forced capa-
city. If we install a producer plant and rate it at its forced capa-
city at 40 lb. of fuel per square foot of grate, it will occupy 1/10
square foot per b.h.p., for such a plant as the New Hampshire would
require, and weigh 30 lb. per b.h.p.
2 From these figures it is apparent that a marine producer-gas
plant can be installed in at least the same space and of certainly
not greater weight than a modern marine steam boiler plant.
These figures refute Capt. Willits' figures of the Westinghouse gas
plant, which occupied 1 sq. ft. per b.h.p., and weighed 28.5 lb. per h.p.,
and the R. D. Wood producer which occupied 1.84 sq. ft. per b.h p.
and weighed 194 lb.
3 Mr. Dinkel asked regarding the small anthracite coals in gas
producers. I can refer him to the generator of the R. D. Wood plant
at Jersey City, which has been operating at the plant of the Erie
Railroad for five or six years on a mixture of No. 1 and No. 2 buck-
wheat coal. The Lehigh Coal and Navigation Companv has in-
stalled a gas producer which has operated on rice coal and has
been running for almost two years. We have two plants in opera-
tion, one at Hartford, Conn., and one near Philadelphia, running on
a fine grade of anthracite coal, mixtures of Nos. 1, 2 and 3 buckwheat.
4 Regarding the factor of safety in gas-engine parts, which Mr.
Moultrop brought up, the six-cylinder single-acting engine referred
to in Par. 33 weighs less than 30 lb. per b.h.p. Of course, that was
a comparatively high-speed engine and delivered 300 b.h.p. at 320
r.p.m. The 1000-h.p. engine will be fitted with cast-steel parts,
in almost every instance where cast iron was used on the steam
plant, and this makes for a big reduction in weight at a very slightly
increased cost.
5 The producer and equipment for the Carnegie, about which Mr.
Wilson asked, is almost finished and will be in the boat when she is
launched on Maj^ 10. The producer shell is about 6 ft. in diameter
and is of copper. The pipe and scrubber are of composition metal,
containing no iron or steel. The only steel or iron parts on the pro-
ducer are manganese steel grates, doors and door frames, near the
hot portion of the fire in the producer equipment, and this manganese
steel is less than one per cent magnetic, when compared to mild steel,
so that of it we have been allowed the use of 1500 lb. The engine
will have bronze cylinders and will not be lined with cast iron, as the
MARINE PRODUCER GAS POWER 207
published reports indicate. We are going to run cast-iron pistons in
the bronze cylinders. The cylinders are comparatively so thin and
so close to the water jackets, that we anticipate no trouble from dete-
rioration. The only steel or iron parts about the engine will be the
cams, rollers and valves. The valves will be of cast iron. The steel
cams and rollers will be hardened and ground. On a commercial
basis, using terms equivalent to mild steel, as far as magnetic force
is concerned, we will have less than 200 lb. total of iron or steel in
that vessel. The published reports make further detail unnecessary.
The boat will be ready to sail July 1.
No. 1239
OPERATION OF A SMALL PRODUCER GAS-
POWER PLANT
By C. W. Obert, New York
Associate Member of the Society
It has been the practice of the packing house of Swift & Company
of Chicago, in the distribution of meats and provisions to retailers, to
establish in different cities distributing depots with the necessary power
equipment for the handling and refrigeration of the products. Some
of these branches in the larger cities are establishments of consider-
able size, and with the extensive cold storage facilities required for
the large stocks carried, require comparatively large power installa-
tions. The new Westchester market, which the company has recently
built in New York at 152d Street and Brook Avenue in the Bronx, is
a notable installation of this kind, involving a 400-h.p. producer gas-
power plant for the operation of both refrigerating and electric
generating machinery, which supplies similar service to a number of
adjoining depots of other houses.
2 The refrigerating duty at present required embraces the opera-
tion of a total cooling system containing over 46,000 ft. of 2-in. pipe,
which reaches a maximum of over 100 tons of refrigeration per 24
hours under the most severe summer weather conditions. Two 65-ton
refrigerating machines were installed for this service, with equip-
ments in duplicate, owing to the great importance of continuity of
refrigeration, particularly in hot weather. A maximum of nearly 90
h.p. is required for compression machines of this size and engines of
100 h.p. were selected for driving them, to provide sufficient capacity
for unfavorable or overload conditions.
3 The electrical load, which includes the operation of several
electric elevators, fluctuates ordinarily between 30 kw. and 50 kw.
but occasionally reaches a maximum of over 60 kw. For this service,
duplicate 75-kw. generators were installed, with driving engines of 100
Presented at the Spring Meeting, Washington, May 1909, of The Amer-
ican Society of Mechanical Enqineeys.
210 OPERATION OF PRODUCER GAS-POWER PLANT
h.p. This was done to secure uniformity of size and detail in all four
of the driving-engines.
4 For gas making, two producer equipments were installed, also in
duplicate. One of these is a 200-h.p. producer, intended for the
supply of one refrigerating machine and one generator engine when
operating at maximum capacity. The other is of 150-h.p. capacity
to permit of closer adjustment of the producer capacity to the load at
other times.
5 The plant arrangement consists of an engine room in the easterly
end of the sub-basement of the market building, and a producer room
adjoining, the entire power equipment occupying a total space, includ-
ing fuel storage, of 48 ft. by about 55 ft. Headroom for the machin-
nery and piping is afforded by the depression of the sub-basement
floor to a level 16 ft. below the street, and the omission of the base-
ment floor in this section, giving thus a clear headroom of 18 ft. The
machinery space was originally laid out as a single room, but as a
result of the requirements of the underwriters, the producer space has
been separated from the rest by a 6-in. hollow- tile fire wall, forming a
producer room 20^ ft. by 24 ft. maximum dimensions. Under the
152d Street sidewalk, there is an 11 ft. by 29 ft. room containing
pumps and auxiliaries for the power equipment and the building
heater; and adjoining this, an 11 ft. by 30 ft. space for fuel storage.
The latter has capacity for over 150 tons of coal, which is dumped into
it through sidewalk coalholes from wagons in the street.
6 The engines are Rathbun vertical, three-cylinder units, of 100
h.p., rated at 280 r.p.m., built by the Rathbun-Jones Engineering
Company, Toledo, Ohio. The two for the electrical service are direct-
connected to 75-kw. generators and the other two through silent chain
drives to the ammonia compressors of the refrigerating equipment.
They are all of the four-stroke cycle, single-acting, enclosed type,
and have 12f in. by 13-in. cylinders, designed for the above rating
when operating on producer gas of not less than 125 B.t.u. per cu. ft.
These engines are throttle-governed, a special form of centrifugal
flyball governor being used, and have each a one-ton fly-wheel at
both ends of the crankshaft.
7 The gas is generated for the engines in a duplicate equipment of
Smith suction producers built by the Smith Gas Power Company,
Lexington, Ohio. Each equipment consists of a simple shell pro-
ducer, a wet scrubber and a dry purifier. While the producers differ
in rated capacity to permit of more accurate adjustment of their
capacity to the power requirements at different seasons of the year,
OPERATION OF PRODUCER GAS-POWER PLANT
211
m
z
o
o
a
o
o
t-H
o
a
03
212
OPERATION OF PRODUCER GAS-POWER PLANT
the scrubbers and purifiers have a maximum capacity of 200 h.p.,
which permits the smaller producer to operate up to the maximum
plant capacity of 200 h.p., if required to do so temporarily. The small
and large producers have 6-ft. and 7-ft. shells respectively, both 12 ft.
in height, their internal diameters being 4^ ft. and 5J ft. respectively,
and they are fitted with shaking grates on the up-draft principle for
operation with anthracite coal. They are not fitted with attached
vaporizers or air pre-heaters, but have an automatic control attach-
ment for regulation of the amount of water vapor to conform to the
power requirement and consequent rate of gasification. The scrub-
bers for cleansing the gas are vertical cylindrical tanks, each 4 ft. in
diameter by 15 ft. high, and the dry purifiers have 4-ft. shells 6 ft. in
height.
irat Floor | [ I
?i^'^f:&^k^.-r^l-ii.^:^' "'■■ '"--" :\. h-^^^^^"
23
Fig. 2 Elevation of Machinery Room in Cross-Section
8 The piping of the plant was somewhat involved by the arrange-
ment of the engines relative to the producers, and, in the Smith
producer system, by automatic vaporizers in the exhaust connections
to utilize the waste heat of the engines for the vaporization of the
water. The vaporizers are located close to the engines and attached
to each vaporizer is an automatic device, through which air is admitted
and superheated for the producer. The air is conducted to the pro-
ducers from these devices by a 10-in. pipe main, extending through the
engine room, and heavily covered with magnesia insulation.
9 The gas is delivered from the producers by 8-in. pipes connecting
from the top of the producer to the bottom of the scrubber shell and
each scrubber has a triplicate connection to its corresponding purifier,
which is a three-part filter. From these the gas is conducted to the
engines through a 5-in. line, with a S^-in, branch to each. The exhaust
connections from the engines to the vaporizers are 5-in. lines and from
the latter, individual discharge pipes are carried up for each engine
through a pipe shaft in the corner of the building to a roof outlet.
OPERATION OF PRODUCER GAS-POWER PLANT 213
It is to be noted that this arrangement of exhaust connections is
effective in so muffling the noise of the escaping gases that they can-
not be heard from the adjoining street and are only barely noticeable
when on the roof close to the outlets.
10 The electrical generators are 75-kw. General Electric direct-
current machines, each rigidly coupled to the driving engine. They
are wound to deliver current at 220 volts, the distribution for both
lighting and power being on the two-wire system. The electrical
circuits are controlled on a three-panel switchboard which contains
the usual equipment of indicating and recording instruments, field-
rheostat switches and generator and feeder switches. The building
is wired separately for lighting and power circuits, and recording watt
meters are connected into the feeder circuits for measurement of the
power delivered. It is to be noted that separate bus bars are provided
for both power and lighting feeders, as well as a switching arrangement
by which the lighting service may be supplied from a generator other
than that carrying the power load, incase the fluctuations of the latter
should interfere with the voltage regulation. This provision has been
found unnecessary, however, as the speed regulation of the engines and
generators is satisfactory under all fluctuations of loading due to ele-
vator operation.
11 The refrigerating equipment was installed on the direct
ammonia expansion system, a feature of which is the connection of all
coils in the coolers in series with those in the freezers, whereby all
ammonia not thoroughly evaporated in the freezer coils will be in the
cooler coils (temperature, 36 deg. fahr.), which permits carrying
the freezer temperature at from 0 deg. to -H 5 deg. without frosting
the compressor. The compressors were built by the Hutteman &
Cramer Company, Detroit, Mich., and are horizontal single-cylinder
double-acting machines, with 14-in. by 30-in. cylinders, each driven at
a speed of 60 r.p.m. by a Renold silent-chain connection from its
driving engine, with a speed reduction of about five to one.
12 The ammonia condenser is located on the roof of the building
and provided with the usual water-cooling sprays. The water supply
for it is obtained from a well extending into water-bearing soil under
the basement floor, and the drainage from the sprays is subsequently
utilized in the scrubbers and in the engine cylinder jackets. One of
the compressor units normally handles the load alone, which leaves
one equipment always in reserve, to provide against the serious emer-
gency of a complete stoppage of the refrigerating service during hot
weather.
214 OPERATION OF PRODUCER GAS-POWER PLANT
13 In operation this plant has proved particularly economical,
largely due to the continuous character of the service resulting from
the operation of the refrigeration plant 24 hours a day, seven da3^s a
week, thereby eliminating standby losses. The average load range
of the plant is ordinarily from 50 per cent (100 h.p.) to full rated load
(200 h.p.), the high and low load factors occurring during the summer
and winter months respectively, when the refrigeration requirements
are maximum and minimum. With the heavier load factor during
the summer months, the fuel consumption has ranged between 3400
and 4800 lb. per 24 hours, the larger figure having been exceeded on
only two days in 11 months, and the consumption per horsepower-
hour as calculated from station fuel records and observed loads,
ranged from 1.4 to 2.0 lb. of coal. The fuel rate has dropped during
periods of continuous high loads, to about 1 lb. per horsepower-hour,
as based on observed loadings, but the daily average under conditions
of ordinary commercial operation is usually greater.
14 The operating conditions during the heavy-load season are indi-
cated in the table at the end of the paper, in which the relation of
fuel consumption to load carried is shown for two weeks of similar
duty. The variations in the amount of fuel charged from day to
day are due chiefly to the differing conditions of the fuel bed in the
producer, the removal of a particularly large amount of ashes on any
day necessitating a heavy fuel charge. No account is taken of cost of
water used in the scrubbers and cooling jackets, as the supply is
obtained from a well on the premises without cost other than that of
pumping.
15 The fuel used is No. 1 buckwheat anthracite that has been
passed over a f-in. mesh and through a yVi^^- mesh screen, with 5 per
cent fineness, and costs $3.50 pergross ton delivered in cargo lots. It is
charged only at the regular cleaning periods, at each of which from
400 to 900 lb. of coal are fed, after the fire has been cleaned down and
the ashes removed from the grate. The fire is cleaned periodically
twice every shift,or four times per 24 hr., and requires about an hour
per cleaning on the average.
16 In this connection it is interesting to note the comparatively
short time required to start a producer into service from the cold,
which has been done repeatedly on short notice in about five hours;
on December 12 when the 150-h.p. producer was placed in operation
to relieve the larger unit, the kindling wood was lighted at 10 a.m. and
the gas supply turned onto the engine at 2 p.m., with only about 12-in.
of fire zone in the fuel bed. The reliability of a suction producer
OPERATION OF PRODUCER GAS-POWER PLANT 215
operating under a continuous and exacting service of this character
is well shown by the duty of the 200-h.p. producer during the summer
season of 1908, which when taken out of service on December 12, had
been continuously in service 24 hours per day and seven days per week
since April 22, a continuous run of 235 days. During that time, it had
received no more attention than the four cleanings and chargings per
24 hours.
17 The operating force for the power plant consists of an engineer
and an assistant engineer and two producer tenders, who work in two
shifts. This force is able to maintain the plant equipment in satis-
factory operating condition, as well as the refrigerating and electri-
cal equipment of the depot, and it is worthy of note that the plant has
not been shut down for any reason since it was started on February 1,
1908, a period of 15 months. Experience during this period indicates
that, contrary to the general opinion, no more attention is required
than for a first-class steam plant, the necessary attendance comparing
very favorably with that of a high-grade steam plant of the same
capacity. CleanHness of all parts of both producer and engine equip-
ments, and careful adjustments, especially of the latter, are imperative
and are the keynotes of successful operation. In order to maintain the
equipment in such condition, a thorough and comprehensive operating
system has been developed which may be of interest.
18 The operating system involves a detailed and thorough
inspection routine that keeps the force well informed as to the condi-
tion of the entire equipment, and a division of duties tending to favor
the maintenance work. To the day operating force is assigned the
inspection and adjustments of the engines and repairs to igniters,
batteries, etc., while the night force has the work of cleaning all
machinery.
19 The regular routine of the day force is in detail as follows:
First upon coming on duty at 7 a.m., an examination is made of all
moving parts of the two engines in operation, and also of oil levels in
lubricators and conditions of water jackets and ignition systems.
There are always two engines in operation, one being a generator
engine and the other a refrigerating engine, which in the periods of
heavier loadings in summer time have a combined load of about 140
h.p. of which fully 75 h.p. is taken by the refrigerating system. Next
the water regulation for the steam supply is noted and then the con-
dition of the suction draft on the producer and also on the scrubber
and purifier, there being three U-shaped draft gages provided for this
purpose, one connected to the gas suction Hne to the engines, the
216 OPERATION OF PRODUCER GAS-POWER PLANT
second to the gas connection from the scrubber to the purifier and the
third to the connection between the producer and scrubber. A
uniformity of suction of from 2 in. to 3 in. of water in these three
gages indicates a proper condition of the three units, while any unusual
suction in any of the connections would indicate an obstruction
needing immediate attention. The latter is always clearly indicated,
as an obstructed condition in the producer, for instance, will raise
the suction to as high as 9 in. or 10 in. of water.
20 Next an inspection is made of the producer, the temperatures
of different portions of the fire being determined to ascertain the con-
dition of the fuel bed, the existence of cracks or fissures or pockets of
unburned coal. To do this, a j\ -in. iron rod is pushed into the fire
through the side peep holes in the producer shell, held there exactly
one minute and then withdrawn, the temperature within being noted
from the color of the rod. If the latter is at a uniform cherry red
temperature throughout its length, this is taken as an indication of an
even fire; but if at a brighter heat or dull in some portions of the rod,
there is evidence of unnecessarily high local temperatures due to rapid
combustion in fissures in the fuel bed, or of a stagnant condition in
dirty or unburnt portions of the fire. The rod is first inserted in the
lowest hole and then successively into the upper holes, in order to
explore the fire in zones. On withdrawing the rod the operator
notes graphically the condition of the fire by marking a line with
chalk on the shell of the producer even with the hole, a straight line
indicating an even temperature, and a broken line sho\ving the dirty
condition, etc. This operation is continued for the four holes and a
fuel curve drawn from it which gives a practical idea of how the dirt
lies in the producer and shows what quality of gas can be expected.
Finding the producer in good order, the scrubber, purifier and connec-
tions are examined for unusual temperatures, condition of water flow,
etc.
21 In the maintenance work, each engine is shut down after every
seven days work of 160 hours for general inspection and cleaning,
and thus on Monday mornings it is necessary to start up the two
reserve units and transfer the respective loads to them. Before
starting up either reserve engine, its igniters are cleaned, which takes
about one hour. With the igniters clear and everything in good
order, the attendant looks at the draft gage, which is equal in impor-
tance to the gage of a steam boiler, to see what gas the engines in
operation are drawing and whether the start can be made without
interfering with their suction. If there are any doubts the gas is
OPERATION OF PRODUCER GAS-POWER PLANT 217
enriched temporarily by putting about four pails of water in the ash
pit of the producer and slicing the fire to work down some hot coals,
which, by turning the water into vapor, increase the hydrogen content
of the gas and enable the third engine to be started without inter-
fering with the others. After getting the engine warmed up, the load
is thrown on and the other engine is shut down. The extra pull on the
producer, due to overload from running the three engines and the
hydrogen added, has usually so enriched the gas that on cutting out a
unit the quality of gas is too rich for the two units operating alone.
To counteract this, it is necessary to give additional air to each of the
units that remain and then, as in starting, there will be no varia-
tion in speed of operation.
22 After the engines have been shut down their inspection is begun
by the removal of the back crank case covers and examination of the
bearings, crank pins, wrist pins, etc., for necessary adjustments.
Besides this the exhaust valves are cleaned and the ignition system
checked. This requires about two days, as but one thing is done at a
time and then only at times when the load on the plant is not heavy.
While the engineer is performing this work, the producer tender pre-
pares to clean and coal the producer, as follows:
23 The method of cleaning is to rake off the ash from the grate
table and then poke down around the shell from the top poke holes.
Having before him the fuel chart which was noted graphically on the
producer shell on coming on watch, the attendant knows what part
of the bed requires most poking. Before opening the ash pit doors
about the shell, water is placed in the ash pit as before and the hot
ashes, dropping down, form sufficient steam to mix with the air
coming through the ash pit door and offset any bad effect therefrom.
This enables the cleaning to be done without affecting the engines.
Having cleaned and poked the fire thoroughly and worked down all
the ash so as to leave it as clean as possible, the coaling is then begun,
count being taken of each hopper of coal charged. The coal is cleaned
by screening if very fine or dirty. Having coaled, the operator slices
across the grate so as to relieve the center of the fire and again puts
water in the ash pit, this time to cool off the grate after cleaning and to
offset the effect of any air that may have gotten in during the opera-
tion. The cleaning usually occupies one hour, the amount of coal
put in ranging up to 900 lb. After giving the producer time to settle
down the ashes are withdrawn from the ash pit, an average of 1^ ash
cans (about 3 bushels) being removed after each cleaning. During the
cleaning operation the operator is always on the lookout for any
218
OPERATION OF PRODUCER GAS-POWER PLANT
change in the engine speed due to weak gas on account of opening the
ash doors. Should this occur he immediately cuts the air supply to
the engine, resulting in a combustible mixture without noticeably
reducing the speed. The producer is now good for 6 hours' operation,
after which the cleaning is repeated.
24 The refrigerating engines are operated for periods of 84 hours
and then gone over. One exhaust valve is taken out of an engine
each week, thoroughly cleaned, and regTound if necessary, thus
insuring attention to each valve once in every three months. Igniters
are cleaned weekly and the batteries and ignition system checked.
The temperature of the fuel bed of the producer is taken twice a day
and a gas analysis is made once a week or oftener if necessary. The
average calorific value per cubic foot of gas is 134 B.t.u., based on
analysis: COj, 8.6 per cent; O, 0.6 per cent; CO, 20.2 percent; H,
18.5 per cent and N, 52.1 per cent.
TABLE 1 RECORD OF LOAD AND FUEL FOR TWO HEAVY WEEKS
ELECTRICAL LOAD jRcfrigera-
'ting Load^
Kw.
hoursi
B.h.p.
hours^
Total
Load
B.h.p. B.h.p.
hours hours
Coal
Charged
Pounds
Sunday, July 25, 1908. . .
Monday, July 26
Tuesday, July 27
Wednesday, July 28
Thursday, July 29
Friday, July 30
Saturday, July 31
Sunday, August 22, 1P08
Monday, August 23
Tuesday, August 24
Wednesday, August 25
Thursday, August 26
Friday, August 27
Saturday, August 28
Totals
332
400
404
390
410
415
403
3.^8
386
393
392
397
391
393
5434
556
600
606
585
615
622
605
2010
2030
2020
2020
2030
2040
2020
549
679
589
588
596
586
590
2030
2020
2020
2010
2010
2020
2020
2566
2630
2626
2605
2645
2662
2625
3600
3900
3540
3180
4020
4320
4080
2579
2599
2609
2598
2606
2606
2610
3660
3420
3600
3540
3840
3540
3720
8266,
28300
36566
51960
1 Recorded by watt-hour meters.
2 Deduced from kilowatt-hours by assuming 80 per cent efficiency for the generator during
light-load periods and 90 per cent for the remaining time.
OPERATION OF PKODUCER GAS-POWER PLANT 219
DISCUSSION
J. A. Holmes. The success of the small producer plant using
anthracite coal is practically assured. Not long since (1905), in
visiting a number of these plants in Cologne, Germany, I found a
newspaper press that had been operated entirely for more than a
year by a small gas-producer plant burning small-sized anthracite
coal; one of the larger hotels there had been using such a plant for a
longer period with entire satisfaction to supply all its electric light
and power; in a large commercial house, electric lamps, elevators
and all other machinery connected with the establishment were
operated by one of these plants. In each of these cases the producer,
engine-driven generators and other equipment in the power room,
were all operated by one man, and the plant was regarded as a success
in efficiency and economy of labor and fuel. In the United States,
also, many producer plants have been using anthracite coal for some
years. In our own investigations at the Government testing station,
anthracite coal has been regarded as a fuel so simple and so easily
regulated that we have done little work on it, turning our attention
mainly to the bituminous coal producer problems.
2 In regard to producer work with bituminous coal, we have in-
vestigated fuels rather than different types of producers. Using every
imaginable grade of bituminous coal and lignite in making short-time
tests, we have encountered many difficulties due to a lack of famil-
iarity with the special manipulations required by certain fuels. These
difficulties, one being to secure a uniform quality of producer gas,
would not be met in using the same fuel year after year. In early
work, with the Taylor producer, we could get gas of absolutely uni-
form character not more than an hour at a time, and the variation in
24 hours was at times from 125 B.t.u. to more than 200 B.t.u. per
cu. ft. of gas, these variations being largely due, no doubt, to inex-
perience in the handling of any special fuel. During the past three
years, however, with more experience, the regularity and efficiency
of this gas have been greatly increased.
3 Another difficulty, and one not entirely separable as yet, is the
slagging or clinkering of the ash in the producer. The ash in certain
coals slags more readily than in others; and different ashes slag more
readily at different temperatures. One of the greatest needs in pro-
ducer development at the present time is that of a regular mechanical
feed of coal and removal of the ashes which now accumulate in some
producers, to be cleaned out after the producer has cooled down. We
220 DISCUSSION
have sometimes found the slag from certain coals, burned at high
temperatures, accumulating irregularly on the brick walls lining the
producer, at the rate of 6 in. to 10 in. during a week's run. If me-
chanical arrangements can be devised, by which the ash may be re-
moved from the base of the producer as regularly as from the base of
a boiler, then the use of a double producer will be largely avoided.
Decided progress is being made in overcoming this difficulty.
4 Still another line of progress is in the reduction of weight and
bulk of the producer making its use possible instead of that of steam
boilers for propelling ships. Mr. Straub's paper indicates what is
being accomplished along this line. Already the anthracite producer
and gas engine have been reduced in size and weight to less than
those of the steam boiler and reciprocating engine; and the outlook
is hopeful for the producer burning bituminous coal.
John H. Norris. I have been connected with the manufacture
of gas engines for a number of years, and the principal trouble we
have had in the operation of gas engines of any size is to overcome the
notion that a gas engine needs no care. At the present time, however,
gas engines are running successfully because in most installations they
receive proper attention. I am glad to see put on record the state-
ment that a gas engine installation needs as close attention as a steam
engine installation.
William A. Bole. The Westinghouse Machine Company has
been working on the gas-producer problem as well as on the gas-
engine problem for some time, and now believes itself ready to offer
gas producers that will be as practical and as easily manipulated
and capable of as long-continued runs as any boiler plant. A pro-
ducer plant of 175-b.h.p. capacity has been in operation at our works
for practically a year, without pulling down the fires. During that
period all sorts of runs have been made, continuous runs at full capac-
ity for ten days or two weeks, and the more ordinary runs in which
producer and engine are shut down at night; and the producers have
burned not only the comparatively good coals of the Pittsburg dis-
trict and the better coals of the Pocahontas region, but several of
the Western and Southwestern lignites and even peats from New
England. The latter have not been so successfully burned, but the
success in burning Colorado lignites has been very decided. This
producer was shut down and cleaned out, simply by shoveling the
ashes out of the water seal, and observations of the condition of the
OPERATION OF PRODUCER-GAS POWER PLANT 221
interior walls showed that it might just as well have been operated
continuously for five years instead of one, or as many years as the
firebrick lining would last. The requirements for continuous per-
formance seem to have been admirably met in this design.
2 This producer is designed for the burning of bituminous coal
alone, and resembles a small producer inverted and placed on top of a
large one, making a double-zone producer especially adapted for the
gasification of bituminous coal without passing tar of any descrip-
tion out of the producer shaft. Apparently the only solid material
emitted from the gas is a small amount of lamp-black which is success-
fully removed by the use of a static or stationary scrubber, and the
cleanliness of the gas is proved by the fact that practically all the
gas was converted into brake horsepower by being employed in the
actual operation of a gas engine, without troublesome deposit of any
kind upon the ports or other parts of the gas engine.
3 Whether such a producer would be available for marine pur-
poses I do not know ; the only question seems to be whether the motion
of the ship would interfere seriously with the descent of the fuel from
top to bottom. The producer has been subjected to practically every
test, and we believe we are about ready to offer it for both large and
small plants.
The author desired to present no closure. — Editor.
No. 1240
OFFSETTING CYLINDERS IN SINGLE-ACTING
ENGINES
By Prof. Thurston M. Phetteplacb, Providence, R. I.
Member of the Society
A great deal has been said recently about the offsetting of cylinders
in single-acting engines and many claims of superiority are made by
those who employ this form of construction.
2 About twenty-five manufacturing establishments in the United
States are building engines in which the cylinders are offset, chiefly
those of the automobile type, and one company is formed for the pur-
pose of making engines in which the offset is equal to the crank radius
and the connecting rod length is about 3f times the crank radius.
3 Among the claims made by manufacturers for offset engines
are: greater power, less side-pressure of the piston on the walls of
the cylinder, better turning effort, less vibration, smoother running
qualities, and when one cam shaft is used, a more convenient mechan-
ical arrangement.
4 On account of the importance of this subject and the lack of
information concerning it, a complete discussion is desirable and is
here presented.
5 The cylinder of an engine is said to be offset when its center-
Une is not in a plane through the center of the crank shaft. The
practice is not new and is applied to both steam and gas engines hav-
ing one or any number of cylinders.
6 In the diagram, Fig. 1, AB represents the stroke, OE the crank
radius, DE the connecting rod, 6 the crank angle, and OC the offset.
It should be noticed that 0 is the angle the crank makes with a line
through the center of the crank shaft parallel to the center-line of the
The full development of the mathematical formulae of this paper, with some
other related matter, is given in an unpublished Appendix, which is on file in
the Library of the Society.
Presented at the Spring Meeting, Washington, May 1909, of The American
Society of Mechanical Engineers.
224
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
cylinder, and not the actual angle passed over from the inner dead
point. The length of the stroke is
AB = R (1/ (a + 1 y- ¥ - v. (a - 1)^ - P)
which is greater than 2R.
R = crank radius.
a = L/R.
L = connecting rod length.
k = offset divided by R.
Fig. 1 Diagram of Crank and Connecting-Rod Train
Thus for 3-in. crank radius the strokes would be as shown in Table 1 .
The distance to the end of the stroke farther awaj^ from the center
of the crank shaft is shortened, thus OM is less than DE + EO or
L -\- R which affects the height of the engine.
TABLE l^^LENGTHS OF STROKES FOR DIFFERENT OFFSETS, 3-IN. CRANK
ENGINE
Ratio L/R
Offset
R
Stroke
Increase
Per cent
Any
eero
6.00000
0.00
3
0.10
6.00375
0.06
3
1.00
6.42279
7.04
4
1.00
6.21180
3.50
5
1.00
6.12930
2.15
6
1.00
6.08730
1.01
7 The dead points are not opposite each other, -so that the crank
angle swept over while the piston makes the out-stroke is less than
that for the in-stroke, causing a quick return motion and an average
velocity for the in-stroke or compression and exhaust strokes greater
than for the out-stroke, or explosion and suction strokes.
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
225
8 An expression for the piston position in terms of the crank angle
6 is developed in the usual way and is
X/R = V {a-\-iy -k^ - coBd - V a" - {k-sin Oy
in which X = the piston displacement from the end of the stroke
farther from the crank shaft.
9 The force of inertia due to the reciprocating parts is equal to
the weight multiplied by the acceleration divided by 32.2. The value
for the acceleration is found by differentiating the expression for the
piston displacement twice with respect to the time. This is done by
expanding the radical V a^ — (k — sin 6y by the binomial theorem,
into a convergent series and then dropping all terms containing a
with a negative exponent of 3 or larger in order to get an expression
that can be easily differentiated. This gives
[a^ - {k - sin)2 ]
i = n -
i a-* k^ + a-» k sin 6 - i a"* sin^ 0
This approximate expression for the radical differs from the radical
for different values of k, a and 6, as shown in Table 2.
TABLE 2 DIFFERENCE BETWEEN EXACT AND APPROXIMATE EXPRESSIONS
l/o=!- (fc -
6)2
— ia"i sin' 6
Difference
1
6
90
6.000000
6.000000
zero
1
6
0
5.916079
5.916666
+ .000587
1
6
45
5.992762
5.992867
+ .000105
.5
6
90
5.979130
5.979166
+ .000036
.5
6
0
5.979130
5.979166
+ .000036
.5
6
45
5.996428
5.996429
+ .000001
.5
3
90
2.958039
2.958333
+ .000294
.5
3
0
2.958039
2.958333
+ .000294
.5
3
45
2.992849
2.992859
+ .000010
.6
4i
90
4.472136
4.472222
+ .000086
.6
4i
0
4.472136
4.472222
+ .000086
.5
44
45
4.495236
4.495239
+ .000003
10 The greatest difference has no significant figure until the fourth
decimal place is reached and this is when A; = 1, which is an unusual
value. Hence it is readily seen that the error introduced by this
approximate form is slight.
11 Substituting this value of the radical in the expression for the
piston displacement and differentiating twice with respect to the time
gives
F/A = 0.00034 W/A N^R (cT^k sin d -i- cos 6 + oT^ cos 2d)
226 OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
which is the expression for the inertia force per square inch of piston
head area when there is an offset.
A = area of piston head.
W = weight of the reciprocating parts.
iV = revokitions per minute.
R = crank radius in feet.
This differs from the similar expression when there is no offset by
the addition of the term a"% sin 6, so that tables for inertia factors
for no offset may be used by adding the value of this term.
12 The expression for the tangential pressure or the turning force
for any offset is
T = Pl sin^+ cos^ sin^-A:
1
i a-^ ¥ + a"' kQmd-\ oT^ sin^ 0
i n which P is the pressure on the piston pin in the direction of the cen-
ter of the cylinder. This is a long expression to solve and a graphical
solution may be followed if preferred. The work of solving the ex-
pressions for inertia force and tangential pressure may be somewhat
lessened by tabulating the quantity a~' k sin 6 which appears in these
expressions.
13 The derivation of the preceding formulae and tables is shown
in the appendix.
SIDE PRESSURE OF PISTON ON CYLINDER WALLS
14 A reduction of the side pressure of the piston on the cylinder
walls is one of the advantages claimed for offsetting.
15 There are two ways in which the side pressure may affect the
single-acting engine : (a) The maximum value of the side pressure de-
termines the length of piston to keep the maximum pressure per square
inch of projected area below a value which is assumed as not too great
to destroy the oil film between the rubbing surfaces; (b) The average
value of the side pressure produces the friction between the sliding
surfaces causing a mechanical loss and some wear of the parts. The
loss in mechanical efficiency is more important than the wear, especi-
ally in the small high-speed automobile engines.
16 The average side pressures may be found by adding all of the
areas between the axis and the curve of side pressures and dividing
by the total length, or the areas themselves may be taken for com-
parison, as they represent the work done on the side of the cylinder
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
227
by the piston, which is lost work and should be kept as low as possible.
Curves of side pressures of the piston on the cylinder walls were con-
structed, it being necessary (a) to assume a gas card, (6) to assume
engine dimensions, (c) to calculate inertia forces and plot curves,
0 13 3 4 5 6.
Fig. 2 Offsetting Ctunders in Singlb-Actinq Engines
Gasolene Card Compression, 70 lb.; Maximum Pressure, 259 lb.; Pressure
ratio, 3.77
(d) to combine inertia forces with gas pressures, obtaining the force at
the piston pin, and then (e) to determine the side pressure component
perpendicular to the center-line of the cylinder for the changing angu-
larity of the connecting rod. The gasolene card chosen is shown in
Fig. 2. As it seemed desirable to investigate two similar cases, one
TABLE 3 CASES INVESTIGATED
High
Specifications Slow
R.p.m 450 1500
W/A lib. 0.70 1b.
R 6 in. = 0.5 ft. 2i in. = 0.208 ft.
O.OOOSiW/A mR 34.4 111.38
228
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
for high speed and the other for slow speed, the dimensions given in
Table 3 were chosen.
TABLE 4 PISTON POSITION FACTORS
Calculated by the Formula
X/R = V(a + 1)2 - A;2 - cos 6* - a + i a "1 A;2 - a -ifc sin0 + i a "isin^©
From Beginning of Stroke Towards the Crank Shaft. Multiply by Crank Radius
TO Find Piston Position. (Note: Crank Radius is Not One-Half of the Stroke)
Crank Angle
L-^R = Z
L -s-fi = 4i
Offset = 0.30 iJ
Offset = 0.50 5
Offset = 0.30 2?
Offset = 0.50 B
4°18'
0
7°11'
0
307'
0
5°13'
0
15
0.023
0.012
0.026
0.018
30
0.129
0.103
0.130
0.111
45
0.309
0.269
0.303
0.275
60
0.525
0.491
0.527
0.492
75
0.804
0.746
0.782
0.742
go
1.070
1.010
1.046
1.005
105
1.321
1.263
1.299
1.26
120
1.525
1.491
1.527
1.49
135
1.723
1.683
1.717
1.69
150
1.861
1.835
1.862
1.84
165
1.955
1.944
1.958
1.95
180
2.0037
2.0102
2.0018
2.0049
188°38'
2.0114
194°29'
2.0321
184°55'
2.0047
188''13'
2.0131
195
2.0065
2 .0303
1.992
2.007
210
1.961
2.001
1.929
1.954
225
1.865
1.918
1.810
1.846
240
1.699
1.779
1.643
1.684
255
1.515
1.585
1.429
1.475
270
1.270
1.343
1.179
1.227
285
0.997
1.068
0.911
0.957
300
0.699
0.779
0.643
0.684
315
0.471
0.504
0.407
0.432
330
0.229
0.269
0.197
0.222
345
0.075
0.098
0.061
0.075
360
0.0037
0.0102
0.0018
0.0049
17 The piston position factors and inertia factors are given in
Tables 4 and 5, and Fig. 3 to Fig. 8 give the curves of inertia forces.
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
229
The full lines represent the slow-speed and the mixed lines the high-
speed cases. Inertia curves and side-pressure curves were plotted
for ratios of l^jR of A.\ and 3, and for offsets of zero, 0.30 i2, and 0.50 i2
for both high and slow speeds, making twelve cases in all. When
there is an offset the inertia curve must be plotted for 360 deg. instead
of 180, since for the return stroke it is not the reverse of that for the
forward stroke, as is the case when there is no offset.
TABLE 5 INERTIA FACTORS
Calculated by Formula (o"*&siii0 + cos^ + o ■* cos 20)
Angle
L/R
= 3
L/R = 3i
L/«=4
L/R = 4i
K = 0.30
K = 0.50
K = 0.20
if = 0.30
K = 0.30
K = 0.40
K = 0.50
15
1.280
1.297
1.229
1.200
1.175
1.181
1.187
30
1.083
1.116
1.037
1.028
1.010
1.021
1.033
45
.778
.825
.747
.760
.754
.769
.785
60
.419
.477
.406
.440
.447
.465
.485
75
.067
.131
.066
.104
.131
.152
.174
90
-.233
-.166
-.229
-.175
-.156
-.134 -.111
105
-.450
-.386
-.451
-.404
-.385
-.364 -.344
120
-.580
-.523
-.594
-.560
-.553
-.535 -.515
135
-.636
-.589
-.666
-.654
-.660
-.645 ! -.629
150
-.650
-.617
-.695
-.704
- .722
-.711 i -.699
165
-.653
-.635
-.703
-.731
-.757
-.751 -.745
180
-.667
-.667
-.714
-.750
-.778
-.778 -.777
195
-.703
-.721
-.733
-.769
-.791
-.797 -.803
210
-.750
-.783
-.752
-.778
-.788
-.799 1 -.810
225
-.778
-.825
-.747
-.760
-.754
-.769 1 -.785
240
-.753
-.811
-.692
-.690
-.669
-.687
-.707
255
-.644
-.708
-.561
-.548
-.513
-.534
-.556
270
-.433
-.500
-.343
-.325
-.288
-.310
-.333
285
-.127
-.191
-.044
-.040
.003
.018
-.040
300
.246
.189
.308
.310
.331
.313
.293
315
.636
.589
.667
.654
.660
.545 .629
330
.983
.950
.981
.954
.944
.933 .917
345
1.228
1.211
1.199
1.164
1.141
1.135 1.129
360
1.333
1.333
1.286
1.250
1.222
1.222 1.222
18 Comparing Fig. 3 with Fig, 6 a slight hump is noticed at the
right-hand side in the former but not in the latter. This is probably-
due to error in the formula, for the small value of L/R since the force
could not be higher near the end of the stroke than at the end.
19 The general effect of offsetting on the inertia curve is shown
in Fig. 9, where the curves for L/R = 3, offsets = zero and 0.50 R,
are compared, the curve for no-offset being in full lines.
230
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
SLOW SPEED
/
/
/
/
/
/
HIGH SPEED
/
Fig. 3 Offset = Zero-
away FROM
CRANK SHAFT
Fig. 4 Offset = 0.30 R.
Curves of Inertia Forces on Piston Position Base
W
Slow Speed: r.p.m. = 450; R = 6; ^ = 1 lb.; L -^ R = 3.
W
High Speed: r.p.m. = 1500; R = 2^ in.; ~ = 0.7 lb.; L ^ R = 3.
Full Lines, Slow Speed; Mixed Lines, High Speed.
Fig. 5 L -^ R = 3. Offset = 0.50 R. Fig. 6 L ^ R = 4i. Offset = zero.
Cdkvbjs of Inertia Forcbs on Piston Position Bask
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES 231
/
/:
^.^^ / TOWARDS
^ /
/
/
CRANK SHAFT /
/
V
/ .
l^^^
___--'
7
/
AWAY FROM
CRANK SHAFT
_,./
7
TOWARDS
" /
:RANK SHAFT /
/
/
/
/
/
/
/
AWAY FROM
CRANK SHAFT
\y
Fig. 7 L ^ R= 4^. Offset= 0.30 R. Fig. 8 L ^ R = 4^. Offset = 0.50 R
Curves of Inertia Forces on Piston Position Base
Fig. 9 Inertia Force Curves Showing Effect of Offsetting
L-5-R = 3, High Speed. Full Lines, Offset = Zero; Mixed Lines, Offset - 0.60 R
232
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
20 The curve for an offset is a little flatter near the end of the out
stroke and the hump is increased near the beginning of the return
stroke, which is probably due to inaccuracy in the formula.
21 The curves of side-pressures are shown in Fig. 10 to 15. The
maximum side-pressure, its cause (whether combined gas and inertia
pressure or inertia force alone) and its location are given in Table 6.
TABLE 6 MAXIMUM SIDE PRESSURES, CAUSE AND LOCATION
MAX. SIDE-PREBSUBE
Ratio L/R Offset
Slow
High
Slow
High
Slow
High
4i
0
25
26
Gas
Gas
1
4i
0.30 iJ
17
24
Gas
Inertia
2
4J
0.50 B
12
28
Gas
Inertia
2
3
0
35i
45
Gas
Gas
1
3
0.30 B
23
39
Gas
Inertia
2
3
0.50 «
20
51
Inertia
Inertia
2
2
22 From this for L/R = 4J, slow-speed, maximum pressure is
lowest with 0.50 R offset, and if the offset were further increased the
maximum side-pressure would probably not be reduced as the values
at the beginning of the second and fourth strokes would increase, and
now they are already 11 so that any increase would soon cause an
increase in the maximum value instead of a decrease. In the case
of L/R = 3 the lowest maximum value occurs when the offset is
0.50 R, but in this case it is possible that the offset is already a trifle
large, as the maximum value occurs at the beginning of the second
stroke, although it is not much greater than that in the first stroke,
being 20 in the former case and 18 in the latter. Hence for the slow
speed the best offset would seem to be about 0.50 R as far as the max-
imum value of side-pressure is concerned.
23 In the ca.se of L/R = 4^, high speed, the maximum side-pres-
sure due to inertia force at the beginning of the second stroke seems
to increase with the amount of offset, while the maximum value due
to the gas pressure in the first stroke seems to decrease with the in-
crease in offset. These values are shown in Table 7. L/R = 4^.
TABLE 7 COMPARISON OF SIDE PRESSURES FOR L/R = 4i
Offset
Zero
0.30 fi
0.50B
Side pressures due to Gas pressure, 1st stroke. . .
Inertia, 2d stroke
26 17 13
15 j 24 ! 28
OFFSETTING CYUNDERS IN SINGLE-ACTING ENGINES 233
^
/
Fig. 10 Curve of Side Pressukes on Piston Position Base
C -T- R = 3, No Offset. Full Lines, Slow Speed; Mixed Lines, High Speed
Fig. 11 Curve of Side Pressures on Piston Position Base
L - R = 3, Offset = 0.30 R. FuU Lines, Slow Speed; Mixed Lines, High Speed
Fig. 12 Curve of Side Pressures on Piston Position Base
L -f- R = 3, Offset = 0.50 R. Full Lines, Slow Speed; Mixed Lines, High Speed
234
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
Fig. 13 Curve of Side Pressures on Piston Position Base
L ^ R = 4J, Offset = Zero. Full Lines, Slow Speed; Mixed Lines, High Speed
Fig. 14 Curve of Side Pressures on Piston Position Base
L ^R = 4|, Offset =- 0.30 R. Full Lines, Slow Speed; Mixed Lines, High Speed
7f^
/^
Fig. 15 Curve of Side Pressures on Piston Position'/Base
L -^ R = 4i, Offset = 0.50 R. Full Lines, Slow Speed; Mixed Lines, High Speed
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
235
24 Plotting curves of these values, the most favorable offset as
far as maximum side-pressure is concerned is 0.16 R when L/R = A\,
This curve is shown in Fig. 16.
25 See Table 8 for values L/R = 3. This would place the best
offset for L/R = 3, as far as maximum side-pressure is concerned,
TABLE 8
COMPARISON OF SIDE PRESSURES FOR
l/r = z
Offset
i
Zero
0.30K 1
0.50 K
Side pressures due to
Gas pressure, 1st stroke . . .
45
25
30
39
1
20
51
as 0.20 R, which would seem to indicate that a greater offset would
be desirable as the ratio L/R decreased. It remains to determine if
possible the best offset as far as the work done in side-pressm-e is con-
cerned.
26 The work done is proportional to the areas included between
the axis and the curve of side-pressures. It seems to make no dif-
ference whether a larger amount of work is done on one side than
on the other, or in other words there seems to be no advantage in
having the work done on each side the same, unless at some time it
might be desired to rebore the cylinder, in which case wear occurring
all on one side might have left the walls too thin or might necessi-
tate the removal of much more metal.
Fig. 16 Curve Showing Variation of Side Pressure with Offset
27 Table 9 shows the results of measuring the areas, namely the
ratio of work done in one case to that in each of the other cases, and
also the actual average side-pressures. For the slow-speed case there
seems to be a decrease of work done on one side, and an increase of
work done on the other side, resulting in a decrease in the total work
done with the offset, which would indicate that the greater the off-
set, the less the loss in work. The average side-pressure also decreases
with the offset, although it is less with no-offset when L/R = 4^
than with 0.50 R offset when L/R = 3.
236
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
28 In the case of the high speed the areas on one side decrease with
the offset while those on the other side increase, but the totals for
L/R = 4^ decrease and then increase, while for L/R = 3 they con-
TABLE 9
Ratio
L/R
Offset
BLOW BPEBD
+ Area
— Area
Total
Area
Average
I Side
Pressure
HIGH SPEED
H-Area
-Area
Total
Area
I Average
Side
Pressure
4i
4i
44
3
3
3
0
0.30 fi
0.50 iJ
0
0.30R
0.50R
1.61
1.09
0.74
2.59
1.55
1.11
2.14
1.99
1.67
3.51
2.70
2.50
6.6
2.33
6.2
1.78
5.2
1.62
11.
4.11
8.4
3.01
7.8
2.47
-1.11
-1.58
-1.90
-1.45
-2.76
-3.40
3.44
3.36
3.52
5.56
5.77
5.87
10.7
10.5
11.
17.3
18.0
18.3
tinue to increase, and of course the same is true for the mean side-
pressure.
29 This would seem to indicate that there is little if anything to
be gained by an offset in regard to work done by the piston on the
walls of the cyhnder when the inertia force is very high, since what
is gained on one side is more than made up in loss on the other side.
30 If it is of sufficient importance to have the work done on each
side of the cylinder the same, we may plot curves of the work done
on each side and note where they intersect, as in Fig. 17 . In the case
of the slow speed we would have the work done on each side equal
when the offset was about 0.40 R and in the high speed this point
would be about 0.36 R.
31 Thermal Cycle. Offsetting increases the length of stroke,
which gives increased expansion to the gas, and increases the piston
velocity on the in-stroke, giving greater inertia to the gas on the
exhaust and reducing the amount of leakage by the piston on the
compression stroke. This refers to the 4-cycle gas engine.
32 Lubrication. The curves of side-pressure show the manner in
which the side-pressure changes sides, which is a good thing for
lubrication. This changing sides would be about the same for off-
set or no-offset except in the case when the offset is equal to the crank
rachus. Here the pressure is almost continually on one side of the
cylinder so that oil would with difficulty be introduced between the
surfaces. Other tilings being even, except for this extreme case,
the reduction in amount of side-pressure should make lubrication
more satisfactory.
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
237
33 Vibration and Balance. Revolving masses and reciprocating
masses may cause vibration in gas engines. Offsetting the cylinders
would not affect the revolving masses at all but does change the curves
of inertia forces, as already shown in Fig. 3 to 8. These inertia-
force diagrams are now combined in different ways according to dif-
ferent arrangements of cylinders, and are compared with similar
curves when there is no offset.
SLOW SPEED
HIGH SPEED
Fig. 17
.:30 .38
Curves Showing Offset when Work is same on Each Side
OF Cylinder
34 The following discussion applies only to the 4-cycIe type of
gas engine, whose arrangements are:
a Single cylinder.
b Two-cylinder vertical.
c Two-cylinder opposed.
d Three-cylinder vertical.
e Four-cylinder vertical.
/ Four-cylinder double-opposod.
g Six-cylinder vertical.
238
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
35 For this comparison the high speed case, when L/R = 4-^, was
chosen, the offset being equal to zero and one-half the crank radius,
36 Fig. ]8 shows the inertia curves for a single-cylinder engine.
These curves must be shown for 360 deg. of crank angle, for they dif-
fer on the return and forward strokes. The curves for no-offset are
shown in full lines and for 0.50 R offset in dotted lines. The difference
between the two curves is apparent.
Fig. 18 Curves of Inertia Forces
ON Piston Position Base. Single-
Cylinder Engine
L -- R = 4J. Full Line, Offset =
Zero; Dotted Line, Offset = 0.50 R.
High Speed Case
Fig. 19 Curves of Free Unbal-
anced Inertia Forces. Two-Cyl-
inder Vertical Engine
L -f- R = 4^; Full Line, Offset = Zero;
Dotted Line, Offset =0.50 R. High
Speed Case
37 Fig. 19 shows the curve of free inertia forces for a two-cylinder
vertical arrangement. These curves are not so very different; the
one for an offset being nearly the same as the other but moved along
a little instead of being symmetrical with a center line perpendicular
to the axis. The maximum values of the forces are about the same.
The vibrations when there is an offset would have unequal periods
but about the same amplitudes. For the four-cylinder vertical case
the ordinates of these curves could be doubled and the same general
difference would exist.
38 In the case of the two-cylinder opposed motor with cranks
at 180 deg., the inertia forces^wouk^be balanced whether the cylin-
ders were offset or not, but in the case of an offset a new couple in
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES 239
a plane perpendicular to the axis would be introduced due wholly to
the offsetting, which cannot be balanced. The couple in an axial
plane due to the cylinders being not in line would be the same, offset
or not, but with an offset there would be added another couple in this
plane due to the offset, which would not be balanced.
39 In the case of a four-cylinder double-opposed motor the forces
would be balanced and also the coujDles in an axial plane, but the
couples in the plane perpendicular to the axis would be doubled while
those in the axial plane due to the offset would be balanced.
40 The case of a three-cylinder vertical arrangement can be dis-
cussed by considering the formula for the inertia forces,
F/A = 0.00034 W/A N^'R {a-'k sin 6 -\- cosd + a-' cos 2 6)
Let the cranks be at 120 deg.; then the crank angles will he 6,6 + 120,
and 6 + 240. Substituting these values in the formula, the part in
brackets reduces to zero, showing that the inertia-forces are balanced.
However, the moments resulting from these forces are not balanced.
By placing two three-cylinder vertical engines together so that the
two middle cranks are in the same plane the six-cyhnder engine is
obtained, in which the inertia forces and couples are both balanced.
41 From this discussion it follows that offsetting the cylinders
has no effect on the vibration due to the reciprocating parts, except
in the case of the 2-cylinder opposed and 4-cylinder double-opposed
arrangements of cyhnders. In these cases the offsetting increases
the unbalanced inertia-force couples by adding new ones.
42 Vibration may be felt from the irregularity of the turning-
effort curves, which for four different cases are shown in Fig. 22.
There is such a slight difference here that it can be neglected, espe-
cially since the turning-effort curve depends so directly on the shape
of the gas card, which may vary considerably. The conclusion in
regard to vibration would be that offsetting does not affect the vibra-
tion appreciably except in the case of a two-cylinder opposed or a
four-cylinder double-opposed motor.
GENERAL CONCLUSIONS
43 The following are perfect^ general conclusions, to be followed
by a more definite comparison of actual engines:
a The length of stroke for a given crank radius increases as
the offset increases.
240 OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
h The length of stroke for a given crank radius for any off-
set decreases as the ratio of L/R increases.
c The increase in length of stroke causes an increase in aver-
age piston speed.
d Offsetting the cyUnders makes the crank and connecting
rod train a quick return mechanism.
e When the cylinders are offset the crank passes over an
angle greater than 180 deg. during the out-stroke of the
piston, and less than 180 deg. during the in-stroke.
/ The average velocity of the piston is greater on the exhaust
and compression than on the explosion and suction strokes.
g Offsetting the cylinders reduces the angularity of the con-
necting rod on the out-stroke and increases it on the in-
stroke.
h When there is an offset, the side-pressure of the piston on
the cylinder walls does not change sides at the end of the
stroke, but just after the beginning and just before the
end of the out-stroke.
{ The place where this change of side-pressure occurs ap-
proaches the middle of the stroke as the amount of offset
approaches the crank radius.
/ With no offset, liigh inertia forces do not greatly increase
the maximum side-pressure during the explosion stroke,
but do increase it considerably during all of the other
strokes, and this effect is slightly greater as the ratio of
L/R decreases.
k With no offset the work done increases with the inertia-
force and as the ratio of L/R decreases.
I For low inertia forces, as far as the maximum value of
side-pressure is concerned the best offset is one-half the
crank radius.
m Considering the maximum value of the side-pressure only,
the most favorable value for the offset decreases as the
inertia-forces increase, for any ratio of L/R, but does not
decrease as rapidly, for smaller values of the ratio L/R.
n For low inertia forces, the work done by the piston on the
cylinder walls decreases as the offset increases, but of
course is greater for smaller values of L/R.
0 For very high inertia forces, the work done decreases
slightly with the offset up to 0.40 of the crank radius for
value of L/R = 4^.
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
241
p For very high inertia forces, and small values of L/R,
there is no advantage in an offset, as far as the work done
by the piston on the cylinder walls is concerned.
q The thermal cycle is slightly benefited by offsetting and
the benefit increases with the amount of offset.
r Lubrication should be sUghtly improved by offsetting the
cylinders.
s Vibration due to the free inertia forces is no different ex-
cept in the case of a two-cylinder opposed or four-cylinder
double-opposed motor.
44 Table 10 gives data of gas engines that have been constructed
and put in operation. The average crank radius is about 2| in., the
TABLE 10 DATA OF GAS ENGINES HAVING CYLINDERS OFFSET
Ii Length
\
1
Ratio
No.
R Crank
of Con-
Ratio
Offset
Offset
Length of
Diam.
Piston
Radius
necting
L/R
Amount
Per cent
Pbton
of Bore
Length to
Rod
Diameter
1
7
231
3.37
7
100
14
9.47
1.47
2
2A
8it
3.48
1
24.
6
5i
1.09
3
2i
8i
3.77
i
16.6
6
4
1.5
4
2i
94
3.8
1
15
51
5
1.125
5
■ 2i
10
4.0
i
15
6i
4i
1.37
6
3
121
4.08
A
18.76
5J
7
2f
9f
4.1
i
21
5J
5
1.075
8
2i
lOi
4.2
i
35
5A
41
1.098
9
2i
12
4.36
i*
34
6i
5i
1.28
10
2i
lOi
4.66
i
38
5i
4i
1.29
11
2i
12
4.8
40
6
4J
1.26
12
2i
12
4.8
40
6i
5
1.25
13
2i
12
4.8
40
6
4J
1.26
14
2i
li
50
15
21
i
U
40
j ....
....
16
3
1
33 ^
....
17
2*
i^J'l„
'.'.'.'.
f
30
....
....
Westinghoxise standard engine has an offset of 50 per cent of crank radius.
ratio of L/R varies from 3,37 to 4.8 and the percentage of offset varies
from 15 to 50. The average diameter of cyhnder-bore is 4.81 in. and
the average ratio of length of piston to diameter is ] ,24.
242 OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
45 For comparison of engines the following dimensions were
taken :
Crank radius = 2^ in.
Diameter of bore = 4| in.
R.p.m. = 1000
Weight of reciprocating parts per square inch of piston head area
= 0.6 lb.
Ratios of L/R = 2>\, 4, and 4^ and an offset, for each of the values
of L/R, the largest amount practicable. These offsets are:
L/R = 4i Offset = zero
« = 4^ " =0.40 i2
" = 4 " =0.30 i?
" = 3J " =0.20 R
46 Tables were prepared for each of the above cases, and values
calculated for crank angles varying^by increments of 15 deg. each.
Each of these tables contained values for the crank angle, piston posi-
tion factor, the actual piston position, the gas pressure, inertia factor,
inertia force, piston pin pressure, tangential factor, and the turning
force from which the inertia curves and turning effort curves were
plotted.
Fig. 20 Curve of Side Pres.sures on Piston Position Base
W
R.p.m. = 1000; j- = 0.60 lb.; R = 2\ in. Full Line, L -r- R = 4^; Offset = Zero
Broken Line, L -^ R = 4J; Offset = 0.40 R. Mixed Line, L -^ R = 4;
Offset = 0.30 R. Dotted Line, L -5- R = 3^; Offset = 0.20 R.
47 Careful comparison of the curves in Fig. 20 will show a slight
difference between them, but not enough to warrant the trouble of
plotting them separately for use in connection with the gas pressures
to find the piston-pin forces from which the side-pressures are deter-
mined.
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
243
48 The inertia forces shown in Fig. 21 were combined with the gas
pressures and the curves of side-pressures plotted as before, with
the results shown in Fig. 20.
49 The maximum values for the side- pressure were determined
and the areas representing the work done by the piston on the cylinde-
Fig. 21 Curves of Inertia
Forces ox Piston Posi-
tion Base
R.p.m. = 1000;
R=2i m;? =0.60 lb.
1 L -i- R = 4i Offset = Zero
2 L ^ R = 4^ Offset = 0.40 R
3 L -T- R = 4 Offset = 0.30 R
4 L -=- R = 3^ Offset = 0.20 R
Fig 22 Turning Effort Curves on
Crank Angle Base
Full Line, L -- R = 4^ Offset = Zero
Mixed Line, L -^ R = 3^ Offset = 0.20 R
Dotted Line, L-T-R=4i Offset = 0.40R
Broken Line, L -f- R = 4 Offset = 0.30 R
walls were carefully measured and recorded (see Table 11). As far as
these quantities are concerned, the best value is L/R = 4|, offset =
0.40 R. The turning-effort curves (shown in Fig. 22) are so nearly
alike that the difference is hardly worth mentioning.
TABLE 11
SIDE PRESSURES AND WORK DONE ON CYLINDER BY PISTON OF
TYPICAL ENGINE
1
MAX. SIDE
PRESBDBB
WORK
DONE IN AREA
UNITS
L/R
Offset
One Side
Other Side
One Side
Other Side
Total
4J
zero
24
7
1.55
0.60
2.15
4i
0.40
12
12
0.93
0.87
1.80
4
0.30
17
12
1.17
0.82
1.99
3*
0.20
i "
1 12
1
1.53
0.81
2.54
50 Table 12 gives a comparison of the four cases chosen. This
table explains itself, but a short discussion may bring out the impor-
tant points more clearly.
244 OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
51 There is a slight increase in the length of stroke, but less than
one-half of one per cent, so that it amounts to very little. The angle
passed over by the crank during the out-stroke is slightly greater than
180 deg. and the greatest gain is 1.7 per cent, which is small. The
first great difference occurs in the length of connecting rod. No. 4
effecting a saving of 2.50 in, or 22.2 per cent.
52 Referring to the next line, the distance from the center of the
crank-shaft to the position of the center of the piston pin at the end
of the stroke, is a measure of the height of the engine and shows a
decrease corresponding to the value of L/R.
53 The maximum side-pressure decreases with the offset and
increases with the decrease in value of the ratio L/R, so the best case
would be No. 2, where L/R is largest and the offset is also largest.
Here a reduction of 50 per cent is gained, which reduced the necessary
length of the piston 44 per cent. No. 4 is the worst case, L/R very
small and the offset also small and then the side-pressure is a trifle
less than it is with no offset. The maximum value of the side-pres-
sure affects the length of the piston and consequently the height of the
engine, and the length of the cylinder, and so the weight of the cyl-
inder and engine, and the weight of the piston and the correspond-
ing weight of the reciprocating parts, and so the inertia force. The
length of the piston is reduced 43.7 per cent in No. 2, 24.4 per cent
in No. 3, and 3.6 per cent in No. 4. The ratio of length of piston to
diameter is rather small in No. 2 but is not unusual in the other cases.
54 If it is not desired to take advantage of the maximum value of
the side-pressure by reducing the length of the piston, it can be made
1.20 times the diameter, a usual value as is seen in Table 10, which
would reduce the pressure per square inch of projected area and so
increase the chances of satisfactory lubrication. The reduction of
this pressure per square inch of projected area is shown in the next
row.
55 In order to find the exact resulting height of the cylinder up
to the top of the piston at the end of the in-stroke it is necessary to
calculate the position of the piston pin in the piston. This is done
in the next row, by making the sums of the products of the areas with
the distance from the piston pin center to their centers balance on
each side of the piston pin. In case No. 2, with L/R = A.\, a 50 per
cent offset might have been used without interference and this would
give better results than 0.40 R offset, but in case No. 4, 0.20 R is
undoubtedly about as much as could be used although it would be
desirable to use more if a very low engine were wanted.
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
245
^
- - kO ^ .3
O ■-" o^ 0> vp
US lO M »-i U5 o^
• • • • (N CO
O (N OO ■* • •
.-H CO
65
■<1< fe^ vO o> •
CD >-c •* (S~^ CO 00
• -MO • .H
N >0 ■ • I I
ft; OS
a; CO «
OS 00 CO
^
^
—
r,^!?
CO
CO Tt<
CM IN (N 6^
r-l ■*
• • ■ OS
e^i
-H ^ 05 <N
■^ ^ CO
-^ <=^. '^ 65
CV| (N iH •"ll
OS >o
CM c t^
o — ■ r^ CO
• O ■^ ,-1
CO • •
r- o c^ t^ CM lO
Oi CO
CM CO
c<i la
O 11
< I
^ £S
«65
^
CO
CD
CO
^
CO
5S
CO
^
05
CM
CM
CO
^s^
CM CO
i-l CO
CO rt
CO 00
CM CM CO
o C^ t>
O • ■
CO .-H CO
00 .-I
tf s
^•2
(D
13 en 2
in u a
o o ■".
Id g
S tf
^2 s i ►^
S g .2 I .2
iJ w
S -o
j3 p "O
Q Q
1-3
^ ^
246 OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
56 The next two rows show the distance from the center of the
crank-shaft to the end of the piston at the end of the in-stroke, which
is a measure of the height and the length of the cyhnder and also of
the w'eight of the cylinder. As regards the height a gain of 15 per
cent may be had in No. 4. No. 2 gives the shortest cylinder, 25 per
cent shorter than No. 1, wMle No. 4 gives one only 2 per cent shorter.
It must be borne in mind that these values are for 1000 r.p.m. and
that the value of the maximum side-pressure will increase with the
speed. However, the low value of the pressure per square inch of
projected area, 15 lb., allows a considerable increase before a dan-
gerous value is reached.
57 The total amount oi lost work is shown in the next line.
No. 2 gives the best value, a saving of 16 per cent, while No. 4 gives
a loss of 18 per cent.
58 In worldng out a satisfactory solution it would seem that one
of two predominating ideas should be followed. Either a very low
engine should be aimed at in which everything is sacrificed to height,
or else the important object is to reduce to a minimum the side-pres-
sure and the work lost due to friction resulting from side-pressure.
59 In the first case, let L/R = 3^, offset as much as possible with-
out interference, and a reduction in height of 13 to 15 per cent may
be had. This means a reduction and a saving in weight of the con-
necting rod, cylinder, valve stems, exhaust pipes, inlet pipes, and
piston. The actual saving in length in the case above is 2f in. There
will be some increase in the work lost in friction due to the increased
average pressure of the piston on the cylinder walls.
60 If a reduction in height is not of primary importance, then a
ratio of L/R = 4^ and an offset of 0.40 R to 0.50 R would seem to
give the best results. This gives a reduction in total height of 8 or
9 per cent, a reduction in piston length of 44 to 45 per cent, a reduction
in cylinder length of about 20 per cent, and a saving in lost work of
about 16 per cent. These reductions would cause a further reduction
in weight of piston, weight of cylinder, weight of valve stems,
weight of exhaust and inlet manifolds, and a reduction of inertia
effects as well as an increased life to the piston, piston rings and cyl-
inder. In this case it might not be desirable to take full advantage
of the reduction in length of piston, maldng it less than the stroke
because the oil hole in the side of the cylinder, if one were used,
would be uncovered at one end of the stroke or the other.
61 In concluding this comparison, the most desirable offset seems
to be as much as can be practically obtained with ratios of L/R = 4
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES 247
and greater, with a decided gain over an engine with no offset for
speeds less than 1400 to 1500 r.p.m. The subject may be summed
up as follows:
62 Improvements due to offsetting, (1) in the thermal cycle, (2)
in the mechanical arrangement, (3) in the turning effort curve, and
(4) in lubrication, are very shght and may be neglected. The real
advantages are:
a A reduction of the frictionai losses due to the pressure of
the piston on the walls of the cyhnder, resulting in a
slight increase in mechanical efficiency and less wear of
the piston, piston rings, and cylinder, and consequently
longer life.
h A reduction of the maximum value of the side-pressure
of the piston on the walls of the cylinder, allowing the
use of shorter connecting rods, shorter pistons, and
shorter cylinders, resulting in a shorter and lighter engine
and in lower inertia-forces due to the reciprocating parts.
The most important of these advantages would be a considerable
saving in weight.
63 The disadvantage of offsetting lies in the fact that the reduc-
tions in average side-pressure and maximum side-pressure grow less
as the speed and inertia-force increase, so that for a speed of 1400 to
1500 r.p.m. there is either no reduction at all or an increase.
Principal Conclusions
64 Offsetting increases slightly the length of stroke and the crank
angle passed over during the stroke toward the crank shaft.
65 The maximum value for the side-pressure of the piston on the
cyhnder walls decreases as the offset increases up to a value of one-
half the crank radius for any ratio of L/R.
66 The work lost in friction due to the side-pressure of the piston
on the cylinder walls decreases as the offset increases up to a value
of 50 per cent of the crank radius.
67 Both the maximum value of the side-pressure and the work
lost in friction increase as the value of the ratio L/R decreases.
68 Offsetting decreases the height and weight of the engine.
69 Offsetting increases the life of the cylinder and piston.
70 Offsetting improves the thermal cycle.
248 DISCUSSION
71 The turning-effort curves when the cylinders are offset differ
but slightly from those for no-offset.
72 The advantages of offsetting as regards the maximum side-
pressure and work lost may be zero or negative for high inertia-forces
resulting from speeds of 1500 r.p.m. or more.
DISCUSSION
WiNSLOw H. Herschel. In December 1907, 1 had occasion to inves-
tigate the question of offsetting cyHnders for large-sized gas engines,
and as the conditions are somewhat different from those of the auto-
mobile engines considered by Professor Phetteplace, the results may
be of interest. I shall consider only the variations in maximum and
average pressure on the cyhnder walls, since, as indicated in Par. 62
of the paper, these are the main questions at issue.
2 For the sake of simplicity I used the graphical method men-
tioned in Tolle's Die Regelung der Kraftmaschinen, page 32. The
computations were based on an actual card from a four-stroke cycle
producer-gas engine, and upon the following data:
R.p.m. = iV = 225
W/A =4.18
22 = 12 in. =1 ft.
L/R = 5
0.00034 W/AN^R = 71.8
This last value, 71.8, happens to be very nearly the average of the
two corresponding values, 34.4 and 111.38, given by the author in
Table 3. Computations were also made using speeds of 450 and 1000,
giving inertia constants of 289 and 1430 respectively, but it soon
became evident that there could be no gain from offsetting under
these conditions, and the investigation was restricted to the speed of
225.
3 As I understand the paper, the author has considered only
vertical engines, or horizontal engines where the weight of the recipro-
cating parts is so small that its direct effect in increasing or decreas-
ing the pressure on the cylinder walls need not be taken into account.
In the present case, however, a distinction must be made between
vertical and horizontal engines. For the latter, when the side pres-
sure acts downward, due to gas pressure or inertia forces, the weight
of the piston must be added, but when the side pressure acts upward,
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
249
the piston weight is subtracted. It should be noted that for a vertical
engine, for a given value of 0.00034 W/AN^R, that is, for a given
inertia constant, as I have called it, it would make no difference
whether this value were obtained with a large value of W/A and a
small value of A^, as in my case, or with a small value of W/A and a
large value of A^, as in the cases used by the author. But on the other
hand it would make considerable difference for a horizontal engine
where the value of W enters into the computation apart from the
inertia constant.
4 By using the same indicator card as for the four-stroke cycle
computations, and disregarding the second and third strokes, I
■iO^i 40^ 50^c 60^ 70f
Offset in Per Cent of Crank
90/
Fig. 1 DiAGKAM Showing Variation in Maximum Pressure on Cylinder
Walls, Due to Offsetting Cylinder
obtained, somewhat approximately, the side pressures for a two-
stroke cycle engine.
5 Fig. 1 shows the variation in maximum pressure on cylinder
walls, or side pressure, due to variations in offset. The ordinates
above the base line ST are proportional to the side pressures. The
line AB shows the maximum side pressure for a vertical engine at
about the middle of the first stroke. As the offset increases the
angle of the connecting rod for this middle position decreases, while
the angle at the end of the stroke increases, so that for large offsets
the maximum side pressure is found at the end of the stroke, with
250
DISCUSSION
values as shown by line BC. Similarly for a horizontal engine we get
the lines D£J and EF. For the maximum side pressure due to inertia
forces we have hne GH for a vertical, and Une JK for a horizontal
engine.
6 It will be noticed that line JK is not parallel to line GH. The
reason for this is that for the hne JK we must use the inertia force
near the end of the second stroke, which gives a downward pressure
on the cyhnder walls to be added to the weight of the piston, and this
downward pressure is not as large as the upward pressure near the
beginning of the second stroke, which was used for the Une GH.
7 In the case of a two-stroke cycle engine, where we must use
the fourth instead of the second stroke, our maximum inertia force
will be near the beginning of the fourth stroke acting upward, so that
the weight of the piston must be subtracted, giving hne QR.
8 If we imagine a somewhat earlier ignition than that shown in
TABLE 1 OFFSET AND PER CENT REDUCTION IN SIDE PRESSURE
Curves
Dominating Factors
Cass
%
Offset
Gain
AB and GH
Gas middle first, inertia beginning 2nd
Vertical late ignition, 2-4 Cycle
37.0
41.5
ABanALM
Gas middle first, gas beginning 1st
Vertical early ignition, 2-4 Cycle
28.7
31.6
DE and JK
Gas middle first, inertia end 2nd
Horizontal late ignition, 4 Cycle
37.7
34.7
DE and QR
Gas middle first, inertia beginning 4th
Horizontal late ignition, 2 Cycle
61.3
54.3
DE and JK
Gas middle first, inertia end 2nd
Horizontal early ignition, 4 Cycle
37.7
34 7
DE and NP
Gas middle first, gas beginning 1st
Horizontal early ignition, 2 Cycle
40.6
37.2
Fig. 2 of the paper, the maximum side pressm-e at or near the begin-
ning of the first stroke will be increased. Whether for this reason or
not, I found that with large offsets the maximum side pressure of the
first stroke was at the beginning of the stroke, acting upward, with
values as shown by line LM for a vertical, and line NP for a horizon-
tal engine.
9 Fig. 1 corresponds to Fig. 16 of the paper and may be used in the
same way to determine the most favorable offset for the various con-
ditions. Table 1 gives the offset and the per cent reduction in side
pressure in each case.
10 The author (Par. 22) finds the most favorable offset to be
50 per cent of the crank length for slow speed, and in Par. 24 and
Par. 25 he finds it to be about 20 per cent for high speed. These
OFFSETTING CYLINDERS IN SINGLE-ACTING ENGINES
251
values may be compared with the first fine of Table 1 ; for the case of a
vertical engine with late ignition, the offset is 37.0 per cent, which is
nearly the average of 20 per cent and 50 per cent, as might have been
expected from the inertia constants.
11 Fig. 2 shows the decrease or increase in average side pressure
or total loss of work from side pressure, in per cent of lost work with
zero offset. While the use of a different indicator card with a later
ignition might have made some difference, it obviously could not
have changed the results so materially as in the case of maximum side
Fig. 2
40S^
o.Cl^
^
a, 30^
\etv
X
y y^
„ 4-t
;ycle, Hi)iizonfni
\
IO5;
/y^^
1 " vJrf.r,,
^^""^ 1 J.'V) 1
!i2£2
r"-\^
\
5 10^
a
— I^o~
SS^-
^"ca;
1
^
\
10^
•iwi
i%
30ji 40!4 5U5J 6US^ 1%
Offset in Per Cent of Crank
Diagram Showing Change in Average Pressure on Cylinder
Walls Due to Offsetting Cylinder
pressure. Thus the curve marked four-cycle vertical may be fairly
compared with the result in Table 9 of the paper, that the most favor-
able offset lies between 30 and 50 per cent.
12 The curves marked 450 and 1000 r.p.m. show the results of
the few computations concerning these speeds not considered in Fig 1 .
13 Both Table 1 and Fig. 2 appear to indicate that more could
be gained from offsetting with a tAVo-cycle than with a four-cycle
engine. But at present it is difficult to make general statements
about this type of engine, and whether or not this advantage will be
attained will depend upon the inertia constant and indicator card
shown by these engines.
252 DISCUSSION
PRINCIPAL CONCLUSIONS
a An offset cylinder may be employed with least benefit on a
high-speed four-stroke cycle vertical engine.
h It may be employed with most benefit on a slow-speed
two-stroke cycle horizontal engine.
c The maximum advantageous offset is limited by the side
pressure near the beginning of the first stroke.
John H. Norris. I have been designing and building both two-
stroke and four-stroke cylinder engines with offset cranks for a num-
ber of years. Our concern was so impressed with the advantage
that in 1886 they bought the patent right to apply the offset stroke
to gas engines. We are still building a few small sizes with offset
crank, but there is no practical gain and as fast as we can re-design
the engines we find we can get better economy and a more convenient
engine, by placing the cylinder directly over the crank shaft. If you
want to reverse the direction of rotation of an engine with an offset
crank, you are in trouble at once. We have built an offset engine as
large as four-cylinder, 14 by 22, and scrapped it. I would like
to say, in connection with the large engine above mentioned, that it
was built in 1896 and was one of a pair that were to run a suburban
electric railroad in the West, on a suction gas producer. We buUt at
that time a suction producer that was reasonably satisfactory. We
have had a great deal of success, however, with our small single and
double-cylinder engines with offset cranks, of which we have built a
large number, though we are just putting on the market engines to
replace them, in which we have placed the cylinder directly over the
center. We used an offset of from 20 to 25 per cent of the stroke.
With this offset the side strain is quite sufficient on the upstroke.
These were all vertical engines.
The Author would suggest that the fact that the cylinders were
offset was not the real cause for scrapping the 4-C5dinder 14 by 22
engine mentioned by Mr. Norris. Of course if stock engines were built,
some to rotate in one direction and some in the other, or if it is desir-
able to build engines that may be reversed, offsetting may not prac-
tically be taken advantage of, as Mr, Norris points out. Furthermore,
in spite of Mr. Norris' experience and his desire to eradicate offsetting
from his product, this practice, in small vertical 4-cycle automobile
engines, at least, seems to be increasing.
No, 1241
PRESENTATION OF PORTRAIT OF GEORGE
AV. MELVILLE
At the Spring jNIeeting, \Yashiiigton, INIay 1909, of The American Society
OF Mechanical Engineers, a portrait of Rear-Admiral George W. Melville,
U. S. N., Ret., painted by Sigismond de Ivanowski, was presented by friends to
the National Gallery. Previous to the presentation of the portrait an address
was made by Admiral Melville on The Engineer in the L'''. S. Navy. This is
given in abstract, together with the addresses of presentation by Walter M.
McFarland, Mem.Am.Soc.M.E., and of acceptance of the portrait for the
Nation by Dr. C D. Walcott, Secretary of the Smithsonian Institution.
THE ENGINEER IN THE U. S. NA\^
By Rear- Admiral Geo. W. Melville, U.S.N., Ret.
Past-President of the Society
Ten years ago my presidential address before the Society had
almost the same theme as my remarks this evening. At that time
the Personnel Law was passed, which amalgamated the engineer
corps with the line or executive officers of the navy, with the under-
standing that thenceforth engineering was to be the function of
these Une-officers. In his report as Chairman of the Personnel Board,
ex-President Roosevelt, then Assistant Secretary of the Navy, said,
"On the modern war vessel every officer has to be an engineer
whether he wants to or not." It is well that these lines should be
constantly in mind, for they set forth the only justification for the
Personnel Law.
2 Remarks have been made to the effect that a line-officer
charged with all these duties would be a hybrid or Jack-of-all-
trades. It is to be noted, however, that our naval officers have to
perform definite duties. The curriculum of the naval school can
be planned to give them a thorough and specific training for the
work they have to do. In this respect these young men have a
decided advantage over the students of any of our great technical
schools, who can receive instruction only in general principles because
they themselves rarely know the particular line of engineering
work which they will follow. It is, in my judgment, just as ridicu-
lous to speak of our modern line-officers, specially trained for the work
254 PRESENTATION OF PORTRAIT OF ADMIRAL MELVILLE
they have to do, as Jacks-of-all-trades, as it would be to apply
this designation to a blacksmith, a lawyer or a doctor.
3 I have not forgotten that I am talking of men who go to sea
and that the line-officers are responsible for the handling of the ships;
in other words, that seamanship is an essential element of the
training. It must be remembered that the modern navy has entirely
dispensed with sails and that it is a misnomer to call the modern
man-of-warsman, a sailor. He is not a sailor, because there are
no sails for him to handle. He is a seaman, because he goes to sea.
Seamanship is an art, proficiency in which comes almost entirely
from practice, so that officers who are given the other portions of
the training in the classroom, laboratory and workshop, get the
requisite proficiency in seamanship from the practical exercises dur-
ing their career as midshipmen at the naval school, and in the
handling of vessels after they graduate.
4 It is natural to inquire how the amalgamation has worked
out in practice. On January 21 of this year, the Chairman of the
House Naval Committee quoted from the remarks of the officer
who commanded the battleship fleet which cruised around the
world, to the effect, " When I got to California, without any engi-
neers, my fleet was in better condition than when it started." It
would seem, however, to have agreed much better with the avowed
intention of the Personnel Law if he had said, "Our cruise was
a great success because every officer was an engineer."
5 The Chairman of the House Naval Committee further said,
" It is the opinion of our naval officers, in command of our fleet
and ships, that this consolidation has been a splendid thing for the
navy, because it makes the man in command of the ship, the master
of the ship, a man who understands all the workings of the sliip.
Before, the command of the sliip was in the hands of the engineer.
We had to make a change in the curriculum of the Naval Academy
whereby the officer of midshipmen there must acquire a knowledge
of engineering, further adding to that the experience which he
must obtain in the engine room as a watch officer. By reason of
these facts, the entire ship is toda}^ under the command of an engi-
neer officer, a man who understands all the duties of engineering
and who is complete master of the ship."
6 I have been told by officers who have recently served on
board ship that one great benefit has resulted from the amalgama-
tion: namely, tliat the idea just expressed in the above quotation
is true; that the commanding officer is now the master of the entire
ship. In my early days, few commanding officers felt any interest
PRESENTATION OF PORTRAIT OF ADMIRAL MELVILLE 255
in the machinery beyond their demand that it should always be
ready for service. If anything went wrong, they washed their
hands of the responsibility, which was naturally upon the special
body of engineers. They now feel the same keen interest in the
machinery that they do in the guns or any other part of the ship,
and the chief engineer of the ship (still so-called) is generally looked
upon as the officer next in importance after the captain.
7 The part of the new regime about which I have felt misgivings
is that thus far there has been no systematic effort to assure train-
ing and experience for every line-officer in connection with the
motive power. Every young officer should be required to serve an
apprenticeship in the engine and fire-rooms, just as he does on
deck, but so far as I have been able to learn there has never been
careful attention given to this point.
8 Having touched upon the general conditions of the executive
side of engineering as affecting the operation and integrity of the
machinery at sea, it is now pertinent to consider the prospects with
regard to present and future designs. Thus far, this work has
remained in the hands of officers specially trained. Unfortunately,
the same condition is found here as mentioned in the previous divi-
sion of the subject. An effort has been made to arouse the interest
of some of the younger line-officers by a course of special training
for this most important work after the present highly competent
and experienced men have retired. There has, however, been no
settled policy for this, and the attempt that was started was inter-
rupted by the cruise of the Atlantic Fleet.
9 I am very glad indeed to bear testimony to the fact that
the recent designs of the Bureau of Steam Engineering have been
highly creditable in every way. In saying this, I feel a touch of
personal pride for the reason that the men who have been doing
this work were formerly my assistants and received most of their
experience during my term of office. I am naturally pleased that
the record which was made during my own term is being maintained.
10 When such praise as this can be given in simple truth, what
can be thought of the official who plans to discredit the men who
have made such a record, and destroy the autonomy of the Bureau
by subordinating it to the Bureau charged with the design of hulls?
I believe you will agree that m}'- service of a lifetime in the Navy
and my record as the head of a great Bureau in the Department,
the longest since the Civil War, entitles my opinion to some weight,
and I want to register my earnest conviction that any such scheme
of consolidation can only bring inefficiency, retrogression and waste.
256 PRESENTATION OF PORTIIAIT OF ADMIRAL MELVILLE
11 There is still another side to engineering in the navy; namely,
the work of the navy yards, which has been prominently before the
pubhc during the regime of the last Secretary of the Navy. Changes
have been made abolishing the separate departments in the navy
yards and consolidating their administration under one officer,
whose work, while a vital element in the building of a ship, was
certainly not the only important part, and moreover was so different
in its nature from the other departments which were absorbed,
that it is obvious he could not be an expert on these other lines.
To me it was so marvelous as to be almost beyond belief that in
this age of specialization a movement so absolutely counter to the
spirit of the age should take place in the name of economy and
reform. If the methods of great shipyards in civil life, or of the
great manufacturing establishments, or the dockyard administra-
tion of other countries, had differed from the methods employed
in our navy yards, a change would at least have been indicated;
if, in these places a system somewhat like the one which it has
been attempted to introduce in our navy yards, had been in vogue,
there could be some understanding of the change; the facts are
however, that in its essential featm'es our navy yard administration
was along the very lines which obtain in foreign dockyards, in the
great shipyards at home and abroad, and in our great manufactur-
ing estabhshments.
12 I am led to believe that the present Secretary is giving the
matter very careful consideration Avith a view to undoing the
tremendous harm brought about by his predecessor, and I trust
he will be well-advised and will restore the yards to their former
efficiency. It ought to be said, however, as a matter of record, that
these changes were made without any consultation between the
late Secretary and the officers most competent from long experience
to know what was best. Indeed, by his own statement, the scheme
was evolved from his own inner consciousness.
13 Our modern navy is essentially an engineering affair. The
vessels themselves are the product of the engineer's brain, and their
successful maintenance and utilization depend entirely on engineer-
ing skill. Ten years ago I said that the change which had been
made, absorbing the engineer corps into the line of the navy and
making every officer an engineer, was a tremendous step forward,
provided a sincere and earnest effort was made to carry out the
scheme which was thus outlined.
14 From what I have said this evening, it will be clear that
PRESENTATION OF PORTRAIT OF ADMIRAL MELVILLE 257
I am not as yet satisfied that this has been brought about. Undoubt-
edly the responsibility for the machinery of our vessels, guns, motive
machinery, electrical machinery, torpedoes, etc., is upon the line-
officer of the navy. He is charged with this duty by law. If the
older officers of the navy had taken hold of this matter with enthu-
siasm, I believe that it would now have been settled and there
would be no question as to the great success of the new officer, the
line-officer of the twentieth century. I am not willing to believe
(and indeed hardly willing to consider) the possibility that naval
officers will neglect any duty with which they are charged, and I
still hope that the scheme will be worked out to a great success.
15 Not much is ordinarily said about the machinists who are
doing good work on board our vessels. They look after the routine
work of repair and adjustment on board ship, but they are without
the scientific training which is required for engineers who are
really qualified for the duties comprehended by that title. If the
line-officers of the navy do not maintain engineering efficiency, it
will then, as the organization now stands, fall upon these machinists
to perform the work of the engineers. In other words, in an organi-
zation whose efficiency is absolutely dependent upon the skill of
engineers, the men relied upon for such work would be relegated to
a position of inferiority so low that they are hardly counted. This
is utterly un-American and can only be matched by absolute mon-
archies or countries as unprogressive in the mechanic arts as Spain.
The speedy destruction of her navy in the war of 1898 was due
to her utter incompetence even more in engineering than in gunnery.
It is inconceivable that self-respecting men in a free country like
ours will attempt to perform work of such vital importance with-
out adequate recognition in the way of rank and position.
16 I will not permit myself to believe that we shall have to
consider this as a practical • question^'because I cannot conceive
that naval officers would fail in their duty, but I feel that both sides
of the question, so vital to our naval efficiency, should be presented.
17 During my entire career in the navy, it was my constant
endeavor to show by my work the importance of the engineer and
to encourage that spirit in my subordinates. I trust I will not be
accused of vanity if I say that I believe my record as engineer-in-
chief added a little to the reputation of engineering. My active
work in the navy is done, but so long as I live my interest will never
slacken and my voice will always be raised to encourage efficiency
in every branch of the service.
258 PRESENTATION OF PORTRAIT OP ADMIRAL MELVILLE
PRESENTATION OF PORTRAIT
By Walter M. McFarlanb, Mem. Am. Soc. M. E.
The honor of being invited to pay a tribute to my dear old Chief,
Admiral Melville, is one which I appreciate highly, as well as the
allied one of acting as spokesman for the donors of the splendid
portrait which is to be presented to the National Gallery this evening.
I admire the Admiral as the fine flower of a splendid type of man-
hood, and his kindness to me for many years has been so like a
father's that with a son's affection I rejoice at this splendid testi-
monial to his personality and his work.
2 Too often the pathway to greatness and fame is marked by
the wreckage of competitors, and even friends, who have been ruth-
lessly thrust aside in the egoism of selfish ambition. Then, there
may be a grudging admission of ability, but there is no love, no true
admiration. When, on the other hand, the hero has always been
the helper and friend of his companions, when he has cheerfully
acknowledged their aid to his success, we have such greatness as
we are celebrating tonight. Then, every member of the profession
feels that the fame of the leader is reflected on the whole body, and
they love the man while they rejoice in his reputation.
3 George Wallace Melville is such a man. He has been one o/
the famous men of engineering so long that we find it hard to remem-
ber a time when his name was not synonymous, as it is now, with
all that represents progress and achievement in our profession.
4 It is a matter of delight to all of us who love him that the
artist, in the picture which is to be presented to the National
Gallery this evening, has faithfully depicted the chief character-
istics which have made him great. These are, in my judgment,
indomitable courage and unbending honesty. It is possible for a
man to have great mental ability and yet fail of true greatness
if he lack these essentials.
5 You all know Melville's Arctic record, which first brought
him an international reputation; there he displayed a heroic
courage which has never been surpassed, and for which Congress
advanced him a grade in the Navy. This, however, was only
one instance of the absolute fearlessness that began with his
earliest days in the service. When he became Engineer-in-Chief,
the same courage, but rather on the moral than the physical
side, was shown. Beginning with his first annual report, he
PRESENTATION OF PORTRAIT OF ADMIRAL MELVILLE 259
spoke out fearlessly, setting forth the truth as he saw it and
striving always for advancement and efficiency. Complaint was
made to President Cleveland of the plain speech in this first report,
but that strong man read it himself and said, " We want more
such men."
6 His professional courage is also remarkable, and, moreover,
a faculty, I believe which is characteristic of all great men,
having once made his decision he does not worry about the result.
Able men of minor rank are always fearful that something may
go wrong and their reputation be injured. The really big man
does not believe himself infallible. He knows that all men who do
things will make some mistakes, and he is strong enough not to
dread them. A notable instance of this kind in Melville's career
was his use of triple screws for the Columbia and Minneapolis. I
saw letters from some of his friends, for whose professional opinion
he had the highest regard, urging him not to make the experiment,
but he had studied the problem carefully, was satisfied with the
correctness of the solution, and persevered. The result was perhaps
the greatest triumph of his professional career.
7 His ability as an executive is of a very high order. The
feature of deciding a case and then refraining from worry is an
evidence. He had a rare talent for choosing able assistants, and
having proved them he left in their hands all the detail work,
thereby giving himself time for careful study of the 'larger problems.
The effect of this was very marked in stimulating the entire staff
to the highest efficiency and zeal. I have known them all personally,
and every man counted it a pleasure to work, without regard to
hours, for the credit of the "Chief" and the glory of the Service.
8 With respect to his professional work, it is notable that his
career as Engineer-in-Chief of the Navy, from 1887 to 1903, is the
longest on record. It covers the building of the "new navy," and
the Spanish war. During this time he was responsible for new designs
of machinery for about 120 vessels, among which were 24 battleships
and 41 armored vessels. Best of all, there were no "lame ducks,"
and no failures.
9 I will mention briefly some details of his more important work.
He was the first to use water-tube boilers in large war vessels and to
determine the actual coal consumption by trials. He was also the
first to use the method of determining trial-speeds, known as the
"standardized screw," which is the simplest, most accurate and
inexpensive, and fairest to the contractor as well as to the gov.ern-
ment.
^ 10 It is^to him also that we owe our first high-speed battleship.
When'* in 1898 the proposals for the Maine, Missouri and Ohio
260 PRESENTATION OF PORTRAIT OF ADMIRAL MELVILLE
were being prepared, he stood alone in his demand for 18-knot
ships. If he had not persisted, we should have been three years
longer behind the other navies of the world in battleship speed.
11 It is very interesting to note also that only a little after this
he proposed an "all-gun one-caliber ship," in other words, what is
now called the "Dreadnought" type. Before I left the Service, I
had often heard him talk of this big ship with ten twelve-inch or
twelve ten-inch rifles and nineteen or twenty knots speed; and about
1899 he submitted a sketch plan of the battery of such a ship to the
Board on Construction. Possibly the same influence which almost
prevented the eighteen-knot battleships prevented consideration of
this more advanced type. At all events Melville was in advance of
the general naval mind, and our country lost the credit which it might
have had for the introduction of this revolutionary improvement
several years before the Dreadnought was produced.
12 During the war with Spain he brought out the repair sliip
and the distilling ship. The idea of the former was not new, but
the Vulcan was by far the most complete vessel of the kind equipped
up to that time. The latter furnished fresh feed-water to the boilers
and enabled a vessel with a storage bunker capacity of 3000 tons
to supply 60,000 tons of water.
13 A clever piece of work at this time was the fitting of new
boilers to some of the old Civil War Monitors to enable them to be
used for harbor defense. For years Melville had advised the Navy
Department that new boilers must be supplied before these vessels
could be used. When the destruction of the Maine made the out-
break of hostilities seem probable, the makers of water-tube boilers
submitted estimates of time and cost for the work. Boilers were
promised in 30 days, but it was necessary to use the standard land
type. As these vessels were not to go to sea, however, this was
satisfactory. The worn-out boilers were cut up and passed out
through the smokepipe (because the armored deck could not be
taken up; the new boilers were passed down the smokepipe in sections
and erected on board; finally each of the boats was given a steam
trial, which was entirely successful.
14 A great deal of experimental work was done under his direc-
tion, all of which is published in his annual reports. The last of
such experiments was a series of tests of oil as fuel, probably the
most comprehensive ever made.
15 My brief sketch of this famous man would be incomplete if
I failed to speak of his personality. The lion-like head and the
frank speech have led some to say that he is one of the old " Vikings,"
PRESENTATION OF PORTRAIT OF ADMIRAL MELVILLE 201
spared to us a thousand years after the others have gone; but if
this leads any to think that he is harsh and cold, there could be
no greater mistake. Like all strong natures, he is pronounced in
his feelings, but he is a man of warm affection, and when he has
once taken you into his heart, you are sure of an abiding-place
there as long as you are worthy. It is often said that no man is
great to his intimates, but I have been with him, day by day, for
years; have seen him under all conditions; and my admiration and
love for him have simply increased as the years go by. I have no
ambition to be a Boswell and I have not kept notes of his doings;
but I have seen the daily workings of a great, kind heart, tender
for the humble yet fearless toward the great; and I can truly say
that I count it a privilege and an inspiration to have been a trusted
friend and helper of this noble man, who has exemplified the highest
type of manhood and added new luster to the profession of engineering.
ACCEPTANCE OF PORTRAIT
By De. C. D. Walcott
It gives me pleasure, speaking for the Smithsonian Institution as
the custodian of the National Collection of Art, to accept from you
for the people of this country this fine portrait of Rear-Admiral
George Wallace Melville, to be exhibited in the gallery of portraits
of Americans who have achieved eminence in their life work.
2 Among the men who have rendered distinguished service to
their country in literature, science, or art, in war or in peace, in
professional or civil life — few have won such well-merited distinc-
tion in so many lines of duty as Admiral Melville. He stands high
in the regard of the Nation as a naval hero, as an engineer of excep-
tional ability, and as a wise and resourceful administrator and
advisor. It is only to be regretted that, under the operation of
law governing retirement, Admiral Melville was obliged to retire
from active duty in 1903, but it is to be hoped that the country
which he has so efficiently and actively served may long be per-
mitted to enjoy the benefits of his counsel.
3 The portrait of Admiral Melville is a most appropriate addition
to this National collection and it is peculiarly fitting that his serv-
ices should be emphasized in this happy manner by a Society which
embraces so distinguished an array of men in the engineering pro-
fession, a Society that for nearly thirty years has exercised a power-
ful influence toward unity of interest and harmony of purpose in
the broad field of American engineering.
No. 1242.
SMALL STEAM TURBINES
By George A. Orrok, New York
Member of the Society
The papers upon steam turbines which have been presented before
the Society have dealt mainly with the larger types of apparatus and
have been written to show the reliability, efficiency and general
desirability of this type of prime mover.
2 This paper treats of the smaller sizes of steam turbines from
the standpoint of the designing and operating engineer, describing
the commercial machines in sufficient detail, with reference to the
service to which they have been applied, and giving certain facts con-
cerning their operation which may be of advantage to the engineer-
ing profession. Curves of steam consumption are given which show
in a general way what may be expected of these machines under cer-
tain conditions.
3 At the present time seven machines are on the market and can
be obtained in various sizes from 10 h.p. to 300 h.p. with reasonable
deliveries. These are the De Laval, Terry, Sturtevant, Bliss, Dake,
Curtis and Kerr turbines. Three other machines are nearly at this
stage of development and patents have been applied for on several
others.
4 Many thousand horsepower of these turbines have been sold
and are in successful commercial service. The following reports of
total sales of sizes from 10 h.p. to 300 h.p. have been obtained from
the manufacturers:
For further discussion of Steam Turbines, consult Transactions as follows:
Vol. 10, p. 680, Notes on Steam Turbines, J. B. Webb; vol. 17, p. 81, Steam Tur-
bmes, W. F. M. Goss; vol. 22, p. 170, Steam Turbines, R. H. Thurston; vol. 24, p.
999, Steam Turbines from the Operating Standpoint, F. A. Waldron; vol. 25, p.
1056, The De Laval Steam Turbine, E. S. Lea and E. Meden; vol. 25, p. 1041,
The Steam Turbine in Modern Engineering, W. L. R. Emmet; vol. 25, p. 782,
Different Applications of Steam Turbines, A. Rateau; vol. 25., p. 716, Some Theo-
retical and Practical Considerations in Steam Turbine Work, Francis Hodgkinson.
Presented at the Spring Meeting, Washington, May 1909, of The American
Society op Mechanical Engineers.
264
SMALL STEAM TURBINES
De Laval, De Laval Steam Turbine Company 70,000 h.p.
Curtis, General Electric Company 70,000 h.p.
Terry, Terry Steam Turbine Company 15,000 h.p.
Kerr, Kerr Turbine Company 10,000 h.p.
Sturtevant, B. F. Sturtevant Company
Bliss, E. W. Bliss Company
Dake, Dake- American Steam Turbine Co
5 All of these machines are of the impulse type: that is to say,
the steam is expanded in a nozzle and the kinetic energy of the jet is
absorbed by passing one or more times through the buckets of the
Fig. 1 High and Low-pressure De Laval Turbine
turbine rotor. In the De Laval turbine only one moving element
and one steam pass are used, which necessitates a very high bucket
velocity. In the Terry, Sturtevant, Bliss and Dake turbines a series
of return passages is provided. The steam returns two or more
times to the same rotor and the bucket speed is much lower. In the
Kerr turbine the steam is used in stages with one bucket wheel in a
stage; while in most of the Curtis machines two or three stages are
used with two or three rows of moving buckets, separated by station-
ary guide blades, in each stage. Compound machines of the other
types have been made but are not as yet produced commercially.
6 By far the larger number of these machines is used in connection
with extra high-speed electric generators, the next larger application
SMALL STEAM TURBINES
265
being to centrifugal fans for high pressures. Centrifugal pumps
adapted to high rotative speeds have been rather generally introduced
in the last few years and it is becoming usual to connect small turbines
direct to these machines. The small space required and the simplic-
ity obtainable in a 100-h.p. turbine at speeds of from 800 to 1200
r.p.m. have been important factors in their introduction.
7 The first of the small turbines to be put on the market was the
De Laval, made by the De Laval Steam Turbine Company of Trenton,
N. J., and introduced in this country about 1896. This machine is of the
Fig. 2 Terry Steam Turbine, 36-in.
pure impulse type, the steam being expanded in the nozzle down to
the exhaust pressure, and the resultant velocity transferred to the
wheel in one steam pass. The bucket speed is high, ranging from
600 to 1300 ft. per sec. Eight sizes of wheels are made, generat-
ing from 10 h.p. to 500 h.p., with one nozzle in the smallest size and
eight or more in the 500-h.p. size.
8 The high bucket speed necessitates the use of gears of special
construction, which have been very successful. The design, construc-
tion and economy of this type have been discussed in vol. 25 of
Transactions, p. 1056.
266
SMALL STEAM TURBINES
9 The Terry turbine, made by the Terry Steam Turbine Company
of Hartford, Conn., has been manufactured for about ten years,
although the commercial machine has been on the market only for
about four years. This turbine is of the impulse type, but the steam
passes through the buckets a number of times before its energy is
absorbed. The case of the machine is parted on a horizontal plane
through the shaft and at right angles to the wheel. The nozzles and
Fig. 3 Terry Turbine Showing Construction
return passes are bolted to the inside of both parts of the casing.
The nozzles are in the plane of the side of the wheel. The return
passages are of brass and are separated by partitions. The wheel
itself is built up of two steel discs held together by bolts over a steel
center. The buckets are built of steel punchings, fitting in grooves
cut in the discs, as shown by the figures. The sizes of wheels manu-
SMALL STEAM TURBINES
267
factured at the present time are 12, 18, 24, 36 and 48 in., and the
number of nozzles varies from two on the 12-in. wheel to eight or ten
on the 48-in. wheel.
10 The Sturtevant turbine, made by the B. F. Sturtevant Com-
pany of Hyde Park, Mass., has been in the development stage for
three or four years and quite a number of machines have been sold.
The present type of turbine may be called " standard, " however, and
four sizes of wheel are built, 20, 25, 30 and 36-in., developing from
Fig. 4 Sectional View of Terry Turbine
3 h.p. to 300 h.p. The turbine is of the multiple-pass type similar
to the Riedler-Stumpf. The casing is cast solid with one end. The
nozzle and return chamber ring are inserted from one side and the
wheel is milled from the solid. The return passages are from eight
to twelve in number and are milled on the inside of the return cham-
ber ring. They are partitioned and are similar in shape to the
buckets. The nozzle lies in the plane of the side of the wheel.
11 The Bliss turbine, formerly known as the American, made by
the E. W. Bliss Company of Brooklyn, N. Y., is of the same type as
268
SMALL STEAM TURBINES
Fig. 5 Wheel and Casing of Stttrtevant Ttjkbine
Fig. 6 Sturtevant Steam Turbine, 30-in.
SMALL STEAM TURBINES
269
the Terry and Sturtevant and has been on the market only a few
months. The casing and steam chamber are cast solid with one side
and the nozzle and return chambers bolted in. The wheel is milled
from a steel casting, or forging in the smaller sizes, and the partitions
Fig. 7 Section of Stxjrtevamt Turbine
separating the buckets are inserted and held in place by three bands
of steel shrunk on the face of the wheel. The return passages are
peculiar in having no partitions. Two sizes of wheel have been built,
the 42-in. and .30-in., but designs have been developed for the 12, 18,
24, 36, 48 and 60-in., covering powers from 10 h.p. to above 600 h.p^
270
SMALL STEAM TURBINES
Fig. 8 Bliss Turbine, 30-in.
Fig, 9 Dake Steam Turbine, 24-in.
SMALL STEAM TURBINES
271
12 The Dake turbine, made by the Dake-American Steam Tur-
bine Company of Grand Rapids, Mich., is a single-stage impulse tur-
bine. The wheel is made of two bucket discs, with milled buckets
and inserted partitions, bolted together over a wheel center. In their
Headlight turbine the governor is enclosed between the sides of the
wheel. The nozzles and return-passages are placed between the
bucket discs. The machine is built in sizes of from 5 h.p. to 100 h.p.,
the diameter of the smallest wheel being 12 in.
Fig. 10 Parts of the Bliss Turbine
13 Coincident with the development of the large Curtis turbines>
the General Electric Company, at their Lynn Works, have developed
and placed on the market a line of small generating sets ranging from
5 kw. to 300 kw. This range is covered by eight sizes, the smaller
machines being single-stage with two or three passes per stage. The
buckets and nozzles are of the well-known Curtis type.
14 The Kerr Turbine, made by the Kerr Turbine Company of
Wellsville, N. Y., is of the compound impulse type. It is generally
built in from two to eight stages. The buckets are of the double
272
SMALL STEAM TURBINES
Pelton type, inserted like saw teeth in the wheel disc. Five sizes of
of wheels, 12, 18. 24, 30 and 36-in., are made and cover a range of from
10 h.p. to 300 h.p. The nozzles are in the plane of revolution of the
Fig. 11 Sectional View of Buss Turbine
wheel and are screwed into the stage partitions and held in place by
a lock nut.
15 As in large turbines, details of these small turbines, to which
reference has been made, show the skill and knowledge of the designer,
and that the same problem may be solved in different ways is well
illustrated by the sections here reproduced.
SMALL STEAM TURBINES
DESCRIPTION OF DETAILS
273
16 Nozzles. The diverging nozzle is used by all makers except
Kerr, whose multi-stage wheel requires a converging nozzle. In the
De Laval, Sturtevant and Kerr turbines, the nozzles are screwed into
their seats; that of the Terry is held in place by a bolt. The nozzles
Fig. 12 Section of Dakb Headlight Tuhbine; Exterior Shown in Fig. 9
of the Curtis, Dake and Bliss turbines are reamed out of the solid.
The larger sizes of the De Laval machine which have been put on the
market lately have a large number of reamed nozzles instead of the
older construction.
17 Buckets. The constructions employed in the Curtis and De
Laval wheels are well known and have been described many times.
274
SMALL STEAM TURBINES
Fig. 13 Curtis Turbine 50 h.p.
Fig. 14 Curtis Turbine in Process of Assembly
SMALL STEAM TURBINES
275
The Terry, Dake, Bliss and Sturtevant buckets are practically semi-
circular in form. The Terry bucket is constructed entirely of steel
punchings assembled between grooves in the two steel discs forming
the sides of the wheel. The Sturtevant wheel is milled out of a steel
casting. The Bliss buckets are milled out, but the partitions are
Fig. 15 Section of Cubtis two-stage, Non-Condensing Turbine, 160 h.p.
inserted and held in place and steel rings are shrunk on. The Dake
buckets are turned out of the solid, the recesses for the partitions
milled out and the partitions inserted; the wheel is then bolted
together. The Kerr buckets are very similar to the original Pelfon
buckets and are inserted in the wheel in a manner similar to the
De Laval buckets.
276
SMALL STEAM TURBINES
Fig. 16 Curtis Turbine, 200 h.p.
Fig. 17 Revolving Element of Curtis Turbine in Bearings
SMALL STEAM TURBINES
277
278
SMALL STEAM TURBINES
1 i
L
^^^7> ^^1^1
K
J^JSSk - "^^^^1
^v
iLdM^BS i^^^^HHI-i '^^^^^^^^^^^1
^^
^■.^
Fig. 19 Kekk Turbine, 18-in,
18 Return Chambers. The Sturtevant returns are milled out of
the solid ring. Bliss casts them in the nozzle piece and finishes them
by hand; Terry casts each one separately, finishes by hand and assem-
bles with bolts; Dake casts the return chambers solid, mills the pas-
sages and covers them with a shrouding.
Fig. 20 Complete Rotating Part, 18-in., 7-Stage, Kerr Steam Turbine
19 Wheel Centers. De Laval, Curtis, Sturtevant and Bliss make
the wheel centers of steel castings or forgings integral with the wheel.
Terry uses a steel casting but bolts the wheel disc to it. Kerr uses
a screwed coupling, the inner part cut in three pieces and keyed to
SMALL STEAM TURBINES
279
280
SMALL STEAM TURBINES
the shaft with round keys, clamping the wheel disc. Dake's wheel
centers are an integral part of the wheel in small sizes, but in the
larger machines are steel castings, in some cases a part of the shaft.
20 Governors. De Laval, Terry, Sturtevant, Bliss, Dake and
Kerr use a fiyball governor on the shaft end, which actuates the
throttle valve through a system of levers. Curtis uses the fiyball
governor on the shaft for small sizes and slower-speed spring-con-
trolled governors of different forms for the larger sizes. The Sturte-
vant, Bliss and Curtis machines are provided with an emergency stop
governor as well as the throttling governor.
Fig. 22 One Stage op Kerr Turbine, Showing Nozzles and Wheel
21 Glands. For non-condensing machines glands are not trouble-
some, as the difference of pressure between the casing and atmosphere
is rarely more than a few pounds. Terry uses a bronze ball-and-
socket gland with a long loose fit on the shaft. Sturtevant and Dake
use a set of ring packing, either cast-iron or bronze. Bliss has a laby-
rinth packing without contact. Kerr has a floating bronze bush with
soft packing behind it. Curtis uses a metallic packing held in place
by a. gland ring, and for condensing service a carbon-ring packing,
steam-sealed.
22 Clearance. In none of these machines is clearance an impor-
tant factor. The clearance between buckets and guide passages on
A 24-in. wheel is usually from ^ in. to ^ in. when hot. Striking or
rubbing is practically unknown.
SMALL STEAM TURBINES
281
23 Thrust. Theoretically, there should be no thrust in any
turbine of these types. Practically, there is always a very small
jy
DE LAVAL
BLISS
Fig. 23 Typical Turbine Buckets
thrust one way or the other. This thrust is usually taken care of by
small thrust collars or washers next to the bearings. Thrust from
282
SMALL STEAM TURBINES
the outside is prevented by the use of a flexible coupling between the
turbine and the machine it drives.
24 Bearings. The bearings are always ring-oiled with large oil
reservoirs, sometimes, on the larger sizes, provided with water jackets
Fig. 24 Section op Wilkinson Steam Turbine, 20-in.
or water cooling pipes for an emergency cold-water circulation. The
lubrication of the thrust is obtained at the same time.
25 Operation. These machines are nearly automatic in their
operation. When the machine is once properly set, the coupling
properly adjusted and the bearings supplied with oil. the machine may
SMALL STEAM TURBINES
283
run for years without an overhauling. The bearings must be looked
after to see that no heating takes place and that the ring is carrying
the oil to the shaft. The coupling should be examined from time to
time to make sure that no thrust is communicated through it to the
turbine. With these precautions a three months' continuous run is
common and a number of turbines have to my knowledge run more
than eighteen months without a cent spent on them for maintenance.
Apparently there is no wear in nozzles, buckets, or return chambers
\s^
y^ ^
i - ^^ -y^
ri^-^-s \->z,u3
y\^ '
>'ater Rat^
.^y.
'*
.<^
-y^^
'^^0^^'
,5.^
5^'^
,'->
.'
^
1500 I
30 40 50
Brake Horse Power
Fig. 25 Steaai Consumption Curvks, Terry Turbine
24-IN. WHEEL, 150-LB. PEE88UKE, NO SUPERHEAT, NON-CONDENSING. TESTED BY WSSTINOHOT7BB
MACHINE CO., PITTSBUKa, PA.
The only wearing parts are the bearings and these are generously
proportioned.
26 These machines may be taken apart and reassembled in half
a day; some of them in two hours. The over-hung machines may be
overhauled in an even shorter time.
27 'New Turbines. The Hachenberg turbine, made by Wm.
Gardam & Son, New York, is a compound impulse turbine resem-
bling in construction the Dow turbine so frequently illustrated twenty
years ago. Some experimental machines have been built, one of
which was tested at Columbia University, and the commercial ma-
chine will soon be on the market.
28 James Wilkinson, of Providence, R. I., has a small steam
turbine nearly in the commercial stage. A number of these machines
are running, and witliin the next few months it is expected they will
be on the market.
284
SMALL STEAM TURBINES
10 15
Brake Horse Power
Fig. 26 Steam Consumption Curves, Sturtevant Turbine
20-IN. WHEEL, SINGLE-STAGE, NON-CONDENSING, 2400 R.P.M.]
29 The Church turbine, lately completed by the Watson-Still-
man Company and tested at Stevens Institute, is another promising
turbine.
T-5000-, 10000
lOOOi 2000
Brake Horse Power
Fig. 27 Steam Consumption Curves, Bliss Turbine, Non-Condensing
TESTED BY F. L. PRYOR AT HOBOKEN, N. f.
O = Two-nozzle, X = Four-nozzle
STEAM ECONOMY
30 The curves of steam economy have in most cases been obtained
from the manufacturers. For the Curtis turbine speed-economy
SMALL STEAM TURBINES
285
10
20 30 40 50
Brake Horse Power
60
Fig. 28 Steam Consumption Curves, 50 h.p. Curtis Turbine
ONE-PUESSDRE-STAOE, THREE ROWS OF BUCKETS, 25|-IN. WHEEL, CURVES CORRECTED TO
150-I.B. BOILER PRESSURE, NO SUPERHEAT, ATMOSPHERIC EXHAUST
curves are given for the 50 h.p. and 200 h.p. sizes. These curves
represent the average of a large number of tests and have been cor-
rected to bring them to standard conditions. The averages were
consistent, and the variation from the average in any case was not
large.
ICO SCO
Turbine 3rat3 Horse Power
Fig. 29 Steam Consumption Curves, 200-h.p. Curtis Turbine
THREE-STAGE, 36-IN. WHEEL, CORRECTED TO 165-LB. ABS. BOILER PRESSURE, NO SUPERHEAT,
NON-CONDENSING
31 The curves for the Terry turbine were plotted from fourteen
tests made at East Pittsburg by the Westinghouse Machine Company.
The curves for the Bliss turbine were plotted from twenty- four tests
286
SMALL STEAM TURBINES
made at Stevens Institute by Prof. F. L. Pryor. The curves for the
Kerr turbine were plotted from tests made by the Kerr Turbine Com-
pany in their testing plant at Wellsville, N. Y.
100 150
Brake Horse Power
Fig. 30 Steam Consumption Curves, 24-in. Kerr Turbine
SIX-STAGE, CONDENSING, VARYING VACUUM, 70-LB. GAGE PRESSURE
7000
20 40 60 80 100 120 liO 160 180 230
Brake Horse Power
Fig. 31 Load Curves of Kerr Turbine
24-IN. WHEEL, 8-STAGE 175-LB. GAGE PRESSURE, NON-CONDEN8IVQ
32 There seems to be no change in steam economy use. It
may be too early to make this statement, but machines running
regularly for three years have shown no increase in steam consump-
tion.
33 The field of the small steam turbine is somewhat narrow when
SMALL STEAM TURBINES 287
compared with the high-speed steam engine. The small turbine has
its place, however, and with the development of a more economical
machine at the lower speed ranges, will have a much wider field. The
turbine-driven centrifugal fan, for both high and low pressures, will
have an increasing use, and the turbine-driven centrifugal pumps
have marked advantages over reciprocating apparatus because of the
absence of shock on the pipe line and their adaptation to space
conditions.
34 The promise of development on these lines has led many manu-
facturers to enter the small-turbine field and the great expansion of
the large-turbine business without doubt presages a like future for
the small steam turbine.
DISCUSSION
Charles B. Rearick. Small turbines are being used extensively
for hot-well service for surface condensers, the turbine driving a
centrifugal pump which carries the condensed water away from the
condenser, supplanting in these cases the usual reciprocating pump.
They are more efficient than the reciprocating pump, they take less
space, there are fewer parts to maintain, and the service seems to
be very popular. They have also proved their worth in larger
installations for driving boiler-feed pumps of the multistage turbine
type.
2 Some of the newest work taken up by turbine drive is for circu-
lating pumps for surface and jet condenser work. We have a represent-
ative lot of such installations using turbines from 50-h.p. to 250-h.p.,
operating such pumps under low heads and at speeds as low as 600
r.p.m., direct-coupled to the rotor shaft of the turbines.
3 The question of economy is not touched upon to any great
extent in Mr. Orrok's paper, except to give some curves which cover
only one condition of service in most cases and cannot very fairly
be compared for the different makes. High economy in small tur-
bine units is in many instances of minor importance. Reliability of
service is most important of all. Under the very low speeds for
The paper on Small Steam Turbines was discussed both at the Washington
meeting, May 4-7, and at the Boston meeting, June 11, 1909. The discussion
is here given in abstract only, eliminating the matter presented at the Boston
meeting which duplicated that given at the Washington meeting. The com-
plete discussion was published in the September 1909 issue of The Journal.
288 DISCUSSION
driving circulating pumps and similar pump work the economy can-
not be especially good; but in nearly all these large power plants
the exhaust steam is all utilized in feed-water heaters, and approxi-
mately 80 per cent of the heat is returned to the boilers.
4 There is only one class of service in which high economy is
absolutely necessary, and that is, when the unit becomes the prime
mover or the main unit for a plant. In that case the turbine, both
condensing and non-condensing, compares well with the engine; for
such work is usually driving dynamos and other high-speed apparatus
and the speed can be chosen for the best economy.
W. D. Forbes. Mr. Francis B. Stevens of Hoboken, who died a
year ago in his ninety-fifth year, seeing in my shop some small Pelton
water wheels, told me that in 1854 he had seen a steam turbine in a
candy establishment in New York, which ran 1200 revolutions for
some twelve years, with little or no attention. It drove a small
fan. Mr. Stevens described the machine as practically the same as a
Pelton water wheel, except that the bucket was cut in two, each half
being placed on a separate disc, and the steam was led to these two
buckets by a " splitter" between them, which of course was stationary.
Each bucket was made fast to the disc, which was of course keyed to
the shaft.
2 What seems strange to me is, if steam turbines are such excellent
things and have been known so long, that they are not more generally
used.
Richard H. Rice. The author describes the construction of
seven different turbines, which may be divided into four classes ac-
cording to the method of using steam, as follows:
a Single-stage, expanding nozzle, one bucket row, one velocity
extraction: De Laval.
b Multistage, conveying nozzle, one bucket row per stage,
one velocity extraction per stage: Kerr.
c Single-stage, expanding nozzle, one bucket row, multiple-
velocity extraction: Terry, Sturtevant, Bliss, Dake.
d Single or multistage (depending on capacity), two to three
bucket rows, mutiple velocity extraction but only one per
bucket row: Curtis.
2 The value of these various methods of using steam is clearly
set forth in the curves presented by the author (Fig. 25 to Fig. 30),
SMALL STEAM TURBINES
289
giving the water (steam) consumption of the various turbines in
Classes b, c and d. In the^diagram in Fig. 1 all these curves are drawn
to the same scale so that they may be readily compared.
3 It will be seen that "with one exception, a very small machine,
which suffers somewhat from this fact, the water-rates of all the tur-
bines in Classes b and c fall rather closely together, while Class d^
Citrvc
A
B
C
c'
D
E
E'
Type R.P.M. Rated H.P.
Sturte^^tult 2,400 20
Terry 2,350 60
Bliss 2,000 :00
2,600 200
IveiT 2,800 150
Curtis 3,000 50
2,000 200
Steam Press.- 150 Lb.
Dry Steam
Atmospheric Exhaust
80
70
f-l
w
A
^
^^
^
Pi
K 60
w
^
B
C
— G-
D
\
^
A
i 50
f
^
"v.
^-
^
^
::::::;:;
1
I40
1
E
"-^
^
-^
B
C
^
D,
c'
30
e'
E
e'
20
i
i
i
4
i
^.
i
Load
Fio, 1 Economy Curves op Small Turbines
even with a 50-h.p. machine, shows considerably better results. This
is much more jaarked in the case of the 200-h.p. machine. Class b
is represented by an eight-stage machine'of 150-h.p. and shows slightly
better results than the machines of Class'c, particularly at light loads.
This result is insignificant as compared with the complication of the
large number of wheels, diaphragms, packings, and length of machine,
290
DISCUSSION
and this complication is therefore apparently not justified. The rea-
son such multistage machines do not give better waterratesisdue to
a considerable extent to the high frictional losses caused by rotating
wheels in a dense atmosphere of steam at comparatively high pres-
sure.
C2
60
58
56
54
L
\
^ Keciprocatiug Engine Tests
A Sturtevant Turbine
X Teir>- Turbine
Q Bliss
X Kerr "
® Curtis "
Pi
n
|48
§ 46
o
1 44
P
|48
140
38
36
34
32
30
\/
L
]
(§)e
f
^
c
]
(f
5)3
4
m
^(^
®
d
lA
i
0
!
1
20 40 60 80 :00 120 140 160 180 200 230
Rated Load B.H.P.
Fig. 2 Comparison op Economy op Reciprocating Engines
AND Turbines
4 The turbines of Class c labor under two disadvantages, due to
using the steam repeatedly in the same Ijuckets; either (as in the case
of the Terry turbine) the steam has to be turned at high velocity
through an angle of 180 deg. in the return chambers, or it has to be
SMALL STEAM TURBINES 291
used in buckets which have in general the wrong angle of entrance;
for it is easy to see that if the bucket angle is correct for receiving
the jet at its highest speed, it cannot be correct when the jet has con-
siderably slowed down.
5 The general principle employed in these turbines was used first
by Professors Riedler and Stunipf in the years 1902-1903, and a
number of these machines were built and put into service. The com-
pany which exploited them, however^ abandoned the principle about
the latter year and since then has built under a Curtis license, this
step having been taken by reason ofthe superior economies obtainable
by the Curtis construction, which the paper seems fully to confirm.
6 At the Detroit meeting, June 1908, Messrs. Dean and Wood
presented a paper giving results of tests on high-speed engines of sizes
comparable with the turbine figures given by Mr. Orrok. Fig. 2
shows the results compared with the full-load water rates given by
the author.
7 Messrs. Dean and Wood tested engines which had been in ser-
vice for some time and many of which had evidently seriously deteri-
orated in efficiency due to wear and leakage. Mr. Orrok confirms
the statement made b}^ the writer, that the turbine does not fall off
in efficiency after similar length of service and in further confirma-
tion of this is the test on a 75-kw. Curtis turbine made by Professor
Carpenter.
8 This fact has also been established by many tests made by the
writer on turbines which have been in use for considerable lengths of
time, and is subject only to the qualification that when steel bucket
constructions are used, as in all the turbines described except the
Curtis, wear may be expected under certain conditions of wet steam
and light loads which will increase steam consumption after a very
moderate length of service. The use of bronze buckets of the proper
composition to resist this deterioration is therefore essential to secure
the best results under all conditions.
Prof. R. C. Carpenter. During the past year I have given con-
siderable study to the results from the use of small turbines, arriving
at practically the conclusion of the author (Par. 33), that the field of
the small turbine is somewhat narrow as compared with the high-
speed steam engine. This conclusion applies to small turbines run-
ning non-condensing, however; when large turbines are operated con-
densing the economy is very high, and I think the results will usually
be superior to those obtained with piston engines.
292
DISCUSSION
2 On testing one turbine, which I think had been in use for three
years, I was pleased to find that my results practically agreed with
those obtained when the turbine was first installed. This seems tf'
indicate that so far as that turbine is concerned, there was no deteri -
oration from use. The general results which I obtained in the econ-
omy tests were substantially those shown on these curves, andindica1(j
that the econom}' is not good compared with the piston engine; the
advantages of the small steam turbine must therefore be other than
simply that of economy. The results of the tests of this machine aie
shown in the table.
3 I have recently had an opportunity of getting figures from a
small turbine operated with a high degree of superheat and running
TEST OF CURTIS TURBINE, 75-KW. CAPACITY, OPERATED NON-CONDENSlxNG
October 13, 1908
Test No 1
Electric load, kilowatts ' 57 . 7
Pressureat throttle, pounds gage 121.
Pressure at nozzle, pounds gage 108.4
Back pressure, pounds gage 0.24
Barometer reading, inches 30 .0
Total water to boiler, pounds 14971
Wet steam to turbine, pounds 14121
Quality of steam, per cent 98 . 3
Dry steam to turbine, pounds 13881
Dry steam to turbine per hour, pounds 3085
Dry steam to turbine per kw-hr. , pounds 53 . 5
Equivalent peri. h.p. (provided 1 kw. = 1.6 i.h.p.) i 33.5
Note — The pressure at the throttle is practically the same as at the boiler, which stood about
80 ft. away.
non-condensing. The results of that test were satisfactory in many
ways: 350 deg. superheat seemed to have aboutthe same effect as 18
in. of vacuum, and a machine having a water rate given as approxi-
mately 50 lb. per b.h.p. went down to about 22 lb. per l).h p. The
small steam turbine has special advantages for many kinds of work
where a high rotative speed and small torque are desirable; for those
kinds of work I believe it will ultimately supersede the small piston
engine.
H. Y. Haden. a field for small turbines not touched by this excel-
lent paper is that of installations where exhaust steam can be utilized
to advantage. There is an installation in Pittsburg of a 150-h.p.
turbine connected to centrifugal pumps, which operates under very
SMALL STEAM TURBINES
293
unusual conditions. Primarily, it takes the exhaust of reciprocating
pumps, without any regenerator, and develops the full power when
exhausting into a vacuum of 25 in., but it is also capable of automatic-
ally talcing high-pressure steam at 125-lb. pressure, should the supply
of exhaust steam fluctuate too much or be entirely cut off. The same
turbine also operates taking steam at 125-lb. pressure and exhausting
into the atmosphere, and it had further to be guaranteed by the manu-
facturers to take high-pressure steam at 65-lb. pressure when exhaust-
ing freely. I believe the above four conditions could not be met by
any reciprocating engine. Of course maximum economy will not be
attained under each, but it is attained under two: that of exhaust-
pressure condensing and high-pressure condensing.
2 The turbine is of the De Laval type, which is particularly adapt-
able for changing steam conditions, and is the only machine in which
any desired plan of operation can be attained by simply changing nozzle
ratios, without any change in the angle of the bucket or velocity of
the pump, and without sacrificing either capacity or efficiency. I
believe there is a large field for turbines operating under conditions
such as the foregoing, whether connected to generators or centri-
fugal pumps — a field where one can depend upon a unit irrespective
of the supply of exhaust steam and without a regenerator or other
expensive auxiliary.
Fred. D. Herbert. Mr. Orrok did not mention tests made on the
Terry turbine at the New York Edison Company's plant some years
ago, in which the steam consumption is much below that shown in
the Terry curve (Fig. 25). The accompanying curve and tables
show the water rates of 50-h.p., 25-h.p. and 12-h.p. Terry turbines
respectively.
TABLE 1 TESTS OF A 25-H.P. TERRY TURBINE
Test
Steam
Back
Quality or
Speed
Load
Per Cent
\ 1
ofTotal WaterWater per
No.
Pres.
Pres.
Superheat
R.P.M.
B.H.P.
Rating
PER HR.
H.P.-HR.
Degrees
1
90
0
62.50
2500
25.52
102.0
1 1068
41.85
5
90
0
71.70
2500
18.60
74.5
851
45.72
2
90
0
82.42
2500
12.39
49.5
680.5
54.93
10
90
0
50.00
2100
24.29
97.2
1068
44.00
11
90
0
60.26
2100
15.816
63.4
808.25
51.09
6
90
0
45.60
1800
22.90
92.0
1077.25
47.10
7
90
0
79.86
1800
10.21
40.8
638.58
62.54
9
90
0
44.70
1800
15.69
62.4
807
1 51.76
294
DISCUSSION
TABLE 2 TESTS OF A
12-H.P
. TERRY TURBINE
Test
Steam
Back Quality or
Speed
Load
Per Cent oi
Total Water Water per
No.
Pres.
Pres. Superheat
R.P.M.
B.H.P.
Rating
PER HR.
H. P.-HR.
Degrees
1
136.5
0 ! 58
2501
13.65
113.6
577.3
42.0
2
136.8
0 15
2513 1
12.08
100.7
547.2
, 45.3
3
128.5
0 12
2490
11.70
95.6
545.2
46.6
4
130.9
0 0
2501
7.23
60.2
421.5
68.3
X 3600
SO •=
1
-+-1
^
-^
^^>^
^
A
K0-
^
^
Back Pressure .68 ^
Superheat 60°
-
\
N
J
R
P,M. 2500
1
K>'
.J
s
^1
^
\
^
V
\
\
k
'
>
'^0,
1^
J
1
^
P
>J
u
~-^
P^
--
-
"■
— —
__
Fig. 1 Curves of 90 h. p. Terry Tuhbine
2 Regarding the statement that turbines are built of the single
stage only, the Terry Company has in operation and under construc-
tion several two-stage condensing turbines, the largest of which is 600
h.p. and the smallest about 125-h.p. Because of the successful re-
sults obtained, the builders claim that the turbine is superior to the
reciprocating engine as an operating proposition, and in none of these
turbines now running has a bucket been replaced.
W. E. Snyder. Almost all the emphasis has been laid on steam
economy. Another point which should receive careful consideration
is the lower cost of maintenance, particularly where turbines are used
for boiler feed, replacing the direct-acting boiler-feed pumps ordinarily
used; for pumping circulating water to condensers; or for driving cen-
trifugal pumps pumping water at comparatively low heads. In all
this work the turbine replaces direct-acting pumps which are very
SMALL STEAM TURBINES 295
expensive to operate, not only from the standpoint of steam-consump-
tion, but from the cost of supi^hes, such as cyUnder oil, pacldng and
pump valves. I have in mind one central condensing plant served by
a direct-acting pump where the costs are from $75 to SlOO a month for
packing, cylinder oil and valves. A similar condensing plant served by
a small turbine involved practically no expense for these supplies.
2 The first question, three or four years ago, before our company
had installed any of these machines, was not a question of steam
economy so much as of reliability of operation. If we put in a tur-
bine to pump water for a central condensing plant, into which exhaust
a large number of steam engines of various kinds used for varied
service, would the turbine break down, just when it was most needed?
For the purpose of investigating this point I went to a plant which
had a small unit in operation driving a generator and which had been
in use for about four years. The gears showed no appreciable wear,
and there had been practically no shutdowns.
3 The result of that investigation was the adoption of turbines in
central condensing plants in a number of works with which I have
to do, and also later for boiler feed and for pumping water under low
heads; and they have proved generally reUable regardless of make.
This is also true in regard to small turbine air-compressor units for
the cupola, etc., in steel works.
4 I think it is, therefore, in the displacement of the direct-acting
pump, always expensive to operate apart from steam consumption,
that the small tiu-bine will find its greatest field of usefulness. In the
Waterside Station in New York, the turbine-driven boiler feed
pumps have been continuously operating for months, running almost
automatically, and requiring practically no attention or supplies.
Steam economies are of course important, yet in large plants where all
of the large units are condensing, the steam from the auxiliaries is
needed to heat the feed-water, and a few per cent more or less in
steam consumption of turbine auxiliaries does not materially change
the total economy of the plant.
5 Other advantages in favor of the small turbine as compared
to direct-acting pumps, are the small'space required and the fact that
they can always be kept clean and present a good appearance.
Direct-acting pumps are usually'very difficult to keep in presentable
condition on account of water leakage and of the excessive use of lubri-
cants. The turbine and pump are entirely enclosed, the case can be
wiped very conveniently, and it presents nothing of the unsightly
appearance which is so often characteristic of the direct-acting
pump.
296 DISCUSSION
F. H, Ball. The conclusions of the author regarding the future
of the small steam turbine may fairly be questioned. On the score
of efficiency the showing made by the several types, even when tested
by the parties interested in their success, is very poor. From these
test figures, it appears that the best performance ranges from about
30 lb. per h.p. per hr. to nearly 70 lb.
2 It must be noted also that very high steam pressure is generally
used, and in some cases superheat. Under these conditions any
good reciprocating engine, even of the single-valve type, run as a non-
condensing compound, would easily develop power on 20 lb. or less:
the user of one of the non-condensing turbines described must
therefore expect to increase his coal bill from 50 to 200 per cent over
a single-valve non-condensing compound engine having the simplest
possible form of valve gear.
3 Moreover, the speed of these turbines, from 2000 to3600r.p.m.,
will generally be considered objectionably liigh. Buyers of electric
motors generally prefer motors of moderate speed, even at the extra
cost, and generally speeds above 1000 r.p.m., even for motors as small
as 10 h.p., are considered objectionable. This same objection must
inevitably be urged against speeds of 2000 to 3600 for engines of con-
siderable power.
CuAS. A. Howard, As far as a comparison of the merits of differ-
ent turbines goes, it must be remembered that the economy is affected
by the bucket speed even more than by steam pressure. In all of
these tests the bucket speeds are different, and any attempt to make a
comparison of the steam economy, as in the diagram by Mr. Rice,
would thus lead only to an erroneous view. While his curves
show in general what may be expected from turbines of this size,
no correct comparisons can be drawn between individual machines.
W. J. A. London.* With reference to Richard H. Rice's compari-
son of Fig. 25 and Fig. 28, showing the steam consumption of the
Terry turbine and that of the Curtis turbine, if the curve of the
Curtis turbine be produced and the peripheral speed of the two
types be made the same, the curve of the Terry turbine will cross
that of the Curtis type at about 1950 r.p.m. See Fig. 1. There is
therefore not much room for discussion of the difference of efficiency
of the two types. Moreover, with a large turbine an increase of a
pound on the steam consumption would increase the cost bill from
* Terry Steam Turbine Co., Hartford, Conn.
SMALL STEAM TURBINES
297
$2000 to $4000 a year, but with a small turbine it would mean an
increase of only from $10 to $25 a year, which would be offset by the
difference in first cost.
2 The greatest value of the paper lies in the fact that, better than a
salesman, it shows to men having reciprocating engines, the great
simplicity of construction of the turbine.- Men famihar with recipro-
cating engines know what to do in case of breakdown, but with a
50
i8
IC
U
12
40
38
3G
3i
32
\
X
\
N
o Indicates Curtis Turbine
*■ " Terry "
^
\
\
^
N
\
k
\
^
k.
"--
-^
" —
30
Carve A 17C
1 1
^
17 18 19 2W0 21 22 23 21
20 27 28 39 3000 31 32 33 31 35 36
Fig. 1. Comparison op Tests in Fig. 25 and Fig. 28 Reduced to the
Same Wheel Velocity.
turbine a failure means a shutdown for several days. In a few years,
however, every engineer will thoroughly understand the construction
of a steam turbine and will be able to make his own repairs. Another
fact which hindered the more frequent use of the steam turbine was
that generators, pumps, blowers and other machines had to be
designed specially for operation with turbines. That this is now
being done is an acknowledgment that the steam turbine is here to
stay.
Chas. B. Rearick. The matter of speed for turbines driving
centrifugal pumps is often a compromise, as the pump speeds
are not always ideal for the turbine. Especially is this true for circu-
lating work where the heads are often only 15 or 20 ft. and the deUvery
20,000 gal. or more per min. It may be necessary to sacrifice some
efficiency of the pump in order to run at a speed suited to the turbine.
In such instances all the exhaust steam may be used to advantage
in heating feed water, while the low cost of^operation and the saving
in oil and supplies will overcome the cost of increased steam con-
298 DISCUSSION
sumption. When efficiency is of importance, as in isolated dynamo
work, the speeds are usually such as to give results quite as good
as those of the reciprocating engine, and in many cases better.
The turbine has demonstrated its ability to give economical results
in all cases where the speeds ire favorable.
2 Regarding the rating of turbines, there is only one point of
maximum efficiency of any design, so far as I loiow, and that is its
maximum load. It is very similar to a gas engine in that respect.
If we want maximum efficiency the turbine must carry its maximum
load. That can be brought about in some cases by the nozzle system
of design in which better economy is obtained at light loads by shut-
ting off nozzles. But where t"bis is done by hand regulation there is
always danger of the load coming on without notice or without the
engineer having opened any of the hand adjustments. Turbines
should therefore be designed to eliminate hand regulation, and to
accomplish this some builders provide for automatic operation of
these valves. While this is successful within a limited degree, these
valves may become leaky in service, and when once leaky the control
of the turbine is beyond the operator, which may result in over-speed-
ing to a bursting point in case the load is suddenly thrown off.
3 It follows that the number of controlling valves should be
reduced to a minimum. If there is only one valve to control in a
machine there is but one valve to look after and to keep tight. Few
operators of turbines appreciate how serious the leakage of the con-
trolling valve is to the proper governing of the turbine, especially on
very light loads or no load. On the other hand, if turbines are kept
well loaded these leaky valves are not noticed and as a result steam
consumption is increased through their use rather than diminished,
unless all the valves are wide open and the turbine is working up to
its full capacity. The turbine which eliminates these dangers, it
would seem, is the better machine, and in small units the difference
in economy is entirely outweighed by the complications and dangers
above cited.
F. B. DowsT. The B. F. Sturtevant Company have for years built
reciprocating engines — single engines, multiple-single engines, and
multistage engines. Later we built direct-current and alternating-
current motors. We first came into touch with the steam turbine
principle in 1883 when our attention was first called to the Wise steam
motor. This motor was an impulse wheel, with four jets, I think,
the steam impinging on buckets, no endeavor being made to expand
SMALL STEAM TURBINES 299
the steam in nozzles. One of our engineers left us at that time to
exploit the Wise steam motor. He returned after a year's sad experi-
ence in connection with the amount of steam that would flow through
a small opening.
2 Our next experience was with the Dow steam motor. Mr.
Dow came to us early in the nineties, I think it was, backed by Mr.
Chisholm of the Chisholm Shovel Works of Cleveland, O. This tur-
bine had previously been developed and used successfully to drive
the flywheel in the Howell torpedo. Associated with Mr. Dow was
Mr. Howard, for some time connected with the Fore River Ship and
Engine Company, as it was then called. The Dow turbine was built
in the Sturtevant works and was really a meritorious machine. It
was what might be called an inward-flow reaction turbine. A motor
of bronze was built and the method which Mr. Dow developed is now
used in balancing our rotors.
3 A little later, during the development of the Curtis turbine,
a representative of Mr. Curtis came to us for journals for high-speed
work. He had trouble in finding a journal box capable of withstand-
ing the high speed of his rotor shaft. A box that we used was very
successful in solving the problem. We knew there was a somewhat
limited field for high-speed motors for use with our fans, however,
and were not quite ready to take up the Curtis turbine commercially.
4 A completed Dow turbine was frequently connected with one of
our No. 6 blowers. The governing device was not developed, but
that was not necessary in order to connect the turbine with a fan.
I believe this turbine was tested in our works, and afterwards tested
at the Massachusetts Institute of Technology for water consumption,
which was found to be high. I have always thought that the Dow
turbine possessed great possibilities and wondered why someone did
not develop it.
5 Experiments with a steam turbine of our own for use with our
fans resulted in the turbine described as the Sturtevant turbine. Mr.
Orrok's description of the rotor is substantially correct, except that
he omitted the fact that in the manufacture of this part we use an
open-hearth steel forging of the best quality. Among the uses of the
turbine are direct connection with generators, with fans for blowing
blast furnaces, and with multivane fans for work on shipboard. Four
fans with geared connection, recently built for a heating system in
the West, have done good work.
6 An interesting problem was recently presented when the Navy
Department insisted on fans for forced draft for the new torpedo boats
300 DISCUSSION
where oil fuel is to be used. This of course demanded a slow-speed
turbine, but we think we have worked out a satisfactory combination
by effecting a compromise between the fan and the turbine element.
It is interesting to recall that a few years ago the Navy Department
did not consider any motive power except a reciprocating engine.
Later the electric motor came into use and now many of the new ships
are equipped with forced -draft fans driven by electric motors.
7 There is not the slightest doubt that the general type of turbine
discussed in Mr. Orrok's paper is here to stay. Engineers like it and
engine builders must get ready to furnish turbine engines.
Chas. B. Edwards.^ We are more particularly interested in the
development of the large marine turbine, but our attention has been
recently directed to the smaller turbines owing to the Navy Depart-
ment specifying them for blowers; there is also a possibility of their
use for circulating pumps and other auxihary machinery on board
ship.. The Navy Department has increased the steam allowance for
turbine installations over what it was a few years ago when it was
limited to 50 lb. per h.p. hour.
2 The great problem in marine installations, particularly for
naval purposes, is that of weight. The navy contracts specify a cer-
tain weight of machine and if we exceed that weight we must pay for
it at the rate of about $500 a ton. In considering the turbine propo-
sition, therefore, we must look at it not only from the mechanical side
but also from the standpoint of weight. One of the difficulties, of
course, is that of the exhaust. The weight of piping, fittings, valves,
etc., runs up rapidly and it is therefore desirable that turbine auxiliaries
should be placed as close to the condenser as possible; and in order
to secure economical results it is also desirable to secure a low veloc-
ity of exhaust in the pipe lines.
V. F. HoLMES.2 The DeLaval Company has recently brought out a
combination high-and-low-pressure steam turbine. Many plants
where condensing water is available have an excess of exhaust steam
from auxiliaries and the question has arisen whether a machine could
not be devised for this class of service. That would necessitate stor-
ing up energy in times of an excess of exhaust steam to carry the
macliine over the periods of limited exhaust steam, and would involve
expense and complications. What is desirable is a machine in which
* Chief Engineer, Fore River Shipbuilding Co., Quincy, Mass.
'Power Equipment Company, Boston, Mass.
SMALL STEAM TURBINES 301
both the exhaust and the Hve steam can be used economically with-
out regenerators and other heat-storing devices.
2 The DeLaval combination high-and-low-pressure turbine is
built with two nozzle compartments, one for high pressure and the
other for low pressure. Each compartment is furnished with nozzles
having the proper ratio of expansion for the conditions under which
they operate. Some of the nozzles are furnished with shut-off
valves for regulation under variable conditions.
3 Two steam connections are provided, one for high -pressure
steam and the other for low-pressure steam, each connection leading
to its own governor valve, which in turn is operated by a separate
governor. The operation is entirely automatic, the low-pressure
governor being set for a speed slightly higher than that of the high-
pressure governor. On the total or partial failure of the low-pressure
steam supply the machine will automatically draw from the high-
pressure steam supply the steam necessary to make up the deficiency.
Also in case of the complete failure of the low-pressure steam sup-
ply, the machine will operate on high-pressure steam, and under this
condition will give practically the same economy as a high-pressuro
steam turbine.
4 The combination high-and-low-pressure turbine is built for
conditions where continuous operation is essential and where the
supply of low-pressure steam is intermittent or is apt to fail com-
pletely. The regulation when changing from cue steam pressure to
the other varies from 2 to 3 per cent, this being on an instantaneous
change from one condition to the other, such as seldom occurs in
actual service. A by-pass valve allows the admission of high-pres-
sure steam into the low-pressure compartment, for operation under
full-load conditions non- condensing. This by-pass valve is not
automatic, and is simply to enable the machine to carry full load in
case of failure of, or during repairs to, the condensing apparatus.
5 Both the low-pressure turbine and the combination high-and-
low pressure turbine are built for steam conditions varying from 5
lb. pressme above atmosphere to 10 in. of vacuum at the steam inlet.
They are also built for low vacuums for conditions where the temper-
ature of the circulating water or existing condensers prohibits the
maintenance of a high vacuum. The steam consumption of the
machine varies somewhat with the sizes and operating conditions;
the average machine operating with steam at atmospheric pressure
and exhausting into a vacuum of 27 in. to 27^^ in. will use from 28
lb. to 32 lb. of steam per b.h.p-hour.
302 DISCUSSION
6 The DeLaval Company is also building a high-speed, low-pres-
sure turbine particularly adapted for direct connection to centrifugal
pumps and blowers. This class of machine is built in both the low-
pressure and combination high-and-low-pressure types, and consists
of the DeLaval wheel direct-connected to the machinery to be driven.
On account of the direct connection of the wheel and the elimination
of the usual DeLaval reduction the machine can be economically
operated only at high speed, and for this reason is not suited to direct-
current generator work, but is particularly adapted for high-speed
pumping and blower work, such as power-plant auxiliaries, boiler-
feed pumps, elevator pumps, etc.
J. S. ScHUMAKER. An error has crept into these figures that I am
sure was not intended. That is, the figures given for the economy
of the Terry steam turbine were obtained from a turbine with nozzles
designed for 100 lb. pressure. But the steam pressure used on the
test was, I believe, 150 lb. One other point that may in fairness be
brought out is that the Terry turbine tests as offered here were made
without representatives of the Terry Steam Turbine Company being
present, while in the majority of the other cases cited the tests are
shop tests.
Prof. Carleton A. Read. I am interested from the fuel side of
the question in the use of non-condensing turbines in small manu-
facturing plants of from 75-kw. to 300-kw. capacity, where there is
an excess of exhaust that can be used only for feed-water heating and
heating the buildings in cold weather. We all agree that it is well
not to have oil in the exhaust if the condensation is to return to the
boilers, but man}^ plants have a good and cheap water supply and
after using as much of their exhaust as possible still have some going
to waste. Nearly all of the tests quoted are from the manufacturers
and without doubt are correct for the conditions under which they
were made, but data as to coal consumption under actual working
conditions would be of interest to the man buying an equipment for a
small plant.
Prof. Ira N. Hollis. One aspect of the subject impresses me as
important. The curves of efficiency used for comparing different
turbines relate particularly to the thermodynamic efficiency of the
machine or the number of pounds of steam per horsepower developed.
It seems to me that where a steam turbine is connected with a pump,
SMALL STEAM TURBINES 303
such as one used for feeding a boiler or for circulating water in a con-
denser, the machine ought to be treated as a whole. From this point
of view, the number of gallons of water delivered per pound of steam
or per pound of coal is an important factor and should be given in
every case. Naturally the pressure against which the water is pumped
is another factor. Ordinary reciprocating engines driving feed
pumps are very uneconomical machines. I have had experience with
pumps that used 100 lb. of steam per i.h.p. or even more. However,
the efficiency of the pumps as a whole for delivering water into a
boiler was never worked out.
2 It may be that the steam turbine is to replace the steam engine
for all purposes about a power station, particularly if the high-pres-
sure centrifugal pump can be developed into a highly efficient machine
in connection with the turbine. It seems to me, therefore that it
would be useful in connection with all tests of turbines used to drive
pumps, to give the combined efficiency of the unit.
Prof. Edw. F. Miller. In looking through these figures of steam
economies it will be noticed that the greater the load the smaller the
amount of steam per horsepower. All the turbines I have had to
do with would stand considerable overload, in some cases 80 per
cent. I would like to know what decides the maker in rating his
turbine. Apparently the economy line runs down as the overload
goes on. Why not rate the turbine higher and get better economy?
John T. Hawkins. I was a pretty old engineer when the turbine
was born and consequently know little about it except what I have
learned by reading and observation. I am not going to try to impart
information but I wish to ask a question. It seems to be a well-
known fact that with the turbine engine, the higher the load the greater
the efficiency within its limits. To what is the fact due that the
turbine is more efficient under high load ?
Richard H. Rice. Just a few words in explanation of why the
turbine water rates decrease as the load goes up and of the effect
of the various governing mechanisms on this action. The impulse
turbine is essentially a turbine of partial admission. In a multi-
stage condensing turbine of this type the buckets in the last stage
are usually designed of the right height and the nozzles of the right
proportion to use the entire circumference of the wheel. In a four-
stage machine, the next to the last stage will use perhaps one-half
304 DISCUSSION
the circumference of the bucket wheel for steam admission. It
could be designed to use all the circumference, but that would
involve undue shortening of buckets. In the second stage there is
a further shortening of the arc of steam flow, and in the first stage,
a still shorter arc is used, perhaps 90 deg.
2 In non-condensing turbines, if we were to attempt to use the
entire circumference of the wheel the buckets would be so small
that the machine would be impracticable and inefficient. We must
therefore use a short arc, decreasing the cost of the governing
mechanism and making a reasonable bucket speed possible. It is
evident that the bucket speed must depend on the size of the machine
and that, in connection with the operating speed, it is the prime
consideration in the cost of the machine. It would be a mistake to
make a 25-kw. machine with the same bucket speed as a 300-kw.
machine, because the former would be so large in diameter and so
expensive as to be impracticable.
3 It follows that one reason why the turbine increases in economy
as the load increases is that a larger circumference of bucket wheel
is used; a smaller percentage of the total power is wasted and there-
fore efficiency increases. Therefore, if the governing mechanism
works by throttling we have this condition : the steam pressure and
area of the nozzle system determine the amount of steam that can
be used in the turbine. In a machine with nozzles wide open, the
latter must be so designed that the turbine will carry maximum
load, as otherwise the turbine would shut down at maximum load.
It follows that nozzles designed for full pressure at maximum load
will be greatly throttled when running with light load, and conse-
quently the efficiency will decrease. Therefore it is advisable to
govern the nozzle system in such a way that nozzles can be designed
for full boiler pressure. By using a larger or smaller number of
nozzles, and hence a larger or smaller arc of wheel, the full economy
of the nozzles is obtained and only the proper number of nozzles are
open for a given load.
Chas. H. Manning. The diagrams confirm the opinion I had
formed of the small steam turbine to the effect that it is a steam
thief. But that its virtues will outclass its sins I am thoroughly
convinced. Recent developments of high-speed centrifugal pumps,
fans and generators open a field for the turbine in which it is sure to
succeed.
2 A small practical point is that almost all of these small high-
SMALL STEAM TURBINES 305
speed machines use the ring oiler, which is in general bad practice. It
has a very small contact on the shaft and any small thing will stop
its running. Furthermore, the rings frequently break. If for the
ordinary ring oiler a chain with a lai'ge arc of contact is substituted,
preferably a window-cord chain, it will never fail and will bring up
ten times as much oil as a ring oiler. While this is a small point,
any machine depends more on the perfection of its detail than it
does on the theory on which it is built.
C. P. Crissey.^ There is one type of the small turbine to which
the author has given scant space; that is, the small condensing
machine. While, perhaps, the field is not so wide for this type as it
is for small turbines exhausting at or above atmosphere, it cannot
be ignored. Practically all marine work requires condensing prime
movers, and many small stations use this type. Only one example
of a condensing machine is referred to by the author, the results of
tests being shown in Fig. 30. It would be a mistake to consider this
curve as representative of small condensing turbines.
2 Small turbines as well as large derive great benefit in economy
from high vacuum, and a vacuum of 28 in. is easily obtained on small
machines of proper design. In a well-designed small turbine the
vacuum shows no greater tendency to fall with the increase of load
than in large machines. Why the Kerr turbine show^s a loss in
vacuum as the load and hence steam flow increase, we are unable to
tell definitely from this paper. It will be noted that the steam is
discharged from the buckets on each side of the wheels. It is there-
fore necessary for one-half of the total flow to pass about the wheels
in order to reach the succeeding nozzles. Excessive velocities and
throttling will occur in the low-pressure stages where the volumes
encountered are great, unless large areas are provided for this steam.
I understand that in the Kerr turbine this throttling is obviated as
much as possible by providing holes in the wheels. These holes, how-
ever, increase the windage loss.
3 One of the reasons for abandoning the Riedler-Stumpf turbine
in Germany was the inability of its buckets to handle large volumes
of steam successfully. The same objection holds against all machines
of the Riedler-Stumpf type.
4 The only small turbines having buckets capable of caring
efficiently for large volumes of low-pressure steam are the De Laval
and the Curtis types. The DeLaval turbine is seriously handicapped
' General Electric Co., West Lynn, Mass.
306 DISCUSSION
by its high rotative speed, while the Curtis turbino, due to its pressure
and velocity stages, is capable of moderate speeds. In order to show
that the results of Fig. 30 are not typical of all small condensing tur-
bines, I will say that Curtis condensing turbines of from 100 to 200
h.p. give economies of 18.5 to 15.5 lb. of steam per b.h.p. hour when
operating with 150 lb. dry steam and 28 in. vacuum.
5 Regarding the curves of this paper, I believe they should be
compared at rated speed, because the bucket angles are designed for
this speed. The rated speed may be taken as the maximum stated
upon the curves.
W. J. A. London. I wish to add ^ something to my remarks
in connection with the curve mentioned by me earlier in the discus-
sion; namely, a comparison between Fig. 25 and Fig. 28. In making
this curve, I used only absolute tests according to the figures men-
tioned and made no deductions whatever except in the question
of relative peripheral speeds. If the curves plotted in Fig. 28 are
reproduced for a series of full-load points on a speed basis, a positive
curve will be formed. On this curve is plotted the two full-load
tests shown in Fig. 25. Now, then, as there are only two points
given in the Terry tests, it is impossible from these tests to obtain the
nature of the curve, but the point I particularly wished to bring
forward was that these two points practically coincide, — one test,
as a matter of fact, being better and one worse, — both of them be-
ing so near the Curtis curve as to make little difference. They are not
so far away as Mr. Rice would have us believe from his diagram.
2 The point has also been raised as to whether emergency gover-
nors were fitted on other makes of turbines besides the Curtis. Par-
ticularly with the Terry turbine and I beheve with the majority of
the other makes, an emergency governor is not provided for the
reason that a positive type of governor is fitted on the main shaft.
The worst that can happen is the breaking of a spring, which would
immediately close the valve. With a governor driven by a gear
shaft an emergency governor is more essential for the reason that
the gears are likely to break; hence some form of governor is used
on the main shaft. If the direct-connected governor on the main
shaft is likely to get out of order, why then is the emergency governor
not likely to get out of order when placed in the same position? Up
to the present time the Terry Turbine Co. has not had a machine
burst, and with the type of governor adopted and the speeds em-
SMALL STEAM TURBINES 307
ployed, the designers consider an emergency governor an unneces-
sary luxury.
R. H. Rice. Mr. London claims that the steam consumption of
the Terry turbine is the same as that of the Curtis turbine, when
operating the Terry turbine at its designed speed and reducing the
speed of the Curtis turbine to two-thirds of its designed speed of
3600 r.p.m. The inaccuracy of this comparison can be readily
understood when it is known that the angles of the buckets in the
Curtis turbine would be radically changed if designed to run at two-
thirds of the present rated speed.
2 In discussing emergency governors, it must be realized that
we are dealing with much higher speeds than those usual with recip-
rocating engines. It has been found best in many plants to install
emergency governors on reciprocating engines. If this is desirable
on slow-speed apparatus, how much more desirable, and even neces-
sary, is it on high-speed apparatus like steam turbines. Many other
accidents besides the breaking of a spring can happen to a positive
type of governor fitted to the main shaft, and any one of these is suffi-
cient to cause a dangerous increase in speed of the turbine. An
emergency governor can be made to possess the utmost certainty
and reliability of action, since its function is to shut down a machine
and not to regulate its speed.
J. H. LiBBEY. The applications of small steam turbines men-
tioned by the author, except for driving high-pressure fans, refer to
uses with auxihary apparatus in a central power station. For this
purpose a small steam turbine must be considered in competition
with a reciprocating engine, and in general the choice will be decided
by the following considerations: First cost, attendance required,
maintenance and repairs, space, economy and influence on design
of power station.
2 First Cost. At present, when the service permits operation at
speeds approximating those for which the turbine was designed, the
cost of the turbine is somewhat lower than that of a corresponding
reciprocating engine.
3 Attendance Required. The attendance required for a small
steam turbine is less than that required for any other type of steam
machinery. It approaches very closely that required for an electric
motor.
4 Maintenance and Repairs. In Par. 32, the author indicates
308 DISCUSSION
that small turbines have been running for only three years. Ob-
viously in this time no data of great value could be obtained to enable
a decision to be made in regard to maintenance and repairs. The
evidence, however, strongly indicates that they will be materially
less than for a reciprocating engine.
5 Space. Steam-turbine-driven apparatus is generally charac-
terized by the small space required. In a great many cases', a tur-
bine unit can be installed where a reciprocating engine would be
impossible.
6 Economy. An inspection of Fig. 28 and Fig. 29 shows that for
the best conditions a turbine can deliver a horsepower with as little
steam as, or less steam than, the same size reciprocating engine. In
installations where the conditions are not favorable, the economy is
reduced. Unfavorable conditions for a steam turbine are low super-
heat, low steam pressure, high back pressure or reduced speed of the
turbine, on account of the characteristics of the driven machine.
The last condition is the most likely to cause reduction in the economy.
7 It should be borne in mind that in the ordinary large central
station where fairly large generating units are installed, the steam con-
sumption of the auxiliaries does not in general amount to more than
10 per cent of that of the main generating units. In such cases, the
auxiliary exhaust will heat the feed water to about 175 or 180 deg.
fahr. A considerable increase of steam consumption of the auxil-
iaries can be permitted before there is sufficient exhaust to heat the
feed water to 212 deg. fahr. In most cases, therefore, the steam
consumption of these small auxiliaries is a matter of secondary con-
sideration.
8 In the various auxiliaries generally used the inherent conditions
which would affect the steam consumption would be in general as
follows:
Exciter, favorable-
Circulating pump, speed low for best results.
Hot- well pump, favorable.
Forced or induced-draft fans, speed low for best results; special
design of fan required.
Feed pump, speed low; however, the steam consumption of
a turbine-driven multistage centrifugal feed pump is much
lower than that of a reciprocating pump of the same
capacity.
9 In this connection Mr. Orrok's statement in Par. 32, that there
SMALL STEAM TURBINES 309
seems to be no change in steam use with length of service is of impor-
tance as it is well known that the steam consumption of engines or
pumps increases greatly with wear of valves, rings, pistons, cyhnders,
etc.
10 Many central stations are toda}^ operating the auxiliaries
with superheated steam. Very few changes are required in the
structure of a steam turbine to adapt it to superheated steam by the
use of which the economy is improved. The reciprocating engine
gains in economy from superheat, but greater changes in the design
are required to obtain satisfactoi-y operation, and the expense of the
engine is therefore increased.
11 Influence of Design on Power Station. Turbine-driven exciters
are generally light in weight and compact. They can be set on the
engine-room floor without a heavy foundation or resulting vibration.
12 Circulating-pump units are Of simple design. In many cases,
a combination of auxiliaries may often be effected. There is on the
market a jet condenser, the centrifugal pump and air pump of which
are on the same shaft with the turbine. When in surface-condenser
work the conditions are such that the speed of the circulating pump
is subject to little variation, the turbine, circulating pump, and hot-
well pump can be mounted on the same shaft. One manufacturer
is prepared to add a rotary vacuum pump to these, either direct-
connected or chain-driven. This arrangement gives practically one
auxiliary for a surface condenser in place of three.
13 Future Designs. The small steam turbine has sufficiently justi-
fied its existence. The future will undoubtedly show types with
improved economy, especially at reduced speeds, simplicity of design,
rugged characteristics, ability to operate without attention, interior
construction that is easily accessible and such that few repairs due
to wear are required.
The Author. The author is greatly pleased with the reception
accorded his paper and the amount of discussion which it brought
out. He must take exception to a comparison of water rates plotted
on percentages of load as abscissae, and a new diagram has been pre-
pared. Fig. 1, showing the results of all the water-rate curves plotted
with bucket speed or peripheral velocity as abscissae, obviously a
much better measure of the performance of these machines. The
author would like to take up the question of improper entrance and
discharge bucket angles in machines of the Riedler-Stumpf type,
as well as the fluid friction question, both mentioned in the discussion
310
DISCUSSION
of Mr. Rice ; but these should be the subject of a mathematical paper and
are not of serious importance in a small turbine. The author feels
that Mr. Ball has failed to grasp the fact that with small rotating
masses speeds of from 600 to 3000 r.p.m. are not as objectionable as
a speed of 150 r.p.m. in a modern four-valve engine, or 100 double
strokes per minute in a direct-acting pump.
2 Replying to Prof. Hollis' discussion, the author knows of many
power plants where entire reliance is placed on turbine-driven multi-
10,000 15,000 ^,000
Peripheral Speed —Ft. per iiiiii.
25,000
Fig. 1 Steam Consumption of Small Turbines Plotted with Peripheral
Speeds as Abscissae
stage centrifugal pumps for feed-water service. He knows of no
case where an attempt has been made to find the coal consumption
of the feed pumps directly; in other words, the duty of the pumps.
It has usually been obtained through the steam consumption with a
knowledge of the evaporation constant of the plant. The use of
Venturi meters in the feed lines and in the steam connections to the
turbine-driven feed pumps would give this duty directly, and a partial
installation of this nature has been made at the Waterside Station of
the New York Edison Company.
No. 1243.
TESTS UPON COMPRESSED AIR PUMPING
SYSTEMS OF OIL WELLS
By Edmund M. Ivbns, New Orleans, La.
Junior Member of the Society
When the Louisiana oil fields at Evangeline were in full operation,
they offered exceptional opportunities for the study of air lifts.
Nearly every known method of piping the wells was in use. The
air plants originally installed were the crudest affairs imaginable,
having been erected in feverish haste during the boom several years
ago. When the production of the fields began to decrease, and the
price of oil also declined, it was realized for the first time that the
operating expenses were abnormal, and that unless greater economy
were practiced, disastrous results would follow. Few changes were
made, however, up to eighteen months ago, beyond the purchasing of
additional equipment.
2 Each concern has a central station or air plant and all the
compressors therein are connected to a manifold from which the air
lines lead to the various wells on the property held by that concern.
The manifold design is such that by manipulating the valves, any
machine may be made to operate any of the wells.
3 Often the air lines reach the wells by a roundabout way, and
have innumerable short bends, valves, double swings to avoid pipe
cutting, and plugged tees instead of elbows. All of this tends further
to decrease the economy of the operation, and taking all things into
consideration, it is little wonder that the eflniciencies of the plants were
low. The size pipe used for these air lines is designed neither for the
amount of air to be transmitted nor for the distance it is to be carried,
but is with one exception 2 in. in diameter.
4 The boilers of the air plants are of 40 h.p., of a portable con-
tracted waist type, and few were covered with asbestos. The boilers
were so set that one-fifth of their lengths projected into the open, as
Presented at the Spring Meeting, Washington, May 1909, of The American
Society of Mechanical Engineers.
312
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
indicated in Fig. 1, in order to avoid the necessity of perforating the
roof to receive the stacks and to provide cooler boiler-rooms, regard-
less of the heat wasted.
5 The redeeming feature in all the plants is the type of compressor
in general use. These compressors are generally of high grade, and
display remarkable endurance. It is common for a machine designed
for 350-lb. pressure to operate under a pressure of 500 lb., and at
speeds far in excess of those for which it was designed. The most
popular type of compressor has the duplex steam end and compound
or two-stage air end. The steam cylinders are fitted with Meyer
adjustable cut-off valves and the air cylinders in some instances with
piston and in others with Corliss intake valves and poppet discharge
valves. Plain speed governors are used and the capacities of the
Fig. 1 A Typical Air Plant
compressors range from 100 to 1000 cu. ft. of free air per minute and
operate at pressures of from 150 to 750 lb. per square inch. The
machine best adapted to the purpose, however, is the 500-cu. ft.,
500-lb. type.
TERMS
6 An explanation of certain terms to be used may not be out of
place.
"Submergence in feet" refers to the number of feet below
the surface of the fluid (after the well has been pumped
down, and is operating under its normal conditions)
that the air under pressure is admitted.
" Per cent of submergence" is the submergence in feet divided
by the total number of feet of vertical discharge line,
measured from the point of admission of the air to the
point of discharge of the fluid.
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS 313
" Volumetric efficiency" of the compressor is the actual amount
of free air that is compressed and discharged by the cyl-
inder, divided by the cubical contents of that cylinder.
"Free air" is air at standard temperature and pressure.
" Pumping head" is the vertical distance in feet (after the well
is pumped down, as before stated) from the fluid level
in the well to the point of discharge.
Gal. per minute X pumping head in feet
"The Constant"
"The Ratio" =
Cu. ft. of free air per minute
Cu. ft. of free air per minute
Cu. ft. of fluid per minute
DESCRIPTION OF SYSTEMS
7 Fig. 2, 3, 4, and 5, illustrate the air lift systems that are and
hi-ve been in use on the oil fields.
8 Fig. 2 shows the Straight Air or Sanders system. The well
top is sealed as shown at A. Compressed air is forced through the
pipe B into the space between the discharge or eduction pipe C, and
the well casing D.
9 When without air pressure, the fluid in the well will stand at
some point- such as E, the level in the air space and the discharge
line being identical. When air is forced through B, the level of the
fluid in the air space is gradually forced down until the end of C is
uncovered. Instantly some of the air escapes into the discharge
pipe C, lowering the air pressure in the air space F. This cau^^es
the fluid to rise in and up the air space and discharge pipe until a
point is reached where air and water pressure balance. Then, more
air coming in, the pressure again rises, the fluid level is forced down
as before, more air escapes into the discharge pipe, and thus the cycle
is repeated. As may be readily seen, the air that rushes into the
discharge line carries the "slug" of water that has just previously
entered.
10 Fig. 3 shows what is commonly known as the Central Pipe
system. The discharge line A is placed inside of the well casing as
before and inside of the discharge is suspended a small air line usually
H in. in diameter. The end of the 1 i-in. Une is plugged and a number
of ^-in. holes are drilled inclining upwards in the last joint of pipe.
Air is forced down through the small air line shown, passes out of
the ^-in. holes, and mingles with the fluid carrying it out through
the discharge line A. It is generally supposed that the fluid in this
314
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
case is discharged because of the aeration of the fluid in the discharge
pipe which in turn is caused by the intimate comminghng of air and
fluid. The weight of the fluid column inside of the discharge pipe
is therefore less in pounds per square inch than that without and the
energy due to this difiference in weight is utilized to lift the fluid and
overcome the various losses.
Fig. 2 Fig. 3 . Fig. 4
Straight Air Lift Central Pipe Return Bend
System System System
11 What is commonly known as the Open End system of air
lift was at one time in quite extensive use on the field. It is
similar to the system just described except that the small air line is
open at the lower end, and of course there are no holes drilled in the
air line.
12 Fig. 4 illustrates a form of the Return Bend system. It is
clj^imed by the inventor that: " It consists in improved processes and
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
315
apparatus whereby the compressed air is delivered in bulk into the
lower end of the water eduction pipe, and the water and air are
caused to ascend through said pipe in distinct alternate la3'ers of
definite dimensions."
13 The use of this system has been discontinued in Evangeline
because, as the field managers told the writer, it failed to produce as
large a quantity of fluid as that produced by other systems.
^—3
SEiCTION f}-j9
Fig. 5 System Combining Features of other Systems Described
14 Fig. 5 shows a patent system which in reality is a combination
of the several systems already described. The claims of the inventor
are: less submergence, and hence less air pressure necessary, decreased
air consumption, or with an equal amount of air, increased fluid yield.
15 Compressed air is forced through a down into the foot piece,
which is placed at that point of submergence shown by test to be
most economical. The well top is sealed and air under pressure
316
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
is also admitted between the casing and discharge pipe on the
water head by means of the branch shown at h. This forces the
fluid to a higher level in the discharge pipe and also prevents fluid
in the air space or chamber from vibrating and foaming. This is
quite an advantage in oil well pumping as the liability of making
"riley oil" is thereby greatly lessened.
16 The footpiece shown in section is made of cast brass and is in
two parts. The air on reaching the foot piece divides and goes up
through the hollow prongs / and g and out the nozzle n. The nozzle
is adjusted to receive the quantity of air to be used by screwing the
upper part s of the footpiece, in or out as the case may be. To
increase the velocity of the fluid in the discharge hne, the footpiece
is restricted and formed into a "venturi" as shown at v.
Fig. 6 Type of Compressor Used
Test No. 1
17 The Crowley Oil and Mineral Company was the first to take
active steps for the improvement of their plant and pumping equip-
ment. They decided to install the patent air lift last described
(Fig. 5). A test of the old system was first made to determine the
amount of compressed air used and the fluid yield. The new equip-
ment was next installed and a similar test made of the same duration
and under the same conditions. The tests and installation were
conducted on Well No. 32, 1805 ft. deep, and located 542 ft. from the
compressor operating it. The air to the well was controlled by means
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
317
of a manifold in the plant and was conveyed to the well top through a
2 in. pipe line which as usual was in poor condition and badly
designed.
18 The system of pumping was that illustrated in Fig. 2. The
well casing was 6 in. in diameter, suspended inside of which was a
4-in. discharge line.
19 The compressor was a duplex steam and compound air type
made by the Ingersoll-Rand Company and designed to compress
1000 cu. ft. of free air per minute to 350 lb. pressure. The steam
end was fitted with Meyer adjustable cut-off valves and the air end
with Corliss intake and poppet discharge valves. The machine is
shown in Fig. 6.
20 The discharge pipe from the well top was run up into a steel
tank of known dimensions and the amount of fluid pumped during the
Fig. 7
test ascertained by direct measurement. Air gages, previously tested,
were placed both at the compressor and at the well top thereby
making it possible to determine the friction losses in the manifold
and air line and also the actual pressure at the well top. Simultane-
ous indicator cards were taken from the steam and air ends of the
compressor, and from these cards were obtained the volumetric and
mechanical efficiencies, the steam and air horse powers, and the
.steam consumption (theoretical) of the machine.
21 The volumetric efficiency .was assumed to be the ratio of
the piston travel during admission stroke to the total piston db-
placement, {— on the indicator card. Fig. 7). Sometimes this
ce
method is inaccurate and unsatisfactory (1) because the enter-
mg air at atmospheric temperature and pressure is heated by
318 COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
contact with the cylinder walls and piston; and (2) because of leak-
age of air from the compression side of the moving piston to the suc-
tion side. Neither the expansion resulting from the first condition
nor the reduction in volumetric efficiency resulting from the second
are observable on the indicator card.
22 The first inaccuracy was partially overcome by placing a
recently calibrated thermometer as far down in the intake pipe of the
compressor as possible, noting the temperature and making the
necessary corrections as will be observed in the log of results.
23 The method of ascertaining the volumetric efficiency, that the
writer would have used, but for his inability to obtain the necessary
apparatus, was in brief as follows:
Connect the air discharge of the compressor to an enclosed
tank. From this tank^ connect to a cooler and from
thence to a second enclosed tank of known dimensions.
Place a regulating valve between the first tank and the
cooler, setting the valve to maintain the pressure in the
first tank at that point at which the efficiency is to be
determined.
Attach test gages to both tanks and a reliable thermometer
to the second tank.
Start the compressor and note the temperatures of the intake
air and of the air in the second tank both at the beginning
and end of the run. Note also the initial and final air
pressures and the reading of the barometer, and the speed
in revolutions per minute of the compressor.
The volume of air compressed is then determined from the
formula:
273 + T/29.92 X P, _ 29.92 X P
~ ^ R V 273 + r, 273 + T,
where
V = Volume of air compressed.
Vi = Cubical contents of the second tank.
T = Room, or intake air, temperature.
Ti = Initial temperature of the air in the tank,
y, = Final temperature of the air in the tank.
R = Reading of the barometer in inches of mercury.
P and Pi = Initial and final air pressures in the
tank.
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS 319
The volume of air thus obtained divided by the total piston displace-
ment equals the volumetric efficiency.
OBSERVATIONS
24 Every thirty minutes for a period of six hours, readings were
taken of the boiler pressure gage, the air gage at the compressor
and at the well top, the r.p.m. of the compressor, the temperature
of the intake air and of the barometer. A set of indicator cards, and
also a sample of the fluid pumped from the well, were taken at each
interval ^^,' |"
25 The temperature of each sample of fluid was noted; it was then
placed in a proper receptacle, and at the end of the test, the weight of
a gallon was ascertained, together with the specific gravity of the oil.
The amount of fluid pumped was determined, as before stated, by
direct i^ easurement, due allowance having been made for the samples
that were withdrawn.
26 When these tests were run, no attempt was made by the writer
to re-design the air lines, or to correct in any manner the numerous
other defects. The old system was tested just as it had been operated,
and the new system was installed and tested under the same adverse
conditions. After both systems had been tested, some few of the
defects were corrected in the manifold and air line design, thereby
insuring more economical operation in the future.
TABLE 1 SUMMARY OF RESULTS
The Crowley Oil and Mineral Company, Evangeline, La.
Old System New System
Duration of test, hours 5.5 6.0
Meani.h.p 122,56 89.19
Mean water h.p 9.97 10.36
Meanairh.p 107.38 79.16
Gallons of fluid per second 0.542 0.608
" " " " hour 1953.6 2188.2
Barrels of fluid per hour 46 . 51 52 . 1
Weight of 1 gal. of fluid 8.7 8.69
Mean temperature of fluid, deg. fahr 111.5 113.2
Percentage of salt water in fluid 87 . 3 86 . 7
sand " 2.2 1.9
crudeoil " 10.5 11.4
Barrels of oil per hour 4 .86 5 .94
Barometer reading, inches of mercury 29 .95 29 .94
Specific gravity of oil 0.9 0.9
320 COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
Old System New System
Constant 63 . 1 101 .8
Size of air line, inches 1 . 25
Total depth of well, feet 1805.0 1805.0
Size of casing, inches 6.0 6.0
Height above ground to which the fluid was pumped,
feet 18.5 18.5
Size of discharge line used, inches 4 4
Total length of vertical discharge line 1513 . 5 1513 . 5
Total length of vertical air line in well 1513.5 . 1493.0
Dimensions of compressors, inches* 10x22x16x20-7^x18x16x20
Number operated 1 1
Kind of fuel usedf Crude oil Crude oil
Gallons of fuel used per hour 48 . 36 35 . 27
Barrels of fuel used per hour 1.15 0.835
Price of 1 bbl. of oil at time of test, dollars 0.90 0.90
Costoffuelforproducinglbbl. of fluid, dollars 0.0222 0.0146
Cost of fuel for producing 1 bbl. of crude oil, dollars. ... 0.212 0.126
♦Type of compressor used, Rand Drill Co. Imperial Type X, Steam Cylinders, compound
air cylinders.
tType of boiler, oil well supply, portable contracted waste.
Test No. 2
wells no. 12, 30, and 32 op the crowley oil and mineral
COMPANY
27 The saving in air volume accomplished by the new system
led those interested to endeavor to operate two wells with one
machine, something before considered impossible in the field.
28 Well No. 30 was forthwith tested, though not with sufficient
accuracy to warrant the publication of the results, and the approxi-
mate pumping head and submergence established. The new system
was then installed with the requisite pipe to equalize the submergence
(hence working pressure) of this well with that of No. 32. How
successfully the working pressures of the two wells were equalized
may be seen by reference to Table 2 of the Appendix.
29 The two wells in question were then connected to one air
compressor with gratifying results. No trouble was experienced in
starting, and the machine furnished air in abundance for steady
operation.
30 Preparations were being made to run the usual test when the
compressor operating Well No. 12 "went dead." This last named
well had been previously tested and equipped with the new system.
This shutdown, of course, would mean a loss of at least a day's pro-
COMPRESSED AIR PUMPING Si'STEMS OF OIL WELLS 321
duction from the well, amounting to quite an item, so the writer
advised that this well be also connected to the machine already
operating No. 30 and No. 32. By speeding the machine up a few
revolutions, the additional load was easily taken care of as may be
more fully noted by reference to the accompanying log (Table 2).
TABLE 2 SUMMARY OF RESULTS
Wells No. 12, 30, 32, Crowley Oil and Mineral Company
Duration of tests, hours 6.0
Mean (total) i.h.p 151 . 1
w.h.p 25.14 •
a.h.p 129.05
Total gallons of fluid per hour 6168.0
« barrels " " " " 146.87
" " " oil " " 16.17
Well No. 12
Weight of 1 gal. of fluid 8.5
Temperature of fluid 118 . 5
Per cent of salt water in fluid 87 . 2
" " " sand " " 1.3
" " " crudeoU " " 11.5
Barrels of oil per hour 6 . 44
Specific gravity of oil 0 . 87
Total depth of well in feet 1705 . 00
Size of casing, inches 6 . 00
" " discharge line, inches 4 . 00
Well No. 30
Height above ground to which fluid was pumped, feet 17.5
Total length of vertical discharge line 1025 . 5
airline 992.58
Weight of 1 gal. of fluid 8 .65
Temperature of fluid 120.2
Per cent of salt water in fluid 88 . 3
" " " sand " " 1.5
" " " crude oil " ".... 10.2
Barrels of oil per hour 4.83
Specific gravity 0.9
Total depth of well in feet 1920 .00
Si«e of casing, inches 6 .00
" " discharge line, inches 4.00
322 COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
Height above ground to which fluid was pumped, feet 18.00
Total length of vertical discharge line 1516 .3
Total length of vertical air line in well 1494 .2
Well No. 32
Weight of 1 gal. of fluid, pounds 8.7
Temperature of fluid 114 .5
Per cent salt water in fluid 86 . 9
I. " " sand " " 1.8
' " " crude oil " " 11.3
Barrels of oil per hour 4 .90
Specific gravity 0.9
Total depth of well, feet 1901 .00
Size of casing, inches 6 .00
" " discharge " 4 .00
Height above groimd to which fluid was pumped 18 . 5
Total length of vertical discharge line 1513 . 0
Total length of vertical air line in well 1493 .0
Size of air lines in well, inches 1 .25
Barometer reading, inches of mercury 29 .95
Dimensions of compressor, inches* 10x22x16x20
Number operated 1
Kind of fuel usedf Crude oil
Barrels of fuel used per hour 1 . 45
Price of 1 bbl. of oil at time of test, dollars 0 . 90
Cost in fuel of producing 1 bbl. of oil, dollars 0 .074
♦Type of compressor used, Rand Drill Co. Imperial Type X, duplex steam cylinders, com-
pound or two stage, air cylinders.
tType of boilers, oil well supply, portable contracted waste.
Test No. 3
well no. 2, mamou power company
31 This test was run in the same manner as those preceding
except that the fluid field was ascertained by means of a two-foot
rectangular weir placed between the earthen fluid and oil pits, the
salt water bleeds of the former having been closed. The old system
used was that illustrated in Fig. 3.
32 The depth of fluid over the crest of the weir was measured by
means of the ordinary hook gage calibrated to read accurately in
hundredths of a centimeter. The weir constant was previously
determined by testing in the usual way, using a sample of the fluid
as pumped from the well.
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS 323
TABLE 3 SUMMARY OF RESULTS
Well No. 2, Mamou Power Company
Old System New System
Duration of tests, hours 10.0 10.0
Mean i.h.p 99.1 62.8
" w.h.p 9.85 13.36
" a.h.p 82.5 50.4
Gallons of fluid per second 0.694 0.849
" hour 2499 .6 3056 .4
Barrels of fluid per hour 54 . 75 72 . 77
Weightofl gal. of fluid, pounds 8.72 8.75
Mean temperature of fluid, deg. fahr 118.3 117.9
Percentage of salt water in fluid 87 . 7 86 . 1
" sand " " 1.2 1.6
"crude oil in fluid 11.1 12.3
Barrels of oil per hour 6.08 8.95
Specific gravity of oil 0.9 0.9
Barometer reading, inches of mercury 29 .94 29 .93
Weir constant 24 .39 24. 39
Pumping constant 97 . 1 202 .9
Total depth of well in feet 1901 .0 1901 .0
Size of casing, inches 6.0 6.0
Height above ground to which fluid was pumped, feet . 3.33 3 . 33
Size of discharge line used, inches 4.0 4.0
Sizeof air line in well, inches 1.25 1.25
Total length of vertical discharge line 1500 . 0 1500 . 0
Total length of air line in well 1489.5 1489.5
Dimensions of compressor, inches* 7^x18x16x16 — 7^x18x16x16
Number operated 1 1
Kind of fuel usedt Crude oil Crude oil
Gallons of fuel used per hour 44.22 30.53
Barrelsof fuel " " " 1 .05 0.727
Price of 1 bbl. of oil at time of test, dollars 0.85 0.85
Cost in fuel of producing Ibbl. of fluid, dollars 0.0163 0.0085
Cost in fuel of producing 1 bbl. of crude oil, dollars .... 0 . 128 0 . 069
*Type of compressor used, Hall Steam Pump Co .Duplex steam cylinders, compound air
cylinders. Plain "D" valves on steam end, poppet valves on air end.
tType of boiler, 72'xl8' horizontal return tubular, manufactured by the Loclcout
Boiler Co.
Conclusion
33 A careful examination of the tests brings out several points
that may require explanation.
34 The loss of air pressure by friction in the small l|-in. air
line in the well, to which the footpiece of the new system was attached.
324 COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
was approximately determined as follows: Pipe connections were
made at the well top, so that by the manipulation of various valves,
the air from the main line could be sent either through the 1^-in.
air line or into the space between the well casing and the discharge
line. By noting the pressure gage readings in each instance, the
friction loss (assuming that there is no loss by friction when air is
forced between casing and discharge) is represented by the difference
in the readings. Corrections were made, of course, for that part of
the discharge line below the footpiece.
35 It was impossible to obtain the actual friction loss in said 1^-
in. line by other means more accurate than those employed. While
some little error may be involved in assuming no friction loss in the
one instance, a comparison of the loss thus obtained with the theoret-
ical loss is quite favorable, the former loss being the greater.
36 Reference to Table 3 will show that the working submergence
of the new system is less than that of the old, in spite of the fact that
there is the same amount of pipe in the well in each case. This is
due to the additional drop in pumping head caused by the increase
of fluid yield. All calculations of submergence and pumping head
were made from the observed air pressures after correcting for fric-
tion losses, etc. The mean of these calculations was verified as far
as possible by actual measurement. This was done by shutting
down the compressor after the well had been in steady operation for
several hours and pulling the discharge line. The point at which the
fluid stood, while the well was being pumped, was plainly defined on
the pipe. The time required after shutting down the compressor
to pull the first "triple" from the well was a fraction less than two
minutes. Comparison of the actual pumping head and submergence
thus obtained with those obtained by calculations from the pressure
gage readings was in each case very close, a difference of 10 ft. 2 in.
being the maximum.
37 Acknowledgment of valuable aid during tests is hereby made
to the following who checked the writer in his various observations:
on Well No.'32,'to'Mr. B. Brand, of the Crowley Oil and Mineral Co.;
on Wells Nol 12, 30 and 32, to Mr. Brand, Mr. J. Murphree and Mr.S.
Bolin, of the Crowley Oil and Mineral Co.; on Well No. 2 of the Mamou
Power Co., to Mr. J. A.Sonet of that company and his able assistants;
and especially is the writer grateful to Mr. J. W. Smith for courtesies
extended during the former's sojourn on the field.
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
325
X
XMao aaj I «>; w t^ « m
AONaiOMJ^ ONIdFIQjJ I » C3 05 OJ OS
i.^jao
Had aoNaoHaHHQg
HONaoHawaflg
«> 00 CO N 00
C5 00 00 Oi 00
t- CO
X W 00 to -H
£1 CO c^ ci M
<N CV| <N CS (^5
•« lo u5 X; 00
«0 CO CO T)< o
00 <0 l>. (N. 00
N CO CO CO co'
N N <N C5 N
00 «
a
346.6
358.05
359.15
359.66
361.28
u) o
«0 00
CO CO
aaanaij I -^ m co ■>!< o
1%%1%
CD «
S 2 !2 2 2
CO Tl< ■^' ijj lO
<0 h- 00 0> O
r^- 1
326
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
i 2
ij
fl
h)
«
a
S
ts
Jz;
§
1
H
>>
W
oa
H
^
ai
rH
^
H
i-J
n
<!
H
O N 00 00 CO
CO 00 <N ^ «>
CO us •*
CO CO CO CO CO
CO CI CO CO CO
CO CO eo j
XNao
0> 0> CD «D CO
CO •* 0> -"It CO
CO CO CO
t^ t^ r~ t^ t^
IN N (M IN IM
t^ CO b~ 00 t^
<N (N (N <N IN
t^ t^ t-
CI IN N
xaaj
Ni aoNaaaawang
05 O) 0)
to lO i-H ^ ,-1
05
rH •^ US ■* 05
05 05 05
CO CO ^ -1 -H
;:^?52g5:l
555
•M -d •»
CO lO ■* ■* lO
t^ t^ t^ I> l>
us CD us us -^
t~ t^ I> t- t~
■* us us
t~- t^ t^
•d 'H 'V
CO ^ ■* •* 00
05 00 Ifl lO t»
t~ r- 1^ t^ t^
t' ■* t~ 05 05
t^ O 00 00 US
t^ 00 t^ t~ t^
us 00 US
CO CO t>.
t^ t^ t>
•d -H -Al.
CO CO ■* ■* Tl(
^ CO CO CO ^
■*■*•<*(
o o o o o
o o o o o
o o o
•d -H 1
to CO t~ t~ m
T-l 05 CO CD 05
05 00 00 00 00
■* us O) t^ CO
05 (N o d 1^
00 05 00 05 00
87.9
88.5
89.1
avajj ONidWQj
US us rt ,-( rH
rt CO us CO .-H
.-1 iH 1-4
CO CO ^ rt -H
t^ t^ 00 00 00
o o o o o
^H 03 CO 05 1-t
00 CD t^ CD 00
o o o o o
III
K
D
(D
(D
«
•<
IPAilV
<N (N O O O
O O O O O
N (N IN (N IN
O us <M US O
o o o o o
<N IN (N IN <N
IN IN (N
jajoooja^ui
t~ t- t^ t^ l^
t^ r- r- t> i^
t~ t~ t» t^ t»
r- t^ t^ t^ t-
r^ t^ t»
t^ t^ ^•
aSjBqosiQ
f- t^ us us 10
o o o o o
N N C^ (N <N
us O t^ O us
O -^ O rH O
CI (N (N IN M
us us us
o o o
IN N N
oiivg
IN 00 US US OS
.-I 05 r>^ t^ c»
00 t> r- t^ t^
O) •<* O 05 00
ess .-1 d 05 t>^
t^ 00 00 W b-
78.6
78.1
79.5
aiQNij*! aaj
t^ r~ t> t~ t^
00 W 00 00 00
t»- i^ t» r» r-
00 X 00 00 00
t^ r- t^
00 00 00
•* •^ •.^
axiiNTi\[
Had SNO^^vo
t^ t^ I^ t^ t-
^ ^ ^ ^ ^
CD CO CO CD CD
CO CO CO CO CO
i^ t>. r- r^ t^
■* n< Ti< •>i< •*
CO CO CO CO CO
36.47
36.47
36.47
HNVX
1ft. 6 in.
'. i .s
: : "*
(avni
-OV) axaNij^ aaj
aiy aaa^ .lo 'i^ "QO
00 •* O O 05
us CO t^ t-^ 00
05 00 t^ t^ CO
CO CO CO CO CO
00 ,-( rH O) t^
CO CO 05 00 05
00 05 00 00 t^
CO CO CO CO CO
382.7
383.9
387.4
•HHv J saaaoaQ
aiy JO aaaxvaadwax
O O O O CO
IN (N IN IN .H
00 00 00 00 00
us CO CQ CO o
00 00 00 00 00
81.0
80.5
80.5
iONaioMiia xNao
aaj oiaxawQiOA
C3 us O <S us
ei r^ d d T^
05 05 05 05 05
us C us us us
.-I (N ^ ^ d
05 05 05 O 05
91.2
90.2
91.0
(xNawaovidsiQ)
axaNij\[ aaj
HIV aan^ .jo "xj -qq
■* US CO CO US
00 (N CD CD C-i
^ ^ CO CO ^
us T«i US us CO
IN 00 Cq IN CO
tJ^ ^ ^ ^ CO
Tt* Tjt ■* ■* •*
436.6
442.5
442.5
awij,
o o
O CO o o o
-1 -1 'H ^ (N
o o o o o
CO O CO O CO
IN CO CO ■* ■*
o o o
O CO o
us lis as
HaawQU
iH (N CO •* us
CO t^ 00 05 o
rt (N CO
>H i-l iH
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
327
iDHSIOIAA^ ONIdRQJ
a
£
<5§
e »
H S
0^ M
0
OQ
2S-ON
08 'ON
ZI ON
28 "ON
OS 'ON
ZX -ON
•d -H -M. i«>ox
Z8"ON
08 "ON
0> O CO U5 ■*
00 0> 0> OS o
-< — I — c ,-1 (N
^ lO i-H t-( lO
t^ 0> t>. CO lO
O 00 O OS 03
lO «o rt ^ m
U3 cc u}
O 0> 03
o»
o
^
^
to
to
oj
^
^
to
o>
to
fft
1^
t^
</l
r/1
t^
to
t^
o
o
o
o
o
o
o
t~
00
00
t^
1^
r-
00
1^
t^
t^
00
t^
t»
t^
o
o
r^
r^
r-
o
r^
1^
t^
o
I~-
r>.
1-
r-
t^
i^
r^
r^
o
o
o
o
o
o
o
o
o
o
o
o
o
•-H
»— '
1—*
r~*
f—*
r^
o
05
o»
a>
Oi
in
05
05
05
lO
05
o>
o>
to o
G5 a
to to
00 t^
(N <N
O lO
00 00
05 Oi
t^ 00 r^ t~. r^
00 00 00 00 00
00 r~ t^
(N <N N
lO o o
lOioioioic tomiO"5Ti<
N rt 05 00 (N
CO CO <N IN N
NCS— lOCO 05(N05.-<0
oo 00 00 00 r~
LO lO
il O -H (N rt
t^ 00 r» CO oo_
lO _ »o
o> o i-H e^ (N
© to t- r^ to
Til IQ lO lO lO
<N CQ M CQ N
to to t^ r- to
G> 0> 0> Oi 0>
SI 'ON
aaaj JO -i j -no OOI
[ aiQNip^ saj SNoaavo
« M
Z8
•ON II^Ai %Y
in e-j o o (N
o o o o o
(N <N C^ (N Cl
IN m O Q N
o o o o o
IN IN (N IN IN
IN M IN
08
■ON IPM »V
CI lO lO OJ IM
o o o o o
(N (N IN (N IN
(N in IN IN <N
o o o o o
IN IN (N (N IN
(N IN N
ZI
•ON IPAi ?V
OO XI CC CO CO
N (N <N W !N
CO 00 00 OO CO
CO (N IN (N <N
00 00 00
N IN CI
J3IOO0J3}a|
ic lo in lO lO
«o to to to to
in in in in in
to to CO to to
in in lo
to to to
t^ lO in m in
in r^ in in in
in in in
9Sj'eT{08IQ
OUVJJ
aiQNiw HS£ oiaij .JO 'Xjj -no
axQNij\[ Haj SNoaiTj) iviox
ZT •ON
ininci'-i-'i' o)ino5co>H ,-ico>h
in in lo in ^
'n< Tjl to 00 «5
to to to to t-
Tii in ■* m in
O Tjt 05 to us
t^ to to to to
in •* in ■*
in in m •* d
t^ t- t>. r- t»
coco'coeoeo coeocococo
t- i> t- t» t^
0000000000 0000000000
00 00 00
ci ci N
Z8 -ON
o o o o o
cO CO CO CO CO
O O O O O
CO CO CO CO CO
O O O
CO CO CO
OS •ON
N N N CI N
CJ CI Cl CI Cl
C< N N
CO CO CO CO CO
CO CO CO CO CO
CO CO CO
C» N (N N IN
N IN CI N IN
CI « N
05 05 O 05 O)
(avQoov) aiiNij^
Haj HIV aaaj .io -xj -qo
•HHyj saaHoaci 'Hiy
aHVXNi J.O anniVHadwax
fKvajv)
iONaiOLJjg oiHxawQqoA
(XNawaovidsiQ) axQNij^
aaj arv aanj jo 'xj 'qq
awix
Haana|i~
to to to to t-
t^ r^ r^ t^ t~
b* F^ to to to
r- t^ t^ t~ w
S S K?
r^ t>« t*
I o • • •
; 03 • _• •
: : :
05 05 CI ■* 00
to <3> to CI •*
•* CI ■*
05 05 O CI •*
rt -H -H O 00
^- r~ t-- t^ to
rt rt CI CI fH
^i-HTeTco •* lo
o r^ CO t^ t^
11 N
to t^
328
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
o
g
CO
o
►J
(M
w
•<
b
fa
n
>
03
2
M
0
n
fa
H
H
a
0)
ft
CO
0
!>,
»
PL,
OJ
h3
P
PQ
O
o
■xl
S
H
<
iONaiOlJjg ONHWQfJ
"«< ic r- «c N
t^ r~ O IN 1-1 CO 1
CO M O -H O
-H -H rH -1 rH
T-H 1-1 CO CO CO d
XNHO
aaj 'aoNaoHawaag
00 00 CO CO CO
00 00 (n' (N (M
CO CO •* •* ■*
CO CO 05 00 00 00
(N N 00 00 00 00
■* ■* CO CO CO CO
laa^
aoNaoaawaag
00 00 CD «D to
(N (M 00 00 CO
CO CO CO CO CO
00 00 CO CO CO
lO lO CD «0 to
CO CD CO 00 00 00
00 00 111 IN C^ IN
CO CO ■* CO CO CO
CO CO 00 00 00 00
CD to 1(5 X5 X5 lO
1
•w •<! -a
00 •* CO O IM
00 05 O O 1-t
-1 rt rH
00 O (N ■* ■* •*
OS O OS OS OS »H
•d -H "V
CO CO •^ "5 in
CO t^ «5 ■* CO iH
lO 1(3 03 1-1 ■*
t^ t> 00 00 O
(N 1-1 r^ to to i>
00 CO r- t^ t' OS
•d -H -M
1-1 e^ to lo to
O O 0) O 05
to CO ^ 1-1 O O
c» OS d <d d d
•J -H •!
OO CO 05 00 t^
^ co' 00 00 lO
05 05 O 05 *-<
1-1 I-l
!N lO X5 CO CO 1-1
1-1 1-1 in CO CO t^
O O OS OS OS o>
avajj ONidivnj
(N C^ ■* >* •*
r~ t^ 1-1 1-1 I-l
Tf rjl t^ N N (N
rH 1-1 in t- t» t>-
CO CO CO to CO
^ —I CO CO CO
OS 03 00 00 00
CD CO in to to CD
CO CD 1-1 ,-1 rH rt
00 00 OS <33 OS 05
Air Pressures
IPAV *V
iC lO ■*■<*' -"J"
■* Tji o >n m in
(N (N t^ f^ t^
lO lO 1^ ■* Tl<
(N IM IN e^ (N
t^ t- CO C^ C<1 N
Tj< •* m m in in
IN IN IN (N (N N
jajoooja^nj
o o o o o
00 00 00 00 00
O O (N <N (N IN
00 00 00 00 00 00
aSjBqosja
o o »o lo lo
to CD m lO IC
IN M IN (N (N
in in o o o o
in in CO to to CO
IN N IN IN (N (N
oixvy
lO 1(5 Tjt lO OS
iH in OS 05 ■* o
-H Tl< CO 00 (^
CO CO t^ CD 1>
05 00 (N ■* m o
CD CO CO to to 00
axamp\r aa<j
aiaaj jo -xj 'qq
t^ rt h- ■* 1-1
U5 CD lO to CD
t^ t^ t> 1^ •* •*
in in in in in in
US U? K) KJ US
in in in in in in
' aXQNIJ\[
aa<j SK[0^^v][)
i
U5 CD lO
r^ OS t^ ^ 05
CD CD
r~ t~ t^ 1^ •<i< ^
•WO 'aia^y^ h3ao
aIa^^^I .■lo iixjhq
-1 C^4 -H O IN
t^ t^ ^~ t» t»
_ rt _ rt o o
t» t^ t^ t^ t^ t^
^^^^^ """"""
(avaxov)
axamj^ aaj aiy
aaaj ao uj -no
to Tft rt ^ CO
Tjt O K3 O lO
M (N c:s CD CD
■* to O CO CO
CO CO •* w ■*
(N in CO •* T)< ^
in in 00 o o c^
d i-< d IN M* CO
00 00 in CD CO >*
CO CO CO CO CO ■^
•UHVjj saaaoaQ
aiy 'SO aaoxvaadnax
I-l lO 00 <N t^
U5 b-.' t^ 00 00
1^ t^ t~ t^ t^
00 CO CO in N o
00 1-1 iH t^ CO in
t^ 00 00 t^ t^ W
AONaiow^a
oiaxawaao^
(N r- O o 1-1
ui •<ii U5 U3 CD
00 00 00 00 00
CO m in t^ t^ 1-1
in ■^ ■^ -^ -^ ui
00 00 00 00 00 00
(XNawaovidsxa)
axnNii\[ aa<j
aiy aaaj jo -xj -no
CO 00 IN O ■*
C« O ■<1< 00 00 00
CO 1-1 00 O CO
i-( ■* cs r~ (N
■* 1j< Tjt 1^ lO
01 o c^i -H rt in
CO r^ CO ■* ■«< CO
1
j awix
t^ <» 6s' o rt
u^ in in in m m
^ ^ ^ ^ ^ ^
1^ 1^ (N CO ■* in
!
aaftwntj
-H N CO •* «S
to t^ 00 05 O *H
1-1 1-1
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS
329
00 ^ Tl< N (O
to o a> <-> lO oi
t- r~ t^ CO t~
<N C<1 (N (M IN
n in ui t> 00 <o
M (N M M IN (N
i-< 00 es) M 64
CO CO CO M ^5 CO
CO CO CO CO CO CO
CO •* tH Tl< -^
'J" lO U5 lO ■* ^
"S CO o ■* ■*
CO CO CO CO CO
■* CO N CO us Tj«
CO CO CO CO CO CO
00 00 M CO CO
OJ 05 CI t^ t^
C^ M C» N W
« N (N N N «
i^ t^ CO r^ CO
CO CO ■* CO Tji
e^ (N (N (N (N
CO O OS -< T»< 1-1
M ffl lO ^ CO CO
(N (N N (N N N
O O N O N
<N eo CO <N O (O
t^ CO CO CO CO CO
US CI lO CO 00
00 a> o> r^ 00
•* ^ 'I' ^ •«ji
-H « t» CO CO e»
5 S S 5 5J S
r»
in
CO
t~
^
CO
00
us
C4
00
t-
a
o
f_H
o
^H
,^
in
in
,_,
nn
lO
la
t~
CO
CO
CO
us
(O
N
CO
(O
<o
CO
to
N.
r».
t^
(O
CO
CO
w
'^
'-'
"-I
rH
o
o
o
1-1
CO
lO
U5
>o
lO
>n
r^
r-
r^
o
uo
»
Oi
Oi
05
on
00
CO
d>
a>
OS
Oi
o
o>
c»
a>
03
o>
OJ
a>
la
CO
o
o
o
o
lO
US
lO
o
o
CO
00
M<
■*
•<J<
■*
r^
r^
r^
•^
IM
(N
M
N
(N
(N
M
(N
(N
(N
N
o
o
o
o
o
O
o
o
o
o
o
00
00
00
00
00
00 .
o
CO
o
o
o
o
us
in
in
o
(N
IN
N
w
(N
C'J
<N
w
(N
a>
o
CO
a>
00
CO
w
C4
US
«
**<
rf
■*
U5
•*
U5
^
M
00
US
•*
00
CO
CO
CO
CO
CO
•*
CO
CO
CO
CO
CO
^
^
Th
00
00
^
^_^
f/l
^
1<
1
t^
00
00
r-
00
w
00
CO
(O
CO
CO
CO
o
CO
o
CO
CO
CO
CO
CO
US
00
C4
O
FH CO
■H CO
00
C4
K
f2
■*
t^
IS
CO 00
t^ CO
1^ t^
04
CO
CO
t»
CO
t^
rt
us us
CO CO
lO
to
00
CO
00
us
00
CO
00
s
CO
00
CO CO
00 00
us <o
00 00
s
o
o
'Jf
o
<*
■«J<
<N (M
TJ. O
(M
IM
IM
o>
(M
00
i
CO
O O
CO CO
r- IM
o »
(M Ci
O
CO
o
(N
s
o
o
o
<M
o
N a
N <N
s
t~
00
a>
o
i-H
esi
•H evj
CO ■««<
UJ
-<e4co^us cot^oooO'H
330 DISCUSSION
DISCUSSION
F. A. Halsey. The air-lift pump is one of those things of which
our knowledge is almost exclusively experimental. Its analysis pre-
sents such serious difficulties that rational equations for the perform-
ance of the apparatus have not been derived by anyone to my
knowledge. Under these circumstances, we are reduced to experi-
ment for the determination of the fundamental laws of the perform-
ance of the apparatus, and these experiments should cover a wide
range in the conditions which lead to variation in the performance
of apparatus, namely, depth of submergence and lift.
2 The most complete experiments now on record were presented in
a paper before the British Institution of Civil Engineers in 1906,*
these experiments being of sufficiently wide scope to supply a satis-
factory guide for the design of these pumps under a considerable
range of conditions.
3 One curious feature attends the behavior of the air-lift pump and
the Taylor hydraulic air compressor, which are essentially the same
apparatus, reversed in action. In each case we have a pair of vertical
pipes communicating at the bottom, one filled with water and the
other with a mixture of air and water — a sort of suds. If the pipes
are of indefinite length the columns will take levels corresponding with
their respective gravities, but if the suds pipe is cut off below this
level it will overflow, the suds will rise, and we have an air-lift pump;
while on the contrary, if the water pipe is cut off below this level, it
will overflow, the column of water will rise, and we have a Taylor
hydraulic air compressor. The point to which I refer is that while the
first of these constructions has a low, the second has a high efficiency.
The highest figure given in this paper for the efficiency of the pump is
about 28 per cent, and while 40 per cent has been reached in exceptional
cases, the average is probably not more than 20 per cent. On the
other hand, the efficiency of the hydraulic air compressor is in the
vicinity of 75 or 80 per cent.
Dr. Sanford A. Moss. The method used by the author for computa-
tion of volumetric efficiency is subject to serious inaccuracies, as he
points out. For the purpose of the tests in question, where relative
rather than absolute volumetric efficiencies are desired, the method
is quite proper. The volumetric efficiencies given must be understood
to be relative and not absolute, however, as the errors may be quite
♦Abstracted in the American Machinist, August 16, 1906.
COMPRESSED AIR PUMPING SYSTEMS OF OIL WELLS 331
serious. I have used the method which the author states he would
have preferred for ascertaining volumetric efficiency, Par, 23, and
found it very good. Everything operated as would be expected and a
reputable value of volumetric efficiency was obtained. In one case,
the method as given in Par. 23 was checked by use of calibrated ori-
fices and also by use of a single tank with rise of pressure, the rate
of increase of pressure when the pressure reached the rated value
giving the desired volumetric efficiency. All three methods checked
very closely. The volumetric efficiency in this case was 58 per cent,
showing the serious losses which can occur due to the inaccuracies
mentioned.
J, G. Callan.* Par. 21 to Par. 23,* with diagram (Fig. 7), refer to
the method of determining the efiiciency of the air compressor, and
Mr. Ivens states in Par. 23 the method which he would have preferred
for determining this quantity.
2 The method of pumping up pressure in a large tank is open only
to the objection that the large tank is rarely obtainable, particularly
when the pressures which it is proposed to use are high, and the vol-
ume large. A more available and almost equally exact method is
afforded by the use of calibrated orifices of which the discharge coeffi-
cient has been duly determined. These orifices can be used in con-
nection with well-lcnown formulae, which involve a knowledge of
corrected pressure and temperature of air delivered to orifice, and of
velocity and temperature of emergent jet, all factors which can be
readily and accurately determined.
3 The losses of a compressor are so considerable and so complex
that the assumption of volumetric efficiency from the indicator card
may be in error by extremely large percentages and is quite as likely
to be misleading as enlightening, unless results are strictly compara-
tive rather than absolute. The value of the present tests would appar-
ently be somewhat affected by the errors in this method of determin-
ing delivery, since certain elements of compressor loss would assume
different percentage values with different discharge rates.
4 This brings up the desirability of some well-understood stand-
ardized method of testing compressors and other apparatus for defiver-
ing air. Commercial loss is certain to arise from the considerable
variations in terminology, and a definition of terms by an authorita-
tive body such as this Society would be of marked advantage. For
example, volumetric efficiency undoubtedly should mean the ratio
' General Electric Co., Schenectady, N. Y.
332 DISCUSSION
between the air actually discharged (when restored to standard atmos-
pheric conditions of temperature and pressure) to the air which would
have been discharged had the compressor cylinder been completely
filled on each stroke with air under standard atmospheric conditions,
and had all the air displaced been completely discharged. The exact
method of determining the approximation of real to theoretical
discharge obviously should be only a secondary consideration in such
a definition, but could properly be taken up as corollary.
5 The segregation and determination of losses in air compressors
of ^iifferent types form an interesting and useful line of investigation
upon which little has been pubhshed. A considerable amount of
work with which I am familiar indicates that these losses are usually
materially greater than the customary assumptions, particularly in
compressors which have been in service for some time. Various
methods h.'ive been devised for independently estimating loss due to
heating of intakc; leakage of inlet, discharge valves and piston, clear-
ance loss, throttling, and losses due to improper setting of mechanical
valve gears, but usually these determinations will involve more labor
than a direct measurement of air output by orifice and impact tube.
It is my belief, however, that the computation which neglects all of
these losses except the clearance loss, particularly where the pressure
is high and the compressor is somewhat worn, is almost sure to be
very gravely in error.
The Author. The discussions of the paper on oil-well pumping
presented at the Washington meeting were centered on the method
of ascertaining volumetric efficiency of air compressors. The object
of the paper was rather to compare the relative efficiencies of various
air-lift systems when being operated on extremely high pumping
heads. However, as all the discussions were agreeable to the writer's
views and methods, there is no room for argument.
No. 1244
THE SPECIFIC VOLUME OF SATURATED
STEAM
By Prof. C. H. I'eabody,' Boston, Mass.
Non-Member
For many years the specific volume of saturated steam has been
computed from the thermodynamic equation
AT dp
dt
in which the quantities have the following significance:
s is the specific volume, for example the volume in cubic
meters of one kilogram.
r is the heat of vaporization in calories.
A is the heat equivalent of a unit of work.
T is the absolute temperature obtained by adding 273 to the
temperature by the centigrade thermometer.
-^ is the differential coefficient of the pressure with regard to
the temperature, the pressure being in kilograms per
square meter.
o is the specific volume of water i,0.01 cubic meters per kg.)
2 For this paper French units are used because the original data
are given in them and comparison with experimental values is con-
venient.
3 All the quantities entering into this equation are now deter-
mined with a certainty and precision that must be considered satis-
factory for engineering purposes and a comparison with experimental
determinations of the specific volume shows an exceptionally good
concordance.
' Professor Naval Architecture and Marine Engineering, Mass. Inst. Tech.
Presented at the Spring Meeting, Washington, May 1909, of The Amebican
Society of Mechanical Engineers.
334 SPECIFIC VOLUME OF SATURATED STEAM
4 To make the exposition of this statement clear it is necessary
to review the experimental data and to state the jDrecision that can
properly be attributed to them.
5 The mechanical equivalent of heat as determined by Rowland'
may be taken as 427 meter-kilograms (778 foot pounds) at 15 deg.
cent., which corresponds nearly with 62 deg. fahr. There have been
more recent investigations which on the whole confirm this result,
though there is some indication that it is a trifle small. The uncer-
tainty may be one in a thousand or one in two thousand.
6 Callendar^ gives for the absolute temperature of freezing point
273.1 deg. cent., with a probable error of one in two thousand.
7 For the range of temperature from 30 deg. to 100 deg. Henning^
gives the equation
r= 94.210 (365 - t) "■^''^'^
in calories at 15 deg. cent. In English units the equation may be
written
r= 141.124 (689 - t) ^-^^^^
Experiments by Dieterici,^ Griffiths^ and A. C. Smith^ confirm his
results and extend the equation to freezing point. The probable
error of this equation is one in one thousand.
8 In his paper, The Total Heat of Saturated Steam, read at the
Annual Meeting, 1908, Dr. Harvey N. Davis gives for the total heat
of steam from 212 deg. to 400 deg. fahr.
H = 7/212 + 0.3745 {t - 212) - 0.000550 (t - 212)-
Transformed into French units this may be written
H = 638.9 + 0.3745' (t - 100) - 0.00099 (t - lOOy
provided that the constant term be taken as the sum of Henning's
value for r at 100 deg. cent, and the heat of the liquid be taken as
100.2, according to a consideration to be taken up later in this paper.
To conform with the conditions already accepted, this equation
should give the total heat in calories at 15 deg. cent., while Dr. Davis
used for the calories 1/100 of the heat required to raise one kilogram
^ Proc. Am. Acad., vol. 15 (n.s. 7), 1879.
' Phil. Mag., Jan. 1903.
'Annalen der Physik, vol. 21, p. 849, 1906.
* Annalen der Physik, vol. 16, p. 912, 19Q5.
•Phil. Tians., 180, p. 261, 1895.
' Physical Review,vol, 25, 1907.
SPECIFIC VOLUME ©F SATURATED STEAM 335
of water from freezing to boiling point. The difference amounts to
2/1000, as indicated by the heat of the Hquid just mentioned
(q = 100.2). Now the total heats at 100 deg. and 200 deg. cent,
are 638.9 and 666.5, and their difference is 27.6 calories, so that
the total effect is less than one-tenth of a calorie.
9 As for the heat of the hquid we have the three following sources
of information:
a Barnes" determinations of the specific heat of water from 0
deg. to 95 deg. cent.
b Dieterici's^ determinations of the same property from freez-
ing point to very high temperatures.
c Regnault's^ determinations of the heat of the liquid.
Barnes' experiments were made by an electrical method for which
great relative precision is claimed, and they showed a good concor d-
ance with RoAvland's work on the mechanical equivalent, which in
reality was an investigation also of the specific heat. Dieterici's
investigation consisted essentially in heating water in a quartz tube,
which was then transferred to the ice calorimeter. His results appear
to be systematically larger than Barnes'; at 95 deg. cent., the dis-
crepancy is y\ of 1 per cent.
10 In 1907 the author endeavored to join Regnault's values for
the heat of the liquid to those deduced from Barnes' values of the
specific heat. Now Regnault's experiments consisted in running hot
water into a calorimeter partly filled with cold water and noting the
rise of temperature in the calorimeter. There were 40 tests in all,
cattered irregularly from about 100 deg. to 190 deg. cent, for the
temperature of the hot water; there were in a way three groups of
tests, one near 110 deg., one near 160 deg., and the third near 190
deg. cent.
11 The average rise of temperature in the calorimeter for the first
group was not far from 9 deg. cent., which item appears to account
for the considerable irregularity of results at that place. The
experiments with the highest temperatures had nearly twice that
rise of temperature in the calorimeter and about half the dispersion
of results.
12 In order to use Regnault's results his values for the heat of
the liquid were recomputed, allowing for the true specific heat of the
water in the calorimeter, and then a diagram was plotted as shown
» Phys. Review, vol. 15, p. 71, 1902.
* Annalen der Physik, vol. 16, p. 593, 1905.
* Memoirs de Tlnstitut de France, vol. 26.
336
SPECIFIC VOLUME OF SATURATED STEAM
by Fig. 1, in which the abscissae are temperatures and the ordinates
are values oi q — t.
13 This allows of the use of a large vertical scale which much
accentuates the apparent scattering of points. A curve was then
drawn to join a curve from 0 deg. to 100 deg. cent., from Barnes'
results for the specific heat of water. This curve passes near the
highest group of points, above the middle group and below the lowest
group.
14 It should be said that Barnes' results were fii'st transformed
to allow for the use of 62 deg. fahr. for the standard temperature,
instead of 20 deg., which he had taken in his report; also that his
-
' /
v7
—3.0
7
-
>
^
-
/
—1.0
/•*.
-
t
.
•/ *. *
.
•* V
-
•;'
y
— OtO
-
100
150
200
- 1 1
1 1 1
1 1
1 1 '
1_
1 1 1
Temperature Centigrade
Fig. 1 Recomputation of Regnault's Experiments on the Heat of the
Liquid of Water
values were slightly increased at temperatures approaching 100 deg.
so as to avoid a break in the curve. The last had the effect of increas-
ing the heat of the liquid at 100 deg. by one one-thousandth.
15 Finally a table of specific heats was drawn off for temperatures
from 0 deg. to 220 deg. cent., which served as the basis of a graphical
integration for the value of g — ^. Fig. 2 gives the curve represent-
ing the final value of this quantity and also a curve representing
values that would be obtained if Dieterici's values for the specific
heat were accepted.
16 The author is of the opinion that the full curve in Fig. 2 shows
verj'- nearly the true value of the property under consideration, and
he has used it to determine heats of the liquid.
SPECIFIC VOLUME OF SATURATED STEAM
337
17 The maximum deviation of a single point from the curve in
Fig. 1 is 0.8 of a calorie, which amounts to | of 1 per cent of the
heat of the liquid at that point. If we could consider that an error
of 0.02 deg. might be attributed to the temperatures in the calorim-
eter it would account for one-third of that deviation. But to take
the most pessimistic view of the situation and charge an error of 0.8
of a calorie against the method, we may still consider that for tem-
peratures above boiling-point the heat of the liquid is always asso-
ciated with the heat of vaporization, and that their sum is more than
-4.80
—5.20
-1.60
100
150
300
Terupe-rature Centigrade
Fig. 2 Values op the Quantity (q-t)
THE FULL CURVE SHOWS THE QUANTITY DEDUCED FROM THE AUTHOR'S COMBINATION OP BARNES*
EXPERIMENTS ON THE SPECIFIC HEAT OP WATER WITH REGNAULT'S EXPERIMENTS ON THE
HEAT OP THE LIQUID, WHILE THE DOTTED CURVE SHOWS EBSDLT8 FROM DIETERICI'B BXPERI-
MENT8 ON THE SPECIFIC HEAT OF WATER.
630 calories, so that the deviation in this light amounts to J of 1
per cent.
18 A more just view is clearly to take the deviation of the worst
group of points. This occurs at 1 17 deg. and is about 0.3 of a calorie,
that is, 0.25 per cent of the heat of theUquid. The most favorable
view is to consider that the upper end of the curve is well fixed by Reg-
nault's experiments, which were then under the most favorable con-
ditions, and that the lower end is tied to Barnes' values, which have
all desired precision. This matter is discussed with some detail be-
338 SPECIFIC VOLUME OF SATURATED STEAM
cause the original experimental results needed to be entirely recast
for the present purpose.
19 But while important from some aspects, the quantities with
which we are dealing are not affected by uncertainties that concern our
main investigation, i.e., the specific volume of saturated steam, for
the maximum variation between the author's value for the heat of
the liquid, and a value determined from Dieterici's investigation,
amounts to 0.8 of a calorie at 200 deg. cent. This is only J of 1 per
cent of the total heat at that place. However, we need for our specific
volume the heat of vaporization, and the discrepancy then becomes
i of 1 per cent.
20 Recent determinations of the pressure of saturated steam have
been made by Holborn and Henning,^ with all the resources of modern
physical methods including the platinum thermometer. They claim
a precision of 0.01 deg. in the determination of temperature and that
25
20
15
10
5
1100 [110 |120_.--flS0 lUO H50 1160 ilTO IISO 1190 |200 i210 \
0 .
Fia. 3 Curve to Extrapolate Pressure op Saturated Steam to 220 Dpo
Cent.
their results reduced to the thermometric scale have a probable error of
not more than 0.02 deg. at 200 deg. cent. Their own experiments cover
the range of temperature from 50 deg. to 200 deg. cent. (122 deg. to
392 deg. fahr.), and they have extrapolated results to 205 deg. cent.
Below 30 deg. they have made use of experiments by Thiesen and
Scheel to extend results to freezing points; these experiments were
not made with the same degree of precision as those by Holborn and
Henning.
21 In order to extend calculations to 220 deg. cent., as has been
the habit in computing steam tables, the author made use of a dia-
gram shown by Fig. 3, in which the abscissae are temperatures centi-
' Annaleu der Physik, vol, 26, p. 383, 1908.
Note. Since these results may not be easily accessible, it may be of interest
to say that they have been transferred directly to Table 3, of the author's Steam
and Entropy Tables, edition of 1909.
SPECIFIC VOLUME OP SATURATED STEAM 339
grade and the ordinates are differences between Holborn and Hen-
ning's value and pressures computed by the following equation:
log V = 5.457570-0.4120021(9.997411296 - 10)* "^o^ +
(7.74168 -10) (9.997411296 - 10)* "^oo
which was chosen as a matter of convenience and because it gave a
curve which crossed the axis near 220 deg. cent, when produced. It
is thought that the extrapolated values are not much in error,
though there is no means of determining this question. Fortunately
this part of the range of temperature, as well as that below 30 deg.
cent., is not so important to engineers.
22 The degree of precision attained by Holborn and Henning in
the determination of the pressure of saturated steam is far beyond
any direct technical requirement, since pressures are seldom deter-
mined closer than one-tenth of a pound; it is, however, requisite, if the
differential coefficient -f is to be determined with certainty and
at
accuracy.
23 Since their results are presented in a table without attempting
to represent it by an equation, it becomes necessary to replace by
dt
J p .
-^ which can be most readily obtained as follows: for a given tem-
J t
perature, for example 100 deg., we may compute the ratio by taldng
two adjacent temperatures, such as 98 deg. and 102 deg., finding the
difference of pressure, which is to be divided by the difference of
temperature; and the result is to be multiplied by 13.5959, because
that is the pressure of one millimeter of mercury on one square meter.
This result is
^=13.5959^'"-^-^°^-^ =369.1
Jt 4
24 A number of elements entered into the determination to use
this method and to take an interval of 4 deg. If the relation of the
pressure to the temperature could be represented by a second-degree
curve, that is, if such a curve were a parabola with its axis parallel
to the axis of pressure, the ratio -f for any interval would be pre-
Jt
dv
cisely equal to ^. A table of values that could be represented by
such a curve would have constant second differences; by second
differences are'meant the results obtained by taking (a) the differences
340 SPECIFIC VOLUME OF SATURATED STEAM
of successive tabular values, and (6) the differences of these differences.
An examination of the second differences of Holborn and Henning's
values showed great regularity between 50 deg. and 100 deg., i.e., for
their own determinations. The second differences increased slowly;
for intervals of 4 deg. the increase was imperceptible, for 6-deg. inter-
vals the increase was barely perceptible, but for 10-deg. intervals
it was very apparent.
25 Now the possible precision of reading the height of a column
of mercury, including allowance for variations of density, is better
than the determination of temperature; consequently the prob-
able error to be considered is that attributed to the determina-
tion of temperature, namely 0.01 deg., consequently the probable
^ J)
error of a single determination of the ratio ~-^. To diminish the
At
effect of local variations this ratio was computed for each degree of
temperature and the regularity of the results thus obtained was tested
by taking first and second differences. Where the second differences
showed irregularity, the values of the ratio were changed to the
extent of 1/1000 in order to improve the regularity of the second dif-
ferences. This process is equivalent to drawing a smooth or fair
curve to represent physical properties obtained by observation.
Aj)
26 Having values of the ratio —^ for each degree of temperature
At
the specific volumes were computed by the thermodynamic equation
in Par. 1. They were in turn tested for regularity by taking first
and second differences: and again the values were changed when
necessary to the extent of 1/1000 to improve the regularity of the
second differences. The combined effect of both fairings is esti-
mated not to exceed 1/500 in any case and the author believes that the
probable error of the final determinations of the specific volumes is
not greater than that amount for the range of 50 deg. to 200 deg.
cent.
27 It may further be said that having computed the values of A'pu
at each fifth degree and plotted the results on a large diagram, no indi-
vidual values were found to vary from a fair curve more than 1/750.
28 Fortunately there are extant experiments on the specific
volume of saturated steam by Knoblauch, Linde and Klebe,* made
with such a degree of precision as to give a satisfactory check on the
computations made by the method described. These experiments
^Mitteilunyen iiber Forschungsarbeiten, vol. 21, S. 33, 1905.
SPECIFIC VOLUME OP SATURATED STEAM 341
consisted in measuring the temperature and pressure of superheated
steam at constant volume, and the results were so treated as to give
the volume at saturation by a sl.'tight-line extrapolation with great
certainty. The experimenters give the following equation to repre-
sent the properties of both superheated and saturated steam;
p V = BT - p (l + o p)
T
B = 47.10; a = 0.000002; C = 0.031; D = 0.0052,
volumes being in cubic meters per kilogram, pressures in kilograms
per square meter, and the absolute temperature being on the centi-
grade scale.
29 For English units the equation may be written
p V = 85.85 T - p {1 +0.00000976 p)
150,300,000 _ ^^333
the volumes being in cubic feet, the pressures in pounds per square
foot and the temperatures in degrees fahr.
30 Knoblauch claims for this equation a mean probable error of
1/500, though admitting individual discrepancies of twice that amount.
This equation applied to the computation of specific volumes of satur-
ated steam shows a good concordance with results, computed by the
thermodynamic equation, the greatest discrepancy being 1/300 at 165
deg. cent. (329 deg. fahr.).
31 Not satisfied with this apparent concordance, which after all
was with an empirical equation which on examination showed some-
what larger variation from individual experimental values at satura-
tion, the author had a diagram drawn of the 32 values of the specific
volume reported by the experimenters. The diagram was drawn to
a very large scale, using temperatures for abscissae and logarithms
of volumes for ordinates, and a fair curve was drawn by aid of a stiff
spline. From readings on this curve the volumes were determined
at 5 deg. intervals, and are set down in the accompanying table
together with values computed by the thermodynamic equation.
32 The greatest deviation of values in this table is 0.2 per cent,
which is precisely the probable error assigned by the experimenters
for their work. It may therefore be concluded that between the
limits of temperature in this table and probably from 30 deg. to 200
deg. cent. (86 deg. to 392 deg. fahr.), the probable error of computa-
tions by aid of the thermodynamic equation is not in excess of
1/500.
342
DISCUSSION
COMPARISON OF EXPERIMENTAL AND COMPUTED VALUES OF THE SPECIFIC
VOLUME OF SATURATED STEAM
0.
VoLDME, Cubic
Metebs
MPEKATURE
Volume, Cxtbic Meters
Experi-
Per cent
Experi-
Per cent
H
mental
Computed
deviation
^
mental
Computed
deviation
100
1.674
1.671
: +0.18
145
0.4458
0.4457
+0.02
105
1.421
1.419
+0.14
150
0.3927
0.3921
+0.15
110
1.211
1.209
+0.17
155
0.3466
0.3463
+ 0.09
115
1.036
1.036
i 0.
160
0.3069
0.3063
+ 0.20
120
0.8894
0.8910
-0.18
165
0.2724
0.2729
+ 0.18
125
0.7688
0.7698
-0.13
170
0.2426
0.2423
+0.12
130
0.6670
0.6677
-0.10
175
0.2168
0.2164
+ 0.19
135
0.5809
0.5812
1 -0.05
180
0.1940
0.1941
-0.05
140
0.5080
0.5081
-0.02
33 This conclusion carries with it the attribution of at least the
same degree of precision to all the properties entering into the ther-
modynamic equation. A little consideration will show that this con-
clusion covers all the properties given in steam-tables including the
entropy. As an apparent exception we have the heat of the liquid
at high temperatures which may be uncertain to the extent of ^ of
1 per cent of itself, but as that quantity is then associated with the
heat of vaporization the influence of such an error will be of no con-
sequence in computations.
34 It may therefore be expected that steam tables based on the
present information will have permanence.
DISCUSSION
Prof. William D. Ennis. The reason for the extrapolation of Fig.
3 is not quite clear to me. The ordinates of this diagram are differ-
ences between the Holborn and Henning values for the pressure of
saturated steam, and those given by the Peabody formula, log p =a —
,t-100
+ c/3
t-ioo
The diagram is extended to include temperatures
above 205 deg., the Holborn and Henning limit. Why would it not
be just as satisfactory, if the Holborn and Henning values are satis-
factory, to extrapolate directly the curve expressing their results?
2 There seems to be little room for uncertainty in any of the
properties of saturated steam, excepting, possibly, the heat of the
liquid. The maximum divergence in values for the former, comparing
the Dieterici and the modified Regnault values adopted by Professor
SPECIFIC VOLUME OF SATURATED STEAM 343
Peabody, occurs at the highest temperatures: at 220 deg. (an extra-
polated point) it is 1.31 cal. or 0.584 per cent. Now if Dieterici does
not claim an accuracy exceeding 0.5 per cent at this temperature, and
since Professor Peabody admits a possible fractional percentage of
error, the true value may be within the limit of estimated error in
both computations. The result of taldng Dieterici's values would be
to make the computed specific volume 0.01 per cent less at 100 deg.,
0.02 per cent less at 140 deg. and 0.045 per cent less at 165 deg., than
those tabulated by the author. The deviation from the Knoblauch,
Linde and Klebe results would be thus generally decreased. There is
still a possibility that the Dieterici values may be more nearly correct.
If the preponderance of error in values of the other quantities enter-
ing into the volume formula were in such a direction as to make the
computed volumes too small, the lower heats of the liquid used by
Professor Peabody might lead to an apparently better result because
of a balancing of opposite errors. It does not seem safe to say defi-
nitely that such may not be the fact. The uncertainty in the heat of
the liquid at 200 deg. is 0.36 per cent rather than 0.25 per cent. The
same uncertainty applies to the entropy of the liquid, and a possible
error of about 0.16 per cent to the entropy of vaporization. The
entropy of saturation at this temperature may then be wrong to as
great an extent as 0.23 per cent, or -gr • This would introduce a barely
noticeable error into computations involving vapor cycles.
3 It is questionable whether permanence, in a matter of this kind,
is as desirable as a standard, a flexible standard. Would it not be
within the scope of the Society to cooperate with national engineering
organizations abroad in the preparation of an international steam
table for saturation and superheat, embpdying the generally accepted
values and subject to modification whenever an undisputed conclu-
sion is reached on the one or two remaining doubtful quantities?
The Author. In reply to Professor Ennis I will first explain
that Holborn and Henning give a table of pressure for each degree
of temperature, and make no use of an equation except as a means
of fairing their results, for which purpose they chose Thiesen's
equation which gives divergent values for higher temperatures.
Now it happens that the equation which I chose gives values which
converge toward Holborn and Henning's values so that it is possible
to draw an extrapolation diagram as shown in my paper.
2 Secondly, I wish to say that Dr. Henning has kindly sent me
in advance of publication an abstract of results which he has recently
344 DISCUSSION
obtained for the heat of vaporization of water from 100 deg. to 180
deg. cent. His memoir will soon be available and will show that
the results which I have deduced from Dr. Davis' values of the total
heat, show a close concordance with these new experimental values.
It is not unlikely that we may have a conclusive determination of the
remaining quantity, heat of the liquid, but as stated in my paper the
concordance of all quantities involved in the computation of steam
tables is even now very satisfactory so that there is no reason to
anticipate any necessity for changing tables for engineers.
No. 1245
SOME PROPERTIES OF STEAM
By Prof. R, C. H. Heck, New Brunswick, N. J.
Member of the Society
The purpose of this paper is to present some recent experimental
results as to two of the fundamental thermodynamic properties of
water and steam, and to make certain comparisons between these
determinations and the older values used in our steam tables. The
two properties considered are, the relation between pressure and
temperature of saturated steam, and the specific heat of water.
THE PRESSURE-TEMPERATURE RELATION
2 This relation is, from the point of view of experimental deter-
mination, the simplest of the properties of steam, and with accurate
instruments and adequate skill can be very precisely measured. For
this reason, the results obtained by various experimenters differ by
relatively small amounts, and in discussing them we take up a ques-
tion in the realm of scientific accuracy rather than one concerning
effectively correct values for ordinary tex3hnical use. For certain
purposes, however, it is most important that this relation be truly
and accurately known.
3 In Annalen der Physik, 1907, vol. 22, p. 609 to 630, is published
a paper by F. Henning, On the Saturation Pressure of Steam, in
which are gathered together all the determinations that have been
made on this relation, from Magnus and Regnault down to that time.
These are compared by means of curves, which show, to a large scale,
their departures from an assumed standard of reference. This stand-
ard is the formula of Thiesen,
(t + 273) log / = 5.409 (t - 100) - 0.508 X lO"' [(365 - t)* - 265*]
760
where t is centigrade temperature and p is pressure in millimeters
of mercury. From the comparison and discussion the conclusion was
reached that up to 100 deg. cent, this formula is to be accepted.
Presented at the Spring Meeting, Washington, May 1909, of The American
Society op Mechanical Engineers.
346 SOME PROPERTIES OF STEAM
while above 100 deg. the determinations of Regnault are best — not as
set forth by his formula, but as worked over by Henning, from a
selection of his more reliable observations.
4 A new and very accurate determination by Holborn and Hen-
ning, over the range from 50 deg. to 200 deg. cent., is fully described in
Annalen der Physik, 1908, vol. 26, p. 833 to 883, in a paper, On the
Platinum Thermometer and the Saturation Pressure of Steam, while
in Zeitschrift des Vereins deutscher Ingenieure, February 20, 1909, is
given a brief presentation and comparison of results. Exceedingly
close agreement is shown between these new observations, the recom-
puted Regnault values, and the work of Knoblauch, Linde, and Klebe
— see Table 3 in Zeitschrift article. The final result is a table giving
p for every degree from 0 deg. to 205 deg. cent., which follows Thiesen's
formula up to 50 deg., and embodies the authors' work from that point.
5 This table is here reproduced in Table 1, but with pressure con-
verted to pounds per square inch and interpolated for every degree
fahrenheit from 32 deg. to 402 deg., or to just past 250 lb. abs. Later
the writer hopes to extend this table, carrying forward the line of the
Holborn-Henning determination in comparison with the observa-
tions of Regnault and others. This can be done even up to a pressure
of 1000 lb. with sufficient accuracy for all practical purposes.
6 In the work of conversion and interpolation, it was necessary to
carry the numbers to a higher degree of apparent accuracy, or to use
more significant figures, than any experimental precision would call
for. Without a mathematical formula, a function of this sort can be
carried forward only by carefully smoothing out the differences until
those of the second order follow a continuous rate of change. In this
operation, the first differences were brought to a sufficient degree of
smoothness to furnish effectively accurate values of the rate of change
dx)
of pwithf; and this differential coefficient, —is also given in Table 1.
It may be considered absolutely correct (as a derivative) within about
four or five units in the last place, while as between successive values
the closeness is much better. This is less precise than might be
desired, but it is accurate enough for use in calculating specific volume,
since the thermal data there involved are not of any greater degree of
reliability.
7 In Fig. 1 is given a comparison between the pressures in Table 1
and some hitherto generally used values. The base is temperature
fahrenheit, the ordinate the difference between the other value of p
and that in Table 1. Curve 1, for the range up to 225 deg. fahr., is
SOME PROPERTIES OP STEAM
347
3 +
CL/
•/
<-♦ *
*
A
S-,
»/""
cc
s
!d
*
X
■^
<f
'X
tt
f ,
^H
S.
\
^
^
a^
^
\
\
V
\
\
V
-^
\
^
\^
\
•
■
^
/
/
"/
r
P z
+ I
2 Q. 3
348 SOME PROPERTIES OF STEAM
drawn to the large scale at the left, and shows how Regnault's formula
drops below the new determination. The curves at 2 have the ordi-
nate scale at the right, only one-tenth as large as that for 1. The
letter R marks the "standard" Regnault curve, here plotted from
the table in Roentgen's Thermodynamics, which happened to be the
most convenient in its manner of expression: note the abrupt change
at about 380 deg. fahr. Curve P shows Peabody's values, which are
based on Regnault, but with revised computations, and depart quite
decidedly from the older table above 325 deg. The scattering of the
points above that temperature is due to the coarseness of numerical
expression, Peabody giving but one decimal place for the higher pres-
sures. The curve is simply sketched through this band of points.
8 Holborn and Henning do not attempt to devise a formula, but
base their table on a method of graphical interpolation. It will be
noted that Curve 1 shows a faint waviness, indicating some departure
from perfect mathematical smoothness; but the extreme smallness of
the irregularities is really a proof of the skill with which the original
interpolation was made.
THE SPECIFIC HEAT OF WATER
9 In Fig. 2 are plotted several important curves for the specific
heat of water — the true or instantaneous, not the mean value.
Curve R shows Regnault's formula, which in fahrenheit units is,
c = 1 + 0.0000222 (t - 32) -|- 0.000000278 (t - 32)^
This curve differs radically from the newer and true determination of
the specific heat over the lower part of the range, as shown by the
other curves.
10 Curve B represents the experiments of H. T. Barnes and
associates; these are described briefly in Proceedings Royal Society,
1900, vol. 67, fully in Phil Trans. Roy. Soc, 1902, vol. A 199; while
in Physical Review, 1902, vol. 15, there is a description of the
determination on supercooled water, which was carried to — 5 deg. cent.,
and the tabulated values for the whole range up to 95 deg. cent. The
body of this work was done by a continuous method, water flowing
through a small tube and absorbing heat which was electrically sup-
plied and measured; for the range below freezing, a method of mixing
was found necessary.
11 Curve P, which begins at 140 deg. fahr., shows the values used
by Peabody above this temperature; below it he accepts the work of
Barnes. Peabody's line — it is almost straight — is based on Reg-
SOME PROPERTIES OF STEAM
349
350 SOME PROPERTIES OF STEAM
nault's experiments: but it hardly seems reasonable to make c thus
an almost straight-Une function of t.
12 Curve D shows the very important experiments of Dieterici,
described in Annalen der Physik, 1905, vol. 16. In these a small
body of water, pure and free from air, was sealed in a tube of quartz.
This little cartridge was heated to a certain desired temperature,
then dropped into a Bunsen ice calorimeter, where the heat given off
in its cooling to 0 deg. cent, is measured. The highest temperature
reached was about 300 deg. cent. The drawback in tliis method is
the relatively large heat capacity of the quartz tube, which has to be
very carefully determined. From 100 deg. fahr. upward, Dieterici
finds that Ms results conform very well to a parabolic equation like
that of Regnault, which for fahrenheit units has the constants,
c = 0.99827 - 0.0000576 (t - 32) + 0.00000064 (t - 32)^
Below 100 deg. fahr., tabulation from graphical interpolation is pref-
erable to expression by formula. A numerical comparison of the
several curves is given in Table 2.
DIFFERENT HEAT UNITS
13 Before discussing these data, something must be said as to the
unit of heat measurement. Regnault intended to use the heat
capacity of water at 15 deg. cent, as the heat unit — in other words, the
15-deg. calorie — but it was not until long after his time that the true
manner of variation of the specific heat over the lower range of
ordinary temperatures was either clearly perceived or accurately
measured. Barnes' values are based on unity at 16 deg. cent., and
it will be noted that the B curve on Fig. 2 crosses the base-line at
just about 16 deg. cent, (the two short vertical cross-lines near 60 deg.
fahr. are at 15 deg. and 16 deg. cent.). The now generally used
numerical values of the mechanical equivalent of heat, 427 m-kg. or
778 ft. lb. are based on a heat unit at 15 deg. cent or 59 deg. fahr.
14 Dieterici's results are expressed in the mean calorie, which is
one one-hundredth of the heat required to raise 1 kg. of water from
0 deg. to 100 deg. cent.; and his specific heat values check up to an
average of unity over this range. Graphically, on Fig. 2, his curve
cuts the 15-deg. cent, ordinate at 0.0012 below the unity base-line.
In a special expei iment, with electrical measui ement analogous to that
used by Barnes, he made the mechanical equivalent of the mean
calorie bear to our standard Rowland value for the 15-deg. calorie the
SOME PROPERTIES OF STEAM 351
ratio of the numbers 419.25 to 418.8, or 1.0011 to 1.00000. Dis-
regarding some micertainties which may exist in the minds of physi-
cists as to the finality of this determination, it seems reasonable, for
engineering purposes, to use this 0.0011 or 0.11 per cent correction
in order to change from one system of units to the other.
15 The amount of attention here paid to this small point is justi-
fied by the importance given to it through the introduction of the
mean calorie to the Society in the recent paper on The Total Heat of
Saturated Steam, by Dr. H. N. Davis. Personally, I think we had
better transform heat values in this unit by means of the ratio just
offered, rather than change our mechanical equivalent of heat from
778 to 778.9.
16 Now the specific heat is the ratio of a certain absolute quantity
of heat to an assumed unit quantity. If we use a larger unit, the
ratio will be smaller, and vice versa. Assuming that the mean calorie
is 1.0011 of the 15-deg. calorie, we change Dieterici's values to the
15-deg. unit if we increase them by 0.11 per cent. This would raise
his curve to the dotted position on Fig. 2, and change his formula to
c = 0.99938 - 0.00005766 (t - 32) + 0.0000006407 (t - 32\)
SPECIFIC HEAT OF WATER — CONCLUSION
17 It is pretty safe to say that the Holborn-Henning results for
pressure and temperature, set forth in Table 1, are final, and that
this relation is now known surely and accurately enough for all pur-
poses of practical science. But in regard to the specific heat of water
we are yet confronted by one of the annoying uncertainties which
have so long surrounded many parts of this subject. Dieterici
claims an experimental accuracy ranging from 0.1 per cent at low
ranges to 0.5 per cent at high ranges of temperature; but his method
is open to the objection that two heat-capacities have to be measured
and their difference used.
18 In spite of some small doubt as to the accuracy of Dieterici's
results, and a faint suspicion that his curve may rise too rapidly,
I am of the opinion that his determination is to be accepted instead of
Regnault's. Further, the idea of an increasing rate of increase in c,
as expressed by a second-degree equation, seems to be far more rea-
sonable than that of a nearly constant rate of increase.
19 It is hardly probable that the heat capacity of water will ever
be so accurately determined that the heat for the external work of
expanding the water will be more than a small fraction of the prob-
able error in heat measurement.
352
SOME PROPERTIES OF STEAM
TABLE 1 THE PRESSURE-TEMPERATURE RELATION
t \ p '• dp/dt
« ' p dp/dt t
p j dp/dt t
p dp/dt
76 0.4433 0.01467 121
77 0.4582 0.01510 122
1.7362 0.04815' 166
1.7849 0.0493 167
5.459 0.1277
32 0.08860.003575
5.588 0.1302
33 0.09220.00371
78 0.4735 0.01554 123
1.8348 0.0505 168
5.719 0.1327
34 0.0960(
). 003845'
79 1 0.4893 0.01600 124
1.8859 0.0517 169
5.853 0.1353
35
0.0999(
). 003985'
80 J 0.5055 0.01646 125
1.9382
0.05295 170
5.990 0.1380
36
0.1039
J. 00413
81 0.5222 0.01694, 126
1.9918
0.0642 1 171
6.129 0.1407
37
0.1081
3.00428
82 0.5394
0.01742 127
2.0466 0.05545 172
6.271 0.1434
38 0.1125
3.00443 1
83 0.5570
0.01792 128
2.1027, 0.05675 173
6.416 0.1462
39 0.1170
0. 004585;
84 0.5752
0.01844 129
2.1601 0.0581 174
6.564 0.1490
40 0.1217
0.004745
85
0.5939
0.01898 130
2.2189 0.05945 175
! •<
6.714 0.1519
1
0.00491
86
0.6132
0.01952 131
2.2790 0.0608 176
6.868 0.1548
42 i 0.1315
0.005075
87
0.6330
0.02008 132
2.3406 0.06216 177
7.024 0.1677
43 0.1367
0.00525
88'
0.6533
0.02065 133
2.4033 0.06355 178
7.183 0.1607
44 0.1420
0.00543
89
0.6743
0.02123 134
2.4675 0.06496 179
7.346 0.1637
45 [ 0.1475
0.00561
90
0.6958
0.02182 136
2.5332 0.06646 180
7.511
0.1668
46
0.1532
0.00580
91
0.7179
0.02243 136
2.6004 0.0680 181
7.679
0.1699
47
0.1591
0.00600
92
0.7406
0.02305 137
2.6692 0.0696 \ 182
7.850 0.1730
48
0.1652
0.00620
93
0.7640
0.02368 138
2.7396 0.0712 : 183
8.025 0.1762
49
0.1715
0.00641
94
0.7880
0.02432 139
2.8116
0.0728 184
8.203 0.1794
50
0.1780
0.00663
95
0.8127
0.02498. 140
2.8861
0.0744 185
8.384 0.1827
51
0.1847
0.00685
96
0.8380
0.02566
141
2.9603
0.0760 186
8.568' 0.1860
62
0.1917
0.00708
97
0.8640
0.02635
142
3.0371
0.0776 187
8.756 0.1894
53
0.1989
0.00731
98
0.8907
0.02705
143
3.1155
0.0793 188
8.947 0.1929
64
0.2063
0.00754
99
0.9181
0.02776
144
3.1956
0.0810 189
9.142
0.1964
55
0.2104
0.00778
100
0.9462
0.02849
145
3.2775
0.0828 190
9.340
0.1999
66
0.2219
0.00803
101
0.9751
0.02923
146
3.3612
0.0846
191
9.542
0.2036
57
0.2301
0.00829
102
1.0047
0.02999
147
3.4467
0.0864
192
9.747
0.2072
58
0.2385
0.00856
103
1.0350
0.03077
148
3.5341
0.0883
193
9.956
0.2109
59
0.2472
0.00883
104
1.0662
0.03157
149
3.6233
0.0902
194
10.169
0.2147
60
0.2661
0.00911
105
1.0982
0.03240
150
3.7141
0.0921
195
10.385
0.2185
61
0.2653
0.00939
106
1.1310
0.03325
151
3.808
0.0940
196
10.606
0.2224
62
0.2749
0.00968
107
1 . 1647
0.0341
152
3.903
0.0960
197
10.830
0.2263
63
0.2847
0.00998
108
1.1992
0.0350
153
4.000
0.0980
198
11.058
0.2303
64
0.2948
0.01029
109
1.2347
0.0359
154
4.099
0.1001
199
11.291
0.2343
65
0.3053
0.01061
110
1.2711
0.03685
155
4.200
0.1022
200
11.527
0.2384
66
0.3161
0.01094
111
1.3084
0.03775
166
4.303
0.1043
201
11.767
0.2425
67
0.3272
0.01127
112
1.3466
0.0387
157
4.408
0.1064
202
12.013
0.2467
68
0.3386
0.01161
113
1.3858
0.0397
158
4.516
0.1086
203
12.261
0.2509
69
0.3504
0.01196
114
1.4260
0.0407
159
4.625
0.1108
204
12.514
0.2562
70
0.3625
0.01232
115
1.4671
0.0417
160
4.737
0.1131
205
12.771
; 0.2596
71
0.3750
0.01269
116
1.5093
0.0427
161
4.852
0.1154
206
13.033
0.2639
72
0.3879
0.01307
117
1.5525
0.04375
162
4.968
0.1178
207
13.29S
0.2683
73
0.4012
0.01345
118
1.5968
0.0448
163
5.087
0.1202
208
13.566
0.2728
74
0.4148
0.01384
119
1.6421
0.0459
164
5.209
0.1227
209
13.84£
0.2783
75
0.4289
0.01425
120
1.6886
. 0.0470
165
5.332
0.1262
210
14,124
0.2819
SOME PROPERTIES OF STEAM
353
TABLE 1.— Continued
t
"
dp/dt
1 '
P
dp/dt
t p
dp/dt
1 t
P i
dp/dt
211
14.408 0.2866
256
33.085
0.5677
301 67.99 i 1.016
346
127.67
1.675
212
14.697 0.2914
257
33.657
0.5758 302 69.01 1.027
347
129.35
1.693
213
14.991 0.2962
258
34.236
0.5840
303 70.06
1.0395
348
131.05
1.711
214
15.290 0.3011
259
34.824
0.6922
304 71.09
1.052
349
132.77
1.729
215
15.594
0.3061
260
36.420
0.6006
305 [ 72.16
1.065
350
1
134. 6li
1.746
216
15.902
0.3111
261
36.026
0.6088
306
73.22
1.0775
1
361
136.26'
1.764
217
16.215
0.3162
262
36.638
0.6172
307
74.31
1.090
352
138.04
1.782
218
16.534
0.3214
263
37.259
0.6266
308
76.40 1.103
353
139.83
1.800
210
16.858
0.3266
264
37.888
0.6341
309
76.51 1.116
354
141.64
1.818
220
17.187
0.3319
265
38.526
0.6426
310
77.64 1.129
356
143.46
1.836
221
17.621
0.3372
266
39.173
0.6513
311
1
78.77 1.142
356
145.31
1.866
222
17.860' 0.3426
267
39.828
0.6600
312
70.92 1.155
357
147.17
1.874
223
18.205 0.3480
268
40.492
0.6688
313
81.08 1.169
358
149.06
1.893
224
18.556 0.3535
269
41.165
0.6777
314
82.26 1.182
369
160.96
1.912
225
18.913
0.3591
270
41.848 0.6868
316
83.44 1.195
360
162.88
1.031
226
19.275
0.3648
271
42.54
0.6960
316
84.66 1.209
361
154.82
1.961
227
19.643
0.3705
272
43.24
0.7052
317
86.86 1.223
362
156.78
1.970
228
20.017
0.3763
273
43.95 1
0.7145
318
87.09 1 1.237
3&3
158.76
1.990
229
20.396
0.3821
274
44.67 ,
0.7239
319
88.34 1.261
364
160.76
2.010
230
20.781
0.3880
276
45.40
0.7334
320
89.60 1.266
365
162.78
2.029
231
21.172
0.3940
276
46.14 ,
0.7430
321
00.87 I 1.280
366
164.82
2.049
232
21.568
0.4000
277
46.88
0.7527
322
92.16 1.296
367
166.88
2.069
233
21.970
0.4061
278
47.64
0.7625
323
03.46 1.309
368
168.96
2.089
234
22.379
0.4123
279
48.41
0.7725
324
94.78 1.324
369
171.06
2.108
235
22.794
0.4185
280
49.10
0.7826
326
96.17 1.339
370
173.18
2.128
236
23.216
0.4248
281
49.98
0.7926
326
07.46 1.364
371
175.31
2.148
237
23.644
0.4312
282
50.77
0.8028
327
08.81 1.369
372
177.47
2.168
238
24.079
0.4377
283
51.58
0.8131
328
100.19 ! 1.384
373
179.65
2.189
239
24.520
0.4442
284
52.40
0.8236
329
101.68 i 1.400
374
181.85
2.210
240
24.967
0.4508
286
53.23
0.8340
330
102.99 i 1.416
375
184.07
2.231
241
25.421
0.4575
286
54.07
0.8446
331
104.41 1.430
376
1
186.31
2.262
242
25.882
0.4643
287
54.92
0.8663
332
106.86 1.445
377
188.58
2.274
243
26.350
0.4711
288
56.78
0.8661
333
107.30 1.461
378
190.86
2.296
244
26.825
0.4780
289
66.65
0.8770
334
108.77 1.477
379
193.17
2.318
245
27.307
0.4850
290
57.63
0.8880
336
110.26 1.493
380
105.50
2.341
246
27.795
0.4920
291
58.42
0.8991
336
111.76 1 1.509
381
197.86
2.364
247
28.290! 0.4991
292
59.33
0.9103
337
113.27 1.525
382
200.23
2.387
248
28.793 0.6063
293
60.25
0.9216
338
114.81 1.542
383
202.63
2.410
249
29.303i 0.5136
294
61.17
0.9330
339
116.36 ' 1.568
384
205.05
2.433
250
29.820 0.5210
295
62.11
0.9445
340
117.02 1.674
I
386
207.49
2.466
251
30.345 0.5285
296
63.06
0.9661
341
1
119.50 j 1.601
386
209.06
2.479
252
30.877 0.5361
297
64.03
0.9678
342
121.10 1.607
387
212.46
2.502
253
31.417 0.5438
298
65.00
0.9796
343
122.72 1.624
388
214.96
2.525
254
31.965 0.5517
299
65.98
0.9916
344
124.36 1.641
389
217.50
220.06
2.548
2ft6
32.621 0.6596
300
66.98
1.0036
345
126.00 ! 1.658
390
2.671
354
SOME PROPERTIES OF STEAM
TABLE 1. — Continued
t p 1 dp/dt
t
P
dp/dt
t
P
dp/dt
t
p j dp/dt
391 222.64 2.594
392 225.24 2.617
393 227.87 2.641
394 230.52 2.664
395
396
397
233.20
235.90
238.62
2.68/
2.71*1
2,735
398
399
400
241.37
244.14
246,93
2.759
2.783
2.807
401
402
249.75
252.60
2.832
2.857
TABLE 2 THE SPECIFIC HEAT OF WATER
Tempehatukh
Regnault
Dieterici
Barnes
Peabody
Cent.
Fahr.
-5
23
32
41
50
69
68
77
86
95
104
122
140
158
176
194
212
248
284
320
356
392
428
464
500
536
672
1.0158
1.0094
1.00530
1.00230
1.00030
0.99895
0.99806
0.99759
0.99735
0.99735
0.99800
0.99910
1.00035
1.00166
1.00305
(1 .0044)
0
1.00000
1.0075
1.0037
1.0008
0,9987
0.9974
0.9970
0.9971
0.9972
0.9974
0.9983
0.9995
1.0012
1.0032
1.0057
1.0086
1.0167
1.0244
1.0348
1.0468
1.0605
1,0758
1.0928
1.1115
1.1318
1.1538
+5
10
1.00049
15
20
1.00116
25
80
1.00201
.'iC
U)
1.00304
1.00425
1.00564
1.00721
1.00896
1.01089
1.01300
1.01776
1.02324
1.02944
1.03636
1.04400
1.05236
(1.06144)
(1.07124)
(1.08176)
(1.09300)
so
60
70
80
90
100
120
0.99940
1.00150
1.00415
1.00705
1.01010
1.01620
140
1.02230
160
1.02850
180
1.03475
200
1 .04100
220
1,04760
240
260
1
280
300
Resnault: from formula, par. 9, Above 200 deg, cent, hia formula is an extrapolation,
Dieterici: from table in original publication, computed by formula from 40 deg, cent, upward.
Bamea: from Phyeical Review, with last value extrapolated.
Teabody: from Steam and Entropy Tables, p. 10.
Dieterici: values in mean calor •^ 'heat units), others in 15 deg. cent, units.
SOME PROPERTIES OF STEAM 355
DISCUSSION
Dr. Sanford A. Moss. In Regnault's original paper on pressure
and temperature of saturated steam was given an empirical formula,
first suggested by Roche, but reconstructed by Regnault and called
"Formula K." This represented Regnault's results more closely
than any other single formula and has also been shown to represent
other experimental results, particularly those for very high pressures
and temperatures. This formula was discussed by Ramsey and Young
in the Philosophical Transactions, vol. 183, 1892, page 111, and also
in London Engineering, vol. 83, January 4, 1907. I have given
some discussion of this matter in the Physical Review, vol. 25, no. 6,
December 1907. It would be interesting if Professor Heck would
give a comparison of the values of his table with those computed by
this formula. If it can be demonstrated, as is the conclusion in the
papers above mentioned, that this formula represents all of the experi-
mental results very closely, it is highly desirable that it be used.
Thermodynamic computations can be carried on very readily, if we
have a single formula for the entire range.
Prof. G. A. Goodenough.* Since the experimental results of
Holborn and Henning have been generally accepted, it seems highly
desirable to have an analytical relation that will express these
results with a sufficient degree of accuracy. For this purpose the
formula of Bertrand,
T
logp = k -n log ^^-^ [1]
seems to be quite suitable. In this formula k, n and a are constants
and T denotes the absolute temperature.
2 The values of the constants can be so chosen that the formula
will give fair results throughout the ordinary range of temperatures
with one set of constants. A closer agreement, however, is obtained
by dividing the temperature range into three parts. The values of
the constants are:
For t = 32 deg. - 90 deg., k = 6.23167, a = 140.1 n = 50
For t = 91 deg. - 237 deg., k = 6.30217, a = 141.43 n = 50
For I = 238 deg. - 400 deg., k = 6.27756, a = 140.8 n = 50
'Associate Professor Mechanical Engineering, University of Illinois, Urbana, 111.
356
DISCUSSION
3 The derivative — takes the simple form
dt
dp
dt
from which follows
dp
lif
pna
T{T -a)
pna
T-a
[2]
[3]
This relation is important since the product T ^ appears in the Clapey-
ron-Clausius formula for steam volume. Table 1 shows a comparison
of the values of p and calculated from Formulae 1 and 2, respect-
dt
ively, with the values obtained by Professor Heck.
4 It will be seen that the maximum difference between the cor-
responding values of p is about 0. 1 per cent. The maximum difference
between the corresponding values of the derivative is greater.
COMPARISON OF VALUES OF p AND
dp
t
P
Beck's value
P
Bertrand'b
FORMULA
1
dp/dt
Heck
dp/dt
Bertrand's
FORMULA
32
0.0886
0.0885
0.003575
0.003588
50
0.1780
0.1781
0.00663
0.006625
75
0.4289
0.4288
0.01425
0.014245
100
0.9462
0.9455
0.02849
0.028575
125
1.9382
1.9387
0.05295
0.052921
150
3.7141
3.7144
0.0921
0.092041
175
6.714
6.711
0.1519
0.15165
200
11.527
11.523
0.2384
0.23843
225
18.913
18.921
0.3591
0.35985
250
29.820
29.834
0.5210
0.52039
275
45.40
45.382
0.7334
0.73248
300
66.98
66.94
1.0035
1.0027
325
96.11
96.07
1.339
1.3390
350
134.51
134.50
1.746
1.7489
375
184.07
184.18
2.231
2.2394
_
400
246.93
247.20
2.807
2.8167
Evidently this is due to the fact that the values of — were obtained
dt
by Professor Heck (and also by Professor Peabody) by a method
involving some approximation; and it is likely that the values calcu-
lated from Formula 2 are the more reliable.
SOME PROPERTIES OF STEAM
357
5 A great advantage of Bertrand's formula lies in the ease with
which it may be used in calculation. This is shown by the following
scheme, which is copied from calculations recently made for the
engineering experiment station of the University of Illinois :
< = 311 312 313 314
7 = 770.58 771.58 772.58 773.58
r-a = 629.78 630.78 631.78 632.78
logT= 2.8868177 2.8873810 2.8879435 2.8885052
2.7998779
log T-a
T
log
50 log
log p = A; - 50 log
T-a
T
T-a~
fc =
T
T-a^
2.7991889
0.0876288
4.381440
6.27756 .
1.896120
0.0875031
4.375155
2.8005659
0.0873776
4.368880
2.8012527
0.0872525
4.362625
los
pna
T-a
p= 78.726
Iogna= 3.8475727.
log pna= 5.7436927
dp
logT— = 2.9445038
^ dt
1.902405
79.874
1.908680
81.036
1.914935
82.212
5.7499777
2.9500998
5.7562527
2.9556768
5.7625077
2.9612550
6 In conclusion, it may be stated that Bertrand's formula seems
to have a wide range of applicabiUty, and with proper choice of con-
stants can probably be used equally well for nearly all saturated vapors.
The Author. The preparation of this paper was an early step in
a special study of the properties of steam upon which the writer
has been engaged for some months; the object of this investigation
has been to formulate more precisely our knowledge of the subject
up to the limit of accurate experiment, at about 400 deg. fahr., and
to extend the relations into the higher ranges, even up to the critical
temperature. It has been found that Thiesen's formula, by modifi-
cation of the two coefficients or constants, can be made to fittheHol-
born-Henning table exactly from 100 deg. to 200 deg. cent., and that
this modified equation holds good up to about 450 deg. fahr. Above
that point the curve begins to rise more rapidly, and even crosses the
original Thiesen curve. Below 212 deg., the relation can be very
exactly extended by means of a small adjustment from the Thiesen
values.
2 The formula presented by Professor Goodenough fits the data
very well; but for precise calculation it has the disadvantage that at
358 DISCUSSION
each point of change in constants there will be a break in the curve
dp .
of the derivative --rT, The Thiesen formula has a forbiddingly com-
plex expression for its derivative, and would be much less convenient
for regular use, if the calculation of — were an operation to be under-
taken frequently.
3 The writer has tried formula K , which is recommended by Dr.
Moss. In the very high range it runs low in p, dropping beneath the
modified Thiesen formula, which itself is not high enough for the
more reliable data.
dp
4 The values of — in Table 1, found simply by the method of
differences, are not precise enough for close work with Clapeyron's
equation, the irregularities resulting from the use of them running as
high as nearly one-fifth of one per cent.
No. 1246
A NF W DEPARTURE IN FLEXIBLE STAY-BOLTS
By H. V. WiLiiE, Philadelphia, Pa.
Member of the Society
There is practically no literature on the subject of stay-bolts, and
this is particularly true of flexible stay-bolts. The increasing size
and pressure of boilers make this subject of vital importance to rail-
roads and to those responsible for the management of that type of
boiler in which the firebox is stayed by a large number of bolts.
2 The boiler of the consolidation locomotive, now the prevailing
type in freight service, contains about 1000 bolts less than 8 in. long
and about 300 of greater length. The large types of Mallet compound
locomotives now meeting with much favor have a much larger num-
ber, there being 1250 short and 300 long bolts in locomotives recently
constructed.
3 In recent years some form of flexible stay-bolt, that is, one
having a movable joint, has been very extensively used in the break-
ing zone of locomotive boilers, but their high cost and the difficulty of
applying them, their rigidity from rust and scale, and the fact that
their use throws an additional service on the adjacent bolts because
of lost motion, has militated against their more general use.
4 It is well known that stay-bolts fail, not because of the ten-
sional loads upon them, but from flexural stresses induced by the vibra-
tion resulting from the greater expansion of the firebox sheets than
of the outside sheets, but notwithstanding the general acceptance of
this theory, engineers have designed stay-bolts solely with respect to
the tensional loads. It is quite general practice, it is true, to recess
the bolts below the base of the thread, and this has effected a
slight reduction in the fiber stress, but practically no effort has been
made to design a bolt to meet the flexural stresses or even to calculate
their magnitude. This is su^p^^"•ing in view of the simplicity of the
calculations to which the ordinary formulae for flexure apply.
Presented at theSpring Meeting, Washington, May 1909, of The American
SociBTT OF Mechanical Engineebs.
360 DISCUSSION
5 Let
F = fiber stress.
E = modulus of elasticity.
I = moment of inertia.
D = diameter.
A^ = deflection.
L = length.
W = load.
We then have
2 F I
W=^-^ (1)
D L
N = ~-^ (2)
Z E I
Substituting
2F U
''-ZED '''
F=3^-°-^* (4)
2 U
This formula shows that the stress increases in direct proportion to
the diameter and decreases as the square of the distance between the
sheets.
6 The application of the formula to service conditions gives the
following stresses:
Conditions : Bolt spacing, 4 in. centers.
Assumed expansion, 4/100 in.
Length of bolt, 6 in.
Type
Diameter of Bolt
Flexural Stress
Iron li in.
Iron 1 "
Iron i "
Spring steel i 1 in. ends /^ in. stem
51,500
45,000
39,400
19,700
7 Iron is universally employed in the manufacture of these bolts
and it is not good practice to exceed a fiber stress of 12,000 lb. per
* Testing of Stay-bolt Iron. H. V. WiUe, A. S. T. M.. vol. 4, 1904.
NEW DEPARTURE IN FLEXIBLE STAT-BOLTS
361
square inch. It is apparent that stay-bolts in the zone which meets
the expansion of the sheets are stressed above the elastic limit and
must necessarily fail from fatigue. Fractures always originate at
the outside sheet at the point where the bending moment due to the
movement of the furnace sheets is greatest.
8 The fractures are in detail, usually starting from the base of a
thread and gradually extending inward. Manufacturers of stay-
bolt material have endeavored to minimize failures and to meet the
unusual conditions of an iron stressed beyond its elastic limit by the
supply of specially piled iron arranged with a view to breaking up
the extension of the initial fracture. For this reason iron piled with
Fig. 1 Section of Firebox
Showing Stay-bolts
Fig. 2 Faggott Piling for
Iron for Stay-bolts
a central section of small bars and an envelop of flat plates has met
with much success for this class of service. In a further efifort to
secure an iron specially adapted to this class of work various forms of
shock, vibratory and fatigue tests have been imposed. No design
has yet been produced however which permits the employment
of material of elastic limit sufficiently high to resist the flexural
stresses, although a large class of material particularly adapted to the
purpose is available.
9 It is obvious that the remedy does not lie in the use of a slow-
breaking material but in the employment, of material of sufficiently
high elastic limit to meet the conditions of service. It is also possi-
362
DISCUSSION
ble to reduce the diameter of the bolt greatly by the use of such a
material, thus proportionately reducing the fiber stress in flexure.
10 Stay-bolt material however must possess sufficient ductility
to enable the ends to be readily hammered over to make a steam-
tight joint and to afford additional security against pulling through
the sheets. To meet these conditions the bolt illustrated in Fig.
3 has been designed. The stem is of the same grade of steel as that
used in the manufacture of springs. It is oil-tempered and will safely
stand a fiber stress of 100,000 lb. per square inch. Its high elastic
limit makes it possible to reduce the diameter to f or ^ in. or even
less. The ends are of soft steel, and it is thus possible to apply and
head up the bolt in the usual manner
jll The employment of a stem of the diameter indicated reduces
the fiber stress in flexure to less than one-half that in the ordinary type
of bolt and it is of material capable of being stressed to a high degree.
It has hitherto been impossible to employ in stay-bolts any of the
steels containing chromium, nicke), vanadium or other metaloids
Fig. 3 Flexible Spring Steel Stay Bolt
possessing properties expecially adapted to this class of work, but
these steels can readily be used in the stem of the bolt described.
^12 The stem of the bolt can be flexibly secured to the end in one
)f the customary ways, but the flexibility of the bolt does not depend
upon a flexible connection. A type of bolt with a relatively inflexi-
ble connection, usually one in which the stem screwed into the ends
with a running fit, met with the most favorable consideration. Such
a bolt is flexible as a spring is flexible, in that it can be deflected to
meet the requirements of service without exceeding the elastic limit.
In fact the stem^may be of a number of pieces, either of plates or
small rods, thus increasing its flexibility.
il3 The actual breaking strength of |the bolt sizes ordinarily
employed is shown in the following statement. These bolts were
recessed to the base of the thread and tested in the same form as that
in which they are employed in service. For comparison the approxi-
mate weights of the usual length of bolt are also given. These
NEW DEPARTURE IN FLEXIBLE STAY-BOLTS
363
weights are for bolts over the entire length, including the squared
ends for screwing the bolts into the sheets.
ACTUAL BREAKING STRENGTH OF
STAY-BOLTS
Type
Nominal Diameter
Actual
Breaking
Weight
Vibrations
. . i 1 in.
32,500
24,500
32,000
20 'oz.
15 "
10/' t
1
6,000
• • 1 I "
5,200
Spring steel stem. . .
. . 1 in. ends ,'„ in. stem
500,000
14 The vibrating test was made by clamping one end of theboltin n
machine and revolving the other end through a radius of ^ in. , the spec-
men bsing 6 in. long from the end of the right head to the center of
the rotating head. A tensional load of 4000 lb. was also applied i(»
the bolts. The best grades of iron bolts break on being subjected to
from 5000 to 6000 rotations, whereas the spring steel bolts wore
■
lllll
^ma^m^^^^^mmmmmmm
1
mm:
Fig. 4 Spuing Flexible and Regular Iron Bolts of Sa.me Tensile SxRENOTri
vibrated 500,000 times without failure, and on some of them tlie
test was continued without -failure to 1,000,000 vibrations. These
tests demonstrated that the bolt is not stressed beyond the elastic
limit under these severe conditions and that the probability of its
failure in less severe conditions is very remote.
15 The extent of the expansion which can take place in the fire-
box of a boiler can readily be calculated.
Distance between stay-bolts, 4 in.
Temperature of inside sheet, 400 deg. fahr.
Temperature of^outside sheet, 100 deg. fahr.
Coefficient of expansion, 0.0000066.
364
NEW DEPARTURE IN FLEXIBI.E STAY-BOLTS
Then the expansion between two bolts will equal: 0.0000066 X (400 —
100) X 4 = 0.0079, and each bolt will deflect 0.00395 in. It has
been shown that this amount of deflection will stress the usual type
of bolt beyond the elastic limit. In practice however one bolt may
hold rigidly, throwing the entire deflection on the adjacent bolt, or
neither bolt may deflect and the sheet will then buckle. Under this
condition the neutral axis will assume the form ABC and the length
AB will equal 2.00395 in. and the sheet will buckle to an extent,
BD = 1/2.00395^ — 2^ = 0.125 in. It is obvious that the repetition of
a force suflScient to buckle a sheet ^ in. must ultimately lead to a
Fig. 5 Showing Manner in which Plates Buckle with Eigid Stays
crack in the furnace sheets. If, however, the bolt deflects, allowing
the sheet to expand normally, the latter will be relieved of these extra-
neous loads.
16 A bolt of sufficient flexibility to deflect under the forces follow-
ing expansion, and of material which will not be stressed beyond the
elastic limit in resisting these forces, will greatly assist in reducing
the cost of boiler maintenance by eliminating broken stay-bolts and
reducing the stresses in the furnace plates. If in addition the bolt
has a smaller diameter the life of the furnace plates should be fur-
ther increased, as such a bolt will interpose less obstruction to the
circulation of the water in the water legs.
NEW DEPARTURE IN FLEXIBLE STAYBOLTS 365
DISCUSSION
William Elmer. The decrease in the diameter of a staybolt, from
15/16 in. or 1 in., which I believe is the present practice, to a small
diameter, as 7/16 in., even if the tensile strength of the material is
increased, brings to mind at once the possibility of twisting off these
small bolts in their mechanical application to the boiler. The writer
hopes Mr. Wille will say something about this,
W. E. Hall. Any one who has had the care of the locomotive
type of boiler appreciates that staybolt maintenance is a source of
intense anxiety, and that the fact that more disastrous results are not
forthcoming reflects great credit on the vigilance of the motive power
departments of our railroads. This result, however, is accomplished
only at high maintenance cost, which, fortunately, has always been of
secondary consideration.
2 Flexibility and length, of which Mr. Wille speaks, no doubt have
considerable influence on the breakage of these bolts. It would be
interesting to have details showing the construction of the bolts,
and just how they were held, in his vibratory tests,
3 It should be noted, however, that his construction calls for a
studbolt of larger diameter than the ordinary staybolt, and in addi-
tion this studbolt must project into the leg of the boiler to give suffi-
cient length of thread-contact of the staybolt proper in the studbolt;
in other words, he increases the diameter (the studbolt), introduces
two threaded surfaces to the strain of flexure, and shortens the stay-
bolt an amount equal to the projection of the studbolt into the leg
of the fire-box. Is this not contrary to his own deductions? Assum-
ing that this construction would decrease, but not eliminate, broken
staybolts, would this construction facilitate their detection, diflScult
under the best conditions, or make it more difficult? The reduced
length might be relieved by making the fit of the thread of the bolt
in the studbolt looser than that of the studbolt in the sheet, thereby
always throwing the point of flexure upon the studbolt at the side of
the outer sheet. But this is a risky procedure hardly deserving of
consideration in boiler practice.
4 The breakage of staybolts is confined almost exclusively to the
upper rows of bolts of the legs of the fire-box. The fracture usually
starts from the top, sometimes from the side, not infrequently around
the circumference and occasionally from the bottom. The fracture
is almost always close to the outer sheet, but a break close to the
366
DISCUSSION
Fig. 1 Blank and Completed Ends of Flexible Staybolt
NEW DEPARTURE IN FLEXIBLE STAYB0LT8
3 67
inside sheet is not unknown. The fracture is always in detail, bar-
ring shamefully defective material or workmanship, or at least up to
the point where the remaining area in contact is not sufficient to with-
stand the strain to which that reduced area of the bolt is subjected.
More or less irregularity of the line of fracture is to be expected. Con-
ditions are not always the same for every bolt : all bolts do not fit the
sheets alike, there is more or less variation in the upsetting, and the
buckling and warping of the inner sheet is not the same for each bolt.
These, together with other minor conditions, representing reasonable
refinement in practice, preclude uniformity of the fracture.
5 But the important points are, that these bolts always break in
detail and always at the root of the thread. Have we any reason to
expect that it would be otherwise? For example, if we wish to break
a piece of metal we first grip it in a vise, notch it close to the jaws of
the vise and bend it back and forth. We do the same with a stay-
FiQ. 2 Flexible Statbolt with End shown in Fig. 1
bolt when we cut the thread with a sharp die (and in this respect the
U. S. Standard is inferior to the Whitworth), hold it firmly in the
outer sheet and subject it to the bending due to the expansion, con-
traction and warping of the inner sheet. Cutting the thread produces
a constructive defect. This feature is necessary as the bolts are now
used and its correction is beyond the scope of mathametics.
6 Fig. 1 and Fig. 2 show designs for bolts made in January 1886.
In Fig 1 the object was to make the bolt of iron or steel, and by a
gradual reduction from the ends towards the center to make the body
of the bolt more flexible; and to make the length of the end such that,
after making sufficient allowance for upsetting, the thread could m^t
project beyond the face of the inside of the sheet. The flare on
the end of the bolt was shown merely as a more ready means of
upsettmg the 'end, especially when the bolt is of steel. As the
368 DISCUSSION
number of bolts breaking at the face of the inside sheet is so
small as to be negligible, and to accommodate for the difference in
widths in the leg of the fire box, the bolts could be made of vary-
ing lengths on a bolt machine and kept in stock. By this method,
if made by upsetting, it would not be necessary to turn the shank
of the bolt. In all cases, however, it is preferable that the thread-
ed end in the heavier sheet should not project into the leg of the
fire box. This construction permits of better circulation, and of
a somewhat longer bolt, and better braces the threaded length
against deflection.
Alfred Lovell. The durability ot stay bolts in high-pressure
locomotive boilers is a prominent factor in the cost of locomotive
maintenance and has an important bearing on the safety of the public.
Any innovation intended to increase their durability or reliability is
therefore worthy of the most careful demonstration and service trial.
The staybolt described by Mr. Wille is of this character, and his paper
clearly shows the desirability of providing a staybolt of high elastic
limit and great flexibility.
2 This is accomplished by the use of metal having the requisite
qualities in a high degree, and yet maintaining in the direction of the
bolt's axis, the desirable features of continuity and rigidity, thus
avoiding the liability of unequal tension of adjacent bolts. This
is undeniably a departure in the right direction, and one that will
receive prompt attention in railway mechanical engineering.
3 One other important feature to be considered, however, in
adopting a new form or in selecting a new material for staybolts,
is ability to resist corrosion under the conditions of use, a feature even
more important with staybolts of the proposed high tensile strength
and small diameter, than with staybolts of ordinary diameter. For
example, the paper gives the actual breaking strength of a 1-in. iron
bolt as 32,500 lb., and of a 3^ -in. spring-steel stembolt as 32,000 lb.
If it is assumed that each is reduced ^ in. in diameter by corrosion,
this will reduce the breaking-strength of the 1-in. bolt by 23.56 per
cent, and of the ^-in. bolt by 48.98 per cent. The breaking-strength
of the iron bolt is then 24,843 lb. and that of the spring-steel stem-
bolt is 16,327 lb. Another J-in. reduction in diameter would bring
the strength of the iron bolt to 18,278 lb., and of the spring-steel stem-
bolt to 5878 lb.
4 Formula 4 of Mr. Wille's paper shows that the flexural stress, the
ordinary cause of failure, decreases directly as ihe diameter, yet it is
NEW DEPARTURE IN FLEXIBLE STAYBOLTS 369
evident that if the bolts are affected by corrosion and in equal amount,
this will eventually reduce the tensile strength of the small bolt to a
point below the proper factor of safety, while the larger bolt has still
tensile strength to spare.
5 Since the new type of staybolt makes it possible to employ
various steels and alloys which cannot be used in the ordinary stay-
bolt, it fortunately provides for a wide range of selection, and quite
probably a steel-alloy or mixture may be provided that will resist
corrosion equally well or better than the iron ordinarily used.
6 The impurities in the water used also very greatly affect corro-
sive action, and it is probable that with the new type of bolt it would
be advantageous to examine carefully the water of the locality where it
is to be used and to select a metal with reference to the character of
the water. Information regarding the resistance of various steel alloys
to corrosion, in water of ordinary purity, and experiments determining
the metal that will give the least corrosion in water having various
impurities, are highly desirable, since the new bolt makes possible the
use of that metal which combines in the greatest degree flexibility,
high elastic limit and resistance to corrosion.
7 A combination of these three qualities, in a staybolt of the type
described, will be an innovation much to be desired in locomotive
practice.
F. J. Cole. In reading this paper the following occurred to me:
a Have any of these staybolts been applied to locomotive
boilers?
b What is their approximate cost?
c With 1-in. ends is there any danger of overstraining the |-in.
diam. spring-steel center in screwing the bolt to place?
It would seem that the torsional stress on the compara-
tively small central part in screwing the bolts to place
when they fit tightly in the sheets would be excessive.
d In upsetting the soft ends, does the |-in. diam. center
afford sufl&cient resistance to avoid injury to the threads
from hammering? It is customary in upsetting the ends
to use a hammer or heavy weight at the other end of the
bolt to afford the necessary resistance and to prevent
injury to the threads.
e In Par. 7 Mr. Wille speaks of 12,000 lb. fiber stress being
good practice. I presume he has in mind the combined
stress due to tension and bending. Current locomotiv*
370 DISCUSSION
practice is represented fairly well by 5500 to 6000 lb. for
tension alone.
The Author. Mr. Hall has fully described the general charac-
teristics of a staybolt broken in service. The able experiments of
Boushinger and Wohler have shown that a detail fracture occurs in
parts which have been stressed by flexure above the elastic limit. If
the parts be threaded the stress may be localized at the base of the
thread which will increase the possibility of failure of parts stressed
at or near the elastic limit, but if the parts be not stressed above the
"fatigue" elastic limit of the material, it will not fail irrespective of
whether it be threaded or of plain section.
2 The bolt described is made of two kinds of material:
a The stem of high-carbon tempered steel.
b The two ends of soft steel.
This arrangement permits the use of a material having such a high
elastic limit and provides an ample factor of safety and still allows
the soft ends to be ri vetted over to make a steam-tight joint. The
bolt referred to by Mr. Hall does not fulfil these conditions, as if it
were made of steel having a sufficiently high elastic limit it would be
impossible properly to rivet the ends.
3 The ends of the composite bolts are not larger than a normal
bolt, those in service being made 1 in. in diameter when used in new
boilers and if in. and 1^ in. when used to replace f-in. and 1-in.
bolts respectively in old boilers. The stems are not made a tight fit
into the ends but are threaded so that they can be tightly screwed into
the ends by hand thus affording some flexibility in the joint. It is
not necessary however to thread the stems as they can be upset and
the ends closed around them, making a ball joint, thus having all the
advantages with none of the disadvantages of the usual type of flexi-
ble bolt.
4 The bolts have been in service for two years in a district notori-
ous for the trouble experienced with corrosion and the bolts are still
intact. Mr. Lovell suggests the possibility of providing special steels
to resist corrosion. The high-carbon tempered steel being dense and
hard is an admirable material to resist corrosion; furthermore as the
stems are oil-tempered they have a coat of enamel which prevented
the slightest amount of pitting even after two years' service.
5 There are a large number of these bolts in service in locomotive
NEW DEPARTURE IN FLEXIBLE STAYBOLTS 371
boilers and no difficulty has been experienced either from twisting of
the stem in screwing into the boiler or from injury to threads in rivet-
ting.
6 The cost of material entering into the construction of the spring
flexible bolt is very much less than that of the usual type of staybolt
since the weight is much less and the materials are the relatively cheap
commercial grades of steel instead of the high-priced irons usually
employed in the manufacture of staybolts. The labor cost to man-
ufacture is however somewhat higher.
7 There have been about 3000 of these bolts put into service
and no danger has been experienced of overstrainin the f-inch
tern in screwing the bolt in the boiler; nor has any difficulty been
sxperienced because of insufficient resistance to avoid injury to the
e bread when the bolts are hammered up.
No. 1247
THE HUDSON-FULTON CELEBRATION
In keeping with the celebration of the discovery of the Hudson
River and the successful application of steam to navigation, the
House Committee of the Society appointed a subcommittee, Edward
Xan Winkle, Chairman, to prepare an exhibit of models, drawings,
letters, books and other items related to early steam navigation, of
interest to the general public as well as the engineer.
The exhibit was the only participation of any engineering organi-
zation, as such, in the celebration, and much credit is due Mr. Van
Winkle for the time and attention devoted to its preparation. The
exhibit was held in the rooms of the Society and was open from 9 a.m.
to 5 p. m. every week day during September and October 1910. A
list of the exhibits follows :
Model of Fulton's boat, the Clermont, as she was at the time of her
maiden trip. After making several trips during the fall of 1807, the Cler-
mont was docked at Browne's shipyard and fitted up for regular passenger
traffic. Loaned by the Smithsonian Institution.
Model of Stevens' boat, the Phoenix, at the time of making her notable
New York-Philadelphia trip — the first ocean voyage made by a steam vessel.
Loaned bj' the Smithsonian Institution.
Model of a steamboat built by John Fitch in 1786, having a peculiar
arrangement of oars dipping into the v.ater something like a canoe paddle.
Despite its clumsy appearance the boat made a trip of 20 miles on the Dela-
ware River. Loaned by the Smithsonian Institution.
Model (9-ft.) of the Deutschland, showing the remarkable development in
steam navigation during the last one hundred years. Loaned by the Ham-
burg-American line.
Portrait of Fulton painted by himself while a pupil of Benjamin West.
Presented to the Sociely by Mrs. R. Anna Gary.
Fulton dining table, of mahogany, 85 ft. long and 5 ft. wide. Presented to
the Society by Thomas Egleston in 1891, who received it from his brother,
George Egleston, into who-se hands it had passed from Mrs. Egleston 's House,
to whom Robert Fulton gave it.
Two autograph drawings bj' Fulton, one of them a brush dr.awing of a high-
level canal, presented by Cornelia J. Carll, showing his artistic ability as well
as his mechanical genius. The other drawing is of the Sound steamer Fulton,
built in 1813, and was presented by Louisa Lee Schuyler.
^ Drawing of the Fulton reproduced on a bronze tablet made in connection
with the Fulton monument erected by the Society over Fulton's grave in Trin-
ity churchyard. The tablet also gives a short description of the Clermont.
374
SOCIETY AFFAIRS
ROBEKT FtJLTON
FROM A PAINTING BY HIMSELF, PRESENTED TO THE AMERICAN SOCIETY OP MECHANICAL
ENGINEERS BY MRS. R. ANNA CART, NOVEMBER 30, 1897
SOCIETY AFFAIRS 375
Photograph of a letter by Fulton describing his experimental boat on the
Seine, with copies of other letters by Fulton and his friends, presented by
Henry Harrison Suplee, member of the Committee on Society History.
Hudson River Guide, published in 1850, describing the various points of
interest on the Hudson and giving the time of sailing of the Hudson River
boats of that period. Loaned by H. J. Gclien.
Oil painting of James Watt, a copy of a portrait by deBreda, now in tlie
possession of John Scott of Hawkhill, Greenock, presented by past members
of the Council.
Oil portrait of Capt. John Ericsson, painted by Ballin of Stockholm, show-
ing the designer of the Monitor at the age of 59. A bust of Ericsson, by Knee-
land, was presented to the Society by James Mapes Dodge, Past-President,
Am. Soc. M. E.
Solid silver model of the Half Moon, loaned by Tiffany and Company.
Model of the Monitor, presented by Thomas F. Rowland, who built the Mon-
itor at the Continental Iron Works.
Models, exhibited by Ericsson at the Centennial Exposition, loaned by the
United Engineering Society.
Early books on steam navigation, from the Library of the Society.
Copy of a letter from Fulton to Boulton & Watt in 1810, ordering an engine
for another boat.
Copies of letters by Fulton's workmen.
Drawings showing the comparative size of the White Star S. S. Olympic,
Clermont and Half Moon; of New York's early water works, corner of Centre
and Reade Streets; a water works note issue in 1774: all loaned by Daniel
Arthur.
Color print of the Half Moon.
Fac-simile of Rules and Regulations for passengers in the Clermont.
Sketch of James Watt discovering the condensation of steam.
Photographs of a letter written by Fulton to Boulton & Watt, describing the
Clermont and ordering an engine for another steamboat. Contributed by
the Smithsonian Institution.
Print of the original oil painting by W. F. HalsoU of the battle between the
Monitor and the Merrimac in the collection of the late Thomas Fitch Rowland,
the builder of the Monitor.
Bills, receipts, etc., written by John Fitch; stock certificate issued by John
Rumsy; copy of New York Herald of 1815 containing items of steamboats;
copy of Washington Gazette of 1821 attacking Fulton: all loaned by Dr. C. S.
Bullock of Stratford, Conn.
Silver Hudson-Fulton Medal, loaned by Dr. Geo. F. Kunz.
Descriptive Guide to the Grounds, Buildings and Collections of the New
York Botanical Garden.
List of Prints, Books, Manuscripts, etc., relating to Henry Hudson, the Hud-
son River, Robert Fulton and Steam Navigation, at the Lenox Branch of the
New York Public Library.
The Indians of Manhattan Island and Vicinity, by Alanson Skinner of the
department of anthropology of the American Museum of Natural Histor}^
The Wild Animals of Hudson's Day and the Zoological Park of our Day, by
W. T. Hornaday, Sc.D., Director of the New York Zoological Park.
All of these loaned by Geo. F. Kunz.
376 SOCIETY AFFAIRS
Considerable interest was manifested in this exhibit and it was well
attended. The total namber of visitors who registered was 355, but
400 is a fair estimate of the total. The greatest registration for any
one day was 52 on Monday, September 27. A wide extent of territory
was represented, including Connecticut, Delaware, Illinois, Maine,
Massachusetts, Maryland, Michigan, Ohio, Pennsylvania, Rhode
Island, Tennessee, Washington, D. C, Wisconsin, Canada, Japan
and Switzerland.
On the morning of September 24, in the presence of officers and
members of the Society and of the Pennsylvania Society, wreaths
were placed on the Fulton Monument erected by the Society in 1901
in Trinity Churchyard, New York. The Pennsylvania Society was
represented by Robert Mazet, Vice-President, and Barr Ferree, Sec-
retary. J. M. Schroeder, Commissioner from Pennsylvania, was also
present.
Members of both societies assembled in the lobby of the church
and, preceded by Dr. Manning, rector of Trinity Church, marched to
the Fulton Monument. Dr. Manning offered prayer and wreaths
were placed on the monument, after which the Lord's Prayer was
recited, followed by the benediction. "Taps" sounded by a Seventh
Regiment bugler brought the exercises to a close.
SOCIETY AFFAIRS
377
378
SOCIETY AFFAIRS
a a
ii a
SOCIETY AFFAIRS
379
o
E-i
380
SOCIETY AFFAIRS
Model of John Fitch's Boat
BUILT IN PHILADELPHIA AND TRIED ON THE DELAWARE RIVER IN 1786
The Phoenix, Built by John Stevens. In 1809 She Made an Ocean Trip
from New York to Philadelphia
rROU PHOTOGRAPHS MADE IN THE ROOMS OF THE SOCIETY
No. 1248
MEETINGS, OCTOBER-DECEMBER
NEW YORK MEETING, OCTOBER 12
At the meeting of the Society held October 12 in the Engineering
Societies Building, Prof. R. C. Carpenter presented his paper on The
High-Pressure Fire-Service Pumps of Manhattan Borough, City of
New York. President Jesse M. Smith presided. The attendance was
192.
Secretary Calvin W. Rice read an invitation from the Institution
of Mechanical Engineers to hold a joint session with them from
July 26 to July 29, 1910. The President then introduced the follow-
ing Japanese commissioners visiting the United States to study
various industries: Dr. Ryota Hara, doctor of engineering and chief
engineer of Yokohama; Rinnosuke Hara, of the Japanese Architec-
tural Society; Junkichi Tanabe, of Tokyo, of the Institute of Japanese
Architects; and Narazo Takatsuji, director of a large spinning fac-
tory. A telegram of regret was received from Kojiro Matsukata^
the leading shipbuilder of Japan.
Those taking part in the discussion of the paper were: Prof. Geo.
F. Sever, WiUiam M. White, Geo. L. Fowler, John H. Norris, J. R.
Bibbins, J. J. Brown, Geo. A. Orrok, Frederick Ray, H. Y. Haden,
Thos. J. Gannon, Henry B. Machen, Richard H. Rice, Chas. A. Hague,
A. C. Paulsmeier, Prof. W. B. Gregory, Wm. 0. Webber and Chas.
B. Rearick.
At the close of the discussion Mr. White showed a number of lantern
slides giving efficiency curves of various pumps designed by the
I. P. Morris Co., Philadelphia, Pa.
Mr. Fowler, described the work of centrifugal pumps in dredging,
and exhibited the following lantern slides, as evidence of the great
suction capacity of these pumps :
A piece of shaft weighing 70 lb. raised and passed by a 15-in. dredging pump;
improvement of New York Harbor, Steamer Reliance.
A piece of tree root raised and passed by a 12-in. pump from 14 ft. of water at
Miami, Fla.; Florida East Coast Railway Company improvements.
382 SOCIETY AFFAIRS
A piece of pig iron measuring 11| in. by 4f in. by 3j in. and weighing 35 lb.,
raised and passed by an 8-in. special cataract wrecking pump from 15 ft. of water
from the wreck of a canal boat sunk at Puas Dock, Yonkers, N. Y., by the Bax-
ter Wrecking Company, New York.
ST. LOUIS MEETING, OCTOBER 16
A meeting of The American Society of Mechanical Engineers and
the Engineers Club of St. Louis was held at the rooms of the latter
organization at 8.15, Saturday evening, October 16, under the direc-
tion of William H. Bryan, Chairman, M. L. Holman and E. L. Ohle,
Secretary, of the local joint committee.
A letter from President Jesse M. Smith was presented, indicating
the sentiment of the Society towards local meetings. This was briefly
responded to by President E. E. Wall, of the Engineers Club of St.
Louis, who also emphasized his belief in the advantages of cooperation.
Prof. R. C. Carpenter of Cornell then presented in abstract his
paper on The High-Pressure Fire-Service Pumps of Manhattan
Borough, City of New York, accompanying it by running comments
and comparisons.
He was followed by Horace S. Baker, assistant engineer of the City
of Chicago, who presented with the aid of illustrations the results of
recent study with a view to adopting high-pressure service. E. E.
Wall, assistant water commissioner. City of St. Louis, outlined the
plan proposed for high-pressure fire-service in St. Louis. He was
followed by H. C. Henley, chief inspector, St. Louis fire prevention
bureau, and vice-president of the National Fire Protection Associa-
tion, expressing views of the fire insurance authorities, entirely favor-
able to the installation of such systems when properly designed and
operated. Chas. E. Swingley, chief of the St. Louis fire department,
on invitation made a few brief remarks to the effect that such systems
were of undoubted advantage in the congested dictricts of large cities,
and expressed the^hope that something might soon be done along this
line in St. Louis. There was further brief discussion by^Edw. Flad,
Prof. H. W. Hibbard, and H. C. Toensfeldt.
Luncheon was served by the Engineers Club of St. Louis. The
attendance was 100.
BOSTON MEETING, OCTOBER 20
On Wednesday evening, October 20, a joint meeting of the Society
with the Boston Society of Civil Engineers was held in the latter
society's rooms, Tremont Temple, Boston, Mass.
SOCIETY AFFAIRS 383
Chas. T. Main, vice-president of the Boston Society of Civil Engi-
neers, presided. Following the routine business of the Society of Civil
Engineers, Mr. Main read a letter from Jesse M. Smith, President of
The American Society of Mechanical Engineers, regretting his inabil-
ity to be present at the meeting, and wishing the Boston members
success for their coming meetings.
A paper by Cav. Gaetano Lanza, professor, and Lawrence S.
Smith, instructor, at the Massachusetts Institute of Technology, on
Stresses in Reinforced Concrete Beams, was read by the former.
Following the presentation of the paper, a discussion by J. R.
Worcester of Boston, Mass., was read. Sanford E. Thompson,
Fred S. Hinds, Henry F. Bryant and Geo. F. Swain contributed
oral discussions.
The total attendance at the meeting was 180, of whom 60 were
members of the Society of Civil Engineers, 50 members of The Amer-
ican Society of Mechanical Engineers and 70 guests.
NEW YORK MEETING, NOVEMBER 9
At the meeting of^the Society in New York on November 9, Pro-
fessor Lanza presented his paper on Stresses in Reinforced Concrete
Beams, and Professor Rautenstrauch his paper on Design of Curved
Machine Members. The discussion on both papers proved valu-
able, the lantern slides shown in the discussion of Professor Lanza's
paper adding much to its interest. Those participating were Sanford
E. Thompson, E. P. Goodrich, Prof. Walter Rautenstrauch, Prof. W.
H. Burr, B. H. Davis, of the Lackawanna Railroad, who showed slides
of a number of concrete arches in railroad work, C. B. Grady of the
New York Edison Company, who showed slides of beams and floor
slabs under test, F. B. Gilbreth, who showed slides of the longest
concrete beam of a rectangular section ever built in a roof, as well as
other beams which had successfully passed through the fire and earth-
quake of the San Francisco disaster. Contributed discussions by
Prof. J. C. Ostrup, E. L. Heidenreich, and.C. E. Houghton were also
presented. Professor Rautenstrauch's paper was discussed by a num-
ber of authorities on machine tool design, as follows : Professor Lanza,
Chas. R. Gabriel, George R. Henderson, Professor Burr and Carl G.
Barth. Those submitting written chscussions were: C. E. Houghton,
A. L. Campbeh, H. Gansslen, F. I. Ellis, E. J. Loring and John
S. Myers.
384 SOCIETY AFFAIRS
ST. LOUIS MEETING, NOVEMBER 13
At the meeting of the Societj'^ at St. Louis, November 13, with the
Engineers Club of St. Louis, a description of the new plant of the
Heine Safety Boiler Company of Boston was presented by E. R. Fish,
Secretary of the Company, under the title, A Modern Boiler Shop.
There was also further discussion of Professor Carpenter's paper on
High-Pressure Fire-Service, continued from the October meeting.
BOSTON MEETING, NOVEMBER 17
A successful meeting of the Society was held at Boston in the Lowell
Building, Massachusetts Institute of Technology, Wednesday even-
ing,'November 17. Two hundred and forty were present at this meet-
ing and the Low-Pressure Steam Turbine was the topic of discussion.
Henry G. Stott of the Interborough Rapid Transit Company gave
an interesting account of the difficulties encountered as well as the
very fine results obtained from an installation recently made at the
59th Street Station of his company, New York. W. L. R. Emmet,
engineer of the lighting department of the General Electric Company,
described the low-pressure turbine situation from his viewpoint and
pointed out the advantages of this type of prime mover for many mill
installations and industrial works in New England. H. E. Longwell,
consulting engineer of the Westinghouse Machine Company, and
Edward L. Clark, manager of their Boston office, both spoke on the
work that company is doing in this field. Max Rotter, turbine engi-
neer of the Allis-Chalmers Company, pointed out in a humorous way
a number of situations where the low-pressure turbine was not a desir-
able proposition. Professor Miller of the Massachusetts Institute of
Technology also discussed the subject.
ST. LOUIS MEETING, DECEMBER 11
A meeting was held Avith the Engineers Club of St. Louis on Sat-
urday evening, December 11, at the rooms of the latter society. The
meeting was called to order by William H. Bryan, member of the
Meetings Committee of the Society and chairman of the joint com-
mittee of the two societies at St. Louis. Prof. E. L. Ohle acted as sec-
retary. There were present fifty-five members and guests.
The paper of the evening was by G. R. Parker of the General Elec-
tric Company, on The Relation of the Steam Turbine to Modern Cen-
SOCIKTT AFFAIKS 385
tral Station Practice, in which the underlying principles of modern
steam turbines were discussed, together with the design of various
prominent types on the market, and the developments made in recent
years in improving capacity and efficiency. Attention was called to
the large turbine capacity which may now be obtained within limited
floor space; to the question of low-pressure turbines and their avail-
ability in supplementing standard reciprocating engines, increasing
both their capacity and economy; also to the work already done in
this direction at the plant of the Union Electric Light & Power Com-
pany in St. Louis, and to prospective work along similar lines in the
same plant. The address was illustrated by lantern slides.
Discussion followed by Chairman Bryan, Prof. H. W. Hibbard,
L. R. Day, E. R. Smith and Prof. E. L. Ohle.
On Saturday afternoon an excursion was made to the Ashley Street
plant of the Union Electric Light & Power Company, for the inspec-
tion of the apparatus and equipment, on the invitation of John
Hunter, chief engineer.
BOSTON MEETING, DECEMBER 17
On Friday evening, December 17, a goodly number of engineers
of Boston and vicinity gathered on invitation of the local members of
The American Society of Mechanical Engineers to discuss the Effect
of Superheated Steam on Cast Iron. The meeting was called to order
by Prof. Ira N. HolHs.
The committee which has been in charge of the meetings, consist-
ing of Messrs. Hollis, Moultrop, Miller, Mann and Libbey, was con-
tinued.
The papers on the subject for the evening were then presented by
their authors. Prof. Edward F. Miller of Boston, Arthur S. Mann of
Schenectady, and Prof. Ira N. Hollis of Boston, and were discussed
by B. R. T. Collins, George A. Orrok, Chas. H. Bigelow, W. K.
Mitchell, John Primrose, L. B. Nutting, Wm. E. Snyder and others.
The general purport of the discussion was rather reassuring to the
users of cast-iron pipe and fittings, and to those who are interested
in the extension of the use of superheated steam, in indicating that
superheated steam per se has no injurious effect upon cast iron fittings,
but that if the pipe lines are properly designed for the greater ranges
of temperature, the fittings made adequate to the pressure and fluctu-
ations in temperature avoided, the use of superheated steam intro-
duces no piping difficulties which can not be easily overcome.
THE ANNUAL MEETING
The thiitieth annual meeting of The American Society of Mech-
anical Engineers was held in the Engineering Societies Building
December 7 to 10, with an attendance of 628 members and 435
guests. For the first time at such a meeting the arrangements for
the entertainment features were entirely in the hands of the local
coEomittee, the members in New York and vicinity acting as hosts,
this method of handling an important part of the annual meeting
being fully justified by the results. A feature of the meetmg was
an afternoon trip on Wednesday, through the new Pennsylvania
Terminal, which brought out a large body of members and guests.
The attendance at the reception on Thursday evening, held in the
ball-room of the Hotel Astor, was nearly 600,
PROGRAM
OPENING SESSION
Tuesday, December 7, 8.30 p.m., Auditorium
THE president's ADDKESS
The Profession of Engineering, by Jesse M. Smith
ELECTION OF OFFICERS
Report of Tellers of Election of Officers and introduction of the
President-elect.
RECEPTION
The President and President-elect, with their ladies, received the
members and guests in the rooms of the Society. Music and refresh-
ments followed the reception.
BUSINESS MEETING
Wednesday, December 8, 9.80 a.m., Auditorium
Annual business meeting. Reports of the Council, Tellers of Elec-
tion of membership, standing and special committees and Gas Power
Section. Amendments to the Constitution. New business.
Luncheon was served to members and guests.
SOCIETY AFFAIRS 387
Wednesday afternoon
Excursion to points of engineering interest. Hosea Webster,
Chairman Sub-Committee on Excursions.
LECTURE
Wednesday, 8.15 p.m., Auditorium
The Era of Farm Machinery, L. W. Ellis, of the Bureau of Plant
Industry of the United States Department of Agriculture at Wash-
ington, D. C. Illustrated by lantern slides.
PROFESSIONAL SESSIONS
Thursday, December 9, 9.30 a.m., Auditorium
measurement of the flow of fluids
Tests on a Venturi Meter for Boiler Feed, Chas. M. Allen.
Discussed by F. N. Connet, Clemens Herschel, Dr. Sanford A.
Moss, Prof. L. S. Marks.
Efficiency Tests of Steam Nozzles, F. H. Sibley and T. S. Kem-
ble.
Discussed by A. F. Nagle, A. R. Dodge, Prof. C. C Thomas, J.
A. Moyer.
The Pitot Tube as a Steam Meter, Geo. F. Gebhardt.
Discussed by Walter Ferris, A. R. Dodge, Prof. W. B. Gregory.
An Electric Gas Meter, C. C. Thomas.
Discussed by Prof. W. D. Ennis, E. D. Dreyfus, A. R. Dodge,
Prof. L.S.Marks.
Luncheon was served to members and guests at the conclusion of
the session.
Thursday, 2 p.m., Auditorium
STEAM engineering
Tan Bark as a Boiler Fuel, David M. Myers.
Discussed by A. A. Cary, Prof. Wm. Kent, Prof. L. P. Brecken-
ridge.
388 SOCIETTf AFFAIRS
Cooling Towers for Steam and Gas Power Plants, J. R.
Bibbins.
Discussed by Geo. J. Foran, W. D. Ennis, H. E. Longwell, B. H.
Coffey, E. D. Dreyfus, F. J. Bryant, Carl G. deLaval.
Governing Rolling Mill Engines, W. P. Caine.
Discussed by H. C. Ord, James Tribe.
An Experience with Leaky Vertical Fire Tube Boilers
F. W. Dean.
Discussed by R. P. Bolton, Prof. Wm. Kent, J. C. Parker, O. C
Woolson, A. A. Gary, Prof. A. M. Greene, Jr., E. D. Meier, D. M.
Myers.
The Best Form of Longitudinal Joint for Boilers, F. W.
Dean.
Discussed by R. P. Bolton, Carl G. Bartb, E. D. Meier, Prof. A.
M. Greene, Jr., W. A. Jones, Prof. S. W. Robinson, Geo. I. Rock-
wood, Sherwood F. Jeter.
Thursday, 2 p.m., Lecture Hall
GAS POWER section
Business meeting and election of officers.
Testing Suction Gas Producers with a Koerting Ejector
CM. Garland, A. P. Kratz.
Discussed by Prof. R. H. Fernald, G. M. S. Tait, H. H. Suplee, L.
B. Lent, S. C. Smith, W. B. Chapman, Edw. N. Trump.
Bituminous Gas Producers, J. R. Bibbins.
Discussed by, G. M. S. Tait, Prof. R. H. Fernald, W. B. Chapman,
H. M. Latham, H. H. Suplee, Edw. N. Trump, H. B. Langer, S. C.
Smith, Prof. Walter Rautenstrauch, G. D. Conlee.
RECEPTION
Thursday, 9 'p.m., Hotel Astor
The Members of New York and vicinity received the membership
of the Society, their ladies and guests, at the Hotel Astor. Dancing
and refreshments followed the reception.
SOCIETY AFFAIRS
389
PROFESSIONAL^SESSION
Friday, December 10, 9. SO a.m.
The Bucyrus Locomotive Pile Driver, Walter Ferris.
Discussed by O. K. Harlan, A. F. Robinson, L. J. Hotchkiss.
LiNESHAFT Efficiency, Mechanical and Economic, Henry Hess.
Discussed by T. F. Salter, Prof. R. C. Carpenter, C. A. Graves,
O.K. Harlan, C. J. H. Woodbury, Walter Ferris, Fred J. Miller, A. C.
Jackson, C. D. Parker, Oliver B. Zimmerman, Geo. N. Van Der-
hoef.
Pump Valves and Valve Areas, A. F. Nagle.
Discussed by Prof. Wm. Kent, A. B. Carhart, Prof. R. C. Car-
penter, E. H. Foster, Chas. A. Hague, I. H. Reynolds, F. W. Sal-
mon.
A Report on Cast-Iron Test Bars, A. F. Nagle.
Discussed by A. A. Cary, T. M. Phetteplace, Prof. W. B. Gregory,
Geo. M. Peek.
COMMITTEES OF THE ANNUAL MEETING
MEETINGS COMMITTEE
Willis E. Hall, Chairman
William H. Bryan
L. R. Pomeroy
Charles E. Lucke
H. de B. Parsons
LOCAL COMMITTEE
William D. Hoxie, Chairman
F. A. Scheffler, Secretary
Wm. L. Abbott
H. P. Ahrnke
Louis Alberger
L P. Alford
G. H. Barbour
G. M. Basford
Edgar H. Berry
Francis Blossom
William H. Boehm
Reginald P. Bolton
Geo. M. Bond
G. I. Bouton
L. P. Breckenridge
Wm. H. Bryan
R. C. Carpenter
H. R. Cobleigh
F. H. Colvin
W. C. Dickerman
Robert M. Dixon
F. L. DuBosque
Frank E. Eberhardt
Harrington Emerson
A. Falkenau
W. H. Fletcher
George J. Foran
E. H. Foster
H. A. Foster
Geo. L. Fowler
R. E. Fox
F. L. R. Francisco
John R. Freeman
H. L. Gantt
Fred J. Gubelman
Willis E. Hall
F. A. Halsey
Geo. F. Hardy
390
SOCIETY AFFAIRS
Henry S. Hayward
G. R. Henderson
F. V. Henshaw
M. L. Holman
W. R. Hulbert
Alex. C. Humphreys
William F. Hunt
F. R. Hutton
F. E. Idell
H. S. Isham
E. B. Katte
R. S. Kent
Walter C. Kerr
Chas. Kirchhoff
J. W. Lieb, Jr.
Henry S. Loud
Fred R. Low
Chas. E. Lucke
R. C. McKinney
F. E. Matthews
E. D. Meier
Fred J. Miller
B. M. Mitchell
Chas. A. Moore
I. E. Moultrop
D. M. Myers
W. W. Nichols
J. H. Norris
H. de B. Parsons
E. H. Peabody
L. R. Pomeroy
H. O. Pond
H. F. J. Porter
W. P. Pressinger
Calvin W. Rice
A. L. Riker
Fred. E. Rogers
H. W. Rowley
W. J. Sando
E. F. Schnuck
Jesse M. Smith
Leo H. Snyder
Albert Spies
E. G. Spilsbury
J. E. Starr
Theo. Stebbins
A. F. Stillman
F. H. Stillman
H. G. Stott
H. H. Suplee
Ambrose Swasey
B. V. Swenson
F. H. Taylor
F. W. Taylor
Stevenson Taj'lor
Edw. Van Winkle
G. T. Voorhees
A. M. Waitt
F. A. Waldron
C. M. Wales
Arthur West
F. M. Whyte
W. H. Wiley
A. L. Williston
Ira H. Woolson
W. L. Clark
W. C. Dickerman
Sub-Committee on Finance
C. A. Moore, Chairman
Alex. C. Humphreys
W. C. Kerr
EXCURSION COMMITTEE
George J. Foran
Percy C. Idell
Hosea Webster, Chairman
Alfred F. Masury
Frederick A. Scheffler
BUREAU OF INFORMATION
F. E. Idell, Chairman
Charles C. Phelps
James V. V. Colwell
ENTERTAINMENT COMMITTEE
Dr. D. S. Jacobus, Chairman
SOCIETY AFFAIRS
391
Sub-Committee on President's Reception
Tuesday Evening
Col. E. D. Meier, Chairman
W. C. Dickerman
Bernard V. Swenson
Francis Blossom
Edward Van Winkle
Louis Alberger
G. M. Basford
Harrington Emerson
Dudley Farrand
W. H. Fletcher
C. H. Foster
Geo. L. Fowler
Willis E. Hall
Geo. F. Hardy
Alex. C. Humphreys
F. R. Hutton
Chas. Kirchhoff
J. W. Lieb, Jr.
R. C. McKinney
Fred J. Miller
B. M. Mitchell
Charles A. Moore
W. W\ Nichols
H. F. J. Porter
Albert Spies
E. G. Spilsbury
John E. Starr
F. H. Stillman
H. H. Suplee
Stevenson Taylor
A. M. Waitt
Ira H. Woolson
Sub-Committee on Reception and Dance
Thursday Evening
Prof. Arthur L. Williston, Chairman
Edgar H. Berry
Wm. H. Boehm
A. P. Boiler, Jr.
Reginald P. Bolton
G. I. Bouton
H. R. Cobleigh
Frank E. Eberhardt
Harrington Emerson
E. H. Foster
Henry S. Hay ward
F. V. Henshaw
F. E. Matthews
David M. Myers
J. H. Norris
E. H. Peabody
H. O. Pond
L. H. Snyder
Theodore Stebbins
A. F. Stillman
J. Stewart Thomson
Edw. Van Winkle
F. A. Waldron
Assignments for the Reception of Members
Tuesday
Afternoon
F. A. Halsey, Chairman
H. P. Ahrnke
H. A. Foster
H. S. Isham
Morning
Albert Spies, Chairman
Percy Allan
C. G. de Laval
F. H. Taylor
Evening
Chas. Whiting Baker, Chairman
W. P. Pressinger
Fred. E. Rogers
H. M. Rowley
Wednesday
Afternoon
G. A. Orrok, Chairman
Geo. B. Caldwell
A. Falkenau
W. R. Hulbert
Evening
H. G. Stott, Chairman
F. H. Colvin
R. E. Fox, Jr.
J. P. Sparrow
S92
SOCIETY AFFAIRS
Thursday
Morning
F. L. Du Bosque, Chairman
L. P. Alford
Geo. H. Barbour
Anson W. Burchard
Afternoon
James T. Whittlesey, Chairman
John J. Boyd
N. B. Payne
R. P. Bolton
Friday Morning
Fred R. Low, Chairman
G. R. Henderson
E. B. Katte
Geo. Dinkel
ladies' reception committee
Mrs. Herbert Gray Torrej^, Chairman
Mrs. H. C. Abell
Mrs. Charles W. Baker
Mrs. G. H. Barbour
Mrs. A. R. Baylis
Mrs. E. H. Berry
Mrs. Wm. H. Boehm
Mrs. R. P. Bolton
Mrs. G. I. Bouton
Mrs. Edward Ciardi
Mrs. J. VanV. Colwell
Mrs. H. Emerson
Mrs. G. L. Fowler
Mrs. F. J. Gubelman
Mrs. F. A. Hall
Mrs. N. H. Hiller
Mrs. H. F. Holloway
Mrs. G. S. Humphrey
Mrs. A. C. Humphreys
Mrs. C. W. Hunt
Mrs. F. R. Hutton
Mrs. F. E. Idell
Mrs. P. C. Idell
Mrs. D. S. Jacobus
Mrs. J. E. Jones
Mrs. J. A. Kinkead
Mrs. G. L. Knight
Mrs. J. W. Lieb, Jr.
Mrs. H. S. Loud
Mrs. Fred R. Low
The Misses Meier
Mrs. C. W. Obert
Mrs. G. A. Orrok
Miss Eugenie Price
Mrs. Calvin W. Rice
Mrs. E. N. Sanderson
Miss Marion R. Scheffler
Mrs. Horace See
Mrs. Jesse M. Smith
Mrs. J. P. Sneddon
Miss Jean Sneddon
Mrs. H. H. Suplee
Mrs. Stevenson Taylor
Mrs. Edward Van Winkle
Mrs. S. E. Whitaker
Dr. Lucy O. Wight
Mrs. Wm. H. Wiley
Mrs. A. L. Williston
Mrs. Jas. Edw. Wilson
Mrs. Ira H. Woolson
Chairmen Committees for the Several Days
Tuesday Afternoon Mrs. John W. Lieb, Jr.
Wednesday Morning Mrs. James Edward Wilson
Wednesday Afternoon Mrs. Ira H. Woolson
Thursday Morning Mrs. J. P. Sneddon
Thursday Afternoon Dr. Lucj"^ O. Wight
„ . , ,, . /Mrs. C. W. Hunt
Friday Mornmg (Mrs. F. A. Hall
SOCIETY AFFAIRS 393
Executive Committee
Mrs. Herbert Gray Torrey, Chairman
Mrs. F. A. Hall Mrs. John W. Lieb, Jr.
Mrs. G. S. Humphrey Mrs. J. P. Sneddon
Mrs. C. W. Hunt Dr. Lucy O. Wight
Mrs. J. E. Jones Mrs. James Edward Wilson
Mrs. Ira H. Woolson
ladies' guides
Mrs. G. S. Humphrey \ chairmen
Mrs. J. E. Jones J
Mrs. Edward Ciardi Mrs. J. P. Sneddon
Mrs. John W. Lieb, Jr. Mrs. S. E. Whitaker
Miss Jean Sneddon Dr. Lucy O. Wight
ACCOUNT OF THE ANNUAL MEETING
OPENING SESSION, TUESDAY EVENING
Vice-President Fred J. MiJler called the session in the auditorium
to order and presented President Jesse M. Smith, who delivered his
address on The Profession of Engineering in which he dealt with
the need of cooperation among engineers, looking toward the mainten-
ance of high standards in engineering practice.
Following the address, Theodore Stebbins, chairman of the Tellers
of Election, presented to the President the report on the election of
oflficers and the following were thereupon declared elected: For presi-
dent, George Westinghouse ; for vice-presidents, Charles Whiting
Baker, W. F. M. Goss, E. D. Meier; for managers, J. Sellers Bancroft,
James Hartness, H. G. Reist; for treasurer, William H. Wiley.
President Smith then called on Past-Presidents Worcester R.
Warner, Geo. W. Melville and Samuel T. Wellman to escort Presi-
dent-elect George Westinghouse to the platform.
After his notification of election and introduction to the members,
the president-elect spoke as follows:
When Mr. Warner, the Chairman of your Nominating Committee, after first writ-
ing on the subject, came to Lenox to ask me to accept the nomination for president
of this great Society, I had already decided that it would be impossible for rae to
have the privilege of accepting; but after he had explained to me the desires of hir
394 SOCIETY AFFAIRS
associates and had represented to me that it was the unanimous wish of all of the
members of your Nominating Committee to honor me at this particular time, and
in so doing to express an appreciation of my efforts and accomplishments in the
engineering field, I with much hesitation consented to accept the nomination and
promised if elected to do everything in my power.
Whether two mistakes have been made — one in yielding to the persuasive words
of Mr. Warner, and the other in my election as your president — the forthcoming
year will determine. I trust I may be able to fulfil your expectations by adding
something to the worldwide reputation of The American Society of Mechanical
Engineers.
With these remarks, I now accept with feelings of deep gratitude the honor
which the members of the Society have tonight unanimously conferred upon me.
There never was a time in the history of the world when honest, wise and con-
servative action is more strongly demanded of us and of all men than now, if \vc
have any desire to preserve the right to comfortably carry on our various
affairs.
I thank you, and I ask your cooperation in my efforts to perform my duties as
your president.
The meeting was then adjourned to the rooms of the Societ}- whein
the members and guests were introduced by Secretary Calvin AA'.
Rice, to the President-elect and Mrs. Westinghouse, who were assisted
in receiving by President Jesse M. Smith and Mrs. Smith, and' Honor-
ary Secretary F. R. Hutton and Mrs. Hutton.
WEDNESDAY EVENING LECTURE
As already stated the lecture on Wednesday evening was on the
Era of Farm Machinery,by L.W.Ellis, of the Bureau of Plant Industry
of the United States Department of Agriculture at Washington, D. C.
The lecture was illustrated by lantern slides. Mr. Ellis first gave an
idea of agricultural progress, by describing some of the most sti'iking
mechanical achievements found on Western farms of the present day.
He first described early farm implements and told briefly of the
transition from hand to machine methods. In 1 SOO wheat was sown
broadcast by hand, after the ground had been plowed with a heavy,
clumsy, wooden plow, requiring as many as eight oxen to pull it.
Sickles cut the grain, and it was bound by hand. During the suc-
ceeding winter it was threshed out either by a flail or by driving
animals over it as it lay in heaps. It was finally winnowed by hand.
Corn cultivation was by the hoe, or a rude shovel plow. The
stalks were cut and the ears husked out by hand. Shelling was done
by scraping the ears against the handle of a frying pan, a bushel in
one hundred minutes.
SOCIETY AFFAIRS 395
Hay was cut with a scythe and was pitched by hand from ground
to cart, and cart to haymow. BaHng and shipping were practically
unknown. Hand methods prevailed in the dairy, the stable, the cot-
ton fields, the potato patch, in fact in every phase of production.
From 1855 to 1894 the human labor consumed in producing a bushel
of corn by the best available methods declined from four hours and
thirty minutes to forty-one minutes, and for shelling it from one hun-
dred minutes to one minute. In 1830, three hours and three minutes
of human labor were requii'ed to raise and thresh a bushel of wheat;
in 1896 ten minutes. Eleven hours were required to cut and cure a
ton of hay in 1860, and but one hour and thirty-nine minutes in 1894.
Power corn shelJers now used have a capacity of from one hundred
to eight hundred ])ushels per day. The cobs are carried to a pile and
the shelled corn delivered into sacks or wagons. The fuel value of
the cobs pays the cost of shelling.
Though hand methods still prevail in some sections, the mower is
now practically the universal means of cutting the hay crop. This
is a modification of the early reaping machines with such factors elim-
inated as are not necessary for cutting the grass. The steel self-
dump rake, the side-deHvery rake and the hay loader, the stacker, and
the baling press are other developments for hay harvesting.
In the extreme West there has been developed the combined har-
vester which seems to represent the greatest possible saving of human
labor. This machine, drawn by from twenty to forty horses, under
control of a single driver, cuts, threshes, recleans, and delivers into
sacks the grain from forty to fifty acres per day. Two men are re-
quired for sewing the sacks. The straw, including all weed seeds, is
distributed over the ground as the team proceeds. On level land the
horses may be replaced by the steam engine, which furnishes power
sufficient to cut a swath up to forty feet in width and to cover from
seventy-five to one hundred and twenty-five acres per day.
For general farm work the internal-combustion tractor may be
said to be rapidly supplanting the steam engine, which, however, has
a great field of usefulness in sections where it is desired to bring large
areas rapidly under cultivation. In older sections, in order to com-
pete successfully with the horse, tractors must bring the cost of ope-
ration close to the cost with horses and at the same time be capable
of a great variety of work. The internal-combustion tractor meets
these conditions better than the steam engine, and is being introduced
at a rate estimated anywhere from two thousand to five thousand
per year.
396 SOCIETY AFFAIRS
The automobile is rapidly finding a place in the business manage-
ment of the farm. It takes from the heavy draft horse the necessity
for long, exhausting trips to town on light errands.
In general, machinery has reduced the cost of producing farm pro-
ducts. It has improved the quality of products by condensing crop
operations within the period when the most favorable conditions pre-
vail. By increasing the acre effectiveness of a man it has reduced
the labor necessary to produce the nation's food supply, leaving it free
to assist in development along other lines. At the same time it has
thrown upon the cities the burden of providing work for an ever-
increasing army of non-producers. It has increased the investment
necessary for the proper organization of a farm, this and the price of
land making it more difficult for a person of small capital to engage in
farming.
As a nation we have occupied nearly all of our naturally productive
area and are confronted with the necessity of providing food for an
increasing population with a constant acreage. In the past, machin-
ery has encouraged extensive rather than intensive farming. Hence-
forth the reverse should be true. If he who makes two blades of
grass grow where one grew before, is a pubhc benefactor, then none
the less is he a pubhc servant who puts into the farmer's hands the
machinery for making such a course attractive.
BUSINESS MEETING
The business session on Wednesday mommg was called to order by
President Jesse M. Smith. Secretary Calvin W. Rice read the annual
report of the Council. The Secretary then read the report of the
Tellers of Election of members, including 166 apphcants for mem-
bership and 21 for advance in grade.
The next in order was the consideration of the proposed amend-
ments to the Constitution. The first amendment relates to C 10
on associate membership, which reads as follows:
C 10 An Associate shall be 26 years of age or over. He must either have the
other qualifications of a member or be so connected with engineering as to be com-
petent to take charge of engineering work, or to cooperate with engineers.
The proposed amendment reads as follows:
An associate member shall be thirty years of age or over; he must have been so
connected with some branch of engineering, or science, or the arts, or industries, that
the Council will consider him quaUfied to cooperate with engineers in the advance-
ment of professional knowledge.
SOCIETY AFFAIRS 397
Another amendment relates to the clause on Junior Membership
which now reads as follows:
C 11 A Junior shall be 21 years of age or over. He must have had such engi-
neering experience as will enable him to fill a responsible subordinate position in
engineering work, or he must be a graduate of an engineering school.
The following addition is proposed by the Committee on Constitu-
tion and By-Laws:
A person who is over 30 years of age can not enter the Society as a Junior.
Both these amendments have been approved by the Committee on
Membership. It therefore remains for the members to vote on them
by letter ballot.
A third proposed amendment to the Constitution relates to the
formation of an additional standing committee. This was presented
at the Washington meeting in the form of a resolution, as follows:
Resolved, That we recommend to the Council the appointment of a Public
Relations Committee, to investigate, consider and report on the methods whereby
the Society may more directly cooperate with the public on engineering matters
and on the general policy which should control such cooperation.
It was moved and seconded that this also be referred to the mem-
bers for letter ballot.
Dr. D. S. Jacobus, Chairman of the Committee on Power Tests, then
made a verbal report. This committee was appointed to revise all
the codes relating to power tests, some of which did not agree with
others, or were not up to date. It had been decided to blend the
whole into one report rather than present a series of reports, as on
engine testing, boiler testing, etc. The first part of the report will
deal with tests in general, caHbration of apparatus, units, etc., while
the second part will be subdivided for the various classes of machines
and apparatus.
Dr. Jacobus also made a verbal report for the Joint Committee on
a Standard Tonnage Basis for Refrigeration. This committee had
made a preliminary report in 1904 and suggested certain units for
measuring the refrigerating capacity of the machinery. They had
also suggested a standard set of conditions under which a machine
should be tested to obtain the refrigerating capacity of that machine.
Later on, the work of the committee was extended, and they were
asked to recommend a method of testing the machines. A prehm-
inary report was also prepared on this portion of the work and had
been before the Society.
398 SOCIETY AFFAIRS
Though the committee had received some favorable discussion on
the report they felt that it was not a complete piece of work, and they
wished that some one would give the committee additional light on
how the report could be made. Furthermore, there were many places
in the report where the committee could not make any definite recom-
mendations, because they did not have enough data at hand.
A resume of the work that has been done by the Committee on Re-
frigeration was prepared and sent to the Congress of Refrigerating
Industries, held in Paris in the fall of 1908, with the request that it be
discussed. In making this resume certain questions were asked, on
which the committee wished to obtain specific information. This
was done in a semi-official way, and after taking up the matter with
the Secretary of this Society, Dr. Jacobus, speaking on the behalf
of the committee, concluded the communication to the International
Committee as follows:
The policy of The American Society of Mechanical Engineers has always been
for the advancement of the arts, and whereas it is only natural that it should take
pride in participating in advancements, it will never look except with satisfac-
tion upon activities of other bodies, even in the subjects on which it has worked.
I feel safe in saying, therefore, that any criticism by the members of this organi-
zation on the work which has been done in connection with the subject at hand
will be gladly received. Criticism leads to the establishment of better and more
up-to-date methods, and what The American Society of Mechanical Engineers is
after, and what I am sure we are all after, is to work hand in hand for the good
of the cause.
I also feel safe in saying that The American Society of Mechanical Engineers
will cooperate in every way in the endeavor to establish some standard set of
rules which shall conform with the views of such able experts as are gathered in
this meeting. It is certainly hoped that the matter presented in this paper
will receive a thorough discussion, irrespective of whether those who take part
agree or disagree with the findings of the committee.
About the same time, a request was made by the committee that
it should be allowed to cooperate with a committee of the American
Society of Refrigerating Engineers, so that if this general committee
recommended certain units, they would really be used by both socie-
ties. A committee of five was appointed by the American Society
of Refrigerating Engineers to cooperate with the committee of five
of The American Society of Mechanical PJngineers. This combined
committee had already held one meeting and sent out a circular letter
to a number of refrigerating engineers, reA'iewing the units that had
been recommended by the Society, and asking for an opinion regard-
ing these specific units. A great number of replies had been received,
SOCIETY AFFAIRS 399
showing how much interest there is in the subject. Most of the re-
pKes said either that the units were acceptable to those who had read
the letter, or that they would leave the selection of the units entirely
in the hands of the committee. The committee therefore has a very
good working basis, and hopes within a comparatively short time to
be able to present the results of its work.
Dr. C. E. Lucke then abstracted the report of the Gas Power Stand-
ardization Committee, of which he is chairman. The report was dis-
cussed by Dr. D. S. Jacobus, Prof. R. H. Femald, A. A. Gary, Edwin
D. Dreyfus and L. B. Lent.
The report of the Gas Power Plant Operations Committee was pre-
sented by F. R. Low in the absence of I. E. Moultrop, chairman of the
committee. The report was discussed by Prof. R. H. Femald, Ed-
win D. Dreyfus, and Arthur J. Wood.
THURSDAY MORNING SESSION
The Thursday morning session was devoted to papers on the meas-
urement of the flow of fluids.
The first paper presented was on Tests on a Venturi Meter for Boiler
Feed, by Prof. C. M. Allen, of Worcester Polytechnic Institute. The
object of these tests with the venturi meter was to determine how
well adapted it would be for use in measuring the feed to a boiler, in
view of the variety of conditions under which it might have to oper-
ate such as the methods of pumping the wate^ through the meter, the
different temperatures of the water pumped, various and fluctuating
pressures and velocities of flow, etc. The results obtained indicate
that such occurrences have practically no effect on the satisfactory
perfoiTnance of the meter. Though there are limits to the satis-
factory operation of a meter, the tests indicate that the venturi
meter is sufficiently accurate for the majority of commercial or engi-
neering requirements.
The paper was discussed by F. N. Connet and Clemens Herschel,
Dr. Sanford A. Moss and Prof. L. S. Marks submitting written dis-
cussions.
The next paper. Efficiency Tests of Steam Nozzles, by Prof. F. H.
Sibley of the University of Alabama, was read by Prof. C. C. Thomas
of the University of Wisconsin. The object of the test was to deter-
mine the efficiency of various shaped nozzles with steam flowing from
a given initial pressure to a known vacuum; also to determine
the effect on the efficiency of changing the angle of divergence.
Two methods were tried out for finding this efficiency: (a) by
400 80CIETT AFFAIRS
the pressure in the nozzle by means of a search tube placed axiaUy
in the nozzle; (6) by the reaction of the nozzle by suspending
it in an air-tight box at the end of a flexible steel tube. The deflection
of the tube caused by the reaction of the nozzle was measured by a
calibrated spring. The results of the tests indicate: (a) that the
reaction is affected by a difference Jh pressure between the muzzle of
the nozzle and the medium surrounding the nozzle; (6) that the effi-
ciencies of the various nozzles were determined within a probable
error of 2 percent; (c) that the efficiency is affected more by the smooth-
ness of finish on the inside of the nozzle than by the exact contour of
the nozzle.
A. F. Nagle, A. R. Dodge and Professor Thomas discussed the
paper, J. A, Moyer submitting a written discussion.
George F. Gebhardt's paper on The Pitot Tube as a Steam Meter
was read by the Secretary in the author's absence. The application
of a pitot tube system as described in the paper is an accurate means
of determining the velocity of steam at any point in a pipe, provided
the values of the various influencing factors are known; and for straight
lengths of piping with continuous flow, under these conditions, it is
an accurate means of determining the weight of steam flowing. Under
average commercial conditions in which the pressure and quality
of the steam fluctuate and an average value must be taken for the
density of the self-adjusting water column, only approximate results
can be obtained, the extent varying with the degree of fluctuation.
Walter Ferris and A. R. Dodge discussed the paper, a written dis-
cussion by Prof. W. B. Gregory being read by the Secretary.
The paper on An Electric Gas Meter was presented by the author,
Prof . Carl C. Thomas, of the University o f Wisconsin. The paper
describes a meter for measuring the rate of flow of gas or air, which can
be adapted for use as a steam meter or as a steam calorimeter. The
operation of the gas meter depends upon the principle of adding elec-
trically a known quantity of heat to the gas and determining the rate
of flow by the rise in temperature of the gas (about o deg. fahr.)
between inlet and outlet. The adoption of this principle of operation
permits the construction of a very accurate and sensitive autographic
meter of large capacity containing no moving parts in the gas pas-
sage; independent of fluctuations in pressure and temperature of the
gas; and capable of measuring gas or air at either high or low pres-
sures or temperatures. The electrical energy required is about 1 kw.
per 50,000 cu. ft. hourly capacity, at the pressures ordinarily used in
gas mains.
SOCIETY AFFAIIto 401
Prof. W. D. Ennis, E. D. Dreyfus and A. R. Dodge discussed the
paper, a wrirten discussion from Prof. L. S. Marks being also read.
THURSDAY AFTERNOON — STEAM ENGINEERING
At the Thursday afternoon session Vice-President L. P. Brecken-
rid c presided. Five papers were presented deahng with different
pha.ses of steam engineering. The first paper, Tan Bark as a Boiler
Fuel, by David M. Myers, described the results obtained by burning
spent hemlock tan bark, the average fuel value of which is about 9500
B.t.u. per lb. of dry matter, which is about 35 per cent of its total
moist weight in the fireroom. The available heat value per pound
as fired is 26G5 B.t.u. One ton of air-dry hemlock bark produces
boiler fuel equal to 0.42 tons of 13,500 B.t.u. coal. A. A. Gary,
Prof. Wm. Kent and Prof. L. P. Breckenridge took part in the dis-
cussion.
J. K. Bibbins then presented his paper on Cooling Towers for Steam
and Gas Power Plants, which contained a critical study of different
types of towers with a description of their distinctive features. The
paper also describes a simple inexpensive type of tower employing a
lath-mat cooling surface and offers suggestions for a combination of
natural-draft and forced-draft types.
The paper was discussed by Geo. J. Foran, W. D. Ennis, H. E.
Longwell, B. H. Coffey, E. D. Dreyfus and F. J. Bryant. A written
discussion by Carl G. de Laval was read by the Secretary.
W. P. Caine's paper, Governing Rolling Mill Engines, was read by
Richard H. Rice. The paper describes and gives indicator cards
and speed curves of a Coriiss engine driving a three-high mill under
two different conditions of governing, (a) under the widest range of
adjustment of cut-off, (6) under a limited range, increasing the econ-
omy and making the engine run much more smoothly and safely. A
table gives the power required for rolling in the mill and the momen-
taiy source of energy, whether from the cylinder or flywheel. A
description is also given of the tachometer used to take the speed
curves. Written cUscussions by H. C. Ord and James Tribe were
read by the Secretary.
The next paper was that by F. W. Dean on An Experience with
Leaky Vertical Fire-Tube Boilers. The author discussed the diffi-
cult.es experienced with some large vertical boilers, somewhat over
10 ft. in diameter, and containing over 6000 sq. ft. of heatmg surface.
The boilers leaked badly very soon after being started and nothing
402 SOCIETY AFFAIRS
that was done improved their condition until the water legs were
lengthened from 2 ft. to 7 ft. 2| in., the boilers thus being raised 5 ft.
2f in. Before they were raised the lower ends of the tubes would
cover with very hard cUnker and become stopped up. This clinker
could be removed only by cutting it off when the boilers were cold.
After the boilers were raised, a Ught clinker that could be blown off
foiTned about the tubes; by removing this by blowing every three or
four hours the leaks were stopped and they have never returned.
Those taking part in the discussion were R. P. Bolton, Prof. Wm.
Kent, J. C. Parker, 0. C. Woolson, A. A. Gary, Prof. A. M. Greene, Jr.,
E. D. Meier and D. M. Myers. A. Bement submitted a written dis-
cussion.
Mr. Dean's second paper. The Best Form of ^Longitudinal Joint for
Boilers, dealt with the defects of the usual form of butt joint used on
the longitudinal seams of boilers, in which the inside strap is wider
than the outside strap. It gave some history of the joint and dis-
cussed some of its defects and suggested a substitution for this form.
The paper was discussed by R. P. Bolton, Carl G. Barth, E. D.
Meier, Prof. A. M. Greene, Jr., W. A. Jones, Prof. S. W. Robinson,
Geo. I. Rockwood. and Sherwood F. Jeter.
GAS POWER^SECTION
The session^of the Gas^Power Section was held on Thursday after-
noon. Chairman F. R. Low presiding. In his address, the Chan-man
referred briefly to the work of the various committees of the Section
and stated that during the year the membership had increased from
247 to 378, a gain of over 50 per cent. Mr. Low also dealt with the
development in the gas-power field during the year, mentioning some
experiments with gas turbines. Gas-engine design, the use of by-
product gases, the development of the bituminous producer, the gas-
ification of peat, and the gas engine in marine work, were also briefly
dealt with.
The report of the Tellers of Election, Edw. Van Winkle, Prof. Walter
Rautenstrauch and J. V. V. Colwell, was then presented by Prof.
Rautenstrauch, the results being as follows: for chairman J. R. Bib-
bins 107; for member of the Executive Committee, F. R. Low 108.
The report of the Gas Power Plant Operations Committee was then
presented by James D. Andrew, and discussed by J. C. Parker, J. N.
Norris and H. H. Suplee. Prof. C. H. Benjamin reported verbally
for the Literature Committee, outlining the work of the committee in
SOCIETY AFFAIRS 403
bringing gas-power literature to the attention of the members. H.
R. Cobleigh and Professor Rautenstrauch also spoke on the work of
this committee, the latter suggesting a plan for better organization
of the committee to deal with literature on the subject.
L. B. Lent reported for the Gas Power Installations Committee
that two forms had been prepared and sent to manufacturers, and
while a good deal of information had been received, not enough was
on hand for a complete report. The committee hoped to have the
material in shape at an early date.
Prof. W. F. M. Goss then presented the paper on Testing Suction
Gas Producers with a Koerting Ejector, by C. M. Garland and A. P.
Kratz. The paper describes a method of testing the suction gas pro-
ducer which is independent of the engine. The engine is blanked off
from the producer and a Schutte & Koerting steam ejector is inserted,
which draws the gases from the producer and delivers them to a scrub-
ber in which the steam used by the ejector is condensed. The gases
then pass to a meter for measuring their volume. Complete data of
calculations and results are given in appendices.
The paper was discussed by Prof. R. H. Femald, G. M. S. Tait, H. H.
Suplee, L. B. Lent, S. C. Smith, W. B. Chapman and Edw. N. Trump.
The paper on Bituminous Gas Producers was then presented by the
author, J. R. Bibbins. The paper describes a double-zone type of
producer and the results obtained in gasifying bituminous coal. Con-
tinuous operation was secured with tar-free gas of reasonable heat
value and producer efficiency and an over-all plant economy of about
one pound of fair bituminous coal per brake horsepower (proportionate
economies for poorer grades). The efficiency and general effec-
tiveness of operation of the producer on low-grade fuel, lignites,
etc., was practically as high as with the higher grades. The following
took part in the discussion: G. M. S. Tait, Prof. R. H. Femald, W. B.
Chapman, H. M. Latham, H. H. Suplee, Edw. N, Trump, H. B.
Langer, S. C. Smith, Prof. Walter Rautenstrauch, and G. D. Conlee.
FRIDAY MORNING
The session on Friday morning opened with the paper by Walter
Ferris on The Bucyrus Locomotive Pile Driver. This paper describes
a new railway pile driver, the leading feature of which is a very power-
ful propelling apparatus and a large boiler, enablmg it to act as a
locomotive and haul its own train of tool cars, boarding cars, etc.,
over the road. A special turntable, consisting of hydraulic lifting
404 SOCIETY AFFAIRS
apparatus and a large ball-bearing, enables the entire pile driver,
including trucks, to be turned end for end or crosswise of the tracks.
0. K. Harlan discussed the paper, A. F. Robinson and L. J. Hotch-
kiss submitting written discussions.
The paper by Henry Hess on Lineshaft Efficiency, Mechanical and
Economic, deFcribed the test of the relative efficiency of a lineshaft
of 2]^ in. diameter, making 214 r.p.m., with bearing load due to
the weight of the parts plus the tension of the belts subjected to known
stress by counterweighting, when running in ring-oiling babbitted
bearings and when mounted in ball bearhigs. The savings in power
consequent on this change ranged fi'om 14 to 65 per cent, with 36
and 35 per cent under average conditions of good practice, due to
belt tensions of 44 lb. and 57 lb. per inch width of single belt respec-
tively. The paper gives data for determining the power savings that
may be expected in various plants, by the use of ball bearings.
Those discussing the paper were T. F. Salter, Prof. R. C. Carpenter,
C. A. Graves, O. K. Harlan, C. J. H. Woodbury, Walter Ferris, Fred
J. Miller, A. C. Jackson, C. D. Parker and Oliver B. Zimmerman.
Geo. N. Van Derhoff submitted a written discussion.
A. F. Nagie's paper on Pump Valves and Valve Areas, called the
attention of engineers to the need of reviewing the common notion
that " valve-seat area " is synonymous with " velocity of flow. " The
purpose of specifications for pumping engines is to secure a low veloc-
ity of flow through the valves, thus reducing the head required to
force water through the pump; but to accomplish this purpose, special
and intelligent attention should be given to the springs of the valves,
rather than to valve-seat areas. If that be done, valve-seat areas
need not be greater than the plunger area' for the vertical triple-
expansion pumping engines so largely used in city pumps. Prof.
Wm. Kent, A. B. Carhart, Prof. R. C. Carpenter and E. H. Foster
discussed the paper. Contributed discussions were by Chas. A.
Hague, I. H. Reynolds and F. W. Salmon.
Another paper by Mr. Nagle, A Report on Cast-Iron Test Bars,
brought out the fact that test pieces, whether cast in separate molds
or in the same mold as the main casting, are not perfect indications
of the character of the iron in the main casting. The results obtained
by the author would indicate a probable variation of 15 per cent
where uniformity might be expected. A. A. Car}' and T. M. Phctte-
place discussed the paper, contributed discussion being by Prof W.
B. Gregory and Geo. M. Peek.
SOCIETY AFFAIRS 405
The meeting closed with the following resolutions, offered by Luther
D. Burlingame:
Whereas The American Society of Mechanical Engineers at its
Annual Meeting, December 1909, desires to express its appreciation
to those who have provided opportunities for entertainment an d on
behalf of the visiting members and their guests thanks for the cordial
welcome extended by the local members and their friends of New York
and vicinity,
Be it Resolved that the Secretary extend the thanks of the So( lety
and express the appreciation of its members and guests to the local
committee for their untiring el'forts, to those who have sent inAita-
tions to visit technical and engineering works and places of inteiest,
to Mr. Geo. Gibbs, chief engineer of the Pennsylvania Tunnel and
Terminal Railroad Co., and to Mr. Walter Kerr, president of the West-
inghouse. Church, Kerr & Co., and their associates, for the opportu-
nity to inspect the new Pennsylvania Railroad station; to Dr. B.T. Gal-
loway, chief of the Bureau of Plant Industry, Department of Agricul-
ture, for the very instructive and entertaining paper on The Era of
Agricultural Machinery, and especially to those ladies who have so
efficiently assisted by extending a generous hospitality to their guests.
EXCURSIONS
As usual at conventions of the Society there were numerous ex-
cursions to points of interest in New York and vicinity, which con-
stituted an important feature of the program for the entertainment
of visiting members and guests. Invitations for these excursions
were generously extended by many firms and individuals, and through
the efforts of the Excursion Committee, Hosea Webster, Chairman,
trips to various plants and industries were arranged, to the represen-
tatives of which the grateful appreciation of the Society has been
expressed.
A list of excursions follows:
Pennsylvania Railroad Terminal and Passenger Station: Invitation by George
Gibbs, Chief Engineer, Pennsylvania Tunnel Terminal R. R. Co., and member of
the Society; Henry R. Worthington Hydraulic Works, Harrison, N. J., by William
Schwanhausser, Chief Consulting Engineer of International Steam Pump Co.,
member of the Society; Ha^-rison Lamp Works of General Electric Co., Harrison,
N. J., by George H. Morrison, General Manager; Interborough Rapid Transit Co.,
central power station at 59th St., New York, by H. G. Stott, Superintendent of
Motive Power, Manager of the Society; Edison factories and Edison Laboratory
at Orange, N. J., by Frank L. Dyer, President of National Phonograph Co., asso-
406 SOCIETY AFFAIRS
ciate member of the Society; De La Vergne Machine Co., New York, by Adolf
Bender, President; New York Telephone Co.; Gramercy and Stuyvesant Central
Offices, by E. F. Sherwood, Superintendent of Traffic; Crocker-Wheeler Co.,
Ampere, N. J., by S. S. Wheeler, President, member of the Society; Westinghouse
Lamp Co., Bloomfield, N. J., by Walter Carey, General Manager; New York Edi-
son Co., Waterside Stations Nos. 1 and 2, by John W. Lieb, Jr., 3d Vice-President,
member of the Society; Astoria Light, Heat & Power Co., Astoria, N. Y., by Wil-
liam H. Bradley, Chief Engineer, Consolidated Gas Co., member of the Society;
BrookljTi Rapid Transit Co., Williamsburg Power Station, by C. E. Roehl, Elec-
trical Engineer; Rockland Electric Co., Hillburn, N. Y.; Singer Building, New York,
by Singer Mfg. Co.; Trenton Iron Co., Trenton, N. J.; Watson-Stillman Co., Am-
pere, N. J.; Metropohtan Life Insurance Building, New York.
Every possible courtesy was extended to the visiting parties in each
case and in some instances special transportation facihties were pro-
vided. At the Edison Laboratory visitors were met by Thomas A.
Edison, Hon.Mem.Am.Soc.M.E., who personally explained many
points of interest about the plant. The Information Bureau, located
in the foyer of the building, under the chairmanship of F. E. Idell,
was of material aid in this connection with the trips.
ENTERTAINMENT FEATURES
The Ladies' Reception Committee, composed of ladies resident in
and about New York, vmder the chairmanship of Mrs. Herbert Cray
Torrey, contributed much to the pleasure of members and guests
(if the Society. Tea was served from four until six o'clock on Tues-
elay, Wednesday and Thursday afternoons during the convention,
in the ladies' headquarters, located in the reception rooms of the
Society on the eleventh floor. Mrs. George H. Westinghouse was
the guest of the committee on Wednesday afternoon.
A number of excursions to shops and hotels were arranged and suc-
cessfully carried out under the guidance of members of the committee
The kindness of Mr. and Mrs. John W. Lieb, Jr., made possible several
enjoyable automobile rides through Central Park and Riverside Drive-
No. 1249
THE ANNUAL REPORT OF THE COUNCIL AND
COMMUrTEES 1909
REPORT OF THE COUNCIL
The Society entered upon a distinct epoch in its history when the
Council approved the recommendation of the Meetings Committee
that meetings of the Society be held periodically in cities other than
New York, thus satisfying a long-felt desire on the part of the
membership, as well as of the Council and the Meetings Committee, to
extend as fully as possible the benefits of membership in the Society.
As a result meetings have been successfully held in Boston and St.
Louis. In the former place, the attendance has been even larger in
some cases than the meetings in New York. The spirit of coopera-
tion has been developed and although these are meetings of the
Society, fellowship in the profession has been promoted in each center
by the participation in the meetings of the membership of local
engineering societies and engineers generally.
Inquiries are constantly being received from other centers for in-
formation respecting the holding of meetings, and every encourage-
ment is being rendered and assistance pledged by the Society to make
it possible for groups of the members in any locality to hold meetings.
Through the policy of conducting these meetings as meetings of
the Society rather than of sections or branches, the solidarity and nat-
ional character of the Society is at once developed. All meetings
are conducted in all places on the same basis with an equally high
standard and before publication in The Journal all papers and dis-
cussions thereon must be approved by the same committee, viz.,
the Meetings Committee; and no papers may be read or discussed
that are not of a uniformly high grade and suitable and worthy
of publication for the benefit of the entire membership.
STUDENT. BRANCHES
The nmnber of student branches affiliated with the Society which
have been formed in colleges and universities during the past year
408 SOCIETY AFFAIRS
show the importance of another movement. Seventeen of these
branches have been established and the reports of their meetings
which have appeared at intervals in The Journal indicate a keen
interest on the part of these organizations and show that here is a
work that the Society may well foster. The basis of affiliation of
these student societies with The American Society of Mechanical En-
gineers is a broad one, and provides for the maintaining of each branch
under its own by-laws subject only to limitations set by the Council of
the Society. The Journal is furnished to each member for the nomi-
nal sum of $2 a year and, in addition, advance copies of papers to be
presented before the Society are supplied gratis for discussion at meet-
ings. Papers for local representation may also be printed and sup-
plied at cost to the affiliated branches. A list of the branches
includes : Stevens Institute of Technology, Hoboken, N. J. ; Cornell
University, Ithaca, N. Y.; Armour Institute of Technology, Chicago,
111.; Leland Stanford Jr. University, Palo Alto, Cal.; Polytechnic In-
stitute of Brooklyn, Brooklyn, N. Y. ; State Agricultural College,
Corvallis, Ore.; Purdue University, Lafayette, Ind.; University of
Kansas, Lawrence, Kan.; New York University, New York; Univer-
sity of Illinois, Urbana, 111.; Pennsylvania State College, State Col-
lege, Pa.; Columbia University, New York; Massachusetts Institute
of Technology, Boston, Mass.; University of Cincinnati, Cincinnati,
0.; University of Wisconsin, Madison, Wis.; University of Mis-
souri, Columbia, Mo.; University of Nebraska, Lincoln, Neb.
HUDSON-FULTON EXHIBIT
The Society's part in the recent Hudson-Fulton celebration was
the preparation of an interesting exhibit of steamboat models, draw-
ings, portraits, books, manuscripts, and other material related to the
development of steam navigation. In making this exhibit, the So-
ciety had the hearty cooperation of the Smithsonian Institution and
of the Hamburg-American line, and was able to place on view
models of early and modern steamboats, the American Museum of
Natural History loaning show cases for this purpose, and members
and friends of the Society also helping to make the exhibit of interest
by loaning or presenting manuscripts, books and drawings. The
American Society of Mechanical Engineers was the only engineering
organization as such taking part in the celebration of engineering
achievement.
Representatives of the Society, together with the Pennsylvania
Society, on September 24th placed a wreath on the Fulton monument
SOCIETY AFFAIRS 409
erected by this Society in Trinity churchyard. The Rev. Dr. William
T. Manning, Rector of Trinity Church, conducted the service.
A description of the improvements in the decorations and the re-
arrangement of the rooms of the Society is contained in the Annual
Report of the House Committee.
The same report contains also a description of the mahogany desk
formerly belonging to Edwin Reynolds, Past-President of the Society,
donated to the Society by Mrs. Reynolds.
THURSTON MEMORIAL
As stated in the Transactions of last year, permission was obtained
from the Alumni Committee of Sibley College, Cornell University,
to place in the rooms of the Society, a bronze replica of the Thurston
memorial tablet at Cornell University. Arrangements for its execu-
tion were made with the sculptor, H. A. MacNeil, a personal friend of
Dr. Thurston, ^.nd the tablet is now in place in the entrance hall.
The figure is about three-quarter life size and below it is the inscrip-
tion.
1839 ROBERT HENRY THURSTON 1903
FiusT President
AMERICAN SOCIETY MECHANICAL ENGINEERS
The committee having the matter in charge were: Dr. Alex. C.
Humphreys, Chairmin, Dr. R. C. Carpenter, Charles Wallace Hunt,
J. W. Lieb, Jr., Fred J. Miller.
The Society was represented by Honorary Vice-Presidents on the
following occasions:
Commencement Exercises of Columbia University, Jesse M. Smith; Inaugu-
ration of Richard Cockburn MacLaurin as President of Massachusetts Insti-
tute of Technology, Worcester R. Warner and Calvin W. Rice; National Con-
servation Congress, Seattle, Wash., R. M. Dyer, M. K. Rogers, W. F. Zimmer-
mann; American Mining Congress, Goldfield, Nevada, Dr. J. A. Holmes; Inter-
national Association for Testing Materials, Chas. B. Dudley; funeral of Edwin
Reynolds, E. T. Adams, F. M. Prescott, E. T. Sederholm, W. J. Sando and
James Tribe ; funeral of F. H. Boyer, G. H. Barrus, F. W. Dean, Gaetano Lanza,
G. H. Stoddard, Dr. C. J. H. Woodbury.
The following resignations were accepted during the fiscal year:
VV. S. Auchincloss, G. W. Blanchard, C. E. Brown, Chas. J. Carney, R. T.
Close, Fred Collins, M. T. Conklin, S. G. Colt, B. J. Dashiell, H. H. Dixon,
410
SOCIETY AFFAIRS
W. L. Draper, Saml. W. Dudley, Thomas Farmer, Jr., W. Flint, M. L. Foucard,
Alex. Gordon, M. M. Green, E. B. Gutherie, O. V. de Gaigne, E. E. Hanna,
W. L. Hedenberg, Jas. Inglis, T. A. Holies, Edmund Kent, C. W. Kettell, C. C
King, F. C. Kretschmer, A. G. Linzee, J. W. Loveland, Jas. H. Massie, F. Mack-
intosh, Alfred Marshall, L. M. Northrup, A. T. Porter, A. S. Pritchard, H. S.
Richardson, L. C. Schaeffer, E. L. Ross, L. N. Sullivan, Marshall L. Whitney,
R. H. Whitlock.
Membership of the following has lapsed during the fiscal year :
M. L. Abrahams, Chas. B. Bruger, H. M. Coale, H. S. Deck, F. H. Davis,.C. M-
Einfeldt. Robt. P. Fritch, J. M. Garza Aldape, A. A. Hale, M. J. Hammers, R.
R. Harkins, L. E. Harper, B. U. Hills, L. A. Holeman, O. H. Klein, D. H. Lo-
pez, Harry G. Manning, Chas. F. Mantine, E. S. Matthews, W. J. P. Moore,
Wm. H. Moulton, A.' W. Mellowes, C. W. Marx, F. J. McMahon, E. C. Patter-
son, F. D. Potter, J. A. Prescott. J. L. Ranch, Fred L. Ray, F. S. Ruttmann,
G. T. Simpson, H. W. Stacy, R. L. Shipman, O. P. Sells, w'm. E. Toelle, W. O.
Teague, Geo. B. Wilson, H. W. Woodward, Chas. H. Young.
The membership has increased during the fiscal year as here indi-
cated :
1
LOSSES
ADDITIONS
1 1
GRADE
1908
Transfer
Resig-
nation
Lapsed
Death
Trans-
fer
Elec-
tion
INCREASE
1909
Honorary
15
1
1
15
Members
2357
18
10
23
35
142
126
2483
Associates
366
11
5
4
4
11
42
29
395
Juniors
786
35
10
11
4
108
48
834
Total
3524
46
33
25
32
46
293
203
3727
AflaUates of C
Jas Power
)tudent Br
Section . .
50
194
150
Affiliates of £
194
The losses by death reported during the fiscal year number the
following:
Honorary Member: Gustav Canet; Members: W. M. Allen, W. H. Bailey,
F. H. Boyer, A. J. Caldwell, K. Chickering, D. H. Gildersleeve, H. F. Glenn,
Thomas Gray, .C. L. Hildreth, W. E. Hill, Robert Hoe, W. S. Huyette, E. H.
Jones, J. Landsing, R. B. Lincoln, Alex. Miller, A. W. K. Pierce, F. A. C. Per-
rine, W. T. Reed, E. Reynolds, R. H. Soule, Geo. W. West, A. R. Wolff; Asso-
ciates: Thomas H. Briggs, Geo. W. Corbin, G. Eberhardt, E. L. Jennings;
Juniors: Albert K. Ashworth, Archibald W. Blair, J. R. Rand, A. E. Wellbaum.
The membership has doubled in the last 11 years. The number of
applications favorably reported during the year 1909 was 290 for
admission. 45 for transfer.
SOCIETY AFFAIKS 411
With the number of men eminent in the profession this is a rela-
tively small increase and on account of the benefits which accrue to
membership and the importance of extending the Society's influence
the members might very properly bring to the attention of engineers
of attainment the desirability of securing membership in the Society.
An amendment to C 45 of the Constitution, involving the appoint-
ment of a Standing Committee on Public Relations, to investigate,
consider and report on methods whereby the Society may more
directly cooperate with the public on engineering matters, and on the
general policy which should control such cooperation, was proposed at
the Spring Meeting and has been approved.
The Committee on Revision and Extension of the Code for Test-
ing Gas Power Machinery, Chas. E. Lucke, Chairman, E. T. Adams,
George H. Barrus^ D. S. Jacobus and Arthur West, was discharged,
and a Committee on Power Tests was appointed by the President,
consisting of D. S. Jacobus, Chairman, Edward T. Adams, Geo. H.
Barrus, L. P. Breckenridge, William Kent, Chas. E. Lucke, Edw.
F. Miller, Arthur West and Albert C. Wood. The purpose of this
committee is to revise the present testing codes of the Society relating
to boilers, pumping engines, locomotives, steam engines in general,
internal-combustion engineS; and apparatus and fuel therefor; to ex-
tend these codes so as to apply to such power-generating apparatus,
as is not at present covered, including water-power apparatus, and to
bring them into harmony with each other and with the best practice
of the day. The committee is empowered to confer with other engi-
neering bodies for the purpose of cooperation.
The Committee on Boiler Code, consisting of J. W. Lieb, Jr. and
Fred. W. Taylor, reported a revision of the Standard Code for Boiler
Tests as desirable in view of the progress made in the art since the
formulation of the code.
A Committee on Standards for Involute Gears, consisting of Wil-
fred Lewis, Chairinan, Hugo Bilgram, E. R. Fellows, Chas. R. Gab-
riel and Gaetano Lanza, was appointed to formulate standards for
involute gears and report to the Council.
The following were appointed members of the Research Committee •
W. F. M. Goss, Chairman, James Christie, R. C. Carpenter, R. H.
Rice, Chas. B. Dudley.
The report of George H. Barrus, P. W. Gates and W. F. M. Goss,
members of the Government Advisory Board on Fuels and Structural
Materials, U. S. Geological Survey, was received and placed on file.
Worcester R. Warner, Chairman, Walter M. McFarland, Morgan
412 SOCIETY AFFAIRS
Brooks, David Townsend and Francis W. Dean were appointetl a
Nominating Committee.
The request of a number of members of the Society for the organiza-
tion of a Machine Shop Section was received and referred for action
to the Meetings Committee, with the suggestion that a sub-committee
to treat the subject be formed rather than a section of the Society.
The invitation extended to the Society by the Institution of Mechan-
ical Engineers of Great Britain, for a joint meeting in England in 1910
has been accepted and a large number of members have already sig-
nified their intention of attending. Arrangements will probably hv
made for the transportation of the party in a single steamer.
The courtesies of the library and rooms of the Society were ex-
tended b}^ the President and Secretary to the Japanese Honorary
Commercial Commission and the professional members attended a
meeting of the Society.
At a gathering of representatives of the four national engineering
societies on April 13, the John Fritz Medal was awarded to Charles
T. Porter, Honorary Member of the Society, for his development of
the high-speed steam engine.
The Society also took a prominent part in the bringing together
in a joint meeting of the four national engineering societies for the
discussion of our national resources. This was the first meeting of
its kind and it is to be hoped that many other occasions will be offered
for cooperation.
FINANCES
The Finance Committee has carefully guided the affairs of the So-
ciety so that notwithstanding increased activities the excess of income
over expense is $4232.79. Of this amount $3010.77 represents 10 per
cent of the reserve fund which for some considerable time in accord-
ance with a resolution of the Council has been transferred annually
from the reserve to the income account.
It is a source of satisfaction to report that the Society is now so
strong that this transfer will be discontinued.
SOCIETY AFFAIRS 41;^
REPORTS OP' STANDING COMMITTEES
Report of the Finance Committee
The Committee submits the statements of the financial condition
of the Society, together with the report of Peirce, Struss & Co., of
New York, certified public accountants, who have audited the books
and accounts.
Peirce, Struss & Co.
Certified Public Accountants
37 Wall Street, New York
November 8, 1909
Mr. Arthur M. Waitt,
Chairman Finance Committee
Dear Sir:
In accordance with your instructions, we have audited the books and accounts
of The American Society of Mechanical Engineers for the year ended September
30, 1909.
The results of this examination are presented in three exhibits, attached hereto,
as follows:
Exhibit A Balance Sheet, September 30, 1909.
Exhibit B Income and Expense?; based on Cash receipts for year ended
September 30, 1909.
Exhibit C Receipts and Disbursements for year ended September 30, 1909.
We beg to present, attached hereto, our certificate to the aforesaid exhibits.
Respectfully submitted,
Peirce, Struss & Co.
Certified Public Accountants
Peirce, Struss & Co.
Certified Public Accountants
37 Wall Street, New York
November 8, 1909
Mr. Arthur M. Waitt,
Chairman Finance Committee
Dear Sir:
Having audited the books and accounts of The American Society of Mechanical
Engineers for the year ended September 30, 1909, we hereby certify that the
accompanying Balance Sheet is a true exhibit of its financial condition as of
September 30, 1909, and that the attached statements of Income and Expenses,
and Cash Receipts and Disbursements, are correct.
Peirce, Struss & Co.
Certified Public Accountants
414 SOCIETY AFFAIRS
EXHIBIT A
Balance Sheet, September 30, 1909
ASSETS
Equity in Societies Building (25 to 33 West 39th
Street) $353 346.62
Equity, one-third cost of land (25 to 33 West 39th
Street) 180 000.00
$533 346.62
Library Books $13 700.60
Furniture and Fixtures 2 966 . 96
16 667 56
New York City 3J % Bonds 1954, Par, $35,000 .... $30 925 . 00
Cash in Bank representing Trust Funds 12 918 . 39
43 843 39
Stores including plates and finished publications 11 600 . 00
Cash in Bank for general purposes $7 444 . 83
Petty Cash on hand 250 . 00
7 694.83
Accounts Receivable
Membership dues $4 924 . 50
Initiation fees 285 . 00
Sale of publications, advertising, etc 4 334 . 55
9 544.05
Advances account of land subscription fund 7 960 . 94
Advanced payments 2 214 . 15
Total assets $632 871 . 54
LIABILITIES
United Engineering Society (for cost of land) $81 000 . 00
Funds
Life membership Fund $35 151 . 07
Library Development Fund 4 902 . 71
Weeks Legacy Fund 1 957 .00 .
Land Fund Subscriptions 1 227 . 88
Robert H, Thurston Memorial Fund 399 . 13
Subscriptions to Annual Meeting 205 . 60
43 843.39
Current Accounts Payable 11 163 . 00
Membership dues paid in advance $494 . 50
Initiation fees paid in advance 50 . 00
544.50
SOCIETY AFFAIRS 415
Initiation fees uncollected $285 . 00
Reserve (Initiation fees) 24 596 . 97
Surplus in property and accounts receivable 471 438 . 68
Total Liabilities $632 871 . 54
EXHIBIT B
Income and Expenses based on Cash Receipts for Year Ended Sep-
tember 30, 1909
INCOME
Membership dues, current $50 273 . 79
Membership dues, arrears 2 355 . 00
Sales gross receipts 8 847 . 39
Advertising receipts 11 997 . 50
Interest and Discount 1 234 . 68
ReserveFund, 10% 3 010.77
$77 719.13
expenses
Finance Committee Office Administration including
Salaries $19 971.91
Finance, United Engineering Society As-
sessments 6 000 . 00
Finance, miscellaneous 983 . 56
^ $26 955.47
Membership Committee 2 392.36
Increase of Membership Committee 147 . 94
House Committee* 1 192.43
Library Committee 2 699. 17
Meetings Committee
Annual Meeting $2 074 . 24
Spring Meeting 1 410 . 52
Monthly Meetings 2 278.19 5 762.95
Publication Committee
Advertising Section The Journal .. $7 026.06
The Journal, except Advertising. ... 13 134 . 80
Pocket List 1 599.59
Revises 523.93
Transactions, Vol. 30. 6 533 . 87
YearBook 1401.30
History 43.65
30 263.20
' From Current Income $1192.43
Reserve Fimd 2500.00
Total Expenses 3892.43
416 SOCIETY AFFAIRS
Research Committee $0 . 58
Committee on Power Test . . 11.25
Sales Expenditures 4 060.99
$73 486.34
Excess of Income over Expenses 4 232 . 79
$77 719.13
EXHIBIT C
Receipts and Disbursements foh Year Ended September 30, 1909
receipts
Membership dues $50 832 . 70
Initiation fees 6 460 . 00
Membership dues and initiation fees paid in advance.. 551 . 00
Sales of publications, badges, advertising, etc 20 833.25
Subscriptions to Land Fund 3 251 .00
Subscriptions to Expense of Annual Meeting 2 188 . 00
Interest 2 072 . 24
John Fritz Medal 123 . 74
Cash Exchanges per contra 575 . 92
. $86 887.85
Cash in Banks and on hand, September 30, 1908 13 708 . 98
$100 596.83
DISBURSEMENTS
Disbursements for general purposes $76 167 . 69
Interest on Mortgage on land 3 240 . 00
Cash Exchanges per Contra 575 . 92
$79 983.61
Cash in Banks and on hand, September 30, 1909 20 613 . 22
$100 596.83
The Committee also submits as called for by the By-Laws a detailed
estimate of the probable income and expenditure of the Society for
the Fiscal year ensuing. This estimate has been submitted to the
careful consideration of each committee concerned and the Finance
Committee has been assured in each instance that the appropria-
tions asked for in the estimate include all needed expenditures to carry
out the work of the different committees as now planned and author-
ized.
It will be noted that the Society is not being operated for profit,
but that practically all of the money received is appropriated for the
development of the Society's various interests, and to enable giving to
eaf^h member a constantly increasing return for his membership dues.
SOCIETY AKFAIK.S 417
ESTIMATE, 1909-1910
Current Income Current Expenses
Dues, Current $52000 Finance Committee $26000
Dues, Arrears 2000 Membership Committee 2400
Reserve Fund, 10 % 2800 Increase Memb. Committee . . 500
Sales gross receipts 5000 House Committee' 1150
Interest 800 Library Committee 2880
Advertising 21000 Meetings Committee 8050
Publication Committee 34900
$83600 Research Committee 500
Executive Committee'"' 600
Power Tests Committee 500
Sales Expenditures 3000
Excess of income over expense 3120
$83600
1 In addition J3000, to be appropriated from the Reserve Fund for the House Committee
for betterments for 1909-1910.
^The appropriation for the Executive Committee for the foreign meeting to be $3000, to be
divided from Current Income at not less than $600 per year for a term of years until can-
celled.
Especial attention of the Council is called to the fact that in con-
nection with entering upon our occupancy of the present refined and
dignified headquarters, a large sum was advanced from the Society's
working capital, known as the Reserve to the Land and Build-
ing Fund from which fund by vote of the Council the interest on the
mortgage for the land is paid. Admittedly the Society cannot afford
to pay for the present headquarters out of its current income unless
the Society is freed from debt ; and it was with the understanding that
sooner or later this debt would be raised, that the Society was justi-
fied and enabled to accept the gift from Mr. Carnegie. During the
past year the total receipts to the Land and Building Fund have
been practically used up for paying the interest on the mortgage, with-
out decreasing the total of the mortgage to the extent of one dollar.
The Finance Committee observes that it has been the custom, by
ruling of the Council, to take 10 per cent of the Reserve Fund
each year to be applied to the payment of current expenses; and^they
recommend to the Council that this custom be discontinued, and -that
the total payments into the Society of initiation fees, which go to
make up the Reserve Fund, shall remain in the Reserve, and that only
by special vote of the Council shall money be expended from this
Reserve.
418 SOCIETY AFFAIRS
The Finance Committee trusts that the time is opportune for the
Land and Building Fund Committee to take steps during the coming
year to raise a portion if not all of the indebtedness amounting to
about S90,000.
It is highly desirable in view of plans for broadening the work
of the Society that our income available for such extension of work be
increased. The organization of our Society is such that the Finance
Committee is charged solely with the responsibility of administering
the Financial affairs of the Society as they find them and not to pro-
duce revenue. All the remaining activities of the Society are for
the expenditure of revenue. The Finance Committee suggests there-
fore that it would be in keeping with good management if a special
committee was appointed to consider the essential feature of the
Society's broader life, viz: the income side, and to see that it is in-
creased to provide for the reduction caused by the discontinuance of
taking 10 per cent annually from the Reserve for operating expenses
and to provide for a broader work in the future,
Respectfully submitted
Arthur M. Waitt, Chairman
Edward F, Schnuck
George J. Roberts \- Finance
Robert M. Dixon [ Committee
Waldo H. Marshall J
Report of the House Committee
The House Committee reports that it has endeavored to make
the headquarters of the Society more attractive, by a rearrangement
of the rooms and by additions to the furnishings.
When the Society entered its new headquarters nearly three years
ago, provisional furnishings were purchased sufficient to carry on the
business of the Society but with no attempt at decorative features.
The original plans of the rooms provided for a large reception hall
which visitors enter from the elevators. In common with the other
floors of the building this hall was open to the main stairway.
A partition cutting off this stairway and another partition separat-
ing the offices has converted this hall into an excellent reception
room.
Sliding doors have been arranged so that the Council Room, the
Library and the Secretary's office give the effect of one large and spa-
cious room.
SOCIETT AFFAIRS ' 419
The walls have been retinted, and new rugs cover the floors. Com-
fortable furniture has been placed in the reception room. There will
be portieres between the rooms, draperies at the windows, and more
comfortable chairs and divans added to the library and Council cham-
ber.
The Committee has aimed to make the rooms homelike and com-
fortable, to make a place which the members will use freely for their
own convenience and for meeting other members or friends for social
or business engagements.
In addition to the large rooms referred to, a small room is especially
reserved where members may attend to their correspondence or hold
private conferences.
Photographs of the Past-presidents have been placed on the walls
of the Library and by order of the Council a similar portrait of each
succeeding President will be added as he retires from office. Name-
plates have been placed on the portraits, paintings and historical
objects, and a very complete catalogue of all these objects of historical
interest has been prepared. This catalogue represents the result of
long and painstaking research on the part of Mr. Edward Van Winkle
of our Committee.
Respectfully submitted,
Henry S. Loud, Chairman '
W. C. DiCKERMAN
B. V. SwENSON )■ House Committee
Francis Blossom
Edward Van Winkle
Report of the Library Committee
During the past year further steps have been taken in the evolu-
tionary process of administering the libraries of the American Insti-
tute of Mining Engineers, the American Institute of Electrical Engi-
neers and that of our own Society, as far as practicable, as a unit.
This process has involved the development of a comprehensive plan
whereby the libraiy of each society maintains only books on sub-
jects in which its membership is particularly interested, treating
all other publications in its library as duplicates. To the American
Institute of Mining Engineers' have been assigned the subjects of
mining engineering, geology, mineralogy, chemistry, metallurgy and a
part of chemical technology. To'the American Institute of Electrical
420 SOCIETY AFFAIRS
Engineers the subjects of electrical engineering, electricity, physics,
mathematics and pure science; and to this Society the' subjects of
general engineering, railroad engineering, mechanical engineering,
civil engineering and a part of chemical technology. This plan has
given satisfaction as a temporary working basis enabling each organi-
zation to complete or supplement imperfect sets from the collections
of the others.
During the year a union card catalogue has been estabhshed, cover-
ing the libraries of the three Founder Societies, which enables readers
to find at a glance all the literature on any subject that may be con-
tained in any of the libraries.
A Library Conference Committee, consisting of the Chairmen of
the Library Committees of the three societies, has under considera-
tion a further important step toward the unification of the three
libraries, involving the organization of the library of the United Engi-
neering Society, to which the three societies shall bear the same rela-
tion as do the Founder Societies in the holding of the United Engi-
neering Societies building and property. Such a plan will enable
gifts of books or periodicals not specifically designated for one society
to be received and taken care of and it may eventually result in the
purchase of books jointly in which the three Societies would have a
common ownership. This plan avoids purchases in tripUcate or
duplicate and concentrates the purchasing power and extension of the
library in a way that will be of undoubted advantage to all who may
have occasion to consult a comprehensive library of engineering liter-
ature, covering all branches of the profession and having available
promptly after publication all the important books.
It is probable that these improvements will necessitate the carry-
ing out of the original building plans for the library, providing addi-
tional shelving in the library room proper, so that all of the volumes
may be readily accessible.
The present status of the Library of The American Society of Mech-
anical Engineers is as follows:
The following titles have been catalogued to date:
Durfee library 570 vol.
A. S. M. E. library 7237 "
Withdrawal of duplicates (not accessioned) 800 "
Pamphlets 1339 "
Total 9946 "
SOCIETY AFFAIRS 421
The additions provided for 1908-1909 and included in tlie above
are as follows:
By gift 168 vol .
By purchase 95 "
By binding of exchanges 197 ''
Total accessions 460
Respectfully submitted,
J. W. LiEB, Jr., Chairman
C. L. Clarke
h. h. suplee
Ambrose Swasey
Leonard Waldo
Library
Committee
Report of the Meetings Committee
To facilitate the work of the present Committee, and it is hoped, of
succeeding committees, a record has been made of its policies and
decisions, some of the more important of which are given below:
The policy of the Committee shall be:
1 Further condensation of papers by the elimination of all superfluous and
irrelevant matter, or matter'previously printed, and of such statements of fact as
are of common knowledge in the profession.
2 The solicitation and selection of such papers, together with the plan of
their presentation at meetings, as may make the Transactions a historical and
up-to-date record of the progress of all branches of mechanical engineering.
3 The presentation of a subject, whenever possible, in such a way as best to
permit of a general and thorough discussion ; and to this end to extend invitations
to those, whether members or otherwise, whose experience has been such as to
bring out the most valuable discussion of the subject.
4 At the Annual and Semi-Annual Meetings, a reduction, when possible, of
the number of professional sessions, and of the number of papers assigned thereto
in order that more opportunity may be given for satisfactory discussion and for
social intercourse between the members. It is the opinion of the Committee that
the professional sessions heretofore have been too crowded.
5 For the sake of uniformity, the adoption of a few rules for the guidance of
authors, these to be based on the experience of the Committee and of the edi-
torial department of the Society, and to offer a review of the rules governing
similar organizations.
6 The adoption of rules tending towards greater uniformity in the actions of
the Commitee; these rules to be such only as concern actions within the juris-
diction of the Committee and subject to such exceptions as in the opinion of the
Committee may seem desirable.
422 SOCIETY AFFAIRS
During the past year, the Committee has submitted to the Council
a number of suggestions relative to changes in some of the methods of
conducting such affairs of the Society as relate to the Meetings Com-
mittee. All of these, with slight modifications, have been accepted
and endorsed by the Council and so far as possible placed in operation.
The selection of a local committee to take charge of all entertain-
ment, apart from the professional sessions, was tried at the last Annual
Meeting with satisfactory results, which we believe long-established
practice will make even better. This is creating greater interest
among the local members, and a feeling of some responsibility for the
entertainment of the visiting members, and places the Annual Meet-
ing upon the same basis as the Spring Meeting, thereby eliminating
what has been heretofore a somewhat inconsistent situation. The
Social and Entertainment Committee will for the first time this year
collect and disburse the fund for this purpose, which will be kept sep-
arate and apart from the^ funds of the Society. This phase of the
arrangement cannot be otherwise than satisfactory.
The resolution of the Committee submitted to the Council, rela-
tive to meetings in mid-season in cities other than New York, was
put into operation immediately upon approval by the Council. In
the opinion of the Committee, this movement is progressing very
satisfactorily and seems to be assuming a natural, healthy growth.
Successful meetings were held at Boston, April 16, June 11, October.
20, and November 17; and at St. Louis, April 10, May 15, October
16, and November 13. This movement, as was desired and antici-
pated, is bringing before the Society much valuable material in the
form of papers and especially of discussion that would otherwise be
inaccessible to tht members. It has resulted in an exchange of
papers, which promises to become more extensive in the future.
The Council's amendment to the Committee's resolution, "subject
to the approval of the Council, " we find from experience to be cum-
bersome. To facilitate these meetings, the Committee must act
promptly upon request from members residing in places other than
New York. With the appropriations for these meetings decided upon
the Committee urges that the Council modify its instructions to the
effect that the Committee may have full authority in compliance with
the original resolution submitted by the Committee to the Council.
The Committee's interpretation that B 21 did not include th§
vouchering of bills covering the expenditures of the appropriations
for its work, has been confirmed by the minutes of the Council of a
few years ago, when the details of such expenditures were placed in
SOCIETY AFFAIRS 423
the Secretary's hands as business manager. The rules governing
office procedure have, however, been changed to define more clearly
this interpretation, resulting in some simplification of the work of
the accounting department.
Last spring a number of members of the Society requested a meet-
ing or conference on the subject of Smoke Abatement. This peti-
tion and the action of the Committee were referred to the Council on
May 28, 1909. This request was for a National Conference with the
elimination of the engineering features as far as possible. After
due consideration the Committee declined to take favorable action.
Subsequent to the above, the Committee received a second peti-
tion asking for a National Conference, but along strictly engineering
lines. In the absence of precedent relative to such a Conference, the
Committee referred the question to the Council. The Committee
has not received, but would gladly receive and carefully consider, a
paper on the subject of Smoke Abatement, if presented along strictly
engineering lines.
We believe the best interests of the Society make necessary a close
working arrangement between the Research and Meetings Committees.
A plan was inaugurated early in the year which it is thought will
bring before the Society more new material than has been heretofore
available. This is accomplished by correspondence with those inter-
ested in original research.
The usual number of meetings were held by the Society during the
past year, all of which are now on record. The Committee begs to
express its appreciation for the assistance and cooperation during the
year of the officers and the several departments of the Society.
Willis E. Hall, Chairman
William H. Bryan
L. R. PoMEROY \ Meetings
Charles E. Lucke j Committee
H. deB. Parsons J
Report of the Membership Committee
During the current year the Membership Committee has held seven
meetings, at which a total of 361 applications for membership have
been considered with the following results:
424 SOCIETY AFFAIRS
Applications void and withdrawn 11
Applications deferred 11
Recommended for membership 339
There were two ballots during the year on which the applicants
recommended by the Committee were voted for. These were at the
Washington meeting 148
New York meeting 187
Total 335
In addition to the most careful consideration which the Secretary
and the Membership Committee can give to the applications for mem-
bership, the cooperation of the whole voting membership is needed
in order to maintain the high standing of the Society. In several
instances during the year action by certain members in giving infor-
mation to the Committee has caused reconsideration of apphcations,
with the result that they have been indefinitely deferred.
A member should not agree to act as proposer or seconder for an
applicant unless he actually knows from his own personal observa-
tion enough of the latter and his work to be able to answer favorably
all the questions on the reference blank regarding him.
The Committee has endeavored to maintain under the By-Laws
the standard of qualifications of applicants for whom they have rec-
ommended to be voted.
The work of the Committee has been greatly facilitated and expe-
dited by the complete and admirable way in which the cases have
been arranged by the Secretary and his staff for presentation to them.
Respectfully submitted,
Henry D. Hibbard, Chairman
Charles R. Richards
Francis H. Stillman >■ Membership
George T. Foran Committee
HosEA Webster
Report of the Publication Committee
The Publication Committee submits herewith the annual report of
its work and of the activities under its control for the past year.
The Committee has held frequent meetings and has earnestly en-
deavored not only to maintain the high standard for the publications
of the Society which has previously been set, but also wherever pos-
SOCIETY AFFAIRS 425
sible to raise the standard to a new level. In its work upon Volume
30 of the Transactions which contains the record of the spring
and winter meetings of 1908, the Committee has given careful study
to the available papers with a view of selecting for that volume only
those of greatest value for permanent record. After due consider-
ation several papers have been omitted and others have been edited
or revised with the approval of the authors. Discussions also have
been edited and in some cases considerably condensed in order to
separate material of permanent value from that which had but a tem-
porary or passing interest.
In compliance with the Resolutions passed by the Council in April
1909, the Publication Committee has undertaken the general super-
vision of The Journal in addition to its other duties, and has adopted
the following general plan for the conduct of this work:
As a general policy, The Journal should be regarded as the news-
paper of the Society and reports of committees, reports of meetings,
professional papers of the Society as a whole or of sections, book
reviews. Society items, etc., should be published as requested by com-
mittees in their official capacity when approved by this Committee,
without charging to the committees or activities concerned any
expense for publication. The Journal has its own expense account
and the appropriation for The Journal should be sufficient to cover
editing and publication of this material.
•No papers, whether for the meetings of the Society as a whole, or
for sections, technical, student or geographical, are to be published
except as formally authorized b}'' the Meetings Committee.
Material from standing committees offered officially will, in general,
be published in the form which these committees desire.
Reports of meetings of the Society and of sections, except when con-
taining strictly professional papers and discussions will, in general,
be published in condensed form.
All matter presented at meetings other than the professional papers
provided by the Meetings Committee, including all discussions, will
be edited under the direction of the Publication Committee. As a
general policy, discussion will be condensed, commercial matter
removed, with a view to presenting only engineering data, opinions
based on experience, historical notes and similar material of value for
permanent record in Transactions.
The advertising section of The Journal which began with the
number of^September 1908, has proven successful. The income from
this source has increased steadily until at the present time there is a
426 . SOCIETY AFFAIRS
gross annual income from it of $21,000; and through the action of the
Council this increased income may be applied to the improving of
the quality of, and to the development of The Journal. Plans for
such development are under consideration, and it is the purpose of
the Committee to make improvements as rapidly as conditions may
warrant.
But the most effective work upon The Journal and that which will
be of greatest benefit to our membership at large is the careful pre-
paration for publication of the professional material presented at the
regular meetings of the Society, and at the meetings of the different
sections. In this great fund of material there is always some that is
unimportant and irrelevant, and much more that could be made of
greater value by skilful editing or by condensation. During the
past year the Committee has done much in this direction that has
resulted in the improved quality of our paper, and also in a consider-
able economy of money, and the papers now appearing in The Journal
are suitable, with little or no alteration, for publication in the Trans-
actions.
In addition to the volume of the Transactions and The Journal the
Committee has issued the annual Year Book of the Society and the
Pocket List of Members.
Respectfully submitted,
A. L. WiLLisTON, Chairman
D. S. Jacobus
H. F. J. Porter \- Publication
H. W. Spangler Committee
g. i. rockwood
Report of the Research Committee
The Research Committee was formally notified of their appoint-
ment under date of April 7, 1909, and at the suggestion of the Presi-
dent, the members were requested to meet during the Spring meeting
of the Society at Washington. Notice was given a short time in
advance of the meeting, and only Prof. R. C. Carpenter and R. H.
Rice were present. These members, however, together with the
President of the Society, Jesse M. Smith, and Charles W, Hunt,
Past-President, and originator of the suggestion that a Research
Committee be appointed, engaged in an informal conference.
S'^A second meeting was called for Wednesday, June 23, 1909, to be
held in New York. There were in attendance the President, Jesse
SOCIETY AFFAIRS 427
M. Smith, R. H. Rice, James Christie, W. F. M. Goss, and the Secre-
tary, Calvin W. Rice. Dr. Coss was chosen Chairman. The Secre-
tary of the Society was recognized as the secretary of the Committee.
The minutes of the informal meeting held in May were read for the
information of the members. After a considerable discussion as to the
scope of the work of the Committee, it was agreed that the Committee
should have information concerning the laboratories of the various
colleges, and other public institutions in America, in which work of
engineering research is proceeding, and to this end tli3 Secretary
was directed to develop a process which would result h: t ! le establish-
ment of such a record in the office of the Society.
It was agreed that the Committee should consider the question of
safety valve efficiency. Arrangements were made for gathering in
existing information upon this general subject, and steps were taken
which will, it is believed, result in a satisfactory outline from which
actual work may proceed. Several other subjects for research, re-
ferred to the Committee by the Council, were laid on the table for
future consideration.
Respectfully submitted
W. F. M. Goss, Chairman
James Christie
R. C. Carpenter \- Research
Richard H. Rice j Committee
Charles B. Dudley J
No. 1250
THE PROFESSION OF ENGINEERING
PRESIDENTIAL ADDRESS 1909
By Jesse M. Smith, New York
President of the Society
Great engineering works existed in raanj^ parts of the world long
before Columbus discovered America. We have but to consider the
ruins left by the Incas in South America and the Aztecs in Mexico to
realize the great work done on this continent in engineering. In
Asia the great wall of China, the temples of Japan, China, Babylonia
and Assyria bear record of the presence of the engineer.
2 In Africa, the vast pyramids of Egypt and the temples on the
Nile are evidences that great engineers existed long before the Chris-
tian era. We marvel still when contemplating the pile of immense
blocks of stone forming the pyramids and try to imagine what form
of apparatus could have been used in placing those great stones one
upon the other.
3 In Europe the Greeks and Romans did marvelous work in roads,
bridges, aqueducts, and various mechanical structures which the
modern engineer may well ponder upon and admire. While we read
much in history of the emperors and kings who reigned when these
great engineering works were produced, we learn little of the men
who produced them, men whom we now call engineers.
4 While engineers have existed for thousands of years it is only
within a comparatively recent time that they have begun to form
themselves into societies for their mutual education and the advance-
ment of the profession of engineering.
5 In England, as early as 1771, Smeaton and his contemporaries
came together to form the Smeatonian Society of Engineers, which,
therefore, according to the calculations of a noted English engineer,
is five years older than the United States. The Institution of Civil
Engineers of Great Britain came into existence in 1818, and was
An address delivered at the Annual Meeting, New York, (December 1909)
of The American Societt op Mechanical Engineers.. .
430 THE PROFESSION OF ENGINEERING
followed by its sister society, the Institution of Mechanical Engineers,
in 1847. La Soci^te des Ingdnieurs Civils de France was founded in
1§48. Der Verein Deutscher Ingenieure was organized in 1856.
6 In this country the Boston Society of Civil Engineers began
its work in 1848. Our elder sister among national societies, the Ameri-
can Society of Civil Engineers, was organized in 1852. The next
member of the family, the American Institute of Mining Engineers,
was born in 1871. Our own Society came into existence in 18S0,
and our younger and very vigorous sister, the American Institute
of Electrical Engineers, came along in 1884.
7 Each of these four national societies, the American Society of
Civil Engineers, the American Institute of Mining Engineers, The
American Society of Mechanical Engineers and the American
Institute of Electrical Engineers, has grown greatly since its organiza-
tion, and each continues to thrive. During the process of upbuilding
of these four great national societies, several other national societies
of specialists in engineering and many local societies of engineers
have been formed, and all of these are also active and thriving.
8 The four greater national societies have an aggregate member-
ship at this time of over 19,000 members. Twelve national socie-
ties of engineering specialists contain more than 13,000 members.
Twenty-three local engineering societies in different cities of the
United States count over 8,600 in their membership.
9 What does this great army of over 40,000 engineers, organized
into many different societies, all for purely professional purposes,
mean? It means that the engineering profession is making Itself
felt in this country of ours, that it proposes to take a prominent place
in the great activities by which the country is being developed, that
it will take its place in public affairs, that it is coing into its own.
10 The national societies are not antagonistic to each other ;^ on
the contrary, they support and give confidence to each other. The
national societies' of specialists are not at war with the other
national societies; they supplement them.
11 The local societies are not in opposition to the national socie-
ties; they extend their influence; they are the outposts of the great
Mrmy. The specialists do not interfere with each other. We are all
specialists to a greater or less extent; but we are all engineers.
12 In the legal profession, some men practice in the criminal
courts; others devote themselves to titles in real estate; others are in
corporation law; others hi patent causes; they all sciuabble with each
other in their practice; but when they meet in their bar association
they arc aU lawyers; they stand by each other and their profession;
they are a power in the world.
THE PROFESSION OF ENGINEERING 431
13 The medical profession is made up of surgeons, oculists,
aurists, general practitioners, specialists of the skin, the heart,
the lungs and every other part of the human anatomy; but when they
come together in their general medical associations they are all
doctors; they also stand by each other and their profession; they
also are a power in the world.
14 In the engineering profession why may not the men who
practice in steam engineering, in machine construction, in hydrauhcs,
in railroad, bridge, mining, electrical and chemical engineering, in
metallurgy, refrigeration, heating! and every other specialty in engi-
neering, come together, stand by each other and their profession,
become known as engineers and be a power in the world?
15 When, in 1SS9, the Institution of Civil Engineers of Great
Britain invited the four national American societies of civil, mining
mechanical and electrical engineers to visit it in London, there was
inaugurated a spirit of friendship and cooperation in the engineering
profession which has grown stronger and stronger as the years have
passed. Following the visit in London, La Soci6t6 des Ing^nieurs
Civils de France, in the same year, invited the American societies to
Paris.
16 Those who were fortunate enough to participate in those
memorable demonstrations of hospitality cannot fail to realize how
greatly the seed of cooperation sown in that year has fructified.
17 In 1900 this Society was again invited by the Institution of
Civil Engineers and the Institution of Mechanical Engineers to visit
them in England, and again invited by the French society to visit
it in Paris. Thus the spirit of cooperation was still further advanced
by these remarkable meetings. On both occasions the sister socie-
ties abroad were untiring in the entertainment of the American engi-
neers.
18 The year 1904 was made memorable^by the acceptance of an
invitation extended by this Society to the Institution of Mechanical
Engineers of Great Britain to hold a joint meeting in Chicago. Thus
the spirit of cooperation and good friendship was again strengthened
and extended.
19 Now the Institution of Mechanical Engineers of Great Britain
has expressed the desire still further to promote this friendly spirit
by inviting this Society to a joint meeting in England in July 1910.
The Council of our Society has accepted this very cordial invitation
of the Institution in the spirit of good will in which it was extended.
It remains for the membership of The American Society of Mechani-
432 THE PROFESSION OF ENGINEEEING
cal Engineers to respond to this spirit and to go to England next
year with its best talent and its best men.
20 The helpful cooperation in professional work which has already
been established with our sister societies over the seas is also be-
coming manifest in our own country. The four national societies
of civil, mining, mechanical and electrical engineers on March 24,
1909, held in this auditorium a joint meeting on the Conservation
of the National Resources, which did much to bring engineers close
together and into cooperative relation.
21 Our Society invited the Boston Society of Civil Engineers to join
in the monthly meetings of the Society recently held in Boston.
The Engineers' Club of St. Louis in like manner was asked to join
with us in the Society's monthly meetings recently held in St. Louis.
In both cases the invitations have been accepted in the best spirit of
cooperation.
22 The engineering societies of the country may be likened to
the members of a large and harmonious family, each member inde-
pendent to do its own special work in its own way, each member
ready to help each of the others, each residing in its own home, but
all ever ready to stand by each other, to work for the common good,
to advance and dignify the profession of engineering.
23 A striking example of the " getting together " of the engineer-
ing societies is found in this building which is the home of our Society.
It is also the home of our sister societies, the American Institute of
Mining Engineers and the American Institute of Electrical Engi-
neers.
24 Under the same roof are grouped together fifteen other socie-
ties of engineering and allied arts. Twenty-five thousand engineers
practicing in all the specialties of engineering may call this building
their professional home. We are hving together here in peace and
harmony. We have brought our books together into a single library
open to the profession and to the public, where every one is welcome.
25 Our meetings are held in the same auditorium and lecture
halls; the doors stand open that all who wish may enter. Our profes-
sional brethren of every society of every country are welcome here.
The large hall at the entrance to the building is a foyer where all
engineers may come together on the same plane, where they may
unite to strengthen each other and to sustain and advance the profes-
sion of which they form a part.
26 The spirit of cooperation which.' now exists must be fostered,
strengthened, made enduring, to the end that as great solidarity will
THE PROFESSION OF ENGINEERING 433
exist in the engineering profession as exists in any of the other great
learned professions.
27 Numbers in membership are, of course, important m the
societies which represent the engineering profession, but a high
standard of membership is of much greater importance.
28 With a considerable number of high-grade technical schools
throughout the country all striving with each other to raise the stand-
ards of engineering education ever higher and higher ; and with the
graduates from these institutions taking, from year to year, a larger
and more responsible part in the great activities of the country, there
is no lack of material from which to form a membership in the en-
gineering societies which will be worthy of the profession.
29 In the Institution of Civil Engineers, as well as in the Institu-
tion of Mechanical Engineers of Great Britain, we are informed, no
person is admitted into the lower grade of membership unless he can
pass a satisfactory examination as to the fundamental principles of
engineering, conducted by an examining board of the Institution.
The rules laid down by this examining board form the standard by
which ihe applicants to membership are measured. If the technical
schools in Great Britain maintain an equally high standard in grant-
ing their degrees in engineering, then the degree may be accepted
in lieu of an examination. In other words, the engineering institu-
tions in Great Britain establish the standard for the degrees granted
by the technical schools. A promotion from a lower to a higher
grade of membership is made only upon a showing of sufficient
experience in engineering to satisfy the rules laid down by the Insti-
tution.
30 In The American Society of Mechanical Engineers, a person
may enter the Society as a Junior upon the presentation of a degree
in engineering from a technical school. But this Society has not,
up to the present, established a standard by which to measure that
degree. I believe the standard for such a degree in engineering
should be established by the Society, and that it should be as high
as that of the best schools of engineering in this country. It will
follow that the schools having a lower standard will soon be brought
up to the higher standard.
31 Promotion to higher grades of membership in our Society is
only made upon a showing of engineering experience satisfactory to
our Membership Committee. This committee is maintaining a high
standard of membership, and I beb'eve that acting under the influence
of the membership and the Council of the Society, it will not allow
that standard to fall, but rather cause it to rise.
434 THE PROFESSION OF ENGINEERING
32 If we are to have a profession of engineering, as distinguished
from the trade of engineer, wc must have a broad education befitting
men of a learned profession, as distinguished from a narrower educa-
tion sufficient for men of a trade.
33 President Lowell of Harvard in his recent remarkable in-
augural address, gave this as his conclusion: "The best type of
liberal education in our complex modern world aims at producing
men who know a little of everything and something well." If that
conclusion be true of a liberal education leading to the learned pro-
fession of the law or medicine or theology, why is it not also true of a
scientific education leading to the learned profession of engineering?
34 If preponderance be given to one part of President Lowell's
conclusion over the other part, certainly knowing "a little of every-
thing" leads to superficiality; while just as surely knowing but one
thing well leads to narrowness. There would seem to be a happy
mean between these two extremes in the education of the engineer.
35 The engineer capable of being at the head of the larger engineer-
ing works must know something of many things, several things well
and one thing profoundly.
36 The engineer, president of a great railway system, for example,
must know something of the alignment and gradients of the perma-
nent way, its construction and maintenance ; something of the proper
location of sidings and stations; something of the system of signals.
of the various kinds of cars, of the quality of water for the locomo-
tives, of the heating and lighting of cars, and many other things.
He must know well that the bridges have been designed for safety
and endurance and that they have been properly constructed. He
must know well that the tunnels are safely protected against
external pressure and falling rocks. He must know well that the
locomotives for drawing the high-speed trains, as well as those for
the heavy freight trains, are of the very best design and capable of
performing their duty with efficiency, economy and endurance. He
must know well how to manage the traffic and keep the accounts.
He must know profoundly how to coordinate all the different parts
of this complex organization so that each part will perform its
proper and full function, to the end that passengers and freight will
be carried safely, surely, quickly and cheaply, and also that dividends
will be paid to the shareholders.
37 The engineer knowing something of many things, several
things well and one thing profoundly, is still one-sided if all this
knowledge is confined strictly to his profession. He will be a much
THE PROFESSION OF ENGINEERING 435
broader man and a better engineer, if in his leisure hours he can turn
his thoughts entirely away from his professional work and toward those
things in nature and art which give that rest and renewal of the pro-
fessional mind necessary to continued work.
38 Engineers have known for many years that tha profession
of engineering is a learned profe'^sion ; the rest of the world is rapidly
arriving at the same conclusion.
39 When in April 1907, this building was dedicated "To the
advancement of Engineering Arts and Sciences," President Hadley
of Yale, where the learned professions have been taught for nearly
200 yea'-s, said:
The men who did more than anything else to make the nineteenth century
different from the other centuries that went before it, were its engineers.
Down to the close of the eighteenth century the thinking of the country was
dominated by its theologians, its jurists, and its physicians.
These were by tradition the learned professions, the callings in which pro-
found thought was needed, the occupations where successful men were venerated
for their brains.
It was reserved for the nineteenth century to recognize the dominance of
abstract thought in a new field — the field of constructive effort — and to revere
the trained scientific expert for what he had done in these lines.
Engineering, which a hundred years ago was but a subordinate branch of the
military art, has become, in the years which have since elapsed, a dominant factor
in the intelligent practice of every art where power is to be applied with economy
and intelligence.
It is encouraging to engineers to have their profession recognized as
a " learned profession " by so great an authority as the president of
Yale University.
40 Enthusiasm and devotion to his profession are characteristic
of the engineer, and from my observation these begin with the
student in engineering and extend throughout his life. President
Wilson of Princeton, in an address at Harvard not long since, dwelt
upon " the chasm that has opened between college studies and college
life. The instructors believe that the object of the college is study,
many students fancy that it is mainly enjoyment, and the confusion
of aims breeds irretrievable waste of opportunity." These conditions,
I believe, exist to a much smaller extent in the technical schools
where engineers are taught, than in the general colleges, where a
liberal education is obtained.
41 Enthusiastic love of work, for his profession's sake, resides in
the heart of the engineer who becomes great. The man who merely
works for wages, and without enthusiasm, does not rise; he remains
a paid servant, and poorly paid at that.
436 THE PROFESSION OF ENGINEERING
42 Where enthusiasm exists, love of work exists; success follows.
Our individual enthusiasm is quickened by the study of the work of
our brother engineers.
43 What engineer while being whisked through the tunnels which
connect Manhattan Island with the lands surrounding it, can fail to
rejoice in his profession as he contemplates the work of the civil engi-
neers, the mining engineers, the mechanical engineers, the electrical
engineers, which, joined together, supplemented each other to
produce success in those marvelous undertakings? The highest
knowledge and skill in each of the four branches of the engineering
profession were called for. and were forthcoming, in the consumma-
tion of this great work. It is not a question which engineers
did the most toward the success of this problem in transportation;
they all did their best; they all did well; each contributed a necessary
part to the success ; they were all engineers working for the advance-
ment of the profession of engineering.
44 Will not every true engineer feel his enthusiasm in his pro-
fession quicken, as he watches the great vessels of trade and the
great vessels of war sweep out to sea, and stops to consider how-
much brains, and long experience, and hard work of many men are
concentrated in each one of them?
45 We marvel still, our enthusiasm is inspired, as we see ponder-
ous steam locomotives and mysterious electric locomotives compet-
ing in the hauling of trains, ever heavier and heavier, ever faster and
faster, and both succeeding.
46 The automobile in its present highly developed and thoroughly
practical form is the result of enthusiastic work of many engineers,
principally within the last fifteen years.
47 The enthusiasm of the engineer is never satisfied. Having
conquered the highway with the automobile driven by the internal-
combustion gas engine, he now proposes to conquer the air with the
aeroplane driven by the same kind of an engine in improved form.
48 The American Society of Mechanical Engineers has before it a
future of usefulness to its members and influence in the profession,
which is unUmited. It only requires that we stand by our tradition
of increasing the membership with men of high quality as engineers;
that the members maintain enthusiastic devotion to good professional
work; that they cooperate with each other in the broadest and most
friendly spirit to produce that solidarity of membership and devo-
tion to high ideals, which will compel the world to class the profession
of engineering with the other learned professions.
No. 1251
THE HIGH-PRESSURE FIRE-SERVICE PUMPS OF
MANHATTAN BOROUGH, CITY OF NEW YORK
DESCRIPTION OF PUMPS AND PUMPING SYSTEM WITH RESULTS
OF TESTS
By Prof. R. C. Carpenter, Ithaca, N. Y.
Member of the Society
The object of this paper is to present a concise description of the
high-pressure pumping system installed for fire service in the city of
New York and the results of a test of the pumping machinery.
2 The system protects the district extending north from City Hall
to Twenty-fifth Street, and east, approximately, from the North
River to Second Avenue. It comprises about 55 miles of extra heavy
cast-iron main, from 12-in. to 24-in. in diameter, with 8-in. hydrant
branches; and two pumping stations so located that they never can
be in the center of a conflagration. At the present time the pumping
stations have a combined capacity of over 30,000 gal. per min.
delivered at a pressure exceeding 300 lb. per sq. in.
THE SOURCE OF WATER SUPPLY
3 The supply of water is ordinarily obtained from the water
mains of the city, which deliver Croton water to the stations at a
pressure of from 14 lb. to 40 lb. per sq. in., depending upon the
demand for water in that district. Both of the pumping stations are
located close to tidal water and connections are made so that sea
water can be obtained in case of difficulty with the Croton supply.
4 The advantage of the Croton water over salt water is that it is
less likely to injure goods, and as the amount required for fire purposes
is only a small percentage of that consumed for the daily supply of the
city its use for fire protection makes no material difference from
financial or insurance standpoints. As this is a matter of consider-
able importance data upon the quantity needed are given in the next
paragraph.
Presented at monthly meetings, New York and St. Louis (October 1909),
of The American Societt of Mechanical Engineers.
438 HIGH-PRESSURE FIRE-SERVICE PUMPS
WATER REQUIRED FOR FIRE PURPOSES
5 The general impression that an enormous quantity of water is
required for fire purposes is erroneous as shown by figures furnished
to Chief Engineer I. M. de Varona by the fire department for the
Boroughs of Manhattan and Brooklyn, years 1900, 1901, 1902, 1903 and
1904, These give the average quantity of water used for fire protec-
tion during these years in the Borough of Manhattan as 74,010,803 gal.
per year, of which 31,056,928 gal. was river water. The daily aver-
age use of Croton water, therefore, for the above five years was 117,-
000 gal.
6 For the Borough of Brooklyn the average for five years was
43,705,568 gal. of which 19,010,928 gal. was river water; daily aver-
age, 67,000 gal.
7 During these five years the greatest quantity used in the
Borough of Manhattan was 99,000,000 gal. in 1901, which included
69,500,000 gal. of river water, leaving 29,500,000 gal. for Croton water,
and Mr. de Varona states (Report of the Department of Water
Supply, Gas and Electricity): "Even if this quantity be made 100,-
000,000 gal. per year, by comparing it with the average daily con-
sumption of about 300,000,000 gal. it will be seen that the total
amount used for fire purposes would be only about one-third of the
amount used for all purposes in 24 hr., forming, therefore, an insignifi-
cant percentage of the total consumption. The quantity needed for
fire purposes (one-tenth of one per cent) may therefore be entirely
neglected as a factor in determining the water supply of the city.
8 "The capacity of each of the pumping stations will be for the
present 15,000 gal. per min. or 43,000,000 gal. per day for the two sta-
tions. By the installation of three additional units in each station,
for which provision is made, this capacity can be increased in round
numbers to 69,000,000 gal. per day.
9 "The two stations, with the motors and pumps as installed,
have a total capacity in excess of that of all the fire engines in the
Boroughs of Manhattan, the Bronx and Brooklyn working under
normal conditions. This comparison assumes the engines to work
on one line of 2^-in. hose, say 500 ft. long, under a pressure of, say
200 lb., and with the capacities as printed in the official blank forms
of the reports of the fire department. It should furthermore be
remembered that provision is made for the installation of still another
pumping station."
HIGH-PRESSURE FIRE-SERVICE PUMPS 439
MOTIVE POWER
10 The power for driving the pumps is transmitted electrically
from several of the electric power and lighting systems located on
Manhattan Island. As the stations of systems are widely separated
and any or all of them are available for motive power the system of
electric transmission was considered more reliable in the case of a
large and general conflagration than power plants maintained directly
at the pumping stations. Each station is provided with two inde-
pendent sets of transmission lines located as far as possible beyond
danger or injury in case of a great conflagration.
11 The cost of erecting and maintaining an independent power
plant would have entailed a greater annual charge than the cost of
the electric current; consequently the present arrangement is advan-
tageous from a financial standpoint.
12 In addition to the charge per kilowatt for the current delivered
there is a charge aggregating $90,000 per year for reserving the first
right of use for the necessary generating machinery for this purpose.
The total cost of maintenance of the system is estimated at $170,000
a year, which amount it is believed will be saved many times over by
a reduction in insurance premiums now paid in the protected district.
13 The electric current is supplied at a pressure of 6600 volts
from the following stations of the New York Edison Company, hav-
ing the capacity indicated: 53 Duane Street, 7600 kw.; 115 East 12th
Street, 1700 kw.; 45 West 26th Street, 400 kw.; 140th Street and
Ryder Ave., 4000 kw.; Waterside Stations No. 1 and No. 2, 196,700
kw. In addition there are feeders extending to the Brooklyn Edison
Company stations which can be called on in case of an emergency
demand.
14 The pumping stations are connected to 18 sub-stations,
equipped with rotary converters and storage batteries, aggregating
a capacity of 124,000 ampere hours at 135 volts, ah enormous reserve.
15 Each station is connected with the main stations of the Edison
Company by two 250,000 cm. three-phase cables laid in ducts, and
two independent reserve feeders extend to the sub-station system
of the Edison Company. With all these precautions, interruption
of the power supply would seem a physical impossibility.
THE DISTRIBUTION SYSTEM
16 The following information upon the distribution system is taken
largely from the department report of Chief Engineer de Varona
440
HIGH-PRESSURE FIRE-SERVICE PUMPS
Hydra n-f yv/fh connection
far ^f reef f /cashing hi/d9.
Fig. 1 Showing Location of Stations and Areas Covered by High-
Pressure Pumping System
the area indicated is served by a system op mains ranging from 24 in. to 12 in. in
DIAMETER "WITH 8-IN. HYDRANT CONNECTIONS
HIGH-PRESSURE FIRE-SERVICE PUMPS 441
for 1905. Fig. 1 shows the system to be bounded by mains laid on
the north through Twenty-third Street; on the east, through Broad-
way to Fourteenth Street, through Fourteenth Street to Third
Avenue, down Third Avenue to the Bowery, down the Bowery to
Chambers Street; through Chambers Street on the south to West
Street; and on the west through West Street.
17 The area actually protected is considerably greater than this
as hose can be extended over a zone 600 ft. wide beyond the limits
of the mains.
18 This district was selected as that in which the fire losses were
the greatest and which most urgently needed fire protection. Plans
have been prepared for the extension of the system southerly to the
Battery, easterly as far as the East "River, and, if necessary, northerly
as far as Fifty-ninth Street, by the simple extension of the mains
and probably the erection of a third pumping station.
19 The pipes, castings and hydrants were tested at a pressure of
450 lb. The specified allowance for leakage in a 10-min. test was
at the rate of 4 gal. in 24 hr. for each lineal foot of pipe joint, equiva-
lent to a leakage of 487,000 gal. for the whole system in 24 hr,, which is
somewhat over one per cen I; of the total specified pumping capacity now
installed. The actual leakage on test was at the rate of 264,000 gal.
per day or about six-tenths of one per cent of the pumping capacitv.
Considering the difficulties of construction and the high pressure, the
results attained were remarkable and reflect great credit on the
engineer in charge.
20 There are sufficient hydrants so that if a block were on fire 60
streams of 500 gal. per min. each, or the full capacity of both stations,
could be concentrated on a block with a length of hose not exceeding
400 ft. to 500 ft., assuming the use of 3-in. hose and l^-in, nozzles.
21 The layout of the mains at the stations both for suction and
delivery is on the loop system; that is, the supply can be taken from
either one of two mains, and the discharge is through either one or
both of two mains. With this system even the breakdown of one
of the discharge mains at the station would only slightly reduce the
pressure at the fire and would not affect the capacity of the station,
as the pumps would be capable of forcing their full discharge through
the short length of a single 24-in. main if made necessary by such an
accident.
22 The mains are of cast-iron, bell and spigot pipe, of the thick-
nesses given in the following table:
442 HIGH-PRESSURE FIRE-SERVICE PUMPS
Unit Tensile Strain
Size of Pipe
Thickness
with 300 lb. pres-
Inches
Inches
sure
Factor of Safety
24
n
1920
0.4
20
a
2000
10.0
16
li
1920
10.4
12
1
1800
11.1
8*
i
1371
14.6
* Used only for hydrant branches.
SUPPLY PIPING
23 At the South Street Station the fresh water supply is derived
from t^^o 30-in. lines, one connected at Chestnut Street to the 36-in,
line on Madison Street, and the other connected at Pike Street to
the 36-in. line on Division Street. These two main feeders, to which
the two 30-in. lines are connected, increase to 48 in. in diameter and
extend independently and directly to the Central Park Reservoir and
are also reinforced by connections with the main feeders in this sec-
tion of the city.
24 An auxiliary salt-water supply, consisting of two 36-in. pipes
about 140 ft. long, brings the salt water from the East River to a
suction chamber located directly in front of the pumping station.
This suction is so constructed that the pipes are always below mean
low water, thus insuring a supply at all times and avoiding the possi-
bilit}^ of a break in the suction caused by air getting into the suction
lines. On the river end of this suction there are constructed heavy
bulkhead screens and in the suction chamber are two sets of bronze
screens which are readily accessible for cleaning. From the suction
chamber there are taken two 30-in. flanged mains to the duplicate
set of mains in the pumping station proper. The vacuum in these
30-in. pipes is maintained by automatic electric vacuum pumps
located on the pump room floor of the station.
25 At the Gansevoort Street Station the fresh-water supply is
derived from two 30-in. mains, one connected at Twelfth Street to
the 4S-in. line on Fifth Avenue, which runs direct to Central Park
Reservoir, and the other connected to the 36-in. line on Ninth Ave-
nue at Little West 12th Street, which increases to a48-in. line and runs
also direct to the Central Park Reservoir. These two main feeders,
in addition to having their supplies direct from Central Park Reser-
voir, are also reinforced by connections with the main feeders in this
section of the city.
HIGH-PRESSURE FIRE-SERVICE PUMPS 443
26 The salt-water suction lines for this station are practically
identical with those for the South Street Station except that the
36-in. lines from the North River to the station are 650 ft. long.
PUMPING STATIONS
27 The two stations, known as the Gansevoort pumping station,
located near Gansevoort Market on the North River, and the South
Street station, located on the corner of Oliver and South Streets near
the East River, are identical in construction and equipment. The
buildings are of simple design, of steel fire-proof construction, with
concrete foundations. The Gansevoort Street building, which is
typical of both, is one story high with basement, 63 ft. 8 in. by 97 ft.
4 in. Each station is large enough for eight pumping units.
MACHINERY
28 There are now five units in each station consisting of Allis-
Chalmers five-stage centrifugal pumps driven by AUis-Chalmers
induction motors and the necessary auxiliary machinery. The
motors and pumps are alike and their parts are interchangeable.
29 The pumps each have a specified capacity of 3000 gal. per min.
of sea-water, working with a suction lift of 20 ft. and a delivery
pressure of 300 lb. per sq. in. The actual capacity as indicated by a
24-hr. test was about 30 per cent in excess of that specified. The
original specifications contemplated the use of six-stage pumps, with
the expectation that sea-water would be used at each fire. Because of
the facts already referred to (Par. 4), that the relative amount of water
required for fire purposes is insignificant and that sea-water may do
considerably more damage to goods than fresh water, a change in the
specifications was agreed to, whereby the pumps should work at best
eflBciency when receiving water from the Croton mains at a pressure
on the intake side varying from 15 lb. to 40 lb. per sq. in.
30 To meet this new condition the pumps were all built with five
stages. All the sea connections and priming machinery as originally
contemplated were installed, so that sea-water can be pumped into
the mains whenever desired. The effect of the change is merely to
reduce the pressure head slightly in case sea-water is used.
ARRANGEMENT OP MACHINERY
31 The floor-plans of the buildings and general layout of machin-
ery, piping, switchboards, etc., are shown in Fig 2. As will be seen
444
HIGH-PRESSURE FIRE-SERVICE PUMPS
Fig. 2 Plan and Elevation Showing Arrangement of Hydraulic and
Electrical Apparatus in Pumping Stations
HIGH-PRESSURE FIRK-SERVICE PUMPS
445
space is provided for three additional units. Working detail plans of
the machinery were furnished by the contractor. The arrangement
shown in Fig. 2 is the same for both stations, the only difference being
that the switchboard and office in the South Street station are on
different sides of the building as compared with the Gansevoort
Street station.
32 The motors and pumps, with suction and delivery branches,
are located on the main floor of the pump room. The switchboard
and switchboard apparatus are placed in an enclosed two-story and
basement gallery.
Fig. .3 Interior View of Station
33 The four high-tension feeders and all other wires entering the
building are brought in through the gallery basement. All terminal
work on the entering wires is located in the basement. On the first
floor of the gallery, which is approximately on the same level as the
pump-room floor, are placed the oil switches, with their controlling
and protective devices, fire-proof cells and compartments.
34 The operating switchboard is conveniently located in the
enclosing wall of the gallery, and is so placed as to allow a man
standing on the pump-room floor to perform all the operations neces-
sary for controlling the apparatus in the station. The bus bars,
446 HIGH-PRESSURE FIRE-SERVICE PUMPS
with their fireproof compartments, are placed on the second floor of
the gallery.
MOTORS FOR CENTRIFUGAL PUMPS
35 The motors are of the constant-speed, wound-rotor induction
type, 3-phase, 25-cycle, 6300-volt to 6600-volt, designed to operate
at about 740 r.p.m. Each pump is direct-connected to its motor
by a flexible coupling which takes care of any variation from align-
ment. In starting, an iron grid resistance is connected in the
secondary circuit and gradually cut out by means of a handwheel
on the motor switchboard panel. When the resistance is all cut
out the rotor is automatically short-circuited and operated by
specially constructed solenoids through a small switch mounted
directly on the shaft of the handwheel above referred to. An
interlocking arrangement prevents the operator from closing the
switch connecting the motor to the line while the motor is short-
circuited.
36 The specifications required the motors to have sufficient
starting torque to attain full speed between 30 sec. and 45 sec. after
starting, with a current not exceeding 150 per cent of that used when
the motor is working under full speed. Each motor was required to
develop not less than 800 b.h.p. when using current of 6300 volts,
25 cycles, and under these conditions to have an efficiency not less
than 92 per cent, a power factor not less than 93 per cent, and a
motor slip not in excess of 2 per cent. At three-quarters load the
efficiency was not to be less than 92 per cent and the slip not to exceed
1.5 per cent. It was specified that the temperature of the motors
should not rise more than 40 per cent on a 24-hr. test at full load,
when measured by a thermometer, the air in the room being 25 deg.
cent.
37 Prof. Geo. F. Sever of Columbia University tested two of the
motors in the shops of the contractor and found them to meet the
specifications and to have a full-load efficiency of 93.2 per cent. The
other motors were inspected and found to be alike and were assumed
to have the same efficiency. The motors were also tested for tempera-
ture rise at the time of the official test to be described later.
MOTORS FOR AUXILIARIES
38 Direct-current motors of 240 volts are provided to operate the
various gate valves in the station and the piston pumps employed foi-
maintaining a vacuum on the salt-water "suction lines.
HIGH-PRESSURE FIRE-SERVICE PUMPS 447
PUMPS
39 As previously stated the pumps were finally constructed with
five stages, each to give a pressure of somewhat over 60 lb. per sq.
in., making the combined pressure of the five stages about 300 lb.
per sq. in. above the intake pressure, which is the maximum working
pressure of the stations at normal speed of 740 r.p.m. This tj'pe of
pump is the simplest now on the market for pumping water either
against a high head or low head, and this simplicity was the deciding
factor which led to the selection of this style of machinery.
40 The pumps are water-balanced by a piston connected to the
last impeller and upon which the water pressure acts, but should
any additional end-thrust occur, it would be taken up by the ball
• bearing provided in the outboard bearing. This ball bearing consists
of two rings of l|-in. diameter steel balls and is water-cooled. The
balancing piston is fitted very loosely in order to keep the friction
losses small, and as a result a considerable amount of water leaks past
it into a chamber at the end of the pump, which is provided with a
discharge pipe and valve leading into the suction. By adjusting
the valve in this pipe the difference of pressures on the piston can be
regulated as desired. The bearings are of the ring-oiled type and are
separated from the pump casing by packing glands which prevent
foreign matter from entering the bearings. The impellers are of
bronze and the shaft of forged steel. All parts of the runners and
diffusion vanes are thoroughly lubricated by oil cups on the base of
the pumps. A feature is the wide base, shown in Fig. 4, which allows
the pump barrel to set low, giving stability.
41 Each combined unit is equipped with automatic and hand
control. The pumps are kept primed for instant service and the
simple operation of a switch on the main switchboard starts the
machine and gives full pressure in about 30 sec.
PRESSURE -REGULATING VALVES
42 A combined regulating and relief valve is interposed between
the discharge pipe and the suction pipe of each pump, and set to regu-
late the discharge of each pump to any predetermined pressure.
43. When the^volume of the water discharged by the pump is in
excess of that forced into^the system, this valve acts as a relief valve
and by-passes this excess into the suction to the pump, the pres-
sure on the main distribution system remaining at the predetermined
448
HIGH-PRESSURE FIRE-SERVICE PUMPS
point. When no water is forced into the distribution system all of the
water discharged from the pump is then by-passed into the suction.
44 The pressure-regulating valves were made by the Ross Valve
Mfg. Co., of Troy, N. Y., and much of the practical success of the
station has been due to the accuracy with which they maintain anv
desired pressure.
■ ■ ' i 1
V -
^* ' —
.1 » «
liJI" . ^,M^
WB^'^A ^^^^
k\ mWir'>Y^'.'«
f^:-^l
\mkM^& S
^ "Wt^
M
k^
^j
Fig. 4 Multistage Pump, Capacity 3000 Gal. per Min.; Maximum Head,
300 lb. per sq. in.
PRIMING APPARATUS FOR SALT-WATER SUCTION LINES
45 The priming apparatus In each station consists of three motor-
driven vacuum pumps, each arranged to maintain automatically a
vacuum of 26 in. in the suction lines. These pumps are of the piston
single-action type, one having a displacement capacity of 300 cu. ft.
per min. for a piston speed of 200 ft. per min. and each of the others
a displacement capacity of 50 cu. ft. with a piston speed of 160 ft.
per min.
46 An air-collecting chamber is connected to each of the salt-
water suction lines and equipped with water-gage glass and vacuum
gage. The air-suction piping Ijetween the air chambers and the air
pumps is provided with a veitical loop sufficiently high to prevent
HIGH-PRESSURE FIRE-SERVICB PUMPS
449
u-^
450 HIGH-PRESSURE FIRE-SERVICE PUMPS
water being carried over to the pumps. The air pumps are inter-
connected to each air chamber.
VENTURI METERS
47 Venturi meters for measuring the discharge of water from the
station and from one main to the other were set by the contractor on
each discharge main and on the cross-connecting main. The meters
of the discharge main are 24 in. in diameter and on the cross-over
main 12 in. in diameter. These meters were standardized under the
direction of F. N. Connet, Manager of the Venturi Meter Sales De-
partment of the Builders Iron Foundry, Providence, R. I., and were
provided with dial-indicating gages and also chart-recorders gradu-
ated to indicate the flow in gallons per minute; and in addition
with an integrating meter which registers the total flow in gallons.
48 The readings during the test were taken by a mercury mano-
meter, graduated to show the capacity in thousands of gallons per
minute. For this purpose a Venturi manometer was attached with a
temporary connection to each of the 24-in. Venturi tubes. The
manometer gave essentially the same reading as the indicating dial
on the main register.
49 The Venturi manometer is practically a tube partly filled with
mercury, one side of which communicates with the upstream pressure
chamber of the meter tube, while the other communicates with the
throat-pressure chamber. The connections with the manometer
are indicated in the diagram, Fig. 6.
50 The sketch shows a 24-in. high-pressure meter tube, its register-
indicator-recorder and manometer. The instruments and meter tube
are drawn to scale, but in the pumping station the meter tube is about
75 ft. distant from the instruments.
TESTS OF MACHINERY
51 The specifications for the pumping system provided for an
endurance test of each motor and pump lasting 24 hr. without stop,
during which time the capacity and eSiciency of the pumps and
motors were to be determined. The tests were to be in charge of an
expert appointed by the commission.
52 The specifications provided for making the test with sea water,
but this was later changed to a test with Croton water under the con-
ditions of actual use. In view of this change the contractor increased
the efficiency guarantee from 70 to 71 per cent.
HIGH-PRESSURE FIRE-SERVICE PUMPS
451
452 HIGH-PRESSURE FIRE-SERVICE PUMPS
53 The original specifications called for a capacity of 3000 gal. of
sea water per minute against a discharge pressure of 300 lb. per sq.
in. and a suction lift not exceeding 20 ft. The total increment of
pressure is equivalent to 308.66 lb. from the intake to the delivery
side. The Croton pressure varies at the stations in different parts
of the day from about 40 lb. to 13 lb. per sq. in. and is affected by the
amount of water being drawn from the mains. Consequently, to
meet the requirements, the delivery pressure would need to be
308.66 lb. in excess of the intake pressure. There is also a further
correction from the fact that sea water is heavier than fresh water
and this correction under maximum conditions might amount to 2.5
per cent.
54 The specifications further provided that the brake horse-
power developed by the motors under test should be computed from
the electrical energy supplied to them, corrected for the efficiency
of the motors as determined by the test. They further provided that
if the aggregate of all stops exceeded one hour for any motor the test
for capacity for such motor was to be run over again for a period
of 24 hr.
55 The specifications also provided that the pumping capacity of
the apparatus and the efficiency of the pumps should be based on
the minimum rate of pumping during any eight consecutive hours of
the endurance test, during which none of the motors were stopped.
56 The discharge of the pumps was determined by the reading
of the Venturi meters, one of which was located in each discharge
fine. These readings were under the direction of F. N. Connet, and
were checked by observers representing the contractors and also the
city.
57 The modified specifications also required that the efficiency of
each pump should be not less than 70 per cent and its capacity not
less than 3000 gal. of sea water when lifted to a pressure equivalent
to 308.66 lb. To determine whether the requirement was met, a
separate test of each pump was required.
58 The efficiency of the pumps was computed by dividing the
horse-power output of the pumps by the horse-power input as
received from the motors. The horse-power input was computed
as follows:
1 • X total wattt^ ^ . f X /r^o o i.N
n.p. input = X efnciencv of motors (93.2 per cent)
746
HIGH-PRESSURE FIRE-SERVICE PUMPS 453
The horse-power output was computed as follows: h.p. output
_ wt. per gal. (8,34) X 2.31 head in pounds X no. of gal. per min.
. 33000
SOUTH STREET STATION TEST
59 The test of the South Street Station was begun at 12:30 p.m.
on September 2, 1908, after about 2 hr. of preliminary running for
the purpose of adjusting the delivery pressure; it was continued with-
out interruption for 24 hr. With the exception of a short stop of
motor No. 2 which was shut down from 2:11 to 2:41 a.m., September
3, to remedy a slight defect in the^^insulation of the field coils, no pump
was stopped. During the time No. 2 was stopped the pressure on the
delivery mains fell to about 300 lb. ; during the remainder of the test
the pressure was maintained at or above the contract requirement,
as will be noted by consulting the last column of Table 1.
60 The average results for each hour for the 24-hr. test of all four
motors are given in Table 2. The smallest delivery for eight con-
secutive hours occurred at the last part of the test, when the
average capacity, as shown by the readings, was 18,447 gal. per
min., and the average efficiency was 72.2 per cent. During this
time the average pressure pumped against was 314.5 lb., or an excess
of about 6 lb. over contract requirement.
61 It will be noted from the last column of Table 2 that there is
considerable variation in the efficiency; that during the first hour the
efficiency was less than 70^ per cent, whereas during the third and
fourth hours the efficiency ^^exceeded 75 per cent. This variation
in efficiency was doubtless caused by variation in the amount of water
by-passed from the pressure to the suction side of the pump over the
balanced piston and through the bearings, and possibly during the
first hour by the discharge of some water through the relief valve
which was pumped but not metered. The valves for regulating the
differential pressure on the balance pistons were nearly closed during
the third, fourth and fifth hours of the South Street Station test, but
were opened the normal amount for the remaining portion of the
test. The amount of water which for maximum difference of pres-
sure may leak around the balance piston of any pump without passing
through the meter could not be accurately determined but was esti-
mated to be in excess of 4 per cent. Hence it appears that slight
changes in the opening of the valve controlling the differences of pres-
sure at this piston must materially affect the efficiency. The normal
454
HIGH-PRESSURE FIRE-SERVICE PUMPS
opening of this valve appears to correspond to an efficiency of about
72.5 per cent.
62 During the test of the South Street Station all the bearings
ran cool with the exception of those on No. 6 pump, which heated up
during the third and fourth hours but were brought to a normal con-
dition without stopping the pump or reducing its load by the appli-
cation of lubricants and cooling water.
TABLE 1 HOURLY AVERAGE OF READINGS OF DISCHARGE AND INJECTION
GAGES ON PUMPS
SoDTH Strekt Pumping Station, September 2 and 3, 1908
Hour
Pump No. 6
Pump No. 4
Pump No. 2
1
Pump No. 1 Pump No. 3
1 ^
Average
Net
Pres-
1
sure
Disc.
Inj.
Disc.
Inj.
Disc.
Inj.
Disc.
Inj.
Disc.
Inj.
Disc.
Inj.
Lb.
i
12:30- 1:15 335.0
21.8
332.3
22.6
329.8
22.5
331. C
24.0 332.9
23.3
332.2
22.8 309.4
1:30- 2:15 347.2
20.9 347.0
22.2
345.5
22.1
343.8
22.6 346.2
22.1
345.9! 22.0 323.0
2:30- 3:15 345.0
20.9 343.8
21.8
342.8
22.1
341.0
22.5 343.4
21.8
343.2 21.8 321.4
3:30- 4:00 344.2
20.9 343.8
21.9
340.3
22.3
339.8
22.6 342.2
21.9
342. li 21.9 320.2
4:30- 5:00 341.7
21.1 341.8
21.9
340.8
22.6
339.3
23.1 342.2
22.4
341.2 22.2 319.0
5:30- 6:00 336.2
22.4 336.3
22.9
335.3
23.8
333.3
25.3 335.7
23.6
335.41 23.6 311.8
6:30- 7:00 337.2
24.4 336.8
24.9
334.8
25.3
333.8
26.3 335.7
24.9
335.7 25.2
310.5
7:30- 8:00! 339.7
25.1 337.8
25.9
335.8
26.6
334.8
27.3 337.7
25.9
337.2 26.2
311.0
8:30- 9:00' 341.2126.6 339.3
26.9
338.3
27.3
337.3
28.8 338.7
27.1
339.0 27.3
311.7
9:30-10:00 344.7 27.6 342.8
27.9
341.8
28.6
343.3
29.3 344.7
27.9
343.5 28.3
315.2
10:30-11:00 342.228.9 341.3
29.1
340.3
29.3
340.8
30.8 342.7
29.6
341.5 29.5
312.0
11:30-12:00 343.730.4 344.3
29.9
343.3
30.3
342.3
32.1 343.7
30.6
343.5 30.7
312.8
12:30- 1:00 345.730.6 347.3
30.1
347.3
30.8
345.8
32.6 347.9
31.4
346.8 31.1
315.7
1:30- 2:15 334.0
30.9 334.5
30.6
*
332.6
33.1 333.9
31.4
333.7 31.5
302.2
2:30- 3:00, 332.4
31.1 331.0
30. 9|
330.8
32.8 332.2
31.6
331.6 31.6
300.0
3:30- 4:00 349.2
31.4 348.3
31. 1*
346.3
31.6
347.3
33.3 349.2
31.6
348.1
31.8
316.3
4:30- 5:00 347.2
31.1 346.8
30.9
346.3
31. -3,
345.8
32.8 349.4
31.4
347.1
31.5
315.6
5:30- 6:00 346.7
28.6 346.3
28.4
345.3
29.3
345.3
30.6 348.2
29.4
346.4
29.3
317.1
6:30- 7:00 342.2
27.6 342.3
25.4
340.3
26.1
340.3
27.1 341.2
26.1
341.3
26 5
314.8
7:30- 8:00 332.7
21.6 332.3
22.1
330.3
22.6
329.3
24.1 331.2
22.6
331.2 22.6
308.6
8:30- 9:00 332.7
20.9 332.3
21.4
331.3
22.1
329.3
22.6 331.2
21.6
331.4: 21.7
309.7
9:30-10:00 331.7
20.6 332.3
20.9
328.8
21.8
328.3
22.8 331.2
22.1
330.51 21.6
308.9
10:30-11:00 334.2
21.4 336.8
21.6
332.8
22.3
331.8
23.3 336.2
22.4
334.4! 22.2
312.2
11:30-12:30 336.021.9 338.0
21.9
333.6
22.6
334.0
23.8 337.7
22.6
335.9 22. R
1
313.3
Readings corrected for error of gage and to center of pumps.
* Pump No. 2 shut down from 2:11 to 2:41 on account of motor.
63 It will be noted from Table 2 that the average results of the
24-hr. test of the South Street Station exceeded the contract require-
ments in capacity, pressure head and efficiency.
64 The horsepower delivered by the motors during the test aver-
aged for the 24 hr. about 920 or about 15 per cent above rating, with-
out excessive heating.
HIGH-PRESSURE FIRE-SERVICE PUMPS
455
TABLE 2 COMPUTATION OF PUMP EFFICIENCIES
South Street Pumping Station, September 2 and 3, 1908
Total h.p.
Net
Hour
Total kw.
from
Gal. per
h.p.
Efficiency
Beginning
r.p.m.
per hr.
motors
rain.
pressure
lb.
delivered
per cent
(input)
12:30 p.m.
3875
4841.4
18334
309.4
3311.6
68.6
1:30
757.0
3851
4811.4
18634
323.9
3523.6
73.2
2:30
755.0
3829
4784.0
19217
321.4
3605.7
75.4
3:30
3819
4771.5
19220
320.2
3592.8
75.3
4:30
755.0
3811
4761.6
19145
319.0
3565.4
75.0
6:30
3837
4794.0
18995
311.8
3457.6
72,1
6:30
765.0
3818
4770.3
18970
310.5
3438.7
72.2
7:30
755.0
3815
4767.5
18980
311.0
3446.0
72.3
8:30
757.0
3863
4826.4
19020
311.7
3461 . 1
71.6
9:30
756.0
3868
4830.2
19120
315.2
3518.3
72.8
10:30
757.0
3873
4838.9
19095
312.0
3478.1
72.1
11:30
756.5
3859
4821.4
19120
312.8
3491.6
72.4
12:30 a.m.
757.0
3870
4835.2
19175
315.7
3534.1
73.0
1:30
756.5
3672
4587.8
18790
302.0
3315.0
72.5
2:30
757.0
3667
4581.6
18776
300.0
3288.4
71.5
3:30
757.0
3890
4860.2
19190
316.3
3543.5
72.8
4:30
757.5
3891
4861.4
19160
315.6
3530.2
72.6
5:30
754.7
3861
4823.9
19110
317.1
3337.7
73.2
6:30
757.0
3865
4828.9
19005
314.8
3492.7
72.4
7:30
754.7
3706
4630.4
18710
308.6
3370.8
73.0
8:30
745.6
3659
4571.5
18100
309.7
3272.5
71.5
9:30
745.6
3651
4536.5
17890
308.9
3226.2
71.. '^
10:30
745.0
3619
4521.6
17795
312.2
3243.4
71.8
11:30
747.0
3618
4519.1
17806
313.3
3256.8
71.8
Average
756.1
72.5
Average eflBciency, 1st period of 8 hr. = 73 . 0 per cent.
Average efficiency 2nd period of 8 hr. = 72 . 3 per cent.
Average efficiency 3rd period of 8 hr. — 72 . 5 per cent.
No. of cycles per 8ec. 12:30 p.m. to 6:30 a.m. » 25.6
No. of cycles per sec. 6:30 a.m. to 12:30 p.m. =» 25.0
TABLE 3 TEST OF INDIVIDUAL PUMPS
South Street Station, September 3, 1908
I'ime
No. of
pump
Gal. per
min.
Pressure
delivery
Lb. per
Sq. In.
h.p.
output
Efficiency
Inj.
Net
of pump
12:58- 1:14
1
3372
344.4
29.3
315.1
620
74.6
1:22- 1:37
2
3809
336.0
27.9
308.1
683
70.1
1:43- 1:58
3
3495
334.0
28.7
305.3
623
73.2
2:03- 2:18
4
3705
334.5
27.8
306.7
662
76.0
2:24- 2:38
6
3740.7
1
344.5
28.8
315.7
689
77.0
Immediately following the 24-hr. test for capacity.
456 HIGH-PRESSURE FIRE-SERVICE PUMPS
65 Immediately after the close of the endurance test of 24 hours,
a short test was run on each motor separately, which was continued
long enough after uniform results were shown to obtain 12 to 15
readings. This test was run for the purpose of ascertaining whether
there were deficiencies in any of the individual motors, and to meet the
requirements specified in the printed specifications for the work, viz:
that each pump should be free from defects, should have a capacity
of 3,000 gal. per min. and an efficiency not less than 70 per cent.
The results of these tests. Table 3, show that the individual pumps
had an efficiency from 4 per cent to 6 per cent in excess of
the average when operated together, and that the capacity for the
specified discharge pressure was considerably in excess of the require-
ment of the specification. It is, I believe, generally the case that
individual centrifugal pumps delivering water into a main singly
show a greater efficiency by from 4 per cent to 6 per cent than the
same pumps delivering together into a single main, due probably to
less loss in eddy currents and friction head, etc.
GANSEVOORT STREET STATION TEST
66 The endurance test of the Gansevoort Street Station with all the
pumps in operation was begun at 9:45 a.m., September 5, after the
pumps had been operated for about 20 min. giving uniform results.
The test was continued for 24 hr. The method of testing and the
various observers were the same as for the tests at the South Street
Station and the results are given in Tables 4 to 6.
67 For the Gansevoort Street Station the efficiency average for 24
hr. was 72.9 per cent, with a variation (excepting the first hour) of less
than one-half of 1 per cent. It fell below 70 per cent during the first
hour, which was due to the opening of an automatic relief valve on
pump No. 2, which discharged some of the water into the suction
before it had been metered. For that reason the efiiciency during the
first hour has not been considered in determining the performance
of the pumps.
68 The least capacity during the eight consecutive hours when
all the water pumped passed through the meters occurs from 10.45
a.m. to 6.45 p.m. The average capacity during this time is 17,419
gal. The average net pressure in pounds is 324.3 which is nearly
16 lb. in excess of the contract requirements. The average efficiency
for the period above is 72.90 per cent
HIGH-PRESSURE FIRE-SERVICE PUMPS
457
TABLE 4 HOURLY AVERAGE OF READINGS OF DISCHARGE AND INJECTION
GAGES ON PUMPS
Gansetoort Stkbbt Pumping Station, September 5 and 6, 1908
Pump _,
XT a Pump ]
No. 6
"^0.4
Pump No. 2
Pump No. 1 Pump No. 3
Average
Net
Hour
Pres-
sure
Lb.
Disc.
Inj.
Disc.
Inj.
Disc.
Inj.
Disc.
Inj.
Disc.
Inj.
Disc.
Inj.
9:45-10:30
342.4
24.7
342.6
25.6
344.8
24.9
346.9
25.2
345.1
25.3
338.4
25.1
319.3
10:45-11:30
347.4
24.7 347.1
25.7
345.4
25.2
348.4
25.4 346.4
25.2
346.9
25.2
321.7
11:45-12:15
348.924.9 348.9
26.2
347.9
25.7
349.4
25.9 347.9
25.9
348.6
25.7
322.9
12:45- 1:15
351.425.7 350.9
27.7
350.1
26.7
352.4
26. 9| 350.4
26.9
351.0
26.8
324.2
1:45- 2:15
352.4 25.7 352.4
27.4
350.9
26.7
353.4
26.7 351.4
26.9
352.1
26.7
325.4
2:45- 3:15
352.925.7 352.9
27.7
351.4
26.9
354.9
27.2 351.4
27.2
352.7
26.9
32,5.8
3:45- 4:15
353.9i26.4 353.9
28.2
353.4
27.7
355.4
27.4) 353.4
27.9
354.0
27.5
326.5
4:45- 5:15
354.426.9 354.4
28.7
353.4
27.9
355.9
28.2 353.4
28.7
354.3
28.1
326.2
5:45- 6:15
349.9 27.4 349.9
29.7
350.4
28.7
351.9
28.9 350.9
29.2
350.6
28.8
321.8
6:45- 7:15
349.928.4 348.9
30.2
349.4
29.2
351.4
29.2' 349.4
29.4
349.8
29.3
320.5
7:45- 8:15
351.928.9 350.9
30.7
350.4
29.7
352.4
29.2 349.9
29.9
351.1
29.7
321.4
8:45- 9:15
351.929.7 351.9
30.9
350.6
29.7
353.4
29.9 350.9
30.7
351.7
30.2
321.6
9:45-10:15
352.4 29.7 354.4
30.9
351.4
30.2
353.4
30.4* 352.9
30.9
352.9
30.4
322.5
10:45-11:15
354.430.4 353.9
31.2
353.4
30.2
354.9
30.91 353.4
31.2
354.0
30.8
323.2
11:45-12:15
354.431.2
353.4
31.7
352.9
31.2
353.9
31.7 352.9
31.2
353.5
31.4
322.1
12:45- 1:15
353.931.2
352.4
31.7
352.9
31.2
354.9
31.7 353.4
31.2
353.5
31.4
322.1
1:45- 2:15
352.431.2
351.9
32.2
352.9
31.2
354.4
31.9
353.4
31.2
353.0
31.5
321.5
2:45- 3:15
350.431.7 349.4
32.2
349.4
31.4
351.9
32.2
350.4
31.7
350.3
31.8
318.5
3:45- 4:15
350.9'32.2 350.4
32.7
348.9
31.2
351.4
32.7 348.4
31.4
350.0
32.0
318.0
4:45- 5:15
350.9'31.9 350.4
32.4
350.4
31.2
352.9
32.2 349.9
31.2
350.9
31.8
319.1
5:45- 6:15
352.431.2 351.9
32.2
352.4
30.7
353.4
31.7^351.9
31.4
352.4
31.4
321.0
6:45- 7:15
350.929.9 349.9
31.2
349.4
29.9
351.9
30.7 '349.4
30.9
350.3
30.5
319.8
7:45- 8:15
349.428.7 349.4
29.7
347.4
29.4
349.9
29.2 ''347.9
29.4
348.8
29.3
319.5
8:45- 9:45
348.127.4 347.7
1
28.9
346.4
28.0
348.4
28.0 346.4
28.6
347.4
28.2
319.2
Readings corrected for error of gage and to center of pumps.
69 The average capacity for the entire test is 17,867 gal. which
was obtained with an average speed of 753.6 r.p.m.
70 Immediately after the completion of the endurance test of 24
hours duration, each pump was tested when operating alone for a
period sufficiently long to obtain 12 to 15 readings after they had
become practically uniform. These tests gave in every case an
efficiency several per cent greater than that obtained when the pumps
were all discharging into the same main.
CONCLUSIONS
71 It appears from the endurance test in each station that the
capacity, efficiency and pressure exceeded the contract requirements
by a large margin, and that during the endurance test no mechanical
458
HIGH-PRESSURE FIRE-SERVICE PUMPS
TABLE 5 COMPUTATION OF PUMP EFFICIENCIES
Gansetoobt Stbeet Pumpinq Station, SEPTSMBsa 5 and 6, 1908
1
Total h.p.
Net
pressure
lb.
1 '
Hour
Beginning
r.p.m.
Total kw.
1 per hr.
from
motors
Gal. per
min.
h.p.
delivered
Efficiency
per cent
(input)
1
9:45 a.m.
740
3671
4586.5
17107
319.3
3188.9
69.5
10:45
749
3589
4484.2
17310
321.7
3251.0
72.5
11:45
750
3591
4486.7
17290
322.9
3259.3
72.9
12:45
752
3591
4486.7
17280
324.2
3270.5
72.9
1:45 p.m.
752
3604
4502.9
17285
325.4
3283.6
72.9
2:45
753
3604
4502.9
17315
325.8
3293.3
73.3
3:45
753
3630
4535.3
17345
326.5
3306.1
72.9
4:45
753
3685
4604.0
17670
326.2
3365.0
73.1
5:45
756
3696
4617.9
17855
321.8
3354.4
72.6
6:45
756
3661
4574.0
17825
320.5
3335.2
73.4
7:45
754
3676
4592.9
17775
321.4
3335.2
73.3
8:45
753
3685
4604.0
17755
321.5
3332.5
72.8
9:45
755
3657
4569.0
17720
322.5
3336.2
72.9
10:45
755
3693
4614.2
17755
323.2
3350.1
72.7
11:45
756
3704
4627.9
17830
322.1
3352.8
72.6
12:45 a.m.
756
3753
4689.0
18195
322.1
3421.4
73.0
1:45
756
3760
4697.8
18310
321.5
3436.6
73.3
2:45
756
3735
4665.5
18315
318.5
3405.5
73.0
3:45
755
3725
4654.0
18290
318.0
3395.5
73.0
4:45
756
374:5
4677.6
18315
319.1
3411.9
73.0
5:45
756
3784
4727.9
18330
321.0
3435.0
72.7
6:45
755
3747
4681.6
18315
319.8
3419.4
73.0
7:45
755
3723
4656.5
18255
319.5
3405.0
73.1
8:45
755
3722
4655.3
18189
319.2
, 3389.5
72.8
Average
1
753.6
i
1
1 1
72.9
Average efficiency, 1st period of 8 hr. = 72 . 9 per cent.
Average efficiency, 2d period of 8 hr. = 73 . 0 per cent.
Average efficiency, 3d period of 8 hr. = 72 . 9 per cent.
No. of cycles per sec. 9:45 a.m. to 2:45 p.m. = 25.00
No. of cycles per sec. 2.45 p.m. to 4.45 p.m. = 26.25
No. of cycles per sec. 4:45 p.m. to 6:45 p.m. = 25.50
No. of cycles per sec. 6:45 p.m. to 7:45 p.m. = 25.00
No. of cycles per sec. 7:45 p.m. to 9:45 p.m. = 25.25
No. of cycles per sec. 9:45 p.m. to 9:46 a.m. = 25.50
or electrical defects were observed. During the test of the South
Street Station one of the pumps was stopped for half an hour to
repair the motor insulation, while during the test of the Gansevoort
Street Station no stop was made. The bearings in both stations
were in perfect condition at the end of the test and the temperature
of the motors not suflficiently high to interfere with the continuous
operation for a longer period. Apparently the endurance test could
have been continued indefinitely without injuriously overworking
or overloading the pumps and motors.
HIGH-PRESSURE FIRE-SERVICE PUMPS
459
460
HIGH-PRESSURE FIRE-SERVICE PUMPS
72 The specifications call for pumping sea water, which most
authorities consider to be approximately 2,5 per cent heavier than
fresh water. The effect of substituting sea water for fresh water
would have been to reduce the capacity of the pump by about 2^
per cent for the same horse-power delivered by the motor, without
sensibly affecting the efficiency. Because of the large capacity
150 300 250 300 350 400
Net Pressure on Pumi) Lb. per Sq. In.
Fig. 8 Characteristic Curves of the Pump for Varying
Discharge-Pressures
RESULTS OF TESTS
shown by the pump, this does not materially affect the results in
relation to the contract requirements.
73 The data and results of the tests at the two stations are given
concisely in the tables. The efficiency is given as computed
for each hour, and shows a slight variation which probably can be
accounted for by changes in the amount of water leaking past the
balancing piston. The individual pump tests at the South Street
HIGH-PRESSURE FIRE-SERVICE PUMPS
461
Station show a variation in efficiency from 70 per cent to 77 per cent,
and at the Gansevoort Street Station from 70 per cent to 79 per cent.
This variation may have been due to the structure of the pumps but in
my opinion is more probably due to variable leakage past the bal-
ancing piston or through the relief valves.
74 Pump No. 6 at the Gansevoort Street Station was tested with
varying openings of the valve in the discharge pipe. The results are
shown in the latter half of Table 6.
TABLE 6 TEST OF INDIVIDUAL PUMPS
Gansevoort Stkeet Station, September 6, 1908
Time
No. of
pump
Elect.
h.p.
input
Gal. per
min. Hg.
Col.
Pressure
delivery
Lb. Pbb Sq. In.
h.p.
output
EfiSciency
Inj.
Net
of pump
10:05-10:31
1 1
916
3800
356.8
35.4
321.4
711
77.6
10:36-10:51
2
877
3800
350.8
35.1
315.7
700
70.8
10:54-11:12
3
920.5
3820
350.4
34.1
316.3
703
78.0
11:17-11:30
4
892
3751.4
352.5
35.6
316.9
695
77.7
11:37-11:53
6
899
3880
350.9
35.2
315.7
714
79.4
11:55-12:03
6
880.3
3457
376.1
36.0
340.1
686
77.9
12:03-12:07
6
929
4500
304.4
34.6
269.8
708
76.1
12:09-12:13
6
946
5070
255.6
33.6
222.0
654
69.4
12:24-12:28
6
952
5500
207.4 ,
33.2
174.2
559
58.7
12:32-12:36
6
927
5588
155.2
33.2
122.0
397
42.8
Immediately following the 24-hr. test for capacity.
PRACTICAL RESULTS FROM THE NEW SYSTEM
75 The high-pressure fire system in New York, which was put
officially into service on July 6, 1908, has been successfully operated
at many fires, but it had a crucial test on January 7, 8 and 9, 1909,
when it was brought into service for five simultaneous fires, three of
them of more than the usual extent and activity, and one particu-
larly so. Information upon the results attained with the system and
the amount of water consumed was given by Chief Engineer I. M.
de Varona and published in the Engineering News of February 11,
1909.
7G The fires occurred at Hudson and Franklin Streets, Hester'
Street and the Bower}'-, Houston Street and Broadway, Sixth Ave-
nue and 17th Street, and Houston Street and the Bowery. The
situation became so dangerous that every engine south of 37th Street,
or 40 engines, were summoned, as well as a force consisting of 12
battalion chiefs and more than 600 men, but there was no need to
use a single one of the engines.
462
DISCUSSION
77 As the violence of the fires increased, additional pumps were
brought into service, so that at one time four pumps and motors were
in commission at the South Street Station and three pumps at the
TABLE 7 SPECIFIED CHEMICAL ANALYSIS FOR PUMP MATERIALS
Nickel steel
Parts of 1 per cent
Phosphorus not to exceed 0.04
Sulphur not to exceed 0.04
Tensile strength at rupture, pounds 100,000
Tensile strength at elastic limit,
pounds I 65,000
Per cent elongation in 8 in 2
Per cent elongation in 2 in 22
Contraction of area per cent 32
Carbon not less than 20 parts of 1%
Nickel percentage I 21 to 24
Medium
steel
0.10
650,000
32,500
22
Steel
forging
0.04
0.04
75,000
38,000
22
32
Steel
casting
0.05
0.05
65,000
32,000
18
24
Gansevoort Street Station, delivering 35,500 gal. per min. against
an average pressure of 225 lb. at the pumps and 205 lb. at the hydrants.
During the operation of the pumps 14,095,000 gal. were pumped as
recorded by the meters, and the current used was 81,450 kw-hr.,
the cost of which was $1222,
DISCUSSION AT NEW YORK
Prof. George F. Sever.* The electrical features of this installa-
tion are of much interest but the reasons for selecting that system
which is now in operation should be given. In the discussion of this
problem both alternating and direct-current power were considered for
the operation of the motor-driven pumps, and alternating-current
power was decided upon. The reasons for such selection I have
noted herewith:
a Absolute simplicity, which is the key-note of the electrical
end of this power installation.
' Professor of Electrical Engineering, Columbia University.
Note. — The high-pressure system was designed by I. M. de Varona, Chief
Engineer of the Department of Water Supply, Gas and Electricity of New York.
It was also constructed under his supervision. The construction of the electrical
machinery was supervised by Prof. Geo. F. Sever as Consulting Engineer. The
details of construction were in charge of Thomas J. Gannon, John P. Reynolds
and Henry B. Machen, assistant engineers of the department. The machinery
of each station was designed a,nd erected by the Allis-Chalmers Co. of Milwaukee.
HIGH-PRESSURE FIRE-SERVICE PUMPS 463
b Commutating apparatus and brushes are entirely absent-
c Induction motors provide very quick starting when it is
necessary to operate the station on a fire signal.
d There is less expense for copper in the distribution system
to insure continuity of service,
e The induction motor is a less expensive apparatus than the
direct-current motor.
/ With the induction motor there are absolutely, no exposed
live circuits in the station, as there might be with a
direct-current apparatus. The final decision was for
3-phase service at 6600 volts and 25 cycles. It was
decided that it would not be desirable to establish a
power house to be operated by the city because it would
be a municipal plant.
2 In order to insure continuity of service there is brought to each
pumping station an independent feeder from each of the two Water-
side stations of the New York Edison Company. There is also brought
to each pumping station an independent feeder from the nearest sub-
station of the New York Edison Company, as follows: to the Ganse-
voort Street station two feeders from the Horatio Street sub-station,
and to the South Street station two feeders from the Duane Street
station of the company. Hence there are really four independent
sources of power supply for each pumping station, assuring practi-
cally no possibility of shutdown.
3 The contract for electric power for the Manhattan stations was
let to the New York Edison Company. This contract provides for
two payments, the first for a reservation of 3250 kw. capacity, of gen-
erating, distributing and controlling apparatus, available at either
pumping station at an instant's notice, or practically without any
notice at all. Thus four pumps can be thrown on with absolutely no
notice to the New York Edison Company that they are to be used.
For that reservation, and care and maintenance of the whole distribu-
ting system, the city pays about $63,000 per year, and the city also
pa5''s one and one-half cents per kw-hr. for all high-tension power
used in each station.
4 Another stipulation in the contract may be of interest to
engineers as it provides for the protection of the city. This stipu-
lation is as follows: "If the contractor, under the terms of this
contract, shall fail to maintain and deliver a continuous and uninter-
rupted supply of electric power when required, the contractors shall and
will pay to the city the sum of five hundred dollars per minute for
464 DISCUSSIOM
each minute's interruption or delay of electric power supply after
the power has been interrupted or delayed for three consecutive
minutes. " So, if they cannot deliver power after an interruption of
three minutes, immediately a charge of $500 per min. is imposed and
is deducted from the bills which the New York Edison Company
renders.
5 The operation of both these stations is extremely simple. The
handle of the oil switch is turned, throwing the 6600 volts directly on
the stator of the motor. By turning a hand wheel, the motor is
brought up to speed in less than 33 sec, and in starting the current is
not supposed to exceed 150 per cent of the full-load current, which
is 64 amperes. As far as I have observed the operation of the stations,
there has been absolutely no trouble from the electrical end, no trouble
with the feeder system, and none with the motors, and I think the
City of New York has two plants which will give it for many years
to come absolutely no trouble whatsoever.
Wm. M. White. The paper deals with questions in which I am
directly interested. The methods employed in making the tests
were probably the best that coula have been selected. There is
probably no more accurate method of determining the quantity of
water delivered by a pump than by the venturi meter, especially
when in the hands of an expert who is familiar with its workings.
The venturi meter, as Professor Carpenter says, has been used for
a number of years; it has been tested in various ways and proved to
give accurate results. The power deUvered to the pumps can be
most carefully measured by electrical instruments.
2 The writer accepts without question the various efficiencies
obtained and^presented by^the author, who states, calling attention to
the variation in efficiencies obtained, that the individual observations
do not agree asj^closely as he would like. I do not think Professor
Carpenter should offer any apology as the results seem to agree very
closely, and certainly are as accurate as are generally obtained on work
of this kind. The efficiencies obtained on^these pumps, though not
the highest that have been obtained, are as high as is usual for
similar conditions of head, capacity and speed. The designers of
the pumps deserve credit for the performance shown by the pumps.
3 I am at a loss to find a reason for the variation in efficiencies of
the pumps, as mentioned in Par. 65,jwhere it is stated that individual
pumps delivering water into a main singly show greater efficiency
HIGH-PRESSURE FIRE-SERVICE PUMPS 465
than the same pumps delivering together into a single main. I
assmne, of course, that the variation in efficiency refers to the pumps
when they are deUvering exactly the same quantity against the
same head at the same speed, whether working singly or in parallel.
In the normal operation of pumps, it would be a fact that when
one pump was operating from a suction main to a discharge main,
the efficiency of that pump would be different from what it would
be when working with another 3ump from the same suction main
and discharging into the same discharge main, because the two
pumps would usually be working against a higher head than when a
pump was working singly. The increased head on the pumps would
mean a decrease of capacity, and the increase of power demanded by
two motors instead of one would mean a shght increase in line loss,
which would again sUghtly decrease the speed and slightly change
the conditions of operation for two pumps over that which would
exist when one pump only was in operation. Of course, under
these conditions, the two pumps would show different efficiencies,
because the efficiency curve of a pump varies as its capacity and
head.
4 I do not believe, however, that this is the condition to which
Professor Carpenter refers. I assume that he has corrected for
this difference, and has obtained from two pumps working in parallel
the same capacities, heads and speeds as though one pump were in
operation, and that under this latter condition he finds the differ-
ence in efficiency in the two pumps. If this be a fact, it is the most
important point brought out from a designer's point of view.
5 I am at this time attempting to duplicate the conditions, to see
whether the efficiencies are different under the same conditions of
capacity, head and speed, as mentioned by Professor Carpenter.
George L. Fowler. A number of years ago I was associated
with Joseph Edwards, who at that time had the contract for exca-
vating the ship channel in New York Harbor, probably one of the
first, if not the first, very large hydraulic engineering projects suc-
cessfully accomplished by the contractor and to the satisfaction of
the Government.
2 The ship channel leading from the Narrows down to Sandy
Hook and out to sea, is about 15 miles long, and runs almost due
south first, turning to nearly due east before reaching Sandy Hook,
and passing through Gedney Channel to the sea. Cutting across it is
466
DISCUSSION
the Swash Channel, not used by any deep-draft boats. When the
work was undertaken New York Harbor was shoal at two points on
the Gedney Channel and the ship channel, where the water depth
was a little less than 24 ft. The Government had a survey made and
an estimate of costs based on material actually removed by the ordi-
nary methods of dredging. Through the open space from Sandy
Hook to Coney Island the whole lower bay is subject to all the winds
coming in from the Atlantic on the east and across Raritan Bay, so
that the water is nearly always rough. Two contractors had at-
tempted the work by ordinary bucket dredging and both had failed.
Fig. 1 Hydrattlic Dredger for Deepening Ship Channels
3 In the ship channel the material was sand and sedimentary
clay, lying over hard sand ; in the Gedney Channel it was gravel, shell
and sand, for two feet overlying hard shingle. Hydraulic dredging
was specially suited 'for this kind'pf work, and many kinds of material
were removed Jrom^the channel besides the ordinary silt.
4 Three sea-going vessels were built for this work by the Joseph
Edwards Company: the ReHance, the Advance, and the Mt. Waldo.
Fig. 1 shows the general arrangement of the ships. At A is the long
drag aft, where the pipe goes into the vessel and where the pumps are
located, each driven by a 192-h.p. engine at 178 r.p.m. The suction
and delivery pipes were 15 in. in diameter, with a shell of 40 in. The
HIGH-PRESSURE FIRE-SERVICE PUMPS
467
pumps delivered 10,000 gal. per min. at a velocity of 1 100 ft. The
efficiency was thus between 65 and 70 per cent, although in later
tests made by the Government, when nothing but water passed
through the pipes, the efficiency rose to as high as 80 per cent.
5 The shoe used is a hook that drags along the bottom, chains
being fastened to the vessel for this purpose. The vessel never
stopped from morning to night, simply running out to sea, dumping,
and coming'^back again to work.
Fig. 2 Detail of End of Suction Line
6 At the point L, Fig. 2, was the heavy shoe that served to dig
into the mud and gravel. At 0 was a butterfly valve, kept open all
the time to admit water above the drag to mix with the material
raised. At the bottom K was another valve which could be opened
in an emergency, in case not enough water was admitted at 0.
468
DISCUSSION
7 The pump itself was of a plain centrifugal type, 40 in. in diam-
eter, with vanes cut away at the center, as shown in Fig. 3. Because
of this arrangement, the material would come in at C and out of the
Fig. 3 Sectional View of Centrifugal Pump for Dredging
vanes at the discharge, without damaging the pump when heavy
substances were drawn in. The three vanes were made with wings
HIGH-PRESSUKE FIRE-SEKVICE PUMPS
469
bolted on, and accessible from both sides. The thrust was taken up
by the bearing at T (Fig. 4) , the nuts marked m being screwed into a head
carried by the bars 0, bringing the thrust plates at the point i. The
reason for threading the nut m was to adjust it to the vanes in proper
relative position to the sides of the pump. That is a simple construc-
tion maintained ever since, with the exception that ball bearings are
now used.
8 Although the pumps were originally intended to take water and
other loose material, such as sand and gravel, they proved capable of
lifting practically anything that came in their way. The three fol-
lowing specimens are interesting as showing the pumps' lifting power:
Fig. 4 Detail of Thrust Beaking of Pump
o A piece of shaft weighing 70 lb. raised and passed by a 15-in.
dredging pump; improvement of New York Harbor,
Steamer Reliance.
6 A piece of tree root raised and passed by a 12-in. pump from
14 ft. of water at Miami, Fla. ; Florida East Coast Railway
Company improvements.
c A piece of pig iron measuring 11^ in. by 4| in. by 3^ in. and
weighing 35 lb. raised and passed by an 8-in. special cata-
ract wrecking-pump from 15 ft. of water from the wreck
of a canal boat sunk at Puas Dock, Yonkers, N. Y.; by
Baxter Wrecking Company, New York.
9 For hydraulic dredging, the Government pays by the scow load
and gets what is excavated. In ordinary hydraulic dredging, like
that in the ship channel, about 15 per cent of the pump discharge was
solid matter. About 40 per cent in excess of the amount deposited
470 DISCUSSION
in the bins went overboard with the overflow, and was carried out to
the flats at the sides by the cross currents, which also carried the loose
material stirred up b}'- the drag. The result was that the Government
obtained an excavation about 70 per cent in excess of what would
have been obtained had all of the material removed from the bottom
been caught in the bins. This, of course, greatly reduced the actual
cost of the excavation. For example: the last contract made on the
ship channel was at the rate of 16|- cents per yard, while with the
allowance indicated, above the actual cost per yard — channel meas-
urement— it was about 11 cents.
10 As for the time of loading, some records indicate that this
ship, 157 ft. long and with a capacity of 650 cu. yd., was loaded in
48 min. ; there are also records of its being loaded at the rate of 16 cu.
yd. per min., of solid matter placed in the bins; and records of its
taking out to sea nearly 4000 cu. 3'd. per day. The vessel was
worked in all kinds of weather, even when tackles had to be used to
board her; and yet the ship was taking her load steadily. Except
in the case of an actual breakdown the work could be carried on for
16 hr. per day.
John H. Norris. In a pumping plant of the character described,
this type of equipment seems in the present state of the art the most
suitable that could have been selected. I would like, in this connec-
tion, to call attention to another type of installation for service of
this kind, though not on so large a scale, which appeals to me as
being more desirable than the electric-driven centrifugal pumping
plant taking its power from a public utilities company.
2 At Coney Island was installed the first plant operated by the
City of New York for fire protection by means of water delivered
into mains under high pressure, with the idea of taking care of a
restricted area where there was great danger from fire.
3 This plant consists of three 150-h.p. three-cylinder, vertical
gas engines direct-connected to triplex pumps, each unit capable of
pumping 1500 gal. per min. against a pressure of 150 lb. These
engines take their fuel from the mains of the local gas company and
can be arranged if necessary to run on gasolene. They are installed
in a building on city property and are arranged to take their water
supply from the city mains or from Coney Island Creek, within 50 ft.
of the pumping station. The engines are started with compressed air,
and the three units can be started up in less than three minutes.
On every occasion they have been found ready for service whenever
HIGH-PRESSURE FIRE-SERVICE PUMPS 471
the demand was made upon them. The cost of this pumping station
was as follows:
Building $10,000
Equipment 37,000
$47,000
The annual operating expenses are:
Labor $13,140.00
Supplies and Repairs 897 . 27
Fuel 150.00
$14,187.27
4 By comparing the foregoing figures it will be evident that for
service smaller than is required in the City of New York, the gas-
engine-operated triplex pump gives an economical equipment that
can be allowed to stand idle for any length of time and yet be ready
for instant service.
5 New York City pays the New York Edison Company an annual
charge of $90,000 for the privilege of calling for sufficient current to
operate the equipment at any time. This item capitalized at 5 per
cent would pay for a good-sized gas-engine plant.
6 The following data were taken from the capacity tests of the
Coney Island units:
Duration of test 14 hr.
Average piston speed of pump 90.3 ft. per. min.
Total head pumped against 156 . 5 lb.
Average pump horsepower for each unit 142.2 h.p.
.Average gas consumed per hour for the 3 units 8914.0 ft.
Average capacity 4512 . 0 gal. per min.
Slip of pump 3 . 45 per cent
Average efficiency of pumps 82 . 00 per cent
J. R. BiBBiNs. Although Professor Carpenter's paper deals pri
marily with multistage pumps, I wish to direct attention to the ques-
tion of motive power, upon which the success or failure of the system
practically depends. We have seen excellent examples of two systems
diametrically opposed in regard to power supply — the electrical and
the gas-driven system. Under certain conditions, both are extremely
serviceable. The first high-pressure installation on a large scale, in
this country, was the gas-driven system at Philadelphia. Although
I have not had an opportunity to follow the results of that station for
the past two or three years, the results obtained and pul)lished for
472 DISCUSSION
the first year or so showed that such a system of gas-driven pumps
merits every consideration,
2 First as to the security of power supply: In Philadelphia the
Delaware Avenue station receives its gas supply directly from a
24-in. trunk main running between two very large gas holders, located
in different parts of the city. Roughly, the pipe line measures four
miles in length, its capacity constituting a considerable reserve in
itself, if both the holders were unavailable. There is no intermedi-
ary apparatus whatever between the pipe line and the engine ; that is,
the plant may draw directly on these two large holders of several
million cubic feet capacity. This constitutes a very safe and reliable
source of motive power which can hardly be paralleled except, per-
haps, by the situation in the New York electric service, where there
are so many stations to draw from.
3 In this connection, I would like to ask whether it is at present
possible to utilize the storage battery capacity in the various sub-
stations for reserve service at the high-pressure pumping station.
It is stated that the storage batteries are available for reserve in
emergencies, such as discontinuance of the main high-tension current
supply. I am under the impression that an inverted rotary requires
a direct-driven exciter to maintain a definite frequency and prevent
racing. Without special controlling apparatus, this inversion would
be impossible in the ordinary sub-station equipment. Possibly special
provision has been made in the New York systems, in which case,
the security of power supply is certainly beyond criticism. In other
words, would it be possible to invert the synchronous converters on
short notice?
4 Second, quick starting: It seems to be a fact that a large part
of the minimum time required for the starting of a fire-service station
is consumed in the operation of the motor-driven by-pass valves. In
Philadelphia these valves are operated from an independent supply,
as in New York, and at least fifteen seconds are required to close them;
whereas the engines are brought up to speed within half a minute
from the time the signal is given, the remaining time being usually
consumed in closing this motor-driven valve.
5 The various tests of the Philadelphia plant showed that each of
the units could be readily put on the fine in well under one minute.
It is an interesting fact that the original underwriters* tests specified
the time limit as twelve minutes for the starting of the first three units,
whereas the whole station can be started in that time, and has been
started in seven minutes.
HIGH-PRESSURE FIRE-SERVICE PUMPS 473
6 During the 36 days of preliminary service trials of the Phila-
delphia station, out of one hundred alarms given, onl3'-four misses were
made in getting any of the eleven units started. In not a single
instance has the station, as a whole, failed to respond to the service, at
least during the period over which my observation extended. This
has been accomplished with the regular operating force of three men.
7 Third, in regard to the cost of service at Philadelphia; The
only data on a large fire available, are those of the fire in the Coates
Publishing House, which lasted about nineteen hours. The average
cost for pumping was about six cents per thousand gallons, including
gas, wages and supplies. The cost of the large East Side service,
cited .in the paper, is about nine cents for power alone, and I think
this does not include the readiness-to-serve factor. On the other
hand, it is patent that the cost of service in either the gas or the
electrical station is relatively unimportant. The main desideratum
is reliability.
8 Finally, I desire to advance an argument for the development of
a new type of pump unit, namely, a high-speed gas-driven centrifugal
pump. Some time ago, in connection with water-works service, I
found great difficulty, even with the present high-speed single-acting
gas engine, in matching engine speeds with those required in centrifugal
pump work However, for the pressure necessary in water-works
practice, about 125 lb., one or two sizes of engines were found to be
directly applicable to multistage pumps, with fair proportion of parts
and good efficiencies. It seems possible to adopt a modified type
of gas engine which would permit the direct connection mentioned.
9 This modification would naturally follow along lines of short
stroke and high piston speeds with perhaps four cylinders. The
engines at Philadelphia were designed with a piston speed of but 730
ft. per min. with a 22-in. stroke. This might be increased to 1000 ft.
per min. without exceeding present-day limits, especially for units
designed for occasional service. Such a unit would find immediate
application in many industries and would combine the high economy
of the gas engine with the simplicity of the centrifugal pump. The
efficiencies shown by Professor Carpenter place the centrifugal pump
in a position of closest competition with reciprocating pumping units.
J. J. Brown. I recently made a series of tests on three 6-in., 8-stage
centrifugal pumps, each designed for 1000 gal. per min. and 560 lb.
pressure at 1200 r.p.m. One of these pumps gave an efficiency from
wire to water of 71 per cent, or a pump efficiency of 76 per cent.
474 DISCUSSION
regret that Professor Carpenter did not give the results of his tests
on the New York fire-service pumps at lower capacities. All of the
tests were made at capacities considerably in excess of that for which
the pumps were designed and they apparently show their best effi-
ciency at approximately 25 per cent over the normal rating. This
increased efficiency at excess capacity seems to be apparent in several
recent tests made on high-lift centrifugal pumps. The 8-stage
machines previously referred to give their best efficiency at 1300 gal.,
or about 30 per cent over rating.
2 Mr. White has raised a question as to the difference in efficiency
between the New York fire-service pumps working in multiple and
as separate units. I think this is occasioned by the variation in
capacity of the pumps when working together on a common suction
and discharge line. I have found it rather difficult to balance two
centrifugal pumps on a common discharge, and pitot tube tests indi-
cate in almost every case a considerable difference between the amounts
of water handled by the individual units under these conditions.
3 I have in mind one installation on fire service, where the pumps
were called upon to deliver against the maximum pressure for which
they were designed and it was only with considerable difficulty that
we were able to cut in additional units. I think that if venturi meters
or pitot tubes had been placed on the discharge of each of the five
pumps when they were working in multiple, a difference in capacity
of the several units would have been shown, which would account
for the difference in eflaiciency observed when the pumps were working
individually and not in multiple.
George A. Orrok. At the time of the award of contract for these
fire pumps, the New York Edison Company was obtaining proposals
for centrifugal feed pumps — a somewhat similar service — and eight
1000-gal. 300-lb. pressure five-stage pumps were purchased. There
was no attempt to obtain a high guarantee for efficiency, but the
builders did state that under the above conditions an efiiciency of
65 to 68 per cent would be obtained. These pumps were of the Jager
type and under test showed an efficiency of about 68 per cent.
2 Fig. 5 shows that the high-pressure fire-service pumps are of the
Kugel-Gelpke type and should be a trifle more efficient because of
smaller friction and leakage. Seventy-one per cent seemed a very
high efficiency and many doubts were expressed regarding the ful-
fillment of the guarantees. The extreme figure of 79 per cent
obtained is probably the result of careful design and extra good shop
H1GH-1'1{ESSURE FIUE-SERVICE PUMPS 475
work and I believe has not been excelled. That this figure came as
a surprise may be explained by the fact that most centrifugal pumps
are stock pumps and not specially designed for the work they have to
do. Pump manufacturers have been more concerned in getting a
line of patterns that will suit standard conditions than in developing
a line of pumps and system of patterns capable of doing the best work.
3 As a centrifugal pump is a mixed-flow or Francis reaction turbuie
reversed, similar care in design and construction would probably
give efficiencies similar to those of the best makes of reaction turbines,
which approximate 90 per cent.
Frederick Ray. The difference in efficiency of the units oper-
ated individually from that obtained when several were operated in
parallel might be due to the different rates of flow through the
ventuii meters under the two conditions. With one pump operating,
this flow would be low and the mercury column reading would be but
slightly over an inch, so that with a given error of observation the per-
centage of error would be much greater than with two or three pumps
discharging through the same meter.
2 Professor Carpenter here replying that the pipe connecting the
two meters was open all the time, Mr. Ray continued:
3 This would equalize the flow in the meters, so that the mercury
column reading when the whole station was running would be
about 6^ times the reading with one pump. It has not been my
experience that parallel operation of a number of pumps has any
tendency to decrease or otherwise change the efficiency obtained
when operated individually. The efficiency should be the same, and
in this case, as the pressures were taken at each pump, any losses in
the piping system due to parallel operation would be external to the
gages and would not show in the calculations. If the pressure had
been taken at the discharge of the whole system, losses in the piping
would affect the results.
4 Many pumps are running under similar conditions, at the
efficiencies given. I have myself obtained efficiencies of 80 per
cent and higher, but I do not rely as much on them as on some a
little lower. I am now testing a 6-in., 2-stage underwriter pump,
having a normal capacity of 500 gal. per min. against 100 lb. pres-
sure, which has developed a maximum efficiency of 73 per cent. .
5 I think the centrifugal pump is the ideal one for fire service,
not only on account of its simplicity and reliability, but also on
account of its characteristic increase in capacity as the pressure is
476 DISCUSSION
reduced. Thus, the 500-gal. underwriter pump referred to will dis-
charge 870 gal. per min. at 60 lb., or enough for four streams at this
pressure. It will give three streams at 90 lb., two streams at 110
lb. and one at 117 Ib.-^all at constant speed without any regulation
whatever.
6 The City of Toronto has recently issued specifications for cen-
trifugal pumps for their general municipal water supply, among which
are several fire pumps capable of discharging against 300 lb. pressure.
These pumps, however, are to be equipped with variable-speed induc-
tion motors, the pressure regulation being obtained by speed variation.
This is superior to throttling regulation from the standpoint of cur-
rent economy and in the case of the New York installation a con-
siderable saving could be made by this means, as most of the fires can
be handled with 200 lb. pressure or less.
H. Y. Haden. a somewhat unusual result obtained from this
type of pump is that as the total head continues to increase beyond a
certain point, the capacity falls off, with the result that the capacity
curve, as given in Fig. 8, shows a backward tendency. It will be
interesting to get the explanation of this.
2 There is unquestionably a large field in fire protection for steam-
turbine-driven centrifugal pumps, and it is to be hoped that the Fire
Underwriters will officially accept this type of fire protection unit.
I believe that a properly designed centrifugal pump, for high speeds and
of few stages, can be used to great advantage when direct-connected
to high-speed turbines.
Thomas J. Gannon/ It was decided to use electricity as power
for the pumping stations, because [the first cost of installation,
yearly cost of operation and maintenance and ^fixed charges
were estimated to be lower, taking into account the intermittent
service. The construction and operation of a steam plant were
entirely out of consideration and the choice lay between gas-engine-
driven and electric-driven pumps receiving power from outside
sources.
2 It was estimated that gas operation of plants equal in capacity
to the present electrically driven plants, would involve a fixed
charge of $50,000 a year, in addition to the cost of the gas actually
consumed. The question as to who should build and maintain
* Engineer, Dept. Water Supply, Electricity and Gas, Manhattan Borough
New York.
HIGH-PRESSURE FIRE-SERVICE PUMP8 477
the necessary large gas mains, the cost of which would approximate
a million dollars, was not definitely settled. That the cost of a
gas-engine-driven pumping plant would have been approximately
double, both for machinery, building and area of land to be pur-
chased, is borne out by the actual costs of the installations in Man-
hattan and at Coney Island.
3 The capacity of the gas-operated Coney Island plant is 4500
gal. of water per min. against a head of 150 lb. per sq. in. The com-
bined capacity of the two pumping plants in the Borough of Man-
hattan, as originally laid out, was 30,000 gal. per min. against a head
of 300 lb., with provision in each station for three additional pumping
units of a capacity of 3000 gal. each, making a total combined capacity
of 48,000 gal. per min. agauist 300 lb. pressure. On actual test,
however, the capacity of the pumps was approximately 20 per cent
greater than the designed capacity.
4 Furthermore, the flexibility of this type of pump permits of an
increased discharge at lower pressures, which gives a capacity of
approximately 5500 to 5600 gal. per min. for pressures between 150
and 200 lb., or a combined total capacity of 55,000 gal. per min.
against 200 lb. pressure. This corresponds to the pressure at which
the station is operated for most fires. In other words, the water
horsepower of the electric-driven as compared Avith the gas-engine-
driven riant is approximately in the ratio of 20 to 1.
5 The cost of the machinery in the Coney Island plant was
approximately S^37^000, and the cost of the building approximately
SI 0,000. The cost of each of the two Manhattan pumping stations
complete, exclusive of land, was practically S240,000. The first cost
of installation of the gas-engine-driven plant is therefore more than
double the first cost of installation of an equivalent electrically-driven
plant, in the city of New York.
6 The high-pressure fire-service pumping stations went into
official operation on July 6, 1908. It was at first decided to put the
stations in service only when called on by the fire department, and
up to and including November 20, 1908, the pumping stations were
called upon to go into actual service for but 17 fires. On that date,
the method of operation was amended so that the pumping stations
are put in service in response to every alarm in the high-pressure
district, and continue in operation awaiting instructions from the
fire department. Under this system, from November 20 to December
31, 1908, the pumps responded to 116 first alarms. From the best
available information, water was used in 55 instances, making a
478 DISCUSSION
total of 72 fires for which the high-pressure service had been used
up to that date.
7 To insure readiness for service at all times, daily tests are made,
of at least half an hour's duration, unless the station has been in
actual operation during the preceding 24 hours.
8 During the first quarter of 1909 the number of aiarms received
was 239, and water was taken from the station for 125 actual fires.
The total amount of water pumped was 17,840,000 gal., and 145,900
kw-hr. was consumed. It was on January 7, 8 and 9 of this quarter
that the three large simultaneous fires mentioned in Par. 75, occurred,
for which over 14,000,000 gal. of water was pumped, leaving about
3,800,000 gal. for the balance of actual fires occurring dm'ing the
quarter. For these three simultaneous fires more than 81,000 kw-
hr. was consumed while the total consumption of power for the
quarter for all fires and testing purposes was but 145,900 kw-hr.
9 As to why a pump running singly develops a higher eSiciency
than when running in conjunction with several others, it is observed
that pumps of the same type do not necessarily develop their best
efficiency at the same speed and pressure. The pump running
singly will naturally develop a pressure which corresponds to its
own design, but when working in multiple, it will have to adjust
itself to the common pressure.
10 As to reliability I have neither seen nor heard of any time
when any one of the ten pumps installed in the Borough of Man-
hattan has failed to respond instantly when called on for service
and to develop the full pressure on the system within one minute's
time. At no time in service have the pumps shut down of their
own accord.
Henry B. Machen.' Among the many difficulties encountered
during the construction of the distribution system, perhaps the
greatest was that due to the congested sub-surface of the street,
which was a source of continual extra expense to the contractor,
and of worry to the man in charge of selecting the location for the
excavation of the trench.
2 The intersection of Sixth Avenue and Fourteenth Street may
be cited as an example, since complete notes are available, due to the
station excavation for the Hudson Tunnels. Here there were nine
gas mains east and west, and nine north and south, belonging to
' [engineer, Dept. Water Supply, Electricity and Gas, Manhattan Borough,
Nnw York.
IIKiH-PRESSURE FIKE-SERVICE PUMPS 470
four different companies; two water mains in each direction; sewers
and their connections on each side of the street; five Edison duct
lines, and five duct lines with large manholes belonging to the Con-
soUdated Telegraph and Electric Subway Company or the Empire City
Subway Company; the conduits and banks of ducts of the Fourteenth
Street and the Sixth Avenue trolleys; and lastly, the columns of
the elevated railroad with their deep foundations.
3 Through this network the high-pressure main had to be so
laid that the construction of the Sixth Avenue tunnel would not
require it to be relaid. The excavation was carried on by tunneling,
with here and there an opening through which the earth could be
hoisted, using a pail let down by a rope. The pipe was lowered
into the trench some distance up the street and pulled through,
piece by piece, inspection of the running of the joint and caulking
being almost impossible, since the space admitted but one man
at a time after the pipe had been hauled in.
4 This condition existed at nearly all intersections of the main
thoroughfares, such as Broadway, Sixth Avenue, Fifth Avenue,
the Bowery, etc., and accounts for the high cost of la5dng the mains,
averaging about $11 per ft. complete.
5 The second great difficulty encountered was in obtaining the
prescribed test, which called for 450 lb. pressure per sq. in. to be
held for 10 min., during which time the leakage was measured.
6 The system contained about 40,000 castings, 30,000 being
straight pipe, tested at the foundry to 650 lb. The specials were not
tested. All these castings, as already stated, were tested in the
ground to 450 lb., the mains being under pressure in sections about one
block long, between gates.
7 During the eighteen months the system has been^^in service,
there have been but three breaks in the mains, all three in castings
which had been subjected to the foundry test of 650 lb., two breaking
at 150 lb. and the third at 300 lb. pressure.
8 To overcome the danger should a break occur [during a fire,
the proposed extensionSjto the distribution system now under contract,
amounting to about $1,500,000, are laid out on what the department
calls the [duplex system. This method of overcoming the difficulty
was first suggested by Mr. Blatt, assistant engineer of the high-
pressure bureau. It consists of laying two entirely independent
systems of mains^and hydrants in alternate streets, the hydrants
of one system being painted red and the other green. The mains are
so laid out that at nearly all intersections of streets hydrants of
l)<>th colors are available.
480 DISCUSSION
9 Should a break occur in either system, the operator at the
pumping station would at once know in which system the trouble
was located by looking at the venturi meters, and by throwing a
switch he would start the closing of two electrically driven valves,
separating one system from the other. Hydrants would then be
available and in service pending the location and isolation of the
damaged section.
10 The section now in operation was designed to give 20,000
gal. per min. on any one block with a loss due to friction from pumps
to hydrant not to exceed 40 lb. The duplex extension will give
the same results, and should either half be out of service by an acci-
dent, there will still be available at the same location 10,000 gal. per
min., with a loss from the pumps to the hydrant in the most unfavor-
able location not exceeding 50 lb.
Richard H. Rice. This paper shows that the installation de-
scribed was made after the most careful study and a very intelligent
choice of the types of apparatus to be used. The choice of the
centrifugal pump for the work described is thoroughly justified by
its simplicity and by the efficiencies obtained. The centrifugal pump
is today the popular means of producing pressure for emergency fire
purposes, as in the fire boats of New York, Chicago, Duluth and San
Francisco, and the new high-pressure service of San Francisco . In San
Francisco twelve of these pumps are now being installed, four on fire
boats and eight for an auxiliary fire installation. On the fire boats
centrifugal pumps are particularly adaptable as they can be run in
series or in parallel. In parallel they give 150 lb. pressure, and in
series the pressure is doubled. This pressure is particularly valuable
where walls have to be battered down, or streams thrown long
distances.
2 The choice of alternating current as the source of power, in view
of the unlimited supply of current existing and the duplicate means of
conducting it into the station, is also justified. In cases where
electricity is not so available as it is in New York, steam turbines
are being installed, and they offer advantages over the gas engine,
where maximum reliability is considered.
3 As an emergency installation pure and simple, I think the
installation mentioned in the paper can be still further simplified.
I believe the speeds chosen for operating the pumps are too low,
and that the pumps contain too many stages. I have had occasion
to make extensive researches in centrifugal pump design with special
HIGH-PRESSURE KIRE-SERVICE PUMPS 481
reference to operation at steam-turbine speeds, and have found that
they can be operated at high speeds with a smaller number of
stages, giving efficiencies comparable with those obtained here,
although the question of efficiency is subsidiary to reliabihty for
this service. Pumps for this service should be designed with two or
three stages at the most, and with considerably higher speed.
4 Pumps can also be designed without balancing pistons, which
are undesirable from the viewpoint of possible interruption of service.
An inspection of Fig. 5, illustrating the construction of the pumps,
will show that the balancing pistons used are quite liable to damage
if water containing sand or other impurities is used, and this damage
would very probably result in stoppage of the pump when it is
badly needed. The use of balancing pistons is unnecessary in such
emergency apparatus and should be avoided.
C. A. Hague. A question has been asked several times with
reference to the results of tests of efficiency on centrifugal pumps
operating singly and in multiple or group. Professor Carpenter
has given the very plausible explanation that the difference in effi-
ciency in favor of the pumps running singly is probably due to the
presence of eddies and disturbances in the pipes when the pumps
are operating together and the absence of such eddies and disturb-
ances when only one pump is at work. In my experience in installing
pumps and condensers singly and in groups I have found them
extremely sensitive to each other in operation, both in taking in
and discharging the water, when more than one pump is working on
a line.
2 In the Manhattan stations, it seems to me that the suction or
inlet pipes and the discharge pipes are coupled too closely for best
efficiency; and also that the inlet pipe close to the pumps is not large
enough for operation in multiple, although perhaps ample for a
single pump when the water is undisturbed by the draft and dis-
charge of several pumps. I have experimented considerably in
that line, and have found that a comparative!}^ large body of water
next to the pumps on the suction side will materially ease the machines
in their performance. The idea is to come up to the building with a
normal supply pipe, and then enlarge it very considerably just where
it enters the building, providing the inlet pipe with a good-sized air
chamber wherever possible. I have tried this several times with
excellent results.
3 Mr. Brown mentioned the difficulty of cutting in with a second
482 DISCUSSION
pump where the first pump was akeady running, a difficulty which
I think is also due to too close connections along the inlet and outlet
lines and a cramped conditior generally. Of course, a disturbance [in
the water column and in the hydraulic horsepower would unbalance
the electric power to a certain extent, perhaps not much, but the
total disturbance may very easily result ^in the loss of several points
in the efficiency.
4 Considering the fact that the city pays by the kilowatt-hour
for its electric current as per switchboard reading, it would be no
more than proper to state the efficiency of ^the machine as a whole,
and not exclusively upon the basis of motor efficiency obtained in
the shop of the makers a thousand miles or so away. In this case
when 100 h.p. in current is supplied to the switchboard, the motor
has shown an output by a competent test of 93.2 h.p. (Par. 37) , the
balance of 6.8 h.p., charged against the city in the power bills, being
lost in heat and friction. Then, all that is charged against the
pump is 93.2 h.p. The 67.57 h.p. shown by the pump for each 100
h.p. at the switchboard indicates only 67.57 per cent total efficiency,
although the 67.57 h.p. indicates 72.5 per cent efficiency of the power
delivered by the motor. I have tested several centrifugal pumping
plants of various sizes and powers, and the total efficiency generally
shows from 64.5 per cent to about 68 per cent and very seldom above
the latter figure.
5 Mr. Bibbins touched upon ;^the possibihties of utiUzing the
centrifugal pump for waterworks service, but uponj investigation
he would find a vast difference between emergency service, where
operating economy counts for little in the face of great danger from
fire, and the steady and necessarily economical service required for
the continual pumping in waterworks stations. To show how decep-
tive a portion of the truth may be, a case is cited where a pumpage
of a capacity of 10,000,000 gal. per day against 110 lb. load could
easily be accomplished with displacement steam machinery by an
expenditure of $10,000 per annum for coal. But an attempt to
drive centrifugal pumps by electricity resulted in a cost for electrical
power, at $6.50 per 1,000,000 gal., of $23,725 per annum; showing a
difference in favor of displacement steam machinery equal to
5 per cent per annum on $275,940. There is no conceivable
difference in cost of machinery, buildings, maintenance, attendance,
or anything else, that would justify such a preference for electricity
and centrifugal pumps over steam and displacement pumps. Note
the following figures:
HIGH-PRESSURE FIRE-SERVICE PUMPS
483
10,000,000 gal. daily, against 110 lb 440 pump-h.p.
120,000,000 steam duty with S lb. evaporation in the
boilers, coal at $2.50 per net ton delivered §9928 per annum
Electric power at S6.50 per 1,000,000 gal. against 110 lb.
means 3,650,000,000 gal. per annum at S6.50 §23,725 per annum
The difference in cost for the element of power is S13,797
per annum, which at 5 per cent would capitalize at $275,940
6 The steam-driven, reciprocating, displacement pumping engine
can show a mechanical efficiency from the power put in through the
throttle to the water-horsepower of the pumps, as high as 96 per
cent, never as low as 90 per cent, under the above conditions. The
centrifugal pump when steam-driven has a corresponding efficiency
of about 65 per cent, and when electrically driven of about 67 per
cent. A comparison of tests is given in Tables 1 and 2 in which it
will be seen that the steam plant saves enough to pay 8.6 per cent
on its entire cost.
TABLE 1 COST OF OWNING AND PUMPING WITH HIGHEST TYPE
AND CLASS OF STEAM PUMPING MACHINERY
One Unit, Steam-Driven, Reciprocating, Displacement Machinery,
Capacity of 25,000,000 Gal. Against 87 Lb.
Pump horsepower 870
Boiler horsepower for triple-expansion vertical pumping engine 450
Engine house and foundations and engine foundations ^
Boiler house and foundation, boiler foundations, chimney, etc
Vertical triple-expansion pumping engine \ $150,000
450 h.p. of boilers
Building for coal supply
CHARGES against PLANT PUMPING ENGINE
Interest 4 per cent
Sinking fund 5 per cent
Depreciation 2 per cent
Oil waste, etc 1 per cent
Total 12 per cent
CHARGES against PLAN'J — BOILERS
Interest 4 per cent
Sinking fund 5 per cent
Depreciation 5 per cent
Total
3 enginecs. 6 firemen. 3 oilers.
Ooal it $2.10 aer net ton
14 per cent
484 DISCUSSION
StTMMARY FOR StEAM RECIPROCATING MACHINERY
Coal per annum $11,957.40
Wages per annum 9,900.00
Capital charges on engine 13,920.00
Capital charges on boilers 1,260.00
Capital charges on buildings 1,548.00
Total charges per annum $38,585 . 40
Cost per 1,000,000 gal $4.11
Cost per horsepower 43 . 16
TABLE 2 COST OF OWNING AND PUMPING WITH HIGHEST TYPE
ELECTRO-TURBINE PUMPING MACHINERY
One Unit, Electric-Driven, Centrifugal Machinery, Capacity 25,000,000
Gal. against 87 Lb.
Pump horsepower 870
Two-stage, electric-driven centrifugal pump
Engine house and foundations and pump foundations
Transformer house and foundations \ $43,750
Transformers, lightning arresters, conductors, controllers and auxil-
.aries
charges against plant — PUMPING MACHINERY, ETC
interesi 4 per cent
Sinking fund 5 per cent
Oil, waste, etc 1 per cent
Depreciation 2 per cent
Total 12 per cent
3 Engineers. 3 Extra men
Electric current, $4.50 per 1,000,000 gal.
Summary for Electro-Turbine Machinery
Electric current per annum $41,062.50
Wages per annum 5,700 . 00
Capital charges on machinery 4,314 . 00
Capital charges on buildings 468 . 00
Total charges per annum $51,544 . 50
Cost per 1,000,000 gal $5.64
Cost per horse power 59 . 24
HIGH-PRESSUBE FIRE-SERVICE PUMPS 485
Thos. J. Gannon. In reply to Mr. Hague I will read the condi-
ditions which occurred on the evening of January 7, when both
pumping stations were put to a crucial test:
7.22 First alarm, Hudson and Franklin Sts.
7.28 Second alarm, Hudson and Franklin Sts.
7.29 Third alarm, Hudson and Franklin Sts.
7.46 Fourth alarm, Hudson and FrankUn Sts.
7.54 First alarm, Bowery and Hester Sts.
8.17 Automatic, Mercer and Houston Sts.
8.19 Second alarm, Bowery and Hester Sts.
8.29 Second alarm, Mercer and Houston Sts.
8.32 Third alarm, Bowery and Hester Sts.
8.40 Third alarm, Mercer and Houston Sts.
8.43 Fourth alarm, Mercer and Houston Sts.
8.45 Fifth alarm, Mercer and Houston Sts.
2 In due time seven pumps were put into operation, with a dis-
charge which reached at times over 35,000 gal. per min., and it was
estimated that over 52 fire streams were in service at the same time.
Each pump responded instantly and remained in service until ordered
shut down. The pressure was ordered gradually increased from 125
lb. to 230 lb., where it was maintained throughout the greater part
of the time that the fires raged. The operating force at each pump-
ing station consisted of but one engineman, one oiler, one telephone
operator and one laborer.
Prof. George F. Sever. A question was asked as to the feasi-
bihty of using the storage battery capacity to invert the rotaries
and provide alternating current, to be spread through the alternating-
current system to the sub-stations, and from those to provide alter-
nating current to the pumping stations. In our preliminary investi-
gation, if I recall the facts correctly, we were assured that this could
be done; giving us another feature of reliability in the operation
of the system. If the Waterside station should go out of business,
we could still get current from the sub-station.
A. C. Paulsmeier.' While the reasons given in the paper for
the selection of electric-driven turbine pumps do not coincide with
the conclusions as to reliability that have been reached in the West,
there can be no question about the careful study given by the engi-
neers who planned the high-pressure fire system described.
1 Chief Enginef r, Byron Jackson Iron Works, San Franciso, Cal.
486 DISCUSSION
2 The pumps show a remarkable efficiency, and one of the principal
points that should commend them to those interested is their great
flexibiUty as to capacity, a characteristic that every fire pump should
possess.
3 The eight fire pumps now being built for the City of San
Francisco are of a combined capacity of 216,000 gal. per min.,
under a working pressure of 300 lb. Each of these pumps is driven
by a 750-h.p. Curtis steam turbine, operating at a normal speed of
1800 r.p.m.
4 In addition there are now being completed four fire pumps
for the boats Dennis Sulhvan and David Scannel, of an aggregate
capacity of 9000 gal. per min. under 300 lb. working pressure, or
18,000 gal. per min. under 150 lb. working pressure, the pumps
being so arranged that they work either in series or in parallel.
The pumps have all been subjected to 24-hr. tests, and while the
data on these tests are not sufficiently complete for pubhcation,
they show that the pumps are not as flexible as to capacity, or
are not as capable of pumping an excess quantity of water, as are the
Manhattan pumps. The reason for this is that the impellers in
the San Francisco pumps are only 13 1 in. in diameter, while the
inlet to the impellers is less than 10 in. in diameter, this opening
being further restricted by the pump shaft, so that it is impossible
to obtain much excess water, no matter how much below the normal
the discharge pressure is carried.
5 In the station pumps now being built the velocities at the
entrance to the impellers have been somewhat decreased, although
it is impossible to make anything like the excess capacity shown by
the Manhattan pumps, which have impellers of such a size that
the inlets may be made anything consistent with good practice.
Prof. W. B. Gregory. It is gratifying to know that efficiencies
ranging from 70 to 80 per cent may be obtained with well-designed
five-stage turbine pumps. The high-pressure fire-service pumps in
New York represent one extreme of conditions, while at the other
extreme is the centrifugal pump used in the rice irrigation territory
of Louisiana and Texas for raising large quantities of water through
comparatively small lifts.
2 The improvement in design of pumps of the latter class in
the last ten years, and especially in the last five years, has made it
possible to specify an efficiency' uf 75 per cent, even with heads as
low as 10 ft. Pmchasers of pumping plants in this section are no
HIGH-PRESSURE FIRE-SERVICE PUMPS
487
longer satisfied with pumping outfits having efficiencies ranging
from 50 to 60 per cent.
3 As examples of the results obtained with pumps of the class
that deals with large volumes of water, the tables are quoted from
recent acceptance tests conducted by the writer, of pumping plants
used for rice irrigation.
TABLE 1 ACCEPTANCE TESTS
Tanrem-Compouxd Condensing Engines, Direct-Connected
Cane and Rice Belt Irrigating Company, Fulshear, Texas, August 12 and 14, 1908
WORTHXNGTON Pt7MP8
Size of pump (diameter discharge pipe), in
Water pumped, gal. permin
Head on pump, ft
Efficiency of engine and pump, %
Efficiency of pump(engine93 %)
First
Lift
I =
Second
Lift
45 ^ 45
47,620 / 46,430
33.90 13.95
69 . 5 73 . 6
74.7 /'9.2
Cross-Compound Condensing Corliss Engine, Direct-Connected
Sabine Canal Company, Vinton, La., May 22, 1909
Worthington Pump
Size of pump (diameter discliarge pipe), in
Water pumped, Ral. per min
Head on pump, ft
Efficiency of engine and pump, %
Efficiency of pump (engine 90 % )
45
44.010
23.2(5
^" 69.5
77.3
Tandem-Compound Condensing Corliss Engine, Direct-Connected'"
Second Lift, Neches Canal, July 16, 1909
Morris Machine Works Pump
Size of pump (diameter of discharge pipe), in
Water pumped, gal. permin
Head on pump, ft
Efficiency of engine and pump (maximum), %
Efficiency of pump (engine efficiency 93.2 %max.).
60,300
10. 1 i'
69.0
75
Charles B. Rearick. Electrically driven fire pumping-stations
for large cities are dependent upon current from an outside source,
usually a large central power plant. It would seem quite practicable
in many cases to locate new fire pumping stations adjacent to some
large power plant having considerable boiler capacity. In such
cases it would be possible to drive the centrifugal or turbine pumps
with steam turbines, and thus eliminate the necessity of large over-
488 DISCUSSION
load capacity in electric generating units for the central station, and
also the liability of derangement of the lines between the power
stations and the pumping stations. The charge for standby service
per annum should be less than for similar electric service.
2 The steam turbines have the advantage of being operative at
any speed, and in this manner will maintain in the discharge mains
any pressure desired. Furthermore, automatic regulating valves can
be used in connection with the turbine to maintain constant pressure
irrespective of demand or flow.
3 It is probable that the cost of installation would be less than
for electric-driven units. The turbine could run non-condensing, as
the question of steam consumption is of small moment for fire service.
Henry E. Longwell. The last paragraph of the paper furnishes
a striking illustration of how purely academic is the ordinary official
efficiency test, and of how little value as a basis on which to predicate
the results that may be expected when the plant is operated under
service conditions.
2 This paragraph gives general figures on the performance of the
pumps during the fire run. There were 14,095,000 gal. pumped,
with a current consumption of 81,450 kw-hr. The average net pres-
sure against which the pumps operated is not stated, but assuming
it was 300 lb. per sq. in., the pump efficiency, after allowing for
the losses in the motor, would be only 40 per cent. However, we
know that for part of the time the pressure did not exceed 225 lb.,
or, considering the pressure in the suction mains, about 200 lb.
net. If the entire quantity of water had been pumped against this
lower pressure, the [efficiency would be well under 30 per cent.
It is therefore perhaps fair to assume that the actual average effi-
ciency was not far from 35 or 36 per cent, or say, in round numbers,
only one-half that shown on the official test,, when the load and other
conditions of operation were more favorable.
W. M. Fleming. With the rapidly increasing size and height of
office buildings, the annual fire loss in the business districts of the
cities of the United States is increasing to an alarming extent. The
installation of these tremendously effective fire-fighting systems has
already proved of definite value in the reduction of city fire losses,
and consequently of insurance costs.
2 What was probably the pioneer large and independent so-
called high-pressure fire system in this country was installed at
rnCH-PRESSURE FIRE-SERVICE PUMPS
489
490 DISCUSSION
Philadelphia in 1903-1904. This plant differs in almost every
important detail from the New York system more recently installed ;
yet the general results in each case have been excellent. In Phila-
delphia the plant has so many times proved of great value in actual
service that a much larger fire-fighting system, consisting of pump-
ing units identical with those originally selected, is now being installed
to protect what is known as the Kensington mill district.
3 From the original Philadelphia station at Delaware Ave.
and Race St., a location unlikely to be seriously injured by con-
flagration, Delaware River water is supplied to independent high-
pressure fire-service mains which effectually cover more than 425
acres at the center of the business district. The pumping units
consist of vertical double-acting triplex power pumps built by the
Deane Steam Pump Company, direct-connected to Westinghouse
vertical 3-cylinder 4-cycle gas engines each of 280 h.p. The seven
large pumping units have each a nominal capacity of 1200 U. S.
gal. per min., at 300-lb. pressure, and two small units have a capacity
of 350 U. S. gal. at the same pressure.
4 The general arrangement of the Philadelphia pumping station
is similar to that of the large NtiW York installations (Fig. 1).
Two rows of pumping units occupy the main floor of the station.
The pumps are nearest the center, and the gas engines are located
in the same relative positions thereto as the motors in the New York
pump houses. A platform extending along the sides of the building,
about ten feet above the floor, serves as a working gallery for the
operation of the engine throttles. Space is provided for the installa-
tion of three additional pumping units, and all mains are propor-
tioned with the ultimate probable capacity of the plant in view.
Suitable connections are provided to the mains so that the capacity
of the pumping station may be supplemented by the use of the
city's powerful fire boats, should occasion require.
5 The internal -combustion engines are of the well-known standard
Westinghouse type and require little explanation. Speed regulation
with varying loads is accompHshed by the action of a centrifugal
governor controlling the quantity of combustible admitted to the
cylinders. Ignition is by a very neat type of make-and-break mecha-
nism contained in a cyhndrical plug. Two independent igniters are
provided in each cylinder, and three independent sources of ignition
current are available at all times. The engines are started by the
use of compressed air, which is admitted to one of the cylinders at
the proper time to secure rotation in the direction required until the
HIGH-PRESSURE FIRE-SERVICE PUMPS
491
^
regular cycle of operation is established. The pumps are started
under no-load.
6 The pumps are of the vertical, double-acting piston, triplex
power type, requiring comparatively small floor space and giving a
rate of discharge so smooth and uniform as to make imperceptible at
the hose nozzles any pulsation in pressure.
7 In Fig. 2 is a sectional view of one of the pumps, indicating quite
clearly the extreme simphcity and accessibility of the machine,
and its general construction. All valves are of the poppet type,
readily accessible through handhole openings. Valve areas and
waterways naturally are comparatively large, so that friction losses
Fig. 2 Side and Sectional. End Elevation of Triplex Pumps ton thk
Philadelphia ITigh-Pressure Fire-Pumping Station
are reduced to a minimum. The water ends are thoroughly brass-
fitted in order that the pumps may be readily started after a long
period of disuse. ;
8 There is a connection through a 12-in. check valve, from the
city mains to the high-pressure system, so that the mains and pumps
are constantly primed with a pressure of 60 lb. and are ready for
service at all times. A complete system of fire-alarm boxes and tele-
phones, with underground wires, permits direct communication
between the vicinity of any fire and the pumping station. On the
sounding of the alarm, the station force, consisting of an engineer
and his assistant, can bring the total plant of seven large units
492 DISCUSSION
into service in seven minutes, and have repeatedly done so. Work-
ing pressure is invariably available at the hydrants one minute
from the time of the alarm. Such a result would be impossible
with ordinary movable apparatus.
9 The pumping units are started up under no-load, by the
use of a motor-driven by-pass valve, through which the pump dis-
charges into an overflow, until the normal cycle of operations has been
set up in the gas engine, when the switch is closed, causing the by-
pass valve to close and the discharge to be directed into the fire mains.
10 Experience has indicated that the maximum pressure of 300
lb. is required only for the most extensive fires, and for fires in the
higher parts of tall buildings. The pressure records show that
probably 75 per cent of the water pumped is required at not more
than 150 lb. to 175 lb. pressure. The pressure desired in each case,
is dictated over the telephone by the fire chief, the required pressure
regulation being obtained by proportioning the number of units in
operation to the requirements.
11 The practical results of the use of the Philadelphia fire system
have been: material reduction in fire losses in the protected district,
large decrease in fire insurance rates, and a greater willingness on
the part of property owners in the protected section to erect pre-
tentious office buildings.
12 Though the writer is unable to present a statement as to
the annual saving to property owners by the installation, yet in
view of the low cost of operation of the plant, there can be no question
but that it presents a considerable yearly saving to the city. During
the year 1907, which is perhaps typical, water was deUvered to 16
fires, the longest one lasting 44 hr. The plant responded to 1 16 alarms
at which no service was required. The operating expenses for the
year were as follows:
Gas, 839,488 cu. ft. at $1.00 $839.49
Electric lighting 343.99
Electric power 7 . 98
65 tons pea coal at $3. 50 227.50
Supplies furnished the pumping station for the entire year 1907 1,500 . 00
Total fixed chargesfor 1907 ." $2,918. 96
Summary I
Salaries (Total for entire staff) $8,389.72
Total cost materials 2,918.96
Total operating expenses $11,308 • 68
Total daily maintenance charge, salaries and operation . S31 . 12
HIGH-PRESSURE FIRE-SERVICE PUMPS i93
13 No mechanical defects have yet developed in either engines
or pumps, and practically the only replacements have been a few
rubber valves for the pumps and ignition details for the engines.
14 While no definite comparison can be made between the small
plant in Philadelphia and the comparatively large plants in New
York, which have not yet been in operation for an appreciable length
of time, the operating expenses of the Philadelphia plant seem
likely to prove much less for a given quantity of service. This Is
largely due to the so-called "readiness-to-serve" charge made by the
company furnishing power to the New York plants. To this charge
must, of course, be added the cost of the current actually consumed.
15 Unfortunately no mechanical efficiency test has ever been
made on any of the Philadelphia pumping units. Judging from
tests of similar machinery, an efficiency of 80 to 85 per cent is to be
expected from pumps of this character operating against 150 to 200
lb. pressure. If this is the case, knowing that 75 to 80 per cent of the
water to be used will be required at pressures not to exceed 175 lb.,
it would seem that the plant efficiency in Philadelphia would prove
greater than in New York, where we understand that the water must
be delivered through reducing valves from 300 lb. to any lower
pressure required.
DISCUSSION AT ST. LOUIS
Horace S. Baker* presented some very complete notes on the
proposed high-pressure system for Chicago, an abstract of which is
given herewith. After telling of that city's need of a high-pressure
system, Mr. Baker illustrated the effect of such an installation on
insurance rates by citing the reductions brought about in other
cities, details of which are given in Table 1, herewith.
2 The costs of maintaining and operating the proposed system
for Chicago should not be more than the following figures, and prob-
ably much less:
Operating costs of three pumping stations, including interest and
depreciation $180,000
Interest on cost of distribution system, 4 per cent of $3,000,000 120,000
Depreciation of distribution system, 2 per cent of $3,000,000 60,000
Maintenance of distribution system .50,000
$410,000
' Engineer, Department of Public Works, Chicago
494
DISCUSSION
Effect on
Insurance
Rates
Reduction of
o •
a ;
o '
CO .
d :
3
73
a.
>- 2
aP4
3S
"c
a
o
3
-0
Probalily Las
Prevented
Increase
Prevented
c
.£
'. "S
. (U 3
. t£-3
• 3 'i
• « C5
:^ 2
Connection
With Bdilp-
INQ
2
.£
& •
.a :
-a •
•3 •
o :
Ql C.
> C CO
J -J d
2
—
. -a
n 3
o
o i
? s
.5 a
en o
c
_o
c
c
o
c^
CO iC
lO
o
(© t-
c^
•<i<
or
H3j; ISOQ
^
■* a>
«o
o
tc
o
lO o
00
cc
CO
CO
C '^
c
cc
<o Cv
CO
tH
>--
p^
CM ro
05
CO
CO
CO
-H O
C'
a
o c
o
Q
o
1^
(7
o
cr
r~
naxsjig JO
IT
o >*
o
o
o
CM
cd'
I8O0 avxox
M
CO 00
*
09
>o
c»
sxNVHaAu
oe
■*
a>
03
03 CO
CO 00
.iO •0|I
■^
o
o
o
o
c
CO
^
C) CM
•SNJ ■SNIVJ^
1
i
'r
CM
1
1
c-
1
1 1
jio sazig
oc
<N M
IM
00
00
ol:
00
oo
o -o
c
c
o
CD
■*
r«
o
o
o r-
a
irj
(N
c
lO
c^
o — ■
SN1VJ\[ ao
ff
t-
r-
iO
oc
■N
rp
CM t^
•xj avaNii
S
10
rli"
w"
CO
lo ao
CM c^
iA
O" lO
.-1 ■*
•saq 'aHas
8F
o
o
s
s
o o
-H 00
^
■^ o
CO 'O
-aaaj -xvjj
(N CC
n
CO
IM
■-I
CO
1
O C
Q
o
O
c
• o
•NIK
o c
c
o c
1 aa<£ -IVQ
]
t-.
d
1-1
CO
o
o
; "o
s§
c
t
j
o .
2
CS -t-
-
■(- is
S Pu
■e £
1 '
sis
2 c.
IS
3 £ 5
s.
a
J H 1^
Ecv
W M
•fa -:<
IM
^ O
N
^3
f» 1
a
•c
<
(
C
C
i I
00 t£
C5 C
00 O
00
1
. 0
CO
CO
00
CO CO O
a o> 00
o -. -
Ph
fi
A P-
. o
c
> c
) c
> c
II
1 §1
o o
o o
o
o
o c
o c
i Noixvinaoj
5
> "■
o c
o o> o
O CO
oc
aaxvmxsa
» t-
(M C
(M (N 00
< ■* (N •*
CO
*
CO
1-
1
'■
< <N
i
>)
c
I
3 2
1 C
3 1
5 a
: £
I" -
3 Z
5
c
c
c
'c
a
u
; c
1 E
£1
•c
2
c
HIGH-PRESSURE FIRE-SERVICE PUMPS
495
1
65
o
lO ♦♦
3
o
3
•a
J3
^ a
J3
3
u
0)
4)
3
J3
^
O
a W
o
-3.2
o
a
O
3
o
o
25
»
o
t3
:z
g
o
o 3 S
4) O TJ
il'i
"1-
.9
(0-3
§«
a
o
3
<i
U3
m a
C o
02 o
M 3
S 3
J5
u
a
s
to
«
i-
o
N
T»<
■*
«o
«o
o
■*
«o
•*
r^
CO
•<»<
r~
CO
Ci
*^
M
m
o
C)
00
r~
■o
o
o
CO
•o
00
r^
00
CO
■*
•o
CO
N
(N
CO
o
^o
^Q
^^"
O
o
o
o
§
CO
O
o
o
Tl<
O
o
lO
o"
o"
CO*
Q
o*
C<3
lO
o
■*
UO
CO
t^
«5
o
(N
o
<o
»^
"J
s
CD
00
"^
:
^O
^
<o
>*
O
o
o
o
CO
I-l
C4
C4
C<l
evi
CO
^
^
4.
ob
o!,
oi,
o
o
o
o
Q
o
o
o
o
<o
■*
w
o
CO
o_
03
o
00
CO
lO
in
Os"
oi
o"
lO
o" '
CO
c^
o
•^
o
g
~^8~'
~i""
§
s
CO
.-1
CO
•H
iH
CO
to
^o
o
w
o
o
o
o
t^
o
o
1
»c
■*
o_
o
00
l|
CO
o
c»
co"
OS
■>!)<■■
o"
;
CO
'-'
•-'
1
n
j
3
, « g
ft >■
^ a
g si 5
•J H M H
a 3-S
MO
all
- a
a
3 3 >
c^
01
"1
3
(l4
^
a
em
.a
: '
lO
OC
CO
t^
■*
o
S
o
c
o
OJ
t»
3
c
c
05
«
00
u
3
O
'^
• j
O
Q
c
Q
o
o
o
o
Q
c
o
o
o
c
c
o_
o
o
o
o
o
o
1
q"
^
U5
lO
o"
oo'
a
00
T-t
eo
cq
lO
o
rt
N
»-l
IH
ev
i-T
;
i
'■ "a
• OJ
d
t-
a
Z
r'
t
c
2
1 03
E S
13
%
a.
a
§
1
i
c
c
J
1'
1
>
j
c
1
I
)
j,
1
1
8
i
i
1
h
- §
a-3 2
M
a>
■c
•iS
,a
H
a
o
".
a
o
73
o
o.
!".
^
e
.O
"o
.a
3
£
5
<0
>
1
"o
o
1
s
1
o
M gL o
e
g
•
■*•
**
1
»
496 DISCUSSION
TABLE 2 COST DATA FOR VARIOUS TYPES OF APPARATUS
NAME or STBTEM
Manhattan.
/ Electric
\ Centrifugal Pumps. . . .
S139.250 S22,333 | $4,642
Coney laland / G^s Engines. . . ,
\ Triplex Pumps.
Philadelphia / ^^ Engines . . .
\ Triplex Pumps .
! Steam Turbines ,
Centrifugal Pumps. . ,
and boUer Plant
OUFuel
San Francisco.
Estimate 2 . . . .
Hartford. . .
Estimate 1 .
' Gasolene Engines . . .
Turbine Pxmips I 20,000 gal.
Rope Drive !
Steam Turbines i 300 lb.
Centrifugal Pumps 12,600 gal.
Coal Fuel
OK.
J *■ a
O IS K
2 O H
Hartford f Gas Engines. . . ,
Estimate 2 \ Triplex Pumps.
Chicago, Estimate . / Steam Turbines
Chicago, Estimate
Chicago, Estimate .
300 lb.
12.600 gal.
250 lb.
\ Centrifugal Pumps j 10,000 gal.
( Gas Engines [250 lb.
\ Triplex Pimips \ 10,000 gal.
/ Electric Motors 250 lb,
\ Centrifugal Pumps 10,000 gal.
737,848
257.620
377,905
263,005
248,112
122,882
14.186 ., 10,444 I 3,152
11.978 2 8.571 1.316
34,630 31,111 1,732
30,595 I 36.892
1.529
45.320 20.466 3,597
8,648 29.992 686
37.400 I 26.300
24,626 24.811
3,740
2,463
57,700 12,288 5,770
> Exclusive of Interest and Depreciation.
3 In the light of current practice as shown in Table 2, it seems
advisable to consider and estimate on the following types of pumping
stations:
a Steam turbines and centrifugal pumps.
6 Electric motors and centrifugal pumps.
c Gas engines and triplex pumps.
HIGH-PKESSIIRE FIRE SEUVICE PUMPS 407
TABLE 3 STEAJVI TURBINE PUMPING STATWN
Approximate Estimate of Cost
(^apacity 10,000 gal. per min.; pressure 250 lb per sq. in.
1 Excavation:
Pump pit '. .' 2300 cu. yd.
Boiler room 3865" '" _
Stack 565 "
Conveyor tunnel 70 " "
6800 cu. yd. at $1 86,800
2 Concrete:
Retaining walls for pump pit 616 cu. yd.
Boiler room foundations 453 " "
Stack foundations 430 " "
Pump house foundation 101 " "
1600 cu. yd. at $7 11.200
i Building:
Pump room, 60 ft. by 54 ft. = 3240 sq. ft.
Boiler room, 78 ft. by 84 ft. = 6552 " "
— n
9792 sq. ft.
.\ssume 10,000 sq. ft, by 30 ft. = 300,000 cu. ft. at 15 cents 45,000
4 Foundations for pumps and turbines, 150 cu. yd. at SIO 1,500
.') Four 2500-gal. centrifugal pumps at $5,000 20,000
G Four 600-h.p. steam turbines at $12,000 48,000
7 Boilers, 2400 h.p. at $15 36,000
Chain grates, hoppers, conveyors, etc 15,000
Stack 8,000
Suction piping from city main and tunnel 6,500
Discharge piping g_000
Steam piping 7,500
Condenser 6,200
Boiler auxiliaries, heater, purifier, pumps, etc 9,000
Two 20-in. venturl meters and recorders 3,000
$228,700
Add 15 per cent 34,305
Appro.kimate Estim.'^te op Operating Expense
$263,005
Interest, 4 perceutof $263,005 $10,520
Depreciation, 4 per cent of $263,005 10,520
Coal:
200 hr., 5 tons at $2.50 \
8560 hr., I ton at 2.50 / ^^'2°°
Oil, waste and supplies 1,500
R«palrs 2,500
Labor:
Men, cost per annum three 8-hr. shifts:
1 engineer $6600
1 oiler 4500
1 fireman 3000
2 coal passers .5400
1 janitor 700
20,200
Total $58,440
498 DISCUSSION
TABLE 4 GAS-ENGINE PUMPING STATION
Approximate Estimate of Cost
Capacity 10,000 gal. per min. ; pressure 250 lbs.
1 Excavation:
Retaining wall 68,400 cu ft.
Main pit 58,089 " "
Engine foundations 5,096 " "
Pump foundations 7,056 " "
Tunnel 5,496 " "
144,137 cu. ft.
= 5,339 cu. yd. at SI $5,339
Concrete:
Retaining wall 11,520 cu. ft.
Retaining wall footing 23,040 " "
34,560 cu. ft.
= 1,280 cu. yd. at $7 , 8,960
3 Building: 82 ft. by 79 ft. by 30 ft. = 19,430 cu. ft. at 15 cents . . 29,151
4 Foundations for pumps and engines, 450 cu. yd. at $10 4,500
5 Seven 1500-gal. triplex pumps, for 250 lb. pressure at $8900 62,300
6 Seven 300-h.p. gas engines at $10,000 70,000
7 Freight and erection .- 7,000
8 Suction pipes from city main and tunnel 6,500
9 Water discharge pipes 5,000
10 Gas connections 8,000
11 Air compressor plant 2,500
12 Gasolene tanks and piping 3,500
13 Two 20-in. venturi meters and recorders 3,000
$215,750
Add 15 per cent 32,362
$248,112
Estimate of Operating Expense
1 Interest, 4 per cent on $248,112 $9,924
2 Depreciation, 4 per cent on $248,112 9,924
3 Gas: 200 hr. at 18 cu. ft. per h.p. at $0.85 per M 6.426
4 Labor: 3 engineers at $2200 = $6600
6 asst. engrs.at 1500 = 9000
1 janitor at 600 16,200
5 Oil, waste and supplies 1,000
6 Repairs 1,000
Total $44,474
HIGH-PRESSURE FIRE-SERVICE PUMPS 499
TABLE 5 ELECTRIC PUMPING STATION
Approximate Estimate of Cost
Capacity 10,000 gal. per min.; pressure 250 lb. per sq. in.
Excavation:
Pump pit 63,936 cu. ft.
Retaining wall footings 8,640 " "
Pump foundations 2,048 " "
Building wall 1,692 " "
76,316" " =2,826cu.yd.at$l $2,826
Concrete :
Wall of pump pit 15,264 cu. ft.
Footings 7,892 " «
Bldg. foundation wall 920 " "
Bldg. foundation footings 329" "
24,405 cu. ft.
= 904 cu. yd. at $7 6,328
3 Building:
Pump room, 36 ft. by 56 ft. = 2016 sq. ft.
Switch room, 16 ft. by 56 ft. = 896 sq. ft.
2912 or say 3000 sq. ft.
3000 sq. ft. by 30 ft. = 90,000 cu. ft. at 15 cents 13,50o
4 Foundations for pumps and motors, 150 cu. yd. at $10 1,500
5 Four 2500-gal. centrifugal pumps at $5000 20,000
6 Four 600-h.p. 3-phase induction motors at $10,800 43,200
7 Suction piping from city main and tunnel 6,500
8 Discharge piping and valves in station 5,000
9 Switchboard and wiring in station 5,000
10 Two 20-in. Venturi meters and recorders 3,000
$106,854
Add 15 per cent 16,028
Total $122,882
Approximate Estimate of Operating Expense
I Interest, 4 per cent of $122,882 $4,915
I Depreciation, 4.3 per cent of $122,882 5,284
I Power bill:
Ready-to-serve charge, $25 per kw. = $37,500
$0,005 per kw. per hr., 200 hr. of full load $1,500 39,000
t Labor, 3 shifts:
3 engineers $6600
6 asst. edgineers 9000
1 janitor 600 16,200
5 Miscellaneous: oil, supplies, etc 1,500
S Repairs 1,000
$67,899
500 DISCUSSION
TABLE 6 ESTIMATED COST OF PROPOSED CHICAGO SYSTEM
Mains, Valves and Hydrants
I
District No. Cost
1 $477,508
2 329,321
3 152,018
4 128,457
5 1 109,178
6 314,569
7 82,791
8 1 178,420
9 ■ 146,432
10 118,916
11 113,268
12 ! 85,852
13 j 75,918
14 175,811
Total $2,488,459
Engineering and contingencies 373,269
I $2,861,728
4itations at $250,000 - 1,000,000
1 $3,861,728
No allowance made for land.
River crossings are assumed to be made as follows: (a) North branch in present Grand Ave.
water pipe tunnel; (i) Maio River in proposed LaSalle St. water pipe tunnel, to be built by
Chicago Railways Company; (c) South branch in present Harrison St. water pipe tunnel.
4 For the purpose of estimate it seems proper to assume a station
of a capacity of 10,000 gal. per min. against 250-lb. pressure, the
working pressure to be probably 150 to 200 lb. To avoid the crip-
pling of a station by the shutdown of any unit it seems advisable
to consider units of 2500 gal.
5 In discussing the various types of installations proposed, Mr.
Baker cited the advantages of each type. The direct-acting duplex
pumps are rugged and ready for immediate service, but their steam
consumption is large. The independent boiler plant necessary,
moreover, would be costly to build and to operate.
6 The gas-engine station has the advantage of lower first cost,
and no cost for power when not in operation. Though failure of the
gas supply is unlikely, gasolene could be used with a change of
HIGH-PRESSURE FIRE-SERVICE PUMPS 501
adjustment, or by running normally on illuminating gas with low
compression, which would be somewhat uneconomical. A gas-pro-
ducer plant might be installed, though this is somewhat open to the
same objection as the boiler plant.
7 Though electric motors are supplied from an outside sources
the large number of generating stations and feeders makes the electric,
supply as rehable as the gas supply. The first cost and the operating
expense of an electric station are low, though the standby charge is
high.
8 Connecting the system to stand pipes and to the sprinkler systems
in buildings had been recommended in Chicago and is the practice in
Winnipeg, Man., and Providence, R. I., and also with the gravity
system in Newark, N. J., Worcester and Fitchburg, Mass. The fire
systems of New York City and Philadelphia are not connected in
this way. The objection to these connections is that great loss of
water might result from broken pipes in the buildings. This could
be avoided, however, by placing a controlling valve in a brick chamber
outside the curb.
Edward E. Wall' outlined the proposed fire system for St. Louis,
which contemplates the installation of six or eight 5-stage centrifugal
pumps, electrically driven, at a station on Chestnut St., from which
the fire service mains will radiate north, south and west. The supply
for these pumps will be taken from the distribution system, a 36-in.
main being laid directly from the Bissell's Point pumping station
to the Chestnut St. station, and connected to the present distribu-
tion system by a number of by-passes. Connections will also be
made between two 20-in. mains on Fourth and Seventh Sts., to the
supply for the pumps, so that in case of failure of the 36-in. main,
the pumps may be suppUed from this source.
2 It would be practicable to draw the fire pump supply directly
from the Mississippi River by building an intake, but this would
probably cost more than the laying of the 36-in. main, and would
necessitate a charter from the Government. It would also raise the
question of obstructing navigation, since it would be necessary to
carry the construction well out into the channel, to insure an ample
supply of water. Supply from the river direct would also preclude all
connection with the distribution system, as it would be unwise to
risk the contamination of the city's water supply by river water.
' Asst. Water Commissioner, St. Louis.
502 DISCUSSION
3 The pumping capacity of the station at Bissell's Point will be
over 100,000,000] gal. of water every twenty-four hours, which is
more than twice the amount ordinarily consumed; the excess being
suflficient to supply more than 30 fire-streams through 3-in. hose con-
tinuously, assuming 300-lb. pressure at the fire pumps.
4 The 5-stage centrifugal pumps proposed for the Chestnut St.
station will have a capacity of 150,000 gal. per hr. each, against a
pressure of 300 lb. per sq. in. It is proposed to connect the station
with the power plants of the Union Electric Light and Power Com-
pany and the United Railways, so that two sources for power will
be available.
5 The three discharge mains from these pumps will be 24 in. in
diameter, the district supplied by them to be gridironed by a system of
12-in. mains laid on the enclosed streets and occasionally connected,
at crossings only, by by-passes, that the breakdown of one main
may. not necessitate the cutting out of any other line. The pipe
used will be cast iron, extra heavy, with bell and spigot joints,
double-grooved. All fire-hydrant leads will be 8 in. in diameter.
6 The system will be under the ordinary distribution pressure
when the fire pumps are not in use, so that for small fires the hydrants
will be available for use; when the fire pressure is put on the system,
the check valves on the by-passes will prevent additional pressure
from coming on the distribution system.
7 While the arrangement of machinery for the pumping station,
and the details of operation, have not been definitely decided upon,
it is possible that gas engines may be used instead of electric motors.
The questions of automatically starting and stopping the pumps,
maintaining the pressure during a fire, and the general details of
operation of the station, as well as the minor points of weight of
pipe, design*of hydrants, etc., have all to be worked out. It is esti-
mated that the cost of this system will approximate $3,000,000.
H. C. Henley*, speaking on the advantages of high-pressure fire
systems, said that they were chiefly valuable for the numerous
powerful streams which can be quickly brought into service and concen-
trated to advantage. For the prevention of conflagrations and for
keeping serious fires from spreading, more powerful streams are needed
than can be supplied by portable fire engines without considerable delay.
To obtain such streams from fire engines, it is necessary to " Siamese"
two or more lines into one nozzle, requiring considerable time; and
' Chief Inspector, St. Louis Fire Prevention Bureau.
HIGH-PRESSURE FIRE-SERVICE PUMPS
503
if a change in the location of engines becomes necessary, consider-
able time is again lost in re-assembling the hose lines.
2 The high-pressure system permits the use of hose of large diam-
eter— 3 in. and 3J in. — and direct connection to hydrants furnishes
a supply to nozzles of large area, without the necessity of siamesing
two or more hose lines. The 2-in. nozzle is best adapted for use
with high-pressure systems, this nozzle, under 75 lb. nozzle pressure,
discharging approximately 1000 gal. per min. A nozzle of this
area provides very effective service, as the loss of pressure, due to
friction in fire hose, decreases as the area of the hose is increased.
The data given in the table are derived from experiments by John
R. Freeman, and show the pressure required at the hydrant in
PRESSURE REQUIRED AT HYDRANT TO OVERCOME FRICTION LOSS
Hose Diameter
2ilN.
3 IN. ' 3J IN.
Hose lines
Single
Siamesed
Single Siamesedj Single
Siamesed
Smooth bore nozzle
liin.
2 m.
Uin. 2 in.
Uin.
2 in.
Length of hose line, ft
100
150
200
250
300
400
121
139
158
176.5
195
232
139
170
201
232
263
325
92 101
99.5 113
107 125
114.5 137
122 149
137 173
84.5
87.5
91
94.5
98
105
88
93.5
99
104. 0
110
121
For the 2-tn. nozzle It Is assumed that two hose lines of the length given are siamesed together.
hose streams of various lengths, to overcome friction loss and main-
tain 75-lb. nozzle pressure, the nozzle being at the same level as the
hydrant.
3 High-pressure systems should be considered as auxiliary protec-
tion and there should be no attempt at abandonment of engines
or other apparatus.
4 Direct connection from a high-pressure system to interior
standpipes, sprinkler equipments and open sprinkler systems, should
be made through Siamese connections and not through direct pipe
connection.
5 The inability of portable steam fire engines to furnish a stream
efficient to cope with serious fires is made apparent by tests made
by the engineers of the National Board of Fire Underwriters. The
steam fire engines for test were picked at random from the equipment
of many of the best city fire departments in the country.
504 DISCUSSION
t
Number of engines tested 102
Nominal capacity, gal 69,800
Actual capacity, gal 55,900
Percentage of efficiency 80
In many cases the efficiency of individual "steamers" is less than
50 per cent.
Edward Flad. It appears to me that a cast-iron pipe is rather
dangerous for high pressure. A cast-iron pipe tested under 300-lb.
pressure will often break at 75 lb. A wrought steel pipe is much
more reliable, and if properly coated, should last 25 or 30 years
under ordinary conditions. If steel pipe is absolutely reliable we
can afford to relay it at the end of 25| years" rather than to use
cast-iron pipe, which is liable to break.
2 In answer to a question by Mr. Flad as to the flexibility of the
joint used in Baltimore, Professor Carpenter replied that it is flexible,
in the sense- that it can be laid at an angle; it is not flexible so far
as change of form is concerned.
H. S. Baker asked what kind of steel pipe would be used in Balti-
more, Professor Carpenter answering that it is extra heavy steel
welded pipe, ^-in. thick, the ends being expanded into semi-spheres,
an 8-in. or 12-in. pipe being expanded just enough to get a ring in
it, and the whole bolted on the outside by external bolts, very like
a steam pipe.
Prof. H. Wade Hibbard. It is a fact that a cast-iron water
main has been in satisfactory use in city service for twenty years
and then a piece has blown out. It seems to me that the use of cast-
iron pipe should be prohibited for this special emergency purpose of
fire protection on account of its unrehability. In fact, in one of the
high-pressure systems using cast-iron mains, leaks have been known
to take place'and the pumps to run for a considerable interval, some
hours, I will say, and the pressure could not be maintained under
test, until it was finally discovered that the water had been pouring
out into a very large excavation and flooding it, unknown to those
operating the pumping station. Steel will show approaching deteri-
oration as cast iron will not.
2 Steel 'pipe ought to be good for thirty'years^of] service. That
period of Licrvice should be sufficient, and cities having such pipe
HIGH-PRKSSURE FIRE-SERVICE PUMPS 505
hould then be willing to replace it, having had more reliable protection
during that period of years, than cast-iron pipe could possibly give.
H. C. Henley asked if there had been any attempt made to pre-
vent the pipes from deteriorating through electrolysis. Professor
Carpenter answering that the Baltimore system is a continuous
metallic structure, from one end to the other, and he beUeved would
be thoroughly protected from electrolysis; or at least, better than by
any other system.
E. E. Wall. It is a fact that actually and not figuratively, steel
pipe must be handled with gloves when it is laid, because the
coating has to be very carefully preserved and can hardly be repaired
if it is broken in handling before the pipe is laid. This is a very
serious objection to the laying of steel pipe on account of exposure
to corrosion after it is laid.
W. H. Reeves. Owing to the magnitude and prominence of these
plants, the pump performances should be of interest to those desiring
information on centrifugal and turbine pumping machinery. The
foremost consideration in the art of building machinery of this class is
accuracy in design. Without accuracy in design it is not possible
to secure the maximum efficiencies within reach. A closely designed
pump should deliver exactly its contract number of gallons against the
contract pumping head, and the capacity should not run over nor
under. From a pump builder's point of view the misfortune of falling
short of the contract capacity needs no discussion here, but the other
misfortune of ruiming over on capacity may not be so clearly under-
stood. One effect of running over is an overload on the motor,
engine or steam turbine driving the pump, and another result is that
the average efficiency of the equipment in daily operationuis below
what it should be, for it it runs over in capacity its maximum efficiency
does not occur at its contract capacity.
2 It will be noted that each of these pumps had a contract capacity
of 3C00 gal. per min., against a total head of 308.66 lb. per sq. in.
Table 2 of the paper shows the performances of the five pumps at the
South Street pumping station. This table does not show the average.^,
but it will be found that each pump averaged approximately 3761 gal.
])er min. against a mean total head of about 313.1 lb. per sq. in.
Although the head was about 5 lb, above the contract condition, the
pumps exceeded the contract capacity by about 25 per cent. This, no
506 DISCUSSION
doubt, caused the motor overload mentioned in Par. 64. The contract
conditions implied 540 h.p. actuall}'' delivered, and at the guaran-
teed pump efficiency 770 b.h.p. would be needed. The delivered
work under test was 686 h.p., and according to the test efficiency of
72^ per cent, 946 b.h.p. was used, that is, approximately 23 per cent
excess motor load.
3 There appear to be no data on tests made on contract con-
ditions. As the pumps were tested at a great excess in capacity it is
quite probable that the efficiency would have been lowered several
points if the pumps had been throttled to the agreed capacity and
head. The tests as per Table 2 show about 686 h. p. delivered and
946 b.h.p., or a pump loss of 260 h.p. For a considerable range it is
probably safe to assume this 260 h.p. loss to be fairly constant.
Assuming this to be correct and adding this loss to the 540 h.p.
delivered represented by the agreed contract conditions, would give
800 b.h.p., thus showing a pump efficiency of but 67^ per cent. If
these pumps had been accurately designed, undoubtedly they would
have shown as high efficiency under the contract conditions as was
obtained with excess capacity condition.
Prof. E. L. Ohle. There seems to be quite a difference in opinion
among engineers as to the reasons for the variation in efficiency of the
pumps when working singly and in multiple. It seems to me that the
reason is the one suggested by Professor Carpenter. It is practically
impossible that all should work at the same speed, as they are
independently driven. If then the pressure in the main should exceed
the pressure which any pump was capable of delivering, the runner of
that pump would simply revolve without delivering any water. This
seems to be borne out by the experience of one pump builder, as stated
by J. J. Brown.
The Author. The discussion of the paper has been so volum-
inous that there is really but little needed from the author. In most
of the discussion additional information of value has been contri-
buted which I am sure will be appreciated by members of the Society.
2. The difficulties in connection with an installation of the kind
described in the paper, involving a complete system of piping and
hydrants capable of withstanding high pressiu-es, as well as the nec-
essary pumping machinery, are well brought out. I think the gen-
eral conclusion will be that the piping difficulties to be overcome,
especially when cast iron is employed, are very serious and require
fflGH-PRESSURE FIRE-SERVICE PUMPS 507
special skill and the best of material. Attention has also been called
to the fact that the city of Baltimore has adopted a system in which
steel pipe is employed in order to overcome the difficulties due to the
breakage of cast-iron pipe.
3 The discussion has disclosed the construction of several stations
in which the motive power has been obtained from gas engines, and
the advantages, disadvantages and expense of such installation.
4 It is pointed out that although the centrifugal pumps are cap-
able of operation at the high efficiencies shown by the paper yet at
the lower heads at which they are frequently operated the efficiency
would be less. I do not believe there is any serious commercial dis-
advantage because of that fact, since it is true that the cost of opera-
tion of a fire station is principally due together items than the cost of
power. A fire station is required to be, above all things, reliable, and
it is of very little importance whether or not the pumping be done
under the most economical conditions for the reason that the total
cost of pumping is only a small portion of the operating expense.
5 It is claimed by one of the discussors that the test should have
been made by the city at the exact capacity called for and the efficiency
should have been based on the result of such a test. This doubtless
would have produced a lower efficiency than that obtained. In the
light of the information now at hand, there would have been no
injustice in such a requirement, but at the date of making the con-
tract matters were different and such a requirement would have
imposed a penalty on the builders, which would have been of no ad-
vantage to the city. The reason for that opinion is, that at the time of
taking the contract the information regarding multi-stage pumps oper-
ating at high heads was qui»;e meagre. Mr. Sando, the designer of the
pumps, secured all the data he could both in this country and in
Europe. The result of his investigation led him to believe that it
was to the advantage of the city and of the builders to put in a pump
of such capacity that it would surely meet the requirements in that
respect. It was believed that this would result in a considerable
increased capacity over contract requirements. The motors were
designed with an equally liberal capacity so that the machine was
intended, even in the beginning, to be capable of a continuous large
overload. The statement that the motors showed any evidence of
being overloaded is in error, possibly because a certain remark which
I made was misunderstood. It strikes me that the city is the principal
gainer by such a system of design and that as a consequence it owns
considerable more pumping capacity than was called for in the speci-
fications, and so far as I Imow, without extra cost.
508
DISCUSSION
(5 1 believe that with the present data it would have been possible
to design both^pumps and motor to carry 25 per cent less load with
the same efficiency as was obtained by the larger pumps and motors.
In that case, a test at the specified capacity would have been a fair
one.
[The following curves show the development of the runners, guide wheels
and guide vanes of the pumps installed in the New York high pressure
pumping stations. — Editor.]
Developed Cylinder Section
I through M-N
Developed Cylinder Section
through X-Y
sifc
DEVELOPMENT OF
SUCTION GUIDE VANES
DEVELOPMtNT OF VANES
IN DISCHARGE RUNNER
Direclio;: of IJuuner
DEVELOPMENT OF DISCHARGE
GUIDE VANE
DEVELOPMENT Oh VANES IN
STATIONARY GUIDE WHEELS
<*=+
J
510 DISCUSSION
7 The interesting question brought out by these tests regarding
the higher efficiency obtained with a single pump as compared with all
the pumps discharging into the mam, has not been satisfactorily
answered. Such results, however, seem to have been noted by every
engineer who has made similar tests.
8 In Par. 65 of the paper I made one suggestion concerning this
point. I have since thought that the variation in construction or in
detailed shape of the discharge volume might possibly account for
some of these differences. It is hardlj"^ possible that all the pumps
can be made exactly ahke and smaP. inherent differences, which
would be obliterated in the operation of all the pumps together,
might account for the higher efficiency of the pumps operating
singly. As suggested by Mr. White, the measurements were of a
character which did not consider the pipe resistances, and the figures
given apply to the delivery from the pump before the water was
subjected to pipe resistances in any case.
No. 1250
STRESSES IN REINFORCED-CONCRETE BEAMS
COMPARISON OF EXPERIMENTAL RESULTS WITH RESULTS
OBTAINED FROM THE USE OF THREE THEORIES
OF DISTRIBUTION OF STRESSES.^^^j;^ .
By Pkof. Gaetano Lanza, Boston, Mass-IT" „
Member of the Society
^Lawrence S. Smith,* Boston, Mass,
' Non-Member
Many experiments have been performed^on the breaking strength of
reinforced-concrete beams, and in the course of them many obser-
vations have been made to determine quantitatively some of the
phenomena attendant upon the application of the breaking load, and
also upon that of smaller loads. A^evertheless it is well known that
the observations made thus far are not sufficient to furnish the means
for determining the actual distribution of the stresses, and hence for
the deduction of reliable formulae for the computation of the direct
stresses, shearing stresses, diagonal stresses, deflections, position of
the neutral axis, etc., under a given load.
2 The test of the vahdity of such formulae should be their agree-
ment with the results of experiments when the loads employed are
about one-fourth or one-third the ultimate loads, because, when the
loads are greater, the ratio of stress to strain varies very considerably
for the different fibres, while for loads smaller than one-fourth of the
ultimate, unknown initial stresses are Uable to exert so great an influ-
ence as to interfere with the deductions.
3 The object of this paper is to make a comparison of (a) the posi-
tion of the neutral axis, (6) the stress in the steel, (c) the stresses in the
concrete, and (rf) the deflection, as determined by experiment, with the
same quantities as computed by three well-known theories of the
• Instructor Massachusetts Institute of Technology.
Presented at the New York monthly meeting (November 1909) of The
American Society of Mechanical Engineers.
512
STRESSES IN REINPORCED-CONCRETE BEAMS
distribution of the stresses. The comparison was made in the cases
of eleven beams, in the testing of which the necessary observations
were taken. Of the eleven, five were tested in the laboratory of Ap-
plied Mechanics of the Massachusetts Institute of Technology, and
six in the laboratory of the University of Illinois.
4 The reinforcement consisted in each case of one or more longi-
tudinal bars placed near the bottom of the beam, and equal loads were
applied at the two points which divided the span into thirds.
Fig. 1 Theory A: Disitubution of Strains and Stresses at a Cross-
Section
5 The three theories employed in making the calculations, all of
which assume that at any given section the strain in any fibre is pro-
portional to the distance of the fibre from the neutral axis, will be
denoted by A, B and C respectively, and may be described as fol-
lows, the notation used in lettering the figures being explained sub-
sequently :
A This theory, which is very extensively employed, makes
the assumption that at any given section none of the
concrete below the neutral axis can be relied upon to re-
sist tension; and further that the stress is proportional to
the strain not only in the steel but also in the concrete.
6 This method is used by those who employ it to determine (a) the
position of the neutral axis, (6) the stress in the steel, (c) the stress in
the concrete, and sometimes the shearing stress at the neutral axis ; but
practically no attempt is made to compute the deflections by it.
Nevertheless for the purpose of comparison, deflection formulae de-
duced on this basis will be given. The distribution of the strains
and stresses at a cross section is shown in Fig, 1.
STRESSES IN REINFORCED-CONCRETE BEAMS
513
B This theory, which was proposed by Prof. A. N. Talbot,
also makes the assumption that at any given section none
of the concrete below the neutral axis can be relied upon
to resist tension; but instead of assuming the propor-
tionality of stress to strain in the concrete, the assump-
tion is made that the stress at any fibre can be represented
graphically by the corresponding abscissa of a parabola
drawn through the neutral axis; the axis of the parabola
being at right angles to the section, and its vertex at the
end of the abscissa which would represent the crushing
strength per square inch of the concrete, were the plot
continued to such a height as to correspond to this crush-
ing strength.
£"cc'
Fig. 2 Theory B: Distribution op Strains and Stresses at a Cross-
section
7 The quantities calculated by this theory are the same as in case
A, and again deflection formulae will be deduced on this basis for the
same reason as there stated. The distribution of the strains and stresses
at a cross section is shown in Fig. 2.
C The third theory is that proposed by M. Consid^re. He
claims that whereas in a plain concrete beam, the concrete
on the tension side cracks when the extension has reached
0.01 to 0.02 per cent, in a reinforced-concrete beam the
concrete on the tension side can undergo many times this
extension without cracking.
8 Among the tests which he cites in confirmation of this view is
the following: He says that he subjected one reinforced-concrete
beam to a load that produced in the lower fibre of the concrete an
514
STRESSES IN REINFORCED-CONCRETE BEAMS
elongation of 0.063 per cent as determined by measurement, and
another such beam to a load that produced in the lower fibre of the
concrete an elongation of 0.13 per cent, that he then removed the
Joads, chipped off the concrete below the reinforcement, and removed
the reinforcing bars, after which he smoothed off the lower surface
of the remaining portion of the beam and sawed out a concrete plank
from the lower side. He says that not only did this plank not fall
to pieces, but that on loading it transversely it bore as much as would
be expected from a plain concrete plank of the same dimensions.
9 In view of the above, M. Considere suggests that the distribu-
tion of the stress at a section is as shown in Fig. 3, the compressive
strength being represented by the triangle OABO, and the tensile
strength in the concrete by the trapezoid OCEFO, the value of CD
£ /"
¥iG. 3 Theory C: Distribution of Strains and Stresses According to
Considere
being equal to the yield point of the concrete in tension ; and that for
greater elongations the tensile stress does not increase.
10 However, inasmuch as the assumption of this distribution
would lead to great complexity in the calculations, he proposes as a
sufficiently close approximation that for the trapezoid OCEFO in
Fig. 3 we substitute the rectangle OHEFO. In this paper this ap-
proximation will be made in obtaining the formulae on the basis of
Theory C.
11 Before obtaining the formulae needed for 'naking the calcula-
tions, the notation used throughout will be explained.
Let a^ = strain in concrete at upper fibre of beam.
(Xg = strain in steel reinforcement.
E = ratio of stress to strain in concrete. In B this will
denote the initial ratio of stress to strain.
E^ = ratio of stress to strain in steel.
STRESSES IN REINFORCED-CONCRETE BEAMS 515
r = —
E
(Tq = compressive outside fibre stress per square inch in con-
crete.
(7a = stress per square inch in steel,
ttg = area of section of steel reinforcement in square inches.
^ = <73 flg.
h = breadth of beam.
n = — 5
b
h = distance from top of beam to centre of reinforcement,
inches.
\ = total depth of beam, inches.
y^ = distance from top of beam to neutral axis.
p = radius of curvature of vertical longitudinal section of
neutral layer.
W = total load applied.
M = bending moment at section at the center of the beam.
I = span, inches.
V = deflection at distance x from left-hand support.
Vq = greatest deflection, i. e., deflection at middle.
The above is the notation needed for A.
12 In B the same notation is used with the following in addition:
Let y = distance of any fibre above the neutral axis.
y^ = distance above neutral axis at which the fibre would
be subjected to the crushing strength.
X = strain of fibre at distance y above neutral axis.
a^ = ultimate compressive strain of concrete.
a = stress in fibre at distance y above neutral axis.
a^ = ultimate compressive strength of concrete.
, = "!■
,1
n, =
a'
3 r a„
&(3-9)
c?! = distance above neutral axis to point of application of
resultant of compression.
c?2 = distance below top of beam to point of application of
resultant of compression.
13 In C the same notation is used with the following- in addition:
Let t = yield point of concrete in tension.
d = \-h.
516 STRESSES IN REINFORCED-CONCRETE BEAMS
FORMULAE
14 Taking up the three theories successively, the formulae needed
to make the computation of the values of y^, o^, o^ and v^ will now be
given, the deduction being left to the reader.
Method A:
Vo = Vv? + 2nh- n (1)
3^ (2)
(3 /i - y,) a,
6^ (3)
(3 h-y,) by,
, ._ '"^ (4)
^o 1296 A
where
A=E [ra, (h - y,y + -f
15 In order to find y^, a^, and a^ by Method B, we have the
equations below all of which include q.
Method B:
Vo = ^ n\ +2nih - n^ (5)
M (6)
««\'^-^M73"-g)
M (7)
^ ^^ \ 3 (2 - g) ^ 12 (2 - g) y^
Hence before we can find the values of y^, o^ and a^ we need to deter-
mine q, and this will have to be done approximately. For this pur-
pose we can use the equation:
2 o "^^ (8)
16 Plot a curve having ° 's for abscissae, and ^^'s for ordinates.
then using for a^ a first approximation to its value, determine a first
STRESSES IN REINFORCED-CONCRETE BEAMS 517
approximation for q. Then determine a second approximation for
o-Q and from it a second approximation for q, etc.
17 In the calculations made here, with the load approximately
one-third of the breaking load, the value that has been employed is
q = 0.2.
18 For the deflection we have
23 WP ._,
Vo= — (9)
1296 A
where
A==E ^ra,{h-yX--^byo'^ ^^-]j
19 In order to obtain y^ by Method C we need to solve the equa-
tion of the fourth degree in y^.
Method C:
by„' + 2\b ih+ d) + 3 ra, \ y„' - 3 \ h {h + df
+ 6 ra> I - 2^^^ I y„^ + 6 ra, I ^^^- + 3 /i^ - rf^ j 2/,
{ 2M ^
-Qra^h I - + h^ - d' \ =0 (10)
20 The solution can be readily" effected graphically for any numer-
ical case by writing u equal to the entire left-hand side of the equation,
and plotting the resulting curve with ^^'s as abscissae and w's asordi-
nates ; then the value of y^ where this curve crosses the axis of abscissae
will be the value of 2/0 desired. Of course the equation has four roots,
but the one required can be easily identified as it must give a neutral
axis that lies within the section.
21 In solving this equation, some value of t, the tensile yield point
of the concrete, must be used. Considere suggests 170 lb. per sq.
in. for the concrete used by him, which was about six months old and
of a composition of nearly 1 — 2.5 — 2.5.
22 In the calculations made in this paper, t = 100 lb. per sq. in.
has been used, as the concrete was from 30 to 60 days old and its
composition was 1—3 — 6. After y^ has been found we can find a^
and (7^ from the following equations respectively :
,.a.3>^._M-<5(^-!,)3A+Ji« (11)
6 0
aM' 34-^. „ ^ + ,j (^_ _ y^^ 2h-K-y„ __^^ ^j2j
518 STRESSES IN REINFORCED-CONCRETE BEAMS
or more easily from the formulae
tbih, -y„)
2 Vo
Os = raJ'-y'^ (14)
For the deflection we have
«=-i^^-i! 05)
1296 A
where
\ '2 ^ 12 i
COMPARISON OF THE VALUES OF y^, a^ O^ AND V^ AS COMPUTED BY THE
VARIOUS THEORIES WITH THOSE DETERMINED BY EXPERIMENT
23 This comparison is exhibited in the tables. The first five
beams were tested in the laboratory of applied mechanics of the
Massachusetts Institute of Technology, and for these we have used
E^ = 28,000,000 and E = 2,335,000, and hence r = 12. The last
six beams were tested in the laboratory of the University of Illinois,
and for these we have used ^3 = 30,000,000, and E = 2,000,000, and
hence r = 15. All eleven were made of 1 — 3 — 6 concrete, the ages
being given in the tables. All were loaded with two equal loads ap-
plied at points dividing the span into thirds.
STRESSES IN KEINFORCED-CONCRETE BEAMS
519
TABLE 1
DETAILS OF REINFORCED-CONCRETE BEAMS
All Beams Loaded at Third Points
Designation
oi- Beam
Rods 1
Age
Days
b
Inches
h
Inches
A,
Inches
Span
Feet
E
Steel as
Area in
Square
Inches
Massachusetts Institute of Technology
A— 1
53
8
10
12
1
It
1.00
1.25
A— 2
49
8
10
12
1
u
1.00
1.25
B— 3
43
8
10
12
2
n
1.125
1.41
C— 5
35
8
10
12
4
it
1.00
1.25
E— 9
54
8
10
12
2
It
1.53
1.91
University of Illinois
11
65
8
10
11
12
4
i
0.785
0.99
27-'04
63
12
12
13i
14
4
It
2.25
1.56
28
60
8
10
11
12
4
?
1.77
2.22
33
60
8
10
11
12
3
i
1.325
1.66
36
60
8
10
11
12
[1
?}
1.473
1.S4
45
61
8
10
11
12
/3
\2
\}
1.473
1.84
* Reinforcement of area above center line of steel, per cent,
t Square.
t Twisted.
TABLE 2 DATA FROM TESTS ON REINFORCED-CONCRETE BEAMS
All Beams Loaded at Third Points-
Actual OR (Plot)
Deflection
Designa-
tion
OF Beam
Break- Nearest
iNG Load* § LoAnf
inches
Load Con-
sidered
W
vo (Plot)
Massachusetts Institute
of Technology
A— 1
15 000
5250 5.4
726
7941
4000
0.0731
A— 2
16500
5250 5.3
650
7644
4000
0.0749
B— 3
15950
5250 5 . 5
565
6615
4000
0.0660
C— 5
16240
4600 4.6
781
8246
4000
0.1015
E— 9
, 21000
6250 5.1
776
7563
5000
0.0769
University
of Illinois
11
11000
4000
4.8
740
11700
4000
0.175
27-'04
26900
9000
6.8
680
8250
9000
0.162
28
14300
5000
5.8
760
7800
5000
0.141
33
14400
5000
4.9
580
9000
5000
0.137
35
15000
5000
6.0
660
6750
5000
0.100
45
12400
4000
6.0
660
6750
4000
0.150
* Exclusive of weight of beam.
t Used in plots and in calculation for y„, ost a,,-
520
STRESSES IN REINFORCED-CONCRETE BEAMS
TABLE 3 RESULTS OBTAINED BY EXPERIMENT AND BY COMPUTATION
Massachusetts Institute of Technology Beams: r = 12, E = 2,335,000.
University of Illinois Beams: r = 15, E = 2,000,000.
Desig-
nation
OF
Beam
Actual
OR Plot
B
\q = 0.20
C
t = 100
Actual
or Plot
B I C
q = 0.20 ] < = 100
Massachusetts Institute "of Technology
A— 1
5.4
4.18
4.29
4.94
726
801
760
759
A— 2
5.S
4.18
4.29
4.94
650
801
760
759
B— 3
5.5
4.36
4.47
5.10
565
774
734
640
C— 5
4.6
4.18
4.29
5.06
781
703
666
653
E— 9
5.1
4.86
4.97
5.45
776
844
802
810
University of Illinois
11
4.8
4.15
4.27
4.93
740
670
634
645
27-'04
6.8
5.88
6.01
6.60
680
711
678
690
28
5.8
5.49
5.59
5.98
760
669
638
662
33
4.9
5.00
5.10
5.54
580
722
686
717
35
6.0
5.17
5.28
5.71
660
701
667
675
45
6.0
5.17
5.28
5.83
660
561
534
548
TABLE 4 RESULTS OBTAINED BY EXPERIMENT AND BY COMPUTATION
Massachusetts Institute of Technology Beams: r = 12, E = 2,335,000.
University of Illinois Beams: r = 15, ^ = 2,000,000.
Designa-
'^s
\
tion Of
Beam
ACIUAL .
OR Plot
B C
g= 0.20 t = 100
Actual
OR Plot
\ q= 0.20 1 / = 100
Massachusetts Institute of Technology
A— 1
7941
13420
13510
9333
0.0731
0.1163
0.1193
0.0921
A— 2
7644
13420
13510
9333
0.0749
0.1163
0.1193
0.0921
B— 3
6615
12000
12100
7374
0.0660
0.1076
0.1105
0.0865
C— 5
8246
11755
11840
7650
0.1015
0.1163
0.1193
0.0887
E— 9
7563
10720
10810
8110
0.0769
0.1105
0.1139
0.0942
University of Illinois
11
11700
14190
14210
9950
0.175
0.1778
0. 1824
0.1500
27-04
8250
11160
11240
8468
0.162
0.1822
0.1878
0.1573
28
7800
8296
8386
6672
0.141
0.1346
0.1400
0.1216
33
9000
10880
10960
8662
0.137
0.1592
0.1647
0.1406
35
6760
9842
9926
7609
0.100
0.1501
0.1549
0.1329
45
6750
7890
7941
5878
0.150
0.1201
0.1240
0.1033
STRESSES IN REINFORCED-CONCRETE BEAMS
521
TABLE 5 COMPARISON OF RESULTS
Per cent op Variation of Results on Tables 3 and 4 from the Actual Values Determined
BY Experiment
Designa-
Vo
^a
tion of
Beam
A
, B
1 c
j < = 100 .
A
B
C
t = 100
Massachusetts Institute of Technology
A— 1
-22.59
-20.56
-8.52
69.00
70.13
17.54
A— 2
-21.12
-19.05
-6.67
75.60
76.74
22.09
B— 3
-20.73
-18.73
-7.27
81.41
82.92
11.47
C-5
- 9.13
- 6.74
10.00
42.55
43.53
-7.23
E-9
- 4.71
- 2.55
6.86
41.74
42.93
7.24
University of Illinois
11
-13.54
-11.04
2.71 21.28
21.45
-14.96
27
-13.53
-11.61
-2.94 35.27
36.24
2.64
28
- 5.35
- 3.62
3.11 6.36
7.51
-14.45
33
2.04
4.09
13.06 20.83
21.78
- 3.76
35
-13.89
-12.00 1 -4.84 45.81
47.05
12.72
45
-13.84
-12.00 -2.84 16.89
17.65
-12.91
.\verage. . .
-12.40 \
-10.32
0.24 } 41.52
42.55
1.85
Values less than the actual are called negative.
TABLE 6 COMPARISON OF RESULTS
Per Cent of Variation of Results on Tables 3 and 4, from the Actual Values
Determined by Experiment
Designjv-
<^o
^0
IION op
Beam
A
B
C
« = 100 j
A
B
C
t - 100
Massachusetts Institute of Technology
A— 1
10.33
4.68
4.55
59.09
63.20
25.99
A— 2
23.23
16.92
16.77
55.27
59.23
22.96
B— 3
36.99
29.91
13.27
63.03
67.44
31.07
C— 5
- 9.98
-14.72
-16.39
14.58
17.54
-12.61
E— 9
8.76
3.35
4.38
43.70
48.11
22.50
University of Illinois
11
' - 9.46
-14.34
-12.83
1.60
4.23
-14.29
27
4.56
- 0.29
1.47
12.47
15.92
- 2.90
28
-11.98
-16.05
-12.89
- 4.54
- 0.71
-13.76
33
18.28
18.28
23.62
16.20
20.22
2.63
35
6.21
1.06
2.27
50.01
54.90
32. GO
45
-15.00
-19.10
-16.97
-19.93
-17.34
-31.15
Average. . .
6.19
0.88
0.66
26.50
30.25
5.76
Values less than the actual are called negative.
522 DISCUSSION
REMARKS AND CONCLUSIONS
24 The results seem to warrant the statement in Par. 1 that " the
observations made thus far are not sufficient to furnish the means for
determining the actual distribution of the stresses, and hence for the
deduction of reliable formulae for the computation of the direct
stresses, shearing stresses, diagonal stresses, deflections, position of
the neutral axis, etc., under a given load." It follows therefore
that whichever of the theories is adopted for practical use, it can be
regarded only as a sort of working hypothesis.
25 It seemed therefore desirable to compare the results of these
three well-known theories with those obtained by experiment. This
comparison can best be made by a detailed study of the tables, especi-
ally Table 5 and Table 6.
26 However, it seems plain, as far as the evidence of these eleven
tests goes, that in deducing the values of y^ and o^ theory C gives
results much nearer those determined by experiment than either A
or 5, and the same is true to a lesser degree in the case of v^, whereas
the differences are not so marked in the case of a^.
27 It also seems hopeless to obtain a reliable deflection formula
without taking into account the tension in the concrete.
28 Of course the computations are more easily made when A
is used rather than B or C, but in the cases of B and C the complexity
is not so great when designing a beam as when determining the stresses
in a given beam under a given load.
DISCUSSION AT BOSTON
Chas. T. Main. All engineers, civil, mechanical or any other, want
to know the most accurate way of figuring the stresses in reinforced
concrete. What I am more anxious to know is that the proper ingre-
dients are used, with proper mixing and good workmanship, so that
we may be reasonably sure of a factor of safety in the finished work
somewhere near what was intended. I have done no work of this
sort without constant supervision, and am obliged to say that I have
done no work that has been a source of pleasure to me. All of the
building materials in common use are, I think, more certain in results
than reinforced concrete. It is quite necessary to improve in the
use of this material and in workmanship, in order to produce work
which will inspire confidence.
Sanford E. Thompson. Professor Lanza's paper is of much value
as a means of comparing the various formulae used in designing rein-
STRESSES IN REINFORCED-CONCRETE BEAMS 523
forced-concrete beams, with the behavior of test beams under load.
Of the three theories the straight-Hne theory A is the simplest, and
to the writer this still seems the^best from a practical standpoint.
2 The formula derived by this theory as now used for determining
the depth of a reinforced-concrete rectangular beam (using the nota-
tion adopted by the Joint Committee on Concrete and Reinforced
Concrete*) may be expressed simply as
M
b
and the ratio of steel required is Ag == pbd
-xl
where d = depth of beam from compressed surface to center of
steel, in inches.
C = a constant for a given steel and a given concrete.
M = moment of resistance or bending moment in general, in
inch pounds.
h = breadth of beam, in inches.
Ag = area of cross-section of steel, in square inches.
p = ratio of cross-section of steel to cross-section of beam
above the center of gravity of the steel.
3 Theory B, where the stress is taken as varying according to a
parabola, is perhaps more exact than theory A, but at the same time
more complicated and diBBcult in practical application. Theory C
agrees more closely in the earlier stages of loading with the tests,
although tests made both in the United States and in Europe indicate
that Considere was not entirely correct in his assumption that steel
when combined with concrete permits the concrete to stretch to u
greater degree than when not reinforced. However, at earlier stages
of loading the cracks in the concrete do not extend up to the neutral
axis, so that more or less of the concrete is resisting tension and
assists the steel in taking the stress. For this reason a method tak-
ing into account the tensile value of concrete gives results closer to
the tests at early periods of loading than either formula A or B.
There are, however, quite important reasons, as will be shown in sue
ceeding paragraphs, why theory A is preferable.
'The Joint Committee is composed of representatives from the American So-
ciety of Civil Engineers, the American Society for Testing Materials, the Ameri-
can Railway and Maintenance-of-Way Association, and the Association of
American Portland Cement Manufacturers.
524 DISCUSSION
4 Reinforced concrete is a complex material, which if properly
used gives very safe and satisfactory structures. It is not, however,
of a kind to which hair-spHtting accuracy may be applied. In select-
ing a formula to use, the aim should be to choose one which wiU give
results always on the safe side and at the same time not very wide
of the mark. Referring to the paper, formula A gives results on
the safe side, while C errs nearly as often on one side as on the other.
5 The behavior of a reinforced-concrete beam under load may be
divided into two stages, the earlier stage where the concrete under
the neutral axis bears tension, which gradually merges into the later
stage, when the tensile strength of concrete is overcome and all the
tensile stress is taken up by the steel. In the earlier stage the stress
in steel increases proportionally to the moment, while in the later
stage the increase in stress in steel is composed not only of the increase
proportional to the moment, but also of the stress which in the previous
stage was carried by the concrete and after its cracking transferred
to the steel. Thus, for example, if a certain load W stresses the steel
up to, say 16,000 lb. per sq. in., an addition to the load of less than
W will double the stress. Therefore, a beam designed for a load which
would produce an actual stress in steel of 16,000 lb. per sq. in. would
have a factor of safety smaller than the ratio of that stress to the
elastic limit of the steel. It is safer, then, to base the design on the
results at the breaking load rather than on the results at earlier
stages of loading, and to use theory A, which at the breaking load
corresponds closely to the tests, and so be sure of the required factor
of safety. In designing, working stresses and working moments
should be used in the formulae.
6 The strongest argument against computing the concrete to bear
tension, in practical design, is the fact that reinforced-concrete floors
and other structures usually have to be buDt with joints between two
days' work. The bond of the concrete on the joints is imperfect, and
consequently the tensile strength of concrete at that point is small
and cannot safely be counted upon in design.
7 Theoiy A is very simple and clear. It has been adopted quite
generally in Germany and England, and I believe also in France,
although that is the home of Considere, while the Joint Committee
in this country has recently adopted it.
8 Theory A when used in figuring deflection does not give very
satisfactory results, but this is not an important factor in reinforced-
concrete design. When necessary to compute deflection, a more com-
plicated formula may be used which considers the tensile strength of
STRESSES IN REINFORCED CONCRETE BEAMS 525
concrete. The best of such formulae known to the writer are those
derived by Professor Thulhe of Austria, which are based on more
logical assumptions than are the formulae of Consid^re.
9 It must not be forgotten that the computation of the stress in
the middle of a supported beam is only one part of the theory of rein-
forced-concrete design. Just as important as the design of the beam
in the center, since reinforced concrete is usually built continuous over
several supports, is the design of the ends of the beam, and of no
less importance is the part designed to resist the tendency of the
diagonal tension to produce diagonal cracks.
10 It maybe said then in conclusion, that although not correspond-
ing strictly with tests, the ordinary straight-line theory is the one
which will probably be used for some time to come because of its
simplicity, and because reinforced-concrete beams, designed accord-
ing to this theory, with due regard to other details, will produce, with
good workmanship, structures which are unquestionably safe and
conservative.
11 Except for a few isolated examples, it is less than ten years
since reinforced-concrete buildings began to be erected; the 16-story
Ingalls building in Cincinnati was built in 1903, and still stands as the
most notable example of a concrete office building. And yet, as has
been stated by Professor Burr, we already know more about concrete
columns than about steel columns; the tests have been more exact,
and more nearly conform to practical conditions. The beam theory
is still in the stage of development, and tests and mathematical
demonstration which tend toward more economical and rational
detailing are welcome. Nevertheless, we may say with surety that
buildings all over the country which are being designed by the
common formulae with conservative stresses, and erected with proper
care, are safe and conservative.
F. S. Hinds said that he had had a very profitable experience in the
last two or three years in the construction of a large office building
built entirely of reinforced concrete, erected for the Phelps Publishing
Company at Springfield, Mass. The building covers an area of
30,000 sq. ft. and is eight stories above the sidewalk. In the con-
struction of the building it was demonstrated that good work can
be done with reinforced-concrete, and that there was no mistake
in selecting concrete for both the interior and the exterior of the
building.
2 His observations had led him to believe that this con-
526 DISCUSSION
struction in buildings even higher than eight stories will yet be
seen. In fact, there is such an oflBce building in Cincinnati, 16
stories above the sidewalk, showing that reinforced- concrete can be
used in competition with the steel frame.
3 Answering a number of questions by Desmond FitzGerald,
Mr. Hinds said that the concrete for the building was mixed by
machine, crushed stone of "pea" size being used. The proportions
of the mixture were 1-2-4, just enough water being added to make the
mixture solid and yet make it flow easily. The ramming of columns
was not done in the usual way, but the concrete was settled by means
of four or five poles. Both round and twisted rods were used, held
in place by small wood blocks which were withdrawn as the mixture
was poured into the form.
4 Continuing, Mr. Hinds said that the great secret in concrete
work is in getting the rods in the proper places. Supervision and
careful preparation of the mixture and handling of materials will
bring the best results. An oil paint and cold water paint without
plastering have been used on the inside of the building, showing how
smoothly the surface was finished.
5 In answer to a question Mr. Hinds said that moisture was
prevented from going through the walls by their thickness — none
being less than 8-in. thick — and by the density of the concrete.
He had seen no cracks whatever in the reinforced-concrete proper,
the only crack in the building being one near the top of the elevator-
well partition, caused by expansion and contraction. Here and
there a small crack appeared in the granolithic floor.
Prof. C. M. Spofford.* I presume we all agree with the previous
speakers that concrete should be handled carefully, as it is subject to
great variations. I feel, however, that merely to be careful is not
enough; we should determine the theories as correctly as possible.
and use them to eliminate so far as possible such uncertainties as now
exist.
2 I am surprised that the C formula, as Professor Lanza has
called it, gives results closer to those of actual experiment than the
other formulae, and hope that the present data may be extended
by further tests and computations. As far as actual use in design
is^concemed, any one of these theories may be safely used, provided
a liberal factor of safety is employed, but further study and in-
* Massachusetts Institute of Technology.
STRESSES IN REINFORCED-CONCRETE BEAMS 527
vestigation along the lines indicated may enable us to determine
more precisely what the factor of safety should be.
J. R. Worcester.! The careful study which the authors have
devoted to these eleven beams is of great value, and their deductions
show how much can be learned from a few experiments made with
care and recorded with scientific accuracy.
2 It seems to the writer, however, that a few other points of
interest in the tables are worthy of comment; as, for instance, the
fact that in two of the beams, A-1 and A-2, alike so far as dimensions
and amount of reinforcement are concerned, there appears to be a
variation of 0.1 in. (1.9 per cent) in the actual location of the neutral
axis; of 76 lb. per sq. in. (12 per cent) in the stress in the concrete; of
297 lb. per sq. in. (3.9 per cent) in the stress in the steel, and of 0.007
in. (10 per cent) in the deflection.
3 Another remarkable variation in the behavior of beams appar-
ently alike is that of No. 35 and No. 45, where the latter with 80 per
cent of the load of the former had the same actual deformations in
steel and concrete, indicating the same location of neutral axis, and
at the same time 50 per cent greater deflection.^These great differ-
ences may perhaps be due to the fact "that No. 45 was cracked before
the test began, and therefore possibly should be excluded from such a
comparison as this, though the cracking did not prevent the beam
from developing fairly satisfactory strength. These striking instances
of variation in observed results, where every precaution was taken
to make the conditions identical, render it important to select theo-
ries of computation safe for the worst results found experimentally.
4 Speaking from a practical standpoint, several of the elements
compared are not of vital importance. The location of the neutral
axis is used only as an intermediate step in the process of calculation,
and, if fairly correct results can still be obtained, error in this part of
the calculation is not serious.
5 Then, again, the deflection is rarely of great importance. It is
comforting to know that beams do not deflect as much as if the con-
crete had no tensile strength, but practically this is as far as we are
usually concerned.
6 The actual compressive stress in the concrete may also be
eliminated from consideration in actual construction, if only we can
limit the area of steel to such a percentage that we are sure failure
!J. R. Worcester, 79 Milk St., Boston, Mass.
528 DISCUSSION
from the compression of the concrete will not occur until the steel has
been stretched beyond the elastic limit. In this connection it is
worthy of note that the beams quoted were with one exception more
heavily reinforced than is usual at the present time. With 0.8 per
cent of steel, or even with 1 per cent, it is safe to base our calculations
for moment upon the stress in the steel only.
7 The element then about which the most interest centers is the
stress in the steel, and it is important that we should adopt a method
of computation which gives this with the least error practicable, and
with that on the safe side.
8 Looking at Table 5 with these considerations in mind, we find
little difference between methods A and B, both giving results well
on the safe side. Method C, while averaging very closely to actual
results, gives errors on the wrong side in five out of the eleven cases
cited; in one case, and that the one most resembling usual practice,
having an error of nearly 15 per cent on the unsafe side.
9 It is noticeable also that the loads assumed are considerably
less than what would usually be considered working loads for the
beams in question. Following almost universal practice at the
present time, the stress in the steel as computed would be allowed
to go to 16,000 lb. per sq. in. This would permit loads on the Uni-
versity of Illinois beams as follows:
No. 11, 5,000 lb in place of 4000 lb.
No. 27, 12,000 lb. in place of 9000 lb.
No. 28, 10,000 lb. in place of 5000 lb.
No. 33, 7,000 lb. in place of 5000 lb.
No. 35, 8,000 lb. in place of 5000 lb.
No. 45, 8,000 lb. in place of 4000 lb.
Only these six are quoted because the essential facts regarding
them are given in the bulletins of the University of Illinois, while
we have not at hand the details of the tests at the Massachusetts
Institute of Technology.
10 The diagrams of these beams indicate under the above loads
the stresses in the steel indicated herewith in Table 1, using the
authors' modulus of elasticity, 30,000,000 lb. In the same table are
given the stresses in the steel as calculated by methods A and C, an 1
the percentage of error by each method.
11 Comparing these results with those obtained by the authors
as shown in Table 5, we find that the common method of computa-
tion, A, gives considerably closer results to those observed than under
the lower loading. The error ranges from 5 to 20.6 per cent, with an
STliESSES IN HEINFORCEl) CONCRETE BEAMS 529
TABLE!— STEEL STRESS UNDER HEAVIER LOADING
BsAU No. Load Used
Stress in Steel, Lb. Per Sq. In.
Error op Calcdlation
Per Cent
Actual
By A
By.C
j
By A 1 By C
11 i 6,000
27 12,000
28 10,000
33 7,000
35 8,000
45 8,000
1
15,600
13,500
14,700
12,600
13,800
15,000
17,700
14,900
16,600
15,200
15,750
15,750
13,600
12,600
15,000
12,900
13,900
13,900
+ 13.5
+ 10.4
+ 12.9
..+20.6
+ 14.1
1+ 5.0
-12.9
- 6.7
+ 2.0
+ 2.4
+ 0.7
- 7.3
Average error
+ 12.76
- 3.6
average of 12f per cent, always on the safe side. On the other hand,
by the Consid^re method, C varies from + 2.4 per cent to — 12.9
per cent, with an average of 3.6 per cent on the unsafe side. This
would indicate that there is no advantage in adopting the more
laborious method, involving the solution of an equation of the fourth
degree, at least so far as proportioning the steel is concerned.
12 The chief difference between the two methods, as explained
in the paper, is in the assumption in the Consid^re method of a certain
value for tension in the concrete below the neutral axis, and the dis-
regard of this in method A. There is no question that under ordinary
conditions the concrete has a small amount of tensile strength while
the loads are small, but there is grave doubt as to the safety of rely-
ing upon a crystalline material under such conditions. Many con-
ditions in actual construction may tend to destroy the tensile strength.
There may be set-joints near the center of the beam; there may be
voids near the bottom where the mortar has leaked out; there may be
incipient invisible cracks extending to an unknown distance. It is a
fortunate circumstance that ease of calculation is on the side of the
safer method, for this is a powerful incentive to its adoption.
13 The statement at the opening and close of the paper that "the
observations made thus far are not sufficient to furnish means for
determining the actual distribution of the stresses," etc., is undoubt-
edly true, speaking literally and with scientific accuracy. At the
same time it should be borne in mind that we are dealing with a crude
product which cannot in practice be made with scientific accuracy.
It is doubtful whether absolute knowledge of the laws of distribution
530 DISCUSSION
of stress in a theoretically perfect material would be of any great
advantage in designing structures of every-day material. The impor-
tant question is whether we know enough to design our beams with
entire safety and reasonable economy. To this query the writer
would unhesitatingly give an affirmative answer. The investigation
of these beams tends to confirm this opinion, which is also supported
by the constantly accumulating experience with actual construction.
We would therefore venture to add two other conclusions to those
advanced by the author, namely:
a Experiments indicate that, though precise determination
of the laws of stress distribution may be impossible in the
present state of our knowledge, sufficiently close approxi-
mations may be made for all practical purposes.
b The simple method of calculation, by neglecting ^tension
in the concrete and assuming a straight-line distribution
of the compressive stress, is the easiest to apply and gives
satisfactory results for the determination of the stress in
the steel.
Prof. Geo. F. Swain. I notice that Professor Lanza has used a
value of E = 2,335,000 for the beams tested at the Massachusetts
Institute of Technology, while for the beams tested at the University
of Illinois he has used a value for E of 2,000,000. The beams tested
at the Massachusetts Institute of Technology were from 35 to 54
days old, while the beams tested at the University of Illinois were
from 60 to 65 days old. The modulus of elasticity ought to increase
with age, other things being equal, yet in these tables the reverse
is assumed. This fact might account for some of the peculiarities
and the results. Professor Lanza does not state whether he measured
the modulus of^elasticity.
2 In Table 2, I think the heading of ^the column "Nearest one-
third Load, " is a little confusing. Those figures are not very close
to one-third the load, and beam C-5, which has a larger load than the
first three beams, has a smaller value in the third column. I suppose
the third column simply me^ns the loads for which computations
were made, and that the loads were appUed in such increments that
the figures given represent the nearest third of the load for which
computations were made. Yet it seems rather confusing that for
a load of 16,240 lb., the nearest one-third should be given as 4600 lb.
STRESSES IN REINFORCED-CONCRETE BEAMS 531
3 With reference to the three theories, 1 have never believed in
Considcre's main contention, namely, that ])y reinforcing concrete
such great strains could be produced without fracture; though his
explanation is in a certain degree plausible. If a body is stretched
so that the molecules are a certain distance apart, nothing can pre-
vent fracture. Ductile material like steel draws down at the point
of fracture and is stretched much more there than on the average
through the length of the piece. If concrete were a ductile material,
its adhesion to the steel bars might prevent any such phenomena as
drawing down and thus distribute the strain; but concrete is not a
ductile material, i nd there seems to my mind to be no possibihty
of the great stretch mthout fracture which Considere claims.
4 As to the results obtained by the three formulae, I think those
given in the tables were precisely what might be expected, because
these loads were only large enough to be called working loads; that
is, they were nothing like the ulti nate load. As a matter of fact there
was tension in the concrete, und( r which condition the steel would be
relieved; we would therefore expect that in case C the stress in the
steel would be very much less than in the other two cases. In practice,
also, there is undoubtedly tension in the concrete unless cracks
occur. The results of tests made by the Boston Transit Commission
show large tensile stresses in concrete beams without reinforcement.
5 However, the question is what to do in designing. In practice
there may be cracks in the concrete, not due to stress, but to the
moving of blocks on which the rods are set, making the cement run
out, or due to shrinkage or joints or other causes; for which reason it
seems to me that in practical designing, engineers are not justified
in assuming any tension in the concrete.
Henry F. Bryant.* Mr. Worcester stated (Par. 9) that on apply-
ing his reasoning to the University of Illinois experiments, the nearest
one-third load for 1(3,000 lb. of stress on the steel would be found to
be nearly double that given in the paper as approximately one-thiid
the breaking load. This emphasizes the question of the yield point.
The rather common practice, as Mr. Worcester states., is to take from
12,000 lb. to 16,000 lb. on mild steel and \\-ith this to use about 500 lb.
as the concrete compressive strength, which, with concrete of 2000-
Ib. compressive strength, gives a factor of safety of four or possibly
five. If the yield point is the critical point in the steel, we are using
a factor of safety of only between two and three in the steel. Mr.
'Engineer, 334 Wasluhgton St.', Bo'ston, and'Brookline, Mass.
532 DISCUSSION
Worcester's analysis of the Illinois experiments would indicate that
instead of breaking at three times what would be considered a safe
working load, the beam would break at not over twice the load. I
think that using mild steel and a factor of at least four, and figuring
that the yield point is the critical point of the steel, we should apply
to the steel something like 7500 lb. or 8000 lb., with 500-lb. compression
on the concrete. That means a little larger percentage of steel than
is common practice, though it is not unusual to adopt this reasoning
with high-carbon steel. I am very glad to see that these experiments
point that way.
H. E. Sawtell.i Consid^re's theory of stress distribution agreeci
very well with the actual tests at about working loads on the eleven
beams mentioned in the paper. We know, however, that his theory
will not agree with breaking-load results as well as either the straight-
line or the parabolic theory, each of which considers that concrete
takes no tension stress. We should adopt a theory which will agree
quite closely with tests at breaking loads, but which will always be
on the safe side for intermediate loads. We can then get a real
factor of safety.
2 Referring to Par. 24, it seems likely that when applied to floor
beams, a formula will remain only a sort of working hypothesis if our
theories are to be based upon test beams which are not more like the
beams used in actual practice, and if our compressive value for con-
crete is based upon plain concrete. The present uncertainty may
appear to favor the side of safety, but on the other hand, when too
many assumptions have to be made, there is little real satisfaction in
working with the material.
3 Tests on rectangular beams are necessary for determining as
nearly as possible the stresses and deflections in slabs and separately
molded beams, butMo not seem to solve the problems of beams and
girders as used in actual construction. Let us first note some of the
stresses as they exist in a beam in actual construction, assuming Fig.
1 to be the cross section of a beam at its place of maximum flexural
stress. The slab steel is placed at the beam, as a great many designers
consider necessary, in order to resist fully and reliably the negative
slab stress, etc., at the beam. These slab rods always are only a few
inches apart, and pass through the top of the beam concrete at right
angles to the compressive stress of the beam.
* Structural engineer, with Chas. T. Main, Boston, Mass.
STRESSES IN RKINFORCED-CONCRETE BEAMS
533
4 Assuming that the concrete in both beam and slab is poured at
the same time, we know of course that for some distance each way
from the beam the slab will work with the beam in resisting compres-
sive stresses. Assumptions are made as to what part of the slabs
will work safely with the beam, and then the beam is calculated for
and designed as a T-beam. In doing this the full working stress for
concrete in compression is used. The concrete at G, E and F has a
large share of the compression to take care of. Also, as a result of
placing the slab steel at the top, as it passes over the beam, the
concrete at G, E and F is again put in compression, this time at its
full working value, but at right angles to the compressive stress in
the beam.
P
A
fSrx-r/'ace. |
Fig. 1 Cross-Section of Beam at Maximum Flextjral Stress
5 Again, the maximum vertical shear in the slabs is along the lines
BB' and A A', this shear, it will be noted, being through concrete
already doing double duty in compression. The concrete at the sur-
face is at the place of maximum compressive stress of the beam and
it also has a maximum tensile stress due to the negative slab moment.
6 The total compression at G, E or F is very much higher than
we would willingly put upon plain concrete as a working stress, while
the concrete at points E or F is in a worse condition. At the sur-
face the material is nearly cracking from a tensile stress, even under
working loads, and it cannot be of much service in compression where
it is most needed Ipy the beam.
7 If these conditions are correctly noted, and if the actual stresses
534 DISCUSSION
are to be kept down to the unit working stress ol 'plain concrete, then
it will be necessary either to assume a much lower unit stress for con-
crete when designing T-beams, or to design a rectangular beam whose
effective top surface does not extend above the slab rods shown in
Fig. 1. But is it necessary to use the value of plain concrete when
designing T-beams? Are we not justified in saying that concrete at
G is confined, and being reinforced, has a much higher ultimate
strength than plain concrete?
8 The compressive strength of concrete in beams is increased in
two ways (a) by lateral restraint, brought about by the surrounding
compressive forces; (6) by reinforcing its shearing resistance, which
may be greatly assisted by placing the rods H at the points shown in i
Fig. 1. These H rods are to be used only at and near the place of
maximum moment in the beam and should be quite close together.
9 But how much does this increase the strength? As bearing
upon the subject, an extreme case may be cited from a paper by Leon
S. Moisseiff read before the American Society for Testing Materials.
The compressive strength of cubes of concrete, reinforced in every
direction by a large percentage of metal in the form of nails, was
increased to two to three times the strength of plain concrete. Some
designers have already noticed an increase of strength under similar
conditions and are taking advantage of it, but are making assump-
tions regarding its amount for different percentages of reinforce-
ment.
10 So far as the writer knows, no T-beams have been tested with
their flanges reinforced and loaded in such a way as to carry their
loads to the beam and thus to strain the beam in the same manner as
in actual practice. It seems that future tests should be along such
a line, if greater economy is to be reached in design and our knowledge
is to become more exact with fewer assumptions made.
11 In conclusion, it would seem as though the slab concrete were
overstrained at E and F, where it is used for T-flanges, for negative
slab compression and for vertical shear from slab loads. Unless it
can be ascertained whether lateral restraint, and the use of the rods
as shown, will increase the strength necessary to resist this strain
safely, it would be better not to calculate for T-beams, but to make
the rectangular section sufficient to meet the stress. Even this
rectangular section should be designed with a conservative concrete
compressive stress, because its top surface is generally considered at
about the point where the slab rods pass over it, this including the
concrete at G.
STRESSES IN BEINFORCED-CONCRETE BEAMS
535
12 Fig. 2 shows the cross section at the center of a T-beam, and
a method of loading which seems to give promise of results which will
come nearer to showing how beams in actual construction are stressed
than rectangular beams whose compressive side is wholly plain con-
crete. The load over the stem should be less than the flange loads;
and should agree with actual floor loading where the slabs carry most
of the loads to the beam and produce tension in the rods and concrete
at the surface over the stem, compression at the under side of the
slab at the stem and shear near the stem. As tie rods are always
Fig. 2. Cross-Section at Center of T-Beam, showing Method
OF Loading.
used in practice it would be well to use them here. They are shown
by dots in the diagram. The slab rods in this case are bent to act
as anchors, and the tie rod at the edge is wired to them on the
inside.
13 It is acknowledged that the loads on the flanges do not stress
them quite as they would be stressed in a floor system; but if the com-
pression, tensile and shear stresses are not more than those that
would be produced, were the slabs continuous, it is thought that as
their stress is at right angles to the beam this difference will make no
practical difference with the results on the beam.
536
DISCUSSION
DISCUSSION AT NEW YORK
E. P. Goodrich.* The several theories which were the basis of the
formulae used by Professor Lanza are approximations to actual con-
ditions, and are made the basis for calculating special points in con-
struction work. The first method is used primarily because of its
ease of application to ordinary conditions, and the factors now intro-
duced into the formulae are based almost exclusively on the results
of actual tests. vijFor instance, in the particular series of tests made
at the Massachusetts Institute of Technology the ratio of the modulus
of elasticity as found by experiment to the computed value is only
eight and a fraction. On the other hand, diagrams of Professor
Fig. 2 Stress-Strain Diagram for Tension and Compression
Talbot's beam tests, in which the position of the neutral axis is shown,
give a ratio of more nearly eighteen, showing that the factor intro-
duced has no real relation to actual conditions. It is the adaptation
of a formula to tests, rather than the use of a formula to check various
kinds of investigations. Occasionally the straight-line formula has
been used to compute deflections and stiffness, as was reported not
long ago in an article published in Engineering News; but as to the
accuracy of this use there has been some adverse criticism.
* Consulting Engineer, 1 Madison Ave., New York.
STRESSES IN REINFORCED-CONCRBTE BEAMS 537
2 As has been said, Consid^re's theory was based on certain experi-
ments, the accuracy of which has also been questioned. Professor
Morsch of Zurich argues both for and against them in his book entitled
Eisenbetonbau, describing certain experiments with concrete beams,
in which he determined the stress-strain diagram for both tension and
compression, finding some such conditions as that shown in Fig. 1.
If in any beam section, the neutral axis be established, and the actual
stresses laid down graphically above and below this neutral axis at
any point, and if the centroids in each section are determined, and
the distance between them measured, the moment which must theo-
retically be sustained by the beam can be computed. Morsch tested
some specimens both in compression and tension, and also in bending,
and computed the theoretical bending moment and ultimate strength
by methods similar to Consid^re's, using a practically constant stress
in the concrete below the neutral axis. ' He found that the theoretical
bending stress in kilograms per square centimeter was 20.7, while
that found as an average of three actual experiments was 21.4, show-
ing a very close agreement in this particular instance.
3 In the case of three other beams in which the percentage of
steel varied from one-half of one per cent to very nearly two per cent,
MSrsch made a similar computation based entirely on a stress relation
similar to that of Fig. 1. He found the resultant of the two tensile
stresses, in the concrete and the steel, then measured the distance on
his diagrams between the centroids of compression and tension, and
computed the moment, which was found to correspond cosely with
the test conditions.
4 Another series of tests of considerable interest is that made by
Dr. Miiller for his doctor's thesis for the Hanover Technical High
School. He treated concrete beams in a manner similar to that of
Professor Lanza, except that he used thirteen points in the depth
of the beam, and measured by three methods the actual strain rela-
tion which existed at different times. In all his work he used simply
a safe working stress, to the limit allowed by the German govern-
ment regulations. He found that in a solid bfrm the stress vaiied to
a certain extent, was very nearly of the straight line type when meas-
ured at all his thirteen points ; while with a beam in which he built in
fourteen artificial cracks by putting sheets of metal close together in
the beam, he found that the stress relation^more neaily corresponded
with Consid^re's theory These artificial cracks produced a variable
stress between the sections, so that the stress in the steel was actually
less between the cracks, some of the stress being thrown into the con-
538 DISCUSSION
Crete, as illustrated graphically in Fig. 2, in which the ordinates above
the base measure the tensile stress.
5 The question of shear has been mentioned, but its effect upon
deflections has not been discussed. The writer believes this is very
important, because of two series of tests which he made some years
ago on beams, one series of which was reinforced only by horizontal
rods, and the other by vertical stirrups also. The deflection was
three or four times as much in the case of the beams without the verti-
cal steel — shear reinforcement — as in the case of beams with con-
siderable vertical reinforcement. Each series had exactly the same
amount of steel in tension. Of course theoretically the vertical
stirrups could not affect the tensile stresses in the bottom of the beam.
The ordinary theory by which deflection is computed does not include
a factor for shear, which actually does have some effect on the deflec-
tion, both theoretically and, as shown by these tests, practically. It
must be taken into account, as well as the tension in the concrete, if
the actual conditions in the beam, especially with regard to stiffness
and deflection, are to be considered.
Fig. 2. Variable Stress Produced by Artificial Cracks
6 It seems necessary that some relation between deflection and
stress should be definitely determined, because deflections can be
more easily measured in any beam test than any other phenomena
Almost every novice determines the deflection, although he does not
know the relation between it and the stresses involved. It is only
through discussions such as this that some true basis can be reached
for the computation of the stresses involved in continuous members.
7 There is another point concerning which the writer has made
some experiments. By means of plaster of Paris, ordinary sharp
(jarpet tacks were applied to the sides of a beam, with the points
sticking outward. The beam was loaded centrally, and the actual
deflection curve was simply picked through a piece of paper from time
to time as the load was increased. The curves were then enlarged
and used as a basis for comparison with the theoretical elastic curve
of a beam loaded centrally. There was a very large discrepancy,
STRESSES IN REINFORCED-CONCRETE BEAMS 539
which was more nearly coordinated when it was assumed that the load
was distributed over a length something like one and one-half or two
times the height of the beam. It is to be hoped that experiments will
be made in regard to the deflection of beams and the distribution of
stresses, so that some true relation can be determined, between this
element, which is easily measured, and the other elements which are
usually unknown: that is, in regard to the relation between deflections
and the actual stresses of compression and tension.
Prof. Walter Rautenstrauch. I regret that more observations
are not recorded and plotted in the paper and that the methods of
making the computations and obtaining the data are not given. It
would be interesting to plot the variation of deflection with load as
observed, and as computed by the three formulae selected for com-
parison.
2 I would ask Professor Lanza how he made his observations for
the strain in both concrete and steel and also how he determined from
these the neutral axis of the section. If these data were submitted
it would be possible to make a comparison with results obtained by
assuming other possible values of E, for example, and thus to ascer-
tain to what extent the differences reported might be due to assumed
and possible actual values.
3 As concrete construction is for the most part monolithic, and
very few beams of the particular kind tested are used, I believe it is
of much broader interest to investigate methods of measuring strain
and computing stress than formulae for simple beams. It is a fact,
I believe, that all the data reported in this paper as actual stresses in
concrete — actual stresses in steel — were obtained, not actually, from
direct observations, but rather from relations between stress and
strain assumed to exist in the concrete or steel. The same I believe
is true in regard to the determination of the neutral axis. If Pro-
fessor Lanza will tell us what assumptions he made in determining
these values we will be in a better position to judge their worth.
4 I need hardly call attention to the fact that the modulus of elas-
ticity for concrete in tension and compression is quite variable. It
seems to depend upon the age of the concrete and the intensity of the
stress. I believe it would have been of some value to take a slice from
the end of these beams and obtain a stress-strain diagram, in order to
( ompute the several values of E and the limits of stress for which each
value of E is constant. Otherwise the actual values of the stress are
not much more reliable than the values as computed by the formulae,
since both are computed from assumed relations.
540 DISCUSSION
5 It is interesting to note that Formula B is based on a rational
assumption concerning the variations in compressive stresses above
the neutral axis. The fact has been well established that the stress
varies as the ordinates of a parabola, and not as the ordinates of a
straight line. On the other hand, I am inclined to doubt the state-
ment of Considere that the concrete on the tension side can undergo
an extension much greater than 0.02 per cent without cracking, when
the beam is reinforced, whereas when not reinforced the concrete
cracks when the extension is from 0.01 to 0.02 per cent. The mere
fact that a reinforcing rod is present does not seem sufficient to change
the physical properties of the concrete.
6 I believe Professor Turneaure has shown Considere to have
been wrong in this assumption. It is not at all unlikely that Con-
sidere removed a piece of concrete in which no cracks had developed.
Furthermore, if cracks are allowed to develop on the tension side —
and this has frequently been observed in beams under working load —
might not this crack gradually extend under repeated loading and
seriously impair the safety of the structure?
B. H. Davis.* Certain practical considerations may be cited to
illustrate the difficulties confronting the experimenter seeking a
rational solution of the deflection problem. Shrinkage is the worst,
or perhaps the most indeterminate factor to be eliminated, since it
spoils so many carefully performed experiments, being a large cause
of the lack of uniformity so generally noted in experimental data.
2 The shrinkage of a concrete block 8-in. square by 2-ft. lohg has
been shown to shorten appreciably a bar -^v-in. square embedded in it
and accurately measured before and after the setting of the concrete
around it. This produces an initial tension in the concrete and an
initial compression in the steel. In the case of a beam reinforced in
only one plane, as perhaps some of the beams tested may have been,
these initial strains may largely account for the lack of uniformity in
the results obtained.
3 The shrinkage of concrete in setting, nearly always a variable
factor, has almost completely upset the theory of arch-ring deflections
when the arch centering is struck. Some settle very considerably
upon striking the centering, especially when the arch ring is a mono-
lith from skewback to skewback, while others settle hardly at all
' Assistant Engineer, Lackawanna R. R., Hoboken, N. J.
STRESSES IN REINFORCED-CONCRETE BEAMS 541
when alternate voussiors are made and allowed to set and shrink
before the ring is keyed. Shrinkage, it has been proved, almost
entirely causes this lack of agreement between the theoretical and
the actual deflections when arch centers, are struck. It would there-
fore seem logical to assume that the same cause figures prominently
in the deflection phenomena of beams.
4 The shrinkage of a beam of large cross section, acting in oppo-
sition to that of a smaller beam, has been known to crack the weaker
member from top to bottom, breaking up any dependence that might
otherwise have been placed upon the concrete in tension, before the
beam had been called upon even to support its own dead load.
5 In designing for a given load by the commonly accepted straight-
line formulae for obtaining stresses in steel and concrete, and using
the prescribed unit stresses of the building code, a certain factor of
safety results. In other words, an overload of two or three times the
load assumed in the design, may be applied, and when removed, the
structure should be just as capable of supporting the working load
for which it was designed as before the overload was applied.
6 Now, granting the conclusion of the author, in Par. 27, that
tension in the concrete materially affects the deflection and strength
of beams (between certain limits of load), would it not still seem
unwise to take advantage of this tension factor in any design where
the assumed load limits might be overstepped at some time, leaving
the beam to serve the remainder of its period of usefulness without
the tension factor counted upon in its design?
7 Almost every design is over-stressed sooner or later, occasion-
ally by test load, but more often, perhaps, because of the enthusiasm
of some shop foreman in showing what his building will stand in the
way of abuse. For example, loaded cars of gravel and broken stone,
and later a 600-class standard-gage locomotive, were run across a
machine and erecting shop floor that was designed for a uniformly
distributed load of considerably less than one-half the concentrated
moving loads applied, this without any apparent damage to the
floor.
8 Settlement, which very often upsets carefully made calculations,
causes even more indeterminate stresses in reinforced-concrete than
in other types of construction, this being due to the continuity and
the monolithic character of the material. This fact further empha-
sizes the necessity for conservatism in working formulae,
9 Construction joints, put in as they usually are, at points of
maximum moment, make any reliance upon the concrete in tension
542 DISCUSSION
entirely out of the question where such joints occur. It is not
generally conceded that construction joints so located do materially
weaken a beam except in shear.
10 A beam, accidentally cracked entirely through near its middle
while being placed in a testing machine, tested higher than the aver-
age of several other beams of the same size and^'reinforcement, show-
ing that a plane of fracture approximately normal to the center line
of a beam had not, in this particular case, unfavorably affected the
ultimate strength of a beam equally loaded at its third points.
11 Until more is definitely known concerning the shrinkage of
concrete and the many other stresses in reinforced-concrete beams at
present indeterminate, as a matter of conservatism it would seem bet-
ter to disregard tension in concrete as a moment-resisting factor.
Chas. B. Grady/ Professor Lanza and Mr. Smith have clearly
brought out the fact that three of the formulae used for the design
of reinforced-concrete beams are approximate with a load of one-
third the breaking load. The writer will say a few words in refer-
ence to the use of these formulae in the design of beams.
2 Formulae A and B, which are used by a large number of
engineers, do not allow anything for the tension in the concrete and
therefore must give for rectangular beams results which are mere
approximations up to a point at which the concrete fails to act in
tension, but the writer believes that if a comparison had been made
at say double the load used, Formulae A and B would have given
better results, and possibly nearer those found by actual test, than
Formula C, especially for the value of a^ (stress in steel per square
inch).
3 In tests of similar beams made at Cornell University by Messrs.
Paulus, Tripp and Davis, the average variation in the values of a^ (stress
in steel per square inch) deduced by formula A from those found
by experiment was 34 per cent with a load of 4000 lb., and less than one
per cent with a load of 8000 lb. The above figures are for five beams
having an average breaking strength of 13,200 lb.
4 The errors in values deduced by Formulae A and B are more
liable to be on the side of safety than the errors in values deduced
by Formula C, and while there is no doubt that Formula C will
give more accurate results when the stress in the steel is compara-
tively small, it is the opinion of the speaker that Formula C, and
* Asst. Mechanical Enginepr, New York Edison Co.
STRESSES IN REINFORCED-CONCRETE BEAMS 543
other formulae making allowance for the tension in concrete, should
be used with caution.
5 It is the practice of many engineers to design reinforced-
concrete beams in accordance with certain working stresses and to
endeavor so to proportion the beam that it will fail by tension,
that is, by either breaking or stressing the steel to a point consider-
ably past its elastic limit, thus making the factor of safety a function
of the stress in the steel. In such cases, no matter how much the
concrete has helped out the steel under working conditions, when the
beam is overloaded the steel must take care of practice Ily the entire
tension; and therefore the writer believes that it is wiser not to
introduce a value for the tension in the concrete into the formulae
used in the design of reinforced-concrete beams.
6 The speaker beheves that the formulae for deflection deduced
by Professor Lanza and Mr. Smith will be of great value to engineers,
and that any one of the three formulae will give results accurate
enough for practical purposes in figuring the deflections of T-beams,
more of which are used in buildings than rectangular beams.
Frank B. Gilbreth. The most important subject related to rein-
forced-concrete, from the standpoint of the mechanical engineer,
is the design of forms, for it is the forms that afford the greatest
opportunity for the saving of money, and the consequent reduction
of price per cubic foot of new buildings.
2 Beams have been designed and built of rectangular section and
over 64 ft. 0 in. long, and have been perfectly satisfactory. The
most successful building of today as well as of the future must be
designed with regard to the economical design and use of forms, and
not to the greatest saving in the quantity of steel and concrete used . The
forms are the most expensive single item of reinforced-concrete work.
3 It is by no means rare to see designs for saving concrete where
the value of the concrete saved amounts to much less than the cost
of the special or odd-sized forms required.
Prof. Wm. H. Burr. Much has been said about the disagreement of
theoretical results with the results of experiments. That is an obser-
vation which may be made, I beheve, in the case of every mateiial
which has ever been used by the engineer; scarcely more so of con-
crete, either plain or reinforced, than of other material. When a com-
parison of this kind is made, I think we should bear in mind, first,
what theory is used.
544 DISCUSSION
2 The so-called common theory of flexure probably is not strictly
applicable to anyj^reinforced-concrete beam which has been broken.
It is a theory which applies to a beam of very small depth, compared
with the length of span. This is not the kind of beam usually found
either in plain or reinforced-concrete, and usually not even in steel.
It is not a matter of surprise, therefore, that such a theory does not
give the results found by experiment.
3 It seems to me we shall have to proceed with reinforced-concrete
beams precisely as with beams of other material, viz., use a simple work-
ing hypothesis for the purpose of securing a formula in which empiri-
cal quantities may be used. That is the case with wrought-iron and
steel beams, with timber beams, and with all other beams, and it is
markedly so, even to a greater extent, with columns.
4 The three theories. A, B and C, may be considered in view of
the varying conditions at different stages of stress. It would be
difficult to show from any results of tests of concrete, that the law of
distribution of stress in theory B is justified. There are some tests
which show a graphic relation between the intensities of stress and
strain, which approximates a parabolic curve, but probably no nearer
than a circular curve or some other. The majority of tests show that
line much more nearly straight than parabolic within the limits of
stress found in ordinary concrete beams.
5 It is true that concrete has considerable tensile resistance, when
it possesses any, but I think there are few engineers who have used
much plain or reinforced-concrete, who would be willing to trust the
tensile part of the beam to carry load, and to be so recognized in
the working formula.
6 The result of the slight contraction of concrete, possibly not
within the first two months, perhaps not within the first year of its
life, is to create fine hair cracks. We do not know how far. these
enter the mass; they may be only skin-deep, but in some cases they
are much deeper. Hence if the beam should show a continuous con-
crete structure on the tension side for the first two or three months,
it does not follow that it is going to remain so. If we are to recognize
such a possibility, and it seems to me we would not be justified in
neglecting it, the only safe procedure is that usually followed, of
neglecting tension in concrete. That does not mean that concrete
may not sometimes have considerable tensile resistance. It simply
means that such resistance caimot safely be recognized in ordinary
concrete work.
STRESSES IN RBINFORCED-CONCRETE BEAMS 545
7 These cracks may be veiy much reduced by continual wetting
of concrete after it has been put in place. That ic one direction in which
the concrete work may be improved. We do not wet the concrete
nearly enough after the forms are taken away. If it were feasible,
concrete should be kept thoroughly wet from three to six months
after being put in place. This is not practicable; but after the forms
are taken away, the concrete should be kept soaked with water just
as long as possible. The contraction will be less and there will be
fewer hair cracks, but it will be impossible to eliminate them entirely.
8 We should be sensible, as engineers, in connection with rein-
forced-concrete work, precisely as we are or ought to be in everything
else, and use the simplest possible, formula, i. e., the straight-line
formula, and not strain after some ultra-refinement which, when we
come to examine it, has little or no solid basis. We should resort to
proper theories and select a simple working hypothesis, and then use
the test beams to determine such theoretical coefficients or quanti-
ties as will make the resulting formulae represent actual results as
nearly as possible.
Prof. J. C. Ostrup.^ Within a short time, from fifteen to twenty
years, at most, reinforced-concrete has gained an enviable position
in the construction world, and unquestionably, in spite of many
inherent shortcomings, will better its reputation in the future among
both engineers and laymen. It is, therefore, to be regretted that
the trend of the authors' paper is toward a negative rather than a
positive support.
2 It is a well-known fact that the greater number of the deductions
and working formulae obtained from the science of applied mechanics
are based upon certain assumptions which to a greater or usually
less extent circumscribe the use of such formulae. The errors resulting
from these^v/undamental assumptions vary considerably with the
different engineering materials with which we deal; they often
vary considerably even with the same material, changing somewhat
with the extreme fibre stress, the manner of application of the load,
etc. The assumptions made in regard to the behavior of structural
steel are probably nearer the absolute truth than for any other engi-
neering materials, so near, in fact, that many engineers have come
to regard the theory of steel design as following an unassailable
mechanical law. Nevertheless this is not so.
' Profeaaor Structural Engineering, Stevens Institute of Technology.
546 DISCUSSION
3 On the other hand, the theory of reinforced concrete is based
upon many assumptions, some of which can be better defended than
others, and some of which have undergone, and will continue to
undergo, modifications from time to time. It is also based upon
many widely varying experiments which the experimenters them-
selves have been struggling to reconcile. Some of the most impor-
tant of these assumptions, together with a brief account of their
probable effect, are:
a That the applied forces in bending are perpendicular to
the neutral axis.
4 This is incorrect, of course, inasmuch as the neutral axis under
deflection follows a curve resembling a parabola. The resulting error
is, however, extremely small.
b That a sectional plane, true before bending, also remains
true after.
c That each fibre acts independently of adjacent fibres.
5 The last of these assumptions is particularly faulty, inasmuch
as the ordinary reinforced beam usually has its reinforcement vary-
ing in amount, both horizontally and vertically, throughout its length.
In other words, unlike a rolled-steel beam whose moment of resistance
is uniform from end to end, the reinforced beam is not uniform in
strength, the stronger parts tending to assist or restrain the weaker.
The error from this assumption cannot be evaluated.
d That the concrete and the reinforcement will stretch or
compress together without breaking the contact bond
between them.
6 This condition, when complied with, as it infallibly must be in
all cases, unquestionably sets up secondary local stresses, the magni-
tude of which cannot be even guessed.
e That there are no initial stresses.
/ That the stress-strain curve for compression is a parabola.
7 The fulfillment, or the non-fulfillment, of the last two assump-
tions, is probably what causes the greatest divergence between theory
and tests. A concrete beam is a casting, in a sense. If the mixture
were perfectly uniform throughout, there would most probably not
be any initial stresses due to the chemical action of setting. This is
evidently not possible ; hence throughout the beam there undoubtedly
exist initial stresses of uncertain magnitude. This fact, in itself,
STRESSES IN REINFORCKD-CONCRETE BEAMS 547
would surely affect the stress-strain curve, but in addition we must
consider the variable modulus of elasticity for the concrete. This
varies not only in the same beam, according to unit stress in the
extreme fibres, but also in beams of the same identical composition
according to its depth, i. e., to the relation between the extreme fibre
stress and the average fibre stress.
8 In addition to these mechanical considerations, we have many
physical considerations governing the strength of concrete and rein-
forced-concrete beams. Such physical conditions must largely depend
upon the personal equation of the engineer in charge; they may
be guarded against, and their effect minimized but not wholly
eradicated. When present, their influence can only be surmised.
9 To make this a little clearer, let us assume a case where a num-
ber of beams were to be prepared for a testing machine and where
great uniformity naturally would be sought; to insure which, only
one grade of cement, one of sand and one of broken stone, would be
employed. Next let us look into some of the more important points
affecting the strength of concrete, as follows:
a Condition of the cement; whether all the bags in a cargo
are of the same age, or manufacturing batch; quantity
of carbonic acid contained; degree of moisture (since the
outside bags in a stack, and even the outside layer in the
same bag, often absorb considerably more moisture than
the inside).
b Uniformity of quality of the sand; whether or not it con-
tains in spots, loam, clay or other impurities, etc.
c Uniformity of the broken stone; whether or not the stones
are alike in strength and texture; whether or not they are
broken to a uniform size, etc.
d Quality or purity of the water; method of mixing the con-
crete, or difference in methods of mixing from batch to
batch, even by the same gang.
e Tamping and placing of the concrete, including the often
unavoidable variations in the degree of flexibility of the
support between the ends and the center of the beam
while the concrete is being tamped.
/ Workmanship. A man is not a machine, consequently the
materials mixed and the beams made, even by the same
gang, will often vary considerably in spite of precautions.
May not ])eams vary much more when made by different
sets of workmen?
548 DISCUSSION
10 Besides the foregoing points affecting the mechanical laws
governing the strength of the concrete, there are others; but enough
have been indicated here to show that, when tested, a variation in
their strength must exist.
11 Since each experimenter must base his deductions upon the
results of his own observations, a divergence in theres ulting formulae
is the natural result, and furthermore, were he to repeat the same
tests under similar circumstances, his second results, in view of the
foregoing, would vary from his first. With all this in mind, is it any
wonder that closer agreement between the various working formulae
most generally in use, has not so far been reached? To an unbiased
mind the wonder is that the divergences are not even greater.
12 Returning to the conclusions of the author, he states in Par.
24 "... . the observations made thus far are not suffi-
cient to furnish the means for determining the actual distribution of
the stresses, and hence for the deduction of reliable formulae
etc. " This may be strictly true in theory, but will hardly be gener-
ally accepted as a matter of fact. On the contrary, it is quite within
good reason and good practice to deduce reliable formulae, even where
the action of some of the minor points involved is in doubt, so long
as the effective range of such points is known. In this connection it
may be recalled that concrete and masonry structures, centuries old,
are still standing and doing effective service, though they were designed
from formulae and data far less reliable than those now at our dis-
posal.
13 The author further says: "It follows therefore that which-
ever of the theories is adopted for practical use, it can be regarded
only as a sort of working hypothesis." This, of course, is a sweeping
condemnatory statement which, if it can be aj)plied to the theory of
reinforced-concrete construction, can, it is believed, be equally well
applied to the theories underlying any form of construction; for no
amount of theory, unaccompanied by practical experience and sound
judgment, will prevail, either in the mechanical or in any other engi-
neering field. This fact cannot be too strongly emphasized.
14 In Par. 26 the author states that theory C gives results in
closer agreement with experiments than does either A or B. This is
undoubtedly true, but so far as the evidence in Tables 5 and 0 is con-
cerned , any one of the three theories is based upon " reliable formulae "
or, what is more to the point, the designs resulting from their use
would be wholly reliable. As a matter of opinion, the preference
gjipuld be for A or B, since they are nearly correct in regard to the
STRESSES IN REINFORCED-CONCRETE BEAMS 549
unit stresses in the concrete, — the weaker material, — whereas they
give somewhat smaller stresses for the steel than those expected.
lo It is equally true that no"- reliable deflection formulae can be
deduced without taking into consideration the tension in the concrete.
We can, however, go a step further, and state that such formulae, to
be correct, must also include a provision for a deflection increment
due to shear.
16 In concluding these ,emarks, the writer would suggest a cau-
tion to such alarmists as are prone to appear from time to time
against a useful and excellent building material. No public good
can result from arousing the apprehension of either engineer or lay-
man with respect to reinforced concrete, and those of us who have
had the opportunity of using it for a number of years cannot help but
be impressed with its increasing serviceability and scope.
E. Lee Heidenreich.^ The tests at the Massachusetts Institute
of Technology, as well as those'^at the University of Illinois, were
based upon a concrete mixture of 1 : 3 : 6, while those of Considere
are based upon a mixture of 1 : 2^ : 2^. I have repeatedly at meet-
ings of the "Joint Committee" urged the desirability of employing
stronger mixtures, and mixtures of a " maximum density " rather
than a certain proportion; and I believe that with such stronger mix-
tures Formula C will come still nearer to a correct interpretation of
stresses and strains. If so, is it not natural to hope that in our rein-
forced-concrete building constructions, lesser dimensions of beams
and girders, thinner floor slabs, and consequently a reduced item of
dead load will result, also materially reducing the present disadvan-
tages of heavy columns and foundations?
2 The most wonderful constructions of i^ tanks, reservoirs and
bridges in Europe have resulted from mixtures of 1 : 3 or 1 : 5, prop-
erly graded. Why should not our beam tests be based upon such
mixtures, notwithstanding the fact that at first glance they may not
appear commercially advantageous for building constructions? I
wish to place myself again on record as advocating a larger percentage
of cement and a mixture representing a maximum density of the
ingredients.
Prof. C. E. Houghton. The paper adds to our knowledge of the
probable magnitude and sign of the errors due to the use of formulae
' Special Engineer, N. Y. C' & H. R. R. R., New York City.
550 DISCUSSION
deduced from a simpler theory. When the size of a structural mem-
ber has been calculated by the use of a formula known to give a greater
value to the unit stress than actually exists, the designer need not
worry about the safety of that member. If in addition the probable
magnitude of the error is known, corrections may easily be made
where it is considered necessary to reduce the cost or weight of the
member.
2 The neglect of tensile resistance in calculations of the strength of
reinforced-concrete beams finds a parallel in the common practice for
the calculation of the strength of riveted joints. The friction between
the plates unquestionably adds to the strength of the joint, yet as far
as the writer knows, no theory has been accepted in American practice
that considers this friction as acting. This friction, like the tensile
resistance of concrete, may vary from zero to a maximum value,
and therefore should be neglected, as neither can be depended on for
additional strength.
3 All formulae^for the strength of reinforced-concrete beams con-
tain a factor whose value is the ratio of the modulus of elasticity of steel
to that of concrete, and any error made in the assumption of that value
affects the result in the same proportion. The modulus of elasticity
of steel is practically a constant term, but that for concrete varies
through a wide range of values depending to a certain extent on the
proportions of cement, sand and broken stone used in the concrete.
4 With the large possible variation of this ratio in mind, it would
seem reasonable to suppose that the probable error, either in assum-
ing a straight-line law for the variation of the compressive stress in
the concrete, or in the neglect of its tensile resistance, will be less than
that due to the choice of the value of this ratio. What is needed is a
value for this ratio, determined by applying the formula derived
from the straight-line no-tension theory, to the results of a great many
tests on specially prepared beams.
5 The number of variable conditions that would affect the results
in any such investigation is so great that unless one of our national
engineering societies will undertake it there seems to be but little
prospect of obtaining anything more than an approximate value
based on the results of compressive tests on concrete.
Wm. Wallace Christie. The writer is particularly interested
in the applications of reinforced-concrete in engineering work, and
has had to do with the designing of a great many floors, foundations
STRESSES FN REINFORCED-CONCRETE BEAMS 551
and other work. He agrees with Professor Burr, and others not
prepared to accept or consider a theory of design of concrete-steel
l)eams allowing tension in the concrete, or an increase by reinforce-
ment of the ability of the concrete to resist tension.
2 After concrete work has been erected for a time, hair-cracks,
and others more decided, often develop in the beams. An example
of this has already been cited: a 70-ft. or longer concrete girder,
with its center, at least, resting on hard pine timbers.
3 With the large factor of safety necessary in the design of con-
crete-steel beams, one cannot go very far wrong in using any of the
three methods mentioned, but the writer prefers a straight-line for-
mula.
4 The paper deals in particular with beams, which in practice
are seldom used, except as lintels, or over openings in building walls.
The experiments conducted with these beams will not give the results
obtainable by the use of T-beams, and the writer doubts whether
the test of a single T-beam, made in the test room, will develop the
same strength or other features, as a test made on a similar T-beam
which is part of a floor system. The beam tested in the laboratory
is not joined tightly with the rest of the floor, while in actual con-
struction the iron would necessarily be secured to the other parts
of the floor system.
The Authors. The data and the results of observation for the
first five beams, which have been asked for, are contained in a paper
by Gaetano Lanza, published in the proceedings of the American
Society for Testing Materials for 1906.
2 The modulus of elasticity of the concrete was obtained from
tests made upon seven 8 in. by 8 in. by 60 in. plain compression pieces
of the same age, materials and mixture as the beams. The values of
E are as follows:
2,479,000
2,223,000
2,367,000
2,264,000
2,670,000
2,623,000
2,341,000
Average 2.424.000
552 DISCUSSION
In our paper we have used 2,335,000 in order to permit of the use of
r = 12.
3 It may be added that the neutral axis was determined for each
load from the strain diagrams (which are shown graphically in the
paper referred to) at the intersection of the plotted line with the
vertical datum line. Numerical details of the strains will be given
in Appendix No. 1, as they seem to be desired.
4 As reference has been made to evidence tending to discredit
Considere's'^ theory" regarding the ability of concrete to stretch when
reinforced, it may be well to say that it is neither the object of the
paper to discuss this question, nor toHake sides for or against
this theory. The history of the^main part of the controversy is as
follows:
5 The theory was attacked by Kleinlogel in an article published
in Beton u. Eisen, Hefte 2 and 4, 1904, in the light of certain tests
which he had made. The two tests of Considere (Par. S) were made
as a refutation of Kleinlogel's argument. An account of them
may be" found in^ Considere's book on reinforced concrete. A sub-
sequent'^reply^ by Kleinlogel, and a reply to this by Considere, are to
be found in Beton u. Eisen, but no new matter is given.
6 In Beton u. Eisen, Hefte 11-1905, Professor Ostenfeld gives an
account of the results of some computations made by him upon the
beams tested by Kleinlogel, and in the light of these he says, "Thus
far I regard Kleinlogel's tests as a beautiful though unwilling confir-
mation of Considere's theory." To this Kleinlogel repUes in Beton
u. Eisen, Hefte 1-1906, but this reply contains no new evidence.
7 Fear seems to be expressed by some that pointing out the very
considerable discrepancies between the results of computation made
by theory A, and the results obtained by experiment, is equivalent
to a condemnation of all structures where theory A was used in the
computations. No such condemnation, however, is intended by the
authors. They believe, however, that the more we realize the facts
in any case, the better prepared are we to use our judgment as engi-
neers, in designing any construction.
8 Most of the arguments advanced in support of the entire suffi-
ciency of theory A may be summarized as follows:
a The calculations can be more easily made.
b That the mere fact of neglecting the tension in the concrete
results in safetj'', though practically all admit that the
concrete does resist tension in the early stages.
STRESSES IN REINFORCED-CONCRETE BEAMS 553
c The use of construction joints, which often take the form
of a vertical joint at the middle of the span when work
on a given floor extends over a period greater than one day.
9 These matters will be considered in the same order:
a There is no doubt that the calculations are more easily made
when theory A is used.
h Whichever of the three theories is used, it is not customary
to calculate by means of it, the sti esses which produce
diagonal cracks, and it is a fact that in a very large percen-
tage of the beams that have been tested, the failure has
been due to these diagonal cracks. Hence it seems to us
that until we have ai rived at some means of making calcula-
tions to deterailne these stresses and strains in such a way
that the calculated resuLs shall have a fair degree of agree-
ment with the results obtained by experiment, we can hardly
claim to have an all-sufficient theory. Moreover, in the
case of beam A-1, the only one for which the shear has
been figured, it is greater when determined from theory
Cthan when obtained from theory A, the difference being
in one case 57 per cent.
c When a construction joint is introduced, the beam is neces-
sarily weak, and until tests of such beams are made, we
cannot claim to know what theory will apply to them.
10 Other considerations which it would seem worth while to dis-
cuss are the following:
a The presence of initial stresses due to shrinkage.
b The variation in the value of the compressive modulus of
elasticity of concrete,
c The recommendation made by some that the formulae to
be used be based upon loads larger than one-third the
breaking load, and by some upon the breaking load.
d The question of so proportioning the reinforcement that the
breaking shall be due to the tension in the steel exceeding
the elastic limit.
11 Discussing these in order we have:
a The presence of initial stress is of course a great source of
uncertainty in reinforced-concrete, as well as in cast iron,
and hence we should expect irregularities due to this
cause, the amounts of which are veiy difficult to estimate.
Whether their influence is still large or not at one-third
554 DISCUSSION
the breaking load, is a debatable question, though it must
be comparatively less at one-third than at smaller loads.
On the other, hand with loads greater than one-third the
ultimate, the ratio of stress to strain becomes quite vari-
able, and any rational formula becomes inaccurate.
b In the light of the experiments made by different men and
in different places, it would seem to the authors that the
variations of the modulus of elasticity for compressive
stresses in the concrete, not more than one-third the
ultimate, would not be very excessive.
c In the case of steel or other beams it is well known that the
ordinary formulae do not apply when the stresses in any
of the fibres have passed the elastic hmit; hence the
difference between modulus of rupture and outside fibre
stress at breaking.
d Regarding the question whether theory A will agree better
with experiment when the percentage of reinforcement is
kept so low that the elastic limit in the steel will be exceeded
before any fibre of the concrete has to bear a stress equal
to its crushing strength, the only evidence in the paper
is the following: in one case the percentage of reinforce-
ment was as low as 0.99 per cent, and in three others,
1.25 per cent, and in these three cases the discrepancies
of theory A are large.
12 In general, it seems to us that thus far not enough systematic
work has been done by way of experimenting and calculating in order
thai we may have more accurate knowledge about a number of
matters, among which may be mentioned:
a The actual distribution of stresses not merely in the case
of longitudinal reinforcement, but also with diagonal
and other reinforcements, and also in T-beams.
b A study of the diagonal tension, not only at the ne'utral axis,
but elsewhere.
c A study of the conditions necessary that the breakage may
always be due to the reinforcement exceeding the elastic
limit, and whether diagonal cracks occur in those cases.
d A study of the effect of construction joints.
13 There only remain for discussion a few additional matters
raised by different gentlemen. While it appears from the last
STRESSES IN REINFORCED-CONCRETE BEAMS
555
column of Mr. ^^'ol•cestel•'s table that method C gives average results
on the negative side, it must be remembered that they depend upon
the value taken for t (the tensile strength of the mixture). This
table, as well as Table 5, clearly shows that if a slightly lower value of
t had been used for these six beams, their average error would have
been a positive one, and also smaller than that by using A.
APPENDIX NO. 1
STRAINS FOR THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY BEAMS
The strains were measured at four points in the depth of the beam
on each side as described in the paper before the American Society
for Testing Materials, already referred to. Columns 1, 3, 4, 2 in the
following tables give the strains for these points. Points 3 and ]
were one and five inches, respectively, above the center of the beam,
while points 4 and 2 were one and five inches, respectively, helow
the center.
BEAM A-1
One 1-in. Plain Rod
Initial Load 1250 Lb.
Age 53 Days
Breaking Load 15000 Lb.
556
DISCUSSION
BEAM A-2
One 1-in. Twisted Rod
Age 49 Days
Initial Load
1250 Lb.
Breaking Load 16500
Lb.
Loads
'
Strains. 1st
application
Lb.
1
3
4
2
2250
0.000044
0.000012
0.000003
0.000033
3250
0.000082
0.000012
0.000027
0.000093
4250
0.000138
-0.000013
0.000077
0.000174
5250
0.000172
0.000016
0.000073
0.000251
6250
0.000216
0.000018
0.000108
0.000358
8250
0.000317
-0.000004
0.000202
0.000595
10250
0.000405
-0.000009
0.000271
0.000835
12250
0.000505
-0.000063
0.000391
0.001039
BEAM B-3
Two J-iN. Plain Rods
Initial Load 1250 Lb.
Age 43 Dats
Breaking Load 15950 Lb.
Loada
Lb.
2250
4500
5250
6250
8250
10250
Strains. 1st application
1
3
4
2
0.000073
0.000013
0.000017
0.000081
0.000100
-0.000003
0.000059
0.000175
0.000144
0.000015
0.000060
0.000223
! 0.000195
0.000002
0.000096
0.000289
0.000398
-0.000020
0.000182
0.000428
0.000519
-0.000066
0.000301
0.000587
BEAM C-6
FoDR S-iN. Plain Rods
Initial Load 600 Lb.
Age 35 Days
Breaking Load 16240 Lb.
Loads
Strains
Lb.
1
3
4
2
2600
0.000083
0.000018
0.000026
0.000087
4600
0.000219
-0.000024
0.000133
0.000296
6600
0.000337
-0.000067
0.000239
0.000532
8600
0.000444
-0.000059
0.000297
0.000751
10600
0.000542
-0.000091
0.000406
0.001023
12600
0.000631
-0.000137
0.000525
0.001272
14600
0.000765
-0.000209
0.000653
0.001525
STRESSES IN REINFORCED-CONCRETE BEAMS
557
BEAM E.9
Two J-iN. Twisted Rods
Initial Load 1250 Lb.
AOE 54 Datb
Breaking Load 21000 Lb.
Load
Strains. Ist application
Lb.
1
3
-0.000012
4
2
2250
0.000037
0.000029
0.000037
4250
0.000107
0.000003
0.000046
0.000134
5250
0.000155
0.000008
0.000060
0.000175
6250
0.000202
0.000004
0.000081
0.000256
8250
0.000275
0.000004
0.000122
0.000402
10250
0.000403
0.000010
0.000161
0.000541
12250
0.000486
0.000003
0.000212
0.000680
No. 1253
THE DESIGN OF CURVED MACHINE MEMBERS
UNDER ECCENTRIC LOAD
By Prof. Walter Rautenstratjch, New York
Member of the Society .
Machine members, such as frames for punches, shears and riveters,
hooks and the like, when subjected to load are generally supposed to
behave like beams originally straight and subjected to the same con-
ditions. The usual analysis applied to such beams in determining
the proportions required to withstand safely a given stress assumes
that the maximum tensile stress at a in Fig. 1 = load considered as
uniformly distributed over the section + the stress due to the eccen-
tricity of the load. Symbolically expressed
W Wle
where
f^ = maximum intensity of tensile stress,
W = load on beam.
A = area of section.
I = eccentricity of loading.
e = distance from gravity axis of section to point under stress /j.
I = moment of inertia.
2 This analysis is unfortunately prevalent in textbooks on the
design of machine elements and strength of materials, and has been
accepted generally because of long standing. However, it does not
agree with the results of experiment on members of this kind ; in fact
such experimental results are so different from results calculated by
this formula that no confidence whatever can be placed in it and safe
proportions can be obtained only by the use of a large factor of safety.
Presented at the New York monthly meeting (November 1909) of The
American Society of Mechanical Engineers.
560
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD
3 The writer has recently pubHshed* the results of a series of
experiments which are remarkable in their disagreement with the
results obtained by the formula. The crane hook was taken as an
example of a beam of this sort and experiments were conducted on
ten hooks ranging from 2 to 30 tons rated capacity. All hooks were
furnished by the manufacturers. In Table 1 the results of the ex-
periments are compared with the results by formula.
Machine-Tool Frame Considered as a Loaded Bea«
4 It is very evident that the assumptions on which the above for-
mula is based are not correct, and that machine members designed
on this basis' have, a; much^smaller^f actor of safety than is generally
supposed. While this has been known in some quarters and attempts
have been made to bring about an adjustment, no theory which has
been developed seems to fit the case better than that evolved by
TABLE 1 COMPARISON OF RESULTS
Description op Hook
Load at Elastic
Limit bt Test
Pounds
Load at Elastic
Limit bt Standard
FORMDLA
Pounds
30-ton cast steel
56,000
115,000
20-ton cast steel
30,000
70,000
16-ton 'cast steel
48,000
145,000
15-ton wrought iron
16,000
73,000
10-ton cast steel
18,000
43,000
10-ton wrought iron
16,000
26.0CC
5-ton cast steel
18,000
62,301
5-ton wrought iron
14,000
20,800
.3-ton cast steel
8,500
14,900
2-ton cast steel
4,700
14,Q0G
*An Investigation of the Strength of Crane Hooks, American Machinist,
October 7, 1909.
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD 5C1
E. S. Andrews and Prof. Karl Pearson of London University. This
analysis is deserving of much more attention than it has received,
and it is surprising that even some who have had access to it have
made the statement tjiat the old theory was sufficiently accurate for
the usual case of design. It will not be my purpose to give a complete
derivation of the new formula, which has been published elsewhere^
in complete detail, but rather to show how the results of the analysis
may be made directly applicable to design.
5 The investigation referred to gives the following expression for
the tensile stress at the most strained point in the principal section of
beam :
where
- r. I + 1
J
fi = tensile stress at most strained point of section, pounds
per square inch.
W = load on hook, pounds.
A = area of section, square inches.
I = distance from load line to gravity axis of section.
p = radius of curvature of belly of hook at gravity axis.
e = distance from gravity axis to point of maximum tensile
stress.
dA
;-, = a function whose value is equal to I — ^ in which
14-'
A
dA is any differential area of section, a distance y from
the gravity axis, 2" denoting the sum of all the operations
indicated by the symbols.
J', = a function whose value is equal to /-^ -
dA
1 + ^
in which y is any distance from b equal to ^ — - at which an
A
at whi
ordinate such as de is erected.
'Technical Series 1, The Draper Company's Research Memoirs, 1904.
562
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD
6 The functions ;-, and j-^ are to be found for any section, as shown
in Fig. 2. The half-area of the hook is akhlc. All possible ordi-
nates of the nature of de determine the curve aehje. Clearly then
area aehjc
' ^ ^ area of section
Likewise, if at any distance y from b an ordinate is erected whose
Fig. 2 Cross-Section of a Hook
length is
dg
(-f)'
= df, all possible ordinates of this nature form
^, ,, , , area aehjc — area afhic
the curve afhic, and 7-j — ; — - —
^ area of section
7 The above analysis was applied to each of the hooks tested,
with the results recorded in Line 3 of Table 2. An inspection of
TABLE 2 ANALYSIS OF HOOKS TESTED
Load at Eiastic Limit, Poonds
RATED CAPACnr
By test
By standard formula
By new formula. .
3a<roN
O
20-TON
IS^roN
2A
^
10-TON
12
1^
66,000 30,000 48,000| 16,000 18,000 16,000
115,000! 70,00flj 145,000 73,000| 43,000 26,000
55,080| 29,925 50.57o| 16,000 16,600 15,000
5-TON
o J'
iS.OOOj 14,000
52,3001 20,800
18,9501 14,100
3-TON
8,500
14,900
8,600
2-TOM
I '4,700
14,900
4.400
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD 563
this table will show how nearly the analysis of Mr. Andrews and Profes-
sor Pearson fits the case and how far from correct are the results
from the old formula. The new formula appears then to be based on a
correct theory and to be perfectly safe for use in the design of all
machine members of this general type.
8 In its present form it is a rather unwieldy instrument in the
hands of a designer, but it may be made more applicable to design
than might be thought at first. Upon examination it will be seen that
the functions y^ and 7-, are constants for all sections of similar form,
that is, for all sections the proportions of which may be expressed as a
function of some unit of dimension, for example, the radius of curva-
FiG. 3 Stakdard Hook Section
ture. Under the same circumstances the entire expression within the
braces is a constant. The equation for a series of si?es and sections
W W
may therefore be written fi = ^K, or A = ~rK. The area is a
/<
function of the unit squared and therefore we may write A = CV, or
-#i -"41
Applying this to the case of a series of hooks ranging from the mini-
mum to the maximum to be manufactured, a standard form of section
may be laid out as in Fig. 3, and the constant established. For the
hooks tested by the writer the following values for the constant were
found :
564 CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD
30-ton hook, cast steel 3 . 00
20-ton hook, cast steel 3.10
15-ton hook, cast steel 3 . 23
15-ton hook, wrought iron 4 . 29
10-ton hook, cast steel 3 . 49
10-ton hook, wrought iron 3 .42
5-ton hook, cast steel 3.12
5-ton hook, wrought iron 3.12
3-ton hook, cast steel 3 . 78
2-ton hook , cast steel 3 . 74
Average 3 . 43
9 To make the case representative of present practice let such
ratio of proportions be assigned to the section shown in Fig. 3 that
C = 3.4. The design of a series of wrought-iron hooks to sustain
loads of from 2 to 40 tons with a limiting intensity of tensile stress of
30,000 lb. per sq. in. will require the following computations:
, , '80000
40-ton hook, r = 3.4 .. = 5.54
^' 30000
r
, r=3.4^^
c, r = 3.4 J
z, r = 3.4 ^1
i
, 60000
30-ton hook, r = 3.4 .^ — = 4.7
30000
, , , 40000
20-ton hook, r = 3.4 ^ = 3.94
30000
, , ,"20000
10-ton hook, r = 3.4 ^ = 2.76
^' 30000
, , ,10000
5-ton hook, r = 3.4 ^ = 1.95
30000
, 4000
2-ton hook, r = 3 . 4 - = 1 . 23
^' 30000
10 The proportions obtained above will be for loads giving a maxi-
mum stress at the elastic limit of the material. For cast steel differ-
ent values will necessarily be obtained. The establishment of such a
standard would lead to a very simple process for the determination
of the principal section of a hook for any capacity; the proportions of
the shank and other parts of the hook may readily be established on
the same basis. The bottom of the hook, being subjected to much
wear, cannot of course be proportioned on the basis of the stress analy-
sis. The above standard section selected as an average representa-
tive of present practice is not, however, the most economic form of
section from the standpoint of equal maximum tensile and compres-
sive stresses. It has been pointed out by Professor Pearson that a
i
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD
565
section with such proportions is approximately an isosceles triangle
with a radius of curvature of 1.75 of the height. The more nearly this
form could be approached the less would be the weight of hook for the
same capacity.
11 Professor Goodman* points out that for hook sections the
functions y^ and y^ ^re expressed approximately as follows:
r2 =
ke
i.¥^
r, = 1 + i.ir2
where k = radius of gyration of the sections, the other symbols being
as before noted.
Centre of radius
at Curvature
Fig. 4 Design for a Punch Frame
A =212 7 = 14,533 *=' = 68.56 i = 38.56 p = 16 e = 8.66
12 It will be found much more convenient to use these empirical
expressions, which give quite accurate results, than to determine the
value of the functions by the more tedious graphic method.
13 In applying these empirical formulae to punch and riveter
frame sections the writer has found that the results are not accurate
but that the values are better expressed as follows :
ke
^' 0.7^
n = 1 + i.ir2
For example, consider the design for a punch frame shown in Fig. 4.
'Institution of Civil Engineers, Proceedings, vol. 167.
566 DISCUSSION
Computing the values for the functions ;-, and y^ by the graphic
method, y^ = 1.4, j^ = 0.405. Whereupon the intensity of stress
according to the new method of analysis for a force of 90,000 lb. at
the punch will be
W
It ^ A
I
1
e\ f ' 1 j _,. 1 ^ = 8500 lb. per sq. in.
The values of y^ and ^2 by the empirical formula are 1.44 and 0.4
respectively. Whereupon the intensity of stress becomes /^ = 8500
approximately. According to the old formula used almost exclu-
W Wle
sively in textbooks, the value of fi is expressed by 4. , whence
A I
U = 2450.
14 The above empirical formulae are derived from the results of
computation of two sections. I am not prepared to state that they
will work out in all cases and must therefore caution anyone against
using these values to check the results by the graphic method. It
may be clearly seen that were the punch in question designed for a
limiting intensity of stress of 2450 by the old formula, there would
actually be a maximum stress of 8500 lb. per sq. in., which is hardly a
safe value for cast iron and particularly for a large casting.
DISCUSSION
Prof. Gaetano Lanza. A careful perusal of the articles of Messrs.
Pearson and Andrews in The Draper Company's Research Memoirs,
containing the formulae referred to by the author, reveals no flaw in
the deduction of the formula for the greatest tensile stress at the sec-
tion of greatest bending moment, provided it is regarded as a formula
which gives the relation between the load on the hook and the tensile
stress mentioned, and provided the section of greatest bending
moment remains plane.
2 To determine in all cases, however, the relation between the
load corresponding to a greatest stress at the above stated section, equal
to the tensile elastic limit, and the elastic limit as determined by the
methods of measurement employed, would, in my opinion, require a
set of tests upon a series of hooks varying in their proportions to a
much greater extent than those mentioned by Professor Rauten-
strauch. in which the formula of Professor Pearson would make the
CURVED MACHINB MEMBERS UNDER ECCENTRIC LOAD 567
two loads cited nearly equal. An example of such a case, in which
this result does not hold, is a set of hooks tested under the direc-
tion of Prof. C. E. Fuller, which were really open links of circular form,
made by bending hot and annealing square bars, the side of the square
being 0.75 in., where jOq = 3 in., and where the load attheelastic limit,
as determined by a method similar to that employed by Professor
Rautenstrauch, was 1100 lb.
3 For these hooks we should have ^-j = 1.0074 and 7-3 = 0.00658.
4 The greatest tensile fibre stress at the section of greatest bend-
ing moment, if computed by the ordinary formula, would be 48,600
lb. per sq. in. and, if computed by the theory of Messrs. Andrews and
Pearson, would be 59,300 lb. per sq. in., whereas the tensile elastic
limit of the material was 30,000 lb. per sq. in.
5 In seeking an explanation of these apparently discordant facts
the following observations should be kept in mind:
a In a straight beam we would naturally expect the elastic
Hmit as determined by measuring deflections to be
greater than that corresponding to a greatest fibre stress
equal to the tensile elastic limit, the excess varying with
the span.
b The methods used in all the experiments cited have been
practically the measurement of deflections.
c The deflections, whether of beam or hook, cannot be deter-
mined by computation from the stresses at the section of
greatest bending moment only, but depend also upon the
stresses at the other sections.
d In the hooks tested by Professor Rautenstrauch the sec-
tion of the hook is a varying one in which the stresses at
sections other than that of greatest bending moment have
not been examined.
Hence it seems to me that before we can consider that a complete
solution of this problem has been attained, we need
a A more extended series of tests which shall include a con-
siderable number of hooks of each kind.
6 An experimental determination, both for beams and hooks,
of the relations between the elastic limit as determined
by deflections, and the load corresponding to greatest
fibre stress equal to the tensile, or compressive, elastic
limits, in the case of varying spans and other proportions.
568 DISCUSSION
Chas. R. Gabriel. The results of tests of crane hooks and the
figures obtained by the old and new formulae, to which Professor
Rautenstrauch calls attention, are very important as regards crane
hooks and similar members of machines. If such members are not
as strong as computed by the usual formula for combined bending
and tension it is none too soon for engineers to be made acquainted
with the fact. This is especially so because of the fact that metal
beams of solid cross section, similar to the cross section of a crane
hook when subjected to simple bending, show greater strength than
that due to computation, at least when subjected to a breaking test.
This excess of strength is so great, especially in beams of cast iron of
certain cross sections, as to justify confidence in lesser dimensions for
straight beam members than those that would be prescribed by calcu-
lation based on the tension and compression moments of beam sec-
tions. Similar excess over calculation, of ultimate breaking resist-
ance by test, exists in shafts subjected to torsion.
2 One would naturally expect to find a similar excess of strength
in crane hooks when put to test, but the results to which consideration
is invited show quite the reverse to be the case, and are none the less
valuable because disappointing.
3 As regards machine members, such as the overhung frames of
presses, punching and shearing machines, etc., the large majority of
such frames require to be rigid under their working loads, to an
extent that renders them perfectly safe from failure by breaking. A
great many points have to be considered with respect to dies being
thrown out of line by the springing apart of the upper and lower arms
of the frames. A small amount of such deflection would in some
cases be sufficient to cause the shearing of expensive punches by the
dies, rendering them unfit for the accurate work intended. In some
few other cases, such as riveting, a comparatively large amount of
deflection is permissible, and in some instances the proportions of a
frame may be considered with respect to safety from rupture alone.
4 The cross sections of overhung frames must of necessity differ
a great deal in different machines, also the relative amount of over-
hang or throat, depth of gap and general form of frame, whether
curved similar to a crane hook, or extending straight up and down
comparatively short or long distances. Various kinds of cross sec-
tion such as solid rectangular, T, H, box or combination of box and
rib, all have their appropriate uses. The successful designer has at
times to depart considerably from formulae that have been in use and
must combine much practical judgment and observation in his work.
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD 569
Factors of safety must vary from 3 to 50 or more, and stresses accord-
ingly.
5 It is hardly to be expected that a formula for strength of crane
hooks can be immediately applicable to all the various cases of over-
hung machine frames, but we judge it might be apphcable to small
frames of sohd section and short overhang. Frames having a long
overhang, such as that represented by Fig. 1 in the paper, would in
our opinion be a more trustworthy subject for the application of the
useful bending formula than frames having relatively a much shorter
overhang, such as that indicated by the dimensions in Fig. 4. This is
because the greater the overhang the more significant becomes the
simple bending moment and the less significant the direct tension in
the back of the frame.
6 Referring to Fig. 4, it is noticeable that the metal in the back
of the frame is very thin. In frames where rigidity is the prime con-
sideration, we believe it is a common error of designers when using
cast iron to place too little material in the back. This no doubt
arises from the known high compressive resistance of cast iron, with-
out regard to its elasticity under compression; frames being designed,
accordingly, with regard to resistance to breaking rather than with
regard to resistance to deflection. We have known of many cases
where frames could be greatly stiffened merely by taking metal from
the front web and putting it on the back web.
George R. Henderson. That we get a rather greater strength
than would be expected from the Unwin formula, especially in the case
of hooks, agrees with my practical experience. A few years ago we
purchased some 60-ton cranes, and when it came to the detail of the
hook to lift the 60 tons, the design submitted by the manufacturers
was for a hook smaller than we thought good practice would
accept. We calculated to reduce the total strain due to the vertical
stress and the bending moment to about 12,000 lb., which we con-
sidered would give a factor of safety of five with the material used.
It was pointed out that the hook did not conform to the specifica-
tions, and that a larger hook was desired. These larger hooks were
provided and they looked gigantic.
2 A little later the question came up again, when the manufac-
turers stood on their dignity and claimed that the hook was stronger
than my calculations showed, and to confirm their case referred to
tests at the Watertown Arsenal, which we all consider pretty good
authority. The hook tested was rated as a 20-ton hook, but it had
570 DISCUSSION
been subjected to a weight of 162,000 lb., at which it merely bent
but did not break.
3 These tests were to determine the ultimate strength, whereas
the paper deals with the elastic limit; but practically, I think, the
ultimate strength interests us as much as the elastic limit. By the
regular Unwin formula, which has been somewhat condemned, the
stress per sq. in. in the hook, when weighted to 162,000 lb., at
which it simply opened, would indicate 142,000' lb. per sq. in. fibre
stress, which of course is absurd. So, from the actual tests, it is
very evident that the hooks are considerably stronger than the Unwin
formula could indicate. In discussing this matter with well-known
machinery builders, we found that while the strain on the hooks
might figure at 17,000 lb. per sq. in. from the formula, and show a
factor of safety of only three, actually the factor of safety must
have been five or six.
4 If possible, I would hke to know how the author can reconcile
these facts, with the practical ultimate strength tests in connection
with the elastic limit.
Prof. Wm. H. Burr. Professor Rautenstrauch has added a very
interesting chapter to the literature of this subject, but there is per-
haps a little more to the matter than has been indicated, and it bears
a good deal upon what has been said by the last speaker. Doubtless
the analysis based upon Professor Pearson's paper, as an analysis,
is a decided improvement upon the Unwin formula, but again there
comes in the same question raised in connection with reinforced-
concrete beams. This analysis, whether by Professor Pearson or
Professor Unwin, is based upon what is ordinarily known as the
common theory of flexure, which belongs accurately only to straight
beams of very small depth in comparison with the length.
2 Hooks and all such members as those shown by the author
are exceedingly short as beams, and they are also curved. These
conditions completely demoralize the analysis as based on the com-
mon theory of flexure, and it is not a matter of surprise that hooks
should show so much gi eater carrying power than the computations
would indicate. In fact, it is precisely in line with what we find in
other short beams.
3 The pins at the panel points of pin-connected bridges are
designed by the common theory of flexure. Yet if one should com-
pute the extreroe fibre stresses in those pins at some panel points as
they have existed, they would be found to run up not only to
CURVED MACHINE MEMBERS UNDER ECCENTRIC T.OAD 571
142.000 lb. per sq. in., but to 180,000 or 190,000 lb. in structural
steel. A partial explanation lies in the fact that an analysis is used,
which, strictly speaking, does not apply to these conditions. The
hook and all such members, as well as bridge pins, are short, thick
beams to which the usual theory of bending does not strictly apply.
4 Again, one will find that in bridge specifications, the regular
working fibre stresses in pins are permitted to be at least 50 per cent
greater than in the tension members of the truss; that is, one may have
a working stress of perhaps 14,000 lb. in bars, and a fibre stress in
tension of 18,000 or 20,000 lb., sometimes even 24,000 lb. in pins.
This is due to a fact I have already mentioned, that as a matter of
accurate analysis, the common theoiy of flexure should not be used
in connection with such members; but there is nothing else to be
done.
5 That again brings me back to the same point made in connec-
tion with concrete beams. The proper procedure is to settle upon
some sensible working formula, just as we do in connection with the
pins in bridges, make tests of such members, and deduce from these
tests such empirical quantities as may be properly used in the formula,
so as to make the results of the analysis in that way conform to safe
and sensible practice.
A. L. Campbell.* Table 2 of Professor Rautenstrauch's contribu-
tion shows an excellent agreement between actual test conditions
and the results obtained by the formula which is the basis of his
discussion.
2 A much simpler formula is used by the writer for similar com-
putations. A crane hook or the frame for a punch is really a tension
member with an exaggerated eccentric load. The maximum unit
stress in such a tension member may be proved equal to
A \ K
using the author's notations. The radius of gyration, R, is
equal to a I • Applying this formula to the frame shown in Fig. 4
gives ff = 7600 lb. per sq. in. This stress is 90 per cent of that given
by the more complex formula.
*The Solvay Process Co., Detroit, Mich.
572
DISCUSSION
Frank I. Ellis. While the paper, together with the article in
the American Machinist to which it refers, covers very fully the design
of hooks, giving results which agree remarkably with actual tests, its
application to shear housings is not quite clear.
2 We note primarily, that in the derivation of his formula the
writer has assumed the entire area to be in tension, i. e., the neutral
axis to lie entirely without the section. While this condition is almost
universally correct in hooks, it will seldom be encountered in shear
housings, but still it appears to have important bearing on the form
of the equations.
r
Gravit3- Axis
-p=co '
7 I V
Fig. 1 Frame with Infinite Radius
3 Another point which is not quite clear, but is a matter of
great importance, is the assumption of the value of ,o, the radius of
curvature of the gravity axis of the section. In the case of a hook,
this of course is obvious, but in machine members, such as
shear housings, this seems far from being the case. For instance, in
a housing of the general form of sketch shown in Fig. 1 herewith, we
would have an infinite value of p. This would reduce the formula to a
case of simple tension, which is obviously incorrect, giving stresses
that would be very much less than would be obtained by actual test.
On the other hand, if we consider an extreme case as in Fig. 2, where
the value of p is very small, the stress as calculated by the formula
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD
573
would be very much in excess of what could possibly exist in the
actual casting.
4 The example of a shear housing which Professor Rautenstrauch
has chosen as an illustration appears to be at variance with our
experience. The stress calculated by the new formula is almost
three times that obtained by the usual methods of computation. In
our experience, cast-iron shear housings in which the calculated
stress is 3,000 lb. per sq. in., never break except through defects in
the casting, a condition which could hardly exist if the actual stress
were in" the vicinity of 9,000 lb.
Fig. 2 Fbahe with Small Radius
E. J. liORiNG.* The figures in the paper show such striking dis-
crepancies from these obtained by the usual methods of calculation that
his analysis of the problem merits the most careful consideration.
These results clearly show that the stresses, and particularly the
maximum stress in a curved piece under the combined direct and
bending load to which such hooks and gap frames are subjected, can-
not properly be deduced from the simple combination of direct and
bending stresses as determined by ordinary analysis from the stresses
in a single plane, but may be influenced to a greater extent by con-
ditions outside of the section plane, such as the relations connecting
that plane with those nearby on either side.
2 This difference between straight and curved members arises
from a different distribution of stress due to the variation of length
of fibres at different parts of the section as taken between similar
adjacent sections.
' Loring Speed (Jauge Co., 76 Highland .\ve., Somen'ille, Mass.
574 DISCUSSION
3 The usual deduction for stress in straight members commonly
applied to this problem assumes that:
a Planes remain planes after bending.
h Strain is proportional to the distance from the neutral axis.
c Stress is proportional to strain and therefore that the stress
is proportional to the distance from the neutral axis.
4 The assumption (c) that stress is proportional to strain is true only
as referring to unit strain, as long fibres will yield more under a given
stress than shorter ones. In the case of straight members the adjacent
minimum sections are parallel and the elementary fibres therefore
all of equal length, and the assumption may be applied. In the case
of a curved member, which I would define as one in which the locus
of the centers of gravity of the minimum cross sections is a curved
line, these sections are not parallel, but radiate from a center of curva-
ture so that the fibres are not of the same length throughout the sec-
tion, and a correction must be made for the variation of the length
of fibre before this assumption can be applied. This point has
generally been overlooked or considered negligible, and in this is to
be found the explanation of the difference in results. I might add
that this exemplifies the danger of applying a formula to conditions
which it was not intended to represent.
5 I am not certain that I can agree with the author in the use of
the theory of lateral contraction in the analysis. I cannot at this
moment see why it is any more necessary in the case of the hooks
tested than, for example, in the case of the test bars from which he
deduced the fibre stresses. Taking only the common assumptions,
with the correction for the length of fibre, as above noted, it is possible
to obtain results in very close agreement with those given in the
formula recommended by Professor Rautenstrauch. In place of the
usual straight-line diagram of stress on the section these assumptions
give the stress at any point as varying according to
y ^^ yp (See Fig. 1)
or
1 _,_2/ io + y
using the symbols of the paper, and from this may be determined the
important fact that for the case represented by the hooks, where the
line of application of the load contains the center of curvature, the
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD 575
' 7.J c, furvature^"
Load 20.000
10,000 g*
Fig. 1 Diagrams for Trapezoidal Sections of Straight and Curved
Members for Equal Intensity of Stress
676 DISCUSSION
neutral axis contains the center of gravity. In other words, instead of
the stress at the gravity axis being equal to the distributed stress as is
true for straight members, the stress at this point in a member with
this degree of curvature is zero, and this represents the manner in
which the stress " piles up" toward the inner edge of a curved member.
This condition of stress at the center of gravity would be represented
in the analysis of the paper by the condition ^-^ = 1 + j^- The empiri-
cal formulae recommended give y^ = 1 -{- \.\ y^. ^^^^ the data on the
hooks give a variation from ^, =1 + I.V7 Y2^^Ti = 1 + 0.88 ^'a with
an average oi y^ = 1 + 1.015 y^ so that it will be seen that this is
approximately true by Professor Rautenstrauch's analysis; that the
gravity axis is the neutral axis for this degree of curvature just as it
is for transversely loaded beams.
6 I have applied the variation of stress
y
1+^
given above to the solution of an assumed section and find that the
stresses and their manner of variation are substantially identical, for
this particular case at least, with those given by the author's method.
I believe that an analysis can be made along this line that will give
results very close to those shown and be more generally workable.
The differential expressions for the net stress on the section and the
moment of the stress are similar to those for a beam with the addition
of a factor
1
p
w • V 2/ d il ^, . yV^dA
W varies as 2. ~ Wl varies as 2 ^
It may perhaps be possible to deduce some general expression to be
used as a factor of correction for curvature to be used with the usual
methods.
7 The effect of the curvature is less, the greater the ratio of radius
of curvature to the depth of section. In the case of hooks this means
greater strength where the contour of the inner edge is elliptical
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD 577
instead of circular, so that the curvature at the most strained section
is less. As the curvature tends to "pilej'up" the stress toward the
inner edge, greater strength may be had by giving the hook a closer
approximation to a T-section, by which'' means the metal is massed
better where the stresses are abnormally high. It would also appear
that a high gap is stronger than a low one for the same depth, since
a lesser degree of curvature is'^possible.
8 I must disagree with the statements in Par. 2 except as limited
to curved members; also with the statement in Par. 8 that j-^ and ^-j
are constants for all sections of similar form, except it be modified to
say " of similar form, curvature and load distance. "
9 In determining the maximum stress by the method which the
author has proposed, the function ^-j is the most important factor,
and this function is obtained from the difference in area of two
derived curves; the difference is small and the less the difference
the greater the maximum stress. It would seem that there is great
opportunity for inaccuracy in|^ determining this factor. It also
appears to me to be simpler to take
(rh)'
^dA
as originally stated, for the purpose of the computation, rather than to
use the value derived from it, of
r2 = ri - ^
for the reason that having the quantities for the determination of j-^
for various points of the section, that is, the values of ( 1 + ^1,
it will be simpler merely to multiply these by the respective distances
from the gravity axis, plot the curve and integrate for the net area,
rather than to proceed by raising the denominator to a new power
and passing through all the processes anew.
10 It is stated that the standard section selected for the compu-
tation of constants for the empirical formula is not the most economic
from the standpoint of equal tension and compression stresses. This is
578 DISCUSSION
true even if the member is straight, in which case, considering the
trapezoid only and omitting the curved ends, the maximum stress in
compression is 85 per cent of the maximum stress in tension. All
other parts remaining the same, for equal intensities of stress in the
edges for a straight member, the half width of the narrow edge should
be 0.095 r, as may be very readily demonstrated. The geometrical
relations for the correct proportions of a trapezoidal section for equal
intensity of stress in a straight member are so exceedingly simple that
I want to give them here, particularly since, so far as I know, they
have never been published. This relation is that the sides extended
intersect at a distance from the far or narrow edge equal to the
distance of the load line from the near or wide edge, and for the
solution of this case we have
d
^6%^
where d = depth of section
y — distance of load hne from the near edge
F = load
/ = maximum stress at the near edge or far edge (equal)
and A; is a design constant = ratio of depth of section to width of far
edge.
1 1 For the case of equal stresses in^a curved member of trapezoidal
section with center of curvature on the load Hne, a similar relation
may be deduced from the analysis that I have here suggested, but
is not quite so simple : the point of intersection of the sides is given by
the following construction: Lay off on the axis of symmetry and
toward the far edge a distance from the near edge equal to the dis-
tance from the near edge to the center of curvature and load line. If
this distance is greater than the depth of section, equal stresses may
be had. If this distance is equal to the depth of section, i. e,, if the
point thus laid off is on the far edge, equal stresses require a triangle
with this Doint as the apex. If the point is bevond the far edge.
divide the distance to that edge in thirds; then the stresses are equal
when the sides extended intersect at the nearer point of division, one-
third of this distance from the far edge.
12 It will be noticed that this construction gives the radius of
curvature for this limiting case equal to 1.33 times the depth of sec-
tion^ instead of 1.75 as given by Professor Pearson. I have investi-
^Or depth of section equal to gap depth.
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD 579
gated this case for both degrees of curvature by the method involving
the lateral contraction and find that using the formulae given by
Professor Rautenstrauch the curvature of 1.33 times the depth,
measured to the gravity axis, gives a stress on the inner edge of 1.091
times that on the far edge. A similar operation for curvature of 1.75
as recommended by Professor Pearson, by his own method gives by
my computations a ratio of stress of 0.912. A sharp triangular sec-
tion such as this is, however, of little or no importance in actual con-
struction, and the method of determining the proportions which I
have given will, I think, be found to be of much more general appli-
cation. I am unable to state at the present time whether a section
having equal intensity of stress on the two edges is or is not the most
economical of material; but presumably it is.^
Prof. C. E. Houghton. The agreement between the elastic limit
as calculated by the proposed formula and that as derived from the
tests is, to say the least, wonderfully close, and the wide variation
between the experimental values and those calculated by the use of
a theory that has been in common use for many years leads one to
ask "Why are there not more failures in crane hooks?"
2 Objection has been made to the tests because the hooks were
not loaded beyond the elastic limit. This seems to the writer to be
a mistake. What the engineer is mostly interested in is the effect
of loads that produce stresses within the elastic limit, since the great
majority of the formulae used for the calculation of stresses are based
on theory that no longer holds true after the elastic limit has been
exceeded.
3 Professor Burr has pointed out that the simple theory of flexure
does not'apply to curved members and Mr. Gabriel notes that stiffness
and not strength is the controlling factor in many of the open-side
machine frames. May not the fact that cast iron is used in the
majority of such frames be another reason why the flexure formulae
cannot be expected to give correct results? The well-known fact
that the physical properties of any cast iron vary with the rate of
cooling, and that the tensile strength and modulus of elasticity are
not constant at all depths from the surface of any cast-iron member,
but vary throughout any given section, leads one to ask "Is it not
more reasonable to use the simpler formulae in the calculations for
'Since writing the foregoing, Mr. Loring has found that the method sug-
gested by him for the determination of the stresses — or a very similar one
— ia given in sjme detail in Hiitte. from some German source dating 1902.
580 DISCUSSION
strength and to provide against possible errors by that useful and elastic
term — the factor of safety?"
H. Gansslen.* The author's tests prove the correctness of Andrews
and Pearson's new formula for figuring crane and coupling hooks.
All the experimenters, however, seem to have limited themselves to
these hooks, for which the formula appears to have been gotten out.
Hook's law of the direct proportionality between stresses and strains
also underhes the new formula and the fact that this law holds practi-
cally good on wrought-iron, steel and similar materials would to
some extent explain the good agreement of the results of tests and
calculations by means of the new formula.
2 The author points out that the formula is applicable to punch
and riveter frames. To generalize thus I believe is hardly wise at
present, as all the various formulae for figuring curved beams are
more or less empirical and each of them is naturally proved to be
true for a certain limited field of calculations only. Hook's law does
not hold true for copper, cast-iron, bronze, stones, artificial and
natural, etc., and this law giving the modulus of elasticity as constant
is the basis of the formula.
3 Engineers know that the old formula for figuring a curved
member in the same way as a straight beam gives factors of safety too
small, but that we are now underestimating the stresses in the throat
of punch press frames 8500 -^ 2450 == 3^ times is surely saying much.
4 However, there is no use disputing the new formula in so far
as tests have verified it and it is to be hoped that the author will have
the opportunity of entering other fields of research besides that of
crane hooks. That of press frames would be a desirable one.
5 I have not come across a case where a punch press frame
figured in the usual, but wrong, way could have been 3^ times under-
estimated, roughly considered by comparing the pressure exerted with
the general behavior of the frame.
6 The old theoiy of flexure as applied to and compared with tests
of cast iron has shown its inapplicability and this should make us all
the more cautious in adopting the new formula for cast-iron press
frames before having on hand the results of tests that would justify
us in so doing.
John S. Myers.2 The author's presentation on the design of
curved machine members and his article in the American Machinist
• Mechanical Engineer, 404 Fisher Bids;., Chicago.
2 John S. Myers, 2456 AUnond St., Philadelphia.
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD
581
of Oc'tobtM- 7, 1909 dealing exclusively with crane hooks, seem to indi-
cate that the new theory is applicable to punch and riveter frames of
the type shown in Fig. 1, where the throat is semi-circular, being
struck with a radius having its center at 0. In order, however, to
find the radius of curvature of the gravity axis of the principal section
it would seem necessary to plot points such as A, B, C, D, E, draw
a curve through them, then, by trial, find the center 0' of a circular
arc which will pass through C and most nearly fit the curve for points
intermediate between B and D.
^FiQ. 1 Frame with Semi-Circular
Throat
Curve A BCD represents the gravity axis
of the section. Point 0 is the center of
the throat radiios. Point O' is the center
of a circular arc which approximately coin-
cides with the gravity-axis curve for points
between B and D.
Fig. 2 Fraaie with Wider Gap
THAN Fig. 1
Curve ABCC'DE represents the grav-
ity axis. Between points C and C this
curve becomes a straight line; hence
2 If the above is consistent with the assumptions upon which the
theory is based, it will be seen that the point 0' is not necessarily
coincident with 0, and that to find the value of p^ a layout must be
made and the gravity axis of several sections determined. It
is also seen that p^ is not strictly a function of the throat radius nor is it
equal to OF + CF as one would at first suppose. This adds more
comphcation to the problem, which is already vexatious.
3 Again, such frames are not always made with the throat struck
with a single radius; in fact, this is the exception rather than the rule
for a large class of machines, which have a wider ''gap" to
accommodate the work and are more Uke that shown in Fig. 2. Here
the curve representing the gravity axis is a straight line between
582
DISCUSSION
points C and C , in consequence of which ^o^ = oo and it would there-
fore seem that the new theory did not apply to this portion of the
frame. Now, if this be the case, and we design that portion of the
frame between OH and O'H' according to the old theory of straight
beams, but design section 01 according to the theory of curved beams
under discussion, it would appear from an inspection of the results
given by Professor Rautenstrauch that section 01 should have about
three or four times the flange area of section OH. Of course the
metal at the corners could be thickened, as indicated by the dotted
line at /, but it would be out of the question to double or treble the
usual flange thickness, which is what the new theory seems to indi-
cate as necessary.
A PofiT/o/si or
T
-D'
mm
'M'f'-MWMW/r'^m^^
r^
w..
i\
'C
Fig. 3 Fia. 3a
Fig. 3 Showing How the Rapid Transition of Stresses Induces Local
Stresses. Fio 3a. Pboposeo Section
4 It would be very interesting to know how the new theory could
be properly applied in such a case; whether, for instance, it is entirely
applicable at the section OG but gradually merges into the old theor>
at sections OF and OH; or whether it has not, as yet, been sufficiently
developed to be generally applicable to sections other than those at
right angles to the line of action of the force.
5 Generally speaking, a structural engineer never puts in curved
tension or compression members because he knows that force either
travels in straight lines or else produces bending strains; but the
average designer of machinery seems to delight in curved ribs, bent
levers, and the like. The average mechanical draftsman makes
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD
583
layouts as if he held the opinion that force travels along a curved rib
in a manner somewhat similar to water flowing in a pipe and that it
will, therefore, follow any devious or sinuous course in which he may
choose to distribute the metals. Most C-frames seem to be designed
on the foregoing assumption and, while it is an exceedingly difficult
piece of mental gymnastics to follow the mathematics of the new
theory, it is, however, quite easy to see that there are stresses induced
in curved ribs which are usually ignored.
6 To illustrate the foregoing, Fig. 3 shows that portion of the
frame of Fig. 2 which lies between hues OF and OH. Now, let T
and T^ represent the total tensions in the flanges on sections OF and
I I
NEW Wf
j THEORY. I •
B
riHi^
Suggestion.
c
Fio. 4 Distribution of Stre8se3 Under Different Theories
OH respectively. By combining T and T^ graphically it is seen that
a resultant force, F, must, in some manner, be supplied to establish
equilibrium. The most direct way of supplying such a force is by
the addition of a rib as indicated by the dotted lines at R which will
distribute part of F into the web and deliver part of the force at the
compression flange where there is a smaller, opposing resultant force.
In the absence of any such rib the necessary force must be supplied
by the web, partly through a local bending and distortion of the flanges
as indicated by the dotted lines at D and D' and partly by a concen-
tration of stresses towards the central portion of the flanges as indi-
cated at C, this concentration being a direct result of the deformation
at D'.
584 DISCUSSION
7 In supplying a rib R, if it was intended to carry the entire force
F it would be necessary to make it about If times the average thick-
ness of the flanges, but since the web can readily take half, or more
than half, of the load, it would seem that a rib of | or ^ of the flange
thickness, narrowed down at the center as shown in Fig. 3a would be
entirely sufficient, especially if the web be judiciously thickened and
liberal fillets used.
8 As I understand the new theory it does not recognize any such
concentration of stresses as indicated at C in Fig. 3 but, on the con-
trary, assumes a more rapid concentration towards the extreme fibres
in a manner somewhat similar to that shown at B in Fig. 4. Now
in view of the close accord between the new theory and the results of
Professor Rautenstrauch's experiments, I am quite ready to believe
that diagram A represents quite closel}^ the actual conditions for
straight beams of solid section, and that diagram B represents the
most plausible theory for curved beams of solid section ; but that for
beams composed of heavy flanges and a light web the probable dis-
tribution of stresses is more nearly Uke that suggested by diagram
C, and that so far as the curved form of the beam is concerned, it is
not the curve of the neutral axis we are interested in but the curve
of the flanges, and that this results in local bending and concentration
of the stresses as already pointed out.
9 I have no well formulated theory to advance in explanation of
my belief in a distribution of stresses like that indicated by diagram
C but have sufficient faith in it to calculate sections of this nature
by the very simple process of considering the stress to be uniformly
distributed over the flange area and entirely neglecting the web ; then
at points where there is rapid transition of stresses, supplying ribs,
thickening up the web and allowing a lower flange stress and liberal
fillets. This procedure may sound crude to a scientific man, but it
has, at least, ease of application in its favor and may yet be shown
to be actually more scientific than the more laborious methods
usually pursued. As yet, I have not had the temeruy lo apply tnis
method to large work but would like to have the opinion of those
who have had experience along these fines.
The Author. The test reported by Professor Lanza is interesting,
but I do not feel justified in replying without a review of the entii-e
data on the experiment. The point made by him in Par. 5 in regard
to deflections, is somewhat misleading. I did not propose in my
experiments to determine the relation of total deflections to the max-
CURVED MACHINE MEMBERS UNDER Et'CENTRIC LOAD 585
imum stress in the hook, but rather to find the load at which the
total deflection ceased to follow the straight-line law. Since the total
deflection is dependent on the deflection of all the sections, it is
rational to suppose that when any variations occur they are due to
the fact that the "fibres" in the most strained section have been
stressed beyond the elastic limit. This is all we wish to know. The
most strained section is without doubt the main horizontal section.
The examination of the bending moments in other sections is of
no value in these determinations.
2 Referring to Mr. Gabriel's remarks: I regret that so man}-^
designers persist in applying the formulae for determining maxi-
mum intensity of stress beyond their limits of application. No com-
putations can be made to determine ultimate breaking strength and
I see no reason why anyone should be surprised that there is a dis-
agreement between the "results of computations" and the results
of test. I did not choose to consider the matter of rigidity, which
the title of the paper would lead one to believe should be included.
Rigidity is, of course, a controlling factor in die work.
3 Mr. Henderson's remark that his practical experience with
hooks leads him to believe that a rather greater strength exists than
can be expected from the Unwin formula, qualified by his report
of certain tests, would lead one to believe that he has made use of
Unwin's formula outside of its field of application. Unwin's formula
indicates nothing beyond the elastic limit. There exists no method
of analysis which enables us to determine the relation between the
load on the hook and the resulting maximum intensity of stress
when that stress is beyond the elastic limit of the material. In
reply to the statement that "the ultimate strength interests us just
as much as the elastic limit, " I would say that I believe designers
will be treading on much safer ground when they confine themselves
to proportioning parts with a factor of safety based on the elastic
limit rather than the ultimate strength.
4 Mr. Ellis says in the second paragraph " We note primarily that
in the derivation of his formula the writer has assumed the entire
area to be in tension, i. e., the neutral axis to lie entirely without the
section. While this condition is almost universally correct in hooks,
it will seldom be encountered in shear housings." No such assump-
tion is made, nor is it universally correct in hooks. I believe that
Mr. Ellis is also mistaken in his remarks on the particular form of
the equation when <> is infinite. When p is infinite the case is not that
of simple tension but rather as expressed by Unwin's formula.
586 DISCUSSION
5 Mr. Loring's explanation of the two analyses, I regret to say
is incorrect. Both analyses are founded on a determination of the
relation between unit stretch and intensity of stress, but the real
difference is found in the methods of evaluating the unit stretch.
The older formula gives the unit stretch as
y'
^y = ^3 +
while the newer analysis gives
P'
P' Po
P9
where
^y = unit stretch of any fiber a distance y' from the gravity
axis.
-^y? = unit stretch at gravity axis.
p' = radius of curvature at gravity axis after stretching.
Po = same before stretching.
yo = modified y' after stretching.
The newer analysis retains terms of the same order of magnitude
as ^y and therein lies the difference. The theory of lateral contrac-
tion is rationally applied in this analysis, its application being un-
necessary to the test piece, since direct measurement of stress is made.
6 Par. 2 in the paper is obviously Umited to curved members.
The similar form referred to in Par. 8, includes the radius of curva
ture. The method in the paper for determining yi an^ X2, I believe
will be found more convenient than those proposed by Mr. Loring.
7 Professor Houghton will agree with me that a more correct
analysis for straining action will permit a more intelligent use of the
factor of safety.
8 Mr. Myers is quite correct in his remarks on the value of p^.
The analysis, however, does apply to the case of straight beams
where |0o = 00 , for which case it reduces to the form of the Unwin
formula. The formula has not as yet beenf sufficiently developed to
determine its usefulness in establishing proportions for other than
those sections at right angles to the load. The difficulty of deter-
mining the stretch on sections at an angle to the load will leave this
problem unsolved for some time. It is, however, rational to suppose
that the flange on obhque sections should be thickened, but to what
CURVED MACHINE MEMBERS UNDER ECCENTRIC LOAD 587
extent has not yet been determined. In regard to the behavior of
a T-section, I would state that Professor Pearson has found experi-
mentally that it is subjected to the same laws as a soUd section.
This indicates that the suggestion of Mr, Myers in Fig. 4 can hardly
be accepted.
9 I judge from Professor Burr's remarks that he discredits the
analysis by Professor Pearson on the basis that it is founded on the
common theory of flexure, that is, it is not applicable to beams of
very grt^at depth compared with the length. I believe that if Pro-
fessor Burr had given more thought to the matter he would not have
made this statement. In view of the experimental results obtained
by myself and others in verification of the theory and the lack of any
data in verification of Professor Burr's statement, I am still inclined
to believe that Professor Pearson's analysis is correct.
1
No. 1253
TESTS ON A VENTl RI METER FOR BOILER FEED
By Prof. C. M. Allen, Worcester, Mass.
Member of the Society
A reliable and accurate hot-water meter has been in demand for
a good many years. The principle of most of the cold-water meters,
where there are moving parts in the water, is not at all adaptable for
hot-water work. A hot-water meter for boiler-feed purposes must
stand not only the variation in temperature but also considerable
variation in pressure, and quite often it has to stand a certain amount
of watei-hammer, this depending somewhat upon the style and con-
dition of pump used. The Venturi meter, having no moving parts
to get out of order and being of material which will stand the ordinary
corrosive effects, should make a reliable hot-water meter.
2 The object of these tests was to determine the accuracy of a
Venturi meter to be used for measuring boiler feed under a great variety
of conditions. The plan was to make a complete series of tests
upon a small Venturi meter under all the probable conditions that
would ever be met in boiler room practice. The tests were made un-
der varying temperatures and velocities; under varying pressures,
intermittent and steady ; using a triplex power pump in good condi-
tion, and with one plunger out of commission, a duplex steam pump
in good condition and in poor condition, and an injector.
3 The meter was installed m the steam engineering laboratory of
the Worcester Polytechnic Institute and set up the way most conven-
ient not only for weighing the water passing through the meter, but
also for heating the water before it went through the meter, and pump-
ing it in in various ways. The meter used was built by the Builders'
Iron Foundry of Providence, R. I., and is what is ordinarily called a
2-in. meter, the upstream and downstream ends being 2 in. in diame-
ter and the throat f in. in diameter. The main part of the meter is
of cast iron and the internal portions are lined with brass. Surround-
ing the upstream end and throat are annular chambers between
Presented at the Annual Meeting, New York (December, 1909), of The
American Society of Mechanical Engineers.
590 TESTS ON A VENTURI METER FOR BOILER FEED
the brass sleeve and the iron casing. Six holes are drilled through
the brass lining into these annular chambers at about equal distances
around the circumference, in order to give the actual pressure heads
in the meter at both throat and upstream end. From the outside
of these annular chambers were pipe connections to a manometer
tube which consisted of a glass U-tube containing mercury. There
were the necessary valves and pet-cocks to manipulate the meter,
blowing out the air whenever it accumulated. The general layout
of the apparatus is shown in Fig. 1.
4 The apparatus was set up so that the meter could be supplied
from a l^^-in. metropolitan injector, a 4^-in. by 2f-in. by 4-in. duplex
pump or 4-in. by 5f-in. triplex power pump; or from a pressure
tank supplied from the city mains or by a large duplex pump. These
pumps were arranged to take their suction from a pit 12 ft. long, 6 ft.
wide and 4 ft. deep, directly beneath the Venturi meter. A 1-in.
steam line was put in to heat the water. There being about 300 cu.
ft. in the supply pit, a very even temperature could be maintained.
In order to keep the discharge from the pumps constant, a suction
well was supplied, kept at constant level by an additional pump from
the main pit. The discharge from all the pumps used was carried up
a vertical 2-in. pipe, at the top of which was an air chamber 4 in. in
diameter and 3 ft. long, with a valve so inserted that it might be cut
out whenever desired. ' From this vertical pipe ran a line containing
the Venturi meter, a thermometer-well for determining tempera-
tures, and a valve for throttling water in order to get any desired
pressure. At the end of this line was a swinging end that discharged
into either of two 5000-lb. weighing tanks.
5 The first tests were made with cold water in order to determine
the coefficient. These were made with steady pressure, securing a
very constant flow. Water was run through the meter until the con-
ditions had become constant. One tank was weighed while the water
was being discharged into the other. The tests were started by divert-
ing the discharge into the weighing tank and taking the time. Read-
ings of the Venturi meter were taken every thirty seconds for low
velocities and every minute for higher. The tests were ended by
diverting the discharge into the other tank, taking the time and weigh-
ing.
6 When hot water tests were made, it was found that a certain
amount of water evaporated ; evaporation tests were therefore made,
which proved this amount to be a negligible quantity.
7 In order to compare the workings of the meter under the various
TESTS ON A VENTURI METER FOR BOILER PEED
591
592 TESTS ON A VENTURI METER FOR BOILER PEED
conditions, it was decided to determine the coefficients for this meter
under these conditions. The discharge of the Venturi meter was
figured from the regular fo-rmula, using a coefficient of one, and the
actual weight obtained from the weighing tanks was divided by this
value to obtain the real coefficient.
8 The following temperatures were used during the tests : 80 deg. ,
120 deg., 140 deg., and 180 deg. fahr. Water for these tests was
supplied by the triplex power pump with different velocities through
the meter. Tests were made at 140 deg. with and without the air
chamber, the water being furnished by the triplex power pump ; then
at 140 deg. with one plunger of the triplex pump disconnected so as
to produce fluctuations in the velocity and pressure of the water sup-
plied to the meter.
9 In order to duplicate more nearly the conditions of boiler feed,
an air chamber and check valve were placed in the pipe line in the
downstream side ofj the Venturi meter (not shjown in Fig. 1). Be-
cause of the air chamber at this end the pump fluctuations could pass
through the meter to a much more marked degree than if the dis-
charge was merely throttled by a valve. Under these conditions tests
were run with water supplied by the injector, taking suction from a
special supply tank; with the injector, however, the temperature of
the water varied of necessity with the velocity through the meter.
CONCLUSIONS
10 The chief difficulty encountered in making this series of tests
was in getting the^true average^readings of the manometer. With
the higher velocities through the meter, the fluctuations could be
easily dampened by closing the valves to the manometerj^tube, but
with the lower velocities any error in reading was so large in propor-
tion to the entire head as to make a considerable difference' in results.
It may be said, then, that the meter is not accurate for velocities of
less than 10 ft. per sec. in the throat of the meter, which corresponds
to a discharge of 0.03 cu. ft. per sec, or about 6140 lb. per hr., so that
this meter would be best adapted for measuring water for a boiler
plant of above 200 h.p. The coefficients are materially lower below
0.03 cu. ft. per sec. The principal feature shown by the cooler water
tests (80 deg. fahr.) is the low value of the coefficients of the meter,
the average, excepting the values for velocities below 10 ft. per sec,
being 0.978. This coefficient might have been expected, however, as
in a small meter the ratio of area of cross section to surface is much
lower.
TESTS ON A VENTUHI METER FOR BOILER FEED
593
1 1 These experiments clearly show that the meter is as accurate for
hot water as for cold. The maximum error in discharge, as figured
from manometer deflections using the mean coefficients for that tem-
perature, is as follows: 80 deg., 1.39 per cent; 120 deg., 1.5 per cent;
140 deg., 1.9 per cent; 180 deg., 0.82 per cent. The average error
is well within 1 per cent.
12 Of the pumps tried with the meter, the triplex gave by far the
best results, and it may be confidently stated that the Venturi feed-
water meter would give very satisfactory results in a plant using the
power pump. Even with one plunger disconnected, the maximum
variation was only 2.4 per cent.
13 In tests with the injector, the weighed calculation from the
Venturi formula, using the mean coefficient, shows variations from
actual weight of 3 per cent. The average error is inside 2 per cent.
Fig
08
97
96
50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 300
Tenipeiatuie, Degrees Fahrenheit
2 Curve Showing Variation of Venturi Coefficient with Rise in
Temperature
14 The results from 4^i-n. by 2|-in. by 4-in. duplex pump show up
better than might be expected as this pump was in very poor condi-
tion, the piston rods being so worn that the pump took air through
the stuffing boxes at the^^water end.
15 These tests represent the worst conditions^ which would be met
with in boiler feed. The pump would start building up pressure,
then pause until the pressure had fallen. The check valve, which
was located only about 6 in. downstream of the Venturi meter, was
opening and closing constantly. For a discharge of more than 0.03
to 0.04 cu. ft. per sec, the coefficients were very consistent.
16 It must be remembered in considering these tests that the Ven-
594 DISCUSSION
turi meter itself probably worked more accurately than the tests
would indicate, as every change in velocity through the meter is
accompanied by its corresponding change in head on throat and up-
stream end, and that a continuous recording device attached to the
meter would probably cut down the error considerably. It is the
opinion of the experimenters that the Venturi meter is a very reliable
form of hot-water meter, provided the proper size is used.
17 These tests were conducted for the most part by George Y.
Lancaster, a post-graduate student in mechanical engineering at
Worcester Polytechnic Institute, and these results are taken from
a thesis submitted by him.
DISCUSSION
F. N. CoNNET. I notice also that the results obtained were not
very satisfactory when pulsations were present and when the throat
velocity was less than 10 ft. per sec. The three reasons for
this seldom if ever exist in actual Venturi meter installations:
a The instrument used in the test was a mercury U-tube or
manometer, containing but little more than a pound of
mercury. The inertia of the mercury was therefore
small and the mercury levels were unsteady. In an
actual installation a registering instrument is generally
used which contains almost 100 lb. of mercury, the mere
inertia of which has a decided "damping" effect.
b The graduations on a manometer scale are quite close to-
gether at low throat velocities. At 10 ft. per sec. throat
velocity, the difference of mercury levels is only li in.
In the registering instruments the movements are in-
creased by a lever so that accurate readings are facilitated.
c During the tests described in the paper, the globe valves
in the two pressure pipes were partially closed to minimize
the mercury level fluctuations, and in all probability the
valve discs were slightly loose on the valve stems. This
therefore allowed the discs to behave like check valves
and permitted a freer flow in one direction than in the
other; consequently incorrect mercury levels would result.
2 The only reason for not always obtaining accurate results with
the Venturi meter for boiler feed is the presence of severe pulsations
in velocity due to the action of the feed pump. The most accurate
results can be obtained when the feed pumps are of the centrifugal
TESTS ON A VENTUBI METER FOR BOII.ER FEED 595
type and many such pumps of the two-stage or three-stage turbine
variety are now in successful use. The pulsations which are due
to the action of the water plungers or to defective valve action in a
reciprocating pump, make it necessary to place a rather large air
chamber directly on the pump, or on the feed line as close as possible
to the pump. If placed on the feed line, it should not be connected
on a tie set in the line but it should be so arranged that all of the water
will pass through it. Furthermore, the cross section of such an air
chamber should be large and the arrangement should be such that the
surface of the water will rise and fall with each stroke of the pump.
There should be a gage glass on the side of the air chamber so as to
insure the presence of a sufficient vacuum of air. These precautions
will render the velocity of the water sufficiently uniform to obtain
accurate results with the Venturi meter.
3 The correction necessary for difference in temperatures with
the Venturi meter is not as great as with mechanical meters,
for the reason that the Venturi meter itself automatically com-
pensates for one-half of the difference in specific gravity. In
other words, if the water be hot and the specific gravity 2 per
cent less than that for which the meter was calibrated, a correc-
tion of 1 per cent is automatically made by the meter and there-
fore a further correction of only 1 per cent [is necessary, whereas,
with a mechanical meter depending upon volumes, a correction of 2 per
cent woul d have to be made if the readings were desired in pounds. The
reason for this difference between the two types of meters is that the
flow through the Venturi meter is proportional to the square root of
the Venturi head and is not directly proportional to it.
4 There are at least three better ways to throttle one or both of
the pressure pipes than by using globe valves. The first and perhaps
the best way is to use a capillary tube, say i-in. inside diameter by two
or three feet long. The second way is to use a needle valve which
is similar to a globe valve, but without a loose valve disc and with
a long tapered point directly on the valve stem. The third way is
to use a cock instead of a valve. Any of these methods of throttling
combined with an ample air chamber permits accurate Venturi meter
readings at throat velocities as low as 2.8 ft. per sec. This extends
the range of the meter from its maximum capacity down to one-thir-
teenth of the maximum.
5 Although a manometer, because of its portability and simpli-
city, is particularly well adapted to the making of short boiler tests,
it nevertheless is not automatic and it shows the rate of flow only at
596 DISCUSSION
the moment of observation, and if this rate fluctuates considerably
from minute to minute, it becomes necessary to take very frequent
readings. For this reason an instrument has been perfected which
has two dials, one for indicating the rate of flow and the other for
continuously recording this rate upon a circular chart paper. A
special planimeter enables the charts to be measured so as to obtain
the total quantity of water. This planimeter multiplies the factor
of velocity by the factor of time and the product, of course, represents
quantity. This type of recording instrument is largely used for
meters 4 in. and smaller in diameter but for larger size meters the
users generally prefer a three-dial instrument of the integrating type
in order that the total quantity of water may be read directly upon
a revolution counter without the aid of the planimeter.
Clemens Herschel. Professor Allen's paper shows, by tests prop-
erly and skilfully made, that the meter is reliable for hot-water
and boiler-feed service, and is new and unique as reproducing in
tests of the meter the curious conditions to which a boiler-feed water
meter is subjected. But for this feature the tests would have been
only a repetition of other tests already made. Not that such repeti-
tions are not desirable, especially when made as accurately and with
the scope and purpose of those given in Professor Allen's paper.
Further series of tests on Venturi meters of all sizes, are in fact still
called for in the interests of exactitude. But they can only in a
general way confirm, not discover.
2 The point to be considered is, that several thousand Venturi
water meters are now in use, the world over. They are the embodi-
ment of the action of one of the laws of nature, and are but little
dependent on a correction by coefficients. They have been tested
in various sizes, from i-in. to 10-ft. main pipe diameter, and operate
exactly alike in all these sizes. They are also used to meter gases,
brine and chemicals, and, as we see from the paper, to meter hot
water. It is indeed a curious circumstance, that while the inventor
and the manufacturers of the Venturi water meter never expected to
see many of these meters of less than 12-in. diameter used in practice
yet the demand for hot-water boiler-feed meters has exceeded in
value that of all the other sizes, for certain periods.
Sanford a. Moss. I understand from Par. 7 that the discharge
of the ^'enturi meter was figured on the basis of cold water with stand-
ard density in all cases, and that the theoretical effect of change of
PESTS ON A VENTURI METER FOR BOILER FEED 597
density was not taken into account in the formula. This would
mean that Professor Allen's curve takes account of the effect of den-
sity changes, as well as all other changes. The actual formula used,
and a sample of the calculations, might be a desirable addition to the
paper.
2 Assuming that the above interpretation is correct, Professor
Allen's curve shows that the actual flow in pounds per hour, with
a given pressure, increases as the density decreases, due to rise of
temperature. Is this not surprising? Theoretically, flow should
decrease with the square root of the density. Of course change in
the orifice friction coefficient, due to change of density, temperature,
etc., might occur to such a great extent as to overbalance effect of
density change. The actual orifice friction coefficient would then
have a greater upward slope than in the chart so as to be over 98 per
cent at 200 deg. Orifice friction coefficients for all density conditions
and all fluids are usually the same for velocities occurring in practice,
which are always above the "critical velocity" where fluid adjacent
to a wall is stationary and where viscosity is a factor. Thus the orifice
coefficient for air is the same as for water, even though the density
is decreased about 800 times.
F. N. CoNNET. If I understand Dr. Moss correctly, he states that
the quantity decreases as the -density increases. With the Venturi
meter this depends upon the character of the graduations. If the
units are cubic feet the readings decreasem proportion to the square
root of the increase of density, but if the units are pounds the readings
increase in proportion to the square root of the increase of density.
One is exactly the reverse of the other.
Geo. a. Orrok. I note that Professor Allen has obtained results
for the coefficient of the Venturi meter similar to those given by
Clemens Herschel in his paper presented before the American Society
of Civil Engineers, December 21, 1887, the lower values of the coeffi-
cient appearing at a velocity of about ten feet per second.
2 The New York Edison Company for some years has been using
\'enturi meters for the measurement of water. We find them accurate
and very convenient. For the last three years we have been using
them in the testing of our boilers, having conducted a series of check
experiments to determine the variations with temperature. Our
condition is considerably better than Professor Allen's, since we use
centrifugal feed pumps and consequently have a steady reading on
the manometer.
598 DISCUSSION
3 In cases where we have both weighed and measured the feed
water our results were remarkably close. On a 7-hr. test, where about
170,000 lb. of water was fed to the boiler, the meter exceeded the weigh-
ing by 631 lb., or approximately 0.37 of one per cent. In another
test, in which nearly 200,000 lb. was fed, the difference was about
0.47 of one per cent. I believe the meter readings are more nearly
correct than the weighing, as there was considerable opportunity
for evaporation from the tanks in which the weighing was done.
The Author. I agree with Mr. Connet in regard to the throttling
of the water in the pipes leading to the manometer. I believe the needle
valve, or a fairly long pipe of small diameter, would be a decided im-
provement over the globe valves which were used in these experiments.
We had not discovered that the movement of the end of the globe
valves affected the reading, but Mr. Connet has had a good deal more
experience along these particular lines, and I am perfectly willing to
believe that this is true and that these fluctuations could be materi-
ally cut down and yet give the true mechanical average. This is
what we are looking for, and it is a good deal better than using maxil
mum and minimum readings and then obtaining the arithmetical
average. The mechanical average obtained by means of throttling
is certainly more accurate because we do not know how long the maxi-
mum deflection continues, relative to the minimum.
2 For the benefit of Mr. Moss, I would state that the density at
different temperatures was considered. The following is a sample
test giving an idea as to how computations were made :
UW= actual weight of water from weighing tank, then
W = Q0waCt\2gh
w = weight per cu. ft. at the temperature
a = area Venturi throat
C = Venturi coefficient
t ■= time in minutes
h = Venturi head
W
c= p=-
60 w a t \'2 gh
W
C = 7
1.48 w t \h
TESTS ON A VENTURI METER FOR BOILER FEED 599
DATA OF TEST
Time 3:40 —3:51; duration 11 minutes. lb.
Weight of tanks at beginning 1158
Weight of tanks at end 5369
4211
Deduct for tank calibration . 20
4191
Add for evaporation 2
Total water 4193
Mean mercury deflection 17 . 24 in.
h = 1.05X 17.24 = 18.1ft.
^h = 4.25
w for temperature of 137 = 61.43
Weight = 1.48 X 11 X 61.43 X 4.25 = 4250
C = = 0.986 coefficient of Venturi meter
4250
No. 1255
THE PITOT TUBE AS A STEAM METER
By Prof. Geo. F. Gebhardt, Chicago, III.
Member of the Society
Steam meters may be conveniently grouped in two general classes,
which, for lack of more suitable names, may be designated as a, series
meters, and 6, shunt meters.
2 The series meter is an integral part of the piping, the entire
mass of fluid to be measured passing through the apparatus. The
St. John's and Venturi meters are the best known of this class. In
the former the volume of fluid passing is determined by the rise and
fall of a weighted plug valve and in the latter the velocity of flow is
determined by the well-known principles of the Venturi tube. Both
are indicating instruments and show only the rate of flow.
3 In the shunt meter only a portion of the steam to be measured
is diverted through the apparatus, the velocity of flow through the
shunt being an indication of that in the main pipe. In this class one
or more small openings ^ in, or less in diameter suffice for attach-
ing the apparatus to the pipe. One instrument suitably calibrated
may answer for any size of pipe. The Pitot tube forms the basic
principle of practically all meters of this class.
4 It is the object of this paper to describe a number of applica-
tions of the Pitot tube for steam measurements as constructed and
tested at the Armour Institute of Technology.
5 The Pitot tube was first used by its inventor, Pitot, in 1730, in
the measurement of the flow of water and since then has been success-
fully used in measuring the flow of many fluids and all true gases.
Considerable difficulty has been experienced, however, in its applica-
tion to vapors condensable under normal conditions of operation;
and so far as the writer knows, no commercially successful instrument
is on the market.
6 Many of the instruments are interesting laboratory devices and
are of considerable value for experimental investigations; but on
Presented at the Annual Meeting, New York, (December 1909), of The
American Society op Mechanical Engineers.
602
PITOT TUBE AS A STEAM METER
account of the great number of variables involved, fall short of being
practical commercial instruments. Fig. 1 illustrates the most com-
mon and the least accurate application of the Pitot tube for measur-
ing the flow of condensable vapor in a pipe. S is the static nozzle
at right angles to, and D the dynamic nozzle facing the current. U
is an ordinary manometer partially filled with mercury. When there
is no flow the mercury in columns N and W will be on the same level
and the upper portions will be filled with condensed vapor. When
there is a flow the mercury will be depressed as indicated and the dif-
ference in height H of the mercury columns in the two tubes will be
Fig. 1 Pitot Tube with Mercury Fig. 2 Simple Gage-Glass Meter
Manometer with Self- Adjusting Water Column
a measure of the velocity of flow in the main pipe. On account of the
great density of mercury and the variation in height of the condensed
vapor above the mercury this application of the Pitot tube has very
little value scientifically or commercially. For example: With dry
steam at 100-lb. gage, a velocity of 8000 ft. per min. would give a
depression H of only 1 in. and an error of 1/100 in. in measuring H
would mean an error of 40 ft. per min. in the velocity.
7 In Fig. 2 is shown the original apparatus designed by Prof.
R. Burnham of the Experimental Department of Armour Institute
of Technology and the writer, in which the water of condensation is
used as a self-adjusting column in place of mercury. This embodies
the basic principle of many of the meters constructed later.
8 Referring to Fig. 2, A and C are two ordinary water gage cocks
PITOT TUBE AS A STEAM METER 603
and G an ordinary glass tube. Gage C is connected to the static
nozzle S and gage A to the dynamic nozzle of a Pitot tube. The
height of water H varies as the square of the velocity of steam
flowing through pipe P and automatically adjusts itself to the varia-
tions in velocity; thus, for decreasing velocities the water in glass G
discharges itself through tube D until the water column H balances
the velocity pressure in pipe P, and for increasing velocities conden-
sation from the upper part of the instrument accumulates and the
water column H rises until a balance is effected for the higher velocity.
9 The velocity of flow is determined by the well known equation :
V = c ^2gH (1)
in which
V = maximum velocity of flow, ft. per sec. Dynamic nozzle
D is inserted in middle of pipe where the velocity is a
maximum,
c = coefficient determined by experiment.
g = acceleration of gravity = 32.2.
H = height of a column of steam equal in weight to water
column H.
Equation (1) may be expressed :
V, = 139c ^/ h^ (2)
in which
Vi = maximum velocity, ft. per min.
d^ = weight of 1 cu. ft. of water in gage glass G.
dg = weight of 1 cu. ft. of the steam or mixture in pipe P.
h = height of column H in inches.
The weight of steam flowing per hour may be determined by substitut-
ing the proper quantities in equation (2) , thus :
W = 58 acr Vlid^^ . (3)
in which
W = weight of steam flowing, lb. per hr.
a = area of the pipe, sq. in. ; other notations as in (2) .
T = ratio of the mean velocity to the maximum.
10 Equations (2) and (3) are general and are applicable to any
size pipe and any pressure and quality of steam. For a given size
of pipe, say 3 in. (extra heavy), and a given pressure of steam, say
70 lb. gage, equations (2) and (3) assume the following simple forms:
V = 2435 cr V h (4)
= 1292 cr \/ h (5)
604
PITOT TUBE AS A STEAM METER
11 Tests with pipes 1 to 6 in. in diameter gave r a value of 0.79
to 0.84. For a 3-in. extra heavy pipe this value was 0.82 and re-
mained practically constant for all velocities and pressures (atmos-
pheric to 100 lb. gage.) Substitute this value of r in (4) and (5) ,
V = 1996 c, V h
w « 1060 Ci Vli
(6)
(7)
in which
V = actual mean velocity, ft. per min.
w = actual weight of steam flowing, lb. per hr.
c, = a coefficient determined by experiment.
Coefficient c^ varied from 0.8 to 1.2 for the simple instrument in Fig. 1.
fewji
T
^===-J^^'^^E^
T -T'-ZT^^zcT-zr^'-rr'-^^j:/
ii^-ir^^^i^-— — _^--ir^-r^
IMMJl
Fig. 3 Influence of Surface Ten-
sion ON Height of^Water in Dy-
namic Nozzle
Fig. 4 Dynamic |Nozzle with Ar-
rangement FOR Maintaining a
Constant Level of Overflow
12 Variation in the value of c^ was found to be due to
a Fluctuation in height T (Fig. 3) , at the end of the dynamic
tube due to surface tension. With a plain tube this varia-
tion amounted to as much as 0.25 in.
6 Variation in density of water column.
c Capillarity in gage glass G.
d Aspiration in static tube at high velocities.
PITOT TUBE AS A STEAM METER
6Uo
13 (a) A number of devices were constructed for eliminating
the variation in height of T in the dynamic nozzle but all proved
inefficient, except that shown in Fig. 4. By serrating the tube as
indicated in A'^, Fig. 4, and surrounding it with a corrugated ferrule
M, thereby forming a series of capillary tubes, the variation was prac-
tically eliminated, amounting to but 0.02 in.
14 (6) The variation in density of the water column is an inher-
ent defect of this type of meter and cannot be remedied in this simple
form of apparatus. The experiments gave a range in temperature
Fig. 5 Static Nozzle Corrected for Aspiration
of 150 to 300 deg. fahr. resulting in a maximum possible error of 6
per cent. For high velocities the fluctuation in temperature is neg-
ligible but for low velocities the range may be considerable.
y/,v//^//
^^Botli iSuifaces'
3 Pipe
cFiuished to liisiJe
/^/////^r--^^
Slots 10' X ]4'
, Flush
piam.of Pipe
- 3 Pipe
,\\\\\\\\\\\V- '^ t.wwwwwwv^
Y: Pipe Tap
Fig. 6 Device for Determining Aspiration Effect
15 (c) Capillarity in the gage glass increases as the diameter of
the glass decreases and may be considerable in tubes of small bore.
With a f-in. tube it amounts to 0.05 in., hence its influence is negli-
gible for high velocities.
16 (cO Aspiration in the static tube is appreciable only wit)i
60&
PITOT TUBE AS A STEAM METER
velocities above 6000 ft. per min. It may be entirely eliminated by
beveling the tube as indicated in Fig, 5. In this device the dynamic
and aspiration effects neutralize each other and only the true pressure
is recorded. The aspiration effect was determined by means of the
apparatus illustrated in Fig. 6. It consists of a special fitting con-
taining a chamber in communication with the main pipe, but so con-
structed that the velocity of flow in the chamber is practically elim-
inated. The difference in pressure between the two openings A and
B is due to aspiration. Further experiments are necessary to show
whether any fixed angle is applicable to all velocities.
17 A simple construction, as illustrated in Fig. 2, with ^-in. pipe
Separator
■w
Pressure Gauge (g)=
Steam Meter
-9 ft. long
Pressure Regulating
Dy Val««
Drip
To Condenser
■ Velocity Regulating
Valve
Fig. 7 Diagrammatic Arrangement of Piping for Testing Steam Meters
connections and |-in. gage glass, fitted with dynamic tube, as illus-
trated in Fig. 4, and static tube as in Fig. 5, is an accurate means of
indicating the true velocity of flow and the actual weight of water
discharged for all velocities above that corresponding to 1^ in. of
water. For lower velocities any error in reading is so large in pro-
portion to the entire head as to make considerable difference in results.
The scale may be graduated to read velocities in feet per minute and
the water rate in pounds per hour.
PITOT TUBE AS A STEAM METER
607
18 The results of a few tests of this simple device are given in
Table 1, and the arrangements of piping for conducting the tests is
shown diagrammatically in' Fig. 7. The results are also plotted in
Fig. 8. The curves give the weights as calculated from equation (7)
and the small circles the weights as determined from the condensed
TABLE 1 TEST OF SIMPLE GAGE GLASS METER
Steam PaEssuaE 70-lb. Gaoe. Steam Dbt. 3-in. Extra Heavy Pipe.
Velocities
Feet Per Minute
h
Weights
Pounds pes Hour
actual
MEAN
MAXIMUM
BT METEB
1
KATIO
differ- 1
actual meteb ebbob
GNCE
2i 2860
3J 3890
5J 4530
71 5360
91 6170
3528
4685
5480
6570
7405
1.23
1.20
1.21
1.22
1.20 I
1516 1545 ' 29 1.88
2062 2052 10 0.48
2400 2400 0 0.00
2852 2878 26 0.91
3270 3240 30 0.92
Average
1.21
1
r=<0.82. Coefficient of Meter — 0.82 X 1.21 = 0.992 or practically unity.
Equation for meter: V - 1996 V K and W ■= 1060 V*.
steam. It will be noted that coefficient q is unity and no calibra-
tion is necessary. Tests on 1-in., 2-in., 3-in., and 4-in. standaro pipe
gave a coefficient oractically of unity. The tests tend to show that
for a given size of pipe and a constant pressure and quality of steam
the actual mean velocity and weight of water may be accurately
determined by equations:
v= 139
i
(8)
w
= 58a i/ hd„ d.
(9)
This is strictly true for continuous flow only. Interrupted flow,
as in connection with reciprocating engines, creates a fluctuating
water column and it is difficult to obtain the mean readings. For
engines making over 100 strokes per minute the height of water is
practically constant but for lower speeds the fluctuation increases
with the decrease in speed. By suitably throttling the lower valve V,
water column H may be made to assume a fairly approximate mean
value for speeds as low as 20 strokes per minute in engines taking
steam full stroke.
608
PITOT TUBE AS A STEAM METER
19 The limitations of the simple gage glass meter for commercia
purposes are :
a It is purely an indicating device and readings must be taken
frequently to obtain average results.
h A scale graduated for a given set of conditions is accurate
/
/
' 1
10
i
1
/
/
/
/
/
9
/
j
/
/
l\
8
1
/
/
/
/
/
J
/
/
/
i
iJ
d
Ik
1
i
V
f
?/
#7
w
?i i
/ fi
to
bo
hi
1
■^ 3
" H'WA
If
t9W
4
f
r/
f
y
/
n
'/
2
/
0
/
/
v>
/
1
/
//
/
_ ^
lieo
reti
;al
y
^
y
' Actiial
^
k
1000 ]500 3000 2500 3000
Weight of Drj- Steam, lbs. per lir.
3500
P'iG. 8 Test of Simple Gage-Glass Meter, 3-in. Extra Heavy Pipe
only for these conditions, the degree of accuracy varying
with the fluctuation in steam pressure, change in quality
of the steam and variations in temperature of the water
column. A convenient form of chart for a given size of
pipe and for a wide pressure range is illustrated in Fig. 9.
The chart is wrapped around the drum, as in Fig. 10, and
set to correspond to the given pressure. The height of
PITOT TUBE AS A STEAM METER
609
Fig. 9 Type op Chart Giving Wide Pressukb Range fob Given Size of
Pipe
610
PITOT TUBE AS A STEAM METER
the water column transferred to the chart by a suitable
pointer gives at once the weight of steam flowing,
c Rapid increase in flow may cause the water in the gage glass
to be blown out requiring several minutes for sufficient
condensation to collect and balance the velocity head.
Fig. 10 Method of Mounting Chabt
d Inaccuracy for velocities below that corresponding to a
H-in. water column, or roughly, 2000 ft. per min. for pres-
sures over 70-lb. gage.
PITOT TUBE AS A STEAM METER
611
e Cannot be used for measuring the flow of highly superheated
steam except for practically constant flow and constant
degree of superheat. With highly superheated steam
a condensation chamber must be fitted to upper gage
cock C, Fig. 2.
20 Fig. 11 shows a modification of the simple gage glass meter
with simple pipe connection, which requires the pipe line to be tapped
in only one place. It may be set at any angle with the horizontal,
thereby increasing the sensitiveness of the readings.
Fig. 11 Gage-Glass Meteu WITH SiKGLii i'xG. 12 Gage-Glass Meter with Dry
Pipe Connection Dynamic Tube
21 Fig. 12 shows an application of the gage glass meter in which
many of the defects of the simple gage glass device are eliminated.
The error due to the variation in water level at the end of the dynamic
nozzle is entirely eliminated by making the tube a "dry" tube; i. e.,
water of condensation is not returned through the dynamic nozzle.
The temperature of the water in the gage glass is practically constant,
all condensation from the static end being discharged through drain
F to the chamber below, and all water carried over by the dynamic
end being discharged into pipe P directly to overflow G. The area
612
PITOT TUBE AS A STEAM METER
and volume of chamber M is so large compared with that of glass W
that sudden variations in flow do not materially affect the level of
water in M and cannot blow the water out of glass W. The only de-
fect in the instrument is the error due to capillarity in glass W, which,
as stated before, amounts to but 0.05 in. for a |-in. tube.
22 The operation is as follows: Velocity pressure is transmitted
through tube D and opening 0, into the body of chamber M. This
pressure acting on the surface of the condensation in the chamber
forces water into glass W until a balance is effected. Condensation
is discharged continuously through pipe P and water seal U of the
Fig. 13 Meter wjth Self- Adjusting AVater Column for Low Velocities
main pipe. Tests of this meter gave practically theoretical results
for all velocities ranging from the equivalent of a ^-in. water column
to 10-in.
23 Fig. 13 shows an apphcation of the self-adjusting water column
for very low velocities, 2000 ft. per min. and under. Glass ball B
rises and falls with the water column, transmitting its motion through
levers L and N, gear sector G and pinion P, to permanent magnet, M.
Th.e latter transmits its motion to magnet M' outside the casing.
PITOT TUBE AS A STEAM METER
613
Pointer C is fastened to magnet M' . Thus any motion of ball B is
multiplied by pointer C and indicated on dial K. Magnets M and
M' are independently mounted on pivot bearings, M within the casing
and subjected to steam temperature, the other outside the casing.
Motion of M is transmitted magnetically to M' through the casing,
thus doing away with stuffing boxes. The relative positions of mag-
nets and casing are illustrated in the lower corner of Fig. 13. On
account of the angularity of the connecting links and the frictional
resistances, small as they are, the dial graduations cannot be con-
veniently calculated but must be calibrated by experiment.
24 This instrument is very sensitive, indicating velocity changes
of 200 ft. per min. When connected directly to the pipe it is subject
~(~\
PI
',
N
N
B
\ '
^
/
^ :~
-\J
0
Fig. 14 Indicating Impulse Metek
to all of the errors of the simple gage glass meter and is altogether top
sensitive for accuracy.
25 All of the devices described above are simple indicating mech-
anisms, and with the exception of the one illustrated in Fig. 13, can-
not conveniently be made autographic.
26 For commercial purposes a steam meter should be autographic,
or better still, integrating. The ideal meter is one which shows at a
glance the weight of steam flowing for any given period; ofltime and
which may be read as one does a watt-hour meter.
27 Fig. 14 shows an application of the Pitot tube for indicating,
autographically, the weight of steam discharged and differing basic-
ally from those just described. A permanent magnet A^ forms the
spoke of a small aluminum wheel A . Rotation of wheel A is resisted
614 DISCUSSION
by spiral spring G. D and S are dynamic and static nozzles, respec-
tively, of a Pitot tube. The velocity head discharges a small jet of
steam through nozzle M and exerts a force on the periphery of wheel
A, tending to rotate it about its axis. The angular rotation increases
with the velocity. The motion of wheel A is imparted through the
medium of magnets A^ and A'"' to pointer R. By means of a suitable
clock-work the angular movement of wheel A may be autographically
transferred to a chart giving a continuous record of the weight of
steam flowing.
28 By permitting the wheel to rotate and by connecting magnet
N' to a series of rotary dials an integrating or total output mechanism
is readily effected.
29 Experiments are now being conducted with the autographic
and integrating devices just described, but sufficient data are not yet
available as to their respective merits.
DISCUSSION.
Prof. W. B. Gregory. The author has apparently developed a
practical instrument of real value. However, it seems to the writer
that the device for determining aspiration effects can not be relied
upon to make determinations of any value.
2 I would like to ask Professor Gebhardt if he has used static
openings about 1/16 in. in diameter drilled at right angles to the axis
of the pipe? Extensive experience with the Pitot tube as a device for
measuring the velocity of water has taught me to avoid irregulari-
ties in a pipe, due to special fittings or other causes, when the static
pressure is taken from the walls of the pipe. An unobstructed length
of straight pipe is absolutely essential to accurate work.
3 The desirability of finding the correct static pressure is apparent
as it seems probable that one constant would apply to reduce velocity
at the center to mean velocity, in any and all sizes of pipe. The
experimental deteimination of the correct angles for the static nozzle,
as shown in Fig. 5, would then be avoided.
Walter Ferris. The remarks of Professor Gregory in regard to
the special fitting for finding the effect of aspiration remind me
forcibly of an experience a few years ago with both a Venturi meter
and a Pitot tube for measuring water. Perhaps the coaclusions at
which I arrived at that time may be suggestive, although possibly not
of direct applicatioa in the case of a steam meter.
PITOT TUBE AS A STEAM METER 615
2 Until quite recently, that is, within a few years, I think it
has been assumed that it was necessary, in the use of the Pitot tube, to
have a static tube close to the dynamic tube, or at least at the same
distance from the walls of the conduit. I believe that William Monroe
White, six or seven years ago, made some experiments demonstrating
that the velocity head taken from the impact side of a pitot tube is
correct, whatever the shape of the nozzle, so long as it is a surface of
revolution. Thus the nozzle may be either cylindrical, or a converg-
ing or diverging cone, and the dynamic head will be correctly indi-
cated, any variations in the coefficient of the pitot instrument as
a whole being due to the shape or location of the static opening.
3 In the Venturi meter, we find that the static pressure is always
taken from the walls of the conduit, where the velocity may not be
over half the maximum velocity, and yet the results from the Venturi
meter are invariably correct to within one per cent, if conditions are
favorable. Therefore a dynamic nozzle, which is a surface of revo-
lution, combined with a static nozzle terminating in the wall of the
conduit (as in the venturi meter) should together form a Pitot instru-
ment which is correct to the formula, and needs no calibration. This
seems to indicate that for a Pitot instrument to measure the flow of
water it is not necessary to take the static head and the dynamic head
in regions of the same velocity, and that the true average static pres-
sure will be indicated through intervening velocities, and correctly
registered, even when the piezometer is located in a region of low
velocity. From this I infer that in this steam meter sufficiently
small static openings in the true smooth wall will probably give cor-
rect results as they do in the water meter, although I have no ex-
perimental data with which to confirm this opinion.
A. R. Dodge. I would like to take exception to a statement in Par.
6: "On account of the great density of mercury and the variation in
height of the condensed vapor above the mercury, this application of
the Pitot tube has very little value scientifically or commercially. "
The General Electric Company has developed a steam meter, both of
the indicating and recording types and hrs built several hundred of
these meters using mercury and condensed vapor above the mercury.
This condensed vapor automatically remains at a constant head.
2 Recently three recording meters, selected at random out of a
lot of fifty, showed a maximum error of less than two per cent. Ninety
per cent of the readings were within one per cent on the three meters,
which had an automatic pressure correction and also a temperature
616 DISCUSSION
correction. These meters can be used on any size of pipe, from 2 in
up to 36 in., the 36-in. pipe, of course, being for atmospheric conditions
of steam. These steam meters we have foimd to be valuable in
improving the consumption of steam in our various plants.
3 We have also experimented with several of the types described
in this paper in which mercury is not used and have found them excel-
lent in many respects, but the use of mercury is not at all objec-
tionable.
The Author. Static openings, about one-sixteenth of an inch in
diameter, drilled at right angles to the axis of the pipe, showed no
aspiration effects at velocities up to 15,000 ft. per min. (the maximum
obtained during the tests) but are unsuitable for the appliances de-
scribed. It is the author's intention to develop a simple meter which
can be constructed of standard fittings and which may be attached
by tapping the pipe in the ordinary way. Such an application neces-
sitates the projection of the static nozzle beyond the inner surface
of the pipe, an arrangement which causes serious aspiration. With a
standard ^-in. nipple projecting i in, beyond the inner surface of a
3-in. pipe an aspiration effect corresponding to 10 in. of water was
noted at a velocity of 12,000 ft. per min. (pressure 100-lb. gage). At a
velocity of 6000 ft. per min. the aspiration amounted to li in. of
water. It was for the purpose of neutralizing this aspiration that
the static nozzle was cut at an angle, as indicated in Fig. 5.
2 Mr. Ferris' remarks are in accordance with experiments con-
ducted by the author, but, as stated above, a static opening terminat-
ing with the inner wall of the conduit is not applicable to the instru-
ments in question. Fig. 1 illustrates such a static opening, but in
the actual construction the nozzle projected i in. beyond the inner
surface.
3 Mr. Dodge's experiments with the use of mercury as an indi-
cating medium are of considerable interest, in that they show the
development of a practicable and accurate steam meter which is
little known to the general engineering public.
No. 1256
EFFICIENCY TESTS OF STEAM-TURBINE
NOZZLES
Bt Prof. Frederick H. Siblet, University, Ala.
Member of the Society
T. S. Kemble,* Cleveland, O.
Non-Member
In 1905 a series of tests was begun at Case School of Applied Science
Cleveland, O., to determine the proper proportions and the efficien-
cies of steam-turbine nozzles for given steam conditions. The final
tests, from which the results given herewith are derived, were made
in an appaiatus designed by T. S. Kemble, of the Chase Machine Co.,
of Cleveland. The writers spent about two years on these tests,
and through the generosity of the company and the facilities afforded
at Case School, were able to procure apparatus of considerable pre-
cision.
THEORY OF NOZZLES
2 The nozzle should be so constructed that the expansion of the
steam will take place between the given initial and terminal pressures
and wholly within the nozzle, the steam filling it completely at all
sections. The available heat energy of the steam less that lost by
friction will then be all converted into kinetic energy and the effi-
ciency will be a maximum. By efficiency we understand the kinetic
energy of the jet per unit mass, divided by the available heat energy
of the steam per unit mass.
3 To find the correct relation between the length and the ratio
of thrust to muzzle area of the nozzle, that the efficiency may be a
maximum, is the problem to be determined experimentally. Let
' T. S. Kemble, Experimental Engineer for the Chase Machine Company,
Cleveland, 0.
Presented at the Annual Meeting, New York, (December 1909,) of The
American Society of Mechanical Engineers.
618 TESTS OF STEAM-TURBINE NOZZLES
W = weight of steam flowing in pounds per second.
V = velocity of the jet in feet per second.
M = mass.
F = reaction in pounds.
g = acceleration due to gravity at Case School = 32.16015
K = kinetic energy of the jet in foot-pounds or B.t.u.
E == total available heat energy of the steam.
Efficiency = K ^ E.
4 From Mechanics we have: K = hMV^ and M = W -^ g. If
W == 1 lb., then
K (the B.t.u.) - 2,^
also
V = -^ for any flow
F and W are the factors to be determined by experiment.
METHODS SUGGESTED
5 Three methods may be suggested for determining the efficiency
of steam nozzles.
a By measuring the force of the jet when allowed to impinge
on an external surface.
h By investigating the character of the jet with a search tube
inserted axially in the nozzle.
c By measuring the reaction of the nozzle when a jet of steam
is flowing through it.
6 The first method involves complications which tend to cast
some doubt upon the results obtained. The force upon the external
surface may be modified by the character of the surface, by eddying
and steam friction, and by the distance traveled by the jet after
leaving the nozzle and before it reaches the surface. When the sur-
face used is a flat plate perpendicular to the axis of the jet, the force
may even vary from a maximum to a negative value according to the
relative location of the plate and nozzle.
7 The second method was tried in a series of experiments to find
the pressure in the jet at various sections of the nozzle. A search
tube was inserted axially in the nozzle and the relation between pres-
TESTS OF STEAM-TURWINE NOZZLES
619
sure and flow was compared with the theoretical relation, as calculated
and plotted from the steam tables. The first search tube used was
iV in. in outside diameter and ly'i in. in inside diameter. One end was
closed and the other connected to a mercury column. A rh-'m. hole
was drilled through the search tube at right angles to its axis. The
Fig. 1 Diagram of Piston Apparatus
readings with this tube varied greatly under identical conditions and
a tube with a larger bore was tried, which gave results much more
consistent but not nearly accurate enough to determine the efficien-
cies at the various sections of the nozrde.
8 The third method provided for the determination of the reac-
tion of the jet in the nozzle, and apparatus was constructed for this
purpose, differing in detail as follows:
a By fastening the nozzle into the outer face of one of a pair
of rigidly connected pistons suspended in a cylinder.
Steam entering between the pistons and flowing out
through the nozzle would produce a measurable reaction.
620
TESTS OF STEAM-TURBINE NOZZLES
b By using a flexible steel tube suspended freely by one end
and having the nozzle attached to a chamber at the other
end with its axis perpendicular to the axis of the tube.
Steam flowing downward through the tube and out of the
nozzle would cause the tube to deflect with a measurable
force.
PISTON METHOD
9 Difficulty was experienced with the piston apparatus, diagram
of which is shown in Fig. 1, due mainly to the static friction of the
Fig. 2 Flexible Tube Apparatus for Measuring Reaction of Jet
pistons when under pressure and to leakage past thepistonswhenthe^
were reduced in diameter to lessen the friction. Modification of this
apparatus to overcome these difficulties would have required almost
complete rebuilding and the flexible-tube apparatus was therefore
tried and with such success that it was not considered necessary to
return to the old one.
FLEXIBLE TUBE APPARATUS
10 The apparatus shown in Figs. 2 to 4 was designed with a
special view to combining the search-tube and reaction methods, by
TESTS OF STEAM-TURBINE NOZZLES
621
which means we hoped to obtain results more accurate than would
result from either method alone.
11 The tests were conducted under a pressure of 155 lb. gage.
There was about 50 deg. fahr., superheat at the boiler, which was con-
nected with the testing apparatus by about 55 ft. of 5-in. pipe and
25 ft. of 4-in. pipe.
12 In the main steam pipe was located an angle needle-valve
operated by a sprocket wheel and chain which made it possible to
hold the nozzle feed pressure very constant. From this point the
Fig. 3 Interior of Box Containing Flexible Tube Apparatus
steam passed downward through the tube A to the chamber 5, (Fig.
4) thence through the nozzle C into the box D, and on through the
passage E to the condenser.
13 The upper end of the tube A was screwed into a diaphragm
on the lower flange of the angle valve. At its lower end it supported
the chamber B which was allowed to move between stops restrained
only by the stiffness of the tube and of the spring F. The motion of .
the chamber B was indicated by a needle which multiplied the motion
about ten to one. The spring ¥ was operated by a micrometer nut
and screw and was calibrated in 'place by^known weights hanging on a
flexible wire cable which extended from the back of the chamber in the
622
TESTS OP STEAM-TURBINE NOZZLES
line of the nozzle axis and down over a ball-bearing sheave. The
tube A , the^chamber B and the nozzle, were all enclosed in the ver-
tical pipe P and the box D, and the vacuum surrounding them wa&
greater than that in^the condenser owing to the "augmenter" action
of the steam jet entering the passage E to the condenser.
iSRi
Fig. 4 Section op Flexible Tube Apparatus
14 The initial temperature of the steam was Fhown by a thermom-
eter inserted in a well in the chamber B and observed through a glass
in the box D. A steam gage was connected by a flexible tube to the
chamber B. The vacuum in the condenser and in the box D was
shown by mercury columns, and another column was joined by a
flexible connection to a hole drilled as near as possible to the muzzle
of the nozzle and perpendicular to the wall. All these connections
TEST? OF STEAM-TURI5INE NOZZLES
623
to the mercury columns were of glass tubing with rubber couplings
which allowed the moving parts to swing freely without friction and
made it easy to observe any accumulation of moisture. By break-
ing the connection air could be let through to dry them quickly.
FORMS OP NOZZLE TESTED
15 The exact dimensions of the nozzles are shown in Fig. 6 and
Fig. 7. They were all of machinery steel, bored taper, and had
entrances rounded off with a hand tool to approximately |-in. radius.
All nozzles except No. 15 and No. 18 were bored smooth and polished.
Nos. 10, 11 and 12 were identical except as to length and angle of
divergence. No. 18 was like No. 11 except that it was finished rough
on the inside between the throat and muzzle, the finishing chip being
Fig. 5 End Views of Search-Tube End and of Condenser End
made with a threading tool having an angle of 120 deg.,the feed being
90 turns per inch. No. 14 and No. 15 were identical except that
whereas No. 14 was bored taper. No. 15 was made in halves, and after
being bored taper these halves were separated and milled longitu-
dinally with a 90-deg. cutter, the cut beginning just beyond the throat
and running deeper toward the muzzle, where^the section became
square, with the same area as the muzzle of No. 14. No. 16 was hke
No. 14 but had a larger throat area so that a needle point could be
introduced to give the same net area as No. 14. No. 9 and No. 13
are search-tube nozzles made with throat and muzzle areas large
enough so that the net areas, with the search tube in place, were equal
to the net areas of the corresponding "reaction" nozzles; that is the
nozzles whose efficiency was determined by reaction alone without
the use of the search tube. No. 9 corresponds to Nos. 10, 11, 12 and
18. No. 13 corresponds to Nos. 14, 15, and 16. The dimensions
of the nozzles are given in Tables I and 2.
624
TESTS OF STEAM-TURBINB NOZZLES
FLOW TESTS
16 Numerous tests were made to determine the rate of flo., jt
steam through the various nozzles. Fig, 8 shows the results of these
tests plotted to a scale of pounds flow per hour. The variations in
flow are probably due principally to moisture in the steam, and to
Noz. No.lO
Fig. 6 Fokms of Nozzles Tested
some extent to leakage from the water to the Eteam side of the con-
denser. The condenser was tested and at no time showed a leak
exceeding two pounds per hour. There was sometimes a trace of
superheat at the nozzle entrance and thk incieased with an increase
in the volume of flow. At pressures of less than 145 lb., moisture
was probably always present. For tb.s reason the values used in the
TESTS OF STEAM-TURBINE NOZZLES
625
calculations were the mean flow values for 145-lb. pressure, and a
trifle less than the mean for the lower pressures. It is to be regretted
that we were unable to procure a calorimeter of sufficient accuracy
for our purpose.
TABLE 1 DIMENSIONS OF NOZZLES IN FIG. 6
Nozzle
No.
Gross
Diameter
Throat
INCHES
Gross
Diameter
Outlet
INCHES
Net Area
Throat
6Q. INS.
Net Area
Outlet
EQ. INS.
Length
inches
Angle of
Divergence
13
14
15
0.3949
0.3038
0.3038
0.625
1.156
1.128
1 IN. SQ.
1.1315
0.0734
0.0725
0.0725
0.0725
1.0005
0.9993
1.0000
1.0055
3A
2H
2H
3
14° 30*
16° 18'
13° 37'
9° 41'
Nozzles Xo.lO, Xo.ll, Xo.l2, ami Xo.lS
Fig. 7 Forms of Nozzles Tested
TABLE 2 DIMENSIONS OF NOZZLES IN FIG. 7
Nozzle
No.
Gross
Diameter
Throat
inches
Gross
Diameter
Outlet
inches
Net Area
Throat
SQ. IN.
Net Area
Outlet
BQ. IN.
Length
INCHES
1
Angle of
Divergence
9
0.3940
1.4505
0.0728
1.6033
6gJ
1 9° 26'
10
0.3039
1.4240
0.0726
1.5926
6i
9° 53'
11
0.3039
1.4231
0.0725
1.5907
5i
11° 41'
12
0.3056
1.4241
0.0733
1.5930
3A
20° 31'
18
0.3039
1.4225
0.0725
1.5893
5H
11° 36'
626 TESTS OF STEAM-TURBINE NOZZLES
17 In Fig. 8 the results are given in pounds per hour for the four
initial pressures. Each small circle represents the result of one flow
test of from 15 min. to 30 min. duration. The vertical dotted lines
represent the flow values that were used in the efficiency calculations.
The flow values for the nozzles of Fig. 6 are a little higher than
the others, as is shown in the upper part of the diagram. The
diagonal lines simply connect together the results found in the same
test. For example, the five circles along the lowest line of the chart
represent the values found for nozzle No. 9 on January 17 and IS,
1908. The vertical position of the observation point is of no signifi-
cance, each initial pressure being placed higher up on the diagram
than the preceding one, as a matter of convenience.
SPRING CALIBRATIONS
18 The accuracy with which the reactions were determined
depended largely on the care taken in calibrating the springs. These
calibrations were first made with the spring in place in the apparatus,
and at the room temperature of about 85 deg. fahr. The readings
were taken while gradually loading the spring to 25 lb. and then un-
loading it, repeating the operation a great number of times and taking
the average extension under any given load as the true extension for that
load. The average extension under a load of 25 lb. was 1.4875 in.,
or 0.0595 in. per lb. In cooling off, tube A, Fig. 4, seemed to warp a
little, so that after about two minutes there was a decrease in the
initial extension of the spring of about 0.0025 in. As it took about
two minutes to shut off the steam and get the initial extension after
each reaction reading, the spring extensions were all corrected by
this amoi/nt.
19 Another factor which affected the spring calibration was the
change in temperature of the spring itself. A thermometer was in-
serted in the spring casing and the spring calibrated at various tem-
peratures by observing the temperature and extensions simultane-
ously. From these temperature calibrations a correction factor
I [0.0002428 (^2 ~ ^i)] was obtained and used to correct there actions
found by using the factor 0.0595 in. per lb.
20 After the thermometer had been inserted in the spring casing
it was noted that as the reaction test progressed the temperature at
first increased and then remained nearly constant regardless of mod-
erate changes in initial and terminal pressures. This was due to the
fact that the box D stood open and cooled off between tests and then
TESTS OF STEAM-TURBINE NOZZLES
627
•sqyni-bs .indfn (:n = dl
628 TESTS OF STEAM-TURBINE NOZZLES
warmed up gradually when the steam was turned on. The average
temperature in the spring casing was about 135 deg. fahr., or 50 deg.
higher than the room temperature at which the original calibration
was made. All the reactions found before this thermometer was in
place were corrected on the assumption that the spring temperature
was 135 deg. fahr. While this did not eliminate the error due to the
fact that the temperature increased during the first part of each test,
it did bring the average quite close to what it should be.
21 Corresponding readings of spring extensions at the beginning
and end of tests were lower and higher respectively than the average
extension. This was due to the above-mentioned difference in spring
temperature. When the spring temperature was read simultaneously
with the spring extension this difference disappeared.
SEARCH-TUBE TESTS
22 After completing the flow tests and the spring calibrations one
other factor remained to be determined before the reaction tests could
be made. This was the determination of the pressure at the muzzle
of the nozzle.
23 The reaction of any nozzle is equal to the summation of all the
components, 'parallel to its axis, of the pressures within the nozzle
and in the chamber from which it leads. If the pressure of the medium
surrounding the nozzle and the chamber is equal to that in the plane
of the muzzle, then the reaction as shown by the pull on the spring
is the true reaction. If the pressure of the surrounding medium is
greater than that in the plane of the muzzle it will decrease the appar-
ent reaction, and if the pressure of the surrounding medium is less
than that in the plane of the muzzle it will increase the apparent reac-
tion. The amount of such increase or decrease will be equal to the
difference in the unit pressure multiplied by the area of the muzzle.
The true reaction of the nozzle is equal to the pull of the spring plus
or minus this pressure difference.
24 The demonstration of this propositiop possibly differentiates
these experiments from those heretofore published, as the writers
do not know of any other case where the combination has been used
in this manner.
25 The muzzle pressure was found by using the search tube with
nozzles No. 9 and No. 13. The search tube here used was a selected
piece of cold-drawn Shelby tube } in. in outside diameter and A in.
in inside diameter, with six holes ^ in. in diameter drilled perpendicu-
lar to the axis. The outside of the tube was polished to micrometer
TESTS OF STEAM-TURBINE NOZZLES 629
measurement. The chamber B was rigidly connected to the back wall
of the box D by the distance piece at J (Fig. 4). The rear end of the
search tube was encased and supported by a tube at L which had on
its outer surface a thread fitted with a micrometer nut, and passed
through the distance piece holding the search tube in the axis of the
nozzle.
26 The holes in the search tube were located in the same plane as
the hole in the wall of the nozzle. One gage was connected to the
box D. another to the hole in the wall of the nozzle, and a third to the
rear end of the search tube. Simultaneous readings of these gages
were taken with varying pressures in the box. These readings were
plotted in Fig. 9 and Fig. 10.
EXPTANAnON OF RESULTS, FIGS. 9 AND 10
27. The diagonal lines represent the box pressures, or external
terminal pressures which were varied from 0.5 lb. to 2 lb. absolute.
The circles represent the pressures at the rim of the nozzle and the dots
those at the center of the nozzle. These are connected by vertieal
lines as an aid in following the chart. The results are plotted for
the four initial pressures as fcund en different dates. The dotted
horizontal lines represent the muzzle pressures used in determining
the true reaction. For example, on May 22 and 23, with an initial
pressure of 100 lb. under a box pressure of O.S lb. absolute the terminal
pressure at both rim and center of the nozzles was 0.7 lb. absolute •
0.649 was the average terminal pressure used in the calculations.
28 To illustrate further, the data for nozzle No. 1 3, April 15 and
16, 1908, with an initial pressure of 145 lb. per sq. in., are as follows:
Rox pressure 0.384 0.84 1.33 1.67 1.71 1.81 1.995 2.315 2.80
Rim pressure 1.665 1.67 1.665 1.67 1.6S 1.685 1.725 1.865 2.23
Search pressure 1.71 1.715 1.72 1.705 1.705 1.72 1.725 1.73 1.73
The three pressures in each column vvere read simultaneously and re-
corded.
29 The box pressure into which the nozzle discharged is repre-
sented arbitrarily by the diagonal line AB, in Fig. 10. The point c
on this line shovvs the box pressure 0.84 lb. per sq. in. given in the
table.
30 The corresponding rim pressure 1.67 lb, per sq. in., (that
given h-y the gage connected vvith the hole in the v all of the nozzle
at the muzzle) is plotted as a circle vertically above the point.
630 TESTS OF STEAM-TURBINE NOZZLES
31 The corresponding search pressure 1.715 lb. per sq. in.,
(that obtained by means of the search-tube \\ith the holes in the same
plane as the hole in the wall of the nozzle) is plotted as a dot vertically
above the point c.
32 The first tests seemed to indicate a higher pressure in the center
of the stream than in the rim. Later a leak in the search-tube con-
nections was discovered and repaired and the tests repeated, vvith the
result that there was practically no differences between the press-
ures at the rim and center of the jet. It was therefore assumed
that such differences as had occurred were due to the leaky search-
tube. As it was difficult to keep the search-tube connections per-
fectly tight, it was decided to use the rim readings.
33 The tests from which Fig. 9 and Fig. 10 are plotted show that,
after the leak in the search-tube had been eliminated, the pressure
at the rim and that at the center of the stream did not differ by an
appreciable amount. They also show that the pressure within the
nozzle did not vary as the external pressure was decreased below that
in the muzzle of the noazzle, and that as the external pressure was
raised above the muz -Ae pressure, it crept in gradually along the wall
of the nozzle, but did not affect the pressure at the center of the stream
within the limits shown.
REACTION TESTS
34 In making reaction tests the following routine was observed:
The barometer was read with every observation, and corresponding
corrections made so that the initial absolute pressure used should
be 145, 130, 115 and 100 lb. per sq. in. This was for convenience
in making computations. One observer maintained a constant
initial pressure by manipulating the needle valve in the steam pipe.
A second observer maintained a constant vacuum in the box D by
manipulating the valves leading to the pump and condenser. A
third operated the micrometer screw which registered the spring ex-
tension and thus held the multiplying-needle opposite an index at
the center of its travel. A fourth and sometimes a fifth man read
gages, and one man was generally occupied in moving about behind
the observers to check observations.
35 As the reaction of the jet forced the chamber B back against
the tension of the spring, the micrometer screw was worked forward
until the tension in the spring balanced the reaction and the multi-
plying-needle indicated that B was swinging freely in the central
TESTS OF STEAM-TURBINE NOZZLES
631
o^niosqv •ui-bg -qi -j-x
ei
tA
r-
oo
Oi
■r-'
rt o
(TJ
■rA
r4 0
o
1
t
-,"
-T
o
o
"1
«'§
l-i
■1
-3^
<\°
?i
JO
,
iil
51
-
6
1
ii-
^
-J
1
\,
111-
\^
\,
t'l
\
1
1—
1
-
s
P-I
c^l
\
(ii
\
11
3R
\
"
i^i
.\,
h-ti
\
■H 1
\
k
1
"
\
S
s
s
1
^
\
\
\
\
"^
\,
1
\,
\,
\,
1
§1
s
s
s
\
S|
\
^
k
3
^
1
\
h
^
•^
f
^
r-
A
^ —
^
— 1
1
\
1
\
'
1
s
^
1
S
s
\
1
1
o
\,
1
•C
\
,1
1
c
\
1
i2
s
■O-l
1
■^
''^'
1
<
%
1
Tn
1
1
1
\
\
\,
\^
1
\
\
L"^
\
\
S
\
\
1
\
'^
s,
\
s
1
\
N^
—
—
-^
\
^
.1
\
■
\
=
\
\,
1
OT
\
^
pi
k
\
:3
K
d
S.
\l
r
^
^
•
i^
kJ-
A^
^
A
;
N
1
F
1
'o'
1
1
o
Q
o m
N ^
§ W
to p
Ph
h3
<!
H
o
H
e« r-l ,-1
— -,-^77^ o
ajniosqv •ui-T)S qTf-J-X o
fa
632
TESTS 0¥ STEAM-TURBINE NOZZLES
•sqv ■q-x_'iiotint!0'a
It
1
7
7
o
1
/
/
OJ
1
/
OS
1
1
'-'
1
/
/
CO
1
/
/
^
1
/
/
t~
~^
/
1
eo
^
'-'
/I
-,.
<
«5
/ 1
--,
1
a
<^
■*
^
1
b
L
Li
tH
■ /
1
/
H!;;;r
-^
n ra
'?)'
1
J
JZ^'J^.^-
^-^
5/
1
il
1
■-^
]
S
•-1 OJ
-7.
\
'/
s5
&
i
"
■■■-,
^
&
ti^
i
Si
^1
A
fli
.
u "^
?r
it
f
c
^■/^i
f,
"-,
^
S o
s/
3]
il
J3
^/i
\
j; T-J
^
^1
>
S
B
:;/4i
:?/'
X -5
—
-n
ii
R
" i
-3n 2 1
.=^/s;
ffl ^
i-ii
S
^
1
■-'7 111
.=^/.il
GO
^/
"^1
^1
d
tJ ^- i
^i-^i
■ o
■7
SI
^1
4i
4/ ^i
l-
^1
p
/
^
s
0 ,
// ^
o
■n'
§,
Si
^i
si
i31
«o
/
r,!
%
1
3^
"!
—
^
o
/
t-
^/
§
o
/
61
i 0
0
.?
21
1/
yi
i
ci
G
»
^.
^i
J
S|
■*
1
^"
^
^1
r
«f
o
1
/
/
1
1
•3I
it
CO
1
/
/
]
o
1
/
/
1
/
r*?'
(N
1
/
/
1
■n'
o
^
c3i
i
^1
'
^L
T-l
i^:
/
""il
si
l\
3:1 1 i 1
^ l-J
(H
n
a
0
u
w
Q
u
J^
t>
H
Q
0^
n
CO
tti
t3
Ph
CO
„
m
u
<!
h1
^
M
0
b)
^
N
N
ta
0
0
^
a
^
(VI
tM
P
0
m
w
^
H
«
Ph
s '^
H
sqv •qT-tioij-i'Baa
TESTS OF STEAM-TURBINE NOZZLES
633
position. When the needle remained quiet for an instant in this
position the observer gave a signal and the thermometer and gages
were read simultaneously. After each reading the steam was shut
, off, and the position of the micrometer screw with the needle in the
central position was noted. The difference between this position and
that when the steam was flowing gave the elongation of the spring
due to the reaction.
1^20 and 2l/-07 ife and 2^ 'OS 2/ll aad IS'US 2;is and 19,'-()8
2/21 and 22/'08
Fig. 11 Preliminary Reaction Tests
36 It will be noted that friction has been entirely eliminated in
this apparatus, except for the trifling amount due to the movement
of the multiplying-needle.
CALCULATION FOR EFFICIENCY
37 A series of preliminary reaction tests was run on all of the noz-
zles with the result that nozzles No. 11 and No. 14 were selected as
634
TESTS OF STEAM-TURBINE NOZZLES
representative for future tests. The results of the preliminary tests
upon these nozzles are given in Fig. 11. Exhaustive tests of long
duration were then made upon nozzles No. 11 and No. 14 with the
results shown in Fig. 13 and Fig. 14, and in Tables 3 and 4. Fig.
12 shows the comparative results of certain of the tests.
38 In Fig. 11 and Fig. 12 the reactions in pounds absolute are
grouped together horizontally. The reaction? are shown for each
of the four initial pressures. The vertical scale is four lines to the
pound. The horizontal position of the points has no significance, the
4/2i and 25/'08 4/27/'08
Fig. 12 Comparison of Tests on Nozzles 14, 15 and 16 with those on
Nozzles 11 and 18
initial pressures being placed diagonally over each other as a matter
of convenience. The diagonal lines connect the reactions of a single
nozzle at the various pressures. The arrowheads indicate whether
the reactions shown on that line were taken when the pressure was
increasing or decreasing. The circles represent the actual reactions
in pounds as plotted from the tests. Two circles occurring together
indicate that two independent readings of the reaction were taken at
the same time.
TESTS OF STEAM-TURBINE NOZZLES
635
39 Fig. 13 shows graphically the result of a complete series of
reaction tests on nozzle No. 11, which is practically the same as noz-
zlesNos. 9,10,11, 12andl3. The vertical scale is 4 lines to the pound,
while the horizontal scale is 20 lines to the jDound. The full diagonal
lines connect together the observed reactions in pounds under a
TABLE 3 COMPUTATION OF EFFICIENCY. NOZZLE NO. 11
I. P.
Abb.
145
145
145
145
145
T. P. Lb.
PER Sq. Reaction I
In. Abs.
0.929
1.029
0.829
0.929
0.929
18.134
17.974
18.294
18.134
18.134
Flow
Lb.
FEB Hr.
553
553
553
558
548
Fm)w
Lb.
PER Sec.
0.1536
0.153b
0.]53a
0.1550
0.1522
Vel.
3796
3763
3830
3762
3831
B.T.U.
288.0
283.0
293.1
282.9
293.3
B.T.U.*
Table
317.4
312.5
322.7
317.4
317.4
Efp.
%
90.75
90.55
90.84
89.13
92.41
145
145
130
115
100
0.929
0.929
0.832
0.735
0.638
18.234
18.034
16.244
14.351
12.45
553
553
498
442
385
0.1536
0.1536
0.1383
0.1228
0.1069
3817
3776
3776
3759
3744
291.2
284.9
285.0
282.4
280.1
317.4
317.4
315.4
313.6
311.5
91.75
89.75
90.36
90.03
89.91
100
100
100
100
100
0.738
0.538
0.638
0.638
0.638
12.29
12.61
12.45
12.45
12.55
385
385
390
380
385
0.1069
0.1069
0.1083
0.1056
0.1069
3696
3792
3696
3793
3774
273.0
287.4
273.0
287.5
284.6
304.6
319.5
311.5
311.5
311.5
89.63
89.94
87.62
92.29
91.36
100
0.638
12.35
385
0.1069
3714
275.6
311.5
88.47
Assuming 2 per ceat moisture
145
0.929
18.134
553
0.1536
3796
288.0
311.9
92.35
130
0.832
16.244
498
0.1383
3776
285.0
310.0
91.93
115
0.735
14.351
442
0.1228
3759
282.4
308.2
91.62
100
0.638
12.45
385
0.1069
3744
280.1
306.1
91.51
B.t.u. are given only to the nearest tenth, and for this reason efficiencies are not accurate in
second decimal place.
B.t.u. 1 = equivalent of kinetic energy of jet in B.t.u.
B.t.u.; => available heat energy of steam.
varying box pressure. The horizontal dotted lines connect together
the same reactions after being corrected for the difference between
the pressure in the box and in the muzzle of the nozzle. The dotted
diagonal line represents the pressure in the muzzle of the nozzle as
found with the search tube and is plotted to the same horizontal
scale as the box pressure. When the box and terminal pressures
636
TESTS OF STEAM-TURBINE NOZZLES
are the same the apparent and corrected reactions are the same, as
is shown by the intersection of the three lines. As the box pressure
increases above that point where the terminal and box pressures
were the same the apparent reaction falls below the true reaction,
and as the box pressure falls below that point where the terminal
TABLE 4 COMPUTATION OF EFFICIENCY, NOZZLE NO. 14
Showing Effect of Ebbor in Determination of Terminal Pressure Flow or Reaction
LP.
Abs.
T. P. Lb.
PER Sq.
In. Abs.
Reaction
Flow
Lb.
PER Hr.
Flow
, Lb.
PER Sec.
Vel.
B.T.U.l
B.T.U.2
Table
Eff.
%
145
1.632
17.821
558
0.1550
3698
273.2
289.8
94.28
145
1.732
17.721
558
0.1550
3677
270.2
286.7
94.23
145
1.532
17.921
558
0.1550
3718
276.3
293.0
94.30
145
1.632
17.821
563
0.1564
3665
268.4
289.8
92.61
145
1.632
17.821
553
0.1536
3731
278.2
289.8
95.99
145
1.632
17.921
558
0.1550
3718
276.3
289.8
95.34
145
1.632
17.721
558
0.1550
3677
270.2
289.8
93. 2 J
130
1.46
15.977
502
0.1394
3685
271.3
288.2
94.15
115
1.288
14.147
446
0.1239
3672
269.5
286.5
94.07
100
1.116
12.295
389
0.1081
3659
267.6
284.7
93.99
100
1.216
12.195
389
0.1081
3630
263 .'3
280.3
93.92
100
1.016
12.395
389
0.1081
3689
272.0
289.3
94.01
100
1.116
12.295
394
0.1094
3613
260.8
284.7
91.62
100
1.116
12.295
384
0.1067
3707
274.6
284.7
96.45
100
1.116
12.395
389
0.1081
3689
272.0
284.7
95.53
100
1.116
12.195
389
0.1081
3630
263.3
284.7
92.47
Assuming 2 per cent moisture
145
1.632
17.821
558
0.1550
3698
273.2
284.7
95.97
130
1.46
15.977
502
0.1394
3685
271.3
283.1
95.84
115
1.288
14.147
446
0.1239
3672
269.5
281.5
95.74
100
1.116
12.295
389
0.1081
3659
267.6
279.7
95.67
B.t.u. are given only to the nearest tenth, and for this reason efficiencies are not accurate in
second decimal place.
B.t.u.) = equivalent of kinetic energy of jet in B.t.u.
B.t.u.a = available heat energy of steam.
and box pressures were the same, the apparent reaction rises above
the true reaction. As an example, the terminal pressure at 115 lb.
initial pressure is approximately 0.7341 lb. When the box pressure
was increased to 0.89 the apparent reaction was 14.10 lb. To this
apparent reaction was added the correction factor (difference in
TESTS OF STEAM-TURBINE NOZZLES
637
•sqv 'qi 'uonnoaa
; 1 1 1
/
1/
\l
/
/I
1/
1
1 1
/!
7
1
/
1
ij
1
/
1
/
1
5l
f?
^i
ri\
4
1
^•i
y
<\
1
^•i
yh
1
"it/
^.\
^
='1
V
a
1
f\
^J
"i
-7
si
31
s
1
'^\
-/
:^i
^
:t^i
^'J
J5
1
11
fl
2|
r/
IM 1
V
s
1
Ojj
!^i \
^1
'^\
1
ll_
1
Icjn
i '
^
u
1
1
§1
//I i
t^l
c?
sl
■f
«
1
?'
■• / 1
•^i
r/
"^1
s
1
k'
^/_^
^
/
li
«i
y
11!
1
?i -^1
"Si
—M-L
^
'
•d
J
SI i^
/
"
«;/
"1
?/
1
1
1
5f
^,
^i
#
^
131
■^
Y
K
1
1
•■
/
"1
/
5
'
o
1
1
1
/
1
/
1
1
/
'!'
^)
-
&
1
1
/
[
/
1
^
^
1/
^y
1
/
1
Si
""■
/
1
/
1
•^
.,^
1
9
f
\-l
1
1
pJO-J^-r
>—
1
1
U:l
1
1
/!
?-'^
/^t
L~~.
^
K
/
1
1
\
'^i
^
~7^
-^
_
y
1
1
i
/!
~"
-^
-.
■/
/
*>
i
<ip
^
!i
S"
-.
/
1 1
1
1
/
1
•?
1
~~
1
1
/
1
,^/
/
1
1
r
*»
/
1
jl
(f)
?'/
1
1
1
1
^1
1
1
;
1
-V
/
1
1
1
1
1
1
_^i
1
I
1
1
m;
1
ii
^i
1
^i
i 2'
1
2l
/
r'l
t
3'
Pi
•sqv •qT-'«on3B9a
638
TESTS OF STEAM-TURBINE NO/iZLES
o
lO
o
o
o
0
osqy ■"
> lo
i o
[•Bs -qi "^'x
O lO o
ej T-i' ,-i
to
o
o
"".
o
KB
O
\
^
\,
^
\
o
\^
§
,-,
\
T'
\
I
\
\T
\
f
C5
\
1
\
1
\i
^
o
\-
\|
i
\
\
^
i\
r|
\
^l
0
N
f-
\
V
\
'^
1
\
\
o —
_^
1
\
S'
r
\
f
—
A
*>—
\
A
-o--'
y-
1
^
\
\
<b--
\
\
\
L
^
t
A
f
' —
—
A
P-
A
\
.w
1
\
\
\
\^
1
\,
\,
\
1
\
s
s
1
■^
s
\,
1
\,
\|
%
^
\,
1
\]
\
1
\,
—
\
1
^
\
0-
A
a
i
1
\
1
\
\
1
\
\
^
1
\
\
^
^
1
\
r,
<ti-
\^
\
•0
k
\
L
\
i
p.
-
1
^
S
1
\
1
N
-°
1
\
]
\
\
n
°
\
\
t
\
\
'
\
\
1
\
*■
k
\=^
\
\
sV-
<J
rr.
\
—
1
1
H
\
o
-
^
^
^1
\i
^
rt'
—a
\
1
'"
s'l
\
•1^
\,
•o
\
4>
\
\
1
\
^
\
\
1
\
o
^
f-
A
*t—
A
\
\
f
-4
i
1
1
\,
1
\,
,
1
\
t.^
P.
—
1
r
\
1
s
1
N
6
1
\
1
1
\
1
\
1
\
<?
=)
H
\
S
¥\
^
•<!>■
—
1 —
\
»« —
\
\
H-
s
A
V
r
~
~'
^
"^
i~
=^i
1
\
s
\
^^i
1
1
-o-
s,
■a
Vd
X!l
M
■p-
A
CS
.1
■^
\
ji^
-Jl
^[
-^i'
\
o
?j^\
S'
fc^l
sh-
\
^
11.1
\
III
^!
Ai
\,
Vi
\
^i
\
K ^
ajniosqv -ni-Tjg -qi-j-i
i-< 11 o
TESTS OF STEAM-TURBINE NOZZLES 639
pressure X area of muzzle) to give the corrected reaction 14.351.
Fig. 14 is like Fig. 13, the values being for nozzle No. 14.
40 The method of calculation may be illustrated by the follow-
ing from nozzle No. 14 with an absolute initial pressure of 145 lb. per
sq. in. The terminal pressure (See Fig. 10) is 1.662 lb. per sq. in.,
and deducting 0.03 this becomes 1.632 lb. per sq. in. The reaction
(Table 4) is 17.821 lb. The flow is 558 lb. per hr. or 0.155 lb. per sec.
reaction X Q 17.821 X 32.16
Velocity = V= „ -^ , = 1^^^ = 3697.6 ft.
flow (lb. per sec.) 0.155
per sec.
72 3697. 6»
B.t.u., = kinetic energyof jet= — ^g = 2 x 32.T6 >r778 = ^73.22
B.t.u.2 = available energy (from steam table) = 289.8
B.t.u., 273.22
Efficiency = ^r^ = oon o = 0.9428 or 94.28 per cent
xJ.t.U.j Joy.o
41 If the terminal pressure had been determined as 1.732 and no
other factor changed, the true reaction as calculated would have
been 17.721 and the resulting efficiency would be 94.23 per cent.
42 If the flow_^had been determined as 563 lb. per hr. with no other
change of values the efficiency would have figured 92.61 per cent.
43 A reaction of 17.921 without other change would have given
an efficiency of 95.34 per cent.
44 If we assumed 2 per cent of moisture, the efficiency would figure
as 95.97 per cent. ^
45 The efficiency wasj^also calculated by the search-tube method,
by first plotting curves (similar to those shown in Fig. 15) showing
the relation between pressure and rate of flow per unit area of section,
with adiabatic expansion and with various percentages of friction
loss. The pressure and flow found by experiment were then plotted
on this chart and the efficiency determined graphically by comparison.
46 The chart in Fig. 15 was designed for finding the nozzle
efficiency by the searchjtube method. The vertical scale is 20
lines to the pound. The horizontal scale is 10 lb. per hour per line.
The chart shows the relation between the flow ^ in pounds per square
inch of section and the pressure in any section jof the nozzle, assum-
ing adiabatic expansion for the lower curve and 5 per cent loss of
640
TESTS OF STEAM-TURBINE NOZZLES
heat for the upper one. The data for plotting these curves were
obtained from the steam tables. By plotting the observed values
of pressure and flow upon these sheets we are able to obtain a graphic
solution for efficiency. For example, in the case of nozzle No. 9 the
terminal pressure at 100 lb. initial pressure was found to be 0.638
lb. per sq. in. and the corresponding flow in pounds per square inch
of section is 238 ± . From the chart, assuming adiabatic expansion,
under terminal pressure of 0.638 the flow would be 248 lb. and
238/248 is approximately 96 = the efficiency.
"
~
N
\
\l
o
N
s
<\
^.
\
s
s
■v.
s
S
N
\
N
^
N
S
\
\
>
N
s
s
^
V
N
s
s
1
V
^
N
N.
X
s
s
\
N
"s
s
N
S
<
N
>,
s
s
N
S
s
[V
■:
■i
-"-J
-
1 [
^c-
coo 580 560 540 5:^0 500 480 430 440 430
Flow Lb.Sq.In. of
400 380 360 340
Section in One Hi-.
320 aOO ^0 2'M 340 230
3.0
1.9
1.8
1.7
1.6 ■
1.5 '
1.4 :
1.3 !
1.2;
1.1
1.0,
0.9
0.8
o.r
0.6
0.5
200
Fig. 15 Pressxtbe-Flow Chakt
RESULTS AND CONCLUSIONS
47 Table 5 gives a summary of the calculations for nozzles No. 9,
11, 13 and 14, and Table 6a summary showing what would be the
effect of error in the observed values.
48 The efficiencies for nozzles No. 9 and No. 13 were obtained
graphically by the method described in Par. 46. The discrepanc}-
between the eflficiencies calculated for the search tube and the reac-
tion nozzles is due principally to the great difference of the effect in
the two methods of calculationca used by a small error in terminal
TESTS OF STEAM-TURBINE NOZZLES
641
TABLE 5 SUMMARY OF CALCULATIONS OF EFFICIENCY FOR NOZZLES
Nos. 9, 11. 13 AND 14
LP. Lb. T.P. 1 p^ow
Nozzle per Sq. I^b. per Lb. per
No- In. Abs. Sq. In. hour
( Abs. 1
1 1
Reac-
tion
Pounds
Theoret.
B.T.tJ.
per Lb.
Calcul.
B.t.u.
PER Lb.
Theoret.
Vel. Ft.
PER Sec.
Calcul.
Vel. Ft.
PER Sec.
Eff.
%
9 100 0.638 ' 385
96.1
115 0.735 442
96.3
130 0.832 498
96.0
145 0.929 553
95.6
11 100 0.638 385
115 0.735 442
130 0.832 498
145 0.929 553
13 100 1.116 389
12.45
14.351
16.244
18.134
311.5
313.6
315.4
317.4
280.1
282.4
285.0
288.0
3948
3962
3973
3985
3744
3759
3776
3796
89.9
90.0
90.4
90.7
98.3
115 1.288 446
97.9
130 1 . 46 502
97.5
145 1.632 558
97.1
14 100 1.116 389
115 1.288 446
130 1.46 502
1 145 j 1.632 ! 558
12.295
14.147
15.977
17.821
284.7
286.5
288.2
289.8
267.6
269.5
271.3
273.2
3774
3796
3798
3808
3659
3672
3685
3698
94.0
94.1
94.2
94.3
TABLE 6 EFFECT OF ERROR IN OBSERVATIONS
Nozzle
Observation
Error op
Corresponding
% Error in
No.
Observations
Efficiency
+
+
_
9
Terminal Pressure
0.1 lb. per sq. in.
8.5 to 14.0
11
Terminal Pressure
0.1 lb. per sq. in.
0.03 to 0.3
13
Terminal Pressure
0.1 lb. per sq. in.
5.4 to 9.2
14
Terminal Pressure
0.1 lb. per sq. in.
0.02 to 0.07
9
Flow
5 lb. per hr.
1.0 to 1.6
11
Flow
5 lb. per hr.
1.6 to 2.4
13 1
Flow
5 lb. per hr.
1.1 to 1.5
14 1
Flow
5 lb. per hr.
1.67 to 2.46
9
Dryness Factor
2%
0.9 to 0.6
11
Dryness Factor
2%
1.6
13
Dryness Factor
2%
0.9 to ^0.7
14 1
Dryness Factor
2%
1.7
11 1
Reaction
0.11b.
1.0 to 1.45
14
Reaction
0.1 lb
1.05 to 1.54
Note. — An error of + 0.1 lb. would be caused in the calculated reaction by an error of 4- 0.1
lb. per sq. in. in the box pressure-reading of No. 14 or by an error of + 0.0628 lb. per eq. in.
in the box pressure- reading of No. 11.
642 TESTS OP STEAM-TURBINE NOZZLES
pressure. An increase of only 20 lb. per sq. in. over the values used
would cover the discrepancy. Corrections for any slight condenser
leak which may have existed would decrease the flow values and bring
the calculated efficiencies closer together.
49 The "calculated velocity" for No. 11 and No. 14 is obtained
as described in Par. 40, from the reaction; and this was obtained from
the series of interconfirmatory tests plotted in Fig. 13 and Fig. 14.
50 The terminal pressures chosen were the minimum observed
values as determined mth the search tube in the nmzzles of the noz-
zles. The dryness factor was assmned as 100 per cent. Assuming
a 2 per cent moisture would make the calculated efficiencies for No. 13
and No. 14 very nearly equal, and very materially reduce the dif-
ference between No. 9 and No. 11.
51 In consideration of the above, taken in connection with Tables
5 and 6, we may assume that the values 91.5 per cent for No. 9 and
No. 11, and 95 per cent for No. 13 and No. 14, are probably within
2 per cent of the true efficiencies. No. 15 and No. 16 show a trifle
less reaction than No. 14 but the flow also appears to be a trifle less,
and there is not sufficient ground for assuming any difference between
the efficiencies of these three nozzles. Neither is there any appreci-
able difference in Nos. 10, 11 and 12. No. IS, with a greater flow and
less reaction than No. 11, shows an efficiency of about three per cent
less. Since no appreciable difference in efficiency is shown either with
a variation in cone angle from 9 deg. to 20 deg., or with such variations
in contour as shown in nozzles No. 15 and No. 16, smoothness of finish
would appear to be a much more important factor than the exact
contour.
52 The specific volume of the steam can be calculated from the
data in Table 5 and is greater than would be calculated by assuming
adiabatic expansion, because the dryness factor has been raised by the
friction.
53 In Tables 3 and 4, lines 1, 8, 9 and 10 contain the same values
as those given in Table 5; and in the other lines assumption is made
of certain values other than the observed values, to show what would
be the effect of error in observation, as is illustrated in Par. 40 to 44
inclusive and tabulated in Table 6.
54 The values in the last column of Tables 3 and 4 are given to
the second decimal place, to show what would be the effect of certain
errors in observation. In Table 5 they are given to the first decimal
place, to avoid conveying the erroneous impression that some of the
values as computed from observations were absolutely identical.
1
TESTS OF STUVM-TURBINE NOZZLES 643
DISCUSSION
Prof. J. A. Moyer. The methods used in these tests are ob-
viously much more accurate than the impact plate devices used by
Lewicki in his experiments with De Laval nozzle and by others who
have conducted similar investigations more recently.
2 The high efficiencies obtained may be surprising to some who
have not followed the latest developments in the designing of steam
nozzles. Results of this investigation confirm in general the results
given by Steinmetz^ and by the writer showing that the efficiency of
a well-designed nozzle for relatively large, as well as for small, limits
of pressure will be above 97 per cent.
3 However, in one respect the investigation is not as complete
as it was hoped it would be. There are not enough data to determine
the effect on the efficiency of varying the length of a nozzle: that is,
nozzles ot different lengths, but with the same taper or angle of
divergence, should be compared. However, the statement is made
in the paper that there is no appreciable difference in the efficien-
cies of nozzles 10, 11 and 12, which, however, do not have the same
taper, but have the same areas at the throat and at the mouth.
It is probable that all of these nozzles were longer than they should
be to obtain the highest efficiencies. More data are needed about
the best length of the nozzle for a given expansion. Lewicki's ex-
periments cover the two extremes: nozzles which are obviously
too short, and those which resemble in proportions the ones used
in this investigation.
4 The error due to moisture in the steam could not readily be
determined, and while it is probably not large, yet this uncertainty
might have been avoided by using superheated steam. The reaction
in a nozzle due to the flow of superheated steam is apparently con-
stant for a varying amount of superheat. This can be shown by the
usual thermodynamic equations for flow and velocity — which deter-
mine the impulse force of a jet — and by the experiments of Lewicki^
on the flow of superheated steam through De Laval nozzles. It
should be observed however, that when these tests were started,
Knoblauch and Jakob had not yet pubUshed the values which we are
now using for the specific heat of superheated steam, and for this
reason alone it was desirable to avoid the use of superheated steam .
1 The Journal, Am. Soc. M. E., May 1908, p. 628. •
' Mitteilungen iiber Forschungsarbeiten, Heft 12, Zaientafcl 9 (c). Verein
deutscher Ingenieure, 1904.
644 DISCUSSION
5 It has not been mentioned by the authors of this paper that
their method can be used to calculate the apparent efficiency of any
nozzle for any initial and final pressures. By measuring the areas
of a nozzle at the throat and at the mouth or "muzzle," the expan-
sion ratio in a nozzle is determined, and by means of empirical equa-
tions, due to Zeuner and others,^ the ratio of the corresponding initial
and final pressures giving the highest efficiency, can be obtained.
This ratio of pressures would correspond to the condition in these
tests where the terminal and box pressures are the same.
6 If the ratio of the initial to the final pressure has been deter-
mined, either of these pressures can be readily calculated if the other
is known. For example, if by measurement of the mouth and the
throat areas, the expansion ratio of the nozzle is found to be, say, 3,
then the ratio of the initial to the final pressure must be nearly
13.3 for the maximum efficiency of the jet discharged from it. For
this nozzle, therefore, with an initial pressure of 200 lb. absolute, the
final pressure should be 15 lb. absolute. From the equations given
in Par. 4 of the paper, the theoretical reaction can be readily calcu-
lated from the available energy corresponding to the pressure limits.
The change in reaction due to final pressures different from those for
which a nozzle is designed is, then, according to the method presented
here, the product of the area of the mouth of the nozzle, times the
difference between the correct final pressure for the nozzle — in this
case 15 lb. absolute — and the pressure in the box, or in practice the
pressure inside the casing of a stage of a turbine. Since reaction and
velocity are directly proportional — ivith constant flow — the apparent
velocity of the jet will be increased or decreased in the same proportion
as the reaction is increased or decreased.
7 In actual practice, however, this does not occur. It has been
obse/ved that if a nozzle is used which does not expand the steam
sufficiently, there is not nearly so much loss in the velocity of the jet
as when the nozzle is too wide at the mouth and "over-expands"
the steam. In other words, it has been found that a nozzle which
is about 25 per cent too large in area at the mouth, will give to the
jet only 90 per cent of the theoretical velocity, while one which is
too small by the same percentage will give within 2 or 3 per cent of
the maximum efficiency obtainable with the pressures best suited.
All this involves something which is not taken into consideration in
these reaction experiments; and for that reason, the results obtained
^ J. A. Moyer, The Steam Turbine, p. 40-41.
TESTS OF STEAM-TURBINE NOZZLES 645
by this method with varying back-pressures may possibly be mis-
leading.
Prof. C. C. Thomas. I have for years been interested in this line
of investigation, and am glad to see this contribution. In Par. 23,
the corrections which the authors make to the observed reactions
seem to me to be somewhat open to question. Aside from this fact, I
cannot quite see the theory upon which the corrections are based;
the fact that the pressures vary considerably in all but perfect noz-
zles, from the center to the walls, and that very considerable irregu-
larities of flow are found in nozzles, makes me doubt the necessity
for making these corrections to the observed reactions.
Strickland L. Kneass. The tests appear to cover ordinary
straight-tapered nozzles, as follows: 1 in 6, 1 in 5.77, 1 in 5, 1 in 4
and 1 in 3. In several cases the net areas vary slightly from these
ratios owing to the displacement of the cylindrical search tube, but
the ratio of the throat area to the outlet area is practically the
same for all nozzles, so that the results relate chiefly to the effect
of the length of the tubes and the friction upon the walls.
Fig. 1 Suggested Form of Nozzle
2 From Table 5 it would appear that nozzle No. 13, which has a
taper of approximately 1 in 4, gives a much higher efficiency between
100 lb. and 145 lb. absolute, than nozzles Nos. 9, 11 and 14, with
tapers ranging from 1 in 3 j to 1 in 6. As far as the knowledge of
the writer extends, there is no logical reason for this result, and he
would attribute the higher percentages to greater precision in the
experiments rather than to any inherent efficiency in the l-in-4
nozzle.
3 The correct contour of a nozzle for the discharge of an elastic
fluid is still a moot question. After an extended series of experi-
ments between the years 1888 and 1891 with steam nozzles of various
tapers, the writer offered the suggestion that the section should be
in the form of a reversed curve, somewhat as shown in Fig. 1 here-
646 DISCUSSION
with. This curve was based on the theory that the acceleration
should be constant during the passage of the steam through the noz-
zle, and that the areas at the several sections should be unit distances
apart. These sections were calculated with due allowance for the
change in the specific volume of the steam during expansion. The
results obtained seemed to confirm this theory and were compared
with the discharge from straight-tapered nozzles in a paper read
before the Engineers' Club of Philadelphia in 1891, The writer's
opinion was further corroborated by F. Hodgkinson before the
Engineers' Society of Western Pennsylvania in 1900. In view,
therefore, of published experiments upon nozzles of special contour
for which advantageous results were claimed, it is surprising that
the authors of this paper did not increase its value by widening the
scope of their experiments, instead of confining their tests to the
oldest and possibly less ejfficient form of tube.
4 Referring again to the experiments of the writer, his conclu-
sions covered the general theorem that there was little difference
in the efficiency of the straight-tapered nozzle, so long as the terminal
pressure of the steam within the tube was the same as that of the
medium into which it flowed, and, further, that the terminal velocity
would be the same under this given condition whether the taper were
1 in 6, 1 in 5, or 1 in 3. This opinion seems to be sustained by the
authors, although the results are not satisfyingly definite, because
different terminal pressures were used with each initial pressure and
the table does not contain the terminal pressures within the nozzle,
so that the comparison cannot be made with the pressure of the
final medium.
5 It is desired that this point be emphasized, for a slight difference
between these two pressures has an important effect upon the results.
It is thought that a more exact method of determining the relative
efficiency would have beea to modify the length so that the terminal
internal and external pressures would always be the same, for when
an attempt is made to introduce minus or plus reaction for correc-
tion, doubt is thrown upon the result. This is especially obvious to
any one who by careful observation of the flow of steam through and
from nozzles of different proportions, has noted the unstable equi-
Ubrium of the jet when the terminal pressure of the medium exceeds
that within the end of the nozzle. Some of the minor discrepancies
may be charged to this item and the writer is somewhat skeptical
as to the accuracy of the results obtained in practice when calcu-
lated under the theorem given in Par. 23.
TESTS OF STEAM-TURBINE NOZZLES 647
6 It would have been interesting if the authors had recorded new
data relative to the action of the steam within the nozzle and deter-
mined the terminal specific steam volume. The writer maintains
that the specific gravity of steam at different sections of the nozzle
does not correspond to that calculated by the thermo-dynamic equa-
tion, and therefore would be glad to have the authors state if the
velocity of the steam, as given in Tables 3 and 5, is equal to the specific
volume based upon the adiabatic equation, divided by the cross-
sectional area.
7 A test of this kind should give the initial condition of the steam.
The authors state that a thermometer was placed in a well at the rear
of the nozzle, but there are no figures in the table giving the percent-
age of moisture. An objection to the construction of the apparatus
can be offered in the liability of condensation of steam in the verti-
cal flexible supply pipe. The steam flows downward under pressures
varying from 100 lb. (328 deg. fahr.) to 145 lb. (356 deg. fahr.) and is
surrounded by steam at a pressure of 28 in. vacuum (100 deg. fahr.)
so that a certain amount is sure to be condensed.
The Authors. It appears to be a generally accepted fact that
under-expansion in the nozzle is preferable to over-expansion. Stod-
ola's Theory of Steam Shock and his search-tube experiments point
very decidedly in this direction. Reaction experiments may even
appear to indicate that under-expansion in the nozzle is in some cases
preferable to using the theoretically correct ratio. This may also be
true; but if the theory advanced in Par. 23 is correct, it is impossible
to accept the results of any purely reaction experiments as giving a
definite answer to this question; and where the pressure in the muzzle
of the nozzle is not taken into account, all the results may be in error.
2 Of course, it is possible to calculate the muzzle pressure by theo-
retical and empirical formulse; but if we are to rely upon theoretical
formulae there is no object in conducting tedious and expensive ex-
periments. Moreover, empirical formulae on this subject are at least
liable to be based in part upon reaction tests which have not taken
into proper account the pressure in the muzzle of the nozzle. Also
when the nozzle discharges into a pressure which is considerably
greater than the theoretical muzzle pressure, violent fluctuations
occur within the nozzle itself, so that the formulse do not apply and
the results of reaction tests may become very misleading.
3 Par. 23 has been called in question from both the theoretical and
the practical standpoint, so that a more extended consideration may
not be out of order.
648 DISCUSSION
4 The first statement, " The reaction of any nozzle is equal to the
summation of all the components, parallel to its axis, of the pressures
within the nozzle and in the chamber from which it leads," can
scarcely be questioned.
5 The net accelerating force F (Par. 4) which produces the velocity
actually present in the muzzle of the nozzle may be divided into two
parts. One part (call Ff) is a summation of components of the forces
with which the internal walls react against the pressure of the steam.
The second part is a force due to the pressure of the steam in the muz-
zle, and acts in opposition to the first.
6 Let Fm be this second part, P^ the muzzle pressure, and A the
muzzle area. Then
F„ = P^A
and
F = Ff-F^ = F^-P^A
Let R be the "true reaction of the nozzle," i. e., the force which is
equal and opposite to F. Then
R = F = Ff-P^A (1)
7 The apparent reaction (called Ra) is the force with which the noz-
zle actually pulls or pushes in the direction opposite to the steam flow
during the test. The apparent reaction of any nozzle is equal to the
summation of the components parallel to its axis, of all the pressures,
both internal and external, upon the walls of the nozzle and of the
chamber from which it leads.
8 That part which is due to the internal wall pressure is equal to
Ff. The external pressure acts, in the direction of flow of the jet,
upon an area which is greater than that upon which it acts in the
opposite direction, the difference being the area of the muzzle.^
9 Let Pe be the external pressure. Then
Ra = Ff-P,A (2)
Combining (1) and (2) we have
R = Ra = A(Pm- Pe) (3)
10 The rest of Par. 23 accords with these equations.
^ Gages connected to various points within the box showed that the external
pressure did not vary in different parts of the box by as much as 0.01 lb.
per sq. in. It must be remembered that the nozzle and the chamber from
which it leads are here suspended within the box into which the jet discharges.
TEST OF STEAM-TURBINE NOZZLES
649
11 It is evident from this that any acceleration or retardation of
the jet beyond the muzzle (due to the pressure into which it is dis-
charged or to any other cause) cannot affect the true reaction, and
that so long as the conditions within the jet are stable so that the muz-
zle pressure can be properly determined, there is no danger of being
misled except by a failure to make the corrections.
12 When the pressure into which the nozzle discharges is consid-
erably greater than the theoretical muzzle pressure, such violent fluc-
uations ensue as to make all corrections impracticable, and the reac-
tion tests under these conditions become worse than useless because
they are misleading. The criticism by Professor Thomas is well
founded with regard to such cases; but does not apply to the testf-
reported in this paper for the reason that these were all made under con-
ditions which did not disturb the stability of the jet within the nozzle.
13 The fact that the corrected reactions shown in Fig. 13 and
Fig. 14 lie in a horizontal line, i. e., are equal, is a further evidence
that the theory upon which they are based is correct, also of the fact
that the jet within the nozzle remained in very stable equilibrium,
and that the creeping in of the external pressure along the internal
wall had no practical effect, while the box pressure varied within the
limits shown.
14 To show further the form of error involved in the failure to
use these corrections, apparent and true reactions have been taken
from Fig. 13 and Fig. 14, and the accompanying table computed.
Noszle
t. p.
Flow
Lbs.
per sec.
Box
t. p.
Reao.
Vel.
i
1
B. t. u.i
B. t. u.»
Table
Eff.
11
145
.1536
0.929
18.134*
3796
288.0
317.4
90.75
14
145
.1550
1.632
17.821*
3698
273.2
289.8
94.28
11
145
.1536
1.632
17.0lt
3561
253.5
289.9
87.43
14
145
.1550
0.929
18.62t
3843
295.1
317.4
92.96
11
100
.1069
0.638
12.45*
3744
280.1
311.5
89.91
14
100
.1081
1.116
12.295*
3659
267.6
284.7
93.09
11
100
.1069
1.116
11.69t
3517
247.2
284.7
86.81
14
100
.1081
0.638
12.77t
3799
288.4
311.5
92.59
* Apparent and true.
t Apparent.
Data obtained from nozzles No. 11 and No. 14, with the box pres-
sure equal to that in the muzzle of the nozzle, are given in lines 1 and
2. These velocities and efficiencies are the same as those given in
Table 5, and require no correction for terminal pressure.
650 DISCUSSION
15 For line 3 the apparent reaction is taken for nozzle No. 11
with a box pressure which would be correct for No. 14, and the appar-
ent velocity and efficiency of No. 11 are calculated from that basis.
16 For line 4 the apparent reaction for nozzle No. 14, with a box
pressure which would be correct for nozzle No. 11, is similarly used.
17 It was found in the experiments plotted in Fig. 9 and Fig. 10,
that the pressure conditions within the nozzle remained stable and
practically constant with such variations from the proper box pres-
sure for each nozzle. Also, by applying the corrections called for in
Par. 23 it is found that these values reduce to the same values as
those obtained in lines 1 and 2, showing that the velocity and efficien-
cies of the jets as they reached the muzzles were not affected by the
changes in box pressure.
18 The acceptance of the uncorrected values would therefore im-
ply an assumption that in nozzle No. 14, with an initial pressure of
145 lb. and a terminal pressure of 0.929 lb.,thejetattained avelocity
of 3698 ft. per sec. in the nozzle, and that after leaving the nozzle
its velocity jumped to 3843 ft. per sec, and that in nozzle No. 11 , with
an initial pressure of 145 lb. and a terminal pressure of 1.682 lb., the
velocity of the jet after leaving the nozzle dropped from 3796 ft. to
3561 ft. per sec.
19 The efficiencies calculated from the apparent reactions, if
accepted in this form, would show that No. 14 is better than No. 11,
not only for its own proper terminal pressures, but for the terminal
pressures found in the muzzle of No. 11 as well. It may be that
such is the case; but there is considerable probability of arriving
at erroneous conclusions if it is assumed arbitrarily, without having
first been proved by very careful experiments which are not in any
manner dependent upon the assumption for their accuracy. There
certainly is no basis for making such an assumption from these data
as it has no bearing whatever upon the subject.
20 Previous to the time when this series of tests was begun, there
had been considerable investigation of nozzles with cone angles up to
12 deg. ; but the action of steam in nozzles of greater cone angle had
not received the same degree of attention. It was therefore decided
to use nozzles with divergence angles of from 9 to 20 deg., it being
then thought that this upper limit might be beyond the value for
highest efficiency.
21 Another set of nozzles tested contained one with a cone angle
of 24 deg. 30 min., which seemed to show an equal efficienc}^ with
those of smaller angle. This set was made of babbitt metal, was not
TESTS OF STEAM TURBINE-NOZZLES 651
perfectly smooth and was somewhat worn with long-continued use,
so that the results could not be thoroughly checked.
22 With the steam conditions given and the ratio of muzzle to
throat area determined therefrom, the only point left for the designer
is the general contour of the nozzle, including the shape of cross sec-
tion, length and angle or angles of divergence. The two sets of noz-
zles shown in Fig. 6 and Fig. 7 were designed with this in mind, each
set having a common ratio of areas; those of Fig. 7 differing among
themselves only in length and consequent angle of divergence, or
vice versa, and those of Fig. 6 differing only in elements of general
contour, not including length.
23 Professor Moyer's statement that " nozzles of different lengths,
but with the same taper or angle of divergence, should be compared,"
is not understood, unless he means to suggest that the whole field of
different steam expansion ratios should have been investigated. This
was not permitted because of limitations of time and other circum-
stances famihar to most investigators. Such an investigation would
not serve to determine the proper length for a given steam expan-
sion ratio, because the different nozzles would not be suited to the
same steam conditions; but it would give the efficiencies for one angle
of divergence with all the pressure ratios to which the various nozzles
were adapted.
24 Each set contained one search-tube nozzle for use in determin-
ing experimentally the terminal pressure in the muzzle, to be appHed
in reaction tests on the rest of the nozzles in that set. The efficien-
cies of these nozzles. No. 9 and No. 13, as calculated by the search-
tube method, are shown in Table 5; but they are not worthy of con-
sideration except as an example of the inaccuracies almost certain
to be involved in this method. The high efficiency given for nozzle
No. 13 is not due to greater precision in the experiments, as Mr.
Kneass suggests, but rather to the great error in the search- tube method
of calculation, caused by a very small error in determining the
muzzle pressure. In Table 6 it is pointed out that a "4- error"
of only 0.1 lb. per sq. in. in determining the terminal pressure would
cause a " — error" of from 5.4 to 14 per cent in the "search tube
computed" efficiency of No. 9 and No 13.
25 These "search-tube computed" efficiencies are evidently re-
sponsible for Mr. Moyer's statement that efficiencies were here found
as high as 97 per cent. Values obtained from reaction tests are lower,
and it is upon these that the conclusions stated in Par. 51 are based.
652 DISCUSSION
26 No. 9 ("search-tube" nozzle) was made with a small angle of
divergence, to be doubly sure that the steam should not leave the
walls before reaching the muzzle.
27 Both the length and the ratio of areas in nozzle No. 10 were
made to correspond as nearly as possible with those in nozzle No. 9
so that the terminal pressure found in the muzzle of No. 9 might be
applied to reaction tests upon the former with the least possible error.
28 No. 11 and No. 12 were made shorter and with a greater cone
angle but with the same sectional areas, in order to find out what
difference, if any, this would make in efficiency.
29 No. 18 was finished rough for comparison with No. 11, upon
which the greater number of tests had been made.
30 No 14 was used to determine the efficiency with a smaller
expansion ratio.
31 No. 13 ("search-tube" nozzle) was made to correspond as
nearly as possible with No. 14, so that the terminal pressure as deter-
mined in the former might be applied in reaction tests with the latter.
32 No. 15 and No. 16 were used to determine the effect of these
very considerable variations in contour.
33 Other forms, such as shorter nozzles or those designed for uni-
form acceleration and upon other theories, may and probably do
give just as good efiiciency as those herein described. It seems
doubtful, however, in view of the uniform results obtained with noz-
zles of such different contour as those covered by these experiments,
whether it would be advantageous to use any form especially diflicult
to manufacture, unless it be for the purpose of controlling the shape
of the jet as it strikes the moving blades of the turbine. This is very
important, as it has a great effect upon the efficiency of action in the
blades.
34 It is to be regretted, as stated in Par. 16, that we were unable
to procure a calorimeter of sufficient accuracy for our purpose, but
such great care was taken to maintain uniform conditions in the boiler
room, and these conditions gave such repeated indications of the
dryness of the steam at the nozzle entrance, that the probable error
introduced is not serious.
35 As stated in Par. 11, the steam pressure was 155 lb. gage and
the superheat about 50 deg. fahr. at the boiler. Steam was throttle
to the required initial pressure just before entering the flexible pipe,
with the result that the thermometer inserted at the noz/le entrance
showed about 4 deg. superheat with 700 lb. flow perhr. and sometimes
a trace of superheat with 500 lb. per hr. It is probably fair to assume
TESTS OF STEAM-TURBINE NOZZLES 653
from this that the steam was dry when used with 145 lb. pressure at
the entrance to the nozzle, and that (in view of the greater throttling
which tends to offset the increased unit radiation from the pipes)
there was always less than 3 per cent of moisture present even with
pressures as low as 100 lb. abs."
36 It may be stated in conclusion that a proper method of deter-
mining the net effect of under and over-expansion in the nozzle would
be as follows:
First: Make a set of nozzles of the same cone angle and finish
with throats identical, and with muzzles of different areas.
Second: Determine accurately the proper terminal pressure
and the true efficiency of each nozzle, by the method
herein described, using a reaction apparatus in which
static and moving friction has been eliminated.
Third: Find the push upon a set of turbine blades, using each
nozzle discharging into its own proper terminal pressure
and into the pressures which are proper for each of the
other nozzles of the set.
Fourth: A comparison of the push exerted under these condi-
tions, bearing in mind the *'true efficiency" of each jet
within the nozzles, will show the net effect of under and
over-expansion.
No. 1257
AN ELECTRIC GAS METER
By Prof. Carl C. Thomas, Madison, Wis.
Member of the Society
The meter described in this paper is designed for measuring the rate
of flow of gas, air or steam. The operation of the meter depends
upon the principle of adding electrically a Icnown quantity of heat
to the gas and determining the rate of flow by the rise in temperature
of the gas between inlet and outlet of the meter. This principle lends
itself to the operation of a meter possessing the following charactei-
istics:
a There are no moving parts inside the meter or in contact
with the gas.
b The accuracy of the meter and its sensitiveness are independ-
ent of the rate of flow of gas, and of fluctuations in pres-
sure and temperature.
c The meter may be used to measure gas at high pressure as
well as at low pressure, and is independent of small fluctua-
tions in pressure, such as those in the discharge from an
air compressor or in the suction of a gas engine.
d The meter produces a continuous autographic record show-
ing the rate of flow and its variation.
e Meters of comparatively very small size have very large
capacity.
/ The meter may be opened for inspection, for blowing out
accumulated matter with an air blast, or for washing with
gasolene, and it can be dismantled to any extent desired
without interfering with the operation of the plant.
2 Fig. 1 shows^the meter as constructed for gas or air measure-
ment, and Fig. 2 shows the exterior of the meter, of which Fig, 1 is a
section. The meter consists of two parts, first, the measuring ele-
ment A (Figs. 1, 3 and 4), through which all the gas passes when the
Presented at the Annual Meeting. New York, (December 1909), of The
American Society of Mechanical Engineers.
656
AN ELECTRIC GAS METER
AN ELECTRIC GAS METER
657
meter is in operation ; and second, a by-pass, B (Fig. 1) , so arranged
that the meter can be readily cut off from the gas main by operation
of the valves C, when it is desired either to operate without the meter
for the purpose of inspecting or cleaning out, or to cut the meter out
altogether for any reason. In certain classes of gas work, rolling
valves, such as are shown at C, have been found to give trouble,
while in other classes of work they are satisfactory. The gate valves
customarily used in gas work can be substituted for rolling valves
as occasion requires, and the by-pass can be made up of ordinary
pipe and fittings instead of being a part of the meter.
Fig. 2['' 1 /lEW Showing the Comparative Size of the Electric Gas Meter
[j (at the Lower Left-Hand Corner) and the Ordinary Wet
Gas Meter of the Same Capacity
3 The meter consists of an electric heater D (Fig. 1 and Fig. 4) ,
formed of suitable resistance material disposed across the gas pas-
sage in such a way as to impart heat uniformly and at a regular rate
to the gas passing through the meter. The temperature of the gas
is thus raised from that, at entrance to some higher exit temperature,
and the rise of temperature is measured and autographically recorded
by means of the two electrical resistance thermometers E (Fig. 1 and
Fig. 4) , on the two sides of the heater.
658
AN ELECTRIC GAS METER
4 These thermometers consist of wire wound upon vertical tubes
so disposed as to come in contact with all the gas passing through the
meter, thereby indicating the average temperature over the cross sec-
tion of the gas passage. The fifteen tubes shown at the right of Fig.
1, and also shown in Fig. 3 and Fig. 4, extending in a vertical direc-
tion over the cross-section of the meter, support the resistance wire
of the thermometers so as to afford a rugged construction. These
H
Fig. 3 Heater Unit and One of the Resistance Thekmometbrs
thermometers are connected to a recorder (Fig, 2 and Fig. 5), which
draws a line on a chart and thus indicates the difference of tempera-
ture between the^two thermometers.
5 A typical diagram is shown in Fig. G, This diagram represents
a gas flow of from 90,000 to 85,000 cu. ft. per hr., taken during a por-
tion of the day when the fluctuation in flow is small, but nevertheless
AN ELECTRIC GAS METER
659
continuous. Every small fluctuation in quantity of flow is recorded
on the diagram.
6 The diagram in Fig. 7 was made during a period in which the
flow varied extensively, the smallest amount recorded being about
17,000 cu. ft. per hr., increasing to 45,000, then to 62,000, to 75,000,
the record ending at a flow of about 32,000 cu. ft. per hr.
7 The record in Fig. 6 was made with a temperature difference of
about 4 deg. fahr. between the two thermometers, and an energy
input of approximately 2 kw. The energy input when the record
Fig. 4 Showing Construction of Hbatbb and Thermometers
in Fig, 7 was made was approximately 1.15 kw. Fig. 6 is a typical
record for a meter of normal capacity of 100,000 cu. ft. per hr., with
an electric input of 2 kw.
8 The principle underlying the measurement of gas by this means
is as follows: If gas is flowing through the heater at a given uniform
and constant rate, and if heat is being supplied electrically, and im-
parted to the gas at a constant rate, a certain definite rise of temper-
ature will be produced in the gas during its passage between the two
thermometers and through the heater, and this constant difference
660
AN ELECTRIC GAS METER
Fig. 5 Recording and Operating Instrument
AN ELECTRIC GAS METER
661
of temperature will be maintained so long as the amount of gas passing
per unit of time is constant. But if the quantity of gas passing per
unit of time diminishes, the heat supplied at the same constant rate as
before will raise the temperature of the gas by a greater amount than
was the case when a larger quantity of gas was flowing and absorbing
the energy liberated by the heater. Conversely, if the rate of flow
increases, the energy being supplied to the heater and delivered to the
gas will not be able to raise the temperature by as great an amount
as when the rate of flow was less. The temperature difference jDro-
duced by a known input of electrical energy thus forms a measure of
the quantity of gas flowing through the meter.
9 The meter may be operated in either one of two ways, of which
the first is as follows: the difference of temperature between inlet and
5 5.
28 A.M.
<
At
10 ut 85
000 CU
ft.per
hour
o
A
)OUt 0(
1200 CU
.ft.pt'i
Iiour
la
\/MV
T^ir
V\J
■^'H
^
k/*yV
^%i
y^
v%,
w
CO
Fig. 6 Autograph Record Showing Gas Flow of about 87,000 cu. ft.
PER HR.
note: this diagram was takkn under approximately steady conditions op flow
ddring the regular operation of one of the plants of the milwaukee gas light com-
pany, the paper in this case was traveling at a rate of 3 in. per hr. the recorder
can be set for any one of three speeds of paper, 3 in., 6 in., or 12 in. per hr. the higher
speeds are desirable as they smooth out the curve of temperature differences. the
scale of temperature differences can also be greatly enlarged if desired.
outlet is kept constant, and the watts required to maintain this con-
stant difference of temperature vary directly as the weight of flow.
The watts input thus forms the measure of the weight of flow of air
or gas, the watts being measured by a recording wattmeter, or in
some cases by an integrating wattmeter. The fixed difference of tem-
perature (about 5 deg. fahr.) is maintained by the action of a device
made upon the same principle as the well-known autographic tem-
perature recorders used in connection with resistance thermometers,
but without the autographic part.
662 AN ELECTRIC GAS METER
10 The mechanism which actuates the pen carriage in the auto-
graphic recorder is so arranged that when the carriage tends to de-
part from the straight-line path indicating a constant difference of
temperature it automatically cuts in and out the resistance necessary
in order to maintain the fixed difference of temperature. This varia-
tion of energy input is accomplished by a small motor-controUed
rheostat mounted on the switchboard. Thus as the rate of How of
gas is increased, the temperature difference tends to decrease, and at
once additional energy is introduced sufficient to heat the increased
weight of gas so as to maintain the constant temperature difference.
This method of operation is advantageous because it does not require
the maintenance of a constant voltage on the line supplying the energy
for heating the gas. The accuracy is thus independent of the small
fluctuations in voltage generally found on electric supply circuits.
11 The second method of operation involves the use of the auto-
graphic temperature recorder, including the graphical part, the dia-
gram from which, representing the variation of difference of tempera-
ture with constant energy input, gives the measure of the quantity
of gas passing the meter. That is, the electrical resistance of the
meter remains constant, and the meter is supplied with current at
constant voltage, which results in constant energy dissipation in the
meter. The difference of temperature between inlet and outlet then
rises and falls according to the decrease or increase, respectively, of
the rate of flow of gas.
12 The first method of operation mentioned is superior to this
second method, inasmuch as the first is independent of any change
which might take place in the electrical resistance of the material
composing the heater. Operation by the second method requires
that constant voltage be maintained across the line, and that the
electrical resistance of the heater shall remain constant, or else
that both watts input and temperature difference shall be recorded.
In the experimental work of developing the meters it has been found
convenient to use this second and more cumbrous method, but in
meters at present under construction the first-mentioned method has
been adopted, thus avoiding the necessity for either constant voltage
or constant resistance, and resulting in simpler ap])aratus through-
out. A record of the watts input is, by the method now used, all that
is required for determining the flow of gas through the meter. The
meters can be arranged to operate with either direct or alternating
current, and the controlling device can be arranged to work with any
desired voltage.
AN ELECTRIC GAS METER
663
13 Fig. 2 shows, at the lower left-hand comer, an electric gas
meter together with its autographic recorder and switchboard con-
trol. This electric meter is used for measuring all of the gas which was
formerly passed through the large wet meter shown in the figure, and
is of sufficient capacity to enable it to measure about three times the
amount of gas for which the wet meter is suited. The electric meter
was placed in this position between a 100,000-cu.ft. gas holder and
the large station wet meter, for the purpose of calibrating the electric
meter and comparing the results, based upon the rate of drop of the
gas holder, with the readings of the wet meter. The curve obtained
from the autographic recorder was thus interpreted by means of the
calibration carried on in connection with the gas holder, the wet
meter and a meter prover of the largest size made. It was found that
the wet meter used in this case was exceedingly accurate. It had
been carefully put in order and calibrated before these tests, and when
operated at loads within its capacity, the readings were entirely relia-
ble. The best evidence of this is given by the results used in plotting
Fig. 8.
14 The specific heat of a given kind of gas appears to be very
nearly constant, since those constituents which vary from time to
time are not those which appreciably affect the value of the specific
heat. But it is desirable to calibrate the meters with a gas having the
same specific heat as the gas which it is intended to measure in a par-
ticular case. The specific heat of illuminating gas is very closely
0.020 per cu. ft. at mean atmospheric pressure and 60 deg. fahr.
temperature, as shown in Fig. 8 and also by the following calcula-
tion based upon a fairly typical analysis. Such variation as com-
monly occurs in the relative amounts of the various constituents
does not greatly affect the specific heat. The following calculation
is for a gas at 760 mm. and 0 deg. cent.
Vol. cu. ft. Weight per Total Weight Specific Heat Specific Heat
cu. ft., lb.
lb.
per lb.
per cu. ft,
COj
0.04
0.11637
0.004658
0.216
0.00100
c^,
0.11
0.0741
0.00815
0.404
0.00329
0,
0.001
0.08463
0.00085
0.217
0.00023
CO
0.331
0.07407
0.02450
0.245
0.00600
CH,
0.1761
0.04234
0.00746
0.593
0.00442
H,
0.303
0.00530
0.00160
3.409
0.00546
N.
0.0389
0.07429
0.00289
0.244
0.00071
0.02111
664 AN ELECTRIC GAS METER
15 The specific heat of blast-furnace gas is practically the same
as that of atmospheric air, and the same is true in a general way regard-
ing producer gas. Thus, taking the following as an average analysis
of blast-furnace gas, the specific heat is found to be 0.0192, while
atmospheric air has a specific heat almost identical with this, or
approximately 0.0191 per cu. ft. This is to be expected, since pro-
ducer gas and blast-furnace gas consist principally of nitrogen and
carbon monoxide.
Vol. cu. ft. Weight per Total Weight Specific Heat Specific Heat
cu. ft., lb. lb. per lb. per cu. ft.
N,
0.60
0.0743
0.0446
0.244
0.0109
CO
0.24
0.0741
0.0178
0.245
0.0044
CO,
0.12
0.1164
0.0140
0.216
0.0030
H,
0.02
0.0053
0.0001
3.409
0.0003
C,H,
0.02
0.0741
0.0015
0.404
0.0006
0.0192
16 The meters have been calibrated with illuminating gas and
with air. A certain amount of water vapor is carried with the gas
or air passing the meter. This vapor forms part of the gas or air,
and is heated just as are the other constituents. The rise of tempera-
ture caused by the heat added in the meter is only a few degrees, and
consequently the water vapor does not experience a change of state.
The temperature of the metal forming the electric heater rises only
15 or 20 deg. fahr. above the temperature of the gas. The question
of latent heat of vaporization of the water vapor therefore does not
enter into the considerations underlying measurement of the gas.
17 While calibration of the meters under actual conditions of ser-
vice is depended upon to obtain quantitative results, yet these meters
are of such a nature that the quantity of gas or air passing through
them can be very closely calculated from a knowledge of the energy
input and the specific heat of gas or air. This fact, that the quantity
of flow can be quite closely calculated, independently of a calibration
curve, makes it possible to check the accuracy of the readings obtained.
18 The development of this meter is a result of experiments which
the writer has been making for some years to determine the specific
heat of gases by heating them electrically. The performance of a
properly constructed heater for this purpose proved to be so entirely
regular that it was apparent that the quantity of gas flowing through
it could be very accurately measured by the method now used in
these meters. The problem is thus the reverse of the problem of
AN ELECTRIC GAS METER
665
determining specific heat by measurement of the electrical energy
necessary to heat the gas. It will be seen by reference to Fig. 1 that
the whole process of heating the gas and of measuring the difference
of temperature between inlet and outlet, is accomplished in a rela-
tively small space which is well insulated so far as heat losses are con-
Ab
>ut 170
OOcu.:
t.per 1
lour
A
A
About 33C00
cu.ft.per
hour
,
\K
V
/
1
Ah
^
sy
/v
About -15000
cu.ft.per
hour
/
,
/^
A^
f Ab
)Ut 610
OOCU.
Lper
lour
About 75000
cu.ft.per hour
VVw
J
Fig. 7 Autograph Record Showing Wide Fluctuations in Flow of Gas
cemed, since the heater and thermometers are contained in a casing
made of hardwood strips and separated from the metallic walls of
the meter by an air space.
19 A typical calibration curve is shown in Fig. 8. The curve
shows the degrees rise in temperature per kilowatt introduced when
666
AN ELECTRIC GAS METER
any given rate of flow through the meter is taking place. It will be
seen that this curve is asymptotic to the coordinate axes, because,
when an indefinitely great amount of gas is being heated, any finite
input of heat will produce only an indefinitely small rise of tempera-
ture ; and on the other hand, when the amount of gas becomes indefi-
nitely small, a finite input of heat will cause an indefinitely great rise
of temperature. The calibration curves obtained are therefore rect-
100000
\
90000
\
\
80000
\
\
C 70000
o
\
y
■a
S 00000
\
>
V
to
■"S 50000
\
>
Nv
O -10000
N
s.
\
s.
30000
s
V
V
"^
20000
"*^
^
>^
—
10000
I
3 4 5 6 7
Deg. rise of temp, per Kw. energy input.
Fig. 8 Calibration Curve, See Appendix for Data
angular hyperbolas. The product of weight of gas multiplied by
degrees temperature rise per watt introduced is a constant, and this
constant, for a given kind of gas, takes the place of a calibration curve
and renders it unnecessary to refer to a curve. The constant as shown
3.412
by Fig. 8 is 170,000, showing a specific heat per cu.ft. of t^ =
0.0201.
20 TheTaccurac}'' of these meters is not affected by changes in
AN ELECTRIC GAS METER 667
pressure of the gas or air, since the unit of measurement is that of
weight rather than of volume; that is, the meter takes cognizance of
the specific gravity, or the amount of "stuff" in a given volume of
the gas. Also variation of temperature of the incoming gas does not
affect the accuracy, because it is a difference of temperature, rather
than a fixed temperature, upon^which the measurement depends.
The meter can be used for gas or air at either high or low pressure,
and at either high or low temperature, provided the materials used
in construction are suited to the conditions.
21 This method of measuring gas seems especially useful in con-
nection with engines operated by gas from producers, blast furnaces,
etc., and in measuring the discharge of gas or air from compressors,
because the small and rapid periodic fluctuations of pressure, due to
the suction of gas engines or to the discharge from compressors, do
not interfere with the steady action of the thermometers. The time
lag of the latter is sufficient to smooth out the curve of temperature
variation, or of watts input, as the case may be, and true average
results are thus indicated.
22 The temperature difference employed when operating with a
constant difference, is approximately 5 deg. fahr. When a curve of
temperature difference is employed,, the temperature rise is from 4 to
5 deg. fahr. when the normal maximum amount of gas is flowing.
This]" difference may be increased to 10 or 12 deg. when the rate of
flow is greatly diminished, and at, 100 per cent overload the tempera-
ture difference is from 2 to 2^ deg. On the autographic record one
inch represents a temperature difference of one degree. The ther-
mometers and recording device are such as to render the records
accurate within 1 per cent. The minute fluctuations showTi by the
curves on Fig. 6 and Fig. 7 are produced by the constantly varying
rate of flow in the gas mains. These can be " damped out " to any
extent desired. The apparatus with which this record was taken
was purposely made sensitive to minute fluctuations.
23 The electrical energy required to operate the meters is approx-
imately 1 kw. per oO,000-cu.ft. hourly capacity. The curves shown
in Fig. 7 represent variations of from 17,000 to 75,000 cu.ft. per hr.,
and were made with an energy input of approximately 1.15 kw. To
provide for more gas and still have the record lie conveniently on the
paper, it is only necessary to increase the energy input by manipu-
lation of the rheostat hand-wheel on the switchboard.
24 The meters are so constructed that the heads can be easily
removed and an air blast used for cleaning out the interior, or the
668
AN ELECTRIC GAS METER
AN ELECTRIC GAS METER 669
entire casing, containing heater and thermometers, can be removed
and dipped in gasolene for the purpose of removing tar or other deposit.
All parts of the meter are of rugged construction, and are of well-devel-
oped materials familiar to engineers. The heater units consist of cor-
rugated strips of resistance ribbon about 1^ in. wide, wound spirally
into discs of such diameter as to fit the inside of the wooden casing.
The number of these discs depends upon the capacity of the meter.
The heater shown in Fig. 4 consists of two discs.
25 The same type of meter, modified as shown in Fig. 9, can be
used for the measurement of steam, and also for determining the
quality or percentage of moisture of steam. When used for measur-
ing the quantity of steam, the steam is first superheated slightly in
a superheater of the ordinary type, after it leaves the boilers and be-
fore passing through the meter.
26 The heater element in the steam meter consists of tubes, as
shown in Fig. 9, made of suitable resistance material and supported
on insulating bushings in the tube plates, the construction being sim-
ilar to that of a surface condenser. The slightly superheated steam
is passed through and around these tubes, and is further heated by
the electrical energy supplied to the tubes.
27 The difference of temperature produced by a given energy
input forms a measure of the weight o^ steam flowing, just as has been
described in the case of the gas meter. In cases where it is desired
to make engine or turbine tests with unsuperheated steam, the steam
can be reduced in temperature after passing the meter, by the injec-
tion of a spray of water. Of course the measurement of superheated
steam is simpler than is the case where superheating is not a feature
of the regular operation of the plant.
28 The amount of moisture carried by steam can be very accu-
rately determined with this apparatus, by passing all of the steam
through the electrical heating material and noting the amount of
energy required to "fry out" the water and cause superheating to
commence. The pointer over the dial of the instrument connected
with the resistance thermometer in the outlet of the calorimeter indi-
cates when the temperature of the steam begins to rise. It is pro-
able that the only way to determine accurately the quality of wet
steam is to pass all of the steam, and not a small sample, through
a calorimeter. It is of course not always p^-acticable to do this, and
in such cases it is necessary to use smaller calorimeters and to sample
the steam.
29 When inserted for cither regular or intermittent use as a steam
670 AN ELECTRIC GAS METER
meter or as a calorimeter, the device can be cut off from the steam
line in the manner already described for the gas meter, and as shown
in Fig. 1.
30 The automatic recording device for the gas meter is so arranged
that in case the flow of gas should be interrupted for any reason the
current is automatically cut off at the switchboard. Also if the flow
of gas becomes so small in amount that the pen reaches within a half
inch of the edge of the paper, the current is cut out. When the gas
has cooled the heater slightly, the current is automatically cut in
again, and if the gas flow is increased the pen goes back toward the
middle of the diagram and operation proceeds normally. If the gas
flow continues but does not increase beyond that at which the current
was cut out, the pen will "hunt" back and forth near the edge of the
paper. It can be brought back toward the middle of the paper by the
introduction of less energy to the meter. The gas meter is thus fully
protected from possible injury due to the com-plete shutting off of
gas supply.
31 At the other edge of the paper, representing the maximum flow
of gas, the operation is similar to that already described. In order
to bring the recording pen upon the range again the electrical input
is increased by manipulation of the hand-wheel on the switchboard.
This applies to operation by the second method described in Par. 11,
in which the temperature difference between the two thermometers
forms the record of gas flow. When the first method is employed,
that of maintaining constant temperature difference, the meter is also
automatically protected by the motor-controlled rheostat, and the
range of the instrument is unlimited and it does not require manipula-
tion by hand. It will be seen by reference to Fig. 7 that the range of
the instrument when operated by the second method of varying tem-
perature difference, is very wide, and takes care of extensive fluctu-
ations of gas flow.
THEORY OP THE METER AND METHOD OF OBTAINING STANDARD
RESULTS
32 The figures given in paragraphs 14, 15 and 19 can be reduced to
standard conditions of temperature and pressure, and the meter read-
ings can be autographically recorded directly in "standard cubic
feet" of gas or air. Let
G = cubic feet of gas per hour
E ^energy in kilowatts
GT
E
AN ELECTKIC GAS METER 671
Then B.t.u. per hr. = 3412 E
T = temperature difference, deg. fahr.
S = specific heat per cu. ft.
Then GST = heat energy equivalent to E, or GST - 3412 E
3412
= „ = a constant K which depends upon the specific heat of the
o
gas.
33 Since the temperature difference T is kept constant, it follows
K K K P
that is constant. Let _ = C. Then G = - ' = CE.
T T T
34 It is now proposed to show by reference to the gas and the air
curves in Fig. 10, that if the specific heat of gas made under given
conditions be calculated from the customary chemical analysis and
the specific heat of the constituents, then this specific heat may be
used for determining the constant C. Fi'om the gas curve (Fig. 10),
which was made with illuminating gas at an average temperature of
59 deg. fahr., and under an average absolute pressure of 6 in. water
and 29.8 in. mercury,
^4-19
K - 170,000 = ~
o
35 Therefore for the condition of the gas when the tests were made
the specific heat per cubic foot S= == 0.0201. If this be re-
170,000
duced to standard conditions of 32 deg. fahr. and 29.9 in. mercury,
then S = 0.021, which is to be compared with the calculated specific
heat (Par. 14), giving S = 0.0211. If the standard conditions are
taken as 62 deg. fahr. and 29.9 in. mercury, the specific heat becomes
0.0198, and the constant becomes
3412
^ " 0:0198 ^ 172,500, nearly
If the temperature difference is kept constant at 5 deg. fahr., then
K ^ 172^0 ^3^^^ = c, or (? = 3450 E.
i 5
36 The cross-section paper on the recording wattmeter is ruled
so that 3450 E is read directly, instead of the watts E. The record is
thus read directly in cubic feet of gas. The regular records of chem-
ical analysis of the gas should be referred to from time to time in order
672
AN ELECTRIC GAS METER
•.tnoq jad jib .to sbS ^aaj orqno =/?,
AN ELECTRIC HAS MBTEK 678
T.o ascertain what percentage variation takes place in specific heat.
It appears, as stated previously, that the elements which vary dur-
ing the operation of a gas-plant are not those whose variation would
produce serious variation^n^specific heat. The variation that does
take place is apparently^ well^^within the limits of accuracy practicable,
or generally considered necessary in the operation of gas plants. By
taking frequent chemical analyses the error can be reduced so as to
be quite negligible.
37 The conditions during the air tests were as follows: barometer,
29.75; pressure, 6.5 in. water; average temperature of air as measured
in the wet meter, 60 deg. fahr. From the air curve obtained under
these conditions (Fig. 10)
3412
K = 188,000, and S = ,^^7^ = 0.0181
loo,UOO
38 Reducing this to standard conditions of 32 deg. and 29.9 in.
mercury, S = 0.0191. This is to be compared with the accepted
specific heat of air under these conditions, or 0.0192 B.t.u. per cu. ft.
This provfdes perhaps the best evidence that could be obtained, as
to the accuracy of these tests, since the specific heat of air is well
known at the conditions under which the tests were made. A more
commonly familiar figure for specific heat of air is obtained by multi-
plying 0.0192 by the number of cubic feet of air per^pound under the
above conditions, or 12.38. The result is 0.2377 B.t.u. per lb. per
deg. and this is to be compared with 0.0191 X 12.38 as given by the
meter, or 0.2365.
39 The constant K for air at 32 deg. and 29.9 in. is therefore
^^-- = 178,630
0.0191
and reducing this to 62 deg. instead of 32 deg.
K = ( 1 X T^ ) X 178,630 = 189,500 nearly
If 7' = 5 deg., - = 3790.
40 The error involved in calhng this constant 3800 is less than ^
of 1 per cent and well within the limits of accuracy possible under the
circumstances. The standard cubic feet of air passing the meter are
therefore G = 3800 E, and the autographic records are arranged to
read accordingly, in standard cubic feet of air per hour.
674 AN ELECTRIC GAS METER
41 The development of a new device requires consideration of a
large number of questions arising out of the conditions of service
proposed. The question of specific heat has been considered in the
preceding paragraphs. The degree of success which has been attained
with this meter in accurately measuring specific heat is due princi-
pally to an extensive experience in this particular class of work, which
has served to point out the way to make an electrical heater in which
heat losses are negligibly small. The arrangement of the meter is
such that the heat given off can go into the gas only, and it necessarily
all goes into the gas, with the exception of a negligibly small loss
which it is not worth while to minimize further. That the gas re-
ceives all the heat, excepting this negligibly small loss, is true whether
or not the heating material has collected deposit of some kind. So
long as the gas can get through the heater, its temperature is raised
proportionately to the heat supplied.
42 The question of the presence of a small amount of water vapor,
as part of the gas, has so far not introduced any complications. It
is conceivable that if the gas carried a large percentage of water the
operation of the meter would be interfered with, — but so would the
operation of a gas engine or a burner. The meter can apparently
measure accurately any gas that can be used by a gas engine. The
absence of moving parts in the meter gives it an advantage over the
engine, and dust can be to a considerable extent deposited before
entrance of the gas to the meter. The heating element and thermom-
eters can be cleaned by dipping in gasolene, without damaging them.
43 Meters at present under construction are being made with the
axis of the cylinder vertical, with a view to greater convenience of
access and in making connections.
44 The first large meter of this type to be installed was put in the
works of the Milwaukee Gas Light Company, and the writer is indebted
to the officials of that company for their cooperation in making exten-
sive tests during the work of development.
45 Referring to Par. 16, for gas or air under the conditions exist-
ing during the tests, of approximately 60 deg. fahr,, 29.8 in. mercury
and 6 in. water pressure, the correction for water vapor introduces a
change in the results of less than one-half of one per cent, and was
therefore omitted. At other pressures and temperatures the correc-
tion for water vapor can be easily made by reference to the charts
commonly used in gas works. An interesting confirmation of the
statement in Par. 16 appeared during the tests, in that the most
AN ELECTRIC GAS METER
675
minute addition of electrical energy caused an immediate rise of
temperature of the gas or air. This was repeatedly tried with great
care, and always with the same result.
APPENDIX
DATA RELATING TO CALIBRATION CURVE, FIG. 8
Time
Average
Wet Meter Cu. Kt. Gas Temperature
Reading Per Hr. Difference
Deg. Fahr.
Average Deg. Temp.
Kilowatts Rise Per
Input 1000 Watts
A.M.
10-05
90148.0
10-10
10-15
90177.5
10-20
90192.0
10-25
90206.0
10-30
90220.5
10-50
90396.0
10-55
11-00
90470.0
11-05
90509.0
11-10
90546.0
11-20
90644.0
11-25
90695.0
11-30
90746.0
11-40
90884.0
11-45
90947.0
11-50
91010.0
p.m.
12-05
91098.0
12-10
91125.5
12-15
91152. T
12-20
91180.0
12-25
91206.8
A.M.
9-30
91418.2
9-35
91493.4
9-40
91568.8
9-45
91643.8
10-00
91857.1
10-15
92074.0
10-30
92291.7
10-45
92506.2
11-00
92717.5
11-15
92925.4
11-30
93131.0
1
17350
45200
6 1200 J J
75600
32640
90240
84960
10.7
4.25
3.25
2.70
6.25
3.90
4.10
1.153
1.150
1 . 160
1.160
2.05
2.05
9.30
3.69
2.80
2.32
5.12
1.90
2.00
676 DISCUSSION
DISCUSSION
Prof. L. S. Marks. The meter described by Professor Thomas
should prove a valuable addition to the instruments used in gas
engine and other testing. The possibilities of error in the indications
of such an instrument must be fully examined.
2 This meter is fundamentally an instrument for determining the
weight of gas or vapor flowing through it and is made to record vol-
umes. It is obvious that these volumes cannot be those actually
flowing but must be the volumes reduced to some standard conditions
of temperature and pressure. The author has not mentioned this
matter in his paper, but it is of considerable importance. A variation
of 5 deg. fahr. in temperature, or of 0.3 lb. in pressure, under ordinary
atmospheric conditions, would result in an error of 1 per cent in the
indications of the instrument if it were assumed to record actual vol-
umes flowing. The calibration of the instrument by passing through
it a known volume of a gas at known pressure and temperature, can
easily be reduced to a calibration under standard pressure and tem-
perature conditions.
3 In Par. 16 the author refers to the effect of water vapor car-
ried in with the gas. He states that, in consequence of the small rise
of temperature, the water vapor does not experience a change of
state, and that, consequently, the latent heat of vaporization does
not enter into consideration. It is obvious that he is considering here
the case of a gas which not only is saturated with water vapor, but
also is bringing with it minute particles of water in suspension. Under
these conditions — and they are conditions which may easily obtain
with blast-furnace gas which has just passed through the washers —
the indications of the instrument will be rendered completely useless.
4 If the gas should enter at a temperature of 70 deg. fahr. it
would contain 0.001148 lb. of water vapor. After passing through
the meter with a rise of temperature of 5 deg. the same weight of gas
could contain 0.001 198 lb. of vapor; that is, there would occur a vapori-
zation of 0.00005 lb. of moisture for every cubic foot of gas passing
through the meter. The latent heat of vaporization at these tem-
peratures is about 1050 B.t.u., or 0.0525 B.t.u. will be used in con-
verting the water into vapor. As the total heat required for raising
one cubic foot of the gas 5 deg. fahr. is only about 0.1 B.t.u., we
have here, obviously, the possibility of an error of the magnitude
of 20 or 25 per cent in the indications of the instrument in the
AN ELECTRIC GAS METER 677
case suggested by the author where the gas is supersaturated with
vapor.
5 The accuracy of the instrument depends primarily on the accur-
acy with which the volumetric specific heat of the gas can be deter-
mined, and upon the constancy of this quantity while the meter is in
operation. For the correct determination of the volumetric specific
heat it is necessary to know the volumetric composition of the gas
and the volumetric specific heat of each of the constituents. The
author has stated that the volumetric specific heat of each kind of gas
is very nearly constant and the calibration of the instrument is based
upon that assiunption; that is, it is proposed to calibrate the instru-
ment with, for example, producer gas, and then to use that calibration
when the instrument is used at other times with producer gas. It
will be interesting to examine how nearly correct this assumption
is. In the four analyses of producer gas, three of them by Mr. Bibbins,
and one b}^ Messrs. Garland and Kratz, I have worked out the volu-
metric specific heats of these gases, using the physical constants given
by the author, and I have also taken at random two analyses from
tests which I have made on a large anthracite gas producer. The
results of the calculations are as follows:
6 For the two lignites in Mr. Bibbins' paper, the values of the
specific heats are 0.01920 and 0.01899, which agree very closely with
the average stated by the author. For the bituminous coal in Mr.
Bibbins' paper, the value is 0.01899, and for the bituminous coal in
the paper of Messrs. Garland and Kratz, the value is 0.0186. My
own tests with anthracite give values 0.01826 and 0.01848, respec-
tively.
7 It is quite evident from these figures that there is considerable
variation, which may be as great as 5 per cent in the volumetric specific
heat of producer gas. It may possibly be, as these figures seem to
indicate, that the specific heat can be stated with greater accuracy
if the type of coal is also specified, since there seems to be a relation
between the volatile contents of the coal and the specific heat of the
producer gas; but this point has not been suflaciently investigated to
permit of any definite conclusions.
8 I have attempted also to see whether the value given for illumin-
ating gas is constant. Only one illuminating gas was considered —
that in Cambridge, Mass. — the analysis having been made by the
chemist of the gas company. The specific heat calculated from this
analysis is 0.02278. The specific heat calculated by the author is
678 DISCUSSION
0.02111. The value which he states as being practically constant
for illuminating gas is 0.020. There is a variation of over 10 per cent
between these values, so it would seem that it is not practicable to
calibrate this instrument with illuminating gas at one place and assume
it to be accurate when used with illuminating gas at some other place.
9 Moreover it must be recognized that such an analysis as that
given by the author for illuminating gas is only approximate; the heavy
hydrocarbons aie never fully analyzed and some kind of guess must
be made as to their composition and specific heats. It cannot even be
accepted as true that a calibration made with any particular illumin-
ating gas will hold at some later date for gas from the same source.
I have found variation in the composition of the Cambridge gas which
would certainly cause a variation of two or three per cent in its
specific heat.
10 It appears to me then, that this instrument cannot be accepted
for accurate measurement unless analyses are being made of the gas
that is going through the meter. In scientific testing, such analyses
will naturally be undertaken and consequently the instrument should
be extremely valual )le in such cases. I would like to know what experi-
ence the author has had with this instrument in the measurement
of volumes when the flow is variable as, for instance, when gas is
flowing through a single-acting, four-cycle gas engine. In this case
the flow will occur approximately for only one-fourth of the whole time
of the test. The author's contention that the indication of the instru-
ment would be accurate under these circumstances seems reasonable,
but it would be valuable to know whether, and to what extent, his
statement has been verified by actual investigation.
Prof. W. D. Ennis. I do not quite follow Professor Thomas'
explanation that the proper correction has been made for fluctuations
in the pressure of the gas. A change of, say, five per cent in the pres-
sure, measured above the zero of pressure, would correspond roughly
with a change of five per cent in the absolute temperature, without
any addition whatever of heat. A change of five per cent in absolute
temperature would mean a very large change in Fahrenheit tempera-
ture.
2 A more important point is suggested by the statement in Par. 4:
" These thermometers consist of wire wound upon vertical tubes so dis-
posed as to come in contact with all the gas passing through the meter,
thereby indicating the average temperature over the cross section of
AN ELECTRIC GAS METER 679
the gas passage. " If that is what the thermometers do, I question
whether they indicate the average temperature of the gas, because
more gas is passing at a point in the middle of the pipe than at points
near the circumference. Do the thermometers indicate the average
temperature of the whole weight of gas, which is the temperature that
we must have in order to calculate the weight of gas flowing?
Edwin D. Dreyfus. Certain fuel gases — particularly blast-fur-
nace, coke-oven and producer gas — carry with them a considerable
quantity of finely divided solid matter, which in turn forms deposits
in the piping or in any piece of apparatus through which the gas
passes.
2 From their construction, it would seem that the grids in the
meter would favor the formation of deposits of this sort, and I would
like to ask whether Professor Thomas has made any trials to deter-
mine what effect, if any, such deposits have on the accuracy and gen-
eral reliability of the instrument.
3 In cases where the gas is carried long distances through over-
head mains — as in many blast-furnace plants — the temperature of the
gas will be influenced largely by the temperature of the atmosphere, as
between the summer and winter months the gas temperatures might
easily vary as much as 50 deg., and the variation in temperature
would have a decided effect on the moisture content. It seems prob-
able that the moisture content of the gas is the most disturbing fac-
tor affecting the accuracy of the instrument. If this be so, then it is
desirable that the actual significance of this factor should be deter-
mined by trials made over as wide a range of conditions as we may
reasonably expect to meet in ordinary everyday practice.
A. R. Dodge. I would like to ask Professor Thomas if he has
made calculations in regard to the amount of power necessary to oper-
ate this meter when used as a steam meter. The specific heat of steam
being greater than that of gas and air, the amount of power required
is considerable. For instance, Thomas meters on the large turbines of
the New York Edison Company would require about 545 kw. at
normal load, quite a percentage of the total output of the turbine.
The Author. Bearing upon the questions asked in the discussion,
I would say, that in addition to the description of the meter given
in the paper, I have given in Fig. 10 completed curves showing the
680 DISCUSSION
results obtained in calibrating the meter with both illuminating gas
and air, reduced to standard conditions of 29.9 in. mercury and 62
deg. fahr. These curves show the method of using the meter for
measuring directly standard cubic feet of gas or air at some convenient
assumed conditions of pressure and temperature. The results of
measurement by the method described in the paper may be considered
as given either in standard cubic feet, or in weight of gas passing the
meter.
2 These meters are essentially applicable to the measurement
of a dry gas or steam, that is, a gas or steam which is either saturated
or superheated. Our experience with the gas meters has thus far
been with illuminating gas and with air, and these are exceedingly
easy of measurement. The gas or air we are measuring is saturated,
carrying its full quota of water vapor. The smallest quantity of
heat introduced causes an immediate rise in temperature of the gas.
If the gas carried a spray or mist of water, the measurement would
be in error to a certain extent, because of the difference in specific
heat between the water vapor and the gas. The extent of the error
would depend upon the percentage of water vapor present.
3 For gas or air under the conditions existing during the tests,
of approximately 60 deg. fahr. and 29.8 in. mercury, and 6 in. water
pressure, the correction for water vapor introduces a change in the
results of less than ^ of 1 per cent, and has therefore been omitted.
For other pressures and temperatures the correction for water vapor
can be easily made by reference to the charts commonly used in
gas works.
4 An interesting confirmation of the statement in Par. 16 is that
the most minute addition of electrical energy to the gas or air causes
an immediate rise of temperature.
5 Regarding variation of specific heat, the meter prover shown in
Fig. 1, herewith, has been developed. It consists of a small electric
heater which is placed in the outlet of the meter and discharges into
a portable gas-holder such as is used for proving large meters. By
this means a small known quantity of the gas is heated, and the spe-
cific heat actually determined by direct measurement. This deter-
mination can be made as often as desired until the variation of spe-
cific heat and satisfactory average values have been determined. So
far it appears that the specific heat in a given installation is practi-
cally constant from day to day and from one time of day to another.
The fact that it is possible thus to determine the specific heat experi-
AN ELECTRIC GAS METER
681
mentally affords a most valuable check upon the specific heat deter-
mined by calculation from chemical analysis, since the methods used
in the latter are at best largely approximations.
6 As to dust and impurities collecting on the heater: A meter
now in operation for some months has been used for measuring in the
neighborhood of 100,000 cu. ft. per hr. of illuminating gas. The
Fig. 1
Gas Meter arranged with Calorimeter for Determining
Specific Heat
heater has been taken out once, and in handling it a small amount of
grease was found on the heater material. Otherwise the interior
of the meter was clean. In handling very impure gas it will of course
be necessary to clean out the meter occasionally, simply in order to
provide sufficient area for the passage of the required amount of gas.
All the heat generated in the heater necessarily goes into the gas.
682 DISCUSSION
The operation of heating and measuring difference of temperature is
all accomplished in a very short length of travel of the gas. This
perhaps answers the question regarding the heat-insulating effect of
deposits which may be formed on the heater. The rise of temperature
of the material of the heater is only 15 or 20 deg. fahr. This tem-
perature rise might be effected by a considerable deposit on the heater,
but the heat generated must necessarily be liberated from the heater
and given up to the gas, resulting in no error in gas measurement.
7 As to variable flow, the best evidence is presented by the curves
and calculations on the chart. The entire regularity of operation,
during experiments conducted under circumstances very favorable
to accuracy of observation seem to show that no error is introduced
by non-uniformity of flow. If such a cause of error existed, it seems
probable that it would have been found during experiments such as
have been made with this meter, covering the wide range of from
6000 cu. ft. to about 127,500 cu. ft. per hr. The meter is now being
built so that the gas passes in a vertical direction through the heater
and thermometers, and this would seem to favor regularity of dis-
tribution over the cross section of the passage. The change from
horizontal to vertical position was however dictated by convenience
of attachment and in order to obtain accessibility, although it seems
favorable to the above-mentioned consideration.
8 During the air tests extensive fluctuations of pressure took place,
due to the pulsations of the blower supplying the air. These were
so great at times as to cause the water to be thrown completely out
of the pressure gages, but the results obtained remained entirely
regular, as shown on the chart. A small meter has been used on a
single-acting three-cylinder four-cycle gas engine delivering from 30
h.p. to 60 h.p. The meter was constructed of sheet iron, and although
the pressure fluctuations'^ were such that the sides of the heater
"panted" continuously ,^{the^ measurement of gas was accurately
accomplished.
9 Answering Mr. Dodge's question regarding the amount of
energy required to measure steam with these meters, we are using 5
deg. fahr. temperature difference, which can be measured to an accur-
acy of 1 per cent and the energy required is 1 kw. per 1000 lb. of
steam per hr. Taking a water rate of 12 lb. per h.p-hr., 1 kw. would
measure the steam used for about 80 h.p.
10 Stated generally this meter seems to be particularly suitable
for the measurement of dry saturated or superheated gas, air or steam.
AN ELECTRIC GAS METER 683
The substance to be measured should be dry, but it may be of any
pressure and temperature which the materials of construction will
stand, and the measurement is independent of fluctuations of pres-
sure and temperature. The recording mechanism can be placed in
any convenient position, as, for instance, in an office, instead of near
the meter, and the graphical record is thus continually observable.
It is not necessary that a graphical record should be taken. An ordi-
nary integrating wattmeter showing the amount of energy it has re-
quired, to maintain the constant temperature difference of 5 deg.
between inlet and outlet of the meter, sufl&ces as a record of rate of
flow, though the variation is best shown by an autographic record.
No. 1258
TAN BARK AS A BOILER FUEL
By David Moffat Myers, New York
Associate Member
It is the object of this paper to give the chief characteristics of wet
spent tan as a boiler fuel. It is believed that tan bark contains a
larger percentage of moisture than any of the other moist fuels. It
is perhaps safe, therefore, to assume that its correct treatment as a
fuel may indicate to a certain extent the most efficient means of
dealing with other moist fuels.
2 The writer has been engaged in testing, remodeling, operating
and constructing steam plants depending partly or principally upon
tan bark for their fuel ; and in connection with this work the following
tests and observations were made, not only to determine the value
of the fuel, but principally to indicate the most favorable conditions
for its complete combustion and to improve the economy of the plants
visited.
PHYSICAL CONDITIONS
3 The nature of tan bark is too well known to require minute
description, but it may be stated, as relating directly to its use
as a fuel, that its condition in the fireroom varies with the method
of its preparation for leaching, and its treatment in the leaches.
Bark which has been finely disintegrated, and has been blown a long
distance through a pipe or flue, will reach the furnaces almost in the
state of a wet powder, difficult to burn. On the other hand, with less
thorough disintegration and a shorter fan drive, the tan will be in
larger pieces which allow a freer passage to the draft. The tan in the
fireroom varies in temperature according to its final treatment in the
leach house and according to the distance it is conveyed from the
leach to the furnaces. Under good conditions it often reaches the
fireroom at a temperature of 110 deg. fahr. The amount of moisture
Presented at the Annual Meeting, New York, (December 1909), of The
American Society of Mechanical Engineers.
686 TAN BARK AS A BOILER FUEL
in the tan varies with the leaching process and usually runs from 62
to 70 per cent. Oak tan is easier to burn than hemlock because it
, is harder and does not become soggy and pack on the grates, but ad-
mits the draft more freely than the hemlock.
CALORIMETER TESTS
4 The calorific tests given in Table 1 were made by Dr. Henry
C. Sherman of Columbia University, the samples of tan being dried
before burning.
TABLE 1 CALORIFIC VALUES OF DRIED SPENT HEMLOCK TAN
Sample No. B. t.d. per Podnd Sampij: No. B. t.u. per Pound
1
9406
7
9519
2
9378
i 8
9450
3
9463
! 9
9504
4
9500
10
9850
5
9516
11
9472
6
9482
Average
9504
Note — Samples 7, 8 and 9 each contained about 6 per cent of oak tan.
EFFECTS OF LEACHING
5 Calorific tests were made to determine the effect of leaching
on the fuel value of the tan, and it was found that the percentage of
tannin left in the bark does not affect its calorific value per dr}-- pound.
As a check a further test was made, comparing a sample of leached
bark and one of unleached bark. The result showed only a slight
difference of heat imits between the two and in fact a trifle more heat
in the leached bark per dry pound. In other words it can be stated
that the degree of leaching to which tan bark is subjected does not
affect its fuel value, except inasmuch as actual weight is subtracted
by the leaching process, so that a smaller quantity of fuel reaches the
fireroom.
6 As regards the loss of weight due to leaching, 100 lb. of air-dry
bark fed to the mill will produce 74.7 lb. of tan in the fireroom under
good leaching conditions. This tan contains 65 per cent moisture,
e. g., 100 lb. of air-dry bark ground at the mill will result in about
213 lb. of spent tan, containing 65 per cent moisture, in the fire-room;
i. e., the weight of the spent tan is 2.13 times the weight of the bark
ground.
TAN BARK AS A BOILER FUEL 687
MOISTURE
7 Ten of these samples contained moisture varying from 63.6 per
cent to 68.27 per cent, the average moisture being 65.5 per cent.
Many more moisture tests were made, but these ten samples are
typical and represent about the average conditions in this respect.
Taking the average calorific power of dried hemlock tan at 9500 B.t.u.,
and the average moisture at 65 per cent, and calculating the loss due
to moisture, we have: Loss by moisture = 0.65[(212° — t) + 966 +
0.48 (T — 212°)]. For t we have the temperature of the moist fuel
= 100 deg., and for T the temperature of the flue gases, which in tan
burning averages about 500 deg. Substituting these values gives:
Moisture loss = 0.65 [(212 - 100) + 966 +0.48(500 - 212)] =
660 B.t.u.
8 Since a pound of fuel is 65 per cent moisture, 0.35 lb., or the
dry amount, must evaporate 0.65 lb. moisture. The calorific value
of 0.35 lb. of dry tan is 0.35 of 9500 B.t.u., or 3325 B.t.u., which
is the total B.t.u. in a pound^f moist fuel. Subtracting the moisture
loss we have 3325 B.t.u. - -MS B.t.u. = lls^B.t.u., as the avail-
able heat in a pound of fuel as fired.
TAN BARK COMPARED TO COAL
9 As previously deduced, the weight of the bark as ground X 2.13
= the weight of the spent tan. Therefore the available heat for
boiler purposes obtained from^ pound of average air-dry hemlock
bark is 2.13 X -sl^ B.t.u. = -Sel© B.t.u. As compared to coal of
13,500 B.t.u., 1 ton of hemlock bark ground at mill = O.^tons coal.
CHEMICAL ANALYSIS
10 The chemical composition of dry tan is as follows:
Per cent
Mineral ash 1 . 42
Hydrogen 6 . 04
Carbon 51.80
Oxygen , 40.74
100.00
688 TAN BARK AS A BOILER FUEL
This sample was hemlock tan containing 6 per cent oak tan. When
dried at 110 deg. cent., it lost 66.77 moisture.
EVAPORATIVE TESTS
11 The actual evaporative power of tan was taken from the
results of 22 complete boiler tests burning tan alone. Eight of these
were thermal-efficiency tests.
12 Four complete thermal-efficiency tests were also made using
both tan and coal for fuel. The above tests were conducted in accord-
ance with the code of The American Society of Mechanical Engineers,
and in different parts of the country, including the South, the East
and the Middle West. The type of boiler in every case was the hori-
zontal tubular. The furnaces varied considerably in design and
economic results; but they were all of the Dutch oven type, set in
front of the boiler; and all were fired from the top through feed-
holes, with the exception of the furnace designed by the writer, the
results and design of which will be described later. All the above
tests were made under natural draft. Tables 2 and 3 show one of
the 24-hr. boiler and furnace efficiency tests, selected to show what
may be considered ordinary conditions and results in tan burning.
13 By designing and constructing a special furnace for the pur-
pose the writer obtained a thermal efficiency of furnace and boiler of
71.1 per cent, burning hemlock tan, efficiency based on available
heat of fuel. The principal features of this furnace, which was sub-
TABLE 2 BOILER TEST "X" TANNERY
Data and Results of Evaporative Test on Boiler No. 6 to Determine
Horsepower Developed and Efficiency of Existing Methods and Con-
[^ditions for Burning Wet Spent Hemlock Tan as Fuel for Steam Gbnera-
j^TION
Kind of fuel Wet spent hemlock tan
Kind of furnace Wide high oven, 2 rows firing eyes
Method of starting and stopping test Alternate
Grate surface, square feet ^ 92
Water heating surface, square feet 2089
Heating surface -r- grate surface 22.7
t _ ,
total quantities
Date of trial November 5 and 6, 1902
Duration of trial, hours 24
TAN BARK AS A BOILER FUEL 689
Weight of tan as fired, pounds 53,000
Percentage of moisture in tan 65 . 6
Total weight of dry tan consumed, pounds 18,232
Total weight of water fed to boiler, pounds 69,825
Factor of evaporation 1 . 1216
Equivalent water evaporated into steam from and at 212 deg., pounds 78,316
HOURLY QUANTITIES
Wet tan consumed per hr. pounds '. . 220S
Dry tan consumed per hr., pounds 1S23
Water evaporated per hr., pounds 2909
Equivalent evaporation per hr., from and at 212 deg., pounds 3263
AVERAGE TEMPERATURES, PRESSURES, ETC.
Steam pressure by gage, pounds 70.25
Temperature feed-water entering boiler 126 . 5
Temperature escaping gases from boiler 471
Force of draft between damper and boiler, inches water 0.39
Temperature outside air 64
Temperature of tan as fired 83
HORSEPOWER
Horsepower developed 94 . 0
Rated horsepower, at 15 sq. ft • 139
Percentage rated horsepower developed 68
ECONOMIC RESULTS
Water evaporated under actual conditions per lb. tan as fired, pounds 1 .32
Equivalent evaporation from and at 212 deg. per lb. tan as fired, pounds .... 1 . 48
Equivalent evaporation from and at 212 deg. per lb. dry tan, pounds 4 . 30
EFFICIENCY
Calorific power of dry tan per lb., British thermal units 9463
Percentage moisture in tan as fired 65 . 6
Available heat in 1 lb. of wet tan as fixed, after subtracting loss due to
evaporating moisture, British thermal units 2443
Efficiency of boiler including furnace, based on available heat in fuel,
per cent 58 . 2
Efficiency of boiler including furnace, based on total heat in fuel, per cent ....43.9
2 rows of eyes, 3 eyes in each row. Eyes were fired according to regular
method employed at this tannery, i. e., they were refilled aa they burned low,
a large amount of tan being fired each time. 51 eyes were fired during the
24-hr. test.
690
TAN BABK AS A BOILEB FUEL
TABLE 3 FLUE-GAS ANALYSIS
Daupsb Widb Opxn
No.
Time
CO2
0
CO
Firing
%
%
%
1
9:45
10.8
7.5
30 minutes after
2
10:21
4.2
12.3
30 minutes after
3
11 :48
11.6
7.0
14 minutes after
4
12 :29
5.0
15.0
22 minutes after
5
1:37
4.6
15.1
0.5
While firing
6
2:06
5.0
13.0
0.3
14 minutes after
7
3 :05
9.2
40 minutes after
8
4 :08
9.8
8.4
While firing
9
5:30
12.6
6.6
0.1
6 minutes after
10
6 :20
12.6
7.0
0.5
26 minutes after
11
6:55
11.6
7.4
0.2
25 minutes after
12
7 : 50
13.0
6.0
0.6
20 minutes after
13
8:55
13.0
7.2
0.1
While firing
14
10:09
11.0
9.1
6 minutes after
15
11 :07
12.5
6.9
0.1
29 minutes after
16
12 :23
11.5
7.5
0.8
23 minutes after
17
1 :07
7.6
11.8
1.1
67 minutes after
18
2 : 10
12.5
7.7
0.0
24 minutes after
19
2 :50
6.0
13.5
20 minutes after
20
4 : 00
4.3
15.7
58 minutes after
21
5:30
3.0
16.4
0.3
While firing
22
6 :02
11.0
8.8
0.0
27 minutes after
23
6:37
9.0
9.8
0.0
6 minutes after
24
7:35
8.7
11.0
0.2
9 minutes after
Average
9.2
8.2
0.32
sequently installed in a number of plants, were: a Large combustion
space over the burning fuel, h Automatic stoking with rotating
comb shafts, c Oppositely inclined grate surfaces converging down-
ward to a set of shaking and dumping grates, d Drying on dead
plates over which fuel passed before receiving air supply, e Con-
centrated draft from opposing grate surfaces to a focus of combustion,
caused by parallel spacing of longitudinal flat grate bars with beveled
edges. / Reverberating draft action resulting from concentrated
draft-currents and the curvature of the arch, thus directing flames
back upon dead plates. The full results of a test on this furnace are
given in Table 4.
EFFECT OF PRESSING AND BURNING MIXED WITH COAL
14 It was desirable to learn the exact economic effect of burning
coal with pressed tan and also the result of pressing the tan to elim-
inate some of the moisture. For these purposes three thermal-
TAN BARK AS A BOILER FUEL 691
TABLE 4 BOILER TEST: AUTOMATIC STOKER FURNACE DESIGNED BY THE
WRITER
Data and Results of Evaporative Test on Boiler No. 1 to Determine
Horsepower Developed, and Efficiency op Boiler with Automatic
Tan Stoker
Kind of fuel Wet spent hemlock tan
Kind of furnace Automatic stoker
Method of starting and stopping test Alternate
Grate surface, square feet 61.5
Water heating surface, square feet 1795
TOTAL QUANTITIES
Date of trial August 14, 1903
Duration of trial, hours 12
Weight of tan as fired, pounds 18,600
Percentage moisture in tan 65 . 3
Total weight of dry tan consumed, pounds 6449
Total weight of water fed to boiler, pounds 33,004
Factor of evaporation 1 . 085
Equivalent water evaporated into steam from and at 212 deg., pounds 35,809
HOURLY QUANTITIES
Wet tan consumed per hr., pounds 1550
Dry tan consumed per hr., pounds 537
Water evaporated per hr., pounds 2750
Equivalent evaporation per hr. from and at 212 deg., pounds 2984
AVERAGE TEMPERATURES, PRESSURES, ETC.
Steam pressure by gage, poimds 43
Temperature feed-water entering boiler 155
Temperature inside furnace 1100
Temperature in combustion chamber 1475
Temperature gases escaping from boiler 493
Temperature outside air 71
Temperature of air entering ash pit 78
Temperature of tan as fired 101
Force of draft between damper and boiler, inches water 0.4
HORSE POWER
Horsepower developed 86 . 5
Rated horsepower of boiler at 15 sq. ft., per h.p 120
Percentage rated horsepower developed 72 . 1
ECONOMIC RESULTS
Water evaporated under actual conditions per lb. of tan as fired, pounds ....1.77
Equivalent evaporation from and at 212 deg. per lb. of tan as fired, pounds. . . 1 .93
Equivalent evaporation from and at 212 deg. par lb. of dry tan, pounds. . . .5.55
692
TAN BARK AS A BOILER FUEL
EFFICIENCY
Calorific value of dry tan per lb., B.t.u 9850
Percentage of moisture in tan as fired 65.33
Available heat per lb. of tan as fired, B.t.u 2623
Efficiency of boiler, including furnace, based on available heat in fuel, per
cent 71.1
Efficiency of boiler, including furnace, based on total heat in fuel, per cent ... 54 . 4
Firing was even and continuous with automatic stoking. Bed of fuel maintained
on grates was about 9 in. thick.
TABLE 5 FLUE GAS ANALYSES
Damper Open
No.
Time
CO2
%
0
%
CO
%
1
2 :25
4:30
7: 15
9:05
11 : 10
11
12.8
12
13
13
7.5
5.6
7.8
6.2
6.4
n
2
0
3
0
4
0
5
0
12.4
6.7
0
TABLE 6 ANALYSIS OF SPENT BARK
Chemist's No. S.-736, Analyzed August 17, 1903
%
Moisture
Total solids
Total soluble solids
65.33
5.11
4.24
Non-tannins. . . .
Available tannin
Reds . .
2.35
1.89
0.87
efficiency tests were conducted, all on the same boiler and furnace,
the conditions as far as possible being the same in all tests, and the
furnace operated by the • same firemen. The furnaces and boiler
were of the general type already referred to. The combustion cham-
ber under the boiler differed by having a system of air admissioo
through the bridge wall. The opening in the wall for admitting air
to this system was set one-fourth open in the first test and remained
so in all three tests, each of ten hours duration. An effort was made
in each of the two tests without coal to equal the boiler output of the
test in which coal was used; but as shown below this could not be done.
TAN BARK AS A BOILER FUEL
693
15 The furnace arch was equipped with two rows of firing-eyes of
three eyes each, and two eyes on the same side of the furnace were
usually fired at a time. The firing was heavy, the cone of tan nearly
reaching the arch each time after firing. The furnace had a grate
surface of 11 ft. 4 in. by 7 ft. 0 in. = 79.3 sq.ft., and the perpendicu-
lar distance from the grate surface to the highest point of the arch
inside was 48 in. The boiler was 6 ft. by 18 ft. and contained 2089
TABLE 7 COMPARATIVE RESULTS OF THREE TESTS
Test No. 1
Test No. 2 Test No. 3
Thermal efficiency of boiler including furnace
based on available heat in fuel, per cent
Evaporation from and at 212 deg. per lb. tan as
fired, pounds
Evaporation from and at 212 deg. per lb., dry
tan , pounds
Percentage rated h.p. developed ,
Type of furnace
Heating surface -H grate surface
Number of firing-eyes
Depth of combustion chamber.inches
Kind of firing
Firing intervals, minutes
Kind of fuel
Moisture in tan, per cent
Steam pressure, pounds
Temperature feed-water
Flue temperature ,
Force of draft in uptake, inches water ,
Duration of trial, hours
Flue-gas analyses, averages per cent CO2
Per cent O
Per cent CO
63.4
1.98
5.43
135.5
26.3
6
48
deep
18.8
pressed tan
and coal
60.4
76
163
561
0.51
10
13.8
4.9
0.15
59.4
1.90
4.71
92.
wide and high
26.3
6
48
deep
18.8
pressed tan
59.6
74
160
486
0.49
10
10.9
7.8
0.93
59.
1.65
4.54
72.4
26.3
6
48
deep
19.4
unpressed tan
63.6
63.4
161
445
0.43
10
10.5
8.6
0.68
sq.ft. heating surface, the ratio of heating surface to grate surface
being 26.3 to 1.
16 Nine flue-gas analyses were made during test No. 1, seven in
each of the following tests, and the averages of these results are shown
in Table 7 which gives the most essential results of the three tests.
17 The results of this series of tests show:
a That the burning of coal with pressed tan increased the
thermal efficiency from 59.4 per cent to 63.4 per cent.
b That the burning of coal with pressed tan increased the
boiler output from 92 per cent of rated capacity to 135.5
694 TAN BARK AS A BOILER FUEL
per cent of rated capacity, the force of draft differing by
only 0.02 in. water-gage.
c That the degree of combustion as indicated by the amount
of CO2 was raised from 10.9 per cent COj with pressed tan
to 13.8 per cent CO2 with pressed tan and coal.
d That the flue temperature was increased from 445 deg. with
unpressed tan and 486 with pressed tan to 561 deg. with
pressed tan and coal.
cf That the thermal efficiency based on available heat obtained
with pressed tan was 59.4 per cent as compared to 59
per cent with unpressed tan.
/ That the burning of pressed tan gave a boiler output of
92 per cent rated capacity as compared to 72.4 per cent
with unpressed tan, the draft being only 0.04 in. water-
gage stronger in the test of greater capacity.
g That since the thermal efficiency was practically the same
with both pressed and unpressed tan the principal advan-
tage of pressed over unpressed tan in this particular
case lies in the intrinsically greater calorific value of
the pressed tan owing to its reduced moisture. In this
particular case, in which the unpressed tan contained 63.6
per cent and the pressed tan 59,6 per cent moisture, the
actual gain in available heat-units was 4.5 per cent over
the unpressed tan or IJ per cent gain in heat value for
each per cent decrease in moisture.
TAN PRESSES
18 Tan presses for reducing the moisture in actual practice result
in a total economic gain of only about 7 per cent; and against this
must be charged the power to run them, maintenance, repairs and
often the disadvantage of noise and vibration in the fireroom.
19 The question of installing a press must be determined prin-
cipally by local conditions, such as amount of tan available, amount
and proportion of coal used and its cost, intelligence of labor, aver-
age moisture in the tan as delivered to the fireroom, design of furnace
and grate, cost of changing such design, etc. Especially when the
firing is poor a press is of much value, for certain firemen who can
get excellent steaming with pressed tan cannot raise steam with
unpressed tan. When a press is introduced for tan burning and the
moisture is reduced, the rate of combustion with given grate and draft
TAN BARK AS A BOILER FUEL 695
is increased. Therefore to burn the pressed tan in the time required
to bum the wet tan, it is necessary to reduce the grate surface pro-
portionately.
20 For example, two complete 24-hr. evaporative tests were
made on the same boiler and furnace, with the same firemen and with
the same conditions as nearly as possible, the first with unpressed and
the second with pressed tan, with the following results:
Test No. 1, without press, 64.4 per cent moisture, consumption
14,565 lb. dry tan with 0.42 draft.
Test No. 2, with press 60.4 per cent moisture, consumption
19,879 lb. dry tan with 0.38 draft.
Thus the rate of combustion on the same grate was increased 36.5
per cent. (It may here be noted that the boiler-horsepower developed
in the first test was 89.3 h.p., and in the test with pressed tan 116.8
h.p.)
21 Therefore in this case to burn the tan in the same time as
formerly it would be necessary to reduce the grate surface about 27
per cent. This is based on the above drop in moisture from 64.4 per
cent to 60.4 per cent, or 4 per cent. The drop in moisture with
fair pressing will run between 6 and 7 per cent, representing a
total gain of available heat of 7 to 8 per cent. With good pressing
the grate surface may be properly reduced one-third.
22 There have been cases of failure with pressing the tan simply
because the precaution of reducing grate surface was not taken.
Tannery superintendents have said, " We have tried presses and had
to throw them out, because the tan would not last throughout the
day, and we had to use more coal to make up the deficiency. "
The point here is that a tannery will almost invariably consume or
waste every pound of steam its boilers will supply. With the press
in operation the steaming was increased proportionately as long as
the tan " lasted, " no provision being made to correct the grate sur-
face, and consequently the result would be exactly as stated by the
superintendent. This has led to, or confirmed, the erroneous opinion
held by some that the moisture in tan aids its combustion.
EFFECT OF SMALL COMBUSTION SPACE OVER THE FIRE
23 A special test on a low-arched furnace was made to obtain
comparison with high arches. A furnace containing nine feed-holes,
696 TAN BARK AS A BOILER FUEL
and having a height from arch to grate of 26 in., was tested, with
the following resultsi;
Efficiency of boiler and grate, per cent 26 . 4
Evaporation from and at 212 deg. per lb. tan as fired 0. 63
Percentage rated h.p. developed 39
Heating surface h- grate surface 28
Depth combustion chamber under boiler 3'6"
Intervals between firing, minutes 6.3
Kind of tan Hemlock
Percentage moisture in tan 67 . 2
Temperature flue-gases 422
Draft at damper, inches 0.37
24 These miserable results are to be attributed to the low arch
and the consequently inadequate combustion-space between the
fuel and the arch of the furnace. Owing to this small space the vel-
ocity of draft and gases was very high, thus removing the volatile
fuel-gases from the furnace before their combustion could be fairly
begun, and allowing no heating or mixing action in the furnace, and
when they struck the boiler the combustion was checked or stopped.
25 The worst evaporative results ever obtained on various de-
signs of higher furnaces, on tan alone and with poor firing, are at least
100 per cent better than this, and these other furnaces ran from 3 ft.
0 in. to 5 ft. and 6 ft. high; it is perhaps significant that the highest
thermal efficiency was obtained in the furnace with the greatest
amount of free combustion space between the fuel bed and the top
of the arch.
26 During recent years tanneries have been using more extracts
and less bark for their liquors. At the same time a greater amount
of machinery has been introduced and it has been found beneficial
to use more heat in leaching. All these changes have resulted in
greater steam requirements and less spent tan bark for its generation.
Therefore more coal is required in the boiler room and stricter econ-
omy of fuel is necessary.
BURNING A MIXTURE OF TAN AND COAL
27 Among other methods of burning the two fuels, the ordinary
coal-burning setting has been tried. The grates are usually of the
shaking variety and are set directly under the boiler, and generally
only 24 in. to 30 in. beneath the shell of the boiler. The coal and tan
'Average flue-gas analyses: CO2, 4.3 per cent; O, 11 per cent; CO, 20 per cent.
TAN BARK AS A BOILER FUEL 697
are fired alternately by hand. Owing to the large volume of tan
as compared to its heating value (spent tan weighs about 33 lb. per
cu. ft.), the fire doors have to be opened very often, thus admitting
a great excess of cold air to the fire. The fire bed has a strong tend-
ency to burn through in spots and form blowholes. The tempera-
ture-retaining effect of the Dutch oven arch is lost, and instead, the
fuel gases come into direct contact with the shell of the boiler, which
cools them before combustion is fairly under way. The result of every
such case the writer has observed has been a dull, smoky fire of greatly
varying temperature and never good combustion. In a certain
plant where this method was thoroughly tried out, it proved so unsuc-
cessful that it was discontinued and the grates were replaced by Dutch
ovens in which the mixture of tan and coal was burned, the fuels being
mixed before firing.
28 The above illustration affords a good comparison of combus-
tion with and without a brick arch over the fire bed with hand firing,
and goes to show that a brick or refractory arch is a necessity for good
combustion, when the two fuels are so related in quantity that their
heat values are about equal. As a further demonstration, experi-
ments with automatic stoking may be cited. The mixed fuel was fed
into a chain-grate stoker operating on induced draft. This was done
under the most favorable conditions possible and under the super-
vision of an engineer representing the chain-grate company, with
the assistance of the writer. The steam pressure and the fur-
nace temperature dropped rapidly, and it was soon necessary to
eliminate the tan and feed coal alone, to prevent the fire from going
out. Experiments with different methods of mixing and feeding the
two fuels all resulted in killing the fire.
29 Like experiments were conducted on a Detroit stoker set in
front of boiler and having a brick arch completely over the fire. No
scientific tests were conducted, but the mixed fuel easily carried the
load formerly carried by the coal, it being necessary only to increase
the speed of the stoker. The steam pressure was increased and as far
as the eye could detect there was not the slightest decrease of furnace
temperature, the combustion being clear and to all appearances
excellent.
30 From these experiments it seems entirely safe to state that
for efficient combustion of a mixture of tan and coal in ratio by weight
of 5.2 to 1 a refractory arch over the greater part of the fire, and pref-
erably over the entire fire-bed, is not only advisable but necessary.
It follows that tan bark of the usual moisture and heat content can-
698 TAN BAEK AS A BOILER FUEL
not be burned without the application of a refractory arch, or some
device of similar function, such as a large combustion chamber lined
with refractory material.
31 Wood is sometimes burned in combination with tan in the
same furnace. Good results are obtained when the wood is ground
to about the same fineness as the tan, when it is known in the saw-
mill districts as " hog feed. " When wood in slab or log form is
burned in the same furnace with tan the results naturally are usually
very poor, owing to blowholes formed in the fire bed and consequent
large excess of air. A case of this description came under the writer's
observation. "Edgings" from a nearby sawmill cost less than coal
per 1000 B.t.u., and it was the custom to fire the edgings by hand into
the front fire doors of the tan furnaces, the tan being fired from above
in the usual manner. A large amount of labor was required and the
furnace fronts required constant repairing. By substituting coal
for the wood in this case, and mixing with the tan before firing, a
very good saving was effected in fuel, labor and repairs.
32 In regard to draft for tan burning, in 26 evaporative tests on
tan-burning furnaces the average force of draft between boiler and
damper was 0.45 in. water gage.
33 The average draft, in 13 tests on tan where less than the rated
capacity of the boiler was developed, was 0.42 in. water gage between
boiler and damper.
DRAFT AND GRATE SURFACE
34 The average draft, in 12 tests on tan where rated capacity up
to 60 per cent over-capacity was developed, was 0.47 in. water gage.
This comparison is significant only in a very general way, owing to the
great variety of ratios of heating to grate surfaces in the different tests,
and the different methods and time intervals of firing. The highest
force of draft found in any case between damper and boiler with
chimney draft, furnaces burning tan only, was 0.63 in., and the lowest
0.32 in.
35 The lightest draft with which rated capacity was obtained was
0.32 in., and in this case the ratio of heating surface to grate surface
was 21.8 to 1. In the test where 160 per cent rated capacity was
developed the force of draft was 0.51 in. and the ratio of heating sur-
face to grate "surface was 19.3 to 1. The greatest ratio of heating
surface to grate surface which developed full rated capacity was 29.9
to 1, giving 110 per cent rated capacity with force of draft in uptake
TAN BARK AS A BOILER FUEL 699
0.42 in. The firing in this case was excellent with intervals of 7.4
min.
36 Owing to space required, facility in cleaning fires, minimum
radiation from furnace and expense of building and repairing, it is
desirable to make the furnace grate area as small as possible and still
conform to capacity requirements. It is therefore necessary to know
the boiler-horsepower developed per square foot of grate surface.
The force of draft, method of firing and ratio of heating surface to
grate surface, are all factors in this result, also the moisture in the
tan bark. Table 8 gives the most important factors with results of
nine actual tests.
TABLE 8
RESULTS OF NINE TAN-BURNING TESTS
Test No.
Tan
MOIST-
DRE
%
Draft
Inches
H.S.
G.S.
Firing In-
TBHVAia
Minutes
B.H.P.
PER
SQ. FT.
G. S.
1
. . . . Oak
0.63
0.63
29
15
2.35
2
. . . . Hemlock
0.61
19.3
8
2.07
3
. ... Oak
0.64
0.68 ,
29
5
2.37
4
. ... Oak
0.64
0.61 i
34.8
1.53
5
. . . . Hemlock
0.65
0.61
17.4
50.8
0.92
6
. . . . Hemlock
0.32
21.8
5.2
1.51
7
. . . . Hemlock
0.68
0.39
29.9
7.4
2.34
8
. . . . Hemlock
0.64
0.42
22.7
30
0.97
9
. . . . Hemlock
0.65
0.40
29.2
Continuous
1.31
37 Inspection of this table shows that one factor above all the
others influences the b.h.p. developed per square foot of grate, and
this factor is the method of firing. Thus the tests in which the time
intervals between firing are small show a marked increase in b.h.p. per
square foot of grate. This test is upheld by many other tests of
the writer. From the table we may take the following average
results :
For oak tan with 60 per cent moisture, 0.61 in. draft, in up-
take of horizontal tubular boiler, and good firing of 15
miu. or less between firing, 2.08 b.h.p. per sq.ft. of grate
surface.
For hemlock tan with 65^ per cent moisture, 0.44 in. draft in
uptake of horizontal tubular boiler, and good firing of say
17 min. or less between firing, 1.52 b.h.p. per sq. ft. of
grate surface.
700
TAN BARK AS A BOILER FUEL
GRATE SURFACE FOR COAL AND TAN
38 The above results are for burning tan alone. At the present
time, however, it is very general practice, as previously set forth, to
burn coal with the tan. The best results in most cases are obtained
when the coal is mixed evenly with the tan, and the mixture burned
in a Dutch oven with shaking and dumping grates. This practice
modifies the boiler horsepower per square foot of grate. The grate
surface often allowed is at the rate of about 3j b.h.p. per sq. ft. of
grate surface, the ratio of heating surface to grate surface ranging
from 50 to 1 to 33 to 1.
TABLE 9 RESULTS OF TESTS OF BURNING COAL WITH TAN IN DUTCH OVENS
Firing
Weight of
B.H.P
Tan
Uptake
H. S.
Inter-
Coal to
PER
Test
Tan
Grates
MoiSTDBE
%
Draft
G. S.
vals Min-
utes
Weight of
Tan as Fired
SQ. FT.
OF
Grate
a
Hemlock
Shaking
68
0.40
33.5
20
Ito 8.9
2.52
b
Hemlock
Stationary
60.4
0.51
28.1
19
Ito 6.1
2.53
c
Hemlock
Stationary
66.7
0.67
37.3
6.3
1 to 46.4
2.70
*d
Oak
Shaking
67.4
0.7
29.4
47
Ito 2.9
2.66
e
Hemlock
Stationary
66.5
1.05
32.8
Ito 22.8
1 1.33
♦Tubular Boiler, coal fired independently in front d^ors of double arched tan furnace. The
47-min. interval refers to the tan only.
39 Of course, as with tan alone, the development per square foot
of grate depends upon the same factors, with the additional factors
of the proportion of coal mixed with the tan, and the B.t.u. and charac-
ter of the coal. The figures in Table 9, selected at random, show
the results in terms of b.h.p. per square foot of grate surf ace obtained
in actual tests of burning coal with tan in Dutch ovens. It is of in-
terest here to note that when the ratio of weight of coal to weight of tan
as fired is 1 to 5.2, then the heat developed by each fuel is practically
the same. This is figuring on 13,000 B.t.u. coal, and "tan as fired"
containing 2500 B.t.u. available heat, which is good average value.
40 In further reference to draft, a test which showed excellent
results was made on a plant operating on induced fan draft. A small
proportion of coal was used in the tan furnaces and a thermal efficiency
of 66.1 per cent was found under regular working conditions, the CO^,
in the flue gases averaging nearly 12 per cent. The writer has also
observed an oak tan-burning furnace operated temporarily on fan
TAN BARK AS A BOILER FUEL 701
draft with closed ashpit. The combustion was apparently very good;
a hot fire was maintained, but additional care was required to keep
blow-holes from forming on the grates.
41 A reliable report has also been obtained on the installation of
a steam-induced forced-draft system, applied under the grate of a
tan-burning furnace. According to the report, the steam used for
the draft condensed on the grate bars, with low temperature due
to contact with the large mass of wet tan, and the concern that made
the installation after considerable experimenting removed the appli-
ance and gave up tan burning.
GRATES
42 Grates for tan burning are made in different patterns, but
usually contain from 20 per cent to 30 per cent air space, the actual
opening between the bars being ^ in. to ^ in. Larger spaces than
these allow the tart to fall through into the ashpit. A shaking grate
is hardly necessary for burning tan alone, owing to the small percent-
age of ash. No clinker is formed, only a very fragile crust on the
grate obtains in three or four hours. In some cases fires are cleaned
only once in 12 hr. with good results. Unlike a coal fire, a tan fire
should be shaken or disturbed very little. If a slice bar is used on a
brightly burning tan fire, dense smoke results and the flame is killed
for some time.
43 The temperature in the throat of a properly designed tan
furnace, burning tan alone, will reach 1500 deg. fahr.
44 The depth of fuel on the grate varies with the design of the
furnace and the method of firing. In ordinary practice the tops of
the cones of tan directly beneath the firing eyes vary from 2 ft. to 5
ft. above the grate surface, while the depth of tan where the cones meet
will be from 6 in. to 18 in., depending on the design and firing of the
furnace.
45 When spent hemlock tan forms a cone from the conveyor-
discharge on the fireroom floor it forms an angle of slide of about 55
deg. to the horizontal. Inside a hot furnace, however, this angle
is about 45 deg., and in laying out a tan furnace the latter angle is
used in figuring the distribution of fuel on the grate. The angle of
slide of tan on a sloping grate is 39 deg. to 41 deg., and these angles
were used in designing the automatic tan [furnace preferred to pre-
viously.
46 Among other experiments with this fuel the writer has made
702 TAN BAHK AS A BOILER FUEL
tests to compare results of double settings with single settings. By
double setting is meant a setting so arranged that two furnaces dis-
charge their gases into a common combustion chamber beneath two
or more boilers. Tests showed no tangible results in favor of either
method as regards evaporation or fuel economy. It is possible
however to imagine a case where the fire in a furnace might be in poor
condition, when if the setting were a double one, the second furnace
might maintain in the common combustion chamber a temperature
suflficiently high to ignite and burn fuel gases discharged from the dull
fire which otherwise would escape unburned. It is the opinion of the
writer, however, that owing to the impossibility of laying off a single
boiler for cleaning or repairs, a double setting is less desirable than
a single setting. Furthermore, good firing and handling would in
any case offset any possible advantage of the double setting.
47 The combustion of tan, as indicated by flue-gas analyses
made with an Orsat apparatus, compares most favorably with the
combustion of coal In factory plants where the same amount of atten-
tion is paid to the boiler room. In fact, the percentage of COj runs
higher than in the ordinary coal-burning plant. For instance,
the average COj of seventeen evaporative tests on different tan-burn-
ing furnaces was 11.3 per cent; the lowest average CO2 in any of the
tests was 4.3 per cent, which represented unusually and abnormally
poor conditions; and the highest average CO2 in any one of them was
16.9 per cent. The CO averages about 0.5 to 0.6 per cent and ranges
from practically zero up to 2 per cent.
SUMMARY
48 The following statements concerning tan bark as a boiler fuel
have been demonstrated by the foregoing data and confirmed by
constant checking in actual practice:
49 Moisture. In condition for firing, wet spent hemlock tan
usually contains close to 65 per cent of moisture.
50 Available B.t.u. Bomb calorimeter tests on many samples
of spent hemlock tan give an average value of about 9500 B.t.u. per lb.,
sample being dried before burning. The available heat per pound
as fired, after subtracting moisture loss, is about 3o6o B.t.u.
51 Ratio weights. Since the weight of the spent tan in the fire
room is 2.13 times the weight of bark ground at the mill, 1 lb. of
ground bark produces an available heat value of 5676'B.t.u. Hence,
TAN BARE AS A BOILER FUEL 703
as compared to coal of 13,500 B.t.u., one ton of hemlock bark is
equivalent to 0.4« ton coal.
52 Effect of leaching on B.t.u. The heat value of spent hemlock
tan is not affected by the degree of leaching, except inasmuch as the
actual weight is affected.
53 Chemical corn-position. The chemical composition of hemlock
tan (sample containing 6 per cent of oak) was:
C = 51.8, H = 6.04,0 = 40.74, Mineral ash = 1.42 per cent
As a fuel analysis this is of interest as regards the small amount of
ash and the large amount of oxygen.
54 Boiler tests. The actual operation of tan bark as a fuel is
shown by complete boiler tests made under working conditions,
records of which are given above.
55 Improved efficiency. A considerable improvement in efficiency
was produced by a specially designed furnace providing automatic
feeding, large combustion space over fuel and special draft admis-
sion.
56 Tan Presses. The use of presses for reducing the moisture
in the tan before firing may be good economy if the amount of tan
compared to the amount of coal used is considerable, and providing
the grate surface is properly reduced to meet the demands of the
more rapid combustion. In practice the grate surface is sometimes
reduced one-third on the introduction of tan presses.
57 Under usual tan-burning conditions thetotalgain in available
heat from the use of a press is about IJ per cent for each per cent
drop of moisture content, and this drop rarely exceeds 7 percent, i. e.,
a reduction of moisture from say 65 to 58 per cent.
58 The writer is of the opinion that the combined boiler and fur-
nace efficiency would naturally rise considerably when tan presses
are used, owing to increased furnace temperature and better combus-
tion, although as shown by Table 7 the thermal efficiency was raised
only from 59 per cent to 59.4 per cent, the moisture in the tan being
63.6 per cent and 59.6 per cent in the two tests respectively. This
may be explained, at least partially, by the fact that the grate sur-
face was not reduced for burning the pressed tan.
~l59 Addition of coal. As a result of special comparative tests,
made under similar conditions on the same furnace and boiler, the
addition of about one pound of coal to six of pressed tan increased
the combined furnace and boiJer thermal efficiency from 59.4 per
704 TAN BARK AS A BOILER FUEL
cent to 63.4 per cent. The addition of coal also resulted in raising
the average CO2 in the flue gases from 10.9 to 13.8 per cent. The
percentage rated capacity of the boiler was increased from 92 per
cent (without coal) to 135.5 per cent with coal.
60 Although the above precise statements were obtained from a
single pair of accurate tests, the writer's subsequent experiences in
many plants confirm the results as generally typical; and it may be
safely stated that under usual working conditions the addition of coal
to tan produces a decided increase in degree of combustion and a con-
sequent rise of thermal efficiency. The increase of boiler output
attending the use of coal with tan is also a factor of economy not to be
disregarded.
61 Ample combustion space. One of the most important factors
in designing furnaces for tan-burning, is that of ample combustion
space. In usual tan-burning practice this means a high arch over
the furnace. Low-arched furnaces are conducive to bad combus-
tion, whereas furnaces providing an arch at a considerable height
above the fuel on the grates give decidedly better combustion and
higher efficiency. (See tests and discussion, par. 23 et seq.)
62 Refractory arch. A refractory arch or similar combustion
arrangement is necessary for the economic combustion of tan. Tan
in its usual condition cannot be burned in a common coal-burning
setting without an arch separating the fire from the cooling surface of
the boiler shell or tubes. Even when tan is enriched with coal to a
point where the two fuels form equal heating values (about 1 lb. coal
to 5.2 lb. wet spent tan) good combustion cannot be obtained in an
ordinary coal setting. This holds true both for hand-firing and auto-
matic stoking. (See tests and discussion, par. 27 et seq.)
63 Draft. The force of draft required in general practice for tan
burning is somewhat higher than that required for producing the
same boiler output with bituminous coal. No definite figures can be
given for force of draft as related to tan burned per square foot of
grate surface, owing to the widely varying designs of furnaces and
methods of firing. It may be stated, however, that for full capacity
not less than ^-in. draft should be provided for in the uptake of a
horizontal tubular boiler, when the boiler is set so that the gases
flow under the shell and through the tubes directly to the stack or
breechen.
64 Horsepower per square foot of grate. The boiler-horse power
obtainable per square foot of grate surface, depending upon the rate of
combustion of the tan, is an extremely variable quantity; but aver-
TAN BARK AS A BOILER FUEL 705
aging the results of nine typical boiler tests it may be said that under
usual conditions for oak tan the boiler-horsepower per square foot
of grate is about 2.08; and for hemlock tan, 1.5 boiler-horsepower
per square foot of grate is an average figure. As demonstrated
by results of tests given in Table 8, the one factor which above
all others influences the boiler output per square foot of grate is
the method and comparative excellence of the firing, very short
firing-intervals producing the higher output results.
65 Horsepower with coal and taji. When coal is burned in the
same furnace with the tan the development per square foot of grate
varies of course with the richness of the mixture and with the other
usual conditions, and it is best in designing furnaces for any par-
ticular plant to use the results of tests made under similar conditions.
In four cases out of five selected at random the boiler-horsepower
per square foot of grate was 2.5 to 2.7.
06 Draft area and air spacing. Grates for tan burning provide
20 to 30 per cent draft area and the spacing is from ^ in. to ^ in.
67 Cliiiker and ash. Tan forms no hard clinker, only a light fragile
crust, and a very small amount of ash (less than 1^ per cent based on
dry matter), and a shaking grate is unnecessary except when coal
is fired with the tan.
68 Furnace temperature. Excellent combustion of tan has given
a temperature of 1500 deg. fahr., in the throat of the furnace,
69 Single vs. double setting. As a result of comparative tests made
on single vs. double settings for tan burning, the writer believes that
the single setting should have the decided preference. A possible
though not a decided improvement in combustion is sometimes
obtainable with the double setting, but this is more than offset,
especially in the small plant, by the advantages of convenience for
cleaning, repairs and adaptation to load given by the single setting.
70 Use of wood with tan. When wood is used as a supplementary
fuel to tan and fired in the same furnace, the best results are obtained
by first grinding the wood to the size known as "hog feed " The
firing of wood in log or slab form is disastrous both to the furnaces
and to combustion.
7 1 Flue gas analyses. Basing comparison on flue gas analyses, tan
burns with a higher combustion than coal, under equally fav^orable
conditions. The large amount of moisture in the tan produces a
comparatively low furnace temperature, even with good chemical
combustion, and acts against an equally high combined efficiency of
urnace and boiler.
706
TAN BARK AS A BOILER FUEL
72 The highest thermal-eflSciency test obtained gave 71.1 per cent
efficiency (on boiler and furnace, based on available heat in tan), and
fiue-gas analyses during this test averaged 12.4 per cent COj. The
approximate relation of CO2 to efficiency as obtained under actual
running conditions is shown by the curve, Fig. 1, which was plotted
49:>
it
i i
a 13
I12
/^
^
^
48
©
480
J
G
)
-
u
'44
5
— 1
■—'
®4.'i7
a ^0
0
i- q
s4
71
©
50
B
—
—
—
~
58 59 60 61 62 63 64 65 66 67 68 69 70
Combined thermal efficiencv of boiler and furnace
71
Fig. 1 Curve Showing Relation of CO g to Efficiency in Tests on
Tan Alone
from seven thermal efficiency tests made on various designs of fur-
naces burning tan alone. The figure given at each point plotted is the
temperature of the escaping gases during the test.
73 The following illustrations, reproduced from working drawings
and sketches, will give some idea ofithe construction of furnaces for
burning spent tan bark, sawdust, and bagasse.
74 Fig. 2 is a working'drawing of the self-feeding tan-burning fur-
nace designed by the writer and referred to in Par. 12 and Table 4.
75 Fig. 3 shows the application to bagasse burning of special
grates shown in Fig. 4. The stokers are done away^with in this
case, the fuel being fed by gravity to the feed chutes with weighted
flaps which are used all over the islands of Cuba and Porto Rico.
This burner has not yet been apphed to bagasse burning.
76 Fig. 4 shows the construction of the grate bars, which provide
the horizontal draft opening tending to produce the draft action
referred to in Par. 13.
77 Fig. 5 and Fig. 6 show the types of tan furnaces found by the
writer in common use throughout the country. Fig. 5 shows what
was known as the old Hoyt furnace. It was originally designed when
tan bark was so plentiful that it was necessary to burn it. The
writer has found these furnaces with inside lengths as great as 24 ft.
TAN BARK AS A BOILER FUEL
707
o
<
•z
ei
o
M
»
m
<
Eh
09
PS
H
H
fa
O
Z
o
H
>
H
55
O
o
Q
<
a
708
TAN BARK AS A BOILER FUEL
TAN BARK AS A BOIT.ER FUEL
700
^
Fig. t Detail of the Grate Bars of the Myers Furnace
Fig. 5 The Early Hoyt Furnace for Burning Tan Bark
I
6 X 18 BoUer
2089 sq.ft. heating surface
J_
Fig. 6 A Tan Furnace with Six Feed Holes
THE SETTING HAD AIR ADMIS3IO.V IN THE BRIDGE WALL AND A BAFFLE ARCH IN THE COMBUS-
TION CHAMBErt. VERY GOOD RESULTS WERE OBTAINED.
Fig. 7 A Furnace with Shaking Grates for Burning a Coal and Tan
Mixture
AIR SPACES OVER FIRE ARCH AND IN WALLS OP FURNACE AND BOILER WALLS. DISTANCE FROM
GRATE TO TOP OF ARCH INSIDE SHOULD NOT BE LESS THAN 4 FT.
710
TAN BARK AS A BOILER FUEL
on the grate surface. Fig. 6 shows a more modem type of burner
designed to give a more even distribution of the fuel on the grates.
78 Fig. 7 shows a more up-to-date furnace designed for the hand
Fig. 8 Cross Section of Fur-
nace WITH Hump-Back Grates
AND Bearing Bar
Fig. 9 Cross Section of a Double-Arch
Tan Furnace of the Thompson Type
ON which a Test was Run
] n n n n rj
! I !
U U iJ u u. u
Fig. 10 Bush Tan Furnace with Multi-Tube Feed
firing of a mixture of coal and tan, the coal being mixed with the tan
before entering the furnace, which is supplied with shaking or shaking
and dumping grates. When coal is mixed with tan in any consider-
TAN BARK AS A BOILER FUEL
711
a
o
<
M
<
H
z O
o !!5
— M
< «
-J OQ
< p
H
>
<
O
n
712 DISCUSSION
able proportion, more air is required for combustion, the best air
spacing in the grate bar being found to be I in. The percentage of
draft area for this purpose sb ;uld be about 40 to 50 per cent, depending
upon how large a percentagt of coal is used with the tan.
79 Fig. 8 shows what is known as a hump-back grate, which has
been installed in different tanneries for the purpose of increasing the
consumption of fuel in a given furnace. For instance, in a plant that
had trouble in consuming all its tan bark, the writer merely took out
the grate bars and put in a ridge bar as shown and converted the grate
surface into the hump-back form. The result was that the consump-
tion of tan bark per furnace was increased from 12 tons per day on
the dry bark basis to 15 tons.
80 Fig. 12 shows what was known as the Thompson type of tan
furnace. The MacMurray furnace, with a convex grate surface and
feed pipes, is a type quite a number of , which |the writer has seen in
operation.
81 Fig. 10 is another form of tan furnace which gave good results
in a plant in the South. The hump-back form of grate is reversed
something like that used in the writer's stoker furnace, except that
the tan is fed through a number of feed holes along the upper edges of
these grates. This furnace was designed by the foreman in a Southern
tannery .
82 Fig. 11 shows a design of the writer's for an adjustable gravity-
feed furnace for burning tan or sawdust. The feed chutes are rect-
angular in section and contain adjustable chutes to regulate the depth
of tan on the grates for any condition of draft, etc.
DISCUSSION
Albert A. Gary. The furnace described by Mr. Myers consiste
of an extension in front of the regular boiler setting, with a numbes
of circular stoke holes, or openings through the top arch, over thr
grate. No little trouble has been experienced with this construction,
due to the destruction of the lower end of these circular fire-brick
tubes through which the fuel is charged to the furnace.
2 If these stoke holes were always completely filled with fuel,
so as to prevent inrushes of air, this destructive effect could be materi-
ally checked. However, as the method of charging fuel by hand
is an intermittent one, the upper end of the cone of spent tan bark,
soon after charging, drops below the level of the top of the aroh, the
inrushing air meets the hot furnace gases at these points and intense
TAN BARK AS A BOILER FUEL 713
combustion results. For this reason, and due to the fact that when
the excessive moisture in the fuel rises as a vapor against the arch
rapidly abstracting its heat (to become superheated steam), the fire
brick cracks and disintegrates, finally resulting in a chipping off of
the brick-work of the reverberatory arch around the lower end of
the stoke holes. Repairs are therefore frequently necessary.
3 A continuous automatic feeding device, which would keep
these stoke-holes constantly filled with the moist fuel, would undoubt-
edly do much to relieve this trouble by preventing an excessive
infiltration of air at frequent intervals of time. Mr. K. McMurray
of New York, has devised a verj^ ingenous method for overcoming this
trouble in hand-stoked furnaces. Fig. 1 herewitli shows both front
and side sectional elevations of this furnace.
4 In the stoke hole is fitted a circular lining of cast iron which does
not extend to the level of the inside of the arch. The lining is
finished with a shoulder which diminishes the diameter of the opening
by about two inches. A tube or open thimble drops into this frame,
being held by a rim cast around its upper end. The lower end of the
thimble extends about a foot into the furnace.
5 The fuel charged into the stoke hole falls through the thimble,
and forms a cone-shaped pile below it on the grates. When the stoke
hole becomes uncovered, the in-rushing air causes the intense com-
bustion to take place, not on a level with the brick-work, but at
a level below the thimble, and the life of the fire-brick arch at the
stoke-hole openings is thus greatly prolonged. The ends of the cast-
iron thimbles burn off gradually, but they cost v ly little, and may
easily ])e pulled out and new ones inserted in their place.
6 Another trouble met with in this type of furnace is the rapid
burning away of the fuel next to the side walls and the consequent
large infiltration of air from the ash pit. This trouble has been
largely overcome by reducing the width of the furnace about a foot
at the grate level, as shown in the front sectional elevation. The
ledges formed on either side of the lower part of the furnace support
the cone of charged fuel on each side, thus keeping the grate effect-
ually covered with fuel.
7 In this construction it will also be seen that instead of using a
flat grate, the grate bars are curved so that the grate surface is
higher at the center of the furnace than at the sides. This design
decreases the thickness of the fuel bed under the stoke holes and
causes a thickening of the fuel bed at the sides of the furnace.
8 Since water can be evaporated in the furnace itself only at
714
DISCUSSION
cc
o
A
0
W
a
o
p
(6
Pi
C3
TAN BARK AS A BOILER FUEL 715
a great loss, every practicable facility should be utilized for depriving
the wet fuel of its moisture. Mr. Myers has mentioned the com-
paratively small gain from pressing the moisture out of the spent
tan. I have used special rolls for extracting the water from moist
fuels, with a desirable gain resulting. These rolls are of cast iron
and run in pairs, one roll being about 12 in. in diameter, the other
about 14 in. and both held together by heavy springs. As both rolls
are revolved at the same number of revolutions per minute their
surface speeds are necessarily different. The faces of the rolls are
roughened by having a shallow checker work pattern cast upon them
The fuel is fed to the rolls continuously, and due to the tearing or
macerating action between the faces of the rolls more than double the
amount of water is thus worked out, as compared with the press
results given in Mr. Myers' paper.
9 In one case, where the chimney was located some distance
from the boilers, a wrought-iron rectangular flue was used to connect
them, a shallow iron trough being formed on the surface of the flue
by having the edges on the two vertical sides continued above the
level of the top. The other three sides of the flue were cover;! in
the usual way. The moist fuel was fed upon the chimney end of the
flue and was drawn by a conveyor towards the boiler and over its
top, whence it was delivered on top of the extension furnace. A
small evaporation of moisture took place, sufficient to make this
device desirable. The heat from the top of the boiler and the exten
sion furnace may also be used in this way. The waste heat from the
boiler may also be used to pre-heat the air delivered to the ash pit.
I know of no condition where pre-heated air can be used to better
advantage than with moist fuels.
10 Mr. Myers has spoken of the advantage of the high furnace
over the low furnace. My experience thoroughly endorses this.
When the moist fuel is charged into the hot furnace, a cloud of steam
is evolved, which when crowded down upon the burning fuel in
a low fiu-nace hinders combustion. A sufficient amount of steam
would eventually extinguish the fire.
11 In addition to the effect of moisture described in Par. 7 and
Par. 8, the large space occupied by the steam in the combustion
chamber interferes with the combination of oxygen and the combus-
tible gases evolved from the fuel.
12 In the flue-gas analysis obtained with moist fuels, of course
the water in the gases condenses and is not accounted for in the
analysis given.
716 DISCUSSION
William Kent. I consider this the most important paper on the sub-
ject of tan bark as a boiler fuel which has appeared in over thirty years.
The only other paper that I know of is one by Professor Thurston
published in 1874 in the Journal of the Franklin Institute. He made
some boiler tests on tan bark for fuel, using two different styles of fur-
nace, some of his results being better than those given by Mr. Myers.
I think that still better results are yet to be obtained from the use
of tan bark as a fuel, by compressing out as much as possible of the
moisture and using the waste heat of ^"gases to dry the bark before it
is put in the furnace. For burning ^the bark we must have a large
fire-brick combustion chamber and give plenty of time to the burning
of the gases, and then we will get as near the theoretically possible
economy as can be expected.
2 The principal cause of poor economy in the burning of tan bark,
besides the difficulty of securing good combustion in the furnace, is
the amount of heat that is carried away in the shape of superheated
steam in the chimney gases. If the bark, after being partly dried
by compression, were further dried in a rotary drier by the waste
heat from the chimney gases, there would be a very important gain in
economy.
3 I have made a calculation showing the theoretical results that
may be obtained in burning tan bark of different degrees of moisture
under certain assumed conditions, the results of which are given here-
with. The dry bark is assumed to have the following composition :
C = 0.50; H = 0.06; 0 = 0.40; N and ash = 0.04. Substituting
in Dulong's formula, 14,600 C + 62,000 (h j, its heating value
is 7920 B.t.u per lb. Bark containing 20 per cent moisture would
have a heating value of 0.80 X 7920 = 6336 B.t.u.
4 Assuming the chimney gases to escape at 600 deg., the heat
required to evaporate the water from 62 deg. and to superheat the
steam to 600 would be (212-62) -f- 970 + 0.48 (600-212) = 1306,
or for 20 per cent moisture, 261 B.t.u. per pound of tan.
5 The 0.06 lb. of H in a pound of dry tan will unite with 0.06
X 8 = 0.48 0, making 0.54 lb. H2 0, which escapes as superheated
steam carrying away 0.54 X 1306 = 705 B.t.u. for each pound of dry
tan or 0.80 X 705 = 564 B.tu. for tan with 20 per cent moisture.
6 Assuming 25 lb. of air to be required per pound of C + H in
the fuel or 25 X 0.56 = 14 lb. of dry tan, the heat carried away by
this air heated to 600 deg. is 0.24 X 14 X (600-62) = 1808 B.t.u.
per pound of dry tan or 1446 B.t.u. for tan with 20 per cent moisture.
Using the figures thus found the following Table 1 is constructed.
TAN BARK AS A BOILER FUEL
717
TABLE 1 THEORETICAL EFFICIENCY OF TAN BARK FUEL
LosseB of heat due to
Sum of
losses
Net heat
value B.t.u.
Efficiency
per cent
Lb. Evap.
Mols- B.t.u.p«r
ture lb. wet tan
Moisture
H in fuel
Heating
a«r
per lb. wet
tan
0.20 6336
0.30 1 5544
0.40 4752
0.50 1 3960
0.60 ! 3168
0.70 2376
0.80 1584
261
392
522
653
784
914
1045
564
493
423
352
282
211
Ul
1446
1266
1085
904
723
642
362
2271
2 J 51
2030
1909
1789
1667
1548
4065
3393
2772
2051
1379
709
36
64.2
61.2
57.3
51.8
43.5
29 S
2.5
4.19
3.50
2.81
2.11
1.42
0.73
0.03
7 Suppose that tan with 60 per cent moisture were dried to 20
per cent before being put into the furnace, using for this purpose the
waste heat of the chimney gases, we would then have 0.40 dry tan +
0.60 moisture dried to 0.40 dry tan + 0.10 moisture, 0.50 water being
removed. Suppose the moisture and the waste gases left the drying
chamber at 300 degrees. Each pound of water dried out would
take (212 - 62) + 970 + 0.48 (300-212) = 1162 B.t.u. and 0.5 lb.
would take 581 B.t.u. The H in the 0.40 lb. of dry tan would make
0.216 H, 0, which would take away 0.216 X 1162 = 251 B.t.u. Heat-
ing the air would take 0.40 X 14 X 0.24 X (300 - 62) = 320 B.t.u. The
sum of these is 1152, which subtracted from 3168, the total heating
value of tan with 60 per cent moisture, leaves a net value of 2016
instead of 1379, the figure given in the table. The efficiency would
be 2016 -T-3168 = 63.6 per cent, instead of 43.5 per cent, and the
evaporation from and at 212 deg. 2016 -j-f970 = 2.08 lb. instead o
1.421b.
Prof. F. R. Hutton. In 1874, the late Robert H. Thurston pre-
sented a paper on The Efficiency of Furnaces Burning Wet Fuel,
before the American Society of Civil Engineers.^ At that date few
engineers were paying attention to fuel economy, and there was
little widespread knowledge as to the details by which it would be
obtained. There was of course no formulated code for boiler testing.
This paper introduced the wTiter at that time to the problems of
boiler testing, and recorded for the first time for him the formulsR for
the barrel type of steam calorimeter.
2 The two furnaces examined were designed to meet the same
requirements as are assumed in Mr. Myers' paper; but the press
Trans. Am. Soc. C. E., No. 102, Vol. Ill, 1874, p. 290.
718 DISCUSSION
which may be expected to expel a proportion of the water absorbed
from the leaching process was not in use, and no data were given as
to the proportion between the dry bark ground at the mill and the
weight of wet leached fuel delivered at the fire room. The Dutch
oven type of furnace was in usC; consisting of a fire-brick chamber
covered with a reverberatory arched roof. The fuel was fed in at the
top of the oven through two holes in the length of the grate. The
grate was of fire brick moulded to obtain a semi-cylindrical surface
to the upper and lower surface of each bar unit, the concave side being
downward towards the ash pit. A large proportion of the finer tan
lumps was expected to fall through the holes in the arched bars of
the grate and complete their combustion there on the ash-pit floor.
3 But it is very plain from the results of the tests that the fur-
naces were on very much the same plane of efliciency as those reported
by Mr. Myers, since the respective results of evaporation from and at
212 deg. per lb. of combustible were for the Crockett furnace 4.41
lb., for the Thompson 5.68 lb. and for the Myers' furnace 5.43, 4.71
and 4.54 lb., if equal accuracy be assumed in the old testi as com-
pared with the new. This is open to doubt, however, as certain
figures were assumed or deducted from other experiments and were
criticized in the discussion of the results.
4 The present paper is especially interesting to the writer, because
it represents the work of a furnace designed by Mr. Myers which
seems to incorporate some eminently sound principles. I think all
will agree that the three cardinal principles for the complete and
smokeless combustion of a reluctant fuel involve the following:
a Time enough for access of oxygen in the air to the carbon
gas from the fuel.
b Temperature enough for the rapid and complete chemical
union of this oxygen with carbon and hydrogen.
c Room enough for each atom of fuel gas to meet the oxygen
atoms with which it is to unite.
The practical attainment of these results is made more diflScult when
the fuel is wet and in small particles of light weight.
5 We have the conflicting conditions of a hot fire and a slow rate
of combustion to combine with an intensity of draft which shall
not be high. Mr. Myers does this by using the step-grate idea, so
as to admit the necessary air horizontally between the overlapping
bars, whereby the dropping 'of fine fuel into the ash-pit is prevented:
but in addition and as a special excellence of the design, the grate
TAN BARK AS A BOILER FUEL 710
is made^to consist of two sections facing each other with their planes
parallel to the long axis of the Dutch oven and the shell of the boiler.
They are, as it were, upon the inclines of a truncated capital letter V.
6 The bark is fed by a measuring stoker cylinder, which drops a
determined volume upon the whole length of the upper bar at each
partial revolution, and this fall of new material displaces downward
some of what has been drying and growing ready to ignite from the
previous charges. At the bottom of the truncated V is a dumping
grate from which the residue of ash may be released at intervals.
7 The consequence of the inclination of the two grate sections,
with a horizontal inflow of the air^ seems to be the same as is pro-
duced in a successful form of burner for acetylene gas. The two
currents of gas and draft appear to meet in the center or in the axis
of the V and an intense combustion takes place there, the heat of
which reverberates downward from the arch of the oven, raises the
temperature of the upper layers of fuel, and stimulates the rate of
the union of combustile with oxygen. Such a furnace of course is not
subject to the alternations of the "famine and feast" conditions
when excess of wet fuel deadens the fire and causes a smoky and slow
combustion, alternating with high heat and good flame and followed
in turn burned-out spots in the fire until fresh charge comes in
through the holes.
8 The system is also most effective for bagasse, the wet juicy
fibre of the sugar cane after passing through the pressing rolls. This
is more difficult to stoke mechanically than the comminuted bark,
but the requirements for it'i successful combustion are very satis-
factorily met. Sawdust and scrap from wood-working shops are
also burned in furnaces of this design with less danger from sparks
at the stack.
9 Referring to the summaries by the author, it should be plain
that ^the grate surface must be reduced if the tan is press-treated
(Par. 56) to remove moisture. Bulk for bulk, there are more heat
units per-unit of volume or of weight after a volume or weight of
water has been expelled than there were when the tan was saturated
and not pressed. If the fire is hot enough to dissociate the oxygen
and hydrogen which compose the water, the heat for such dissociation
is drawn from somewhere; doubtless from the flaming gases, where the
process takes place, and of course they are cooled, and perhaps killed.
9 If such oxygen and hydrogen recombine, nothing is lost, and
perhaps a mechanic-thermal advantage is reaped because the hydro-
gen flame is longer than the carbon flame. If for any reason such
720 DISCUSSION
dissociated hydrogen does not get a chance to recombine from lack of
temperature or time or room, there is a loss. Mr. Myers' results
should serve to check the claims still advanced at intervals, that the
combustion of steam-gas is a source of great possible economy.
The Author. Mr. Gary in his discussion has described the
McMurray tan furnace, one of many ;different types and designs
now in use. All the ordinary forms of tan furnaces feed the fuel
through holes in the top or arch over the grate. The number and
arrangement of these feed holes vary in the different designs, but
they all form a bed of fuel composed of cones of tan. For this reason
they are all subject to the objection made by Mr. Gary, i. e., that the
fuel burns away most rapidly around the bottom of these cones where
the depth of fuel is least. The central parts of the cones offer great
resistance to the draft so that active combustion takes place on only a
small percentage of the entire grate surface. This necessitates large
grate surfaces and large furnaces with attendant radiation losses.
2 Another objection to the cone method of feeding the fuel,
especially when only a single row of feed holes is employed, is that the
fire is actually divided into a number of small fires around the bottom
of the cones. This multiplicity of small fires, separated by heaps of
wet tan of low temperature, results in lowering the furnace tempera-
ture and in retarding combustion.
3 In furnaces of this type with careless firing the writer has seen
fully one-half of the gi'ate surface doing no work at all in the way of
any active combustion. These ill effects are best eliminated by very
frequent feeding of the tan in small amounts, so that the percentage
of wet tan in the furnace at any time is very small compared to the
actively burning mass. High furnace temperature is thus maintained,
more grate surface is active and the rate of combustion per square
foot is greatly increased. The result is less grate surface required,
smaller radiation loss due to smaller furnaces and greater ease in hand-
ling and cleaning the fires.
4 In general, the greater the number of feed holes the higher will
be the rate of combustion and the smaller the furnace required.
Rapid firing in small amoimts to equal the rate of combustion in the
furnace is productive of best efficiency with any of the usual types of
tan furnaces.
5 Tan presses of different makes, but all of the same type de-
scribed by Mr. Gary, have been experimented with by the writer. It
was found that with careful adjustment and attendance the presses
TAN BARK AS A BOILER FUEL 721
would equal the performance quoted by Mr. Gary but that under
tannery conditions of indifferent attendance and unskilled labor the
presses do not maintain their efficiency.
6 The interference of the steam — gas evolved from the fuel with
the union of the combustible gases with the oxygen must be overcome
by providing large combustion space, preferably over the fuel bed, by
special baffles or by special draft action as in the writer's design of
automatic furnace shown in Figs. 2 and 3 and referred to by Professor
Hutton.
7 The chemical composition of tan is assumed by Professor Kent
to be practically the same as that given from an actual analysis in the
author's paper. The heating value according to Dulong's formula is
7920 B.t.u. per lb., whereas the results of a large number of tests in a
bomb calorimeter by Dr. Sherman, shows the heating value of a
pound of dry hemlock tan to be close to an average of 9500 B.t.u.
8 I have carefully read the record of tests on tan burning furnaces
made by Prof. R. H. Thurston, and presented in a paper before the
Franklin Institute in 1874. Professor Kent states that some of the
results there given are higher than those determined in recent practice
by the writer. The two evaporative results by Thurston are given
as 4.24 lb. equivalent evaporation from and at 212 deg. in the boiler
per pound of combustible for the Thompson furnace, and 3 . 19 lb. for
the Crockett furnace. The corresponding figure obtained by the
writer in his automatic furnace was 5.55 lb.; that is, over 31 per
cent better than Thurston's best result.
9 The writer finds that the evaporations of 5 . 68 and 4 . 41 for the
Thompson and Crockett furnaces respectively were obtained by
Thurston by adding to the evaporation in the boiler the amount of mois-
ture in the fuel evaporated from and at 212 deg. A similar addition to
the writer's evaporation in the boiler of 5 . 55 lb. would make an evap-
oration of 7.75 lb. including the moisture in the fuel. The latter
figure is therefore the one to be compared to Thurston's result of 5 . 68
lb. On the same basis of calculation the economic result of present
best practice is over 36 per cent higher than the best result recorded
by Thurston.
10 Moreover the highest result in the Thurston test was obtained
by a rough volumetric approximation of the weight of the fuei used.
It was not weighed to the fireman as in all the author's tests. Fur-
thermore, both the weight and temperature of the feed water were
merely approximated and assumed to be coiTect in the Thompson
722 DISCUSSION
furnace test; whereas these values in the author's tests were all
observed and recorded in a most accurate and systematic manner.
11 The accuracy and reliability of these old tests is very much
to be doubted, as Professor Button suggests. But even if taken
at their full values it is seen that the results of present practice have
exceeded the old results by over 30^per cent.
12 Actually the present results are probably even higher than
thisji from^,a jComparativej^standpoint, for the^ reason that in the old
days of tanning, the moisture in the tan was less than^in present
practice.][^This consideration would have given the Thurston tests
a decided advantage in the shape of a greater available heat value
of the fuel. Thurston gives the moisture contents of the fuel as fired
as 55 and 59 per cent, whereas the moisture in the writer's automatic
furnace test was 65.3 per cent.
13 This increase in moisture is due to radical changes in the pro-
cess of leaching the bark. Where formerly the bark was treated
with cold, or nearly cold, water it now is leached at temperatures
as near the boiling point as possible, and is subjected to the leaching
process two or three times as long as in the former methods. This
is on account of the high price of bark nowadays, which makes it
pay to leach out as much of the tannin as is practically possible.
Some tanneries to-day leach their bark so thoroughly that only ^
to 1 per cent of tannin remains in the spent tan.
14 The author desires to add that all results and data given in
his paper are results of actual tests made under working conditions.
No assumptions or theoretical calculations are involved in the con-
clusions. The feed water was in every case measured by means of
two tanks or barrels set above a reservoir from which a separate feed
pump supplied the boiler. Feed connections were so separated that
it was physically impossible to pump the water elsewhere than in
the boiler being tested. All connections involving a chance for
leakage were blanked off. Valves were never assumed to be tight
but were proved so during the entire test by means of an open-T
arrangement which would show any leakage.
15 The temperature of water entering as well as leaving the
measming barrels was taken at frequent regular intervals. The
barrels were calibrated by weighing when filled to their overflow
pipes with water at the temperature which the feed water had
averaged during the test.
16 The fuel was in every case weighed in equal amounts to the
fireman. A sample corresponding to each 200 lb. was taken, kept
TAN BARK AS A BOILER FUEL 723
in closed receptacles and at the end of the test was mixed, and quar-
tered down to a quart or two quart sample which was sent in sealed
jars to Dr. Sherman for determination of B.t.u. and moisture.
All readings and observations were obtained with like regard for
accuracy of results.
17 In Par. 3 Professor Hutton also compares the best results
obtained by the writer with those of Professor Thurston; but as before
pointed out, the results are on a very different basis and are not
comparable, unless the moisture in the fuel is also added to the equi-
valent evaporation obtained in the boiler. If this is done the follow-
ing table gives a correct comparison :
POUNDS EQUIVALENT EVAPORATION FROM AND AT 212 DEG.
Ikcludinq Wateb In Fuel
Excluding Watbb in Fuel
Thurston Tests
5.68 for Thompson furnace
Myers Test Thurston Tests < Myers Tests
7.75 for Myers furnace 1 4.24 Thompson furnace 5.55 forMyers furnace
11
4.41 for Crockett
furnace
1
6.63 for present 3.19 for 4.30 for present
ordinary furnace Crockett furnace ordinary furnace
The table shows that when compared on the same basis of efficiency
the art of tan burning has been greatly improved over the old methods,
both with improved and ordinary furnaces.
18 Thermal efficiency is of course the safest and most accurate
basis of comparing results of various boiler and furnace settings,
and the highest result yet obtained in a reliable witnessed test in
tan burning was 71.1 per cent. This is based on available heat in the
fuel as fired after allowance is made for evaporating the moisture
in the fuel. This, test, which .was Jmade on the automatically stoked
furnace before referred to, showed an efficiency of boiler and furnace
of 54.4 per cent, based on the total heat of the fuel.
No. 1259
COOLING TOWERS FOR STEAM AND GAS-
POWER PLANTS
WITH PARTICULAR REFERENCE TO THE POSSIBILITIES OF THE
NATURAI^DRAFT AND AUXILIARY-DRAFT TYPE
By J. R. BiBBiNS, New York
Member of the Society
The object of this paper is to bring to the attention of the members
of the Society a subject which has received relatively little attention
in the past, but which the author believes merits the careful study of
all engineers interested in future power-plant development. The
cooling tower has been looked upon as a makeshift, and its use has
been correspondingly restricted. This, however, is largely due to
the extremely .limited information of an exact or technical nature,
available to the general public, relative to depreciation and per-
formance under unfavorable weather conditions. And, further, it is
the author's belief that the present high prices* constitute the
greatest obstacle to the more widespread adoption of the cooling
tower in both turbine and gas-power plants.
2 Believing that interest may be aroused in this subject by a
more widespread dissemination of engineering data, the author will
I)resent for discussion a type of tower with which some personal
experience has been acquired, and suggest a type of combined fan
and natural draft suited to most efficient running on peak as well as
light loads. It is not the intention to discredit the cooling tower
in its present forms but rather to bring about a more general recog-
nition of its inherent advantages.
' Some recent quotations from a number of builders of forced-draft towers
suitable for a load of several thousand kilowatts, (not inclvding the motor or
engine for driving the fan), ranged between $4.80 and $6.93 per kw., as much
as the entire condensing equipment.
[^Presented at the Annual Meeting. New York, (December 1909), of The
American Society of Mechanical Engineers.
726
COOLING TOWERS FOR POWER PLANTS
PRESENT FIELD
3 There is a continual demand for cooling towers from inland
power stations where the condensing water supply is costly or re-
stricted. Turbine-driven plants, as a rule, operate with higher
vacuum than engine-driven, with the result that the perform-
60 H AVERAGE MONTHLY
TEMPERATURE AND HUMIDITY
50° PITTSBURG, PA.
1904-5-6
40°
0 5 10 15 iiO 25 30 5 10 15 20 35 31 5 10 15 20 25
Fig. 1 Cyclks of Average Temperature and Humidity at Pittsburg, Pa.
ance demanded of the tower must be proportionately better. In the
past, cooling towers have generally been associated with badly run
plants and low vacuum. This, however, is clearly a question of de-
sign and adaptation and not an inherent fault. To be sure, service
requirements are not easily met (see Fig. 1) . Not infrequently atmos-
COOLING TOWERS FOR POWER PLANTS
727
pheric temperatures of 90 deg. to 100 deg. fahr. are encountered, with
cooling water at 80 deg. to 90 deg. fahr. and humidity above 85 per
cent saturation. Yet the auxiliary plant must be moderate in bulk
and the power consumption low. Furthermore, it must be capable of
overload capacity to tide over daily peak loads and periods of
TABLE 1 WEATHER CONDITIONS, PITTSBURG, PA.
DATE
AVERAGE TEMPERATURE
DEO. FAHR.
AVERAGE HUMIDITY
PER CENT
1904
1905
1906
52.8
52.1
52.0
73
68
70
52.3
70.3
Maximum Temperatdre and Humiditt Ranges, Summer of 1906
TEMPERATURE
ABOVE
90 deg.
85
80
75
70
Below
Days in Month
9
14
6
2
14
10
4
1
AVERAGE FOR
MONTH
0.6
10.5
20.5
27.7
23.7
Days in Month
AVERAGE FOR
MONTH
JUNE
JULY
august
90 %
3
1
3
2.3
80
14
10
15
13.0
70
8
11
11
10.0
60
3
8
2
4.3
50
2
1
1.0
Below
..
—
Data from Pittsburg Weather Bureau.
excessively hot and humid weather, all with small investment
cost.
4 Curiously enough, there is an active demand for cooling towers
in the South, e. g., Florida, where the atmospheric*conditions are the
most unsuitable ; also in the Western mining and coast regions. For-
728
COOLING TOWERS FOR POWER PLANTS
tunately, low humidity prevails here as a general rule (e. g., Colorado
Springs ranges around 50 per cent).
5 In gas-power work, the demand for cooling towers is especially
pressing. The large quantity of water required for engine jackets
and for gas cooling and washing entails a heavy expense if water is
scarce and costly. Gas-engine discharge water, being quite pure,
should not be wasted, but cooled and returned to the plant. Even
with deep well pumps supplying sufficient water for cooling, a station
Fig. 2 4500-h.p. Natxjral-Forced-Draft Cooling Tower at Gary, W.Va.,
FOR a Low-Presstjre and High-Pressure Turbine Installation
THESE TOWERS HAVE AUXILIARY FANS IN THE STACK RIVEN BY PELTON WHEEIS AND SMALL
TURBINE-DRIVEN ROTARY PUMPS LOCATED IN THE POWER STATION, AND OPERATED WITH
NATURAL DRAFT WHEN NOT HEAVILY LOADED
is handicapped by a large expense for auxiliary power consumption.
In one instance inj^an Arizona mining plant, the only water available
for cooling" was so impure as to make it necessary to install a com-
pletely closed cooling system for the engine jackets, in which no evap-
oration took place, simply cooling by conduction. In city light and
power plants not fortunate enough to be located on water frontage,
cooling towers built upon the roof have been utilized for engine cool-
COOLING TOWERS FOR POWKR PLANTS
729
Fig. 3 1500-h.p. Open-Type Steel Tower at San Luis Potosi, Mexico, for
A Ga8-Po"wer Central Station
NB TOWER SERVES THE ENGINE JACKETS. THE OTHER THE PRODUCER SCRUBBER. WIRE-
SCREEN SIDE CAama 1» tJSBD to reduce the water lost by windage. cooling, 10 TO
30 DEG. HUMIDITY GENERALLY LOW— ABOUT 50 PER CENT.
730
COOLING TOWERS FOR POWER PLANTS
COOLING TOWERS FOR POWER PLANTS
731
o
732
COOLING TOWERS FOR POWER PLANTS
ing. The expense of buying city water for this purpose would other-
wise be prohibitive except in large plants.
REPRESENTATIVE INSTALLATIONS
6 As examples of present cooling-tower practice in connection
with high-grade power properties, the following may be mentioned:
a At Gary, W. Va. (Fig, 2) , three towers 25 ft. in diameter by 18
ft. high serve a plant of both high and low-pressure turbines
Fig. 6 -4000-kw. Natural-Dkaft Towek at Butte, Mont., Steam
Turbine Station. Wood Construction.
recently installed in connection with a non-condensing
engine plant furnishing the mines with light and power.
A unique feature is the induced-draft fan located in the
stack and driven by a Pelton water motor, which is served
in turn by a small turbine-driven centrifugal pump in the
COOLING TOWERS FOR POWER PLANTS
733
power house. This equipment will be referred to later in
connection with the combined natural-forced-draft type.
The 13th & Mt. Vernon Street station of the Philadelphia
Rapid Transit Company is another low-pressure turbine
plant employing forced-draft cooling towers.
At the central station at San Luis Potosi, in the Mexican
highlands (Fig. 3), separate cooling towers serve engine
TABLE 2 WEATHER CONDITIONS, BUTTE, MONT.
.\vEnAGE Tempekature and Humidity, 1894-1904 (Weather Bdreau, Helena, Mont.)
MONTH TXMPERATURE
DEGREES
HUMIDITY
PER CENT
MONTH
TBMPEBATDHE HUMIDITY
DEGREES PER CENT
January 23 . 7
February 23.9
.March 28.3
April 40.0
.May 49.0
June 52.0
66.8
68.6
60.8
54.3
54.9-
50.0
July
August. ...
September .
October . . .
November.
December .
62.8 45.3
62.9 43.4
.52.0 1 53.6
45.1 64.4
33 . 7 62 . 7
26.6 70.5
Av. for year 41.6
58.9
Tempeeatukb Ranges
BETWEEN
DEGREES
DAYS
ABOVE
DEGREES
DAYS
PER CENT
YEAR
HOURS
PER CENT
YEAR
70- 75
25
70
99
27
276
3.15
75- 80
24
75
74
20
196
2.22
80- 85
28
80
50
13.7
158
1.8
85- 90
19
85
22
0.6
51
.58
90- 95
2
90
3
0.82
9
.097
95-100
1
95
1
0.28
1
.01
Total
249
62.40
691
7.9
Date tvoai M. H. Gerry. Jr.
and producer systems, the make-up water being furnished
from deep well pumps. This plant has been in opera-
tion since 1904.
d Western plants of inexpensive construction are the Mt.
Whitney Power Company (Fig. 5), and the Colorado
Springs Light & Power Company (Fig. 4), both turbine
plants.
e Perhaps the best example of the adaptability of cooling
towers is an equipment designed and built by the Helena
734 COOLING TOWERS FOR POWER PLANTS
Power Transmission Company for its auxiliary turbine
station at Butte, Mont. (Fig. 6).
7 This latter tower was made the subject of an exhaustive pre-
liminary study and a subsequent test by the company's engineering
organization, and through the courtesy of M. H. Gerry, Jr., chief
engineer and general manager, the writer has been able to place the
complete report at the disposal of the Society for future consideration.
The tower (shown in Fig. 6) serves a turbine plant of 4000-kw. cap-
acity at 28-in. vacuum. The designers state that after two years'
experience the results coincide closely with the theoretical deductions
made before its construction.
SPECIAL PHASES OF COOLING TOWER OPERATION
8 Two important factors fortunately contribute to the effective
operation of a cooling tower:
a One factor is the well-known characteristic of a natural-
draft tower considered as a " chimney" — ^increase in cap-
acity with increase in temperature head (see Fig. 12).
9 In steam work, especially with high vacuum, the general range
of discharge temperatures is relatively low; in gas-engine work, on
the other hand, it is high. Pistons are today operated at tempera-
tures of 140 deg. to 160 deg., cylinders from 120 deg. to 150 deg., and
occasionally higher. Owing to the small volume of water in the
minor circuits, such as valves, packings, etc., these temperatures have
little effect upon the average outlet temperature of the engine, which
ranges from 115 deg. to 130 deg. in the large engines, and 140 deg.
in the smaller sizes and verticals. This would correspond to a very
poor vacuum in a steam plant, not more than 24 in. to 26 in.; prac-
tically out of the question in turbine work. However, this high
temperature results in a high rate of heat dissipation in the tower per
unit of cooling surface, with a correspondicg reduction in bulk of
tower.
h The second factor relates to develops nts in the efficiency
of the steam-condensing plant.
10 The function of a condenser is, primarily, 1 1 1 at of a water heater
and the measure of its efficiency as a condensing vessel is the differ-
ence between the temperature of the exhaust steam and that of the
discharge water. A theoretically perfect condenser would heat the
outgoing cooling water exactly to the temperature of the incoming
steam. But in practice from 10 deg. to 50 deg. difference exists,
COOLING TOWBRi FOR POWER PLANTS
736
depending upon the type of condenser and the volumetric latio of
water to steam. A good surface-condensing plant with dry-air pump
should operate at 28-in. vacuum with a temperature difference of 15
deg. ; often it is more, and the author has seen 25 deg. to 40 deg. dif-
ference in some of the largest stations in the country. A good baro-
metric or centrifugal jet condenser, with dry-air pump, should oper-
ate with a temperature-difference of 10 deg. to 15 deg. Although
it is possible for this type to operate on less — perhaps 5 deg. to 10
deg. — commercial practice rarely concedes such results.
11 A very recent development in air pumps has made it possible
to operate on a still smaller difference (from 2 deg. to 5 deg.) with a
130
SHOWING APPROXIMATE MAXIMUM TEMPERATURES OF COOLINC
AIR PERMISSIBLE FOR VARIOUS VACUA
'
w"
Temp.Difference
in Condenser
40°
/
f
f
: r' f
0°
/
120
/
/
/
/
/
0<"
San
pie 5
Air
/
/-
H
/,
/
/
lid
1 i f
or 70
^ /
' 1
iii
/
/
'^p 1 ■§/ 1 ^/ ^/
'S'/ 1 ^/ ■^/ ^/
9«"
/
1
6
/ 1 ^/.//.^/
100
S
a
/
/
0
1 / v/ >/
Assumptio
Vac. Temp. ]
as
latio
(ti 0
O.90
B
3
3
O
>
n
/
/
f
/■f/
LLcUes Deg. Water
Hg. Fkhr. Steam
19 80 100
28 101 60
£7 115 40
/
/Vv
Of)"
/
7
7
7
/ \y7
sd
.t.u per Ib.exh.steam
tT tooltd to inlet air
1
{ 1
70
Temperatu
eEi
tran
1 i
t Aii-Fahr.
20- 30 JO" 50- CO'" TO- bO'' 90° 100'
Fig. 7 Showing Maximum Air-Inlet Temperatures for Various Vacua
reasonable water ratio, and even to approximate theoretical condi-
tions. All this in the right direction. The smaller this tempera-
ture differential, the higher the maximum inlet temperatures per-
missible for a given set of conditions — both water to condenser and
air to tower. The curves in Fig. 7 show this relation in approxi-
mate form — vacuum possible with varying condenser and fixed cool-
ing-tower performance.j|,For example, with 28-in. vacuum, 20 deg.
differential in the condenser, water ratio 60, and 15 deg. cooling in the
tower, the highest possible temperature of outside air would be 65
deg. fahr. With warmer air, the vacuum would necessarily fall.
736 COOLING TOWERS FOR POWER PLANTS
Under the same conditions, with 10 deg. differential, a maximum
air temperature of 76 deg. would be permissible; and with 5 deg., 81
deg. inlet air. It is therefore apparent that the tendency of modern
condenser development toward higher efficiencies will materially
assist in the successful operation of cooling towers under extremely
adverse conditions.
ELEMENTS OF DESIGN
12 The most important elements entering into the design may be
considered under the following heads, having special reference to the
enclosed type of cooling tower, which for a given floor space has by
far the greatest cooling capacity:
a Type of cooling surface.
b Water distribution system,
c Draft and air distribution.
13 The following are a number of essential points that seem to
the writer to have a most important bearing upon any type of tower
designed for maximum duty and efficiency.
a All tortuous or unduly obstructed passages should be
avoided. It is of no advantage to give ample spacing
in one part of the tower and contract it in another, unless
sufficient stack height is provided to overcome the addi-
tional resistance.
b Avoid free falling water. It should be distributed so as
to descend clinging to some form of wetted surface.
c Avoid open spaces in the mat work, usually occurring at
points where it is difficult to fill in between the frame of
the tower. This will "short-circuit" and invariably
diminish the effectiveness of the working sections.
d Reduce working section to minimum possible height, add-
ing extra stack if necessary. The power required to
elevate the water is important, and the working height
of the tower is lost, even in a closed-condenser circulating
system.
e Baffles or variable spacing are often necessary to obtain
imiform air distribution.
/ A settling basin of liberal depth is always advisable in
order that entrained air may separate. In all jet-con-
denser installations, this is extremely important owing to
the amount of air returned to the condenser; and even in
COOLING TOWERS FOR POWER PLAjNTS 737
surface installations, this air will find its way back to the
condenser via the feed water: result, impaired vacuum.
g All wooden mat surface is subject to swelling. Means
should be taken to insure permanent alignment; other-
wise serious reduction in draft area and capacity may be
encountered.
h For maximum effectiveness, a cooling surface is required
which provides uninterrupted descent of water, in a thin
film at all times in intimate contact with ascending air.
If any interruption is necessary, the descending sheet
should be guided into place to avoid free fall.
PRESENT TYPES
14 The various types of cooling systems now in use naturally group
themselves into a few general classes:
a The simple spiral-spray nozzle discharging into an open pond.
15 A prominent example is the 10,000-kw. Wyoming Avenue tur-
bine station of the Philadelphia Rapid Transit Company, where this
cooling pond is employed during a portion of the summer months.
It has been suggested that the sprays be mounted upon the power-
station roof, thereby taking advan+age of the inclined surface of the
roof for extra cooling effect, suitable gutters returning the water
to the cold well. There might be some hesitancy about installing
a reservoir on the roof; but in one notable instance, the recently
designed gas-power station of the Duquesne Lighting Company,
Pittsburg, Pa., the roof reservoir forms a very effective part of the
cooling system. Here a small cascade type of tower assists in cool-
ing. Without other agency this simple system requires only 10 to
20 per cent make-up water.
h The simple tray type. Fig. 3, with water dripping through
perforations, and cooling entirely by means of transverse
air currents from the side.
16 Here no direct draft is possible, and the tower has no direct
cooling surface. The trays operate simply to arrest the lall of the
water. In this respect, the type is a simple mechanical refinement
of a rough frame tower filled with brush, such as has often been
employed in temporary power work. It is, however, comparatively
inexpensive, and under some conditions, may be utilized to advan-
tage. The tower of the Potosina Electric Compan}', San Luis Potosi,
Mexico, is built entirely of structural materials and cools from 10 deg.
738 COOLING TOWERS FOR POWER PLANTS
to 30 deg. with very low humidities. Although encased in netting
to prevent loss by spraying,^ as|much as 10 per cent of the volume
passing through the tower is carried away during a brisk wind. The
inexpensive construction is shown in Fig. 5, using horizontal wire
screens instead of perforated trays and without wind screens.
c The simple cascade type, constructed either of wood or of
corrugated sheet, in which a considerable part of the
cooling is by actual conduction.
17 In the case of the original gas engine service plant of the Union
Switch & Signal Company, Swissdale, Pa., this cascade system mate-
rially assisted in the work of cooling the gas engine jacket water, but
the absti'action of heat through the concrete walls of a large reservoir
was largely responsible for the cooling. This cascade system seems
to have been overrated. In one prominent plant, the author under-
stands it to have been a decided failure; in any form, it is extremely
primitive and not in accordance with effective design.
d Another representative of the simple types of construction
is the multiple cascade. (See Fig. 8 a)
18 Here the fall of water is simply interrupted at short intervals,
and no cooling surface is installed. It is evident that successful
operation is dependent entirely upon the accuracy with which the
trajectory of the falling particles can be predetermined in the spac-
ing of trays and maintained in the subsequent operation of the tower.
This would require an absolutely constant head.
19 The tower at Colorado Springs (Fig. 4) utilizes the construc-
tion, as in Fig. 8 6, a horizontal slotted surface with wind shields to
prevent spray loss. This tower gives 40 deg. cooling in fair weather.
The humidity however is very low, around 50 per cent (relative).
e Several American towers are constructed simply of horizon-
tal lattice work, usually of cypress, the numerous tiers
being staggered in order to break more effectively the
fall of water (See Fig. 8 c).
20 In some, the upper and lower faces of the lattice work are
beveled (Fig. 8 c) to lessen the resistance of descending water and
ascending air. Cooling water is distributed by atomizing nozzles, by
numerous spray pipes, or by Barker's mill.^ This type evidently
does not lend itself readily to natural-draft work, owing to the
serious resistance offered to the draft by the lattice work.
* Radial arm distributor propelled by lateral reaction of its own jets.
COOLING TOWERS FOR POWER PLANTS
739
I
t
III I I I I I I 11
7?
m
-«fa»-» A
1^ 'I X -«).«« j'liPi.A.Wi^-jyr
^
M , :
Stream Lines
II _l U LI □ □
J J J jiG,[a
ELEVATION
TioiifU
F^'V i ■ r ^ -T-
./\ A A A
J
J ^ 3 I jL^tream Line
r~)r"^ , pJ: Tiough
=+=♦=
^-._^^o_o_^
•Distiibutina
Oriflces
^^-^LJ^ rirL_Steam Openings
(= Unit Section ofTower »
8c/
j ^ I i I i
I ! I ! I
y V V 'i' Y
Q 0 0 0 0 0
ITT
r
Fig. 8 a,b, c, d Types of Cooling Surfaces Employed in Various Towers
/ A modification of the multiple cascade system, used in a
German tower (Fig. 8 d), endeavors to utilize partly the
inclined deflecting surface as a cooling medium, although
it is a question whether this is of much effect owing to
740
COOLING TOWERS FOR POWER PLANTS
"■ Discharge Troughs sloinng
to center of tower
Draught area restricted 50?! at troughs
Fig. 8 e, /, g Types of Cooling Surfaces
the fact that the ascending air in all cases impinges on
the lower surface of the deflectors, and not on the upper
wetted surface.
g Another German design (Fig, 8 e), which has been intro-
duced into this country, advances one step in introducing
vertical cooling surface in transverse tiers. But most
important is the attempt to guide the water downward
COOLING TOWERS FOR TOWER PLANTS 741
in the form of a film, by forming each slat with a saw-
tooth edge, meeting the lower transverse slats and guid-
ing the water streams thereon.
21 The designer has evidently appreciated the necessity of avoid-
ing free fall of water and deliberately piped the water from the flume
to small troughs serving the upper row of slats.
22 An American builder has modified this system (Fig. 8/) by
practically discarding the numerous tiers of slats for vertical ones
extending halfway down the tower, turning them 90 dsg. at the
middle of the tower, ostensibly for the purpose of equalizing air dis-
tribution. However effective this may be, we have here the desir-
able elements of continuously wetted surface and no free fall of water.
h A well-lmown American type, resembling the multiple-
surface German tower (Fig. 8 e), employs numerous tiers
of galvanized iron or tile cylinders, with a distributor at
the top, of the Barker's mill type, propelled simply by
reaction of the issuing jets.
23 Although highly effective in fan-type towers, there is much
free falling water owing to the non-continuity of cooling surface ; and
in the author's experience, there is some objection to the Barker's
mill distributor in the difficulty of maintaining ball bearings in proper
condition.'
i Several builders employ continuous galvanized-iron sur-
face from top to bottom of tower, either in the form of
corrugated sheathing or of wire mesh, the water being
carefully guided to the sheets so as to avoid free fall. The
principle is right. With the close spacing permissible,
a most intimate contact of air and descending film may
be maintained.
/ Coming now to exclusively wooden-mat construction, an
example of the attempt to combine in a single-slat con-
struction all the above-mentioned desirable features, is
that shown in Fig. 8 g.
24 This resembles the form employed by Mr. Gerry, at Butte, Mont. ,
although it is sketched from a design by Mr. Moser, of the New-
house Mines & Smelters Company, Newhouse, Utah. This tower is
' In a Detroit station, the entire condensing plant lost its vacuum on several
occasions at peak load owinii to the stoppage of this distributor; and, finally,
a three-deck phosphor-bronze ball bearing had to be designed to withstand the
corrosion from the ascending vapor.
742
COOLING TOWERS FOR POWER PLANTS
\f\/ood Hirfition ,<
in Loivei
Distribution Pipes
Fbks Supporfir
Pkrtfcrms
Bnck Piers.
' 'CPU Cops
Timber Platform
Fig. 9 Experimental Natdral-Draft Tower at Detroit, Mich.
COOLING TOWERS FOR POWER PLANTS 743
of the natural-draft type with a side fiume communicating with
numerous transverse ducts which discharge upon continuous vertical
slats, the saw-tooth construction being employed to guide the water
on to the wetted surface. At the bottom, instead of allowing the
water to fall freely into the receiving basin, each descending sheet
is caught in a small trough and conveyed to the center of the tower,
where it descends without retarding the ascending current of air.
These distributing troughs reduce the effective draft area of the
tower by about 40 per cent; but, on the other hand, the reduction in
area is lairly uniform throughout the tower, and the area correspond-
ingly diminished. That this type is extremely effective is proved
by the results of the tests at Butte.
LATH MAT CONSTRUCTION
25 In a design originated in Detroit, Mich., shown in detail in
Fig. 9,* an attempt was made to subdivide the cooling surface into
sections or tiers, while maintaining the advantages of continuous
vertical surface. This it was thought would facilitate the construction
and repair of the tower; it was also hoped to avoid the distortion of
the mat surface occasioned by the swelling' of the timber, which it is
hard to avoid when long slats are employed.
26 This tower was designed under the direction of Alex Dow,
Mem.Am.Soc.M.E., general manager of the Detroit Edison Company,
largely in order to try out the natural-draft type under conditions of
central station operation. There were many features which could
have been improved upon, but that the type of mat surface employed
was extremely effective is shown by the results of tests made in 1902.^
The important point in design was reduced cost of construction.
With the exception of the sheet-steel shell furnished by a local boiler
' Described in Engineering News, March 20, 1902.
* For example, this difficulty was experienced in a large gas-engine cooling
tower in Texas. Plain horizontal platforms were used, with boards spaced far
enough apart for the water to drip through. After some time in service, the
timber had swelled to such an extent as practically to close off two-thirds of the
tower, deflecting the greater portion of the water to the sides, where it descended
without being cooled to any extent. This trouble, of course, is not so serious
in the vertical slat tower; yet in a Boston plant employing rough boards set
vertically on edge, the boards so swelled and warped that they practically closed
the intervening air passages at certain points. This distortion may be noted some-
what in the Colorado Springs tower ( Fig. 4).
' Conducted by the author and by Messrs. Armstrong and Richardson.
744
COOLING TOWERS FOR POWER PLANTS
maker, the tower was built by unskilled labor employed about the
station, and its total cost, including shell, concrete and brick work,
material and labor, was in the neighborhood of S1350, serving a
1000-h.p. engine-driven plant. The shell was designed self-support-
ing with an independent internal frame work for bearing the weight
of the mass. Wooden sheathing could have been used to ad-
vantage, however, and the entire tower constructed by unskilled
labor.
SKETCH OF LATH MAT-ASSEMBLY
All Mats installed at inclination of s"
Mat tiers incline alternately to right and left
Backing Strip supporting
Mat between Joists
Effective Air Opening
between Laths
1«; Laths per Running foot of Mat
■::o sq. ft. elfective cooling surface per running foot
Fig. 10 a Details of Sectional Mat Surface
27 The mat surface was constructed of common wood lath,
assembled on a form, with iron nails protected from corrosion by
being imbedded in the wood. Mat details are shown in Fig. 10 a.
These lath mats produced a very desirable form of cooling surface.
The rough surface kept the descending stream in constant agitation,
and there was sufficient slope to prevent free falling water for any
great distance, and also to constrain the ascending air to slice upward
through the interstices, thereby bringing into use both sides as well
COOLING TOWERS FOR POWER PLANTS
745
v ft-Ji Distance
Piece
Fig. 10 b Details of Sectional Mat Surface
746
COOLING TOWERS FOR POWER PLANTS
as both edges of the lath. Thus a cooHng surface of approximately
20 sq. ft. per running foot of lath mat was obtained. The various
tiers were readily assembled in succession, working from the shell
inward until full. Uniform water distribution was effected by means
of the pipe-spray system, with laterals spaced like the mats below.
28 Two series of tests^ were made at Detroit at different times,
first with only the two upper tiers, and finally with all the mats in
position. Tables 4 and 5 and Figs. 11 to 13'show the relation between
DISTRIBUTING PIPES
SECTION
I
I ,,
^ 2^ Dia. 12 Thd's per Inch
'^M
-311i<J^
->f2J^' Din. 12 Thd's per Inch
-^e
sa
H^ieTT
for Drain
i Pipe
-l'7K^ ^
2 Pipe
m
B^E3B-
XHole
Fig. 10 c Details of Distributing Pipes
the various quantities observed. It was very noticeable that the
rate of heat dissipation in B.t.u. per sq. ft. per hr. was considerably
higher for the uncompleted tower with only about three-fifths of
the mats in operation. However, by the addition of the remaining
mat surfaces the tower was enabled to work on lower water tempera-
ture, and we should 'therefore expect a lower rate of heat dissipation.
This would indicate that the upper tiers of towers were more effective
* In this plant the condensing system was not well adapted to economic working.
Air and circulating pumps were direct-coupled, making it impossible to control
the tower water separately from the condensation. There was considerable air
in the system from a long run of exhaust piping; and mth no dry-air pumps,
a vacuum of 24 in. was normal practice. But the condenser was operated
with a temperature differential of 47 deg., so that vnth an efficient condenser, a
vacuum of 28 in. might have been obtained with the same tower performance,
16 deg. cooling, 71 deg. cold well.
COOLING TOWERS FOR POWER PLANTS 747
TABLE 3 COMPAU.VTIVE DATA, FORCED AND NATURAL-DRAFT TOWERS
TYPE
FORCjED-DRAFT
STATION A
Rated Engine, i.h.p 1500
Cooling surface, sq. ft 34,780
Surface per h.p 23 . 2
Space occupied, cu. ft 7064
Space occupied, sq. ft 175
Cost complete SSOOO
Cost per h.p % 2.60
Auxiliaries, e.h.p 13^
DIMENSIONS
Delivery pipe above ground 29 ft.
Height over all 40 ft.
Height mat section 17 ft.
Height stack 12 ft.
Height outlet 10 ft. 10 in.
Diameter tower 14 ft. 10 in.
Diameter fan 9 ft. 3 in.
Oin.
Gin.
Oin.
6 in.
NATURAL-DRAFT
STATION C
900
24,500
27.2
10,850
200
$1350
% 1
50
35 ft. 8 in.
53 ft. 9 in.
25ft. Oin.
18 ft. 1 in.
9 ft. 0 in.
16ft. Oin.
TABLE 4 TESTS OF NATURAL-DRAFT COOLING TOWER. DETROIT
Incouplete. Thbee-Fifths Subface In8TAI<LED
Temperature, Deo. Fahr.
Quantities
j
■< .
^ii
TIME
AIR
HOT
WELL*
COLD
WELL
WATER
COOL-
ING
TOTAL
HEAT
HEADt
TOWER
WATER LB.
PER HR.
HEAT
DISSIPATED
B.T.U.
LB. PER HR.
tfi B *
M <D K
<%^
CIRCULATING
TER PER SQ.
LB. PER HR.
LOAD
KW.
1
2
3
4
6
6
7
' 8
9
10 i
11
12 noon
34
102
89
13
68
375,000
4,880,000
332
25 1
270
1.30
35
106,5
90
16.5
71.5
("375.000
^ \ 370,200
6,108,000
415
24.8
r3i5
\290
2.30
35
106.5
87.5
19
71.5
375,000
7,120,000
484
25
315
3.30
35
113
88.5
24.5
78
375,000
9,000,000
613
25
350
4.30
32.5
100
84
16
67.5
399,000
6,384,000
434
26.6
365
5.00
28.5
103.5
88
15.5
75
445.500
6,900,000
470
29.7
485
6.00
26
125
94
31
99
417,000
12,930,000
880
27.8
655
7.00
24
121
94
27
97
427.000
11,532,000
785
27.4
570
8.00
24
123
i
94.5
28.5
99
427.000
12,174.000
827
27.4,
600
♦Assuming a more efficient condenser, say 10 deg. difference, the probable vacuimi would be
26 deg. to 27.5 deg. This condenser actually operated at 40 deg. to 50 deg. difiference.
tTotal heat head = air heating + lost head.
tOnly three-fifths cooling surface installed.
^Difference due to rapid change in load.
748
COOLING TOWERS FOR POWER PLANTS
TABLE 5 TESTS OF NATURAL-DRAFT TOWER, DETROIT
Complete, Five-Fifths Surface Installed
Engines: Two 400. i.h.p. 300. kw. Mcintosh & Seymour tandem compound
engines, overhung generators.
Condensers: Worthington surface (admiralty type) 1600-sq.ft. reciprocating
wet-air pump and circulating pump.
Tower: Wood mat construction, 24,500-sq.ft. evaporating surface, ex-
clusive of shell.
Test: March 15 to 16, 1901, 4 p.m. to 4 p.m., 24 hr.
A.M. P.M. AVERAGE
Weather: Barometer (abs.), min. 30.22 30.07; 30.14 30.27
Temperature air, deg. IS. 5 25; 30 25
Relative humidity, per cent 76 82; 38 72
Load: 600 kw. max. to 50 kw. min. Average 244 . 9 kw
Engine efficiency = 92.5 =875 i.h.p. max. Average. .354.8 i.h.p
Steam: Weight of condensed steam per hr., lbs 5910.6
Temperature exhaust steam, deg. fahr 134 . 38
Temperature condensed steam, deg. fahr 108.78
Weight of steam per hr., max. load, lbs 13,500
Vacuum (abs.) 25 to 19, average about 22
Vacuum corresponding to temperature exhaust steam. ... 25
Vacuum possible with good condenser (10 deg. difference) . 28
Water: Circulated per hr., lbs 293,536
Temperature hot well, average, deg. fahr 87. 50
Temperature cold well, average, deg. fahr 71 . 27
Vaporization loss per hr., lbs 5,970
Results: Condenser surface per kw., sq. ft 2 . 66
Steam per kw. hr., lbs 24.3
Steam per i.h.p. hr., lbs 16.66
Circulating water per lb. of steam, lbs 49 . 6
Steam per sq. ft. condenser surface per hr., lbs 3.7
Circulating water per sq. ft. tower surface, lbs 12.0
Difference in temperature between exhaust steam and dis-
charge, deg. fahr 47
Coohng: Max. 20 deg., min. 3 deg -5 deg. Average 16.23
Heat dissipated perhr., B.t.u 4,769,000
Heat per sq. ft. tower surface, B.t.u 195
Heat per sq. ft. per 1000 lb. water, B.t.u 0.665
Evaporation: Circulating water, per cent 2 . 03
Engine steam, per cent 101
Tower: Surface per kw. (average load 245 kw.), sq. ft 100
Surface per kw. (max. load 600 kw.) , sq. f t 408
Surface per 1000 lb. steam max. load, sq. ft 1 . 82
Surface per 1000 lb. steam average load, sq. ft 4.14
Surface per 1000 lb. circulating water per deg. max.
cooling, sq. ft 5 . 22
COOLING TOWERS FOR POWER PLANTS 749
than the lower. The heat dissipation during the tests on the com-
plete tower ranged from 200 B.t.u. to 300 B.t.u. per sq. ft. of surface
per hour under normal conditions, and this could undoubtedly have
been increased in a carefully constructed tower with suitable con-
denser apparatus.
29 In general, the tower showed very little difference in efficiency
summer and winter, rather against expectations. Apparently the
increased evaporation possible in the higher air temperatures of
summer offset the greater conduction of heat in the colder air of
winter. In very hot weather, a negligible effect from conduction was
apparent, from the fact that at certain maximum loads the vacuum
fell rapidly, indicating that the capacity of the tower had been reached,
due to complete saturation of air, while in cold weather the vacuum
would hold up better at the same load. This shows that with air
fully saturated and evaporation checked, the dissipation of heat by
conduction in hot weather was quite insufficient to give an appre-
ciable margin of overload.
30 This tower was in constant use for a period of about four and
a half years, cooling all of the condensing water for the central station.
Depreciation was at first thought to be a serious factor, but later,
when the tower was finally dismantled, the frame work and mats
were found to be in excellent condition. The only parts showing
deterioration were the upper sheets of the tower shell lying above the
distributor, where corrosion had taken place owing to the alternate
wetting and drying of the surface during the last six months of service
in 1905 (10-hr. operation) The mats themselves were as sound as
when put in. After a few months' service, the mat surface became
coated with scale due to the incrustating properties of the water.
This scale would accumulate, crack off and fall to the settling basin.
31 Although fairly successful, this experimental design might
have been considerably improved. By straddling the supporting
joists in the manner shown in detail in Fig. 10 b, the various tiers of
mats may be brought together into a practically continuous surface
from top to bottom, thus entirely preventing the fall of water.
At the bottom , the obstruction to draft may be prevented by
employing deflecting troughs under each mat, to convey the water
to the center of the tower, as in the Moser tower. Fig. 8 g. A better
distribution sj^stem in the form of horizontal slotted laterals discharg
ing upward and over-flowing directly on to the respective mat sec-
tions is shown in Fig. 10 c.
32 In any system of stationary jets, it is extremely diSicult to
750
COOLING TOWERS FOR POWER PLANTS
\
s
^
^\
o
\
a
\
\
fl \
to \
\
I
"^
\
\
\
3
o
*3
\
1
\
\
\
53^
\
A
t;
\
4
2-0
«> o
QU.
^<^
q:
\
^3
\
A
\
■ii'
\
\
'^ \
P.
\
\
3 ^
1
\
\
•^yj
1
\
h-H
\
1^
^
\
\
o
V
>
<
\
\
s
o
8
o
S
?
S
\
'<>
V
V ^
laxss
PIOO
)J9
\
y
S
•jqi
5J "Sac
iSmrc
)00
o
\
tf
COOLING TOWERS FOR POWER PLANTS
751
obtain uniform distribution of water over the entire tower at different
rates of flow. With the slotted pipes, it is an easy matter to open
or close the slots so as to distribute uniformly, and as they are laid
horizontally, this adjustment is permanent. It is also easy to free
the laterals from foreign matter, as is not the case with jets. This
is because of the ample section of the laterals ; whereas any system
using a large number of small distributing pipes or apertures involves
ultimate trouble from clogging.
Cooling Deg-.-F.
Fig. 12 Lost Head of Natural and Forced-Draft Towers
THE EVAPORATIVE COOLER
33 In gas-engine work it is often necessary to economize water
to the greatest possible extent. In an Arizona mining plant employ-
ing gas engines, where the mine water was so foul and acid as to
prohibit entirely its use for cooling jackets, an evaporative cooler was
recently constructed of ordinary hot-water radiators arranged in
series-parallel, with air forced over the surface by a motor-driven
752
COOLING TOWERS FOR POWER PLANTS
fan. The well-known counter-current system was employed, and
the outfit was fairly efficient, the jacket water being cooled 15 deg.
with a power consumption of 5 per cent of the output of the engine,
34 It occurs to the author that by keeping the radiator surface
continually wet the effect of evaporation as well as convection might
be utilized in cooling. The foul mine water may sometimes be used
for this purpose without contaminating the j acket circulation. With
an expenditure of 2f per cent in evaporation, an increase of 24 per
cent in cooling would be obtained, assuming the air entering and
Fig. 13 Comparative Results, Natural, and Forced-Draft Towers
lea'\Hng to be fully saturated (Fig. 15). This system has been
attempted in connection with steam condensers, but apparently
without much success. The principle seems entirely logical, but
the difficulty of maintaining tight joints with thin-walled tubes of
sufficient diameter to permit of the passage of the proper amount of
air, would seriously detract from the effectiveness of this apparatus
by reason of air leakage. The low vacuum shown during tests of such
apparatus largely confirms this supposition. For gas-power plants,
however, the type seems admirably suited.
COOLING TOWERS FOR POWER PLANTS
753
STANDARDS OF DESIGN
35 The cooling tower should be design(;cl with the same flexibility
as other good power-plant apparatus, as regards capacity under
various conditions of operation; it is subject to the same peak-loads as
the prime mover. As a matter of fact, relatively more heat must be
Fig. 14 Proposed Combination Tower With Natural and Auxiliary-Fan
Draft
abstracted by the tower during peaks owing to the higher steam con-
sumption of a steam engine per horsepower-hour on overloads. Con-
sider, for example, a normal central station load. The evening peak
seldom extends over three hours, and usually the most severe demands
754
COOLING TOWERS FOR POWER PLANTS
on the generating apparatus occur within a period of one hour.
Here, then, a definite overload capacity in the cooling tower is as
desirable as in the engine or boiler; and some means should be
employed to relieve during these peaks the tower which would have
ample capacity to operate unaided during the remainder of the day.
36 Again, consider the comparatively short periods of unfavor-
able weather in normal climates. Reports^ from the Butte plant
(Table 2) reveal a mean temperature of 41.6 deg. fahr. Yet there
were 99 days of the year in which the temperature was above 70 deg. ;
50 days above 80, and three days above 90 deg. fahr. Taking 70
140
130
bbl20
Ol
P
jmIIO
2100
g 90
a 80
^
^
^
^
_^
y^
^
y
^
/
/
/
/
ABSOLUTE HUMIDITY OF SATURATED AIR
IN GRAINS PER CU. FT.
CARPENtER-HEATING-VENTILATION
CALL--SMITHSONIAN TABLES.
/
/
'
/
/
0 10 20 .30 40 50 (10 70 SO
Vapor Content — Grains per cii. ft.
Fig. 15 Vapor-Content in Air at Various Temperatures
deg. as an empirical standard, it is apparent that for two-thirds of
the year this temperature would not be exceeded. But careful
hourly observations at Butte show that a temperature of 70 deg. was
exceeded only for 691 hr. throughout the year; i. e., 7.9 per cent of
the actual time.
37 A study of Pittsburg conditions shows similar results (Table
1). Average throughout the year, 52.3 deg. The temperature dur-
ing June, July and August averaged 10.5 days per month above 85
deg., and the humidity, 15.3 days per month above 80 per cent.
* Provided by Mr. Gerry.
COOLING TOWERS FOR POWER PLANTS 755
Although the actual hours of maximum are not available from the
Weather Bureau reports, it is safe to say that these unfavorable
atmospheric conditions existed not more than one-tenth of the daily
period, or 2.5 hr.
TABLE 6 OPERATING DATA, OPEN-SCREEN TOWERS
Mt. Whitney Power Co., Visalia, Cal.
Tower designed for 1500 kw. — 2000 kw. at 27-in. vacuum
Horizontal screen surface, sq. ft 9550
Circulating water handled, gal. per hr 1 ,720,000
Rate of circulation, lb. per sq. ft. per hr 1500
Dimensions, ft 30 by 47 by 15 high
10 tiers galvanized iron screens 5 mesh per in.
Cost of tower including concrete form $2,000
Observations, October 23, 1906
Maximum load carried, 5.20 p.m., kw 1130
Temperature atmosphere, deg. fahr 55
Depression, wet bulb thermometer, deg 8.5
Relative humidity, 50 per cent absolute, gr. per cu. ft. .. . 2.35
Temperature incoming hot water, deg. fahr 110
Temperature outgoing cold water, deg. fahr 100
Cooling, deg. fahr. (minimum for day) 10
Vacuum carried (ref. 30-in. barometer), in. Hg 26.6
Difference between temperature steam and condenser dis-
charge
Possible vacuum (10 deg. difference in condenser)
Maximum cooUng for day (730 kw.), deg. fahr 16
Data from Hunt, Mirk & Co.. Engineers, San Francisco, Cal.
38 It is apparent from the above that the problem of maximum
capacity in cooling towers involves a condition of peak load existing
only 5 per cent of the time, and high temperature only 8 per- cent of
the time. Moreover, these maximum demands will not generally
occur at the same hours of the day. In the example cited in Par. 11,
0 deg. difference in the condenser, the maximum permissible air tem-
perature would be 81 deg.; the more efficient the condenser, the
higher the allowable air temperature. Yet at Butte, the temperature
of 85 deg. was exceeded during only 22 days of the year, or 51 hr.
This is equivalent to 2.5 hr. during mid-day and less than 0.6 of 1
per cent of the total time.
39 Is it, therefore, good engineering to design a cooling tower
installation with a vacuum-produciug capacity large enough for any
756 COOLING TOWERS FOR POWER PLANTS
and all emergencies; or, on the other hand, to provide auxiliary means
for assisting during these brief periods of maximum demand, while
keeping the proportions of the tower within reasonable limits for
normal operation? Might not even a considerable impairment of
vacuum under the most unfavorable operation be better tolerated
than the increased expense of equipment suited to maximum de-
mand?
40 This, of course, applies particularly to natural-draft towers.
Flexibility already exists in the forced-draft tower through the
speeding of the fans; but even here there are some drawbacks owing
to the high velocities already employed for normal working. Any
large increase in the velocity of the fan may seriously disturb the
uniformity of air distribution over the tower surface and give rise
to eddies destructive to efficiency. That this condition exists, is
very plainly shown by a survey of the discharge velocity by means
of an anemometer. Examination of one defective installation by
this method revealed the fact that fully one-third of the area was
practically ineffective and that reverse currents actually took place
in some parts. The air-distribution problem is exceedingly impor-
tant, and more so in the forced-draft than in the natural-draft tower,
where low velocities favor uniformity.
"booster" type of tower
41 The natural-draft tower is of itself ill-adapted for operating
with affixed temperature head. It thrives on the weakness of the
condensing system. The lower the vacuum, the better the tower
works, because of thelincrease in temperature head. And as this
is clearly a problem of chimney design, the only way out of the diffi-
culty is apparently by some method of auxihary draft. As the speci-
fic heat of air is about 0.23, it is evident that an increase of 25 per
cent in heat dissipation would require roughly double this increase
in quantity of air, in order to maintain the same temperature condi-
tions. This, however, is well within the capacity of a comparatively
small fan auxiliary.
42 There are two methods of accomplishing this result:
a By locating in the stack an induced-draft fan which nor-
mally remains idle.
h By installing at the base'of the tower a forced-draft system
so designed as to supplement the natural draft without
causing a back-flow.
COOLING TOWEBS FOR OPWER PLANTS
757
43 The method ofja^is used in the cooling towers at Gary, W.
Va.' (Fig. 2). , Although'dcsigncdjfor constant service, the arrange-
ment [of thej^stack fan is precisely as suggested. Whether the
Pel ton type of motor, direct-connected to the fan, is superior to
belt or chain drive from a motor mounted on the outside of the tower,
is a question of mechanical convenience: the essential elements are
present.
44 The second suggested arrangement is crudely shown in Fig.
14. The auxiliary air is delivered to the tower through four " L "-
nozzles supplied from a concrete duct surrounding the base of the
tower. With this arrangement, the natural draft under the base of
30 . 40 50 60 70 80 90 lf)0
Fig. 16 Humidity Chart for Wet-Bulb Thermometer
the tower might tend to be reversed owing to the back-pressure
resulting from the blast. It is believed, however, that with a fairly
open mat structure such as has been described above, the introduction
of four auxiliary blast ducts would serve only to entrain more air
and further assist the tower in the absorption of heat.
45 Of the two systems, the former has the advantage of being
already put into practice. However, there is to be said in favor of
the latter that no working parts, such as fan bearings, belt trans-
mission, etc., are in the current of vapor; and with a tower operated
intermittently, corrosion is an important matter, as was pioved by
the deterioration of the upper-sheets at Detroit. Furthermore this
system lends itself more readily to a square or rectangular-shaped
tower, which may be desirable in large sizes.
758 DISCUSSION
CONCLUSIONS
46 Improved types of cooling towers are in active demand for
high-grade power plants both steam and gas, especially for low-pres-
sure turbine installations.
47 Effectiveness of working is largely dependent on efficienc}^ of
condenser, i. e., minimum temperature difference between steam
and discharge water is desirable to increase the temperature head on
the tower.
48 Cooling towers are particularly adaptable to gas power plants.
The bulk of the tower is reduced by the high temperature head avail-
able with hot jacket water.
49 Elements of most effective design: avoid free falling water;
maximum retardation of descent with minimum obstruction of draft ;
insure uniform distribution of water and air; provide the maximum
exposed wetted surface for a given bulk and an interrupted descent
of fluid film.
50 In locations of low humidity simple forms of construction usu-
ally serve the purpose, except where ground space is valuable.
51 Sectional lath mat type of tower well adapted to natural-draft
work. The construction suggested is simple, durable and inexpen-
sive.
52 Normal rate of heat dissipation obtained by lath mat construc-
tion, 200 B.t.u. per sq. ft.
53 Auxiliary fan " booster " suggested as the best means of obtain-
ing the desired overload capacity — a combination of natural and
forced-draft. Overload conditions (high temperature, humidity or
load) usually last but a small percentage of the time — 1 to 5 per cent.
Natural draft suffices for the major portion.
DISCUSSION
Geo. J . FoRAN. Evidently Mr. Bibbins has intentionally restricted
his discussion to the subject of the paper, the cooling tower. He has,
however, presented certain tables which, without discussion, are liable
to be misleading with reference to the condensers and general cooling-
tower condensing situation.
2 The paper discusses the tower quite fully, but classifies the con-
denser as good, bad or worse without discussion. This is made pos-
sible by assuming that the various condenser results obtained are sim-
ply a question of condenser design. This permits the inference to be
COOLING TOWERS FOK POWER PLANTS 759
drawn that the various results can be obtained at the same, or prac-
tically the same, cost, which is incorrect. Some of the results stated
are possible of attainment, but would not show profitable investment.
3 It is impossible to differentiate the tower and condenser quite so
completely as in the paper. Each is strongly influenced by the pos-
sible range in operation of the other, and I would like to show just how
the relative sizes and consequent costs of the plants will be modified
by the results desired.
4 Observers agree that the heat transferred through condensing
surface varies directly with the mean temperature difference between
the two sides of the tubes. Whether this mean should be arithmetical
or geometrical is immaterial for the present discussion, and for sim-
plicity I have selected the arithmetical mean.
5 It is unnecessary to assume condensers of varying grades of
design and efficiency; in fact, it hopelessly complicates the question,
and for my discussion I have assumed a condenser of uniform design
and maximum efficiency with a varying amount of surface, which will
permit us to obtain the various results tabulated by Mr. Bibbins.
6 The fairly universal practice for high-vacuum work for the past
few years has been that for a 15-deg. rise in temperature of the incom-
ing circulating water, during its passage through the condenser, it will
be brought to within 15 deg. of the temperature corresponding to the
vacuum. The proposition is frequently made to add only 10 deg. to
the water and bring it to within 10 deg. of the vacuum. This is per-
fectly feasible, but we must see what this involves.
7 It means, first, that if we must carry away the heat from the
steam by increasing the temperature of the circulating water 10 deg.
instead of 15 deg., we must have 50 per cent more water with conse-
quently larger and more expensive circulating plant and piping.
With a 15-deg. rise to within 15 deg. of the vacuum temperature, the
mean temperature difference between the steam and water side of the
tubes will be 223/^ deg. With a 10-deg. rise to within 10 deg. of the
vacuum temperature, the difference will be only 15 deg. or, in the lat-
ter case, 50 per cent more surface will be required.
8 Following the 28-in. vacuum line in Fig. 7, it will be noted that
Mr. Bibbins has added practically 15 deg. to the condensing water and
has given three curves — one for a good condenser with a temperature
difference of 10 deg. ; a very efficient condenser, 5 deg. ; a perfect con-
denser, 0 deg.
9 Let us consider only the perfect or maximum-effect condenser
with varying surface to produce the results named. For the 0-
760 DISCUSSION
deg. curve the mean difference between the steam and water side of
the tubes will be 73^ deg; for the 5-deg. curve this becomes 123^^ deg.
and for the 10-deg. curve, 173^ deg. Or, if we should take the case
where we add but 10 deg. to the water, these three mean differences
would become 5 deg., 10 deg., and 15 deg. respectively, so that the con-
denser for the 0-deg. curve would have twice the surface required
by the condenser on the 5-deg. curve and three times the surface
required for the 10-deg. curve.
10 While there are several plants which report a circulating deliv-
ery temperature at approximately the temperature of the vacuum, it
is evident that no plant should depend upon such a performance to
obtain the economical results upon which the plant investment is
based, as this would requu-e absolutely perfect test conditions in
every day operation; it would give no leeway at all and would result
in too wide a variation in performance for a slight falling off in operat-
ing eflficiency. Even a slight air leak would result in lowering the
temperature in the vacuum space 5 deg., with a consequent loss in
heat head and reduction in heat transference, owing to the presence of
the air itself. These matterc must be considered in addition to the
question of cost.
11 Again, following the 28-in. vacuum line in Fig. 7 until it inter-
cepts the 10-deg. curve, it will be found that it calls for water at 75
deg.^ the 5-deg. curve calls for 80 deg. and the 0-deg. curve for 85 deg.
All these conditions assume that these results depend only upon the
condenser, and if 1 understand the table correctly, call for the same
quantity of steam and water, the temperature of the circulating water,
it will be noted, being raised 15 deg. in each case. The author also
assumes that the twater is cooled to the temperature of the out-
side air.
12 Although I am sure that the author does not intend to convey
the apparent meaning, the further statement is made that this calls
for a fixed cooling-tower performance; in other words, as I understand
io, that the size of tower and the performance will be the same, to cool
a given quantity of water through the same range in temperature,
irrespective of the temperature of the air.
13 Let us follow this a little further, and in line with the general
assumptions, assume for this purpose that the hot air leaves the tower
at the temperature of the hot water and 100 per cent saturation. By
reference to psychrometric tables it will be seen that each cubic foot of
air at 70 deg. temperature and 70 per cent humidity, when increased
to 85 deg. and 100 per cent, will take on 7.15 gr of moisture, whereas
COOLING TOWERS FOR POWER PLANTS 761
a cubic foot increased from 47 deg. and 70 per cent to 62 deg. and 100
per cent, will take on only 3.575 gr. of moisture; that is, although the
temperature is increased 15 deg. just the same, the air carries away
but one-half the moisture at the lower temperature, showing that
twice the air capacity of tower efficiency will be required at the lower
temperature. This is better understood when we consider that
within the usual air temperature ranges, the moisture-carrying capac-
ity of the air is doubled for each 22-deg. rise in temperature. To be
brief and to avoid confusion, I have used the ordinary nomenclature,
which is scientifically incorrect. We all understand that it is the
space and not the air which is saturated, but this splitting of hairs
would not affect the point under discussion.
14 I have purposely neglected the several minor considerations as
they affect the question to a very small extent. For example, the
volume of the air entering the tower at 70 deg. and 70 per cent humid-
ity, and leaving at 85 deg. and 100 per cent humidity, is increased
nearly 53^2 pe^ cent, due partly to the increased temperature and
partly to the reduced pressure of the air itself, owing to the increased
saturation and vapor present. It is well known that the cooling
tower performs its work principally by the withdrawal of heat from
the main body of water which provides the latent heat for the evapora-
tion of a small portion of the water carried away in the form of vapor
as increased hum.idity of the cooling air.
15 Temporarily omitting th.e]'perfect plant, let us consider an
average operating plant in a location having air at 70 deg. and 70 per
cent humidity. The usual cooling-tower turbine plant would carry a
vacuum of 27 in. with water cooled from 100 deg. to 85 deg. If it is
desired to cool this water from 90 deg. to 75 deg., this would permit of
carrying a vacuum of 27-% in. with the same amount of surface and
water, but would require an increase in the quantity of air and of
tower capacity of approximately 50 per cent. If it is desired to cool
the water through only 10 deg., that is, from 85 deg. to 75 deg. and to
bring the water within 10 'deg. of the vacuum (28V4 in.) this would
call for 50 per cent more [water, 50 per cent more surface and over
100 per cent more air and cooling tower capacity than for the usual
27-in. vacuum plant.
IG There are"^ hardly tw^o plants which have quite the same deter-
mining factors. The determination as to the advisable vacuum and
plant must be decided in each case, but there are few plants where the
conditions would warrant the installation of a plant to produce the
maximum vacuum under the most severe conditions.
762 DISCUSSION
17 With reference to the type of tower with fans in the stack, as
shown in Fig. 2, the Worthington Company installed their first tower
of this type with rope fan drive, in 1900, and recent reports indicate as
good results as when the tower was installed. As a general proposi-
tion, however, there are several questions to be considered in compar-
ing this type. There is a saving in the number of fans over the arrange-
ment with the fans below the tower filling, but the fan operates in the
hot, highly saturated air, is more or less inaccessible and out of sight.
and therefore will not receive the best of attention. It requires good
installation and is more difficult to maintain in good condition owing
to the fact that it is an exhauster. Any of us would prefer to install
a pressure fan rather than an exhauster; the capacity of the fan in the
stack must be somewhat larger for the reason that as neither the circu-
lation nor the surface efficiency is improved, the total volume of free air
required is the same, this being handled at a less pressure and higher
temperature and humidity.
18 Comparing the fan and natural-draft towers, there are few, if
any, locations where high results are desired, where the natural-draft
tower could be selected. A little calculating will convince any
engineer that the draft is principally due to the wind velocity over the
tower. Study of the meteorological tables will show that in most
power centres, except in very few locations, the wind velocity is much
greater in winter than in summer — just the opposite of our require-
ments. This is clearly demonstrated in the operation of any fan
tower from the fan speeds permissible at different seasons. It must
be remembered that with a tower of the same height the wind assist-
ance is the same for either type of tower. There are many locations
where a so-called combined tower can be used if the additional
expense is warranted, but strictly speaking, the operation cannot be
combined. It must be used either as a natural-draft tower or as a fan
tower, but if the fan is operated at all, all the air must pass through it,
whether the fan is located above or below the filling.
19 I do not see how there can be any induction in the tower shown
in Fig. 14. The object of the tower is to get sufficient pressure below
the filling to force through the requisite amount of air, but this pres-
sure must be uniform in the entire space below the filling in order to
obtain complete surface efficiency, and under such conditions air
would leave rather than enter the tower through any additional open-
ings to the outside air.
20 The Worthington Company make a so-called combined tower
which permits of two water levels in the cold well. At the lower level
COOLING TOWERS FOR POWER PLANTS 703
the air enters through the fan at rest and below the lower plates of the
tower shell above the water. At the higher level the lower plates are
sealed and all the air enters through the fans, which can be operated
at the speed necessary to supply the additional pressure required by
the low wind draft. This is also accomplished by the use of additional
draft doors.
Prof. William D. Ennis. Will Mr. Bibbins explain in more detail
the derivation of the curves in Fig. 7? The tower must provide cool-
ing sufficient to absorb the heat liberated with the exhaust steam,
viz., 939 B.t.u. per pound. The amount of cooling in each case would
then be 939 divided by the weight of circulating water per pound of
steam. On this basis, the maximum temperatures of entrant air
agree closely with the curves at 27-in. and 28-in. vacuum, but are
about 1 deg. higher than the curves indicate at 29 in., and 2 deg. or
3 deg. higher at 26 in. The curves should apparently be more nearly
straight.
2 The paper gives unusually complete and valuable data on many
phases of cooling-tower operation, but it is to be regretted that the
matter of loss of water has not been dealt with in more detail. This
is perhaps the most vital question. Manufacturers are sometimes
asked to guarantee a limit of loss, but it would be just as logical to
ask for a guarantee as to the value of n. A rough estimate often
offered is that the loss will not exceed the amount of boiler feed water.
3 Mr. Bibbins gives data from three plants : that at Duquesne, in
which the makeup water was from 10 to 20 per cent; the Potosina
plant, in which the loss of vapor by windage was occasionally as much
as 10 per cent of the volume of water passing through the tower;
and the Detroit natural-draft plant, in which the vaporization loss
was 2 per cent of the water passing through; practically equal to the
weight of boiler feed. The average cooling per hour was (293,530 +
5910.6) X 16.23 = 4,860,018 B.t.u. Each pound of water vaporized,
if we neglect the cooling effect of the air must then have absorbed
4,860,018
c7^rr~= 816 B.t.u. This is the nearest to a reasonable result I
5970
have ever seen in a cooHng-tower test.
4 Usually, and this apparently applies to the two other cases
cited by Mr. Bibbins, the loss of water is far greater than theory indi-
cates as necessary. The cooling of the water is accomplished by (a)
the absorption of heat by the air and (6) the evaporation of a portion
of the water. When the minimum temperature of the air equals or
764 DISCUSSION
exceeds the maximum temperature of the water, the first effect
becomes zero. When the air is initialh'^ saturated, the second effect
becomes zero, except as the air is heated during its passage. Under
the hmiting condition at which there is no direct transfer of heat t-u
the air, the necessary volume of air is increased, and the loss of water
does all of the cooling; but the proportion lost need not exceed, in
theory, the quotient of the range of cooling by the heat of vaporiza-
tion, and the use of screens enables us even to reclaim some of the
otherwise lost vapor. Why is it that almost invariably the make-up
water greatly exceeds the amount thus computed as necessary? It is
inferred from Par. 34 that Mr. Bibbins has considered this question of
cooling by evaporation, in which case some exposition would be desir-
able.
Henry E. Long well. Very early in 1884. under the direction of
John C. Dean, of Dean Brothers Steam Pump Works, I made draw-
ings for a cooler that was built for the Kane Milling Company, Kane,
111. I am told that it was the first one erected in the United States,
and it is, at any rate, a well-authenticated case of a very early in-
stallation. The plant was operated for only two years, being then de-
stroyed by fire, but so far as I can remember the installation acted in
a very creditable manner, especially considering the primitive state
of the art at that time.
2 There are probably many engineers who will take issue with the
author if he means that the coohng-tower field is yet comparatively
unexplored. For ten years or more the cooling tower has been on a
strictly scientific basis. Its design and construction constitute a
branch of engineering that is just as distinct and as well developed as
any of those which deal with other specialties such as gas engines,
steam turbines and the like. When we consider that one builder
alone has constructed about 2000 cooling towers which in the aggre-
gate are capable of cooling condensing water for about 3,000,000
horsepower, we must admit that this device has progressed a long way
beyond the rudimentary stage.
3 It is not excessive cost or lack of knowledge that has restricted
the use of cooling towers in the United States. It is because nature
has been so good to us that the conditions in which cooling towers are
desirable or necessary are comparatively rarer than in the less favored
and more congested European countries,' where these devices have
reached the highest state of development.
4 I regret that the author has not presented in exactly the same
COOLING TOWERS FOR POWER PLANTS 765
form the two tests of the cooling tower described. In Table 4 is
given a complete log of the principal observations made at approxi-
mately hourly intervals; in Table 5 we have only the average of all the
observations made over a period of 24 hours The two tests were
made under such widely different conditions that they afford no proof
as to whether the performance of the tower was any better or even as
good with its full complement of cooling surface, as it was with only
three-fifths of it. During the test with only three-fifths of the cooling
surface installed, the average load was nearly 80 per cent greater,and
the average quantity of water circulated per hour was nearly 35 per
cent greater than on the test with all of the surface installed.
5 Referring to Fig. 11, the indications are that the added cooling
surface served no useful purpose. Indeed if the diagram means any-
thing at all, it means that for the same temperature head the product
of the heat dissipated per square foot of surface per hour multiplied
by the proportion of the cooling surface installed, is practically a con-
stant; also, that for equal temperature heads, the number of degrees
cooling is practically the same.
6 In Fig. 12, in which temperature head is plotted against degrees
of cooling, the lines corresponding to three-fifths and five-fifths
surface, coincide '^so nearly that one could hardly say that they
depart from each other by more than the limit of the normal error of
observation.
!■• 7 Fig. 13 at first sight seems to indicate that at hot-well tempera-
tures below 120 deg. the cooling was considerably greater with five-
fifths than with only three-fifths. But we know that on the test with
only three-fifths of the surface, the amount of water circulated was
very much greater than with five-fifths surface. Compari-.ons of
this sort are misleading unless the quantity of water circulated per
hour and the temperature" of the incoming air are the same in both
cases.
8 The inconsistency of the curves in Fig. 13 will become apparent
if we extend the straight line curve for three-fifths surface until it
cuts the line of zero cooling. This will indicate that at a hot-well tem-
perature of a little above 85 deg. the water would not be cooled at all,
although we know from Table 4 that the temperature of the incoming
air was at no time higher than 35 deg. The inference would be that
water entering the tower at a temperature below 85 degrees would be
warmed by coming in contact with air at or near the temperature
at which water freezes.
9 The indications are that the tower is too small for the work, and
766 DISCUSSION
that its performance is limited, not by the amount of cooling surface,
but by the weight of air that can pass through it in a given time.
After all, it is the air that carries off the heat, and the quantity of air
passing through the tower is just as important a factor as is the area
of the cooling surface.
10 The tower described occupies 200 square feet of floor space,
and is rated at 900 h.p Assuming 15 lb. of steam per h.p-hr., the
tower would have to cool sufficient water to condense 13,500 lb. of
steam hourly. A natural-draft tower designed by one of the mo>t
experienced builders of this class of apparatus, would for this same
duty occupy a space about 29 by 24 ft., or nearly S'j/o times as great as
that occupied by the towers described. It would also be from 7 to 10
ft. higher, which would give a more powerful draft.
11 Referring again to Fig. 12, it will be seen that the temperature
of the water leaving the natural-draft tower is from 40 to 70 deg.
above that of the incoming air. On this same diagram are curves
which purport to show the performance of the forced-draft tower
briefly referred to in Table 3. It would appear from these curves
that the forced-draft tower under favorable weather conditions cools
the water to within 3 or 4 deg. of the atmospheric temperature. Under
unfavorable weather conditions it appears to cool the water to within
15 to 35 deg. of the temperature of the atmosphere.
12 The cost of the forced-draft tower is given as $2.60 per h.p. as
against $1.50 for the natural-draft tower. However, if the compara-
tive results as shown in the diagram (Fig. 12) are dependable, it would
appear that the forced-draft tower was well worth the additional cost,
and a little bit more.
13 In Fig. 7, the author purports to show the maximum tempera-
ture of inlet air permissible for various vacua. This diagram
really shows the maximum temperature of cooling water to produce a
given vacuum on the assumption that we limit the number of pounds
of cooling water per pound of steam condensed, to the arbitrary
figures set down in the lower right-hand corner of the diagram. The
temperature of the atmosphere is not necessarily the limiting tempera-
ture to which the water may be cooled. It is well known that with low
humidities, cooling towers may reduce the temperature of the water
to several degrees below that of the atmosphere. And there is no
law of nature that stipulates that we may circulate no more or less
than 100 lb. of condensing water per lb. of steam to produce a 28-
in. vacuum, or GO, 40 and 30 lb. per II). of steam to produce respec-
tively vacua of 27, 26, and 25 in.
COOLING TOWERS FUR POWJOR PLANTS 767
14 I would point out that the diagram Fig. 1 shows that on two
days in June 1906 the average temperatui'e exceeded 90 deg Accord-
ing to Table 1 on the following page there was not a single day during
that month on which the maximum temperature reached 90 deg., to
say nothing of the average. If there were 10 days in the month of
June 1906 on which the temperature exceeded 75 deg., it is difficult to
see why there must not have been at least as many days on which it
exceeded 70 deg. The quantities set down in the columns headed
"Average for Month" require some explanation to make them in-
telligible.
15 The theory of cooling towers is ^^imple, and any one who has a
reasonable acquaintance with that branch of natural science which
deals with heat, may easily know it a little or even very well. As far
as the theory itself is concerned it would be hard to improve on the
clear, concise and generally masterly presentation of the subject by F.
J. Weiss, inventor of the well-known Weiss condenser, which appeared
in a book entitled "Kondensation," published in Germany about
ten years ago. But as in all branches of engineering, the coefficients
by which theory is reduced to practice are the property of the few
who by special application and practical experience have come to
know the subject profoundly.
Barton H. Coffey.^ The advent of the turbine with the high cost
of fuel in steam plants and the increasing cost of water for cooling pur-
poses in urban installations of refrigerating apparatus, are making the
cooling tower a necessary means of economy.
2 As the author remarks, the literature upon the subject is scanty;
in fact, with the exception of C. 0. Schmitt's paper before the South
African Association of Engineers in 1907, there is scarcely anything
extant that I know of, worthy of the name.
3 I do not wholly agree with Mr. Bibbins' presentation of the
meteorological conditions to be met by cooling towers, as given in Fig.
1 and Table 1 . The comparison of average humidity and temperature,
as given by the weather bureau, is a little misleading, as the humidity
observations are made at 8 a.m. and 8 p.m. only. In lieu of hourly
humidity measurements, I think it better to take the average aqueous
pressure at 8 a.m. and 8 p.m., as it is known that this quantity changes
slowly, and from this the hourlj'- humidities can be calculated. It will
then be found, of course, that as the temperatures advance toward
midday, the humidity falls, thus tending to maintain average thermal
'With Edwin Burhorn, 71 Wall Street, New York.
768 DISCUSSION
conditions with respect to cooling towers and explaining the approxi-
mately uniform results actually obtained. The mean aqueous pres-
sure for July, covering a number of years, works out about as follows :
Table 1 Mean Aqueous Pressure
Actual Aqueous
Pressure
Boston, Mass 0.542 in. mercury
Philadelphia, Pa 0.614 "
SaltLake City, Utah 0.296 "
St. Louis, Mo 0.648 " "
At St. Louis, therefore, where the mean maximum temperature for
July is 88 deg., the Relative humidity would be 49 per cent against a
mean humidity of 66.1 per cent, as given by the tables, which is dis-
tinctly a more favorable condition for cooling towers.
4 While on meteorology, I would like to call attention to the
statement in Par, 156, that the tray or atmospheric type of tower cools
only by means of "transverse air currents from the side", the obvious
deduction being, that without wind this type of tower fails. In
fact, in a dead calm the efficiency of all forms of tower falls off, but
this condition is of small practical account, as in the interior region the
percentage of calm rarely exceeds 2 per cent and on the seaboard
is practically unknown. However, in a dead calm the towers still
continue to work, due to an ascending column of warm air and
aqueous vapor over the tower and a corresponding horizontal inflow
of cool dry air. This condition must exist, otherwise the entire space
surrounding any tower would become filled with warm saturated air
and all cooling would cease. In a forccd-drafc tower for example, the
fan would be simply circulating air having no capacity for absorbing
heat. Apropos of this, I have records from an atmospheric tower on
refrigerating work for the entire month of September 1907, taken with
recording thermometers, in which the cooling water from the tower
was^maintained at an average of 75 deg., never exceeding 80 deg.,
with a cooling range of about 10 deg.
5 In Par. 136, among the elements of design, Mr. Bibbins advises,
"Avoid free falling water. It should be distributed so as to descend
clinging to some form of wetted surface." I would like to know the
basis for this statement, as probably by far the largest number of
towers in use throughout the world employ the principle of finely
divided falling water, as, for instance, the various forms of atmos-
pheric and chimney towers in Europe, South Africa and this country.
OOOLTNO TOWERS FOR POWER PLANTS 769
6 As 75 to 85 per cent of the cooling is due to evaporation, which
can take place only at the surface in contact with the air^ the form of
cooling surface is of great importance. In a cooling tower with free-
falling water, the cooling surfaces consist of the hurdles or decks and
the exposed surface of the falling water. Experiments show the
weight of a drop of water to be about three-fourths of a grain,
the diameter of the corresponding sphere being 0.178 in. A gallon
of water properly distributed will therefore expose about 54 sq. ft.
of surface. If we know the flow ^per second and the time of fall in
seconds,, properly corrected for atmospheric retardation, we can cal-
culate the exposed surface in the water, which, added to the fixed
wetted surface, gives the total cooling surface in the tower. The eflfi-
ciency of the surface in the falling water is greater than the fixed sur-
face, due to the greater velocity of the air relative to the water surface,
due to the motion of the drops.
7 The question of type of surface, in my opinion, is one of expedi-
ency to be determined by the conditions of operation. Fixed surface
is undoubtedly more compact and when skilfully designed opposes
less resistance to air currents. On the other hand, it involves weight,
greater difficulties in distribution, and where oil is present in cooling
water, it becomes coated, the capillarity is destroyed and the water
film is reduced to streams, thus greatly lessening the water surface
exposed.
8 If the atmospheric form of tower is to be employed, it is hard to
conceive of any form of surface, save drops, that would be exposed to
the wind from any direction; and where space is available for sufficient
surface, the temperature reduction called for can always be attained.
9 In a test by the speaker of an atmospheric tower circulating^
440 gal. per min., with air at 93 deg. and humidity 34 per cent, the
water was cooled from 80 deg. to 74 deg., or within 3 deg. of the wet
bulb, which is the limit of atmospheric cooling.
10 With reference to IVIr. Bibbins' remarks on the effect of tem-
perature range on the size of the tower, I beg to submit a few figures
on the volume of air required at 80 deg. and 80 per cent humidity to
absorb 1000 B.t.u. when the air can be heated to the following final
temperatures and saturated :
Table 2 Volume of Air Required to Absorb 1000 B.t.tj.
Class of Work Final Temp. Air Cu. Ft .
Refrigeration 88 deg 985
Steam Condensing 27 in. vac 100 " 429
Steam Condensing 26 in. vac 110 " 267
77C DISCUSSION
This shows cne enormously increased quantities of air required as the
lower ranges of cooUng are approached, and also shows the particular
advantage of the atmospheric tower for refrigeration work, in saving
the power necessary to handle this large volume of air.
11 For example, with air at 80 deg. and 80 per cent humidity, to
cool 600 gal. of water per min. to 80 deg., would require about 70,000
cu. ft. of air per min., requiring about 17 brake horsepower in a fan
tower. An atmospheric tower of like capacity, having 960-sq. ft. wind
exposure, would receive 248;000 cu. ft. of air per min. at a velocity of 4
miles per hour. In steam condensing with a limited space, the
forced-draft tower is, of course, the only available type.
Carl George de Laval, The author states that the present
high prices constitute the greatest obstacle to the use of cooling
towers, and, further appears to give the impression that the cooling
tower is a makeshift and not a permanent apparatus.
2 There are three classes of towers, forced-draft, natural-draft
and a combination of both, the last-named being used either way,
depending on the season of the year. The selection of the tjrpe
should depend on climatic conditions, cost, etc., a dry climate being
best suited for a cooling tower.
3 The author states that the costs range from $4.80 to $6,93 per
kw., which appear to be slightly higher than market prices, the reason
perhaps being that the author had imposed severe conditions when
asking for bids on cooling towers, thereby increasing the costs.
4 Let us assume a plant of 1000-kw., consuming 19,000 lb. of
steam per hr., basing the condenser performance upon the ordinary
10-deg. difference in a counter-current jet condenser, and upon a
27-in. vacuum, with air at 70 deg., and 70 per cent relative humidity.
A cooling tower with interlocking pipe filling can be built approximately
19 ft. by 35 ft., fitted complete with fans, for about $5 per kw., and
a wood-filled tower about 21 ft. by 35 ft, for about $4.50 per kw.
5 The author is correct in stating that installations are not being
sufficiently studied, and this, no doubt, is the principal cause for the
failure of cooling towers and has prevented a more general adop-
tion of them. It is not sufficient merely to obtain information as to
maximum load, steam consumption, maximum temperature and hu-
midity, but it is necessary to know whether these maximum load con-
ditions must be met at the conditions of maximum temperatures and
humidity, and if so, for how long a time.
6 Let us assume that bids are asked for a cooling tower for 8000
COOLING TOWERS FOR POWER PLANTS 771
kw., the conditions named being an air temperature of 75 deg. and
75 per cent humidity, 27-in. vacuum, no time being stated when this
load of 8000 kw. is likely to occur, and what its duration is. The
real facts may be that this load comes in winter only, and that in
summer probably not over 5000 kw. would be required during the
evenings, while the summer mid-day load might not be over 20C0 or
3000 kw. Under such conditions a tower calculated for a 5000-kw.
summer load would be ample for an 8000-kw. winoer load, and if the
installation was made on the basis of 8000 kw. the year round, the
cooling tower would be too large and expensive, and the cost per
kilowatt of maximum load would be too great.
7 The maximum mid-day temperature and humidities likewise
should not be the basis of consideration with maximum loads, as the
electric lighting plant maximum during summer should instead be
based upon 8 p.m. temperature and humidities. One sometimes
sees the requirement to handle maximum loads at an atmospheric
condition of 90 deg. and 80 per cent relative humidity — a condition
that may never be reached in the particular locality where the tower
is to be installed.
8 Most of the towers described by the author appear to be home-
made or makeshift towers, for instance, the tower shown in Fig. 6,
installed at Butte, Mont., having a cross-board filling and a
wooden stack for natural draft. The design is such that it will
lose much of its efficiency as it continues in service, and the
boards, as well as the upper stack, will warp, admitting cold air
above the filling and tending to kill the draft upon which such a tower
depends for its efficiency. The warping of boards will also cause
leakage through the sides of the tower, the leakage being carried by
any strong breeze, and thrown against surrounding buildings and
territory, where during winter it may freeze into a heavy mass.
9 Referring to preceding discussion on the design of towers for
maximum atmospheric conditions, one will note in the temperature
ranges in Table 2, for the Butte tower, that the atmosphere was over
80 deg. during less than 3 per cent of the total time of the year, so
that such conditions can hardly be used as a basis for calculation.
Atmospheric conditions at Pittsburg during the four months from
May 15 to September 15 average approximatel}'^ 70 deg. and 70 per
cent, which appears to be about a standard basis for cooling towers.
10 Par. 9 refers to the use of cooling towers for handling jacket
water of gas engines, the temperatures being about the same as those
encountered in ice plants, and higher than in the case of steam con-
772 DISCUSSION
densation. Several installations show this temperature to be from
156 cleg, to 111 dog., and 130 dcg. to 80 deg.
11 Par. 10 and Par. 11 state that the difference between the theo-
retical steam temperature and the temperature of the circulating
water varies from 10 deg. to 50 deg. The usual jet condensers and
surface condensers give about 15 deg., and cooling towers for recipro-
cating engines are usually based on a 24-in. vacuum, with circulat-
ing water cooled from 125 deg. to 100 deg. and an air temperature of
70 deg. and 70 per cent relative humidity. Counter-current con-
densers give about 10-deg. difference, the circulating water being
handled under the same conditions of vacuum, with a temperature
range from 130 deg. to 105 deg., instead of from 125 deg. to 100 deg.,
which of course gives an easier condition for the cooling tower.
12 It is a well-knowTi fact that an efficient condenser must be
installed in order to get good work from a cooling tower, it being an
advantage to the tower to have the temperature of the hot water and
the cold water as high as possible. For instance, taking examples
of the two conditions, both 1000-kw. plants consuming 19,000 lb.
of steam per hr., at 24:-in. vacuum, and an air temperature of 70 deg.
and 70 per cent relative humidity, oile plant being based on a 40-
deg. difference between the exhaust steam and the outlet circulating
water, which requires the water to be cooled from 100 deg. to 75
deg.; the other plant being based on a 10-deg. difference between the
steam and the water, the water being cooled from 130 deg. to 105
dcg. In the former case, for the same load, vacuum and air tem-
perature, we require an interlocking pipe-filled tower, 22^ ft. in dia-
meter by 35 ft, high, having four 96-in. fans; whereas in the latter
case with only a 10-deg. difference we can do the same work with the
tower 13 ft. 6 in. in diameter by 35 ft. high, having one 120-in. fan.
The efficiency of the condenser therefore makes a very decided differ -
ence in the size of cooling tower.
13 Under c in Par. 13, the author apparently refers principally
to towers with wood sides, having a wood structure within the out-
side boarding. It is very important that the filling nmst come close
to the side of the tower. Particular care should be taken in erecting
towers to see that pipes are first laid around the outside edge as closely
to it as possible; otherwise, there will be a short circuit of cold air
around the side of the tower and a loss of efficiency. This condition,
while bad enough in the forced-draft tower, is much worse in towers
of natural-draft type, because this air will seriously reduce the draft
by mixing with and cooling above the filling the heated air upon
which the draft depends.
COOLINO TOWERS FOR POWER PLANTS 773
14 As to height of working section, it is true, as the author states,
that the height is important, and the distance of the elevation of
the water should be kept as low as pos^il^le. A pipe-filled tower is
13 ft. 4 in. deep with a drop at the bottom of from' G ft. to 11 ft.,
according to the size of the tower. With a distributor operating
head of 5 ft., this gives the largest towers a maximum pumping head,
plus friction in the piping, of 29 ft. 4 in. against approximately 38
ft. as required in the experimental natural-draft tower at Detroit,
shown in Fig. 9. The horsepower necessary to pump the water the
additional 8 ft. 8 in. in height will offset the usual fan horsepowers,
making a natural-draft tower of jthis type more expen.sive to operate
than a fan-draft tower.
15 As to the mat of wood swelling and being thrown out of place,
I would state that towers have been built with a cross-board wood
filling, and four of these have been in satisfactory operation since 1904.
In these towers were used 2 in. by 2 in, verticals at intervals through
the filling, with the boards nailed so as to hold the filling in place and
prevent distortion or formation of large open gaps through warping
of the fining.
16 The cooling towers illustrated in Fig. 3, are furnished with per-
forated pans and have free-falling water, the sides being screened.
This tower depends for its efficiency upon a cross breeze and is very
inefficient in still air as the air cannot rise within the tower on account
of the pans. A strong breeze will blow most of the water out through
the sides of^the cooling tower, in spite of the screen. The tower
shown in Fig. 4, occupies considerable space, and also requires addi-
tional space in the immediate vicinity because of loss of water through
windage. The tower illustrated in ;Fig. 5 is evidently much less
efficient than that in Fig. 4, because of the large amount of free-falling
water. The free-falling or splashing of water is a very inefiicient
method of cooling. Water for proper cooling should always be brought
down in contact with a surface so that it will descend slowly and thus
have close and intimate contact with surrounding air.
17 In Par. 26, the author gives the total cost of the Detroit tower,
erected complete and including filling, distributing pipes, founda-
tions, etc., at $1350. It appears ^that the steel work, if made of
at least No. 10 gage, would weigh approximately 20,000 lb., which
at 6^ cents per pound, which is about as low a rate as mentioned,
would require an expenditure of $1300 for steel work alone. The
lath filling and the work of assembling and installing this tower, would
cost about $400; the timber supports, distributing piping, etc.,
774 DISCUSSION
about $250; concrete foundations an additional $250; or a total cost
of $2200. Assuming a load of 1000 h.p. with 19,000 lb. of steam per
hr., vacuum 24 in., with a temperature difference of 10 deg., the
circulating water being cooled from 130 deg. to 105 deg., air at 70
deg. and 70 per cent humidity, a pipe-filled cooling tower of the fan
type, measuring 13 ft. 6 in. by 35 ft. could be installed for about
$2500.
18 The results of the test given in Table 5, with atmospheric
temperatures of from 18^^ deg. to 30 deg., are not complete for a
natural-draft tower, as such towers fall off in efficiency very rapidly
when the air temperature is raised. The results at temperatures from
70 to 80 deg., would not be so favorable.
19 In Par. 30, the condition of scale covering the wooden filling
would be experienced in any tower, and is usually encountered where
well water is used to make up in coohng towers for refrigerating plants.
The scale forms a protecting coating in a pipe tower and prevents
possible rusting of the pipe filling.
20 In Par. 32, the author refers to possible advantages of a slotted
pipe as compared with spouts on a distributor arm, in regard to clog-
ging. The spouts used by some first-class designers are 1 in. in dia-
meter and are consequently much less liable to clog than are pipes
having a i-in. slot in the top.
21 In Par. 15, the author refers to the use of sprays over a pond.
This seems a very simple apparatus, but it must be realized that the
sprays require from 15 to 20 lb. pressure at the nozzle and so con-
sume more power than required for circulation through a cooling
tower of the fan type, and in most cases as much power as is required
both for the circulating of the water and for the driving of the fan.
22 The arrangement of cascade or cooling sprays on a roof as
described by the author is not new. The installation was in use by
J. H. Stut of San Francisco, previous to 1892, being placed upon the
roof of a factory. Galvanized troughs, 5-ft. wide were arranged
in tiers on a slight incline so that the water traveled back and forth
a distance of about 2000 ft. before being returned to the condenser.
An arrangement of falling from one trough to another, these troughs
being spread out upon a roof, was used at the old Budweiser Brewery
in Brooklyn previous to 1890. The sprays and roof troughs are alike
open to the objection that if there is a strong breeze the water is
carried all over the surrounding neighborhood and if there is no breeze,
a heavy fog quickly collects at the point of spray and thus greatly
reduces the amount of cooling.
COOLING TOWERS FOR POWER PLANTS 775
23 Referring to various types of filling illustrated in Figs. 8a to
8^, Fig. 8a offers too serious an obstruction to the draft within the
tower, closing more than 40 per cent of the space necessary for verti-
cal circulation of air, as against 3 per cent covered by interlocking
pipe filling or 25 per cent by wood filling. The cascades as illus-
trated must fall as shown in the sketch in order to operate efficiently,
that is, the water must strike the pans on the next lower section of the
filling; but this they will do only if the amount of water supplied is
practically constant, otherwise it is liable to spill over several rows of
filling, and result in quick descent and consequent loss of efficiency.
24 The filling illustrated in Fig. 86 is that used by Henry W.
Bulkley, and depends upon a cross steam of air, as in the tower shown
in Fig. 3. It is open to exactly the same objection as the latter tower.
25 The filling 8c will cause large quantities of free-falling water
between the several courses and will result in inefficient operation.
The filling Se' , a wooden cross-board type, is apparently good. It
requires additional expense in placing, but evidently will save some-
thing in fan horsepower. The filling 8d offers a bad obstruction to
the draft on account of deflecting the air alternately to the right and
left. The water also will evidently flow down the top side of the
board; whereas the air impinges most strongly against the lower side
of the board.
26 The filling 86 is the same as 86' and is good. The filling Sg
is open to the objection of having no redistribution, — the water dis-
tributed at the top, however unequal it may be, must remain unequal
from top to bottom. The filling 8/ has one redistribution at the cen-
ter. Otherwise it is open to the same objection as 8g.
27 In Par. 23, and also in the footnote, it is stated that ball
bearings are difficult to keep in good condition. Ball bearings are
not used in modern towers, a floating water-step bearing being used
instead.
28 Referring to Par. 41 and Par. 42, a combination type tower
may run with natural draft about eight months during the year.
At a plant in Newark, N. J., a combination type tower with side doors
operates ( n an 800-kw. load nine months of the year with natural
draft, and requires 25 h.p. during the remaining three months of the
year.
29 As to Par. 46, efficient condensers are more badly needed than
efficient cooling towers. Cooling towers have reached as high an
efficiency as can be expected, but most plants now operating with
direct jet condensers delivering into the towers, could obtain much
776 DISCUSSION
higher vacuum or handle greater loads at the present vacuum, if
condensers of the counter-current jet type, or the more efficient
baffled-surface condensers were substituted for the condensers origi-
nally furnished.
30 As to Par. 48, the temperatures are practically the same as in
ice plants. In order to get the temperature head mentioned, it is
more economical to circulate the water for the ice plant first through
the ammonia condenser and then through the steam condenser
delivering to the cooling tower, than to have two towers handling
separately the water of the ammonia condenser and that of the steam
condenser.
31 The open wooden towers referred to in Par. 50 are not restricted
to points of low humidity; but as already mentioned, they require
much open ground, not only on account of their size, but also for
wind effect, and that surrounding buildings may not be drenched
with water blown from the tower.
32 As to Par. 51, the tow^er best adapted to natural-draft work
is the one which offers the least resistance to the ascending current
of air. In Par. 52, no temperatures are given to substantiate the
statement of heat dissipation by lath mat construction.
33 As to Par. 53, one cannot endorse the fan booster or induction
type when combination towers can be made that will give better
results and that are surely preferable for overload conditions.
34 The largest number of towers in this country are of the forced-
draft type, while European practice tends towards natural-draft
towers. It is thus apparent that there can be no standard of type
or size^ because of difference in climates; each installation must be
considered as a separate problem.
E. D. Dreyfus. In Par. 10, Mr. Bibbins says, "But in practice
from 10 to 15 deg. difference exists, depending upon the type of con-
denser and the volumetric ratio of water to steam. " I wish to supple-
ment this by adding that it does not depend altogether on the volu-
metric ratio. Another important factor is the effectiveness of air
removal. Lower vacuum makes it possible to operate with a dimin-
ished volumetric ratio as the temperature rise is increased.
2 Exception is to be taken to ]\Ir. Foran's remark that perfect
condenser operation entails much greater experience, which might be
implied as generally applicable. This is true only of surface con-
densers. In cooling-tower practice, the conditions are extremely
favorable to the use of the more simple jet type. The more efficiently
COOLING TOWERS FOR POWER PLANTS 777
this latter type is operated, that is, the nearer the discharge water
is brought to the temperature of the exhaust steam, the smaller is
the volume of water necessary, since volume and temperature rise
are component factors of the B.t.u. extraction. Therefore, with less
volume of water handled, the size of the condenser may be reduced
and consequently furnished at a smaller cost.
3 A remark made by the author in presenting the paper, that
an inefficient condenser and an efficient cooling tower go hand in
hand, bears further explanation, although the statement was somewhat
modified subsequently. With an inefiicient condenser, the vacuum
is not likely to be very good, and therefore, with the higher tempera-
ture prevailing in the condenser, the water might pass to the tower
at a higher temperature, making it easier for the latter apparatus
to perform its work. On the other hand, the statements might be
applied with equal, if not greater, force to efficient condensers which
are able, for the same condensation, to create higher vacua, besides
heating the discharge water up to the same final temperature head
as the inefficient type, there being little or no terminal difference in
an efficient design at its normal capacity. Moreover, considering the
benefit accruing to the prime mover, a smaller volume of water may be
used and worked at the same temperature as in the inefficient type
of condenser, thus increasing the possibility of the tower. I would
qualify the above statement to the extent that it deals with a com-
parison of condensers designed for the same vacuum, and evidently
would not hold for a case where a very poor vacuum was admissible.
4 It might be well to state here that a near approach to the theo-
retical vacuum is not an impossible condition in actual operation.
This implies, of course, that the character of the condenser design,
the counter-current type with an efficient air ] pump, fulfills the
requirements. In a test which I conducted last fall on a 1000-kw.
low-pressure turbine equipped with a counter-current jet condenser,
the following results were obtained: At three-fourths load with 83
deg. injection water, a vacuum of 28.20 in. (30 in. barometer) was
maintained, and the water left the condenser at a temperature of
96.8 deg. The temperature corresponding to the vacuum was 97.6
deg., giving practically one degree terminal difTerence.
5 I have observed that temperatures of the water leaving the
tower were several degrees colder than the atmospheric temperature
in warm weather, the difference being as much as ten degrees at
times.
6 With the increasing recognition the cooling tower is receiving,
778 DISCUSSION
it would be desirable to have the Society define a standard basis of
measuring the efficiency of the apparatus. There is a conspicuous
lack of harmtmy of opinions as to what constitutes the governing
characteristics of tower performance.
T. C. McBridb. In the earlier parts of the paper the author would
lead us to believe that cooling towers have not received the scientific
attention warranted. Reference to the literature on this subject
and the work that is being done hardly confirms this statement. A con-
siderable number of manufacturers have for some years past been
supplying cooling towers designed on scientific lines, and the pro-
posals submitted by them, particularly on fan-type towers, are in-
telligently framed and leave no points whatever open to guess work.
2 The paper very properly calls attention to the intimate rela-
tionship of condenser efficiency to cooling-tower performance, but
in doing so is extremely unfair to the condenser — in fact, in speaking
of different types of air pump, the author almost leads us to believe
that some are so superior to others that the vacuum they create is
of a kind superior to that created by other air pumps.
3 Condenser engineers now agree that the efficiency of condensers,
with regard to the comparison of discharge-water temperature with
theoretical vacuum temperature, is as much a question of the average
temperature of the vapor in the condenser as its design. The average
temperature of the condenser is necessarily determined by the amount
of air present therein, and is a direct function of the ratio of the air-
removing capacity of the air pump and the volume of air reaching
the condenser with the steam. The merit of the air pump cannot
therefore be determined either from the vacuum obtained or from the
relation of the discharge- water temperature to the theoretical vacuum
temperature, but is wholly a question of the capacity of the air
pump to handle air at the least expenditure for power, maintenance,
interest on first cost and depreciation.
4 It is true that the question of condenser efficiency and air-
pump efficiency is somewhat involved with that feature of condenser
design having to do with the reduction of air-pump suction tempera-
ture, but as all condenser designs should take care of this feature it
may be eliminated from the comparison of types of condensers or
types of air pumps. It is conceded that the author's division of
condensers and air pumps into good, indifferent and l)ad classes,
in accordance with the vacuum and discharge-water temperature
obtained, follows lines which have been generally accepted in the
COOLING TOWEKS FOR POWKR PLANTS 779
past; but a view from an engineering standpoint must consider the
impurities in the steam in the shape of air and non-condensable
vapors, before judging any particular type of condenser or air pump.
The Author is exceedingly grateful for the interest shown in the
paper and the practical nature of the discussion, which has served
to clear up several ambiguities and to extend the subject into channels
of inquiry representative of everyday commercial problems.
2 Mr. Ennis deprecates the loss by windage of considerable
volumes of circulating water, in excess of that supplied by condensed
steam. Theoretically, without windage loss, there should be practi-
cally no make-up water required, as an exact thermal balance has
been established. But this loss does occur in both forced-draft
and open-tray type towers, and often to a serious extent. However
this is simply a point in favor of the closed natural-draft type of
tower, in which the velocities are reasonably low and hence small
tendency exists to abstract water from the cycle.
3 The high loss in the Duquesne Lighting Company plant, it should
be explained, is not due to windage. The hot jacket water can only
be partially cooled, consequently enough must be thrown away to
lower the temperature by the addition of fresh cold water. The loss
at Potosina, however, was entirely due to windage.
4 The curves in Fig. 7 may very possibly be slightly in error,
as they were necessarily based upon arbitrary assumptions — hence
no attempt was made at absolute accuracy.
5 Mr. Foran evidently has had in mind the surface condenser
in discussing possible and probable temperature differentials, whereas
the author has referred more particularly to the barometric or jet
types, especially in Fig. 7. This should have been stated more clearly
in the paper. Generally speaking, it is possible with the jet type
to work with much lower differentials than with the surface type.
Mr. Foran's deductions regarding the extent of surface required to
meet special conditions are therefore entirely proper. This very
difficulty which is experienced with surface condensers in meeting the
conditions imposed by the best cooling-tower practice, only empha-
sizes in the author's mind the inherent advantages of the jet types.
6 The term "fixed cooling-tower performance" could not apply
to the construction of the curves in Fig. 7, as it is here used in the
sense of efficiency rather than size. The use of "performance" here
was in reference to relative cooling effect (deg. fahr.) — not capacity
for absorbing heat — for the sake of eliminating another variable
780 DISCUSSION
in the construction of Fig. 7. The size or capacity for a given con-
dition is simply a function of a heat quantity (B.t.u.) absorbed from
the exhaust steam. For a given type of surface and draft velocity,
the rate of absorption is fairly constant — a parallel to the constant
rate of heat transmission through the tubes, as cited by Mr. Foran.
7 In reference to the Detroit tests. Table 5, it should be noted
that the condensing plant was not well adapted to the work in view,
being an equipment temporarily retained in service from an old
plant, too limited in surface and without means of operating air and
water pumps individually, as required for economical working. The
poor resultsfrom this particular plant were therefore distinctly attribut-
able to the temporary nature of the installation, and not to an inher-
ent fault in the type itself, as might be gathered from the reports.
8 In his closing remarks, Mr. Foran seems to confine the use of
"natural-draft tower" to the open-tray type. It is quite true that
this has no application where large capacities or the highest efficiency
are necessary. The closed chimney type is not dependent to any
extent upon lateral wind velocity, and may be designed to economize
space effectively.
9 The point raised by Mr. Dreyfus in regard to the effect of
low temperature differentials is well taken. The author's observa-
tion that poor vacuum and good cooling go hand in hand applies to
a given equipment, but the highly efficient condenser with low differ-
ential of course finds the most direct application.
10 The author did not observe or infer^that the cooling-tower
field remains comparatively unexplored, but that certain conditions
have tended to render the subject a closed book. This is not the
case with engines, turbines, boilers, condensers, etc., so the fact that
this condition obtains with cooling towers is not readily justifiable.
11 The two scries of tests could not be presented in identical
form, as the data were not available in such form. However, the
curves, Figs. 11, 12 and 13, were drawn up to facilitate comparison.
The first test covered day and peak loads only; the second, the entire
24 hours, — hence a low average load, as Mr. Longwell observes.
Because the tower shows a low rate of heat dissipation with the entire
surface installed, it should not be inferred that the actual work done
was proportionately lower. Considering abscissae (B.t.u.) as equiva-
lent to load (kw.) it must be apparent that for the same load a much
higher cooling effect was obtained with the cooling surface complete.
12 For equal temperature heads, the cooling is bound to be the
same except when the "lost head" differs, as it does slightly in Fig.
12, This opens up an extremely interesting line of inquiry — a survey
COOLING TOWERS FOR POWER PLANTS 781
of rates of heat dissipation and humidity in each successive zone of
the tower. Which part of the tower does the most work? Assuming
air to be discharged exactly saturated at the temperature of exit,
what spacing of mats is correct to produce a proper gradation of
humidity from, say 70 per cent at entrance to 100 per cent at exit?
13 Regarding the inconsistency of Fig. 13, Mr. Longwell has
forgotten to reckon the "lost head" sho^Ti in Fig, 12 — approximately
40 deg. There is thus a very small discrepancy. However, it is
hardly safe to interpolate in such a case. It is already pointed out
in the paper that the tower is working at a disadvantage, owing to the
extremely poor condenser performance, that imposes an extra burden
on the prime mover as well.
14 The circulating water ratios adopted as a basis of the curves in
Fig. 7, were so adopted to approximate average practice, otherwise
a "family" of curves would replace each single curve shown.
15 Mr. Coffej'" favors the use of vapor pressure in heu of relative
humidity. The author entirely agrees to this method as more scien-
tific. However, absolute humidity expressed in grams per cubic foot
perhaps has a more direct bearing on cooling tower work.
16 The suggestion "to avoid free-falling water" should have
been amphfied in the paper, and Mr. Coffey j ustly directs attention
to it. Comoactness or maximum duty for a given size is so essential
in restricted locations that the atmospheric tj-pe is handicapped, if
not debarred, which he himself recognizes in the closing sentence.
The paper is directed entirely along these lines of maximum duty, and
especially toward the development of the natural-draft type.
17 Mr. de Laval advances the argument that a tower should
not have to be designed, rated and purchased entirely on a peak load
basis. This is entirely in agreement with the author's object in
presenting the combined natural-forced-draft tower with fan auxiliary
for use only during peak loads or during bad w-eathcr.
18 The objections of Mr. de [Laval 'to the construction of the
Butte tower are, however, not well taken, as the construction is more
substantial than as described by him, and several years' service has
not developed the defects he mentions.
19 The tests made at Detroit occurred, it is true, during the
colder season, but in Par. 29 it is stated that the tower showed very
little difference in operation in winter or summer — this on the advice
of the chief operator.
20 Tables 4 and 5 present the temperatures asked for to sub-
stantiate the assertion of a safe rate of heat dissipation of 200 B.t.u.
per square foot per hour for the lath mat construction.
No. 1260
GOVERNING ROLLING MILL ENGINES
By W. p. Caine, Ensley, Ala.
Associate Member of the Society
In considering the conservation of steam-power equipment for
driving rolling mills, we must take into account the two methods of
rolling: the two-high mill driven by a reversing engine, and the three-
high mill driven continuously in one direction; and the relative amount
of power required for each.
2 There is very little variation in the type used for each class of
mill. Twin engines are used for two-high mills and single engines for
three-high mills, usually tandem compounds.
3 The reversing engine for the two-high mill must be powerful
enough to take care of the engine and mill friction and the maximum
torque produced by the piece in the rolls in any position. As these en-
gines are usually twin engines with cranks at 90 deg., each side must be
capable of doing the work alone when the other side is on the dead
center.
4 In determining the size and distribution of the metal in engines
of this type it is the custom to make the dimensions a little larger
and the parts a little heavier than formerly for the same work.
Reciprocating parts are made heavier to stand the shocks, thereby
increasing their inertia, and making necessary heavier frames, bed-
plates, bearings and pins, as well as more rigid adjustments, which in
turn require more attention.
5 As an example of the power sometimes used for an engine of this
type, a certain engine may be cited which was fully capable of deliver-
ing 25,000 h.p. while the actual average work on the steel passing
through the mill could not have required more than 2000 h.p. at the
maximum capacity of the mill. The engine and mill friction, if the
mill were driven continually in one direction, would not fall much short
of 1000 h.p. Assuming that 500 h. p. would be required for the
Presented at the Annual Meeting, New York, (Dccemljer 1909), of The
American Society of Mechanical Engineers.
784 QOTERltING ROLLING MILL ENGINES
reversals, the total average work of the engine would thus be about
3500 h.p., or less than one-seventh of its capacity.
6 As the three-high mill^is^driven continually in one direction,
the energy, stored in [the flywheel makes it possible to do the same
work with considerably less than one-half the maximum power
required in the former case, the amount depending upon the size and
weight of the flywheel. The greater the amount of energy the wheel
can store up, the closer can the maximum power required approach
the average work of the mill, resulting in the more economical use
of steam and a lower cost of equipment.
7 Mill designers do not always give sufficient consideration to this
fact and operators have to deal later with high steam cost and diffi-
culty in keeping the proper steam pressure. Of course, there are
other features to be considered, but economical use of power is a very
important item.
8 For driving a three-high mill a twin engine of the cross-com-
pound type could be used with an intercepting valve such as is em-
ployed in locomotive practice, by which the engine could be started
from any position and handled by a quick-acting throttle valve, so
that it could be brought to a standstill as soon as a piece passed
through the rolls, if another were not ready to enter the mill. Such
an arrangement would go a long way toward answering one of the
principal arguments in favor of the two-high mill: that its engine
uses steam only when the piece is on the mill,
9 If an engine of the type described be furnished with a very
heavy flywheel located between the engine and the mill, the shocks
due to the piece striking the rolls will be taken very largely by the fly-
wheel and not by the engine. Furthermore, if the engine were so
designed that it could not work through the wide range of steam
admission, as is the current practice, the abnormal amount of com-
pression now necessary would be cut down to a large extent, the
parts of the engine would be strained less, and the engine would run
with greater steam economy owing to the cutting out of the high
release pressure during heavy work and the reduction of the number
of strokes during the period of negative work.
10 Going into the details of operation of an engine driving a three-
high mill, we find that the engine first develops just enough power
to take care of the friction of engine and mill; next, the piece
strikes the rolls; and third, the piece leaves the rolls.
11 If the engine were running at a constant speed during this
period, just enough steam would be admitted to the cylinder at each
aOVBRNINC! ROLLINCi Mfl.I. ENCiJNhIrt 7S5
revolution to do the work with the least possible variation of cut-oflf,
resulting in the most economical use of the steam. The more the
speed varies the greater the amount of steam required per horsepower
developed.
12 A constant speed is also desirable for another reason; namely,
the available energy stored in the flywheel is always normal under
these conditions, whereas with a varying speed it would be below the
normal about one-half the time. Should the piece strike the rolls
when the steam pressure is low and the steel cold, the engine would
be more liable to stall if the stored energy were below normal than if
it remained constant.
13 The initial force of this^^blow is absorbed by the flywheel and
the speed of the engine is reduced in consequence. When this has
dropped, say four or five revolutions, the governor has probably so
adjusted the steam valves that the engine is developing its maximum
power and the valves will remain in this adjustment until the engine
is nearly up to speed again. During this time the release pressure
will be high, making it necessary to carry a very high compression,
conditions imder which a non-condensing engine will make the most
noise. Further, if the steam going to waste were utilized the engine
would be capable of doing about one-third more work.
14 On many passes the engine is receiving the maximum amount
of steam at the instant the piece leaves the rolls and the flywheel
absorbs energy through the increase of speed above normal. In the
writer's opinion, this is the time when nearly all of the failures of fly-
wheels on rolling mill engines occur, and anything that can be done
to cut down the amount of energy to be absorbed by the wheel at this
time will increase the safety of the engine and decrease the repairs on
the engine and mill.
15 Having in most cases reached the highest point of speed, the
governor will shut off steam entirely from the cylinder, the engine will
slow down to several revolutions below normal speed before sufficient
steam will be admitted to increase this speed again and the result will
be a wide range of the speed variation during the first period when the
engine has simply to overcome its own friction and that of the
mill.
16 The operations outlined in the foregoing arc very complicated
when two or morj passes occur at the same time.
17 The results of a study of the above details of operation of the
No. 1 rail mill engine driving the four-pass roughing rolls at the Ensley
Rail Mill of the Tennessee Coal. Iron & Railroad Company, during
786
GOVERNING ROLLING MILL ENGINES
a series of tests show how great a difference in the performance of the
engine the writer was able to obtain with one adjusting screw added
to the governor,
18 The engine is a Reynolds Corliss, 52 in. by 72 in. non-condens-
ing, equipped with a long-range cut-off valve gear. '
Fia. 1 Showing Location and Function of the Adjusting Screw D on the
Governor
19 Fig. 1 shows the location of the adjusting screw D on the
governor. Its purpose is to prevent the governor from dropping to
position A which would allow the maximum amount of steam to
reach the cylinder, as determined by the valve gear. In this ease the
steam was admitted nearly three-quarters of the stroke with the
GOVERNING ROLLING MILL ENGINES
787
K >1
S H
s *
788 GOVERNING ROLUNG AULL ENGINES
governor in this position. When th(; pin is in position C the steam is
entirely cut off. The screw D was adju3tcd very slowly while the
engine w«is under load to determine the mo>?t advisable position B.
20 This position must be at a point such that the engine will
carry an average load and also will not stop under a heavy load. It
can readily be seen that by reducing the range of adjustment of steam
distribution the engine will operate with more economical steam
consumption and the greatest strains on the engine and mill will be
reduced.
21 The degree of success attained by this adjustment may be
judged somewhat by the accompanying continuous indicator cards
and tachometer speed curves. Card and curve marked A (Fig. 2)
were taken together before the attachment was used. Card and
curve B (Fig. 3) were taken when the attachment was in use.
22 The speed curves were made by a recording tachometer which
the writer rigged up from a Schaeffer & Budenberg indicating tacho-
meter.
23 This indicating instrument was mounted on a plate, a bevel
gear being fastened to the end of its driving shaft, which in turn drives
a shaft with a worm at one end. The worm shaft is mounted on a
bracket which will swing out of gear, thus disconnecting the indicating
instrument if desirable. The worm wheel driven by the worm shaft
runs loose on one of the paper-feed rolls and when a record is to be
taken a small clutch causes the worm wheel to drive the rolls. Two
rolls are geared together and a third acts as a press roll. The paper
rolls used were such as are furnished for the Uehling pneumatic
pyrometer.
24 The indicating needle on the original tachometer was replaced
by a longer one which reached to the paper. A pencil was attached
to the end of the new needle but there was so much friction that the
records were of no value, so a small tin funnel was fastened to the
needle and a linen thread passed through the hole in the bottom,
protruding about -|- in. The ink in the funnel worked down the thread
to the paper and made a satisfactory record.
25 Below the paper-supporting plate is a vertical plate to which
are fastened two electric bells with the gongs removed and pencils
substituted for the clappers. One of these bells was operated by a
contact made once in a revolution of the engine, the record being
shown at the bottom of the curve A (Fig. 2) . As it was very evident
that the paper would always feed the same amount at each revolution
it was not considered necessary to use this device each time.
GOTBRNING ROLLING MILL ENGINES
789
790 GOVERNING ROLLING MILL ENGINES
26 The other bell was operated on a circuit that had two gaps in
series, one of which was closed when the indicator cards were started
and the other kept closed except at the instant an observer at the rolls
indicated the start and stop of the various passes, by momentarily
breaking the circuit by means of a push button. This last feature also
appears unnecessary, as the speed curves have a pronounced change
of direction at these instants.
27 From card A it will be noted that during the four passes there
are 37 records showing that the engine is taking steam, and 22 records
showing that it is not. These would lead one to think that if every
card were a positive one, and the work were distributed throughout
the entire period between pieces, and the mean effective pressure
averaged, there would be a more economical use of steam, and possibly
with a heavy flywheel the size of cylinder could be reduced. That
would be the ideal condition, which cannot be realized, however,
because of three changing functions: the varying time between
pieces, the varying temperature of the steel and the varying steam
pressure.
28 Card B, Fig. 3, shows the engine doing the same work as before,
on a piece of the same length, and it can be seen that the work is dis-
tributed over 34 revolutions and only two negative cards. With the
engine running as this card shows, it is a comparatively easy proposi-
tion to set the valves for economical steam distribution and it is also
much easier to keep the rods and boxes properly adjusted. The low
terminal pressure is the cause of the engine's running much more
quietly than when card A was taken.
29 The indicated steam consumption of card A is about 43 lb. of
steam per h.p. per hr., against 37 lb. in card B, a saving during rolling
periods of over 20 per cent.
30 Of the speed curves, curve A shows that during the friction
load the engine varies from 66 to 73 r.p.m., with an average of about
69 revolutions, and after the passes the speed becomes about 80 r.p.m.,
an increase of 11 revolutions above normal. Curve A shows also that
the second pass is the heaviest one of the four, the speed dropping to
51 r.p.m.
31 Curve B indicates that during friction load the speed varies
only about 3 r.p.m., and that the highest velocity is 75 r.p.m., or 7
above normal. Some changes were made on the rolls between the
two records so that the third pass was as heavy as the second and the
speed drops to about 45 r.p.m. These curves indicate that the engine
could be speeded up to about 75 r.p.m. and not exceed the speeds
I
GOVERNING ROLLING MILL ENGINES
791
J
<
H
CO
01
>l
<
a>
^
^
to (O
35
ft
o >c
792 DISCUSSION
used before, that more energy was stored in the flywheel and that
the engine would not drop below the speed shown on Curve A.
32 Referring to the table, the constant used in items 7 and 8
is the foot-pounds of work for one stroke of the engine at 1-lb. mean
effective pressure.
33 In the original calculation for item 9 only the flywheel was
considered. This is 22 ft. in diameter, weighs about 130,000 lb.
and has a radius of gyration of about 7.4 ft. Taking the formula
energy -—
and altering it to get the energy stored up or given out at a change
in velocity of one revolution per minute it becomes
wt. wheel Aadius gyration X 2.Ty
E for 1 r.p.m.= ^^ 60^econds /
^ 64.32
Substituting and solving with the values given above
E for 1 r.p.m. = 1220 ft-lb.
As the energy varies as the square of the velocity we would use the
following to represent the amount of energy involved in a change of
velocity
E = 1220 n,^ - V
in which Wj is the higher number of revolutions of the engine and n^
the lower number. In checking this up against indicator card and
speed curve results on friction alone, it became evident that 1180
was the proper constant to use to include the inertia effects of the
mill and reciprocating parts.
34 The constant in item 12 was the average foot-pounds of work
per revolution during the friction period as calculated from the indi-
cator cards. The slide rule was used in making the calculations.
DISCUSSION
Henry C. Ord. The conservation of energy as applied to rolling
mills has received very little attention until during the past five or
six years. The power required to roll a given piece was not known
until the continuous indicator and recording tachometer were applied.
aOVERNINQ ROLLING MILL ENGINES 793
The cards from these instruments furnished records from which the
conditions for any stage of the operation could be calculated, giving
complete information as to the variation in power and speed for
different conditions and classes of work.
2 Rolling mill engineers have several reasons for preferring the
two-high mill. As Mr. Cainc says: "The engine uses steam only
when the piece is on the mill." As there is considerable time be-
tween pieces in some classes of work, this is an important item.
Should a piece not enter properly and stick in the rolls, thus stalling
the engine, it is easy to reverse and back out the piece. This con-
dition with a three-high mill would cause considerable trouble and
delay. This is the reason why some of the modem three-high elec-
tric-motor-driven rolls are fitted with an emergency reversing de-
vice. The reversing feature can also be used as a quick salety-stop
in case of accident.
3 The two-high mills are not so complicated as the three-high
mills, and they have less rolls and no reversing mechanism for rais-
ing and lowering the table. In considering the two systems, this is
an item of power that should be charged to the three-high mill.
However, power is not the only consideration; it is usually a question
of the maximum tonnage in minimum time with the least amount of
power.
4 The 25,000 h.p. engine Mr. Caine refers to is a 42 in. and 70 in.
by 54 in. twin tandem horizontal compound-condensing blooming-mill
engine, designed by the writer about four years ago, and built by the
Allis-Chalmers^CompanyTor a blooming mill requiring an average of
about GOOO h.p., which is also about the economical load for the engine.
It was designed for a maximum of 25,000 h.p. under the following con-
ditions: steam pressure," 150-lb. gage; cut-off, f-stroke; vacuum 25
in. referred to 30-in. barometer; r.p.m. 200. This machine has been
described as "the world's most powerful engine." I (believej the
piston speed, 1800 ft. per min., is the world's record.
5 If the engine Mr. Caine has experimented on was tested under
the same conditions as regards pressure and work, with and without
the function of the adjusting screw, we would expect different re-
sults than those sho^Ti. The controlling device has no control over
the engine before the load is increased, until the speed falls to that
fixed by the adjusting 'screw. At this speed and power the engine
will be doing the maximum work allowed by the adjusting screw;
consequently this control can be applied onlj^ to engines that have a
longer-range cut-off than is required for the greatest loads they have
794 DISCUSSION
to carry. After the above conditions are studied, it will be evident
that to prevent the engine's being stalled before reaching the latest
cut-ojff for which it was designed, we would have to dispense with
the ser\dces of the adjusting screw.
6 From a study of the speed cm-ves in Fig. 2, assuming that the
height of the governor varies approximately as the speed of the engine,
it will be seen that had the adjusting screw been applied and adjusted
for the maximum load or minimum speed, it would be in momentary
control during the second pass, and from the speed curves given in
Fig. 3, it is seen that it was in action for about the same length of
time during the third pass. As it would take considerable time for
sufficient change in the energy of the flywheel to produce the results
claimed, we believe there are other reasons for the improved condi-
tions shown. When the adjusting screw is in control, the engine
will slow do^vn much more quickly than without it, and the engine
will be stalled by a lighter load; it will also take more time to do
a given amoimt of work.
7 As engineers prefer to have engines with some power in reserve
to take care of the abnormal load, I believe they would hesitate
before using any method of control that eliminates the reserve power
of the motor to which it is applied.
8 From a study of the Indiana Steel Compan3'''s plant at Gary,
Ind., it is evident that conservation of energy as applied to steel
plants has received considerable study. The rolls are driven by
motors, current being supplied by gas-engine generators.
James Tribe. In Par. 5, Mr. Caine refers to a certain engine cap-
able of developing 25,000 h.p. while the average load does not ex-
ceed one-seventh of its maximum capacity. I do not know what engine
he refers to, but a blooming mill engine of unusually large dimensions
and answering somewhat to the description given, was built by the
Allis-Chalmers Company and installed less than two years ago at the
Carnegie Steel Company's South Sharon plant. This was a revers-
ing engine for rolling 28 in. by 28 in. ingots on a two-high mill.
The maximum power, or rather, the maximum possibility, of this
engine, was likewise 25,000 h.p., which was also far in excess of its
average load, but it is doubtful if there is in existence a more efficient
reversing blooming-mill steam engine equipment.
2 In Par. 6, Mr. Caine asserts that in a three-high mill driven
continuously in one direction, the energy stored in the flyT\dieel would
make it possible to do the same work with considerably less than one-
GOVERNING ROLLING MILL ENGINES 795
half the power. There should therefore be some explanation to j ustif y
the installation of so large an engine, at so recent a date, and having
so large a percentage of surplus capacity. There are two reasons
for this: first, because of the stalling action at the moment the rolls
bite the ingot ; secondly, because of the probable increase of speed as
the ingot is released.
3 The reversing engine, for well-known reasons, has no flywheel,
consequently the momentum of the rotating parts is comparatively
nothing. Therefore the stalling action at the instant of biting the
ingot, due to the tremendous impact, which is followed immediately
by an abnormally high tangential resistance at the rolls' surface,
creates a demand for an exceedingly powerful engine. It is just at
this moment that surplus power, or reserve energy, is of the most
vital necessity in order to save time and heat which would otherwise
be wasted while waiting for the engine to recover itself. At this
critical moment the term "horsepower" does not explain the measure
of effort necessary for overcoming this resistance; for as a less power-
ful engine would be almost, if not quite, brought to rest, two of the
power elements, namely, time and space, are for the time being
practically eliminated, and the engine reduced to a simple ''force" act-
ing on the crank pin. Hence, it becomes a question of a turning
moment sufficient here to overcome the resistance, and of regaining
normal speed in the shortest possible time : for loss of time means not
only delay (which is very serious), but loss of heat, and loss of heat
means additional power necessary.
4 In the second place, the engine must be so constructed as to be
capable of permitting 25 per cent increase of speed above normal with
perfect safety, for the reason that at the instant the ingot leaves the
rolls, the slightest delay on the part of the operator in shutting off
steam, all resistance except friction having been suddenly removed,
results in an increase of speed and the safe limit is quickly reached.
These two extreme conditions, full steam and abnormal speed,
never occur at the same instant, in actual operation, but the
engine must be capable of meeting them, and therefore such an engine
may be said to be capable of several times its normal capacity.
5 So far as the gripping and the releasing of the ingot are concerned
the effect is the same whether a reversing or a continuously running
engine is employed; for the energy of a flywheel may to some extent
prevent the stalling action, just as this is accomplished by the sur-
plus capacity of the larger engine. But flywheel energy cannot be
spent without a proportionate reduction in speed, and with loss
796 DISCUSSION
of velocity more time must be taken to regain it than would be the
case where the force of the steam is applied entirely in the mill.
Part of the steam energy would be spent in restoring the wheel energy
and consequently, more time would be consumed in the pass than is
the case in a sufficiently powerful engine without a flywheel. This
loss of time and heat partly offsets the apparent gain ine c j nomy of the
smaller engine. But the more serious loss would be experienced in a
three-high mill, in both time and heat, as well as the additional power
required for raising and lowering the ingot to the two different
levels for each succeeding pass. Considering the shortness of the
passes in blooming-mill work, this delay would be a very serious loss.
6 It therefore seems to me that but Uttle, if any, substantial
advantage can be gained in heavy blooming-mill work by the three-
high mill so long as it is steam-driven. It also seems to me that the
only hope of any improvement in economy over present practice
will be in the use of the present two-high reversing mill, but driven
electrically. In such an equipment, we would have the necessary
power to avoid delay on gripping the ingot, the means for instantly
throwing off the power at the release of the ingot, and also the
continuously running steam engine with a sufficiently heavy fl3rwheel
at the generator.
E. W. Yearsley.' The value of the flywheel as a means to obtain
constant load with intermittent work is well illustrated by Mr. Caine's
experiments. This arrangement has been considerably developed in
conjunction with electrically-driven rolling mills. Where consider-
able speed variation is allowable, and there is a suitable ratio of
pause to operation time, the fl^'wheel may be applied to many drives
with economy.
2 Economical considerations are at present of great importance
in the steel industry. Engines used for driving rolling mills are
usually excessive steam consumers. There is no doubt that their
performance in this respect can be greatly improved, especially for
continuously running mills. In my opinion the electric motor will
be found more reliable and satisfactory for this work, and it will be
desirable to confine the refinements necessary for great economy of
prime movers to an electric generating station.
3 Mr. Caine's method of regulating the governor is somewhat
analogous to that used for controlling the rate of application and the
'Electrical Engineer, Midvale Steel Company, Philadelphia, Pa.
GOVERNING ROLLING MILL ENGINES 797
limit of electric current to a main roll motor, in order to obtain the
similar results of more uniform load, less rapid speed variation,
and protection of the driver. The tests show conclusively the
improvement in steam consumption and performance resulting.
4 As the paper points out, the problem is considerably compli-
cated by variation in the number of pieces passing simultaneously,
also by variation of the interval between passes and its relation to
the time of the pass, and in the temperature and composition of the
material. A speed variation of from 12 to 20 per cent transferring
from 23 to 36 per cent of the kinetic energy of the flywheel, has been
found desirable. With a given torque, time of load, and interval,
this speed change fixes the weight of wheel required. Data of power
performance of rolling-mill drives are rapidly accumulating. This
paper is an interesting addition to such information.
The Author. Mr. Ord seems to have the impression that there
was a great difference in the work done in the two examples given.
As a matter of fact, the area of the piece was the same in each case,
on entering the first pass, and therefore, the total work for the four
passes would be as their relative weights, 2680 lb. for Case A and 2550
for Case B; B having a shght advantage in weight, and A an advant-
age of 5 lb. in steam pressure, so that the work was practically the
same in each case.
2 The valve setting was not altered between tests. The differ-
ence in the behavior of the engine was due to the adjusting screw
alone; and now, three years after these tests were made, this screw
is still in service. This method of engine control does not eliminate
the reserve power; it does cut it down to a point where judgment says
there is still sufficient reserve to answer all requirements.
3 Mr. Tribe asks the reason for building reversing engines with
such a large surplus of power. Such engines are usually driving
blooming mills, where it is no uncommon practice to roll about one-
half of the total number of passes, from bloom to finished product, in
one stand of rolls, the remainder being taken care of from three,
four or more stands of rolls, so that the blooming mill must handle
these passes in very rapid succession in order to get the tonnage.
The engineer handles the throttle and reverse levers, and the roller,
the screw-down and the table rolls. The screw-down adjusts the
distance between the rolls ; consequently it fixes the amount of reduc-
tion on the bloom and the load on the engine is proportional to the
reduction.
798 DISCUSSION
4 The screw-down has no fixed limits for each pass, therefore it
will be set in a very short period, according to the judgment, or lack
of judgment, of the roller. The wi-iter has timed these operations
with a stop watch and found that quite often the adjustment was made
in less than two seconds; that is, the time from the end of one pass to
the beginning of the next. It is quite likely that the screw-down does
not get located where the operator intended; if the reduction is less,
the roller will make some other passes heavier because he does not
wish to add two additional passes. From the calculated results,
from continuous indication cards on an engine of this type, on a
single bloom one pass was noted where no reduction was made, while
another pass required nearly three times the average power. From
this sort of operating conditions, coupled with the desire to get an
engine that wdll not stall under any circumstances, it becomes very
evident why there is a great surplus of power. This also calls atten-
tion to one of the features in favor of the three-high mills, namely,
that the roll designer can distribute the work approximately equally
on every pass, with the proper data at hand.
5 The fact that the reversing engine is man-governed is brought
out. This practically places the speed limit at the rate at which
it would run with a wide open throttle and nothing in the mill ; which
would far exceed 25 per cent of the normal. Speed curve A shows
that our engines run at about 16 per cent above normal, and with
curve B at but 10 per cent above.
6 Mr. Yearsley suggests that the principle involved might be
applied to other than mill engines. The ^vriter can cite an instance
where this was done. Our company has two-crank flywheel hydrau-
lic pumps which are started and stopped by an accumulator. When
the accumulator would drop, the governing throttle valves would
open wide and the pumps would run up to the speed determined by
the fiy-ball governors (50 r.p.m.), and when the accumulator reached
the top limit it would shut off the steam, stopping the pumps verj'^
abruptly. This continual starting and stopping caused considerable
trouble in keeping up the various adjustments, and pins ran hot at
times. Upon my suggestion the engineer in charge adj usted the govern-
ing throttle valves so that they could be only partially opened, and
as a result the maximum speed is just a little above the average, the
pumps running almost continually at about 20 r.p.m., the trouble
with hot pins is no longer experienced, the rod adjustments last several
times as long, and it is my belief that the water valves must give
less trouble.
Xo. 1261
AN EXPERIENCE WITH LEAKY VERTICAL
FIRE-TUBE BOILERS
By F. W. Dean, Boston, Mass.
Member of the Society
In 1905 I made a design for a large vertical fire-tube boiler, two
of which were built, to be placed on a brick fire box provided with a
chain grate. In accordance with a great number of precedents the
water leg was short, being in fact 2 ft. deep below the underside of
the crown sheet. Unusual provision was made for easy circulation
by wide spaces between tubes at every 45 deg. of the circumference
instead of the customary 90 deg., or as in some cases at 180 deg.
The distance from the top of the grate to the underside of the crown
sheet was 7 ft. The following are the general dimensions of the
boilers as they now are:
Inside diameter of smallest course of shell 120^ in.
Inside diameter of largest course of shell 1221^ in.
Inside diameter of water leg 112 in.
Height of water leg 7 ft. 2| in.
Height of brick furnace 5 ft. 8 in.
Distance from grate to tube plate 12 ft. 2 J in.
Outside diameter of tubes 2^ in.
Length of tubes 20 ft. 0 in.
Number of tubes 488
Pressure for which the boiler was designed 165 lb.
I^d of grate B. & W. chain
Size of grate 8 ft. 6 in. by 9 ft. 0 in.
Grate area 76.5 sq. ft.
Water heating surface, say 4900 sq. ft.
Superheating surface 1181 sq. ft.
Total heating surface 6081 sq. ft.
2 The boilers were designed for S. D. Warren & Company and
were used in their paper mill at Cumberland Mills, Me. They were
built by the Portland Company of Portland, Me. Each boiler was
Presented at the Annual Meeting, New York (December 1909), The
American Societt of Mechanical Engineers.
SOO LEAKY VERTICAL FIRE-TUBE BOILEBS
rated at 500 h.p., or substantially 1 h,p. for every 10 sq. ft. of water-
heating surface, and was expected to work at 1000 h.p. a good portion
of the time. Artificial induced draft was used and it was possible to
obtain a draft of 2^ in. of water in the smoke box.
3 At the back end of the chain grate, instead of a water back or a
brick back, a vertical or slightly inclined common grate was used,
against which the unconsumed coke would accumulate, and under
which the ashes would pass and fall upon the ashpit floor. Difficulty
was found in making the coke accumulate uniformly and the ends of
this grate were frequently bare.
4 The boilers were started gently and then operated at high
capacity. After about two weeks a number of the tubes began to
leak at the lower ends. They were expanded, but shortly began to
leak again, and this process was repeated until the tubes were so
injured that they could not be further expanded. They were then
removed and new ends were welded on, but after a comparatively
short time they leaked again. The leaks were more on the back half
of the boilers than on the front. In winter when a nearby door in the
building was open and cold air blew on the vertical grate, when
the ends of the grate were bare, the leakage would increase. After
learning how to keep the vertical grate covered, and keeping the
door closed, the general trouble continued.
5 Knowing the sensitiveness to dirt on the crown sheet of vertical
fire-tube boilers, and as the design permitted access to the interior, these
crown sheets were examined and found to be clean. Thinking that
possibly some invisible oil had entered the boilers in some way, one
of the boilers was boiled out with caustic soda, but with no effect.
6 The opinions of several boiler experts were obtained, but they
differed and were unsatisfactory. One thought that the workman-
ship was poor, another that the design was the worst he had ever
seen. Another thought that the tube plates were too limber and
even recommended riveting crown bars to them to stiffen them.
7 Spring Hill coal from Nova Scotia was used at first, followed
by New River coal from West Virginia. With the Spring Hill coal
the lower ends of the tubes quickly became incrusted with clinker,
and were finally closed by it and a little later the clinker would hang
in stalactites from the tube ends. Not all of the ends would be
closed, but this was the case with a large proportion of them.
With New River coal there was less trouble. Spring Hill coal was
satisfactory under horizontal boilers and never plastered over the
tubes; in the vertical boilers, however, the incrustation was so hard
LEAKY VERTICAL FIRE-TUBE BOILERS 801
that it had to be removed with chisels. This incrustation was of
course moulten earthy matter injected by the draft against the tubes
and tube plate and there congealed by the comparatively low tem-
perature of the metal. In the horizontal boilers it falls to the bottom
of the setting before itjarrives^at the^^tubes. In water-tube boilers
it can be seen^adhering^to the lower tubes.
8 The existence of this incrustation probably furnishes the explan-
ation of the tube leakage. As a large proportion of the tubes became
stopped up the others had to pass all the hot gases, the water about
their ends was probably driven away and they became very mucli
overheated, causing them to over-expand, to become upset, and at
some later time, when they became cooler, to be loose in the holes.
9 As a last resort, when it seemed as if the boilers must be con-
signed to the scrap heap, someone suggested that to length';n the fire-
box and raise the boilers by the amount of the extension might cure
the trouble. One boiler was thus altered, started August 31, 1908 and
run at the estimated rate of 1100 h.p. 24 hr. per day for some three
months, without the slightest leakage, although the tubes were very
thin from over-expansion. The other boiler was then altered, and
started February 25, 1909. When the first boiler was worked at the
estimated rate of 1100 h.p. it consumed 84,000 lb. of New River coal
in 24 hr., burning it at the rate of 46 lb. per sq. ft. of grate per hr.
Neither boiler has leaked up to the time of presenting this paper.
10 The distance from the grate to the tube plate is now 12 ft. 2f
in. There is some hicrustation, but it is light, brittle and easily
crushed, and can be blown off by a rotating multiple tube-blower in
the smoke box. The tubes are conveniently and quickly blown in
this way every three or four hours.
11 The boiler plant at this mill consists of Babcock & Wilcox
boilers, 90-in. horizontal return tubular boilers, and the two vertical
boilers described. The latter were intended to reduce the space
occupied, both on the floor and above. The rear drum of the chain
grate is exposed and the clinker is dropped at the back end, where it
is easily removed without inconvenience to the fireman. Above is
room for the smoke flues and economizer, which with other types
of boiler would have been p) ^ed. in this case, with diflEiculty.
12 After these boilers had been operated long enough to show
that they were reliable and a gO'jd investment, it was decided to test
one of them, with the results given in Tables 1 and 2.
13 While the evaporation is good it is not satisfactory. The
function of a boiler is to absorb the heat generated in a furnace. The
BEFORE ALTERATION
Fig 1 Boiler Before Alteration
LEAKY VERTICAL FIRE-TUBE BOILERS
sas
Fig 2 Lower Part of Boiler Showing Combustion Chamber
After Alteration
804 DISCUSSION
furnace efficiency may be poor and the boiler efficiency good, and
that was the case during these trials. The best furnace result occurs
when the carbon is burned to COj with as little surplus air as prac-
ticable. In these trials the COg was low and some CO was neai-ly
always found. It was impossible to get any better combustion for
reasons which I do not know. Experiments will probably be made
to ascertain this and overcome the trouble. It might disappear with
another kind of stoker.
14 That the boiler efficiency was good is evident from the low
temperature of the escaping gases, when developing over 1000 h.p.,
which is at the rate of less than 5 sq. ft. of water surface per horse-
power. It will be noticed that the evaporation was best on combus-
tible when the boiler was operated at double its rated horsepower.
It was found impossible to keep the horsepower down to 500. This
could only be done by reducing the grate area.
15 An interesting result of the tests is that the superheat was
the same at all rates of power.
16 Returning again to the matter of clinker on tubes, it occurs
on locomotives which burn anthracite coal, and I understand on
locomotives of the Boston & Maine Railroad that burn coke. In the
latter case coke-burning locomotives cannot be used on long runs,
but whether a better quality of coke would be more successful I do
not know. Professor Denton has informed me that clinker accu-
mulated on the boiler tubes of a Transatlantic steamship on which he
was a passenger to such an extent that men were sent into the com-
bustion chambers to remove it. He also referred to a Manning boiler
on which this trouble occurred.
DISCUSSION
Reginald P. Bolton. It appears to me that this design of boiler
was an invitation to the trouble that followed, and it is only neces-
sary to go back into the experience of other people to find out that
others have suffered in the same manner. If the view of the boiler
as presented in the paper is turned horizontally, and it is imagined
that it is a locomotive boiler cut off short, it will be seen that there is
no combustion chamber whatever in it. This boiler was to be put
to a service which might call for a rate of combustion in the furnace
demanding double its rated capacity output, so that the double
aggravation of a very small combustion chamber and verj^ large
ate of combustion, was present.
LEAKY VERTICAL FIRE-TUBE BOILERS
805
TABLE 1 TESTS OF BOILERS
Time
Draft
Tempera-
ture
CO2
April 29, 1909
8.35- 9.00
0.75
575
7.6
11.2
0.6
80.6
9.05- 9.30
full
575
10.0
9.0
0.3
80.7
9.35-10.00
full
575
7.2
12.2
0.1
80.5
10.05-10.30
full
575
7.0
12.8
0.3
79.9*
10.35-11.00
full
580
7.8
11.2
0.1
80.9
11.05-11.30
full
600
7.0
12.8
0.2
80.0
11.35-12.00
full
620
7.8
11.6
0.1
80.5
12.05-12.30
full
620
8.2
10.7
0.2
80.9
12.35- 1.00
full
610
8.2
11.5
0.5
79.8
1.10- 1.30
full
585
7.7
12.3
0.1
79.9*
1.40-2.10
full
600
8.2
12.3
0.2
79.3
2.20-2.50
full
625
7.4
12.2
0.1
80.3*
3.00-3.25
full
585
7.0
13.3
0.0
79.7
3.35-4.00
full
600
4.8
14.8
0.0
80. 4t
4 . 05-4 . 30
full
610
7.8
12.0
0.3
79.9
4.35-5.05
full
615
8.3
10.9
0.2
80.6
5.10-5.25
full
640
8.8
10.6
0.4
80.2
5.30-5.50
full
620
8.4
April 30, 1909
May 1, 1909
8.05- 8.30
0.5
575
8.3
11.2
0.3
80.2
8.37- 9.05
0.5
540
8.3
11.1
0.1
80.5
9.15- 9.37
0.5
13.5
5.3
0.8
80. 4t
9.45-10.10
0.4
500
10.7
7.8 .
0.9
80. 6t
10.20-10.45
0.2
600
10.0
8.7
0.4
80. 9t
11.05-11.30
0.2
510
11.0
7.4
0.6
81.0
11.35-12.00
0.2
520
9.6
8.8
0.8
80.8
12.05-12 30
0.2
525
7.9
11.8
0.2
80. 1§
12.35- 1.10
0.2
500
6.7
12.7
0.0
80.6
1.15- 1.45
0.2
525
7.9
12.2
0.2
79.7
1.50- 2.25
0.2
515
7.1
12.5
0.1
80.3
2.30- 3.00
0.2
485
7.0
11.6
0.1
81.3
3.05- 3.35
0.2
520
6.3
13.9
0.0
79.8*
3.40- 4.00
0.2
560
7.4
8.00- 9.00
520
7.3
13.3
9.15-10.00
0.8
525
7.2
12.8
10.00-10.55
0.8
5.0
14.6
11.05 11.40
0.9
530
7.7
11.6
12.45- 1.30
0.7
630
9.0
'
1.30- 2.30
0.7
560
8.6
10.4
2.40- 3.30
0.7
565
6.0
13.2
♦Cleaned fire.
tThin on backside.
0.0
0.0
0.0
0.2
0.3
0.0
tFire thick and banked.
§Fire thinner.
79.4
80.0
80.5
80.5
80.7
80.8
S06 DISCUSSION
TABLE 2 RESULTS OF BOILER TRIALS
Running Start-and-Stop Method of Trial, Vertical Fire-Tube Boiler, New River
Slack Fuel
Date of trial ' Apr. 29 ' Apr. 30 May 1
Duration of trial, hours 7 8 8
Number of boilers in use one one one
dimensions and proportions
Grate surface 8 ft.6 in. by 9 ft. 0 in., square feet. .
Water-heating surface, square feet
Superheating surface, square feet
Total heating surface, square feet
Ratio total heating surface to grate surface
average pressures
Steam pressure, by gage, per square inch, pounds ....
Atmospheric pressure per square inch, pounds
Absolute steam pressure per square inch, pounds. . . .
Force of draft in column of water between damper and
boiler, inches ■
ATERAOB TEMPEBATURES
External air
Feed-water before entering
Escaping gases after leaving
Steam
Moist coal consumed, poimds.
Moisture in coal, per cent. . . .
Dry coal consumed, pounds..
Wood consumed, pounds
Total dry refuse, pounds
Total dry refuse, per cent
Total combustible, pounds . . .
17
18
19
20
21
22
23
24 Dry coal consumed per hour, pounds.
QUALITY OF STEAM
25 I Moisture, per cent
26 I Degrees superheated
BRITISH THERMAL UNITS
27 I Number of heat units in a pound of dry coal, by oxygen
calorimeter
28 Number of heat units in a poimd of combustible, by
oxygen calorimeter
29 ■ Specific heat of superheated steam at constant pressure
30 j Heat units absorbed per pound of steam generated . . .
31 j Heat units imparted to boiler per pound of dry coal
32 j Heat units imparted to boiler per pound of combustible
EFFICIENCIES
33 i Efficiency based on dry coal, per cent
34 EflBciency based on combustible, per (
FACTORS OF evaporation
36 I Factor of evaporation
76.5
76.5
76.5
4900
4900
4900
1181
1181
1181
6081
6081
6081
79.5
79.5
79.5
129.8
124.3
126.8
14.7
14.7
14.7
144.5
139.0
141.5
1.47
0.28
0.63
38
35
39
50.6
63.6
58.2
609.1
517.3
558.6
372
369
371
24,192
17,030
21,782
3.06
1.45
4.55
23,452
16,783
20,791
0
0
0
2361
1310
1650
10.02
7.8
7.94
21,091
15,473
19,141
3013
1934
2393
0.00
0.00
0.00
17
17
17
14,759
14,759
14,759
IWJrti
15,760
15,760
15,760
0.61
0.61
0.61
1182
1174
1167
9901
10,075
9893
11,009
10,929
10,746
67.1
68.3
67.0
70.0
69.3
68.2
1.224
1.215
1.209
LEAKY VERTICAL FIRE-TUBE BOILERS 807
TABLE 2 RESULTS OF BOILER TRIALS.— Continued
Date of trial
Apr. 29 Apr. 30 May 1
WATER
36 Total water pumped into boiler, pounds
Water actually evaporated, corrected for quality of
steam i
Equivalent water from and at 212 deg. fahr., pouuds
Equivalent water from and at 212 deg. fahr., per hour,;
pounds ,
37
196,441 144,025 178,250
198,185 145,321
242,578 176,633
34,654
EVAPORATIVE PERFORMANCE
Water actually evaporated per pound of dry coal,
pounds
Equivalent per pound of dry coal from and at 212 deg.j
fahr. , excluding economizer, pounds
Equivalent per pound of combustible from and at 212-
deg. fahr., excluding economizer, pounds
COMMERCIAL HOR8EPOWEB
On basis of 34 J lb. of water from and at 212 deg. fahr.,
per hour, by boiler h. p
Heating surface of water surface per horsepower, square
feet
Horsepower per square foot of grate surface, h.p
Rated horse power
Percentage developed above rating
RATE OF COMBUSTION
Dry coal actually burned per square foot of grate sur-
face per hour, pounds
Dry coal burned per square foot of water-heating sur-
face, per hour, pounds
RATE OF EVAPORATION
50 Water evaporated per square foot of heating surface
per hour from and at 212 deg. fahr., pounds
40
42
43
48
49
1005
39.1
0.61
7.05
22,079
640
25.3
0.39
4.50
179,854
217,444
27,193
8.45
8.66
8.65
10.34
10.52
10.46
11.50
11.42
11.36
788
4.87
7.66
6.22
13.14
8.36
10.30
500
500
500
101
28
58
31.3
0.49
5.55
808 DISCUSSION
2 The design of the boiler is radically defective in two import-
ant points, namely, the tubes are entirely too long, and the com-
bustion space was entirely too small. It is now very nearly half
a century ago that the experiments of Geoffroy and Petiet demon-
strated the futility of unduly lengthening fire tubes. These experi-
ments demonstrated the rapid reduction in efficiency due to length of
tubes, under various conditions of draft and rates of fuel consumption.
Almost precisely the same conditions were tested as in the author's
boiler, as follows:
3 A consumption of fuel exceeding 50 lb. per sq- ft- of grate,
under an air pressure of 2.36 in. with the following results:
Evaporation
per Sq. Ft.
Lb.
Fire-box plate. 23.5
First three feet of tubes 5.4
Second " " " " 2.5
Third " " " " 1.33
Fourth " " " " 0.83
Fifth, three feet evaporated only 0.48
Sixth " " " " 0.3
The last two were found by extending the curve.
4 An examination of these results might have dissuaded the
author from the mistake of designing the boiler with such a length
of tube, involving not only inefficiency, but the evident concomitant
of leakage as a result of expansion and contraction. Apart from the
other defective feature, the boiler could have been shortened so as
to reduce the tubes at least five feet in length, and would no doubt
have given better efficiency as a result.
5 The general type of the boiler possesses nothing new or original
unless we may so regard the restricted combustion chamber, by
which the tube plate was brought within seven feet of the grate,
allowing a total capacity of only 535 cu. ft. for the fire and for the
gases of combustion.
6 A very simple computation of the results of the combustion of
40 lb. of coal per square foot of grate area, will show that the volume
of products of combustion would be so great, that only an excessively
heavy draft could force them through the combustion chamber and
tubes, and that incomplete combustion was bound to result.
7 The addition of 5^ ft. to '^the'^height of the chamber, which was
arrived at only after three years' experience with this boiler, nearly
doubled the effective space for combustion, and also removed the
ends of the tubes from the direct action of the blast. It may be
LEAKY VERTICAL FIRE-TUBE BOILERS 809
observed that a Dutch oven would have afforded equal results, at
perhaps less expense.
8 The reason for the adhesion of molten clinker to the ends of
the tubes, need have presented little difficulty, in the light of past
experience, since the ends of the tubes were placed so close to the
fire. This result developed in the fire-tube boilers of H. M. S. Poly-
phemus nearly thirty years ago, and when found in the boilers of
locomotives is due to precisely the same cause.
9 It will be noticed that the best of the tests which were made
after the change of combustion chamber was effected, is that in which
the rate of fuel consumption is least.
10 I agree with the quoted conclusion of the second boiler ex-
pert, referred to in Par. 6, and am at a loss to understand why such
an opinion, thus expressed, was regarded as unsatisfactory. It may
be hoped that the paper may stand as a warning signal to other de-
signers. It requires a great deal of courage to present a paper of
this kind, and the author should be thanked for bringing forward
a record of a failure so that we may profit by the facts.
William Kent. I join with Mr. Bolton in praising Mr. Dean's
courage in bringing forward a report of his failure, and I regret that
some eight or ten years ago I did not bring forward a record of another
similar failure, not my own, but that of some other man, which might
have prevented Mr. Dean's. The New York Steam Company bought
a boiler for their Greenwich Street Station to go in a very small
ground space. It was a very large plain vertical cylinder boiler,
eight or ten feet in diameter, full of tubes about 20 ft. long, and
was rated at 1000 h.p. It had not been in use more than a week or
two when it began to leak. There was no way to clean the flat tube
sheet or to clean the tubes of scale, and the boiler was condemned and
taken out.
J. C. Parker. The reason that the tubes leaked was that when
the boiler was set close to the grate the tube ends were subjected to
wide fluctuations in temperature. The flow of air through a chain
grate increases toward the rear end, and where the boiler was set
higher there was more mixing of the hot and cold currents and, con-
sequently, less fluctuation in temperature.
2 The ciinkering of the tubes would naturally increase the trouble
because of the concentration and increased friction of the gases in
the tubes that remained clear.
810 DISCUSSION
Orosco C. Woolson. This discussion has brought out the im-
portant fact that perfect combustion should take place before the
gases reach the tubes or shell of the boiler,
2 I have been somewhat surprised in toy travels among the
cotton and woolen mills of the eastern states where the manage-
ment have large experience in cotton spinning but'^'are limited in
personal experience regarding what constitutes the'^production of
the highest calorific value of a pound of bituminous coal. One man
of large experience in mill work wanted his furnace fire directly
under the tubes of his vertical boilers, and gave me his reasons. I
told him that I would guarantee him better results if he would dis-
card the idea that the area immediately under and against the tube
sheet should act as a combustion chamber. Let combustion take
place entirely before it reaches the tube sheet and the results will
be much more satisfactory.
3 Secondly, as to the tubes filling with vitrified slag or any other
residuum of combustion, I would suggest that such deposit should
be made to take place under a fire arch, where ^it ,will adhere to the
crown and serve a useful purpose by forming a refractory coating.
This practice is becoming popular, and more so today than ever
before. It is my opinion that by completing combustion under a
properly constructed arch within a properly constructed combustion
chamber, the products of this combustion will be sent to the boiler
in the form of what we will term "caloric ether" and not a mixture
of its original constituents which play no useful part, under the
circumstances, in producing or maintaining heat, but rather are
subject to ready condensation.
A. A. Gary. In my experience with vertical fire-tube boilers I
once found a boiler containing shorter tubes and of a greater dia-
meter than ordinarily found in the Manning type. The fuel used
was a moist anthracite coal, and there was a natural draft of more
than one inch of water in the smoke box over the boiler. The draft
could not be regulated, due to the previous burning out of the steel
plate butterfly damper. The partially burned furnace gases passed
rapidly through the vertical tubes and ignited above the top tube
sheet, thus causing the destruction of dampers and the steel breech-
ing, to say nothing of the reduced^evaporation in the boiler due to
this waste of heat.
2 The trouble was remedied by placing the grates at a greater
distance from the lower tube sheet and arranging baffles in the
LKA^KT VERTICAL FIRE-TUBE BOILERS 811
combustion chamber so as to insure the more complete combustion
of the gases before they entered the tubes. A cast-iron plate dam-
per replaced the former one of steel plate, and no further trouble
has since been experienced.
3 In another ease, the question came' up as to the advisability
of. applying a special automatic furnace, using bituminous coal and
producing very high temperatures, under boilers of the Manning
type. An arrangement which has been used in New York City for
a number of years was suggested and successfully applied.^
4 Fire-bricks, piled on edge with open spaces between the bricks,
were arranged a short distance beneath the lower tube sheet. This
checker work of bricks filled the entire space beneath the boiler, the
openings between the bricks at the center being very much reduced,
so as to cause a decreased flow of gases directly under the center of
the overhead tube sheet. By this means, a very even distribution of
temperature was secured over the entire area of the lower tube sheet
with a slight reduction of heat delivery at its center, the^most sen-
sitive portion of the whole tube area.
5 The author mentions inefficient combustion, which is indi-
cated by the comparatively low percentage of CO2 and high :per-
centage of O, shown in Table 1. As the higher temperatures are
secured by the most complete combustion with the least excess of
air, the question arises, why should such destructive results follow
such inefficient furnace conditions?
6 Pyrometric testing with gas analyses have taught me that
when a furnace is being operated inefficiently, very high tempera-
ture may be found in one part of the furnace while at the same time a
comparatively low temperature may exist in another part. This
may lead to the simultaneous impingement of gases of very differ-
ent temperatures upon various parts of the lower tube sheet, setting
up destructive strains and contributing to such troubles as have been
described by Mr. Dean.
7 The lower tube sheets of boilers of the Manning type are very
sensitive, especially towards the center of the sheet where the water
seems to penetrate with great difficulty, thereby failing to keep
this portion of the heating surface constantly wet.
8 Concentration of heat due to concentration of combustion and
lack of space for this small volume of high-temperature gas to dif-
fuse itself throughout the entire mass of furnace gases before they
reach the tube sheet, is bound to cause trouble., especially when this
highest temperatiu-e is concentrated against the center of the tube
812 DISCUSSION
sheet on the inner surface of which there is apt to be little or no water.
After the center of this sheet loses the supporting effect of the center
tubes, acting as stays, the surrounding tubes are very apt to follow.
9 Concerning the low efficiency of the furnace referred to in
Par. 13, there should be no trouble in remedying this fault. A prop-
erly conducted furnace test (apart from the boiler) with pyrometers,
gas-analyzing apparatus, etc., will show just where the trouble exists
and will point out the needed changes as well as the limitations
under which this type of stoker can be worked with the different
grades of fuel used.
Prof. L. P. Breckenridge. One of the speakers said that the
highest temperature in a boiler furnace is directly over the fire. This
is not always so. We have measured the temperature twenty feet from
the fire and found it higher. It depends on the volatile content of
the fuel and whether the flame has been supplied with a sufficient
amoimt of air early in the process of combustion. It is this that
determines whether the high temperature point is ten feet or twenty
feet away. Many times in our experiments in the St. Louis boiler
trials we have seen that every time the furnace door was opened the
temperature at the rear end of the combustion chamber went up,
because when more air was admitted the combustion was better and
the temperature increased.
2 For experiments concerning the transmission of heat through
a boiler tube, it occurs to me that Mr. Dean has designed one of the
most satisfactory laboratory boilers I have seen. There has been
much discussion of late on the heat transferred through a boiler tube,
as influenced by the velocity of the gases passing thi'ough the tube.
This boiler with its large number of tubes would be just the type
with which to make a test on this point. I wish Mr. Dean would
.burn a large amount of coal per square foot of grate in this boiler
furnace, using, first, all the tubes, and secondly, only one-half the
tubes. If the same amount of coal was burned in each case the ve-
locity of the gases through the tubes would be twice as great in the
second case, and it would be interesting to know the relative amounts
of heat transferred.
3 I hope that some time we may take up the question of the burn-
ing of fuel, making a distinction between the economical performance
of the boiler and of the furnace. We have reached a time when we
can intelligently discuss these questions separately. Anthracite coal,
on account of its high fixed-carbon content, is burned mostly on
LEAKY VERTICAL FIRE-TUBE BOILERS 813
the grate itself. When burning semi-bituminous coal, with 18 to
20 per cent volatile content, a large combustion chamber is re-
quired, and as the volatile content increases the size of the combustion
chamber must be increased. When burning bituminous coal, with
40 per cent volatile content and 20 per cent ash, the fuel actually
burned on the grate is small. The grate supports the fuel and some
coal is burned there, but it is in the combustion chamber that we
burn fully one-half the combustible part of our fuel. It is evident
that more attention must be given the proportions of our combus-
tion chambers when burning high-volatile coals, and especially at
high rates of combustion.
Prof. A. M. Greene, Jr. In London Engineering for October
22 and November 5, 1909, appeared an article by Professor Dalby,
in which he summarized a number of articles referring to heat trans-
ference through plates. I would commend the article to the atten-
tion of all the members of the Society interested in this matter.
2 In London Engineering for February 1909, Professor Nicholson
described experiments showing clearly that only a small part of the
possible heat transmission through plates is utilized. I mention this
to call the attention of the members to the fact that some data are
available on this subject. In this article are given the formulae
for heat transmission which may be compared with the results of
German Experiments recently completed at Dresden (Zeit. des
Verein Deutscher Ing., October 23, 1909).
WiLLLAM Kent. In another issue of London Engineering, a cor-
respondent showed that the idea of high speed of the gases being
favorable to combustion was negatived by the Lancashire boiler, in
which the flues are very large and the speed of the gases low, yet the
economy is as high as in any other boiler.
Reginald P. Bolton. It is mainly a question of the difference
in temperature between the inside and outside of the heating sur-
faces. The lower the temperature of the feed water, and the higher
the temperature of the fire, the greater will be the efficiency of the
boiler.
E. D. Meier. I find myself in substantial agreement on some
points with all the gentlemen who have spoken. I want to say for
Mr. Dean, that he is correct in his conclusion that the precipitation
814 DISCUSSION
which occurs at the bottom of the tubes has a great deal of influence
on the overheating of the tube sheet. The other causes which were
mentioned are also true, but there is no doubt an accumulation of
carbon there. I do not know whether Mr. Dean preserved any of the
precipitate or stalactites, but I believe a large part of it was uncon-
sumed carbon, which will remain at a high temperature for some time.
2 I am reminded of an experience which I had with water-tube
boilers at the Chicago World's Fair. I think there were ten differ-
ent makes of water-tube boilers, most of them sub-horizontal, but
some of the vertical-tube type and some of the bent-tube type. We
were burning crude oil, and all the boilers suffered from the same causes,
— every one lost tubes by burning out. Some were careful enough to
shut down a boiler as soon as they noticed the blisters on the tubes.
3 The boilers which I had at Chicago were afterward placed in
the midwinter fair at San Francisco, and were fired with California
crude oil for seven months without a tube being lost. These boilers
were afterwards sold with the condition that if the customer found
any tube damaged it would be replaced, but not one was found to
be burned. That bears on the subject mentioned by Mr. Dean.
The trouble we found was this: The oil is supposed to be atomized
in the burners, but this is not always the case. Little slugs of oil
would fly up and adhere to the tube, and would spread and slowly
carbonize. They would not burn, because no air could get to them.
One little spot, a half inch in diameter, would become red hot in
spite of all the circulation of water, and would ultimately burn out
and make a blister.
4 When the boilers were installed in California, the oil burners
were placed lower and were directed downward so that the jet would
strike the bottom of the combustion chamber at a distance of six
feet from the front, hence there was no chance of oil [striking
the tubes. Perfect combustion was obtained, and on one occasion
one of the boilers was forced so hard that a picture was taken of the
inside of the furnace by its own heat. I have that photograph still,
to show what can be done. One can see a perfectly white heat and not
a single blister on the tube. In Mr. Dean's case carbon was deposited
and became incandescent, and gave an intense local heat on some
point, which accounts for the failure of the tubes at such point.
5 In regard to the combustion chamber, I agree with Professor
Breckenridge. I have always been a believer in a large combustion
chamber, and one of my early experiences in that direction was when
in charge of a plant having two horizontal tubular boilers, using
LEAKY VERTICAL FIRE-TUBE BOILERS 815
Illinois coal. At that time everybody in the Mississippi Valley
believed in river practice. The boilers, engines and dimensions of
pipes, etc., were according to river] practice. The boilers were set
with the grate twelve inches from the bottom of the she. Ill
raised them to thirty inches, and I was told I would not get any heat,
but I got better results, and the boilers lasted longer. The increase
in 'the distance from the fire to the shell was a great advantage,^and,
of course, incidentally I increased the eflBcicncy )f the boiler.'
David Moffat Myers. In my paper on Tan Bark as a Boiler
Fuel, results of an efficiency test are given in Table 4 in which the
temperature inside the furnace was 1100, the temperature in the
combustion chamber, under the boiler, was 1475, the flue tempera-
ture was 493, and the thermal efficiency was 71.1 per cent.
2 These figures prove that under conditions of good efficiency
it is quite possible to have a higher temperature at some distance from
the fuel than close to it. The combustion of the gases is simply retarded
to a later point of their travel. This might be caused by the com-
bination of a high velocity of draft with a moderate air supply, so
that the oxygen does not come into sufficiently intimate contact with
the fuel gases in the primary combustion chamber, that is, in the
furnace proper. In the case quoted, the CO2 ran almost uniformly
at about 12 per cent, the 0 between 6 and 7 per cent, with no deter-
minable CO.
A. Bement. In the boiler which Mr. Dean describes, I like the
scheme of having the rear end of the chain grate exposed so that it is
accessible. The capacities obtained with these boilers are very large;
the strength of draft, however, is somewhat too much for an ordinary
chain-grate fire. It is my experience that chain grates are not pro-
portioned so that it is possible to carry the requisite thickness of fire
for a draft such as existed in this case. I think this will account for
the low percentage of CO2 in the combustible gases, ^and in thisjs
found the reason why the efficiency was not higher.
2 I would attribute the leaking of the tube ends in the head over
the fire to another cause than that given. Considerable experience
in similar cases leads me to believe that the trouble is caused by
excessive heating on the delicate tube ends in the flue sheet. There
are two thicknesses of metal to be penetrated before the heat reaches
the water; also the opportunity for water to enter among the tubes
and to flow freely over the heated parts is rather restricted. When
816
DISCUSSION
the ordinary return tubular boiler is set with a fire under the shell,
a^large portion of the heat flows"through]the shell, with the result that
the'^temperature of the gases^is much ^'reduced, so that by the time
they^impinge'upon the tube^sheet,^their^temperature is low enough
so that no'damage results. I
Fig. 1 Setting of a Fire-tube Boiler in Which the Tdbes Leaked
Pl4W^^^^^^.<;^^;^^^^^^:%^^:^^^^IT^
Fig. 2 Showing Water Leg to Lower Temperature of Gases Imping -
iNG on Tube Sheet
3 A case of trouble of this kind is illustrated by Fig. 1 and Fig. 2,
the first showing a return tubular boiler set agamst an enclosed fire-
brick furnace, in which the gases first impinged upon the tube sheet,
passing through the tubes to the other end of the boiler, thence find-
T.EAKY VERTICAL FIRE-TUBE BOILERS 817
ing exit by wa}'^ of a chimney attached thereto. When these boilers
were put at work immediate and very serious trouble resulted with
the tube ends; they leaked very badly, the bead getting out of shape
and springing away from the sheet. By means of a little door in the
side of the furnace one could see the water squirting from every tube,
and running away from the setting on the floor in a large-sized stream.
4 A remedy was effected in this case, as shown byFi g. 2, by mount-
ing above and below the furnace a drum which extended crosswise of
the setting, and connected by vertical 4-in. boiler tubes as indicated;
each of these drums being in communication with the boiler,
allowed circulation of water and steam. With this scheme the gases
first pass between these vertical tubes, which are set closely together,
with the result that there is a considerable reduction in the temperature
of the gases before they came in contact with the end of the boiler
tubes.
5 Another case of this character was remedied by carefully
cleaning off the end of the boiler and coating it with an asbestos
cement, which was rounded over and into the boiler tube openings
in such a way that the flue sheet was entirely protected. This
covering lasted about three months, after which it was necessary
to renew it. As it was a house-heating boiler, two renewals a season
served until the ^boiler plant was dismantled. The cure of the
trouble with the boiler having the extended water leg, as shown in
Fig. 2, is due in my opinion to the added heat-absorbing surface in the
deeper leg, as it operates to abstract a much larger quantity of heat
from the gases before they came in contact with the tube ends, than
did the boiler before alteration.
The Author. Replying to Mr. Bolton's remarks, I have heard of
the experiments which he quotes in regard to the rate of evaporation
of different portions of the length of a tube, but I am not at all
impressed with them as a guide. It is well known that the first
surface that receives heat gives the greatest rate of evaporation and
leaves less for the remaining surface to do. Attention to this to the
extent apparently advocated by Mr. Bolton would lead to an absurd
result, for one might go on indefinitely shortening tubes. It should
be remembered that only 16 feet of the 20-ft. length of tubes are
in contact with water, the remainder being for superheating.
2 Apparently Mr. Bolton believes that it is known how long tubes
should be. I do not think that this is kno\vn, for the reason that a
boiler must undergo a wide range of duty; a short tube would do
818
DISCUSSION
for light work and a long one would be necessary for heavy work.
Many^vertical^ boilers with 2i-in. tubes 20 ft. long ^have jbeen used
successfully for years and they are still being built. Mr. Bolton would
evidently prohibit increasing the size of a boiler by increasing the
length of tubes, and would recognize only an increase in diameter as
a means of increasing size. To my mind this is illogical and not con-
sistent with the teaching of successful practice.
One extra heavy
tube with safety
plue 13'0"above
lower tube plate
Fio. 1 Ceoss Section op Vertical Fibb-Tube Boileb
3 Mr, Bolton speaks of the small combustion chamber as the boiler
was first ^installed, but he ignores the hundreds, if not thousands, of
vertical boilers with less^combustion chamber space. I believe that I
am the only person who designs vertical boilers with^the crown sheet
as much as 8 ft. above the grate, and this I have been doing for
many years. In regard to the Dutch oven in front of these boilers,
it would have wholly defeated the object of using vertical boilers.
Besides it would have added undesirable brick^work. ]
4 Mr.^^Bolton easily accounts for the lack of economy of the boiler,
but ignores the perfection with which it absorbs heat. I believe the
LEAKY VERTICAIi FIRE-TUBE BOILERS
819
lack of economy to be wholly due to want of air, and when this is
supplied and properly distributed the economy will be satisfactory.
This would be equally true if the combustion chamber were much
longer. The locomotive boilers tested at 'the'St. Louis Exposition
by the Pennsylvania Railroad have very little combustion chamber
Fig. 2 Sectional Elevation of Furnace of the Author's Firb-tubb
Boiler
820
DISCUSSION
space, and the excellent economy is due to the proper admission and
distribution of air. In regard to the economy of the boilers under
discussion, it should be remembered that it was good, only not as good
as is sometimes the case.
5 Mr. Parker states that the tubes leaked for the reason that they
were set close to the grate and were therefore subjected to wide ranges
of temperature. This is true if we consider the closing of many of
the tubes by clinker and the consequent overheating of those that
were not closed.
6 I agree with Mr. Bement that some other kinds of stoker would
probably not have precipitated the clinker on the tube ends, and this
' I stated in the paper.
Fig. 3 Section of Furnace of the Boiler Shown in Fig. 2
7 Concerning the ability of the water to enter among the tubes,
there are many large vertical boilers, some nearly as large as the one
described, that have far less space for the passage of water among the
tubes, and no trouble results. I know of some that have only one
wide space across the crown sheet, while mine have four wide spaces
entirely across, or eight reaching to the center.
8 I observe that Mr. Bement considers that the cause of the cessa-
tion of the leakage of the tubes of my boiler was the added surface
of the water leg. I cannot feel that this is so. It is inconceivable
to me that the heat near the center of the furnace should be sensibly
reduced thereby. Moreover the absence of the clinker after the
change seems to me ample cause of the improvement, for, as I
have stated in the paper, a large proportion of the tubes were stopped
up, and those that were in service must have been overheated. I
LEAKY VERTICAL FIRE-TUBE BOILERS 821
think that if the boilers had been raised without adding to the water
leg the trouble would have ceased.
9 Whatever the cause of the leakage may have been, I finr) on
January 17, 1910, the date of writing, that the tubes are not leaking;
nor have those of one boiler leaked since August 31, 1908, nor those
of the other since February 25, 1909, each boiler having been worked
constantly to about 1000 boiler horsepower.
No. 1262
THE BEST FORM OF LONGITUDINAL JOINT
FOR BOILERS
By F. W. Dean, Boston, Mass.
Member of the Society
It has been generally accepted in this country for a number of
years, that the best form of butted longitudinal riveted joint for
boilers is that in which the inside strap is wider than the outside, and
which has one or more rows of rivets passing through the shell and the
inside strap beyond each edge of the outside strap. The pitch of
the first row of outer rivets is double that of the rows that pass
through both straps, and if there are other outer rows they may or
may not have a still greater pitch.
2 In England, where until comparatively recently boiler con-
struction has been superior to ours, this form of joint appears to
receive no recognition. It was first devised, as far as I know, by
Dr. E. D. Leavitt, Past-President of the Society, and Edward Kendall,
both of Cambridge, Mass., and was first used by Mr. Leavitt in some
locomotive type boilers designed by him for the Calumet & Hecla
Mining Company. I have a blueprint of this boiler dated 1879. It
is, of course, hazardous to state that this joint was never used before
and it is quite possible that it was used in England, and discarded and
forgotten as poor construction, as I believe it is. It was first used
on an American locomotive by the Baldwin Locomotive Works in
a consolidation locomotive built by them for the Calumet & Hecla
Mining Company, the drawing of this joint having been made by me
when I was in Mr. Leavitt's employ.
3 While every boilermaker has for years been familiar with butt
joints, this form made slow progress towards adoption in this country.
One form of joint used to avoid the butt joint and get something as
good, was a lap joint with an inside strap bent at the edge of the lap
and riveted on each side of it. This was used on locomotives exclu-
Presented at the Annual Meeting, New^York, (December 1909), of The
American Society of Mechanical Engineebs.
824 LONGITUDINAL JOINT FOR BOILERS
sively, and was of little or no value as it was simply a somewhat
elastic bent tie connecting the two parts of the shell plate. Finally,
and fortunately, this joint gave way to the butt joint first described.
4 I believe there has been no case of an explosion of a butt-joint
boiler; at least one due to rupture of the joint. Recently, however,
a boiler at Woonsocket, R. I., narrowly escaped explosion, a longi-
tudinal rupture of the plate on one side of the joint, and within its
limits, being discovered while the boiler was subjected to steam pres-
sure. The steam pressure was rapidly reduced and no explosion
occurred. An account of this is given in Power, January 26, 1909,
and the joint itself is in possession of the boiler-inspection depart-
ment of the Massachusetts district police at the state house in Boston.
5 It has been growing upon me for some years that a one-sided
boiler joint, such as that first described, is poor construction, and
may sooner or later cause a crack in the plate. The Woonsocket
phenomenon has tended to confirm this opinion. It is evident that
unless the outside rivets fill the holes they do very little good, and
when they do fill them they form an overhung connection and to
some extent possess in themselves the now recognized defect of the
lap joint. Moreover the extended inside plate forms a bent connec-
tion between the different rivets at different distances from the center
line of the joint.
6 In many cases designers have placed the outside rivets at a
considerable distance from the edge of the outside strap and this is
constantly overdone. It is obvious, on careful thought, that the
outside rivets should be as near the edge of the outside strap as practi-
cable, thereby diminishing the bent-tie effect. In order to diminish
this effect still further, and also to render the overhung rivets
more effective, the inside strap should be thicker than usual, and this
feature can hardly be overdone. The inside strap should be at least
as thick as the shell plate, and great care should be taken to have •*■'<'
the holes match and the rivets fill the holes.
7 When a joint of this kind is tested to destruction in a testing
machine, it will be found to fail somewhat in detail, the inside strap
bending slightly and the outside rivets being the last to rupture after
yielding a little. In a boiler the joint would beAveaker than a flat
specimen on account of the bent-tie feature. This could l^e pre-
vented if it were practicable to calk the inside strap, as it would
thereby be compelled to maintain the circular form. The theoretical
efficiency of this joint is greater than of any other kind, but in practice
I believe the efficiency is not i-ealized and the defects that I have
described render the joint, in my opinion, undesiral)le.
LONGITUDINAL JOINT FOR BOILERS
825
8 In order to avoid the defects of the one-sided butt joint, I have
adopted and intend to use hereafter, a joint with both straps of the
same width, as illustrated in Fig. 1. This has the merit of having all
rivets in double shear and the strains all taken care of in the best
manner. The efficiency of this joint can hardly be above 84 or 85
per cent while that of the one-sided joint can be theoretically 91 or
92 per cent; but the certainty that the efficiency of the former is
realized in practice is ample compensation for the use of slightly
thicker plates. The pitch of the outer rows of rivets is rather great,
compelling the use of a thick outside strap in order to stand calking
and remain steam-tight. I use an equally thick inside strap in order
to diminish the bent-tie effect. This effect is small, however, as the
o-
-e— -o-
Fig. 1 Recommended Form of Longitudinal Joint
rivets are all near the center of the joint. It can be eliminated by
calking the inside strap, which is practicable with this joint, and is
done in the best marine practice. This assumes that the calking is
effective and will remain so.
9 While this subject is under consideration, it is well to call atten-
tion to the perfection with which the longitudinal joints of boiler-
plate cylinders can be welded, a fact which has been demonstrated
for many years with corrugated furnaces and more recently with
soda digesters. While the joints of corrugated furnaces are in com-
pression those of digesters are in tension, and their proved safety
should be sufficient to overcome any timidity concerning the per-
fection and safety of welded joints. Circumferential joints are not
so easily welded as longitudinal, and it is of course of little importance
in boilers that they should l)e welded.
826 DISCUSSION
DISCUSSION
Reginald P. Bolton. The form of longitudional joint for boilers
which Mr. Dean has described as the best is as old as the time of Bru-
nei, and was tested by him in 1838, and again by Longridge in 1857.
It is a double- welt triple-riveted joint, omitting alternate rivets in the
outer strip, and it has the defect of undue distance for calking between
the outer rivets. It is not so good a joint as it would be when the
triple riveting is continued, instead of omitting the alternate outer
rivet. The other form of joint to which Mr. Dean refers, in which
the inside welt was wider than the outside welt, has stood the test of
many years usage, and I do not know of any case of failure.
2 In discussing the longitudinal joint, we should not lose sight of
the fact that the weak parts of every longitudinal joint are the ends,
where the two shell plates unite and the circular seams meet the longi-
tudinal joint. It is there that weakness develops in all joint construc-
tion. In explosion cases on which I have been engaged, I have found
that trouble has developed at those points, and have noted that rup-
tures commenced there. Therefore, in dealing with the design of
longitudinal joints, the essential feature Seems to me to be its character
where it meets the circumferential seam.
E. D. Meier. I think that the value of this joint depends largely
on the diameter of the boiler that one has in mind. In a Scotch
marine boiler, from 12 to 15 ft. in diameter, the joint would be an
excellent one, especially with the scalloped edges mentioned by Mr.
Dean^. That is a very troublesome thing to do, but in addition to the
advantage of the scalloped edge which Mr. Dean cited, there is the
further one, that it modifies the tendency, common to such joints, to
buckle at the point where the sheets come together. The butt joint
is stiffer there than any other part of the shell and with a change in the
pressure and temperature the buckling ultimately tends to impair the
joint.
2 With a small boiler, 36 in., 42 in., or 48 in. in diameter, the
joint is too large a proportion of the total circumference, and this
action would become worse. That buckling action is distributed by
making the butt plates as thin as possible, and making the inside one
longer than the outside one.
3 The welded joint will be an ideal one when we can be sure of
1 This was referred to by the author in presenting his paper.
Longitudinal joint for boilers 827
a weld that will give 95 per cent efficiency. The difficulty will be to
test it. We do know, however, that when we rivet a' joint and do it
honestly, we have something that can be relied on. Much will depend
on how the material is chosen and how the work of laying up and
riveting is done. The joint should be made by carefully bending the
butt straps at a red heat to the true curve, and rolling the plate itself
true to template. This will make as perfect a joint as possible. For
a large diameter of boiler, I think the joint advocated by Mr. Dean,
especially if the edges are scalloped, is an excellent one, but for^smaller
diameters I prefer the old joint. __
4 Two other points must be considered: first, how the calking
is done, as in many sheets the initial fracture is caused by bad calking;
second, what sort of metal was used, for unless the chemical analysis
of the plates as to minimum of injurious metalloids is firmly insisted
on, trouble is sure to follow even in the best proportioned joints.
Prof. A. M. Greene, Jr. Mr. Dean is probably aware that in
the 1893 report of the Chief of the Bureau of Steam Engineering of the
Navy, it is shown that|the boilers intended for the New York, the
Columbia and the Minneapolis, were all designed on the same plan as
that^which Mr. Dean recommends. The illustration in the paper is
almost exactly similar to those in the report. These boilers were all
installed and have given entire satisfaction.
2 Locomotive ejigineers, however, are using the unequal length
butt strap quite extensively. I know of locomotives in which two
rows of rivets were placed outside of the outer butt strap, and I do not
know of any failure of such joints. If it is a case of getting increased
efiiciency, and still having the outer butt strap arranged for a calking
distance, I do not see why we should depart from the method of
unequal straps to use the equal strap arrangement which cannot give
such high efficiencies.
William A. Jones. I wish to point out the tension which exists
in the outer row of rivets and its effect on the drum shell. This
should have an important part in determining whether the form of
joint which Mr. Dean recommends is really better than if the outer
butt strap were cut back one row of rivets on each side, so that the
rivets at their calking edges would be close together.'
2 We probably all agree that rivets are more reliable in shear than
they are in tension; that the more closely and firmly the edge of^the
outer butt strap is held down, the less calking will be required and
828 DISCUSSION
the less possibility there will be of injuring the shell plates by calk-
ing the butt strap in the shop, and the more remote will be the prob-
ability of subsequent leaks, prompting inexpert men to calk them
again later,
3 If we assume that the inner rivets are about 3 in. apart, then
the outer rivets shown in the joint which Mr. Dean recommends will
be about 6 in. apart, and each rivet will be holding an area of butt
strap of from 15 to 20 sq. in., which, at 200-lb. pressure, will require
from 3000 to 4000-lb. tension per rivet. In addition, each of these
rivets will be required to hold the calking for an edge about 6 in. long,
and the calking will have an advantage over the rivet of about 2 to 1,
due to the leverage which it has because the rivets are back from the
edge. It does not require much thought to see that these rivets
would be better able to do this work if they were twice as close together.
4 The joint which Mr. Dean has shown has five rivets in double
shear on each side, in a length equal to the pitch of the outer rivets, so
that ten times the area of one rivet is the total area in shear in this
length. If, on the other hand, the outer butt strap were cut back so
that the rivets at its edge would be close together and the outer rivets
were in single shear, then the total area in shear would be only one-
tenth less, and the proportion of the circular tension transmitted by
the rivets in single shear could not be more than 11 per cent of the
total in this case.
5 I understand that it is in an effort to improve the action of this
11 per cent of the force involved that this wide outer butt strap is
recommended, and that where four rows of rivets are used instead of
six, this proportion may rise to 20 per cent. In any case, the slight
bending in the shell plate is less, I believe, than the bending tendency
which the tension would produce in the rivets, due to pressure on the
wide outer butt strap.
6 Let us consider the forces acting upon a rectangular area of
plate in a drum shell under pressure. The circular tensions acting
tangentially at the edges of this area are equal in intensity, but act at
an angle to each other, so that each has a component normal to the
chord of the area considered. These normal components exactly
balance the pressure acting on that chord. When the area considered
embraces a half-circle, the normal components become equal to the
circular tension.
7 In the case of the outer butt strap, if all the circular tensions of
the drum could be transmitted to the outer butt strap by rivets at its
extreme edge, the shear of these rivets alone would hold the outer butt
LONGITUDINAL JOINT FOR BOILERS 829
strap to the drum, and the components of the shears normal to the
chord would just balance the steam pressure on that chord, so that no
tension in the rivets would be necessary, except for calking. Mov-
ing the rivets back from the edge of the butt strap makes the shear
act more nearly parallel to the chord, while it does not diminish the
chord, so that shear alone will no longer hold the butt strap in place,
and tension must be developed in the rivets to make up the difference.
8 Transmitting part of the circular tension through the inside butt
strap further increases the tension on the rivets, due to pressure, but
the additional tension in this case maintains the curve in the inner
butt strap by stitching it to the surface which receives the pressure,
and the reaction of the tension at the inner ends of these rivets is thus
provided for.
9 In the case of the outer rivets of the joint which Mr. Dean shows,
reaction of this tension at the inner ends of the rivets must be absorbed
by an abrupt change in direction of the circular tension at those
points, tending to produce corners in the drum shell in order to satisfy
the triangle of the three forces formed by the tension on the rivet, the
tangential tension to the right, and the tangential tension to the left.
If we assume a 42-in. drum, 200-lb. steam pressure, 6-in. pitch of outer
rivets, each of which takes in tension the pressure of 20 sq. in., we have
4000-lb. tension in each rivet due to steam pressure, the inner ends of
the rivets being anchored by an abrupt change in direction of about 9
deg. of 25, 200-lb. circular tension.
10 Evidently, this abrupt change of direction of the total circular
tension may readily distress the plate more in the form of joint which
Mr. Dean recommends than in the usual form of joint with the narrow
outer butt strap, even though a very small part of the circular tension
is transmitted through a rivet in single shear.
11 Mr. Dean's statement that he believes there has been no case
of failure of butt-strap joints, would indicate that there was nothing
wTong with the established form using the narrow outer butt strap.
Certainly the remedy proposed seems more objectionable than a rivet
in single shear.
Sherwood F. Jeter.' It seems that all engineers design joints
with reference to their weakest point, that is, provided the joint were
to be ruptured in a machine. Of all 1 ho explosions that to my knowl-
edge have been due to ruptures, none have occurred in the theo-
' The Bigelow Co., New Haven, Conn.
830 DISCUSSION
retically weakest part of the joint. Most explosions due to rupture
of the sheet have occurred near the joint and were apparently due to
flexure of the metal, which had destroyed its life at the particular
point of rupture.
2 I believe that there is a great need for an investigation as to
what causes the rupture of the plate, and for other than machine tests
of different kinds of joints. An account in Power states that there
have been four ruptures of butt-strap joints of a nature similar to what
was previously alluded to as a ''lap cracldng" of the joint. From the
great number of lap joints in successful use for twenty-five years or
more, it may be judged that something besides a mere lapping of the
plates causes such defects.
The Author. There is very little for me to say in closing, as
my views have been fully set forth in the paper. I am interested in
the history of this joint as stated by Mr. Bolton. I first knew of
it in 1889; it is shown in Thomas W. Traill's book on Boilers, and a
table of sizes of parts is there given.
2 Several of the speakers express doubt as to the tightness of
the joint on account of the wide spacing of the outer row of rivets.
There should be no doubt of this kind, for too many of them are in
use. I know of one joint with li-in. rivets [in 1-in. straps on a
pitch of 9 J in., and another with lA-in. rivets in a |-in. strap on a
pitch of 81 in.
O. 1263
TESTING SUCTION GAS PRODUCERS WITH A
KOERTING EJECTOR
C. M. Oakland, Urbana, III.
Member of the Society
A. P. Kratz,! Urbana, III.
Non-Member
The method of testing the suction gas producer herein described,
and the forms for computation given in the Appendix to the paper,
have been used by the writers to advantage in their gas-producer
tests in the mechanical engineering laboratory of the University of
Illinois. The method of testing has reduced the labor of running
such tests to a minimum, and the forms for computation have greatly
reduced the labor and tedium of the calculations.
2 The tests were made on an Otto suction gas producer rated at
60 h.p. and 8000 cu. ft. of gas per hour. The plant as originally in-
stalled consisted of the producer A (Fig. 1) , the wet scrubber B, the
gas receiver C, and a 22-h.p. engine. In order to facilitate the test-
ing of the plant the connection to the 22-h.p. engine was blanked
and a Schutte-Koerting steam ejector of 12,000-cu. ft. hourly capacity
was placed in the gas main at F. This ejector was used to draw the
gases from the producer and deliver them to the wet scrubber G, where
the steam used by the ejector was condensed.
3 The condensed steam and condensing water passed out at the
overflow M, while the gases passed out through the separator A^ and
into the dryer H, constructed from a gas bell, or holder, filled
with straw, and used to separate the suspended moisture from the
gases before they entered the meters / and /. The meters were of
8000 and 3500-cu. ft. hourly capacity respectively, and were connected
'Assistant, Mechaaical Engineering Laboratory, Univ. of 111.
Presented at the Annual Meeting, New York, (December 1909) of The
American Society op Mechanical Engineers.
832
TESTING SUCTION GA.S PRODUCERS
P
TESTING SUCTION GAS PRODUCERS
833
in parallel for capacities greater than 8000 cu. ft. per hour, the larger
meter alone being used for lower capacities. From the meters tho
gases were discharged into the atmosphere above the roof of the lab-
oratory.
4 A gage box L was adapted to receive thin plates with orifices,
and was used in calibrating the meters, by means of air. The meters
having been blanked from the gas main, compressed air was admitted
at /<", and expanding passed through the meter to be calibrated and
out at the orifice in L. The data for the orifice was taken from the
paper, on the Measurement of Air Flowing into the Atmosphere
through Circular Orifices in Thin Plates and under Small Differences
of Pressure, by R. J. Durleyi. After the calibration, the inlet to the
box was blanked.
o The producer is of the contained vaporizer type, with grate and
without charging bell, the specifications stating that it is only to be
used twelve hours at a time. During some of the earlier tests the
cast-iron vaponzer was cracked. A steam jet was then used to sup-
TABLE 1 TEMPERATURES IN FUEL BED
Time
Zone No.
Temp. 3 in.
From Near
Waix V
Temp, at
Cbnteb, F*
Temp. 3 in.
From Fab
Waix F»
10:05-10:10 a.m.
10:25-10:30
10:43-10:55
1
2
3
2100
2350
2037
2225
2200
2025
2275
2400
ply the moisture, and the vaporizer was blanked off. The weight
of steam was measured by passing the jet through a calibrated ori-
fice in a thin plate.
6 The test was started' with"the 'producer full and with a clean
fuel bed. The coal fired during the test was weighed and at the end
of the test the fire was cleaned, the fuel bed being brought as uear to
the starting condition as possible, and the producer filled. In order
that the error in determining the weight of coal fired in this manner
might be known, the producer when cold was filled a'number of times,
and the weight of coal required was noted. The average of these
weights was taken to be the true weight of coal required to fill the
producer, the probable error infilling with a given weight of coal being
estimated from these results. In running it was endeavored to make
1 Trans., Vol. 27, No. 1098.
834
TESTING SUCTION GAS PRODUCERS
the tests of such duration as to bring the probable error of filling down
to about two or three per cent.
7 The temperature of the gas leaving the producer was taken at
0 by means of a platinum-rhodium thermo-couple and a Siemens &
Halske milivoltmeter, calibrated to read direct in degrees centigrade.
The temperatures in the fuel bed were taken with Hoskins thermo-
couples and galvanometer, the latter reading in degrees Fahrenheit.
Other temperatures were taken with mercury thermometers.
8 The temperatures in the fuel bed were taken in three horizon-
tal zones 10 in., 18 in. and 24 in., respectively, above the grate. In
each zone readings were taken 3 in. from the lining on each side, and
in the center of the fuel bed. The results are given in Table 1.
:jjj^,j^^^/rj^^ji'i;^
'''" " '"^ '
Fig. 2 Sampling Tube fob Taking Samples of Gas Continuously
9 By means of the sampling tube illustrated in Fig. 2, samples
of gas were taken continuously for test by a Junkers calorimeter and for
analysis by Hempel apparatus. The results of the analyses are
given in Table 2.
10 As already stated, the weight of steam fed to the producer was
determined by the use of a calibrated orifice. By means of a small
laboratory aspirator, a sample of the gas leaving the producer was
drawn successively through a calcium chloride tube and a small
TESTING SUCTION GAS PRODUCERS
836
gas meter, the weight of moisture being determined by the calcium
chloride tube and the volume by the meter. The per cent of moisture
determined by this method was used merely as a check, the percent-
age used in the computations being obtained by calculating the
weight of water decomposed from the analysis of the gases and the
analysis of the fuel. The difference between this quantity and the
total weight of moisture carried into the producer, gives the weight
of the moisture in the gas leaving the producer.
11 The volume of gas generated by the producer, and measured
by the meters, was also checked by computing the volume of the gas
generated from the analyses of gas and coal. In the anthracite pro-
ducer, where the loss of carbon in soot and tar is small, probably not
over 1 per cent, this offers an excellent means of checking the gas
volume, and also of computing the weight of air used. The gas analy-
TABLE 2 GAS ANALYSIS BY VOLUME
Per
Per
Per
Per
Per
Per
No.
TiMK
Cent
Cent
Cent
Cent
Cent
Cent
b.t-c.
CO2
O2
CO
CBU
H2
N2
1
6:23- 9:05 a.m.
6.7
0.5
22.9
2.1
12.2
56.6
134
2
9:10-10:37
4.1
0.2
27.9
1.6
11.1
55.1
142
3
10:39-12:25 p.m.
3.3
0.1
28.4
1.5
9.5
57.2
137
4
12:30- 2:52
4.3
0.2
26.9
1.8
10.6
56.2
139
5
3:00- 4:35
3.6
0.1
28.6
1.8
9.0
56.9
139
6
4:40- 6:05
4.1
0.3
27.4
1.8
10.0
56.4
138
Average
4.20
0.23
27.01
1.77
10.40
56.40
138.1
sis, where continuous samples are taken by the form of sampling tube
illustrated, should be accurate within 1 per cent. The greatest error
is likely to be made in the sampling of the coal. With a fine coal,
such as pea or buckwheat, and a sample representing from 10 to 20
per cent of the total weight of coal fired, the error in sampHng should
not exceed 2 per cent. The maximum error in determining the gas
volume and the weight of air used should not exceed 5 per cent, if
the error in filling the producer is 2 per cent. The probable error is
therefore much less. In most of the tests, the volume of gas com-
puted from anal3^sis has checked within 5 per cent the volume deter-
mined by the meters. The meters are known to be accurate well with-
in 2 per cent.
12 In the testing of large producers of the bituminous type, it is
often difficult to measure the gas volume by any mechanical means.
836
TESTING SUCTION GAS PRODUCERS
s
t
.a
\
1
\
^
1
;^
v'l
[
\
.
^
'v
bo
S
Vl
[
/
\
^
>
•o
<
]
s,
1
,
\
S
E
?;
1
V
■'
e
[J
1
.^
Z'
^
0|/
'i
1
2
i
/
V
1*-
^S
rO
9
i\
1
i
|\
■S
\
■"
i:
^.
1 s
i
2
1
^^
^
n
1
M
■o
r^P
j
\-i-
fl
1
11
'Ck
'
u
\
r/
Jl
1
\f~
§
^,
M
S
\
"^v
^,,
V
\
|.
Oi
'^\
'
<
2
''
/
I
J
|\
-
-
^
: 1/
/
55
1
I
1
\^
/
>
oi
.b
A,
1
\
\
k
Q
P-_
A
^
W
/
\
3
,
s
\
a
\
i k^
1
''^X,
f
\
^
)
^
^
/
^^
/"
a
[ \
k 1
f>\
\
\
/
V
y
i
J
1
K_
tH^
r
A 1
'«,
nN.
\
,
,
f
[
\
-
u
■a
i
1
J
^
r'
\
1
V
1
\ /
\
\
s
\
I
i 1
/
£
)
\l
1
\,
s
\
/
fa
m
1
t
J
"N
pj
y
/
J
\
1
^
\
\
'
1
f
\
1
I
/
ll
\
\
^
y
'
H
^
V-
_-
-'
"
X,
V
'n
u
/
\
— •
—
^
y
/■I
\
%
^
'^
'/
\
Tjst
1
p
/
1*^
/
1
' in
1
&
y
'*A
/
I
_g 1
\
^
(
\
'*i^
\
1
J/
V "^
1
%
1
\
\
/
g
IN^
?f
^?\
V
Ti
•iA
lO
,
1
*:,
A,
£
M
1
^
/
/
\
i
I
\
V
U)
i
\
\ei
\ \
iS;
'
•X
■vj
\,
/
•y
\
/
■?
\
w
a
V.
K
A
i
-A,
/
1
u\
■
Hi
X
/
\
\
V
c8
q-
""'>•
^
I'
t.
-1
\j
A
'/
I
.
>
2^
\
■3 J
/*
\
y
a
/\
\
S
>
s
V
1
^
/
\
^
/
-
-
1
\
\
-
\
\
.2
1
Fl
3
\
1
c/
\
c»
\
13
£
'."
H/
,
\
\
?^
Oh
s
\
\
\
s
\
V
Q
_^
U
\
>,
/
/
\
V
\
1
\
-
o
§-
-1^
-SSSSSS-
Sg2§§ S?
3
o
Tf
^
o
T*
9*
ns'sajj nrea
jaiBjwIsni L
jaiTJAVsni
-
^^ .laanpojj SaiMJai
3.1
JS
asjail }B aanssajj
notions
1 1 ! 1 {
ra^I
i 1
1
1 J 1 j J
Mill
1 1
1 S ^ 1 1 2
S S 5 s ?? s
■qi 0001 --loji'AV JaqqiiJOS putJ-y -no OOOI-st;o
TESTING SUCTION GAS PRODUCERS 837
In such cases, if the carbon lost in the soot and tar is estimated from
a sample of the soot and tar, and this amount deducted from the total
weight of carbon in the coal, the volume may then be computed from
the analyses of the gas and coal, and may be relied upon within 5 per
cent, provided the sampling is accurate.
13 In order to facilitate computations, we have prepared three
separate forms, or rather two forms and a guide sheet. Form 1 is
used only for the presentation of the results of the tests. Form 2
contains all items used in the computations, while Form 3 is the guide
sheet containing all of the formulae and their derivation. The items
of Form 3 are arranged in the order of computation. In following
out this method, the average corrected quantities are taken from the
original data sheets and placed on Form 2. The computations arc
then made by following Form 3. After Form 2 is completed, the
results are transferred to Form 1.
14 Referring to Form 1, Item 46, it will be noted that the total
ash and refuse is much less than the weight of ash alone that would
be obtained by computing from the analysis. This is due to the diffi-
culty in cleaning the ash out of the fuel bed, and partly to the loss
of ash in the form of dust, which is carried over into the scrubber.
In this particular coal, which had very little tendency to clinker, the
ash was soft and fine so that it packed in and filled the interstices
between the live coals. A small amount of clinker was formed on the
sides.
15 This tendency of the ash to pack in the fuel bed, while it pre-
vents the accurate determination of the actual weight of ash, does
not, it is believed, materially affect the determination of the weight of
coal as fired, for the reason just given; that is, while the fuel bed may
contain as much carbon at the start as at the close, the bed is much
more compact due to the ash. The weight of ash and refuse is valu-
able principally for the determination of the unburned carbon lost
through the grate.
16 Item 66, dry coal per sq. ft. of grate area per hour, is high;
while the producer was operating only at about 4800 cu. ft. per hour
capacity, this was considerably above its actual capacity. If the
fuel had contained a fusible ash the results as shown on Form 1 and
the graphical log Fig. 3 would have been practically impossible.
17 The heat balance, Form 1, shows the unaccounted-for loss to
be 4.4 per cent. This includes radiation and conduction, which for
this test probably amounts to between 2 and 3 per cent. By refer-
ring to Form 2, Item 126, it wiU be seen that the volume of standard
838 TESTING SUCTION GAS PRODUCERiS
gas, computed from the analysis of the gas and the analysis of the coal ,
checks within about 2.3 per cent of the volume of standard gas as
given by the meters, Item 125.
18 The graphical log sheet (Fig. 3) illustrates the uniformity of
conditions that were maintained throughout the test.
19 Permission for running the producer tests was obtained through
Prof. L. P. Breckenridge, the results being presented through the
courtesy of Dean W. F. M. Goss, of the University of Illinois.
APPENDIX
FORM 1 RESULTS OF GAS PRODUCER TRIALS
1 Test number 25
2 Made by CM. Garland and A. P. Kratz
3 At University of Illinois
4 Kind of producer Otto
5 To determine Efficiency
6 Principal conditions governing trial Uniform load
7 ICind of fuel Scranton-Anthracite
8 Kind of grate Plain
9 Method of starting and stopping test Alternate
10 Type of producer Suction
11 Form of blower-ejector Schutte & Koerting
12 Date of trial 5-29-1909
13 Duration of trial 12 hr.
DIMENSIONS AND PROPORTIONS
14 Dimensions of grate, ft 1 . 25 by 1 . 33
15 Grate area, sq. ft 1 . fi63
16 Mean diameter of fuel bed, ft 1 . 545
17 Depth of fuel bed, ft 2.21
18}^'Area of f uel bed, sq. ft 1.877
19 Height of discharge pipe above grate, ft 2 . 875
20 Approximate width of air spaces in grate, in 0.5
21 Area of air space, sq. ft 0. 722
22 Proportion of air space to whole grate area, per cent 43 . 3
23 Area of discharge pipe, sq. ft 0 . 165
24 Water heating surface in vaporizer, sq. ft
25 Outside diameter of shell, ft 2.833
26 Length of shell from base to top of magazine, ft 7 . 125
27 Ratio of water heating siu^ace to grate area, — to 1
28 Ratio of minimum draft area to grate area, 1 to 48 . 8
AVERAGE PRESSURES
29 Draft in ashpit, inches, water 0.61
30 Suction at producer outlet, inches, water 2 . 04
31 Pressure at meters, inches, water 3 . 76
32 Corrected barometer reading 29 . 15
32. 1 Steam pressure, lb. per sq. in. gage 90 . 5
840 TESTING SUCTION GAS PRODUCERS
AVERAGE TEMPERATURES
33 Of fire room, deg. fahr 82 . 2
34 Of steam leaving vaporizer, cleg, fahr 212
35 Of feed water entering vaporizer, deg. fahr
36 Of overflow from vaporizer, deg. fahr
37 Of water entering scrubber, deg. fahr 57 . 8
38 Of water leaving scrubber, deg. fahr 103 . 6
39 Of gases leaving producer, deg. fahr 1108
40 Of gases leaving scrubber, deg. fahr 84 . 3
41 Of gases entering meter, deg. fahr 68 . 0
FUEL
42 Size and condition Pea-Clean
43 Weight of coal as fired, lb 798.5
44 Percentage of moisture in coal 2.75
45 Total weight of dry coal fired, lb 776 . 5
46 Total ash and refuse, lb 85.0
47 Quality of ash and refuse
48 Total combustible consumed, lb 614
49 Percentage of ash and refuse in dry coal 10.9
PROXIMATE ANALYSIS OF COAL
50 Fixed carbon 78.45
51 Volatile matter 5.99
52 Moisture 2.75
53 Ash 12.81
54 Sulphur, separately determined 1 . 10
ULTIMATE ANALYSIS OF DRY COAL
55 Carbon, C 79.84
56 Hydrogen, H2 2.67
57 Oxygen, Oj 2.37
68 Nitrogen, Nj 0.82
59 Sulphur.S 1.13
60 Ash 13.17
61 Moisture in sample of coal as received 2.75
ANALYSIS OF DRY ASH AND REFUSE
62 Carbon, per cent 38.80
63 Earthy matter, per cent 61 . 20
a SiO
./AlA
\FeA
c MgO
d CaO
TESTING SUCTION GAS PRODUCERS 841
FUEL PER HOUR
64 Drycoal fired per hr., lb 64.7
65 Combustible consumed per hr., lb 51 . 2
66 Dry coal per sq. ft. of grate area per hr., lb 38 . 8
67 Combustible per sq. ft. of grate area per hr., lb 30 . 7
68 Dry coal per sq.ft. of fuel bed per hr., lb 34.5
69 Combustible per sq. ft. of fuel bed per hr., lb 27 . 3
70 Rate of descent of dry coal through fuel bed, lb. per ft. per sq.
ft. perhr 15.6
71 Rate of descent of combustible through fuel bed, lb. per ft.
per sq. ft. per hr 12.4
CALORIFIC VALUE OF FUEL
72 Calorific value by oxygen calorimeter per lb. dry coal, B.t.u. . 13,040
73 Calorific value^by oxygen calorimeter per lb. of combustible
B.t.u 15,700
74 Calorific value by analysis per lb. dry coal, B.t.u 13,125
75 Calorific value by analysis per lb. of combustible, B.t.u 15,800
WATER
76 Total weight of water» fed to vaporizer, lb 267 . 8
77 Total weight of overflow from vaporizer, lb
78 Water* actually evaporated in vaporizer, lb 267 . 8
79 Total weight of water fed to producer, lb 341 . 5
a From vaporizer* 267 . 8
6 In air 51.7
c In coal 22.0
80 Total weight of water decomposed 218.2
81 Total weight of water in gas leaving producer, lb 123 . 3
82 Ratio of water decomposed to water supplied 0.639
83 Weight of water decomposed per lb. gas generated, lb 0 . 055S
84 Weight of water decomposed per lb. of dry coal fired, lb 0.281
85 Weight of water decomposed per lb. of combustible consumed,
lb 0 . 355
86 Weight of water decomposed per lb. of air supplied, lb 0 . 0702
87 Weight of water supplied per lb. of dry coal fired, lb 0 . 440
88 Weight of water supplied per lb. of combustible consumed, lb. 0 . 556
89 Weight of water supplied per lb. of dry air used, lb 0 . 1097
90 Total weight of scrubber water, lb 22,200
WATER PER HOUR
91 Water evaporated per hr. in vaporizer, lb
92 Water evaporated per hr. per sq. ft. of water heating surface
in vaporizer, lb
93 Weight of water decomposed per hr,, lb 18.2
> Steam fed to vaporizer.
842 TESTING SUCTION GAS PRODUCERS
94 Total weight of water fed to producer per hr., lb 28 . 5
95 Weight of scrubber water used per hj., lb 1850
QUANTITY OF AIR
96 Per cent of moisture in air, per cent of dry air 1 . 66
97 Total weight of dry air, lb 3112
98 Total weight of dry air per hr., lb 259.2
99 Weightof dry air used per lb. of dry coal fired, lb 4.01
100 Weightof dry air used per lb. of combustible consumed, lb. . 5.07
101 Weight of dry air used per lb. of dry gas generated, lb 0 . 796
GAS
1 02 Per cent moisture in gas leaving producer, per cent of dry gas 3.15
103 Per cent of soot and tar in gas leaving producer
104 Calorific value of standard gas from analysis (high value)
B.t.u. per cu; ft 138 . 1
105 Calorific value of standard gas from calorimeter (high value)
B.t.u. per cu. ft 137.3
106 Specific weight of standard gas, lb. per cu. ft 0 . 0680
107 Specific heat of dry gas leaving producer 0 . 3281
108 Carbon ratio C/H 14.07
109 Total volume standard gas, per cu. ft 57,500
110 Volume of standard gas per hr., per cu. ft 4,795
111 Volume of standard gas per lb. of dry coal 74 . 1
112 Volume of standard gas per lb. of combustible 93 . 7
113 Total weight of standard gas, lb 3912
114 Weight of standard gas per hr., lb 326
115 Weight of standard gas per lb. of dry coal fired, lb 5.03
116 Weight of standard gas per lb. of combustible consumed, lb. . 6.37
GAS ANALY.SIS BY VOLUME
117 Carbon dioxide, COj 4 . 20
118 Carbon monoxide, CO 27.01
119 Oxygen, Oj 0.23
120 Hydrogen, H, 10.40
121 Marshgas,CH, 1.77
122 Olefiant gas, CjH^
123 Sulphur dioxide, SOj
124 Hydrogen sulphide, HjS
125 Nitrogen, Nj, by difference 56. 40
EFFICIENCY
126 Grate efficiency, per cent 95 . 3
127 Hot gas eflBciency, based on high heating value, per cent 90 . 9
128 Cold gas efficiency, based on high heating value, per cent 78 . 3
TESTING SUCTION GAS PRODUCERS
843
EFFICIENCY BASED ON COMBUSTIBLE
128o Hot gas efficiency, based on high heating value
12S& Cold gas efficiency, based on high heating value
COST OF GASIFICATION
1 29 Cost of fuel per ton delivered in producer room
130 Cost per 1000 cu. ft. of standard gas, cents
131 Cu. ft. scrubber water per 1000 cu. ft. gas
95.4
82.2
6.18
POKING
132 Method of poking From top, slicing from bottom
133 Frequency of poking Three times during run
FIRING
134 Method of firing Hand
135 Average intervals between firing Twice during nm
136 Average amount of fuel charged each time, lb
HEAT BALANCE
Debit B.t.u.
a Total heat supplied per lb. dry coal 13,040
b Total heat supplied by air per lb. dry coal 19
c Total heat supplied by moisture in air
per lb. dry coal
d Total heat supplied by moisture in coal
per lb. dry coal
e Total heat supplied as sensible heat in
coal per lb. dry coal
/ Total heat supplied by water* in vaporizer
per lb. dry coal 385
250
Total 13,444
Credit B.t.u. Per Cent
a Heat contained as sensible heat in dry
gas 1725
b Heat contained in moisture 262
c Heat contained in dry gas (heat of
combustion) 10,240
d Heat in unbumed carbon 618
e Heat contained in ash and refuse as
sensible heat
/ Heat lost in overflow from vaporizer
g Heat lost in radiation and conduction . . . 599 4 . 4
12.8
2.0
76.2
4.6
Total 13,444
100.0
' Supplied in steam.
844 TESTING SUCTION GAS PRODUCERS
FORM 2 RESULTS OF GAS PRODUCER TRIALS
NO. OF TEST 25. DATE 5/29/09. TIME OF START 6.15 A.M.
TIME OF STOP 6.15 P.M. DURATION OF TRIAL 12 HR.
KIND OF FUEL SCRANTON-ANTHRACITE
DIMENSIONS AND PROPORTIONS
1 Dimensions of grate, ft 1 . 25 by 1 . 33
2 Grate area, sq. ft 1 . 663
3 Mean diameter of fuel bed, ft 1 . 545
4 Depth of fuel bed, ft 2.21
5 Area of fuel bed, sq. ft 1 . 877
6 Height of discharge pipe above grate, ft 2 . 875
7 Approximate width of air spaces in grate, in 0.5
8 Area of air space, sq. ft 0 . 722
9 Ratio of air space to whole grate area, per cent 4.33
10 Area of discharge pipe, sq. ft 0. 165
11 Water heating surface in vaporizer, sq. ft
12 Outside diameter of shell, ft 2.833
13 Length of shell from base to top of magazine, ft 7 . 125
14 Ratio of water heating surface to grate area — to 1
15 Ratio of minimum draft area to grate area 1 to 48.8
AVERAGE PRESSURES
16 Average barometer reading, inches Hg 29 . 258
17 Average corrected barometer reading, inches Hg 29.152
18 Draft in ash pit, inches water 0.61
19 Suction at producer outlet, inches water 2 . 04
20 Absolute pressure at producer outlet, inches Hg 29 . 00
21 Suction* at orifice, inches water 90 . 5
22 Absolute pressure' at orifice, inches Hg 104 . 8
23 Pressure at meters, inches water 3 . 76
24 Absolute pressure at meters, inches Hg 29.43
25 Vapor pressure at meters, inches Hg 0 . 685
26 Dry gas pressure at meters, inches Hg 28 . 75
27 Suction at meter for dryer, inches water 2 . 04
28 Absolute pressure at meter for dryer, inches Hg 29 . 00
AVERAGE TEMPERATURES
29 At barometer, deg. fahr 78 . 0
30 Of fire room, deg, fahr 82.2
31 Of fire room, deg. abs. fahr 542 . 2
32 Of steam, deg. fahr 212
33 Of feed water entering vaporizer, deg. fahr
34 Overflow from vaporizer, deg. fahr
35 Rise in vaporizer, deg. fahr
'Steam ressnrep
TESTING SUCTION GAS PRODUCERS 845
36 Of water entering scrubber, deg. fahr 57 . 8
37 Of water leaving scrubber, deg. fahr 103 . 6
38 Rise in scrubber, deg. fahr 45 . 8
39 Of gases leaving producer, deg. fahr 1108
40 Of gases leaving producer, deg. abs. fahr 1568
41 Of gases leaving first scrubber, deg. fahr 84 . 3
42 Of gases leaving first scrubber, deg. abs. fahr 544 . 3
43 Drop in temperature of gases in scrubber, deg. fahr 1023.7
44 Of gases entering meters, deg. fahr 68 . 0
45 Of gases entering meters, deg. abs. fahr 528
46 Of gas at meter at dryer, deg. fahr 80 . 0
47 Of gas at meter at dryer, deg. abs. fahr 540
FUEL
48 Size and condition Pea, Clean
49 Weight of coal as fired, lb 798.5
50 Percentage of moisture in coal 2 . 75
51 Total weight of dry coal fired, lb 776 . 5
52 Total ash and refuse, lb 85.0
53 Quality of ash and refuse
54 Total weight of combustible, lb 614
55 Percentage of ash and refuse in dry coal, per cent 10.9
PROXIMATE ANALYSIS OF COAL
56 Fixed carbon, percent 78.45
57 Volatile matter, per cent 5 . 99
58 Moisture, percent 2.75
59 Ash, percent 12.81
60 Sulphur, separately determined, per cent 1 . 10
ULTIMATE ANALYSIS OP DRY COAL
61 Carbon, C, per cent 79 . 84
62 dydrogen, Hj, per cent 2 . 67
63 Oxygen, Oj, per cent 2 . 37
64 Nitrogen, Nj, per cent • 0 . 82
65 Sulphur, S, per cent 1 . 13
66 Ash, percent 13.17
67 Moisture in sample of coal as received, per cent 2 . 75
ANALYSIS OF DRY ASH AND REFUSE
68 Carbon, per cent 38 . 80
69 Earthy matter, per cent 61 . 20
a SiO^
JAIA
\FeA
c MgO
d CaO
TESTING SUCTION GAS PRODUCERS
FUEL PER HOUR
70 Dry coal fired per hr., lb 64 . 7
71 Combustible consumed per hr., lb 51.2
72 Dry coal sq.ft. of grate area per hr., lb 38.8
73 Combustible per sq. ft. of grate area per hr., lb 30 . 7
74 Dry coal per sq. ft. of fuel bed per hr., lb 34 . 5
75 Combustiblepersq. ft. of fuelbedperhr.,lb 27.3
76 Rate of descent of dry coal through fuel bed, lb. per ft. per sq.
ft.perhr ' 15.6
77 Rate of descent of combustible through fuel bed, lb. per ft. per
sq. ft. per hr 12 . 4
CALORIFIC VALUE OF FUEL
78 Calorific value by oxygen calorimeter per lb. dry coal, B.t.u. . 13,040
79 Calorific value by oxygen calorimeter per lb. combustible,
B.t.u 15,700
80 Calorific value by analysis, per lb. dry coal, B.t.u 13,125
81 Calorific value by analysis, per lb. combustible, B.t.u 15,800
WATER
82 TotaP weight fed to vaporizer, lb 267 . 8
83 Total weight of overflow, lb
84 Water' actually evaporated in vaporizer, lb 267 . 8
85 Weight of water fed to producer,
a From vaporizer' 267 . 8
6 In air 51.7
c In coal 22.0
Total 341 .5
86 Total weight of water decomposed from analysis, lb 218 . 2
87 Totalweightof water decomposed as used in calculations, lb.. 218.2
88 Total weight of moisture in gas leaving producer, lb 123 . 3
89 Ratio of water decomposed to water supplied 0 . 639
90 Weight of water decomposed per lb. of gas generated, lb 0.0558
91 Weight of water decomposed per lb. of dry coal fired, lb 0 . 281
92 Weight of water decomposed per lb . of combustible consumed ,
lb 0.355
93 Weight of water decomposed per lb. of air supplied 0 . 0702
94 Weight of water suppUed per lb. of dry coal fired, lb 0.440
95 Weightof water supplied per lb. of combustible consumed, lb.. 0.556
96 Weight of water supplied per lb. of air used, lb 0 . 1097
97 Total weight of scrubber water, lb 22,200
98 Total weight of water absorbed by dryer, grams 15
^ Steam fed to vaporizer.
TESTING SUCTION GAS PRODUCERS 847
WATER PER HOUR
Water evaporated per hr. in vaporizer, lb
Water evaporated per hr. per sq. ft. of water heating surface
in vaporizer, lb
Weight of water decomposed per hr., lb 18 . 2
Total weight of water fed to producer per hr., lb 28 . 5
Weight of scrubber water used per hr. lb 1850
QUANTITY OF AIR
Relative humidity of air, per cent . 73
Per cent of moisture contained in air, per cent by weight jf
dry air 1 . 66
Total weight of dry air by analysis, lb 3112
Total weight of dry air by orifice, lb
Total weight of dry air as used in calculations, lb. 3112
Weight of dry air per hr. from total used in calculations 259 . 2
Weight of dry air used per lb. of dry coal fired, lb 4 . 01
Weight of dry air used per lb. of combustible consumed, lb. . . 5 . 07
Weight of dry air used per lb. of dry gas generated, lb 0 . 796
GAS
Volume of gas passing through meter at dryer, cu. ft 31 . 06
Volume of standard gas passing through meter at dryer, cu. ft. 28 . 0
Total weight of gas passing through dryer meter, lb 1.9
Percentage of moisture in gas leaving producer, from dryer,
per cent dry gas 1 .74
Percentage of moisture in gas leaving producer, from water fed
to producer, percent dry gas 3. 15
Percentage soot and tar in gas leaving producer
Calorific value per cu. ft. of standard gas from analysis B.t.u.
(high value) 138.1
Calorific value per cu. ft. of standard gas from calorimeter,
B.t.u. (high value) 137.3
Specific weight of standard gas, lb. per cu. f t 0 . 0680
Specific heat of dry gas leaving producer 0 . 3281
Carbon ratio C/H 14.07
Total volume of gas from meters, cu. ft 60,630
Total volume of standard gas, from meters, cu. ft 57,500
Total volume of standard gas, from analysis en. ft 56,200
Total volume as used in calculations, cu. ft 57,500
Volume of standard gas per hr. from total used in calculations. 4795
Volume of standard gas per lb. of dry coal from total used in
calculations, cu. f t 74 . 1
Volume of standard gas per lb. of combustible from total used
in calculations, cu. f t 93 . 7
Total weight of standard gas from total used in calculations,
lb 3912
848
TESTING SUCTION GAS PRODUCERS
132 Weight of standard gas per hr., lb
133 Weight of standard gas per lb. of dry coal, lb. . . .
134 Weight of standard gas per lb . of combustible, lb .
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
GAS ANALYSIS BY VOLUME
Carbon dioxide , COj
Carbon monoxide, CO
Oxygen, O2
Hydrogen, Hj
Marsh gas, CH4
Olefiant gas, C2H4
Sulphur dioxide, SO2
Hydrogen sulphide, HjS . .
Nitrogen, Nj by difference.
GAS ANALYSIS BY WEIGHT
Carbon dioxide, COj
Carbon monoxide, CO. . . .
Oxygen, O2
Hydrogen, Hj
Marsh'gas, CH4
Olefiant gas, C2H4
Sulphur dioxide, SO2
Hydrogen sulphide, HjS . . .
Nitrogen, Nj, by difference.
EFFICIENCY
153 Grate efficiency, per cent
154 Hot gas efficiency, based on high heating value, per cent .
155 Cold gas efficiency, based on high heating value, per cent .
EFFICIENCY BASED ON COMBUSTIBLE
155a Hot gas efficiency, based on high heating value, per cent. .
155 6 Cold gas efficiency, based on high heating value, per cent.
COST OF GASIFICATION
156 Cost of fuel per ton deUvered in producer room
157 Cost per 1000 cu. ft. of standard gas, cents
158 Cu. ft. scrubber water per 1000 cu. ft. standard gas
POKING
159 Method of poking From top, slicing from bottom
160 Frequency of poking Three times during test
FIRING
161 Method ot firing Hand
162 Average intervals between firings Twice during run
163 Average amount of fuel charged each time
326
5.03
6.37
4.20
27 . 01
0.23
10.40
1.77
56.40
7.16
29.25
0.29
0.81
1.12
61.37
95.3
90.9
78.3
95.4
82.2
6.18
250
TESTING SUCTION GAS PRODUCEKS 849
HEAT BALANCE
Debit B.t.u.
a Total heat supplied per lb. dry coal 13,040
b Total heat supplied by air per lb. dry coal 19
c Total heat supplied by moisture in air per lb dry coal.. . .
d Total heat suppUed by moisture in coal
e Total heat supplied as sensible heat in coal
/ Total' heat supplied in vaporizer water 385
Total 13,444
Feb
Credit B.t.u. Cent
a Heat contained as sensible heat in drj'- gas 1725 12.8
b Heat contained in moisture 262 2 . 0
c Heat contained in drj' gas (heat of combustion) 10,240 76 . 2
i Heat in unbumed carbon 618 4 . 6
e Heat contained as sensible heat in ash and refuse
/ Heat lost in overflow from vaporizer
g Radiation and conduction, by difference 599 4.4
Total 13,444 100.0
FORM 3 GUIDE SHEET CONTAINING ALL FORMULAE AND THEIR
DERIVATION
The item numbers refer to the items of Form 2, and are arranged in the order
of computation.
Item 4. " Depth of fuel bed" is to a certain extent arbitrary. In order that
the term may have a fixed and definite meaning we will define it as the
distance between the upper edge of the ash zone and that section of the
fuel bed from which the gases separate and leave the fuel. The upper
edge of the ash zone can ordinarily be readily determined by inspection.
Item 16. This reading is the average of the barometer readings for the test and
is not corrected.
Item 17. Item 16 corrected. Hie following formula may be used:
Let H = corrected barometer reading.
t = temperature, deg. fahr.
h = barometer reading corresponding to temperature t.
ThenH = h (1.00254 - 0.0000790
Item 17. = Item 16 (1.00254 - 0.000079 X Item 29)
Item 18. = Observed.
Item 19. = Observed.
[tern 20. = Item 17 - Item 19 X 0.0735
^tem 21. = Observed.
Item 22. = Item 17 - Item 21 X 0.0735
* Supplied in steam.
850 TESTING SUCTION GAS PRODUCERS
Item 23. — Observed.
Item 24. = Item 17 + Item 23 X 0.0735
Item 25. = Taken from steam tables using temperature in Item 44, i lb.
per sq. in. = 2.04 in. Hg.
Item 26 = Item 24 - Item 25
Item 27. = Observed.
Item 28. = Item 17 - Item 27,- X '^0.0735 j
Items 29 to 48. The observed temperatures should be corrected from the cali-
bration curves before being placed in Form 2. The absolute tempera-
ture = the observed temperature + 460 deg.
Item 39. This item is observed in deg. cent, and should be transferred into deg.
fahr.
9
Deg. fahr. = - deg. cent. -I- 32
5
Each observation must be transferred.
Item 50. Taken from Item 67.
Item 50
Item 51. Item 49 1
100
Item 52. Taken from ash sheet, correction being made for any moisture taken
up in the ashpit.
Item 54. In these tests the total weight of combustible consumed will be taken
as the total weight of dry coal fired.
The weight of ash computed from the analysis— the weight of nitrogen
— I X the weight of oxygen — the weight of carbon contained in the
ash and refuse =
Item 51 X Item 66 Item 51 X Item 64 | Item 51 X Item 63
Item 51 —
100 100 100
Item 52 X Item 68
I
Therefore,
Item 54 = Item 51
[-
100
Item 66 + Item 64 + f Item 63
Item 55. =
100
Item 52 X 100
Item 51
Item 52 X Item 68
100
Items 56 to 69. From chemist.
Item^ 69, a,b, c, and d. The ultimate analysis of the ash will be made only in
special cases to obtain data on the formation of clinker.
Item 51
Item 70. =
hours
Item 54
Itemll. =
hours
TESTING SUCTION GAS PRODUCERS 851
Item 72. =
Item 73. =
Item 74. =
Item 75. =
Item 70
Item 2
Item 71
Item 2
Item 70
Item 5
Item 71
Item 5
Item 76. "The rate of descent of dry coal through the fuel bed," or "the dry
coal per cu. ft. of fuel bed per hour," which is the same, offers a means of
comparing the rate of gasification in different producers that seems to be
better adapted for the purpose than the expressions taken from boiler
pratice, viz: "coal per sq. ft. of grate area," or "coal per sq.ft. of fuel bed,"
the latter having been used in producer practice.
Item 74
Item 76.
Item 77.
Item 4
Item 75
Item 4
Item 78. Taken from chemist's report.
Item 78 X Item 51 - Item 52 X Item 68 X 145.40
Item 79. =
Item 54
Item 80. = { Item 61 X 145.40 + Item 65 X 40.00 + [Item 62 - i of Item
63] X 620.00}
Item 80 X Item 51 - Item 52 X Item 68 X 145.40
Item 81. =
Item 54
Item 113. Total volume of gas passing through meter at dryer. Observed.
Item 114. Total volume of standard gas passing through meter at dryer,
neglecting the effect of moisture.
Let Pj = absolute pressure in inches Hg. nt dryer meter.
<, = absolute temperature, deg. fahr. at d-yer meter.
Vj = total volume of gas passing through meter.
P, V, and T, be the condition of standard gas.
P = 30 in. Hg.
T = 460 + 62 = 522
Then
t, T
■p,v,T p,v, X 522 17.4 p.v,
or F =
Pt, 30<. t,
from which the value of Item 114 follows.
Item 28 X Item 113
Item 114. = 17.4 -
Item 47
Item 118. Not considered in these tests.
852
TESTING SUCTION GAS PRODUCERS
Item 119. One cubic foot of standard gas, that is, gas at a temperature of
62 deg. fahr. or 522 deg. abs. and a pressure of 30 in. Hg., gives up on
combustion, when the products of combustion are brought back to this
temperature and the moisture is condensed, the following heat quantities:
Hj = 328 B.t.u. per cu. ft. of standard gas.
C2H4 = 1480 B.t.u. per cu. ft. of standard gas.
CO = 319 B.t.u. per cu. ft. of standard gas.
CH4 = 1010 B.t.u. per cu. ft. of standard gas.
Item 120. This quantity is the average of all the calorimeter determinations.
Each separate determination by the calorimeter must be computed and
the heating value obtained. The following formula may be used. The
calorimeter readings are taken in centigrade imits with the exception of the
meter readings and pressure.
Let <2 = [temperature of entering water, deg. cent.
ti = temperatm-e leaving water, deg. cent,
r = rise in temperature of water, deg. cent.
W = weight of water used during the intervals = 8 litres for all tests.
G, = cu. ft. of gas used from meter.
t^ = temperature of entering gas, deg. cent,
pg = pressure entering gas inches Hg. absolute, corrected for vapor I
pressure of water (see Item 25) .
H = heatingValue per cu. ft. of standard gas (62 deg. fahr. or 16.7 deg.
cent.'^and 30 in. Hg.)
t^ —■ temperature of standard gas = 62 deg. fahr. or 16.7 deg. cent.
p8 = pressure of standard gas = 30 in. Hg.
G^ = cu. ft. of standard gas.
fi Pg ^ GgPs
t t
g 3
^ t Xp
Where tg and t^ are in absolute deg. cent., f, — t^ = r
Total heat per cu. ft. standard gas in B.t.u. = H
Total heat absorbed by water = W X r
H
WXrX 3.968 = W XrX 3.968
V txp^ )
8X r X 3.968 X t X30
g
G.Xp^X (16.7 + 273)
t XrX 3.29
^
G. xp„
where 3.968 is the cooversion factor.
TRSTINO SUCTION GAS PRODUCERS 853
In this formula it is assumed that the exhaust products are brought back
to 62 deg. fahr. This is not strictly true but the error introduced is
negligible, when the error in the use of the apparatus is considered. There
is another error due to the exhaust products carrying out more or less
vapor of water than was brought in by the entering gas and air.
This error will also be small and may either be positive or negative
depending on conditions. The entering gas will in most cases come from
direct contact with water and will therefore be saturated. The air ordina-
rily will not be saturated. On combustion, moisture will be formed by the
union of the oxygen and hydrogen, there will be a contraction in volume of
the gases due to the combustion, and also a contraction or expansion due
to a change in temperature after combustion. In whichever direction the
change in the weight of moisture in the out-going gas from that brought in
by the entering gas may occur, this change may be considered very small;
for the contraction on combustion will be comparatively small, and this
contraction will partly offset the unsaturated condition of the air used for
combustion. Also the change in temperature of the out-going gas from
that of the entering gas will be small.
The heating values as given in Items 119 and 120 are the high values.
The values obtained from the analysis will be more accurate and will
be used in all computations.
Item 121. The specific weights of the following gases at 62 deg. and 30 in. Hg. are
COj = 0.11610 CH, - 0.04278
CO = 0.07362 CjH«= 0.07370
Oj =0.08418 802=0.16380
Hj = 0.00530 HjS = 0.08682
N^ = 0.07400
Item 121. = [Item 135 X 0.1161 + Item 136 X 0.07362 + Item 137
X 0.08418 + Item 138 X 0.00530 + Item 139 X 0.04278 + Item 140
X 0.0737 + Item 141 X 0.1638 + Item 142 X 0.08682 + Item 143
X 0.0740] Tio
Items 144 to 152, Calculation of the gas analysis by weight from the analysis
by volume. Assume that we have one cubic foot of gas at 62 deg. fahr-
and 30 in. Hg. of the following composition:
Analysis by Weight
)LUMETRIC
: Analysis
Specific Weights
Per Cent
CO, = a
per
cent
0.1161 = PTa
A
W^X a
W
CO = 6
0.07362 = IFb
B
TFb X 6
W
Oj = c
0.08418 = Wc
C
TFc X c
W
Hj =d
0.00530 = PFd
D
_ PFd X rf
854 TESTING SUCTION GAS PRODUCERS
Volumetric Analysis Specific Weights
CH, - e 0.04278 = W^
CjH, =/ 0.07370 = W{
SO^ = g 0.16380 = Wg
H^S = k 0.08682 = TFh
Nj - t 0.07400 = Wi
Analysis by Weight
Per Cent
E
TFe X e
W
F
WiXf
w
G
WgX g
W
H
Wi^X h
w
I
Wi X i
w
Where W = [a X W^ + b X Wi, +c X W^ + e X W e- ■ • ■ +i X W{\ j^
= Item 121.
Item 122. The specific heats of the gases vary according to the pressure and
temperature. As the pressure used throughout the experiments is atmos-
pheric we have only to consider the variation with the temperature.
The following formulae taken fromZeuner, vol. 1, page 147, give the specific
heat for constant volume C^.
«COj, mCv = 6.50 + 0.00774« (1)
HjO, toCv = 5.78 + 0.00572« (2)
02HjN2,CO,toCv = 4.76 +0.00244t (3)
mCp - wCv = 1.9934 (4)
For the specific heat of marsh gas CH^, our other constituent, we will
use the value Cp = 0.6. This is approximate, but as the quantity of CH^
is small the resultant error is consequently small.
In the above formula, m is the molecular weight of the gas, t the tem-
perature in deg. cent., and Cy the mean specific heat between zero and t deg.
cent. Cp is determined from formula (4) . From the above formulae, the
analysis by weight as determined below and the temperature of the gases
leaving the producer, the specific heat of each constituent in a unit weight
of the gas may be determined. The specific heat of the gas will be the sum
of the specific heats of the constituents.
Substituting the value of mCv from formula (4), and the value of m, and
changing to deg. fahr. we have from the above formulae:
ForCOa, Cp =0.19 + 0.0000977< a
HA Cp = 0.426 + 0.000176^ b
Hj, Cp = 3.355 + 0.000678« c
CO, Cp = 0.24 + 0.0000484« d
N2, Cp-0.24 -h 0.00004 84< e
CH„ Cp - 0.6 /
O2, Cp = 0.21 + 0.0000424« g
iMallard and Le Chatellier's formulae.
TESTING SUCTION GAS PRODUCERS 855
Let. a, b, c, d, e, and/, represent the mean Cp for the above gases between 32 deg.
and t deg. fahr. Then the Cp of the producer gas = the sum of the products of
the constituents of the gas by weight X the specific heat of the constituent.
That is,
Item 122 = [a X Item 144 + c X Item 147 + d X Item 145 + e X Item 152
1
+ / X Item 148 + 3 X Item 146]
Item 123
12 3 4
00^ = 02 + 0 CO = C + O CH, = C + 2H2 C2H, = 2C + 2H,
44 = 32 + 12 28 = 12 + 16 16 = 12 + 4 28 = 24 + 4
The total weight of carbon appearing in a unit weight of gas from the above =
3 3
per cent by weight COj X + per cent by weight CO X — + per cent by
weight CH4 X + per cent by weight CoH. X —
400 "^ * 700
The total weight of Hg appearing in a unit weight of gas = per cent by weight
TJ 1 1
, „^ + per cent by weight CH^ X 1- per cent by weight C2H4 X ' —
100 400 700
or Item 123 = [Item 144 X 0.273 + Item 145 X 0.429 + Item 148 X 0.75 +
Item' 149 X 0.858] -4- [Item 147 + Item 148 X 0.25 + Item 149 X 0.143]
Item 124. Observed.
Item 125. Let G = total volume of gas as measured by the meters.
p = absolute pressure of this gas in inches Hg. as observed.
T = absolute temperature in deg. fahr.
t = observed temperature.
The volume of gas G as measured by the meter is saturated with water
vapor at the temperature t.
Let Pi = pressure of this vapor in inches as obtained from the steam
table.
Then as the pressiu-e p is the total pressure of the mixture, the actual
or partial pressure of the dry gas is p — p, = pj.
Let ps, Gg, and Tg. be the condition of standard gas. Then
Gg X Pa G X P2 _ GXp^XTs _ G X P2 X 522 _ GXPa
Ts T T X Pa T X 30 T
Therefore Item 125 equals
Item 124 X Item 26 X 17.4
Item 45
856 TESTING SUCTION GAS PRODUCERS
Item 126. Calculation of the volume of the gas from the analysis of the gas and
the analysis of the coal. Evidently the total weight of the carbon appear-
ing in the]°gas should be equal to the total weight of carbon in the coal
minus the weight that is lost through the grate and the weight lost in soot
and tar. This latter is small for the hard-coal producer and will be ne-
glected.
Let P = per cent carbon by weight in dry coal.
W = total weight of dry coal.
Wi = total weight of ash and refuse.
P, = per cent by weight of carbon in the ash and refuse.
W,= total weight of carbon that should appear in the gas, or the
weight of carbon utilized in the producer.
_ PW - P,TF,
100
This carbon is contained in the COj, CO, CH4, and C2H4.
The proportion by weight of C in CO^ is 3/11, of C and CO is 3/7, of C in CH,
is 3/4 and of C in C^H^ is 6/7.
Therefore the total weight of C contained in a unit weight of gas will be
^ 3/11 A+3/4:E + 3/7F + 6/7 G
' 100
Where A, E, F, and G are the per cent by weight of COg, CH^, CO, and CjH^ from
the gas analysis.
The per cent of this carbon contained in the gas as CO2 is
The actual weight of this carbon will be ~^ X TF,. Since W, is the
TFg X 100 *
total weight of carbon utilized from the fuel.
One pound of carbon on bm-ning produces 3f lb. of COj.
F"j X~^ X 3J = total weight of COj in the gas.
Tr, X 100
Let Wg = the specific weight of CO2 at 62 deg. and 30 in. Hg. See Item
121. The standard volume Fg of COj will therefore be,
^^^— = 7«
100 XW^XWs
Let this volume equal a per cent (from the volumetric gas analysis) of the
total volume of gas delivered by the producer. The total volume of standard gas
from the gas analysis is therefore
lOOJ^^^,^
a
y, ... ^ X TF3
^ a X TF3 X TFs
1
TESTING SUCTION GAS PRODUCERS 857
Item 126 therefore equals
Item 144 X (Item 51 X Item 61 - Item 52 X Item 68)
0.116 X Item 135~X (bT273 Item 144 + 0.75 Item 148 + 0.429 Item 145
+ 0.858 Item 149)
Item 127. Item 126 should be used as a check on Item 125. The difference
between the two values should not exceed 5 per cent. Item 125 should be used
in all computations.
Item 127
/tern 128.=
hours
,„^ Item 127
Item 129. =
Item 51
^ ,^^ Item 127
Item 130. = ~
Item 54
Item 131.= Item 127 X Item 121
Item 131
Item 132. =
hours
Item 131
Item 133.= ^
Item 51
Item 131
/ie7ral34.=
Item 54
Items 135 to 143. From chemist.
Item 104. The relative humidity, or per cent saturation is observed by means of
a hair hygrometer. This may also be obtained from a wet and dry bulb
thermometer, and a set of psychrometric tables.
Item 105. See Kent, page 484, for weights of air and moisture.
Let p = per cent saturation, or relative humidity. Item 104.
n = weight of moisture contained in one cu. ft. of saturated air at
the observed temperature, Item 29.
pn ,
— = weight of moisture in 1 cu. ft. of air as used.
100
If m = weight of 1 cu. ft. dry air at the observed temperature, then
pn pn ^ n
Item 105 = - - X 100 = ^^ = Item 104 X
100m m m
This formula is in error due to neglecting the vapor pressure of water; this is
however, negligible in the present case.
Item 82. Observed.
Item 83. Observed.
Item 84. = Item 82 - Item 83.
Item 86. The weight of water decomposed in the producer is evidently 9
times the weight of hydrogen formed, since 1 lb. of water on decompostion,
yields 1 lb. of hydrogen and 8 lb. oxygen. The total weight of hydronge
formed is 'equal to the totaweight of free hydrogen appearing in the gas,
plus the total weight of hydrogen appearing in the CH4 in the gas minus
858 TESTING SUCTION GAS PRODUCERS
the total weight of hydrogen that is not in combination with oxygen in the
coal.
Therefore, Item 86.=
/Item 131 (Item 147 + 0.25 Item 148) - Item 51 (Item 62 - i Item 63)
\ 100
Item 87. Owing to the difficulty in obtaining the weight of moisture in the gases
leaving the producer with a proper degree of accuracy by the use of a dryer,
j^it will ordinarily be better to use Item 86 for this item.
Item 106. Obtained from the gas analysis by weight, Items 144 to 152 inclusive.
Let A =
per
cent COj
Leti>
= per cent Hj
B =
per
cent Oj.
E
= per cent CH^
C =
-N.
F
= per'cent CO
(1)
(2)
(3)
0 +O2 = CO2
C + 0 =
= C0
H, + 0 = HjO
12 + 32 = 44
12 + 16
= 28
2 + 16 = 18
3 8 ,
3 4
7 ^7 =
1
9 9
Fom equation (1), one lb. of CO2 requires 8/11 lb. of O for its formation
From (2) one lb. CO requires 4/7 lb. of O for its formation.
The total amount of O appearing in 1 lb. of the gas is therefore
8 4 \ 1
~rA + ^F + B]X -—
II 7 I 100
This O comes from that contained in the air, that contained in the coal,
and from the water decomposed. The oxygen contained in the coal, how-
ever, is supposed to be united with hydrogen, and is therefore contained
in moisture which has been allowed for in the water decomposed.
Let W = total weight of gas.
Then the total weight of O used is
--- ( ~ A + ~F + B
100 yii 7
Let W^ = weight of water decomposed. From (3), 1 lb. of water de-
composed liberates 8/9 lb. of O.
Weight of O supphed by decomposition of water = 8/9 W^
Let W3 = total weight of O supplied by the air.
From the above equation"we have,' | ^^ '
8 4 \ TF 8
A + -F + B] — = - W2 + W^
11 7 y 100 9 '
W / 8 4 \ 8
or 1^', = A + ~ F+ ] - W, (4)
100 111 7 B ^
I
TESTING SUCTION GAS PRODUCERS 859
w
The weight of air used is therefore , since the proportion by weight of
(J*4UO
O in air is 23, or
1 r PF /8 4
- - W,
0.23 ■ • 0.231 100 V 11 " 7 "" / 9
(5)
fltem 131 / 8 4
Therefore Item 106 =» - Item 144 + - Item 145 + Item 146
L 100 \11 7
J 0.2
X Item 87 (6)
9 J 0.23
The above computation may be made from the weight of nitrogen appearing
in the gas. The nitrogen comes from the air used and from the nitrogen intro-
duced with the fuel.
Let C per cent per lb. = weight of N2 from analysis
Let W as before = total weight of gas
CW
Then = total weight of Nj in the gas.
W H
The weight of Ng supplied by fuel will be — ^ — ' ,where Wi equals the total
weight of dry coal and H, is the per cent by weight of N^ contained in the
coal. We have therefore,
CW W,H, ^
100 100
where W^ = total weight of N^ in the air.
The weight of air supplied is therefore
0.77 \100 100 / 0.77 " \ 77
or Item 106 = (Item 131 X Item 152 - Item 51 X Item 64) — (7)
The weight of air derived by formula (6) will be liable to error, due princi-
pally to the error in the determination of the total quantity of water decom-
posed,which may be large, and also to the neglecting of the SO2 formed.
The weight determined by formula (7) will be in error, due principally to
the taking of the weight of N2 from the analysis by difference.
The results obtained from formulae (6) and (7) should check within 5 per
cent.
The results obtained by (7) are beUeved to be more accurate and will be
used in all computations.
Item 107. This may be obtained direct from the calibration curve of the orifice.
It should be compared with the two values obtained above.
860
TESTING SUCTION GAS PRODUCERS
Item 108. This will ordinarily be taken from Item 106
J, ,„„ Item 108
Item 109.=
Hours
,, ,_^ Item 108
Item 110.=
Item 51
Item 111.=
Item 54
-, „„ Item 108
Itemll2.=
Item 131
Item 85.= Item 84 + Item 856 4- Item 85c
ta856=^*""iO«XJ*«™l««^
100
/tern 85c.= ^*^"^^^^ 1*^^50 '
100
Item 88. = Item 85 - Item 87
Item 87
Item 89.=
Item 85
Item 87
Item 90. =
Item 131
Item 91.= ^''''^L
Item 51
Item 87
Item 92.= — -
Item 54
Item 87
Item 93.= —
Item 108
Item 94.= ^'^^
Item 51
,, „^ Item 85
Hem 95.=
Item 54
„ ^^ Item 85
Item 96.=
Item 108
Item 97. = Observed
Item 98. = Observed
., -^ Item 84
Item 99. =
Hours
,, ,^^ Item 99
Item 100.=
Item 11
,, ,., Item 87
/iero 101.=
Hours
T. ,«^ Item 85
Item 102.=
Hours
,, ,-^ Item 97
/<em 103.=
Hours
Item 115.= Item 114 X Item 121
Item 116. =
Item 117. =
TESTING SUCTION GAS PRODUCERS 861
Item 98 X 0.2205
Item 115
100 Item 88
Iteni 131
Item 153. The grate efficiency is 100 times the ratio of the total B.t.u. in the
fuel minus the B.t.u. in the fuel lost through the grate; to the total B.t.u. con-
tained in the fuel. Therefore
Item 51 X Item 78 X 100 - Item 52 X Item 6S X 14,540
Item 153. = —
Item 51 X Item 78
Item 154. The hot gas efficiency is 100 times the ratio of the total heat of com-
bustion of the gas, plus the sensible heat of the dry gas, plus the total
heat contained in the moisture, minus the heat given to the producer by
the entering air, by the coal as sensible heat and by the moisture or steam
in the air, or suppHed from any outside source; to the heat of combustion
of the dry coal. Therefore, Item 154 = {item 119 X Item 127 + Item
122 X Item 131 (Item 39 - 62 deg.) + Item 88 [1116 + 0.6 (Item 39-
212)] - Heat from external source} X 100 -^ Item 51 X Item 78.
The heat given to the producer by the air, moisture, coal, etc., maybe
neglected if the room temperature is within 20 deg. of the standard tem-
perature 62 deg. This will ordinarily be the case. If steam is supphed
to the producer by a steam nozzle taking steam from some outside source,
the heat in this steam must be subtracted from the numerator of the above
formula.
Item 155. The cold gas efficiency is 100 times the ratio between the total heat
of combustion of the gases and the total heat of combustion of the dry coal.
That is.
Item 119 X Item 127 ,
Item 155 = ^— — ^^ „„ 100
Item 51 X Item 78
Item 157. =
Item 158.
Item 156 X Item 49
0.02 X Item 127
Item 97 X 1000 Item 97
62.5 X Item 127 0.0625 X Item 127
HEAT BALANCE
DEBIT
Item a. Obtained from Item 78.
Items h, c, d, e, f. Using as a standard the temperature of 62 deg. fahr., the
heat given to the producer by the items 2 to 6 inclusive is in most cases
negligible. The error at a temperature of 100 deg. fahr. is less than 1 per
cent for a producer of the contained vaporizer type. However, the for-
mulae will be given for computation, of these items.
Item b. = Item 110 X 0.24 (Item 30 - 62°F)
862 DISCUSSION
Item 85b X (H - 1070) , „ .i, . . , u . • i .k , .a
Item c = where H = the total heat in 1 lb. saturated
Item 51
steam at the temperature of the fire room.
Item 49 X Item 50 ,^ „„ ^^ , , , ,
Item d = (Item 30 - 62 deg. fahr.)
100 X Item 51
Item e = 0.24 X (Item 30 - 62 deg. fahr.)
Item 82 (Item 33 - 62 deg. fahr.)
Itemf =
Item 51
CREDIT.
Item a = Item 122 X Item 133 X (Item 39 - 62 deg. fahr.)
Item b - Itemni_>ii!«"ii33 ^^^^^^ ^^ _ ^^^ ^^^ ^^^^^ ^ ^^ ^ j^^^^.
Item c = Item 119 X Item 129
Item 52 X Item 68
Item d X 145.40
Item 51
Item e This is very small and may be neglected.
Item f = -^ (Item 34 - 62 deg. fahr.)
Item 51
Item g = Sum of Items on debit side — (Item a + Item b + Item c + Item d
+ Items e[and /.)
DISCUSSION
Prof. R. H. Fernald. In connection with the Government
investigations, the feeling has prevailed ever since the beginning of
the Work in 1904, that gas producers could be tested on practically
the same basis as steam boilers, i.e., without necessarily operating
an engine in connection with the test. This would mean discharging
the gas into the air in a manner similar to the discharge of steam in
boiler test practice. This method of procedure has not been adopted
at the Government testing station because so much prejudice has
existed against the gas producer and gas engine. It has therefore
been necessary that the gas generated at the testing station be utilized
in an engine in order to avoid any discussion relating to the uncertainty
of such operation. This has been particularly necessary owing to
the large variety of fuels that have been handled and the variation
in the quality of gas produced. It is true, however, that from the
TESTING SUCTION GAS PRODUCERS 863
producer standpoint alone the engine is not essential, and the method
suggested by Mr. Garland is ingenious and reasonably convenient.
2 There are a few points in connectionwith this paper upon which
further information is desirable. In Par. 5 it is stated that the weight
of steam was measured by passing the jet through a calibrated orifice
in a thin plate. Methods of determining the quantity of steam used by
gas producers seem to be varied and the results obtained somewhat
uncertain. I believe it would be interesting to know the details of
the method employed by Mr. Garland. In the testing station at
St. Louis the steam used by the pressure producer was determined by
means of a calibrated orifice, but the fluctuations in pressure were such
that the readings obtained were not regarded as absolutely reliable.
During tests covering a period of approximately two years the steam
used varied from 0.28 lb. per lb. of coal fired to 1.13 lb. of steam per
lb. of coal fired. The average for twenty consecutive tests showed
0.69 lb. of steam per lb. of coal fired. It should be borne in mind that
the fuels used for the different tests were quite different in composi-
tion and that the amount of steam required by the different fuels
may have varied considerably; but in spite of this fact the feeling
which prevailed about the plant was that the method of determining
steam by means of calibrated orifices was not entirely satisfactory
unless the pressure of the steam passing into the producer, and the
percentage of moisture in the steam, could be kept constant during
the test.
3 At the Norfolk station, however, the steam required by the
producer was supplied by an auxiliary boiler, so that all water passing
into this boiler could be positively measured. Although the coals
used for the six tests reported below were practically the same in
composition, yet the records show the steam consumption per pound
of coal fired to be decidedly variable, as follows:
(1) 1 . 12 lb. per lb. of coal fired (4) 0.821b. per lb. of coal fired
(2) 1.14 "
« t( «
«
«
(5)0.77 "
i< « «
(3) 1.04 "
« « «
«
((
(6)0.69 "
tt « «
This wide variation shown for these six tests is due entirely to the
methods of operation, and not to uncertainties in measurement, as
might at first be inferred. There is need of systematic and careful
investigations relating to this question of steam per pound of fuel.
At the Pittsburg station the method of determining the amount of
steam used in the vaporizer is by means of a water tank calibrated in
pounds, thus insuring accurate measurement.
864 DISCUSSION
4 In Par. 6 is presented the general method of determining the
amount of fuel used. One phrase attracts especial attention: "at the
end of the test the fuel bed being brought to as near the starting condi-
tion as possible. " In boiler practice where the quantity of fuel on the
grate ^at any one time is relatively small, it is undoubtedly possible,
within a reasonable percentage of error, to determine the condition
of the fuel bed and to make this condition practically the same at the
beginning and close of an eight or ten-hour test.
5 However, the situation is totally different in gas-producer
practice in which the initial fuel supply and the amount of fuel on
the grate at any given time is large compared with the amount
required by the plant during a run of a few hours only. Even though
the conditions at the close of a producer test be made to duplicate
those at the beginning, there is still considerable difficulty in deter-
mining the exact fuel consumption, owing to the lack of accuracy in
determining the true thickness of the fuel bed. In a producer of
250 h.p. rating it is not uncommon to^make an error of from four to
six inches in the true depth of the fuel bed. In a producer of this
size, this will cause an error of about 800 lb. of coal, or about 400 lb.
of coke, according to the condition of the fuel bed at the time. It is
imperative, therefore, that the tests of producer plants be continued
to such length that these errors in measurement will be but a small
percentage of the total fuel consumed.
6 Mr. Garland states that an endeavor was made to make the tests
of such duration as to bring the probable error of filling down to
about two or three per cent. It will be of interest to have explained
in further detail the method of procedure used in .determining the
exact amount of fuel consumed. With a 250-h.p. plant in which the
fuel consumption for a period of 8 hr. amounts to only 1800 lb.
approximately, the error due to inaccurate measurement of the depth
of bed and variations in fuel bed thickness may be as great as 1150
lb. The percentage of possible error in calculating fuel consumption
for short periods is obviously great. With a period of 24 hr. and a
fuel consumption of about 5400 lb., the percentage of possible error,
although much less, is still over 20 per cent.
7 In the producer tested the effective fuel bed volume was
approximately 4 cu. ft., which is equivalent roughly to 250 lb. of an-
thracite pea coal. It is probable that a large percentage of the gas
value of this coal may be given off without materially decreasing the
fuel volume, under certain conditions of fuel bed. In a run in which
the fuel consumption for this producer amounts to only 8001b. with an
TESTING SUCTION GAS PRODUCERS 865
initial bed of 250 lb., it is a question whether the percentage of error
in fuel bed estimates may not amount to 10 or 12 per cent instead
of 2 or 3 per cent.
8 In a recent paper ou this subject published by the United States
Geological Survey, the following conclusions were presented :
a Throughout a test the fuel bed should be maintained in
uniform condition, with regard both to character of the fire
Jand thickness of the bed.
6 Failing in this, special care should be exercised to ^see^that
the fuel bed is in the same condition and of the same
thickness at the close of the complete test or at the end
of a test period, as at the beginning.
c A test should never be started when the producer has been
standing idle for some time with banked fires, as the fuel
bed will not be in the average condition under which it
will be required to work during the test.
d If, as the appointed hour for closing the test approaches,
the fuel bed is not in the proper condition, the time of
ending the test should be postponed until the bed natur-
ally assumes the proper thickness and character. No
yi^forcing of conditions should be allowed simply to bring
the test to an end at a previously determined hour.
9 In Par. 12 it is suggested that the volume of gas may be com-
puted from the analyses of the gas and coal and the statement is
made that this "may be relied upon within 5 per cent, provided the
sampling is accurate." This last clause "providing the sampling is
accurate" seems to contain the essential point. This is an impor-
tant subject and too much emphasis cannot be placed upon the
fact that proper sampling is difficult to accomplish.
10 Reference to the packing of the ash in the fuel bed^suggests
another point which must be very carefully considered in making
fuel bed measurements, viz., the swelling of any coals due to the
appUcation of heat. Frequently in the government tests, the measure-
ments of the fuel bed have caused very misleading impressions due
to the fact that the fuel had swollen materially during the operation
of the plant.
11 In Form 1 a number of items appear under a heading Quan-
tity of Air. Although it is quite possible to determine small quanti-
ties of air with some degree of reliability, yet methods for making
such measurements of large quantities appear to be entirely lacking.
866 DISCUSSION
Furthei decaiis of the methods pursued in this test will, I believe,
prove of interest.
12 In items 128 and 1286, are presented the producer efficiencies
based on dry coal and combustible. It is not apparent why there
should be a difference of 4 per cent in the efl&ciencies shown.
G. M. S. Tait. The usual method of testing a plant for such
a short period would be to operate the producer^ or^two or three days
beforehand so as to bring the fuel bed to an average working condi-
tion, that is, with an average amount of carbon in proportion to ash.
Then a comparativel}^ short run, provided great care was taken as to
the fuel depth, would give fairly reliable figures. Otherwise, when
drawing on a fresh fire and making a run of only twelve hours, it
would be necessary to pull the entire fuel bed at the end of the run
and analyze the contents for carbon and ash, in order that any sort
of accuracy might be obtained.
2 In one of the author's tests, instead of 34 lb. of coal per sq. ft.
of grate area, 8 to 10 lb. would be a normal figure, as 34 lb. of coal per
sq. ft. is entirely impracticable for anything but a very short run on
American fuels. In this test a large part of the coal originally in
the producer was apparently burned to ash, and its consumption was
completely left out of the test, causing very erroneous results.
3 Attention is called to the fact that the ash content in the ashpit
is practically much less than the ash content of the fuel, as shown by
analysis. The balance of the ash is undoubtedly in the fuel bed and
its presence there entirely upsets the basis of calculation for this
paper. It is safe to say that a two days' run would have given a
reversal of the first day's figures.
H. H. SuPLEE. I would like to speak of the unreliability of an
orifice as a means of measuring. My attention^ has recently been
called by a member to the discharge of steam from a boiler in which the
orifice and all conditions surrounding it were identical in several tests.
The amount of steam generated was measured,by carefully weighing
the water, double-checking it in tanks, and yet there was a variation
of ten to fifteen per cent in the results, the steam pressure and the
temperature being kept as uniform as possible. This fact casts a
doubt on the orifice as a means of determining flow.
L. B. Lent. The figures given show thatj.the draft through j^the
producer was practically li in., and yet 38.8 lb. of dry coal was
TESTING SUCTION GAS PRODUCERS 807
burned per sq. ft. of grate area. Still, with this consumption the
producer efficiency seems to be very good. My impression is that
this is a remarkable rate of consumption in producers of large type;
and I would like to know if this is the common practice in smaller
sizes of producers.
H. F. Smith. Regarding the conditions of the fuel already dis-
cussed it seems to me that the author has presented all the necessary
evidence to show that the conditions in the fuel bed were not the same
at the end as at the beginning of the test.
2 In the graphical log in Fig. 3, itVill be noticed that the temper-
ature of the gas leaving the producer at the beginning of the run was
400 deg. fahr., and at the end of the run something over 1300 deg.
fahr. The rates of gas production and fuel consumption were practi-
cally uniform. It is evident that there was some variation in con-
ditions, otherwise this difference in temperature would not have
occurred.
W. B. Chapman. Perhaps I can answer Mr. Lent's question in
regard to the quantity of coal gasified in producers. Producers for
furnace work are usually rated at 10 lb. per sq. ft. of internal diameter
on Pennsylvania coal, but only 7 lb. per sq. ft. on Illinois coal. The
best record I have seen for hand operated bituminous coal producers
was 16 lb. per sq. ft. Mechanically agitated producers gasify from
15 lb. to 30 lb. per sq. ft.
2 The question of the amount of anthracite coal that can be
gasified is very interesting. Engineers from abroad say that two
or three times as much can be gasified as is the custom in this country.
Every gas producer manufacturer in this country having a foreign
engineer in charge has designed his first producer very much too
small. The more experienced manufacturers do not rate their pro-
ducers at more than 10 lb. per sq. ft.
3 It is strange that we cannot get the results said to be obtained
in foreign countries. The difference must be in the coal.
Prof. R. H. Fernald. In reference to the rate of burning per
square foot of grate area, I desire to call attention to the high figures
shown by Mr. Garland. These figures seem to be very unusual for
this type of producer even under the test conditions described. The
highest rate with which I am familiar in commercial operation is
that found in the case of a large installation using lignite as fuel. This
868 DISCUSSION
plant shows a daily rate of 33 lb. per sq. ft. of fuel bed area per hour
during 164iours each day and 48 lb. during the remaining 8 hours.
In this installation the producers are of the down-draft type, but
even under these conditions this rate is, I believe, exceptional.
2 In reference to Mr. Chapman's remarks about the manufac-
turers abroad, I would say that apparently all of them stipulate the
tjrpe of coal that shall be used in their producer. They specify that
the coal must be of such and such a grade, non-caking and with only
such and such a percentage of ash and tar. As nearly as I was able
to ascertain, practically every manufacturer abroad has reached the
conclusion that it is wise to designate definitely the coal to be
used.
3 In one suction producer in Germany, operating on bituminous
coal, I found that the successful manipulation of the plant was due
to the fact that three kinds of coal, mixed in the proper proportions,
were being used. In other words, this type of producer using bitum-
inous coal as fuel was entirely feasible in the home plant of the manu-
facturer, but it would hardly prove a saleable article in this country,
as it would be almost impossible to guarantee the three required grades
of coal at all times. It would also be out of the question to secure
operators at a reasonable compensation who would give the plant
the required attention.
E. N. Trump. The rate of combustion in producers using anthra-
cite coal depends very much upon the size of the coal. Seven tons
per 24 hours, with a producer 7 ft. in diameter, is about the maximum
for No. 1 buckwheat coal. This equals 15 lb. per sq. ft. of grate sur-
face per hour.
2 Burning Western coals in producers, especially Hocking Valley
coal, a high rate of combustion is obtained. I have operated pro-
ducers continuously for a considerable period at the rate of 42 lb.
per sq. ft. of grate surface. This is with a large percentage of steam
in the air, also with mechanical ash extraction, the fuel bed being
thus kept well agitated.
2 Venturi meters give very accurate results in the measurement
of both gas and steam, much more accurate than the simple orifice.
The Authors. It will be well to emphasize thefact that the pro-
ducer under discussion was designed and intended for intermittent
service only; that is, it is not suitable for runs of greater than 12 to
18 hours duration. This is due to the small size of the producer, and
TESTING SUCTION GAS PRODUCERS 869
the absence of charging bell, water-sealed ashpit and mechanical
means for agitating the fuel bed.
2 Owing to the small size of the producer and the absence of means
for thoroughly cleaning the fuel bed from time to time, as above
noted, the accumulation of ash toward the end of 12 or 15 hours of
continuous operation is so great as to necessitate such thorough
cleaning as seriously to lower the heating value of the gas.
3 From the foregoing it will be evident that our test corresponds
to the condit ons under which the producer is normally operated.
Owing to the thorough cleaning of the fires before starting the test,
and the removal of the ash from the grate, a large quantity of green
fuel is brought into the path of the outgoing gases, resulting in their
being cooled. At this time, the temperature of the fuel bed is also
lower, as indicated by the analysis of the gases over the first two hours
of the test. The heating value of the gas is not lowered, for two reas-
ons: first, the descent of the green fuel into the path of the gases
results in the distillation of the CH4 and other heavy hydrocarbons;
secondly, an increase in the percentage of hydrogen results from the
lower temperature of the fuel bed.
4 At the close of the test the fuel bed was evidently at a higher
temperature than at the start. This resulted in increasing the un-
accounted-for loss in the heat balance, but its extent (estimated from
the results of a number of tests) is about one per cent for the present
test. This, it is believed, explains the condition pointed out by
Mr. Smith.
5 Professor Fernald and Mr. Tait call attention to the probable
inaccuracy in determining the v/eight of coal fired on the test. We
have recognized this source of error, and in Form 2 have included such
items as give proof of the accuracy of the work through the stoich-
iometric relations. As the full import of these items has evidently
not been realized, we will amplify them.
6 First, to determine approximately the purely mechanical error
in estimating the weight of coal fired during the present tests, the
producer was filled four separate times, and the weight of coal re-
quired was noted in each case. The average of the four weights was
taken as the mean or true weight of coal required to fill the producer.
The results are given in Table 1, herewith. It will be seen from this
table that the maximum variation from the mean is 8.751b., or 1.7
per cent. This in the test under consideration represents an error
of probably 1.1 per cent.
870
DISCUSSION
7 Mr. Tait seems to think that the presence of the ash in the fuel
bed "upsets the basis of calculation for this paper." The total
weight of ash in the dry coal is 776.5 lb.; 13.17 per cent = 102 lb.,
of which 52 lb. was taken out in the ash and refuse, leaving 50 lb.
remaining in the fuel bed. This would seem to indicate an error of
6.3 per cent, due to failure to remove this ash. Since the ash is soft
and fine it would pack into the interstices between the coals so that
its volume would by no means displace the same volume of coal.
If it displaced one-half its volume of coal it would cause an error of
slightly over 3 per cent. It is probable that its presence caused even
less error than this. In order to bring out the different errors we will
analyze the conditions existing on the test.
TABLE 1 WEIGHT OF GREEN COAL REQUIRED TO FILL THE PRODUCER
Trial Number
Weight
Lbs.
Variation from Average „ /- , tt , ..
„, . , ^ Per Cent Variation
Weight
1
2
3
4
669.25
676.25
683.25
683.25
-8.75 1.70
-1.75 0.28
+ 5.25 0.77
+5.25 1 0.77
Total
2712.00
Average
678.00
8 It is probable that the composition of the producer gas on leav-
ing the scrubber, and at any two points in the cross section of the
main, is the same. In order to eliminate such an uncertainty, how-
ever, we have taken the gas for our samples simultaneously'- from
different points in the cross section of the main and at a point beyond
the scrubber, by means of the sampling tube illustrated in the^ paper.
These samples were taken continuously over the period of the test,
both for analysis and for the calorimeter. The heating value of the
gas, as computed from analysis, is 138.1 B.t.u. After corrections
were made for the error in the meter, the error due to the vapor pres-
sure of water, and the error due to radiation and conduction into the
calorimeter, the heating value of the gas as determined by the Junker
calorimeter was 137.3. Since the heating value as determined from
two separate samples of gas, by two independent methods, and by
two independent observers, checks within 1 per cent, it must be ad-
mitted that the sampling, the analysis and the heating value of the
gas are probably correct within less than 1 per cent.
9 The volume of gas generated by the producer was measured by
TESTING SUCTION GAS PRODUCERS 871
a Westinghouse meter, guaranteed by the company to be accurate
within 2 per cent. However, as a further precaution, the meter was
carefully recalibrated and was found to be accurate within this limit.
A calibration curve was plotted from the calibration, so that the error
in determining the gas volume must have been within 2 per cent,
and was doubtless even closer than this.
10 As shown by a number of tests on the present fuel, the coal was
fairly uniform. A sample representing about 15 per cent of the coal
fired was mixed and quartered until about eight or ten quarts remained.
This was then ground, and again mixed and quartered until sufficient
to fill a quart jar remained. The heating value from this sample
as determined by the calorimeter was 13,040 B.t.u. per lb. The
mean of eight determinations on this same fuel showed a heating
value of 12,900 B.t.u. The probable error in the analysis and in
sampling the fuel, judging from the heating value, is doubtless not
greater than 1 or 2 per cent.
11 We have noted the volume of gas computed from the analysis
of the coal and the analysis of the gases in Form 2, Item 126. This
volume is 56,200 cu. ft. of standard gas, while the volume as actually
measured by the meter, corrected for the vapor pressure of water, is
57,500, showing a discrepanc}- of about 2.3 per cent. The volume
determined from computation was obtained from the formula of
Item 126, Form 3. It is based on the fact that the weight of carbon
in the coal fed to the producer must equal the weight of carbon appear-
ing in the producer gas, plus the carbon lost in the ash, plus the carbon
lost in soot and tar, plus the carbon lost by the absorption of CO2 and
CO by the scrubber water. The carbon lost in the ash is readily ob-
tained, the carbon lost in the soot and tar is not over 1 per cent,
while the carbon lost through the absorption of the gases by the scrub-
ber water is also very small.
12 It may be well to compute the carbon in the gas and compare
this with the carbon fed to the producer in the coal. We will com-
pute the latter first. The total carbon in the coal is 0.7984 X 776.5
= 620 lb. The total carbon in the ash is 0.388 X 85 =33 lb. The
carbon that should appear in the gas is therefore 587 lb. The total
weight of gas from Item 131, Form 2, is 3912 lb.
Carbon in CO, of gas = 0.0716X12/44X3912 = 76.4
a « Qo " " = 0.2925X12/28X3912 =490.5
" CH4"-" = 0.0112 X 3/4X3912 = 32.8
599.7
872 DISCUSSION
Thus 599.7 lb. is the total weight of carbon appearing in the gas as
measured by the meter. 599.7 — 587 = 12.7 lb. of carbon unac-
12.7 X 100
counted for = — ___ _ — = 2.1 percent. As already stated, there
may be an error of 1 per cent in the meter by which the above volume
of gas was determined, the error being either positive or negative.
There may have been 1 per cent of carbon lost in the soot and tar,
but not more than this; there may also have been an error in the analy-
sis of the coal amounting to l| per cent. We estimate the principal
errors in the test as follows:
Per
Cent
Error in filling the producer —2.1
Gas analysis or heating value of gas ±0.7
Volume by meter ±1.5
Coal analysis and sampling of coal ±1.5
Carbon lost in soot and tar — 1.0
Loss in sensible heat in the fuel bed due to the lower temperature at
the start than at the close of the test — 1.0
The total error in the results of the test that would affect the cold-gas
efficiency of the producer, if all the above errors are assumed as ac-
cumulative, equals 0.8. The probable error is 2.7.
13 There are three other errors that may affect the heat balance,
namely, the error in measuring the temperature of the outgoing gases,
the error in the determination of the specific heat of these gases, and
the error in the amount of steam fed to the producer. The error in
measuring the temperature of the gases may be 2 per cent; the error
in determining the specific heat of the gases may be 6 per cent; the
error in determining the steam fed to the producer may be 25 per
cent. If these errors are accumulative, the first two represent a total
error based on the heating value of the fuel of about 2 per cent, and
the third of about 1 per cent. Therefore, if all errors are accumulative,
the total error in the heat balance is about 6.8 per cent; as some of
these errors will be positive and others negative, the probable error in
the heat balance is about 3.5 per cent. As the heat balance shows
an unaccounted-for loss of 4.4 per cent, about 1 per cent being radia-
tion and conduction, the actual error in measuring the coal delivered
to the producer on this test could not have exceeded 3 per cent. We
have therefore been able to run tests of such duration as to reduce the
probable error in filling the producer to 2 or 3 per cent. Furthermore
we believe the results indicate that they are above the average in
TESTING SUCTION GAS PRODUCERS 873
accuracy for this Idnd of work, as we have seen very few tests on pro-
ducers that would stand the above analysis.
14 As Professor Fernald and Mr. Suplee have pointed out, we
have found the use of the thin plate orifice for the measurement of
steam not altogether satisfactory. As the heat supplied in the steam
on most of our tests is small, a large error is permissible in the measure-
ment. Our aim has been to vary the pressure on the orifice so as to
keep the hj'-drogen content of the gas practically constant. It might
be well to state that the orifice was used only while we were obtain-
ing a new vaporizer for the producei .
15 We have found no tendency in the anthracite coal to swell.
We believe that this is a property of bituminous coal containing large
quantities of moisture and of hydrocarbons.
IG As the quantity of air does not enter into the computation
of the more important quantities, it was computed from the nitro-
gen in the producer gas. The formulae for this computation are given
in Form 3.
17 The difference in the eflSciency based on dry coal and the effi-
ciency based on combustible, as noted by Professor Fernald, is due
to the fact that we have used the word combustible as defined in
Form 3, Item 54. The efficiency based on combustible corresponds
to the efficiency based on 100 per cent grate efficiency. It is used for
the reason that it is often desirable to show relations between effi-
ciency and other quantities that are independent of the grate efficiency.
18 The amount of coal burned per square foot of grate area is a
very variable quantity and depends upon the size of the fuel, the kind
of fuel, the nature of the ash, the amount of water supplied, the pro-
portions of the producer, the operation and the length of run.
19 For intermittent work, such as the present producer is adapted
for, and with coals containing an ash infusible at temperatures under
2300 deg. fahr., it is possible to operate at several times the capacity
possible with coal containing a fusible ash which necessitates a low
fuel bed temperature.
20 The rapidity and extent of the reaction of COj on incandescent
carbon depend upon the temperature and upon the catalytic action
of the fuel. At a given temperature and an indefinite time of con-
tact of gases with the incandescent carbon, a definite amount of CO2
and CO will be formed. The lower the temperature the less the per
cent of CO formed and the longer the time required for equilibrium,
so that with low temperature in the fuel bed the time of contact of the
874 DISCUSSION
gases with the fuel must be greatly increased. The same is true for
the reaction of water on incandescent carbon. Harries^ passed water
vapor over incandescent carbon at different temperatures and obtained
the results given in Table 2. These results show the effect of temper-
ature upon the water-gas reaction. Due to the low temperature, the
CO2 is high, the CO is low and the ratio of water decomposed to
water suppUed is small. The latter fact, in the case of the producer,
results in lowering the efficiency, as the undecomposed water carries
out a large quantity of heat.
21 The curves of Fig. 1, herewith, taken from Dr. Clements'^ work
on the rate of formation of CO in gas producers, illustrates the effect
of the time of contact, expressed in terms of velocity in feet per second,
upon the amount of CO formed in passing CO2 over incandescent
anthracite coal. At a temperature of 1100 deg. cent., and a time of
contact corresponding to a velocity of 1 ft. per sec, 11 per cent of
CO is formed. If the velocity is reduced to 0.1 ft. per sec, so that the
time of contact is increased ten times, 70 per cent of CO is formed. If
an indefinite time of contact is assumed, equilibrium is reached at
TABLE 2 EFFECT OF TEMPERATURE ON WATER-GAS REACTION
Temperature
Deg. Cent.
H»
CO
CO2
HiO
674
8.41
0.63
3.84
87.12
838
28.68
6.04
11.29
54.09
054
44.43
32.70
5.66
17.21
1126
50.73
48.34
0.6
0.303
this temperature with 90 per cent of CO formed. This illustrates
why it is necessary to use a small rate of combustion per square foot
of grate area, due to operating with coals requiring a low temperature
for the prevention of clinker formation.
22 If in the example just cited the temperature had been 1300 deg.
cent, in the fuel bed, 70 per cent of CO would have been formed
at a velocity of 0.5 ft. per sec The time of contact would have been
reduced five times, so that the rate of combustion could have been
increased almost five times without appreciably changing the com-
position of the gas or the depth of the fuel bed.
23 In the case of our tests with the Scranton pea coal, we have
* Habers, Thermo-dynamics of Technical Gas Reaction, p. 138.
* Bulletin No. 30, Engineering Experiment Station, University of Illinois.
TESTING SUCTION GAS PRODUCERS
875
been able to vary the coal per sq. ft. of grate area from about 10 lb.
to 45 lb., without appreciably affecting the efficiency of the producer.
At the higher rates of combustion, however, the producer requires
much more attention. If it were not for the fusion of the ash, the
weight of coal per square foot of grate area could be increased in-
definitely by the use of a blast and a sufficiently deep fuel bed.
24 The term " coal per square foot of grate area, " as used in pro-
ducer practice, is not, we believe, a true basis of comparison for the
operation of different producers, for the reason that the coal per
0.30
0.40 0.50 0.60
Feet per .Second
Fio. 1 Velocity of Gas in Feet per Second, Fuel Bed 1 ft. Deep
square foot of grate area depends to a certain extent upon the
depth of the fuel bed. For this reason, largely, we have used a term,
" rate of descent of coal through the fuel bed, " or "coal per cubic foot
of fuel bed per hour," which appears under Items 70 and 71 in
Form 1.
No. 1264
BITUMINOUS GAS PRODUCERS
WITH SPECIAL REFERENCE TO TESTS ON THE DOUBLE ZONE TYPE
By J. R. BiDBiNs, New York
Member of the Society
Several manufacturers have seriously applied themselves for years
to perfecting the bituminous producer. The problem has been diffi-
cult and success elusive; but the improvements of the last two or
three j^ears have been material, and likely to lead to a type univer-
sally acceptable as standard. Outside of the question of pecuniary
reward, much credit is due to these manufacturers for persevering
against material obstacles and personal prejudice, and at an expense
ruinous to any but those possessing large resources.
2 It is the object of this paper to record the results of the most
recent achievements in this direction, and to interpret them in the
light of personal experience. No attempt is made to discuss the com-
mercial aspect, and in this respect the results presented will largely
be left to speak for themselves. These results are drawn from resources
accurate and reliable in so far as commercial tests can be made
to approximate scientific investigation. Beyond this no claims can be
made for refined accuracy.
ESSENTIAL REQUIREMENTS
3 Successful operation of a modem gas engine generating station
prescribes certain requirements in the producer plant:
a Continuous operation, 365 days per year. Any departure
from this condition means reserve equipment, additional
capital outlay and idle plant. Producer designers can-
not escape at this advanced stage of the art a condition
parallel to that of steam boiler practice. For this con-
tinuous service the water-seal has proved adequate, but
Presented at the Annual Meeting, New York (December 1909), of The
American Society or Mechanical Engineers.
878
BITUMINOUS GAS PRODUCERS
o
Q
O
a
12
^
H
O
U
^M
% 3
M
A:'.-
1 "^
^■V'
i
BITUMINOUS GAS PRODUCERS 879
some means of mechanically removing ash should be
developed.
b Plant suited to various kinds of fuels without remodeling,
such as change of grates, etc. Fortunately gas producers
are unusually flexible in this regard.
c Gas clean and free from tar. No engine design except per-
haps some one of the simple valveless types can withstand
the action of viscous tar deposits on the valve seats. Me-
chanical extraction can hardly be considered an acceptable
remedy in this regard.
d Moderate labor requirements. No design will last which
requires excessive attendance and large periods of shut-
down for cleaning or repair.
e Prevention of clinker formations. . Both labor cost and the
uniformity of gas production are affected seriously by
clinker. The obvious remedy is relatively low fuel bed
temperatures.
/ Automatic gas regulation. Large and expensive gas holders
should be unnecessary. Quantity and quahty regulation
of gas may be made substantially automatic by proper
design. An essential requisite is to reheve the producer
attendant of all possible adjustments, as it is almost im-
possible to obtain at the prevailing wage the grade of
intelligence otherwise necessary. The power-driven ex-
hauster has removed a great proportion of the disabihties
of the steam blown producer.
g Minimum auxiliary apparatus. It is manifestly inadvis-
able to nullify the high efficiency of the producer by waste-
ful auxiliaries. For this reason the suction principle has
come into favor. Internal vaporizers provide automatic
regulation quite adequate to the usual fluctuations in de-
mand for gas, thus dispensing with the smaU boiler.
DESCRIPTION OF POWER PLANT, ETC.
4 The tests herein presented pertain principally to the double
zone type of producer. As a complete description of this type was
incorporated in the last report of the National Electric Light Associa-
tion, 1909, constructional details may be dispensed with. Fig. 1
880
lilTUMINOUS GAS PRODUCERS
shows the arrangement in sections, comprising the following essential
parts :
Water sealed ash pit,
Lower coke gasifying zone,
Central belt evaporator,
Upper coking zone for green fuel.
Air cooled top (preheating air blast),
Charging funnel open to atmosphere,
Vapor control valves for top and bottom fires.
Fig. 2 525-h.p. Bituminous Producer Plant, Western Chemical Com-
pany, Denver
Radial poke holes for raking topand bottom walls.
Static cellular washer.
Positive rotary exhauster,
Automatic by-pass regulator valve,
Regulating gasometer for maintaining constant delivery pressure to engine.
5 This system obviously works entirely by suction with the charg-
ing top at atmospheric pressure. The sole adjustment is the relative
position of the vapor control valves, which are set permanently for
any given fuel and require no change for ordinary variations in power
load. These valves determine the relative rates of combustion in
BITUMINOUS GAS PRODUCERS 881
the upper and lower zones, the temperatures, and the rate of settling
of the two fuel beds. While the producer is not supersensitive, intelli-
gent adjustment is necessary to secure the most uniform gas. But the
gas holder is dispensed with entirely, as the production is directly
proportionate to the demand, giving a constant delivery pressure at
the engine.
SCHEDULE OF TESTS
6 This producer plant has been under test in commercial sizes (175
h.p.), at East Pittsburgh, since December 1907, with various fuels and
under various conditions of load. Up to July 1009, a total of over
20-10 hours of operating tests had been run, operating from a mini-
mum of 47 to a maximum of 514 hours continuous tests and on both
10- and 24-hour runs. Over 266,000 lb. of coal were gasified, the fuels
ranging from low grade lignites to the best Pocahontas semi-bitum-
inous coal. Some trials were also made on meadow peat. All the
gas made was tested by means of a standard three-cylinder engine
of 140 h.p., operating also continuously against the resistance of a
prony brake. The gas was measured by wet meters at both' the pro-
ducer and the engine. Determinations were made regularly for cal-
orific value by means of the Junker calorimeter; for impurities by
the Sargeant filter paper method; for composition and heat value
by chemical analysis. Coal was weighed on scales — not measured.
Table 1 shows a complete schedule of tests; of these special Tests F
and G were run to determine accurately the normal standby loss;
Test H to try out the type of washer shown in the sectional drawing,
Fig. 1. It is apparent that this scries of tests is unusually valuable in
indicating results under various conditions of service. The import-
ant results follow, and are discussed seriatim.
DISCUSSION OF RESULTS
7 It should be noted by Table 2 that fuels containing as high as
25 per cent of water were successfully used for power purposes.
The efficiency curve (Fig. 3) fully establishes the fact that the effi-
ciency of heat conversion is practically as high with lignites as with the
cheaper fuels.
8 In the test with Texas lignite an important fact was brought out,
which has especially puzzled theorists for some time, viz., that with a
poor fuel the rate of combustion can be increased sufficiently to per-
mit the same rating of the producer as with better fuel. This re-
'For check purposes, meter calibrated by positive holder fall.
«82
BITUMINOUS GAS PRODUCERS
TABLE 1 SCHEDULE OF TESTS
Load on Pbo-
Dura-
Hr. per
day
DUCBB
'i'est
Date
Fuel
tion
Hours
Remarks
b.h.p.
Max.
1908
A
4/2-4
So. Am. Lignite
72
11
....
....
Purged Gas.
B
4/1&-30
Col. Lignite
314
24
121.7
156.9
Continuous test.
C
5/ 8-23
Pittsburgh Coal
298
24
158.3
129
u
D
7/16-31
« «
370
10
158.5
206-
Intermittent test.
E
8/ 4-25
514
10
170.8
190.7
("Standby test.
F
9/ 1-19
« «
432
Standby
Fires blasted
G
10/12-19
" "
168
"
1 hr. once in 24 hr. day
H
11/ 9-14
1909
" **
10
137 to
204
Washer test and capac-
ity test
I
6/ 1- 2
Pocahontas Coal
46i
22}
75.6
Continuous test.
J
6/ 3- 4
" «
48
24
101.4
« «
K
6/5-6
« (<
48
24
126.5
" "
L
6/ 7- 9
« <<
72
24
150
« u
M
6/30- 2
Texas Lignite
72
24
128
135
« u
N
7/ 7- 8
42
24
157.2
It II
moves a heavy restraint on the development of producers for the
enormous Hgnite fields of Texas, Wyoming, Colorado, Montana and
the Pacific States. In Test N, Table 3, a charging rate of 27.2 lb.
per sq. ft. per hr. was maintained with Texas Lignites and 15 lb. with
Pocahontas coal, both at 150 h.p. load ; with Pittsburgh run of mine
it was slightly higher (18.1).
9 An economy of less than 1 lb. per brake horsepower-hour is
probably below previous results with bituminous producers. This
TABLE 2 TYPICAL PROXIMATE ANALYSES OF FUELS TESTED
Class of Fuel
Moist-
ure
Meadow Peat — Massachusetts
Lignite — South America
Lignite — Northern Colorado
Lignite — Texas
Bituminous — Pittsburgh run of mine. . . .
Semi-Bituminous — Pocahontas run of
mine
38.10
20.05
16.63
24.08
2.03
Volatile
40.54
34.44
33.78
38.55
34.98
1.39 I 16.01
Fixed
carbon
17.86
30.85
42.22
28.76
56.22
74.28
Ash
Sul-
phur
B.t.u.
per lb.
as fired
3.50
1.05
6410
14.66
8035
7.37
9589 ;
8.61
0.57
7974
6.77
1.29
13305
8.32
13983
B.t.u.
per lb.
dry
10340
10045
11500
10503
13590
14170
BITUMINOUS GAS PRODUCERS
883
§ s s
.toiiapHia ;uoo .laj
pouati iJOiis xnn — »
1
1
,
p
\
\
1
1
§
inoH t
1
c
i;na)n()xi3i;S[ '|
V
\*
\
1
sJnOHG
\
\
\
\
\
t\
If II
«
n
jndjno ^If^'a ■ X
BK '
\
^ ^\"'^
.\
o <
3 l^^^O^
a —
\
\,
j
\
/So
3
\\
o
■^^
ii
^(§
5s
y
60
5
\
\
^^
X 0
A
\
1
\
\\
V
o
N
^ 1
_3
\
i;
x
1 3
1 ^
^
V
1
13
\\
J
V
1
\]
V
V
^\
\
WP"^-
\
\!
\
'a\ V: j*\
niSiiac
1
A
"^^
\
fe
%
\
^
^'^^
\
♦A.
\
w^
'^
^
3^>iV^
^^
^
■%-\
S^
3 o
3
oofe
t- ^
a
u
C3
o
o
C4
Oh
o
^5 a
£ f^
w
suotxTlK-JnoH jad n-^-g
884
BITUMINOUS GAS PRODUCERS
corresponds to less than 1^ lb. per kw-hr. in an electric generating
station. An interesting point is the low standby fuel consumption,
which averages in over a week's run 1 lb. per sq. ft. of fuel bed area
per hour. In Test G it was reduced to this amount from 1.49 lb.
(Test F) simply by reducing the natural up-draught through the
idle "producer, by closer adjustment of the valves.
,-8:s8
8 8 8
8 8
g^ GOO
^_
y
V
^
^
-^
r-N
J
L
^
^N
..--.,
^
-'roducer
deg. fa
1 1
— ^
'
from analyses I
Total B.t.u.
o O O O O
__
__
-=
=-
-
/
'*
—
s
/
■^
/
s
/
"^
/
CIh 165
>s-
r—
.s^
-•y
■s./
V ^
J
^
L_
L.
r"
^
L-
Load on lijDg-ine
n
s
/
V
a -3 80
1 »70
^=
_
^
s:
^
^
=
^
^
—
-^
^
_
_
, .Gas Oi
it
Sci
ubt
er Outlet
§•2 60
—
—
—
—
—
—
—
—
^^^
^=
Wate
:'V
—
—
Inlet
9
0 8
S3' 7
1 R
A
/
s
/
—
\
1^
Sc
ub
ler
V 1
-J
-
y
\^
\
>,/
^
f
s
^
1
r
Oullet
Pr
!ssure
v
V
1 6
o
i 8
& 2
1
1
1 j
Kxhaus
er Outlet
Pi'efcsiire
r~
^
rv--
f-^
"-V
Vi T\
y(<;ri'oducer|Outlet
■
—
-"
«^.^
-^
•^
V
k/
1
Press
ire
1
1
Fig. 4 Typical Log, Pocahontas Coal, Last Day of Run, 2\\ hr.
10 Test H, which was a capacity test, shows an 18 per cent over-
load on gas production maintained for nine consecutive hours with
Pittsburgh run-of-mine. Test C with the same coal shows nearly
30 per cent overload.
11 In heat value the gas is not high; but, which is more im-
portant, it is fairly uniform as shown by the typical log, Fig. 4, 5 and
6. The heat value seems to bear a certain relation to the fuel bed
temperature. It is found that if a certain temperature of the gas off-
BITUMINOUS G.\S PRODUCERS
TABLE 3 FUEL DATA
885
Fuel Consdmption
Lbs.
charged
incl.
standby
Lbs. C
Per
BARGED
Hr.
Max.
Lb. perl
sq. ft
per hr.
Test
Per b.h.p. hr.
Per kw
. HR.*
Avei;.
Gross2
Net3
Gross
Net
A
5003
151.6
160.8
15.2
2.3
2.3
3.33
3.33
B
60740
194
228.5
19.4
1.69
2.31
C
33925
181.1
207
18.1
1.14
1.66
D
27157
169.2
222
16.9
1.39
1.19
2.01
1.73
E
35647
184.6
202
18.5
1.24
1.08
1.8
1.56
I
4849
104.2
10.4
1.37
2.0
J
6065
126.4
12.6
1.25
1.81
K
6403
133.4
13.3
1.06
1.54
L
10699
149.1
14.9
.983
1.42
M
16964
233.1
23.3
1.82
2.64
N
11503
272.1
27.2
1.80
2.61
F
8813
14.9
This standby rate applies to previous
tests
G
1680
10.0
Standby rate reduced by reducins up-
draft on fire
' Based on area at level of green fuel.
^ Including standby fuel.
' Standby fuel deducted.
* Based on 92.5 per cent generator efficiency.
TABLE 4 GAS DATA
Cu. ft. per
B.t.u.* per
Test
Cu. ft.
hr. 60 deg.
Max. for
Cu. ft. per
Cu. ft per
cu. ft.t
B.t.u.
uncorrected
fahr. 30 in.
mer.
1 hr.
lb. fuel
b.h.p. hr.
total
effective
A
238,900
7245
8500
47.7
102
113.4
B
3,873,000
12400
16700
63.8
101.4
118.2
C
4,812,200
16210
23300
89.3
102.1
113.8
D
2,074,500
15660
21500
88.2
101.9
110.7
E
2,797,470t
16600
17853
90.2
111
H
17815
21511
(9 hours)
111.4
K
119
109.7
L
120.6
112.2
M
891.300
11933
51.3
93.1
118.7
107.8
N
569.300
13092
15500
48.05
86.75
129.7
118.75
* Total heat values. Water determination not made during some tests.
t Corrected to 30 in. mer. and 60 deg. fahr.
886
BITUMINOUS GAS PRODUCERS
take is exceeded (about 1000 deg.), vitiation of the gas ensues from
excessive combustion. The condition of the fuel bed may be readily
watched by means of a pyrometer (in the discharge) and with proper
adjustment of vapor and draught, temperatures may readily be held
below this limit, especially with lignites.
^§S00
^1400
iftSOO
£""120
150
140
^13000
^12000
90
1-180
£a70
§■0 60
I 13
0 12
W 11
1 10
I 9
^
~~"
■~^
r-
"~
—
"
—
-^
■^
1
Lo|ad
)n
bngine
G
asr
lad
e per Hour 1
/"
—
r—
—
1
r--.
^
/
=-
^"^i
==>
-=
-
—
—
—
—
-=,
^^^
^
i
cr
ibb
jr Waste >Vater
OntilptlOnL
^
;;=
=
=
=
—
—
—
=
—
—
^
—
■77-
Inl
•\.\
Va-'jr
,--'
s
^
\
/
S
^
—
•-.
—
—
S
:iu:
)be
C
Jtle
t
^
\
/-
\
^
—
—
^
_
■~-
Pi
Oi
tie
Fig. 5 Typical Log Texas Lignite, Beginning of Test
12 In^this connection a comparison of these heat values with sim-
ilar values from other plants is interesting, revealing the extreme
range permissible with engines of modem design (Table 10). Both
are fair operating plants, but deliver gas at a considerable variation
from specified value (125 B.t.u.), without occasioning any disturb-
ance in the operation of the engine. Results from the double-zone
producer show that present engine ratings are well suited to the gas,
a higher compression is permissible, and that a high hydrogen content
— as high as 20 per cent — does not necessarily interfere with opera-
tion.
BITUMINOUS GAS PRODUCERS
887
13 Perhaps the most important result is tar-free gas. The im-
purities normally consist of dust and lampblack. By the filter paper
method, Fig. 7, it is possible to detect the least trace of tar, which
quickly discolors through to the second layer of paper. Fig. 7 shows
the maximum deposit from a run on Pittsburgh coal. Note that there
is no discoloration of the second paper.
ga g
^■^600
1-^500
"5 g400
II 300
130
"i V 3-120
H 100
a.
i 130
% 1S5
S 12000
£ 11000
5 10000
1-80
fa TO
9. 8
g S 8 8 _ 8 8
d> 3 ;3 a ' ^ <^
8 ga
_^
s
-
y
/
'^
__
■^
_
\
^
i
/
s
N
^
_
___
Load
on
Tr,
•
....
'""*= '"^
Gas made per Hour
— ■
"—
"—
"^
_
_
__
J
G
a£-
icr
abb
er(
)ut
let
.
^
J
^
W
ite
slJ
t
'
"
Ll^^H
y
—
-«.
^
/
\
^
—
z'
v
^
Sci
ahi
er
Ont
let
Ou
Ele'
-
^
' — 1
1=1
■
^
"
—
U
xlUcer
r~
£
zh
.oa
er
3^
W
Wi
^SSl
re
Fig. 6 Log Texas Lignite, End of Test, 114 hr.
14 Test H, (Table 6), 25 determinations of Pittsburgh run-of-
mine, shows well under 0.02 gr. per cu. ft., which is below the usual
guarantee. In Tests M and N determinations on Texas lignite averaged
0.0193 gr. per cu. ft. These results are borne out by results in the
field. Seventy-three determinations at Denver^ averaged 0.022 gr. per
cu. ft. In these tests all of the gas determinations represent average
gas drawn continuously throughout the day's run. In no case are
' Western Chemical Co.
888
BITUMINOUS GAS PRODUCERS
TABLE 5 PRODUCER EFFICIENCY
Producer Efficiency Based on
Plant Efficiency Based on
Test
Total
EfiFective>
Gross Fuel
Net Fuel
A
j 76.3
B
' 78.8
16.75
C
76.5
16.8
D
1 77.2
13.9 16.05
F
75.5
15.45 17.6
V
13.2
J
14.7
K
17.5
L
17.7
M
76.8
69.8
17.5
N
1 76.95
70.2
17.6
• Efficiency on effective basis from 7 to 8 per cent lower than on total.
TABLE 6 DUST DETERMINATION
Test H Pittsburgh Run-cf-Minb
Gas perhr. cu. ft.
Impurities in
Gas, Gr. per Cd.
Ft.
Load h.p.
Average of Five
Determinations
Max.
Min.
142
14910
0.02079
0.0432
0.0129
137
14310
0.02087
0.0.398
0.0100
170
17740
0.01712
0.0318
0.0062
184
19260
0.01611
0.0287
0.0034
183
19160
0.01718
0.0459
0.0063
204
21511
Test M
Texas Lignite
114
140
28,467
. . . 1,400.000
0.0193
Maximum, gr. per cu
Minimum, gr. per cu
.ft
0.0770
.ft
0.0010
BITUMINOUS GAS PRODUCERS
889
TABLE 7 TYPICAL GAS ANALYSES
Test
Fuel
H2
CO
CH4
CO2
N2
A
17.4
10.4
11.6
15.2
2.4
2.6
1.8
12.4
13.2
9.6
56.6
B
c
Colorado Lignite
Pittsburg Coal
17.6
14.1
54.4
68.9
TABLE 8 CHARACTERISTICS OF LIGNITE GAS
Western Chemical Co., Denver
Date
4- 8-09
4- 9-09
4-10-09
4-11-09
4-12-09
4-13-09
4-14-09
4-15-09
4-16-09
4-17-09
4-19-09
4-20-09
4-21-09
4-22-09
4-23-09
4-24-09
4-26-09
4-27-09
4-28-09
4-29-09
4-30-09
5- 1-09
5- 2-09
5- 3-09
5- 4-09
Average
Heat Value (Total)
No. of
Deter-
mina-
tions
Max.
Min.
129.6
121.5
123.9
115.4
124.0
117.3
128.0
128.6
128.0
117.8
130.0
129.0
126.0
112.4
113.7
106.1
111.0
121.0
113.7
114.7
108.5
108.5
110.0
111.0
123.4
106.4
141.6 i 127.0
131.0 136.0
Average
123.0
118.3
113.1
122.0
lis. 2
122.0
115.5
119.5
133.3
121.2
117.8
114.7
121.0
126.0
115.7
125.0
120.4
133.2
128.5
121.2
Impurities
No. of
Determin-
ation
Max. Min.
0.0290
0.0617
0.0667
0.0264
0.0204
0.0198
0.0260
0.0194
0.0046
0.0095
0.0288
0.0048
0.0064
0.0429
0.0089
0.0099
0.0172
0.0.356
0.0587
0.0.328
0.2325
0.01567
0.0144
0.0081
0.0115
0.0051
0.0048
0.0209
0.0058
0.0018
0.0051
0.0113
0.0018
0.0036
0.0040
0.0020
0.0059
0.0066
0.0216
0.0340
0.0067
0.0735
Avefage
0.0213
0.0284
0.0231
0.0184
0.0114
0.0124
0.0247
0.0126
0.0033
0.0051
0.02005
0.0033
0.0053
0.0179
0.0056
0.0073
0.0123
0.0304
0 . 0463
0.0205
0.1381
0.02227
890 BITUMINOUS GAS PRODUCERS
TABLE 9 SCREEN TESTS— FUEL AND ASH— NORTHERN COLORADO LIGNITE
Western Chemical Co., Denver
.
Fdbl
Ash
Over i Inch
Over y'^ inch
Through -j^g-mch
58 per cent
23i per cent
18i per cent
23.5 per cent
33 . 0 per cent
43 . 5 per cent
1 100 per cent
100 per cent
Samples represent about one bushel of material quartered down from stock pile.
TABLE 10 TYPICAL PRODUCER OPERATION
Plant A
Plant B
Texas Lignite
Pocahontas
Coal
Max.
1
Min.
Avg.
Max.
Min.
Avg.
Date of Observation
Ma
y 20-29,
•08.
Apr. 1-9,
'08.
No. of Determination
20
24
CO2
11.6
0.9
14.8
19.3
2.52
62.4
8.0 1
4.0
12.0
11.84
1.5
5.7
10.0
0.63
13.7
15.3
1.88
55.2
11.6
19.4
12.4
8.8
65.1
j 8.6
11.4
4.7
5.1
53.9
9.9
O2
CO
H2
CH4
N
15.7
9.09
67.6
5.93
Heat Value Total
109.9
89.2
102.7
20.7
166.4
157.8
110.4
101.0
133.1
Heat Value Effective
124.5
Fluctuation B.t.u
56.0
Fluctuation Per Cent = Avg. Value . .
....
....
10.1
21.0
BITUMINOUS GAS PRODUCERS
891
snap samples used. The latter method of testing should be rigor-
ously avoided except for some special purposes, as it affords no indi-
cation whatever of average conditions.
15 In former papers* the writer has described the method of ob-
taining producer efficiency from isolated tests. Fig. 3 shows the close
agreement of this theory with the fact based upon these several dif-
ferent kinds of coal tested. The interesting point is illustrated, that
the gas producer varied only 10 per cent in efficiency throughout its nor-
mal range of load. This type will give approximately 70 per cent
Fig. 7 Comparative Methods of Testing fok Impurities
5 CU. FT. EACHi30 MIN.
0.0259 QH. PBB CU. PT. FILTER PAPER, TWO LAYERS
0.0216 OR. PER CU. FT. COTTON BATTING, TWO LAYERS
efficiency (on an effective heat basis), or 77^ per cent (total heat
basis), at full load. This is manifestly reasonable by inspection of
Table 5. That all fuels should fall so. closely on the heat input lines
at various loads, is a remarkable agreement, and closer than antici-
pated.
16 However, in a paper by the writer^ the same agreement was
found in plotting the results of tests on another type of plant at Rich-
• Norton Test, Vol. 29, Transactions A.S-M.E. 1907; Transactions i A.I.E.E.,
page 1128, vol. 27, 1908
892 BITUMINOUS GAS PRODUCERS
mond, Va. The standby, and three-load determinations, followed
almost on a straight line of heat input to producer.
OPERATING RESULTS
] 7 After the first year's period of tests, this plant was dismantled
for examination. The gasification of 182,472 lb. of fuel showed no
perceptible effect upon the condition of the producer, the walls being
practically intact. This is, due to the complete absence of clinkers
and high temperatures. Experience plainly shows the latter to be
the cause of clinker troubles. The fuel bed normally grades from
small ash at the bottom through pure coke to green coal at the top.
With proper handling, clinkers may be entirely avoided. For ex-
ample. Table 9 shows a screen test of coal and ash at the Denver
installation — 43^ per cent ash through a ife-in. screen.
18 An examination of a long gas main and the engine valves after
the year's run, showed no deposits of tar either near or distant from
the producer, indicating the complete fixation of the volatiles. Ah
condensibles are removed in a static washer, the cells of which seem
to clear themselves automatically of the deposit. For example the
pressure drop or resistance through a month's run increased slightly
more than ^ in. The principal skill in handling this producer is re-
quired in studying the characteristics of various kinds of fuels. Each
must be handled differently for best results. With friable fuels the
" let well enough alone " rule is particularly desirable, as resistance of
the bed may be greatly increased by too much poking.
GENERAL CONCLUSIONS
19 The writer's good fortune in obtaining results directly or in
analyzing results from several types of producer plants, has led him
to convictions on producer practice in general along certain broad
lines. Knowing the facts regarding daily performance from both
shop and field, the statements herein contained are believed to be
conservative. It is not meant to convey the idea that success-
ful working has been confined to any one particular type of plant or
equipment. But those undertakings that have been backed up by
experience gained in tests on a commercial scale are certainly most
sure of success and worthy of support. There is much activity
resembling plagiarism in the power gas field — guarantees based solely
on productions from competitive results, portions of a foreign design
BITUMINOUS GAS PRODUCERS 893
incorporated in an incoherent whole, etc. Such mis-matching inevit-
ably leads to failure without thorough analysis of cause and effect.
In short, the great desideratum is a campaign of investigation at
the factory which shall not be of the shiftless pounds-and-kilowatt
order, but sufficiently comprehensive to facilitate accurate inter-
pretation. Many failures that now lie to the discredit of the internal
combustion system would thus be avoided.
20 It is also patent that personal prejudice seemingly plays a far
too important part in dominating the selection of plants, steam as
well as gas, and its effect appears not only in the selection but also in
the operation of the plant, making it often impossible to secure
results which under proper conditions would be easily within reach.
It is therefore as desirable to analyze reports of poor results ob-
tained from a given plant, as of results which seem so good that
they arouse suspicions of inaccuracy.
21 Finally, a present need in producer work is some reasonably
accurate means of control indicating and compensating for "low
gas, " which may result from poor condition of fuel bed. Automatic
production without the holder storage has now become an accom-
plished fact with the simplest apparatus, but no means ai'e available
for keeping under observation the heat value, except the cumbersome
and delicate Junker calorimeters. Were it possible to alter in inverse
proportion the ratio of air to gas at the engine accordingly, maximum
efficiency could be maintained. But this variable factor has received
practically no attention, and as a consequence producer operators are
working entirely in the dark.
22 The principle of tar-free gas production has its demonstration
in the type of producer under discussion, and it is believed the end
justified the means, even at the shght expense of heat value. Due to
unknown and complex reactions, tar laden or " green " gas possesses
somewhat higher heat value than tar-free gas, due possibly to the
preservation of more unstable hydrocarbons. This might then be-
come a factor in rating, with engines designed with little or no margin.
But good practice today recognizes no difference between gas of 1 1 0
and 125 B.t.u. The margin necessary is present for other reasons,
and there is therefore no practical obstacle in the way of development
of the tar-free type of gas producer in any form.
23 Further investigations are urgently needed as to the pos-
sibility of larger units. If the gas engine is considered behind the
times in the matter of development of large units as compared with
steam practice, the producer is hopelessly so. For rapid future devol-
894 DISCUSSION
opment, large units are an obvious necessity. We have steam turbine
units capable of sustaining continuous loads of 30,000 h.p,, boiler units
of 2000 to 3000 h.p., gas producers of only a few hundred, maximum.
The very multiplicity of units thereb}'- required in the design of a
large station, very seriously militates against the selection of the pro-
ducer gas system of motive power.
DISCUSSION
G. M. S. Tait. The results reported in this paper are entirely in
accord with what we have found, namely, that the gas of the lesser
British thermal units is much more satisfactory for engine practice.
In other words, the efficiency of a gas of 90 B.t.u. is proportionately
double that of a gas containing 600 B.t.u. per cu. ft., the gases in
question being respectively blast-furnace gas and gasolene vapor,
2 I would like an expression of opinion as to the reason for this
great discrepancy in efficiency between the two gases, my own opinion
being that the excessive normal losses are due to the sudden high
temperature developed in the gas of high B.t.u., which is greater
than can be handled by normal piston speeds.
3 The tar washer used in this test appears to be a succession of
water seals and I would like to know what would be the total fric-
tional effect of these seals under normal conditions and on full load.
4 In all producers properly designed the thermal efficiency
appears to remain constant between 20 per cent and 100 per cent load.
I can confirm Mr. Bibbins' experience as to the action of this par-
ticular class of fuel and its desirability for producer work.
Prof. R, H. Fernald. Mr. Bibbins places as his first essential
requirement "continuous operation 365 days per year," and states
that any departure from this condition means reserve equipment.
He also states that the condition for producer operation must parallel
steam boiler practice.
2 It is undoubtedly true that a producer which will operate
continuously 365 days a year would prove a splendid commercial
proposition, but it seems to me that in the requirements outlined the
conditions imposed are much higher than those of any steam boiler
plant and are beyond practical requirements. Every plant of any
size must necessarily have one or more reserve units, as no plant can
operate continuously 24 hours a day 365 days a year. If the producer
described by Mr. Bibbins can approach this operating condition, it
BITUMINOUS GAS PRODUCERS 895
will certainly revolutionize our present day power-plant practice.
It would seem advisable, in the light of the present development of
gas producers, to impose conditions which are less severe.
3 Relating to the adaptability of a single producer to all classes
of fuel, it is well to bear in mind that the government testing
station has practically demonstrated the fact that almost any variety
and grade of our recognized fuels can be handled with more or less
success in a given producer installation without change of details
of design. It is questionable, however, whether such practice lends
itself to the efl&cient use of a wide range of fuels. It is probable that
better results can be obtained by utilizing a producer type to cover
a certain range or variety of fuels and another plant of somewhat
modified design for another range.
4 Mr. Bibbins refers to the excessive labor required by most
producers. At the present time the labor requirements are excessive
for the majority of the plants utilizing bituminous coal. This labor,
however, even under bad conditions of operation, such as those
involved when the fuel is one that clinkers badly, probably does not
exceed that of the average steam installation, although the labor is
of a somewhat different character. During the regular operating
period of the plant this labor may amount to very little; but at the
close of a week, two weeks, or any length of operating period, in the
commercial plants now in operation, cleaning may be an exceedingly
dirty, hot and tedious operation. With the steam boiler plant the
labor is more uniformly distributed. In spite of the more erratic
and more violent labor required at times by the producer installa-
tion, the total cost for cleaning, ash removal, etc., is probably within
the limits of the average steam installation.
5 Experience with a large variety of fuels leads one to question
whether the treatment accorded one fuel in order to prevent clinker-
ing will produce the same results with a fuel possessing totally dif-
ferent characteristics. The impression from the tests carried on at
the Geological Survey testing station is that fuels varying greatly
in composition and in characteristics require widely different treat-
ment. This impression has been obtained from tests on a large variety
of fuels, but the number of tests on each of the different fuels was not
sufficiently large to warrant positive conclusions regarding this point.
European practice, however, seems to confirm this opinion, as practi-
cally every producer manufacturer finds it imperative to specify
coals of certain characteristics for use in his type of producer and
does not guarantee the plant on fuels outside of this class.
896 DISCUSSION
6 In the discussion of the results the point is brought out that
with Texas lignite the rate of combustion in this producer can be so
increased as to permit the same rating of the producer as when oper-
ing on a high-grade fuel. Note is made of the fact that a charging
.rate of 27.2 lb. per sq. ft. per hr. was obtained with this lignite. An
installation in Texas, which I visited a year ago, consisted at that time
of three producer units of 1100 h.p. rating each, or a total of 3300 h.p.
7 Owing to the character and high percentage of the ash, together
with the excessive demands upon the plant each unit was cleaned
every third day, or, what amounts to the same thing, one unit was
cut out of operation during a part of each 24-hr. day. It required
eight hours to cut out the gas from a given unit, to clean thoroughly,
rekindle fires and cut in the new gas. During each 2'4-hr. day, then;
the full plant capacity, rated at 3300 producer h.p. was in operation
16 hr., while only 2200 producer h.p. were in oj)eration the remain-
ing 8 hours. During the entire 24-br. period, however, according to
the operating records, the engines were developing 2800 h.p. The
operating records also showed that the fuel consumption per square
foot of fuel bed area per hour amounted to 33 lb. during the 16-hr.
period and 48 lb. during the 8-hr. period.
8 The statement is made that the economy of less than 1 lb. per
b.h.p.-hr. is probably below previous results in bituminous producers.
It is assumed that this statement is not intended to cover the tests
at the government testing station, which has reported a number of
instances in which the consumption varied between 0.8 lb. and 1
lb. per b.h.p.-hr.
9 Mr. Bibbins states that perhaps the most important result is
tar-free gas. It is undoubtedly true that tar-free gas is eagerly
sought in all cases in which the gas is to be used in engines. In
my own mind, however, it is somewhat questionable whether tar-
free gas, as reported in this paper, means that the gas from any and
all fuels used in this plant would necessarily be free from tar. Experi-
ence with a aroducer of somewhat different design shows tar-free gas
with the m jority of fuels, but in the case of certain fuels the results
are quite threeverse. If the producer under discussion can produce
tar-free gas from any and all varieties of fuel, it is certainly a de-
velopment in the right direction.
10 In the closing paragrai)h of Mr. Bibbins' paper the impr ssion
is conveyed that the steam boiler units of 2000 and 3000 h.p. are
found not infrequently, and that producer units are small in com-
parison with the usual boiler unit. In my opinion the condition at
BITUMINOUS GAS PRODUCERS 897
TABLE 1 CAPACITIES OF PRODUCER-GAS POWER PLANTS
No. of
plants
Total
HoRSErOWBR
Per Cent
OP Total
Average
Minimum
Maxi-
mum
No. ;
i
h. p.
Anthracite Coal:
Over 500 h.p
500 h.p. or less
8
407
7,550
40,550
950
100
600
15
1500
500
1
Total
415
48,100
116
15
1500
88
43
Bituminous Coal;
Over 500 h.p
500 h.p. or less
20
17
49.000
5.150
2,450
300
750
35
6000
500
Total
37
54,150
t,4R0
35
6000
8
1
49
Lignite:
Over 500 h.p
500 h. p. or lesd
3
19
7,275
1,725
2430
90
525
25
3750
250
1
Total
22
9,000
410
25
3750
4
8
All Plants
474
lli,250
235
15
6000
100
1
100
the present time is quite the reverse of this. In European practice
it is not uncommon to find producer units of 1250 and 2500 h.p., and
in the United States units of considerable size are in commeicial
operation, as shown by the accompanying summary of the producer-
gas power plants operating in June 1909. There are undoubtedly
over 500 plants in operation, as the list includes 474 (Taljle l).
11 It is true that many of these larger plants are made up of
several units, but an inspection of the original data shows the follow-
ing single units of 500 h.p. or more:
h.p.
No.
h.p.
No
50C
4
1000
10
625
6
1500
1
750
3
2000
7
One single unit of 3,000 h.p. and one of 4,500 h. p. are reported, but
these figures have not been verified.
12 It is interesting to observe that about 88 per cent of the total
number of installations in the country are operating on anthracite
coal (a few using charcoal or coke) and that bituminous coal and
lignite are used in the remaining 12 per cent. It is not strange,
therefore, that the majority of plants are at present made up of
898 DISCUSSION
relatively small units, although the number of large units is rapidly
increasing as bituminous plants are becoming more common. In
point of size the bituminous plants at present average 12^ times the
size of the anthracite plants. Of the total horsepower approximately
57 per cent is derived from bituminous coal and lignite, and 43 per
cent from anthracite coal, charcoal and coke.
13 Although in large central stations there are many operating
advantages in relatively small units, yet it is believed that in the
near future central station development will demand equipment of
much larger capacity. A consideration of the fuel resources of the
country indicates that in order to keep the price of power developed
from fuel down to a consistent figure j
a|Grades of fuel which warrant transportation, or which may
be defined as "marketable," should be used with the
greatest practical economy.
h The very large percentage of^coal of so-called low grade
which today is left at or in the^mine must befutilized.
c Advantage must be taken of the large^ deposits of lignite
and peat which are found in many sections of the country.
It is undoubtedly true that in general, under conditions which do
not require the use of steam for other than power purposes, the pro-
ducer-gas power plant meets the requirements of a. At present the
only method of advantageously handling the fuels ^mentioned in b
and c is in the gas producer, and the utilization of these lower grades
of fuel on an extensive scale demands concentration of large power
units within close proximity to the fuel supply.
W. B. Chapman. In Par. 3, among|the different requirements
for successful operation, is mentioned the prevention of clinkers.
I think the formation of clinkers can^be'a voided by the prevention of
blow-holes or chimneys which allow the air to blow up through the
fire bed, making hot spots. The average temperature across the
hot zone in a producer is seldom high enough to produce clinkers. It
is only in the neighborhood of the blow-holes that a sufficient temper-
ature is attained to form clinkers. If the excessively high temperature
necessary to the formation of clinkers existed throughout the producer,
a clinker a foot or so thick would form immediately across its entire
width. When ashes are melted they tend to run together, forming
a clinker. The way to prevent this is to agitate the fuel bed contin-
ually, just enough so that the molten ash running down cannot take a
permanent set in large masses, but is constantly kept in small pieces.
BITUMINOUS GAS PRODUCERS 899
2 The successful producer should keep the fuel bed at an even
temperature and uniform density throughout any horizontal plane.
If there is a lesser density' in any particular spot, the air blast immed-
iately makes for this spot, causing an uneven temperature. To
obtain this uniform density and temperature I believe that it is neces-
sary to use some sort of mechanical agitation by hand methods, as
no man or group of men can maintain a fuel bed of uniform density
and temperature throughout any given horizontal plane long enough
to get satisfactory results from soft coal.
3 Another point is that the successful producer should be made
in a variety of sizes. The principles used in the producer described
do not seem to admit of such variety. If this producer is of large
diameter, the draft will go down the walls rather than in the middle,
and the upper zone will not get hot enough in the middle to drive the
tar out of the coal. If the tar is not removed by high heat in the
upper zone, it is sure to get to the engine.
4 A successful producer should not require a delicately balanced
draft, for the "balance" is often difficult to maintain. Uniform
density in the two zones is imperative in double-zone or balanced-
draft operation, as otherwise the draft will vacillate from one zone
to the other according to their varying density or resistance. The
density is apt to change with the loads and with change of operators.
The density will also change when the ashes are removed, as during
this process a cavern is often formed which drops suddenly. In a
producer of this type I have seen the vacuum vary from 2 in. to
18 in. in the lower or up-draft zone, and from 10 in. to 30 in. or more
in the upper or down-draft zone.
5 In Par. 21, referring to the question of varying the air supply
to the engine according to variations in the heat value of the gas,
Mr. Bibbins says: "But this variable factor has received practically
no attention and as a consequence producer operators are working
entirely in the dark." To my mind the proper way of overcoming
this difficulty would be to provide suitable mechanical means for
maintaining uniform conditions in the organization of the fuel bed.
H. M. Latham. I think Mr. Bibbins has struck the keynote in
regard to bituminous gas producers, when he says that the primary
requisites are continuous operation and tar-free gas. There is no
question in my mind that these are the most important considera-
tions. Any producer which satisfactorily meets these requirements
should have a large field of usefulness.
900 DISCUSSION
2 We have already seen from the figures presented by Professor
Fernald, that the bituminous producer is at present the predominant
type, and it seems probable that future development, especially in
large units, will be along this line. In New England the high cost
of anthracite coal suitable for use in producers of the strictly anthra-
cite type, offers serious objections to its employment as a fuel.
3 As regards continuity of operation, while it goes without say-
ing that a certain reserve power should be provided, yet it is fre-
quently convenient and desirable in installations where power is
required every day in the year, to be able to operate without calling
upon the reserve, or in other words, to run absolutely without inter-
ruption.
H. H. SuPLEE. In regard to the question of continuous opera-
tion, I think Professor Fernald will remember that we have had a
number of gas producers running continuously in this country and
elsewhere, not for one year only, but for a number of years, but we
did not call them gas producers; we called them blast furnaces.
But 1 hardly think we care to run our producers continuously.
2 In regard to the prevention of clinkers by keeping the contents
of the producer in motion, that solution was adopted in the Kitson
producer ten or twelve years ago, by means of an inclined grate which
was made to revolve slowly. As a result the contents of the pro-
ducer were kept moving up and down, and at no time did any clinker
form. The producer was discontinued, but for other reasons. The
inventor of that apparatus based it, he said, on the idea that running
water would not freeze, and that in the same way, any substance
would be prevented from solidifying by keeping it in continual
motion.
3 It must be remembered that in the operation of gas engines,
the calorific power of the gas produced is not the essential thing, but
rather the value of the charge actually delivered to the cylinder; and
this can be made almost anything which rasiy be desired, the propor-
tion of air being regulated according to the richness of the gas so as
to give a charge of practically constant heating value.
E. N. Trump. In making tar-free gas all of the valuable by-prod-
ucts are destroyed. If Mr. Bibbins proceeds to burn up the bj'-
products from the gas in the centre of his producer, he will lose from
80 to 90 lb. of sulphate of ammonia per ton of coal, which would pay
for a large part of the coal used in his producer, if it were recovered.
BITUMINOUS GAS PRODUCERS 901
2 As to continuous operation, we have had one plant burning
from 150 to 155 tons of coal per 24 hours, in continuous operation for
the past ten years; the pressure has never been oCf that plant but
once, and then for a period of two hours.
3 If the fuel bed in the producer is agitated, and plenty of steam
provided, clinkering is almost entirely prevented. Agitation can
be produced by continuously extracting the ashes at the bottom,
thus uniformly loosening the bed. Even with a very deep bed
almost no poking is required.
4 Our experience has been with Hocking Valley coal, which
will not coke. With coking coals it is more difficult to prevent the
clinkering, but the agitation by the special mechanism for removing
the ashes prevents clinkering to a great extent.
H. F. Smith. While it is of advantage to run continuously, still
in most plants it is desirable to start and stop the engines. The
majority of manufacturing plants run from eight to ten hours a day,
and it is of equal importance to be able to shut the producer down,
and to start up again in the morning with a reasonably uniform con-
dition of operation, within thirty minutes, say, of starting the plant.
Whether or not the type of producer outlined here|^is^adaptable to
meet that condition is open to question.
George D. Conlee. I would like some information regarding
the possibility of naphthalene formation by the gas producer. In
coke-oven and coal-gas practice, if the heats are suflSciently low to
prevent the formation of naphthalene, an excessive production of
tar results. Either the one or the other will be present.
2 Regarding the possibility of removing sulphur from gas by
reheating, in the manufacture of enriched water gas for illuminat-
ing purposes, the gas is passed through checker brick heated to
about IGOO deg. fahr. The gas is then scrubbed with water, cooled
and passed through iron oxide to remove the hydrogen sulphide.
The passage of the gas through the checkers seems to have no effect
on the hydrogen sulphide, though it may change some other sulphur
compounds to the sulphide.
The Author. In presenting this paper I have had misgivings
that it would be considered by some as unduly optimistic. But ]
hope that I have been absolved from that charge through the simple
showing of as complete facts as were at my command.
902 DISCUSSION
2 The producer under discussion is more or less the culmination
of experiments of many years on different types. It represents the
work of a number of engineers who have all striven for the perfection
of the bituminous type in one form or another, and I feel safe in say-
ing that the results are such as to give us some encouragement that
the problem of gasifying bituminous coal is not as hopeless as sup-
posed.
3 First let me define what is meant by continuous operation.
While I think no commercial plant should have to shut down every
fifth day to clean out, yet 365 days for the plant does not necessarily
mean 365 days for the producer. Taking conditions such as normally
exist in an electric light plant using steam boilers, we should expect
a producer unit to run at least as long without excessive labor charge
for cleaning and recharging. A small percentage of reserve equipment
is always essential, but 100 per cent is certainly not required.
4 If the producer is to stand by itself, there is no occasion for espec-
ial leniency, i. e., we should demand from the designers a grade of ser-
vice equal to that rendered by present steam plants, and from pres-
ent indications this can be obtained.
5 These high rates of combustion — 30 to 50 lb. per sq. ft. grate
area mentioned in the discussion — are interesting, but it must be borne
in mind that sometimes the amounc of coal fired includes the additional
fuel for building new fires. It is apparent from the Norton test that
a very considerable proportion of the total coal fed into the producer
was withdrawn at the end of a normal run , and if the heat equivalent
of this fuel be deducted the rate of combustion will be lowered con-
siderably. So, in comparing intermittent and continuous tj^pes of
producers, it is necessary to take this extra fuel into account, for in the
case of very frequent recharging the net loss is high.
6 The size of producer mentioned by Professor Fernald is rather
extraordinary. I think not many of us realize that 3000-h.p. pro-
ducers are being built. If it was a two-shell producer (the two rated
as a simple unit) it should hardly be compared with the single shell
producer on the same basis.
7 The sensitiveness of the balanced draft method of control has,
I think, been overestimated by Mr. Chapman. While it is stated in
the paper that the two control valves should be permanently set,
I presume it would be recognized that ihese valves are put there to
correct any inequalities or deficiencies in the fuel bed. When the
producer is properly operated the valves need little or no adjustment,
otherwise they must be adjusted occasionally.
BITUMINOUS GAS PRODUCERS 903
8 I do not quite agree with Mr. Chapman's statement that it is
impossible to maintain uniformity of the fuel beds with hand firing.
When the plant illustrated was visited I noted this point especially
bj' the aid of a simple apparatus. This is a doable poker, consisting
of a section of pipe with a solid rod through the center. By shoving
both down into the fire and pulling out the rod and covering the pipe
with a glass at the top, the condition could be noted. It was inter-
esting to see that when the top of the fuel bed appeared practically
dead, just under the surface it was at the proper temperature. 1
did not find the irregular conditions of fuel bed which Mr. Chapman
mentions and I do not think it was merely a coincidence. The ten-
dency towards short circuiting which he fears in large producers is
not as marked as might be expected, excepting with wet peat, possi-
bly owing in part to conditions.
9 As to the sulphide which Professor Rautenstrauch mentions,
I can only say that it has not to my knowledge caused trouble. I
have seen engines running successfully for a time on by-product
coke-oven gas where it was found that by keeping the rods as hot as
possible the deposition of sulphur was avoided and the consequent
corrosion of the rods. As far as I know napthalene has not created
similar trouble. A napthalene formation is characteristic of the
distillate process where the higher hydro-carbons form the greater
percentage of the heat value.
10 It is encountered in by-product coke-oven gas to some extent.
But the difficulties arising from deposition of napthalene seem to be
confined to d(;licate measuring instruments rather than the engine
valves or rods which seem to be at a temperature sufficient to dis-
sipate the accumulation. In producers the heats are run well above
the destructive point.
11 Mr. Smith seems uncertain as to the possibility of the producer
under description retaining its condition over periods of daily shut-
dosATis. Table 1 shows a period of 18 days — 432 hours — during which
the producer was entirely idle for 23-hour periods. After a night's
shutdown 15 minutes usually sufiices to bring the fire into normal tem-
perature conditions.
12 The automatic variation in the proportion of air and gas to the
engine according to the richuess of gas delivered to it is a problem of
engine design relating to regulation of mixture. Designers must
face the possibility of variations in gas from the best producers, and
I do not believe any mechanical agitation of the fueljbed will avoid
this necessity. In a plant employing a 15,000-ft. mixing holder I
904 DISCUSSION
have observed a puff of rich gas (liberated just after charging) make
its way clear through to the engine at regular intervals quite de-
stroying the mixture for the moment.
13 Mr. Trump assumes that the breaking up of hydrocarbons
occasions a serious loss of efficiency not encountered in the generators
of tar-laden gas. Just what are the precise reactions seems to be
unsolved, but in the last analysis only one factor is uppermost — the
comparative efficiency of the two systems. I doubt that much over
70 per cent is obtained from either process and less when the power
consumption of tar extracting auxiliaries is taken into account.
No. 1265
THE BUCYRUS LOCOMOTIVE PILE DRIVER
By Walter Ferris, South Milwaukee, Wis.
Member of the Society
The machine described in this paper is of some engineering interest
as the most substantial and complete railway pile driver yet produced.
Its special claims to consideration as a new development in mechanical
engineering, however, lie in the unusual arrangement and strength of
the self-propelling mechanism, and in the self-contained hydraulic
turntable, whereby the entire machine, including trucks, is quickly
lifted clear of the rails and turned end for end. The propelling
engines, mounted on the car body and delivering more than 250 h.p.,
are connected to the axles of ordinary bogie trucks without inter-
fering with the movements of the trucks in turning curves, passing
over frogs, and the like.
2 The machine was designed to meet the requirement of the
Atchison, Topeka & Santa F6 Railway system, for a pile driver cap-
able of climbing any grade on their line and hauling its own cars of
piles, tools, etc. The self-propelling pile drivers built hitherto are
capable of moving themselves for short distances while at work,
but from lack of sufficient steam capacity as well as engine power must
have a locomotive in constant attendance. The services of this loco-
motive are usually charged against the bridge department of a railway
at the rate of from $20 to $30 per day. After having used several of
the ordinary self-propelling machines, A. F. Robinson, bridge engineer
of the Santa F6 system, prepared specifications calling for a pile
driver of much higher propelling power. This resulted in the designing
by the Bucyrus Company of the machine herein described, which has
been in active service on the Santa F^ lines since January 1909.
3 The general appearance of the machme is shown in the illus-
trations. Fig. 1 shows the machine with leaders folded in shipping
position. Fig. 2 shows the leaders up ready for driving, with the
swinging frame turned across the track, and also shows how the coun-
Presented at the Annual Meeting, New York, (December 1909), of The American
Society of Mechanical Engineess.
906
THE BUCYRUS LOCOMOTIVE PILE DRIVEE
THE BUCYRUS LOCOMOTIVE PILE DRIVER
907
Fig. 2 The Leaders in Positiox for Driving with the Swinging Frame
Across the Track
908 THE BUCYRUS LOCOMOTITE PILE DRIVER
terweight on the opposite side of the swinging frame balances the
weight of the leaders, keeping the machine always in a stable condi-
tion. In this position a pile can be driven 19 ft. from the center of the
track.
4 Fig. 4 shows the machine standing on its hydraulic turntable
with all wheels in the air. In this position and without any blocking
the pile was picked up, put in place in the leaders and driven at a
distance of 32 ft. from the center of the track. It was not desirable
to drive this pile all the way in and the leaders were therefore backed
down to clear the partially driven pile. The principal use of the
hydraulic turntable, which will be described later on, is to turn
the machine end for end when there is no railway turntable or "Y"
available.
5 Fig. 3 shows the general arrangement of machinery. The car
is 40 ft. long, built entirely of structural steel and steel castings.
On the front end is mounted the swinging frame, shown in Figs. 1, 2
and 4, consisting of a pair of parallel trusses supporting the leaders at
one end and a counterweight at the other end with the necessary
parts for raising and lowering the leaders and swinging the entire
frame to the right or left at right angles to the car body. This frame
is swung by a large worm wheel, which also serves to raise and lower
the leaders.
6 The latter operations are accomplished by means of the long
worm-wheel hub projecting upward through the center pintle upon
which the swinging frame revolves, a double-grooved sheave or drum
being ke5''ed to the upper end of the worm-wheel hub. This drum is
provided with a clutch by which it can be engaged with the main base
plate of the revolving frame. When this clutch engages with the
swinging frame the latter moves with the worm wheel. When the
clutch is out of engagement, however, and a brake is applied between
the car body and the swinging frame, the revolution of the worm
wheel does not carry the swinging frame with it, but merely turns the
drum, which is keyed to the worm wheel.
7 The ropes leading from the drum to either end of the revolving
frame are so arranged as to raise or lower the leaders. The details
of the worm wheel," drum, clutch, etc., are clearly shown in Fig. 5.
This figure also shows a large circular base plate on the car for support-
ing the weight of the revolving frame. The latter is provided with four
conical rollers which rest upon the finished upper surface of the base
plate.
8 From Fig. 3 it may be seen that the leaders are mounted on a
FOLDEK No. 3.
TRANSACTIONS THK AMEHUAX StHIETY OF MECHANICAL ENGINEERS VOLUME n
THE BUCYRUS LOCOMOTIVE PILE DRIVER
Plate 3 Side and Fuont Elevations and Partial Plan of Bucyrus Locomotive Pile Driver
THE BUCYRU8 LOCOMOTIVE PILE DRIVER
909*1
910
THE BUCYRUS LOCOMOTIVE PILE DRIVER
BufjDag p, [^,
THK HUCYRUS LOCOMOTIVE PILE DRIVER
911
912 THE BUCYRUS LOCOMOTIVE PILE DRIVER
leader-raising frame by means of a pivot near the center of the leaders.
A screw and nut device takes hold of the leaders some distance below
the pivot and with this they can be inclined either to right or left so as
to drive batter piles. The arrangement for raising and lowering the
leaders acts directly upon the raising frame, which is carried by two
rolling trucks A which roll on the top of the upper chords of the
swinging frame, while the radius arm B takes hold of the lower end of
the raising frame, causing it to move in the arc of a circle as indicated.
The ropes C andZ) over the drum pass around suitable idler sheaves
and are anchored to the sliding crosshead E forming a closed circuit.
From this crosshead the raising arms F take hold of the raising frame,
transmitting the movement of the crosshead to the latter. The ham-
mer-hoist rope, pile-hoist rope and steam pipe (the last-named is
not shown) run up from the car body to the swinging frame through
the large hollow hub of the swinging worm wheel. The steam pipe
is on the center and the ropes are so close on either side that they work
equally well with the leaders in any position with regard to the car
body.
9 The main engines are 11 in. by 12 in., with double cylinders and
Stephenson link motion. From the crank shaft the two drums for the
pile-hoist and hammer-hoist lines are geared in the usual manner with
cone friction clutches. The engines, however, are much more power-
ful than would be required for these drums. The propelling gearing
consists of two inclined shafts leading from the crank shaft of the
engine to the rear axle of the forward truck and the forward axle
of the rear truck. From Fig. 3 it will be seen that each of these shafts
carries on its upper end two bevel gears, while the crank shaft carries
a sliding sleeve with a small bevel gear on one end and a large one on
the other end, the two meshing respectively with the two pairs on the
inclined driving shafts. By sHding the sleeve to one end or the other
a fast or slow propelling ratio is obtained.
10 With the fast gear, on level or moderate grades and with
moderate loads, the machine can readily be driven at 25 miles per
hour and has been driven at 30 miles per hour. With the slow gear
the engines are powerful enough to slip the two driving axles and thus
obtain all the tractive force that can be had with about 80,000-lb.
weight on drivers. The machine can thus be used effectively as a
switching engine and will readily haul its own weight with considerable
additional load over grades of 1^ per cent or more. The acceptance-
test of the first machine built was a run of 32 miles up a grade aver-
aging 75 ft. to the mile, with a maximum of 97 ft. to the mile.
THE BUCYRUS L(K'(>MOTIVE PILE ORIVKR 913
11 The lower ends of the incUned propeUing shafts shown in Fig. 3
are provided with bevel pinions. These pinions mesh with bevel
gears cast in one piece with large sleeves, as shown in Fig. 6. These
sleeves surround the driving axles, a cored hole through the middle
of the sleeves 10 in. in diameter providing about 2-in. clearance around
the axles. The sleeves are supported b}- brackets rigidly attached to
the car body with babbitted bearings. All this gearing is fastened to
the car body only and remains in line without regard to the swiveling
of the trucks.
12 The connection by which the driving torque is communicated
from the propelling sleeves to the axles is also shown in Fig. 6. It
consists of a modified type of universal joint so arranged that there is
nothing to interfere with the axle passing through the middle. The
propelling sleeve carries at one end a large flange with lugs supporting
two pins G; these pins engaging with two bronze bushed lugsi/ formed
on the inner side of the toggle casting /, On its outer side it carries
another pair of lugs J on an axis at right angles to the axis of the pins
G and these lugs / are cormected to a U-shaped driving yoke K.
The open end of this yoke is again pin-connected to a bracket L
which is keyed to the axle.
13 Both pins, G andil/, are made much longer than the lugs which
engage them, to permit end play due to the displacements of the
axle, as shown on the plan view in Fig. 6. As these two pin axes are
at right angles to each other their combined shp will take care of any
movement of translation, while the combined revolution of the parts
around the pins G, M andN provides for any possible twisting. The
wearing parts involved are six steel pins and six bronze bushings,
all of the same size, and all parts are so made that the wearing surfaces
can be replaced without taking the truck from under the machine.
The pins are made hollow and are packed for continuous lubrication.
14 The method of detaching the driving gears when it is desirable
to ship the pile driver in a freight train is slightly indicated in Fig. 3,
at the rear axle of the front truck, where an operatmg lever is shown
taking hold of the bearing which supports the bevel pinion at the
lower end of the forward driving shaft. This bearing and the pinion
are mounted in a sliding support, which enables the pinion to be
drawn out of mesh with the bevel gear, permitting the propelling
sleeves and gears shown in Fig. 6 to revolve freely with no gears in
mesh. The same arrangement is provided on the rear truck,
15 In order to provide the necessary steam capacity for these pro-
pelling requirements, the boiler required is nearly three times the
914
THE BUCYRUS LOCOMOTIVE PILE DRIVER
a
z
W
Q
<
m
Z
o
o
THE BUCYRUS 1 OCOMOTIVE PILE DRIVER 915
size of those ordinarily furnished for pile drivers. The boiler is of
the locomotive type, 54 in. in diameter, 15 ft. 9 in. long, having about
800 sq. ft. of heating surface and designed for 175-lb. pressure. This
pressure is required only for steam economy on propelling runs, as the
engines are so large that all the oidinary movements of the machine
can be made with 100-lb. pressure
16 One of the striking features of the machine is the hydraulic
turntable, which is shown in action in Fig. 4, and in shipping
position in Fig. 1. It is frequently very important that a pile
driver should be able to turn end for end or else to work at either
end indifferently. The latter plan requires that the boiler and pile-
driving machinery shall all be mounted upon a swinging deck, which
can be turned through a full circle and reach either end of the car.
This plan has bj^en thoroughly tried and is satisfactory as far as pile
driving is concerned, but makes it impossible to get a sufficiently
powerful and reliable propelling gear between the engines and the
trucks. In the new machine, therefore, the pile-driving apparatus
is momited on the car body where it can work at one end only,
thus obtaining the powerful propelling drive already described.
To reveise the machine the hydraulic lifting jack shown in P'ig. 7 is
attached underneath the car and under the center of gravity of the.
entire strucf.ure.
17 This jack consists of two ball-race castings having races about
5 ft. in diameter and is provided with 2-in. steel balls. The upper ball
race is carried upon a set of four bell cranks or levers 0, two on each
side of the car, the bell cranks being pivoted upon brackets P attached
to the main car beams. The upper ends of each pair of bell cranks are
connected by a parallel rod. while the rear bell cranks on the two sides
of the car are connected across by a heavy shaft Q. This arrangement
compels all four bell cranks to act in unison, and when they are oper-
ated by the hydraulic cylinders the four pins from which the upper
ball race is suspended move up and down the same distance, main-
taining the turntable at all times parallel to the car, even though the
center of gravity may be quite a distance away from the center of
the turntable.
18 The system of bell cranks is operated by a pair of hydraulic
cyHnders 12 in. in diameter, having about 28-in. stroke. One cylin-
der is located on each side of the car. The cylinders have trunk pis-
tons with sufficient area between the outside of the trunk and the bore
of the cylinder to provide lifting force enough to raise the turntable
away from the track and put it in shipping position. While lifting
916
THE BUCYRUS LOCOMOTIVE PILE DRIVER
THE BUO-RUS LOCOMOTIVE PILE DRIVER
917
918 THE BUCYRUS LOCOMOTIVE PILE DRIVER
the car the pressure acts upon the full area of the 12-m. piston. The
working pressure of about 200 lb. per sq. in. is provided by the boiler
feed pump.
19 The lower ball race, which is suspended from the upper ball
race by suitable clips, is also provided with a set of rail blocks s
which rest on the rails and can readily be placed under the four jack
screws, which are located in the four corners of the lower ball race.
The lower ball race also carries a circular rack, while the upper ball
race has a transverse shaft with a crank on each end and a double gear
reduction to a swinging pinion which meshes with the rack on the
lower ball race.
20 When the machine is to be turned it is necessary only to put
the. rail blocks under the jack screws and run the latter down
until they touch the blocks. The entire car is then raised by pumping
water into the hydraulic cylinders and is turned end for end by hand,
two men working on each crank. In a high wind three men may be
required on each crank. The entire turning operation occupies from
10 to 15 minutes.
21 An important incidental advantage of the turntable has
already been touched upon in Par. 4. Fig. 2 and Fig. 4 show its use
to enable the driver to reach a pile at a long distance from the center
of the track. In this way, should occasion arise, any point within 32
ft. of the track may be reached and the pile driven.
22 The tests made since the first machine was put in operation
indicate that it will fully come up to expectations. The first
machine was built with slow gear only, having a maximum speed
of 15 miles per hour. The results of its test on grades have already
been mentioned. It has since been in constant use on the western
divisions of the Santa F6 and on heavy grades. The fast propelling
gear herein described has now been added and two machines thus
equipped have been built and shipped. On one of these, built for
the Canadian Pacific Railway, the following speed test was made.
The machine hauled an ordinary passenger car from South Mil-
waukee to Racine and return, a distance of 12.6 miles each way. The
run to Racine was made in 31 min., an average speed of 24.4
miles per hour, two miles being made at a speed of 30 miles per hour.
The return run was made in 37 min., making an average speed of
20.5 miles per hour.
23 The shipping weight of the machine without the turntable,
as shown in Fig. 3, is about 147,000 lb.; with the turntable, as shown
in Figs. 1, 2 and 4, about 160,000 lb. It is equipped with either a No.
THE BUCYRUS LOCOMOTIVE PILE DRIVER 919
2 steam hammer or a 3500-lb. drop hammer, or both. The leaders are
so made that cither hammer can be used without change. The
reach for driving piles is 18 ft. ahead of the center of the forward
wheel, or 19 ft. on each side, as already mentioned; while with the
turntable, 32 ft. on either side can be reached. The leaders are 40
ft. long. The construction' is entirely of metal, except the house.
DISCUSSION
A, F. Robinson.' I feel very much pleased with the behavior of this
driver as far as we have gone. I am especially pleased with the last
three drivers, which are equipped with the extra high-speed gear.
Our men find in handling this driver that it saves a good deal of time
over the locomotive, especially in the short moves required in spotting
the pile for dri-vdng and also the short run back to the end of a bridge
to obtain piles.
2 As soon as this machine is thoroughly understood a great many
will be used. This will especially be the case when we use reinforced-
concrete piling more extensively.
L. J. HoTCHKiss.^ There are in use many antiquated pile drivers
which are slow and difficult to handle. In some cases the leaders
must be raised by means of a set of blocks attached to the track
ahead of the driver, the fall line being carried to a spool on the engine.
With such a machine ten minutes may be required to raise or lower
the leaders. Where the work is not too far from the station, and there
are no overhead obstructions, it may not be necessary to lower the
leaders when running to the station. In many places, however, the
leaders must be lowered every time the pile driver goes in, and
raised again on coming out. On a busy single-track railroad this may
cause much loss of time in the course of the day.
2 The time loss may not be merely that directly caused by slow
handling of the machine. In many locations the movement of trains is
such that there are several periods during the day when with a quickly
operated driver there is just time between trains to run out, drive one
or two piles and get in the clear again. With a driver operated as
previously described this cannot be done, as so much time is required
to handle the leaders that there is none left for driving piles. There
are, however, drivers which do not have this objection but which
'Bridge Engineer, Atchison, Topeka and Santa F6 Ry.
*As8t. Bridge Engineer, Chicago, Burlington & Quincy R. R., Chicago, II!.
920 DISCUSSION
must be handled by a locomotive. This is expensive in two ways.
There is charged to the work of pile-driving the cost of engine service,
and the locomotive is kept out of regular train service. In times of
heavy business the latter item is in itself one of considerable impor-
tance
3 The self-propelling feature of the machine described by Mr
Ferris, its large boiler capacity and the arrangement for turning it,
are its most prominent features. As stated by Mr. Ferris, the usual
charge for a locomotive and crew is from $20 to $30 per day, $25
being assumed as a fair average charge. The locomotive will furnish
steam for the driver, making a fireman on the latter unnecessary. In
the case of the self-propelling driver it is necessary to have a fireman,
and as the machine is somewhat complicated, better men must be
employed both as engineer and as fireman than would be needed ordi-
narilj'. For this reason the net saving by cutting out engine service
probablj"^ will not exceed $20 per day. It is not unusual to have from
COO to 800 piles to drive on one division in a single season. If we esti-
mate that 20 piles a day are driven, and this number is well above the
average, 30 days will be req\iired to drive GOO piles. For this period
the charge for engine service would amount to $000, which, is 5 per
cent on an investment of $12,000. It will thus be seen that the elim-
ination of engine service in pile-driving work is a matter of no small
importance.
4 A machine such as Mr. Ferris describes has sufficient po^ er and
steaming capacity to handle its own train a considerable distance.
Where a long haul is to be made the propelling mechanism is quickly
throAvn out of gear and the whole outfit put in a regular train. One
of these pile drivers recently handled a train consisting of four bunk
cars, a locomotive tender fully loaded with coal and water, one car
containing 40 tons of coal, and a way car. This train was taken up a
1 .4 per cent grade more than a mile long. A few days later this driver
hauled 140 tons in addition to its own weight up the same hill at
about 7 miles per hour. The steam gage showed 175-lb. pressure when
the top of the hill was reached
5 The conditions of railroad operation today require that all
possible economies be made both in operation and construction. The
locomotive pile driver of large capacity is a recent development and
one which must still be regarded, to a certain extent, as an experiment.
Experience so far, however, indicates that it is an economical machine,
in that it dispenses with locomotive service and is quickly handled on
all classes of work.
THE BUCTRUS LOCOMOTIVE PILE DRIVER 921
The Author. The railway pile driver is used for two general
classes of work, construction and maintenance For construction
work, in most cases, almost any track machine which is capable of
driving piles will answer the purpose fairly well, because in such work
the machine, if fairly well fixed, 's able to stand for considerable periods
of time at one place, and efficiency as a pile driver is the leading
object.
2 In maintenance work, however, which generally consists in
repairs, such as strengthening the abutment of a bridge which is show-
ing some signs of washing down, or especially in repairs after a washout,
the mobility of the machine is the leading feature. To illustrate
this point, I may say that the first machine of this design which we
built was tried out at a bridge in California which was a mile and
a half from the nearest railroad siding. I happened to be with that
machine at the time, and during the forenoon we ran it out from the
siding to the bridge we were repairing, and back into the siding again,
seven times, to dodge passing trains. During this time twelve piles
were driven, one or two at each trip.
3 The base price of this machine is $11,650 without the turn-
table and the steam hammer. As the turntable and steam hammer,
and electric light plant and other extras are added, the total price
may run to something about $14,000. This represents an increase of
cost to the railroad, above what they have been accustomed to pay
for a pile driver, of $3,C00 to $4,000 for each machine. The experiment
in the case of this machine was quite as much in the line of commercial
engineering as of mechanical engineering. When we built the first
machine we were a good many thousand dollars behind, and it was some-
what doubtful if we would get it back. It looks now as if the machine
would take very well. The operating department of the Southern
Pacific, to whicVj we recently furnished a machine, had previously
charged the bri Ige department $45 a day for the use of a locomotive,
wliich was dispensied with by the use of a machine capable of doing
its own propelling work.
No. 1266
LINE-SHAFT EFFICIENCY, MECHANICAL AND
ECONOMIC
By Henry Hess, Phii.ax)Elphia
Member of the Society
The efficiency to be treated in this paper is that of the line shaft
considered as an element for the transmission of power.
2 The complete power transmission system is made up of the shaft
and pulleys; the belts, ropes or other equivalents; and the journals
supporting all of these.
3 The difference between the power delivered to the system and
that delivered by it is consumed in the work of bending and slipping
the belts and overcoming the friction of the journals. There may
be another loss due to the bending of badly aligned shafting; but as
misalignment should not occur and as the remedy is obvious, it will
not be considered further.
4 The power lost in the bending of the belts and in their slipping
or creeping is but a small fraction of the total loss, and one, moreover,
that cannot be materially lessened; assuming, of course, that belts
are kept properly pliable and not allowed to dry out, become caked
with dust or stiffened with adhesive dressings, all causes of loss of
belt efficiency that no good shopman will allow to exist.
5 There remains the journal friction. In the average plant this
accounts for nine-tenths or even more of the entire line-shaft losses.
Included in the journal friction are the losses at the loose-pulley bear-
ings and the countershafts.
6 The coefficient of friction of plain babbitted or of cast-iron bear-
ings ranges all the way from ^ of 1 per cent to 8 per cent. This range
covers all of the many methods of lubrication in general use. The
better value is rarely realized outside of the laboratory; the poorer
value is by no means as rarely found as it should be. A showing
of 3 per cent friction coefficient is one that the manager may well pride
Presented at the Annual Meeting, New York, (December 1909), of The
American Society of Mechanical Engineers.
924 LINE-SHAFT EFFICIENCY
himself on; while a coefficient of 5 per cent is much more general,
but need not be taken, under existing conditions, as reflecting ad-
versely on attention to details.
7 The remedy obviously lies in the substitution of roller bearings for
plain bearings. In other fields than line shafting this remed> finds
considerable employment; in some the plain bearing has indeed
been superseded almost entirely. This is particularly the case where
the power efficiency is of great importance, as for instance, in the
automobile.
8 While some shopmen still doubt the reliability of the ball bear-
ing, those who have followed the development of modern machinery
know that hundreds of thousands of ball bearings are carrying loads
varying from a few ounces to many tons, day in and day out, at
speeds ranging from a few turns per minute to 10,000 or more.
They realize that it is not a question of reliability 'per se, but one of
selection of sizes suitable for the loads to be dealt with.
9 In line shafting the economic question is to the fore. The first
cost of a ball-bearing installation is greater than a plain bearing equip-
ment. Will it pay for itself by the savings effected and if so at what
rate? What return on the difference in investment can be realized?
That there is a saving is generally known, but accurate figures are
wanted by which a manager can justify his recommendation to those
who control the purse strings and are responsible for dividends.
10 When only the idle running of the line shafts is considered,
answers to these questions can be easily obtained, now that electric
motors are so generally applied directly to line shafts and it is so
simple a matter to take readings of the power delivered to them.
The difference in readings for the same shafts with plain and with
ball bearings represents fairly accurately the saving for the idle run.
11 But line shafts are not put up to run idly; they drive machines
and these machines are sometimes heavily loaded, sometimes lightly
loaded and sometimes idle. While comparative current readings
taken under these conditions may be fully satisfactory to those im-
mediately concerned, this rather crude method cannot lay claim to
that accuracy which is being demanded more and more by the engi-
neering world.
LINB-SHAPT EFFICIENCY
925
PLAN OF TESTSI
12 To supply definite information the author decided that a series
of comparative tests should be made, involving no variables other
than the bearings themselves. In order further to eliminate possible
personal bias in favor of the ball bearings the author called on Messrs.
Dodge & Day to make these tests, giving them carte blanche as to
methods, with instructions confined to a demand for definite and reliable
figures. This investigation was the first undertaken, so far as tho
Fig. 1 View Showing Line Shafting Tested
author knows, under conditions practically those of the workshop,
the sole difference being the substitution of constant loads for the
variables of ordinary working.
13 Besides the change in load due to the operation of the various
^The hangers employed in the tests were made by the Geo. V. Cresson
Company, Philadelphia, and the plain bearings used were of the regular
babbitt-lined ring-oiled type, made by the same company.
926
LINE-SHAFT EFFICIENCY
machines driven from a line shaft, already referred to, there is the
change in load due to variation in belt stress. A preliminary test
quickly demonstrated that reliance could not be placed on ihe use
of tension weighing clamps in putting on the belts. The tension was
found to differ from that determined by the clamp scales. This error
could have been minimized by the use of the admirable methods and
apparatus worked up by the engineers under our Past-President, Fretl.
Machine
Cbuntenhaff
Fig. 2 Arrangement of Countershaft Frames Used During Comparative
Tests of Ball and Ring-Oiling Bearings
W. Taylor, and this plan was given serious consideration until it was
found that the influence of varying humidity and temperature in the
shop was such as greatly to change the tension of the belts even after
they were in place.
14 The complete plan finally decided on and carried through was
as follows: A line shaft of 2i^-in. diameter and 72-ft. length used
LINE-SHAFT EFFICIENCY
927
to operate a series of heavy turret lathes was set aside for the test.
This was alternately equipped with plain ring-oiling babbitted boxes
and Hess-Bright ball bearings. In order to facilitate the exchange
-■—--V
/0^4iL-'O-
.3/ ^/".
■*ss
'•3x§2 "-/I 'x 4'^ Pulley ( Hub Ire bore x 4% long) - 343 R P Af.
'46
'^a'rs
■4x. |p (Leyiafhan) "^SrS Pulley ( Hub 1^ x 4^ long) - 2b6 ff PM.
8%^ 42--^ \'S' ^K.
'74
3x-§g "-10 X 3^ Pulley {Hub //g x 4 long) - 188 P. PM.
•4
34 """dj^ .^^ y. ^,
Jc ^4
^d " ^i "•^ ".^S" 3" '=^?
10x4^ --»0
3' /"
lSx6l
IOx(>-^ '.32X^
'3x1
-=♦?
"36
10x3 Pulley {Hub If^ bore 3 long}- 206 PPM
*72
'^1^ % "'I0x4-'a Pulley ( /f^ bore x 4^ long) - 2/2 P PA/.
*37
-3x^
^-8 A S Pulley (I fig bore x ()^ long Uneven)
zee P PM.
f.„"^" '"^
'7J-
'^""32 '''•^o'x31l Pulley (% bore x 3g length ot hub)- iSSffPM
'-S'x f^' 'lo"x 3% Pulley (if^ bore x 2^ lengfh o-fhub)
322 P. PM
2 if, bore,
<! hub H "iQ
40x122
<-2}^x72'o"LineShaff. 2/4 /fPM.
FiQ. 3 Abranqement of Line-Shapt andjBblt Dbivbs
of bearings, the ball bearing boxes^were placed on the shaft close to
the hangers, making it necessary only to slip the plain bearings out
of the hanger and slip the ball bearings in. Both types were held by
the same supporting screws usual with modem hangers.
928 LINE-SHAFT EFFICIENCY
15 It is perhaps needless to mention that great care was exer-
cised in seeing that the shaft was correctly aligned at the beginning
of each test. It was supported by ten hangers, with an average
spacing of 8 ft. See Fig. 1 to Fig. 3, the last of which shows the
arrangement adopted to secure constant load.
16 The belts from the line-shaft drive pulleys mounted on swings
hung at their upper ends, loaded by ropes attached to their lower
ends, and leading over guide pulle3''S to weights. The tension in the
belt is thus definitely determined by the weight and is independent of
slight variations of belt length from whatever cause. The load on the
journals is therefore constant and definitely known. The only pos-
sible variable is in the friction of the loose puUej'^s on the swings.
These loose pulleys were ordinary tight pulleys picked up around the
shop and arranged for oiling through the set screw holes and by the
addit-on of channels. The friction was kept as nearly constant as
possible by oiling at the beginning of every test. This answered
fairly well except for two tests in which the rather small dimensions
of the hubs gave rise to heating.under the abnormally high belt ten-
sions used.
17 Of such swings eight were employed. All were mounted on
the same side of the shaft to avoid any uncertainty in load conditions
that might have resulted from a possible balancing of pull from oppo-
site sides.
18 Speeds and dimensions of shafts, pulleys, loose pulleys and
loose-pulley hubs, belts and belt material are marked in Fig. 2. The
drive was supplied by a 10-h.p. motor from the floor. The tension of
the main belt was kept constant by a weighted idler pulley bearing
on its driving side.
19 Constancy of line shaft speed was assured by a rheostat in-
serted in the field of the 110-volt, direct-current, shunt-wound motor
and the use of a Warner tachometer connected to the motor. The
electrical measurements taken were the voltage across the motor ter-
minals and ammeter readings of the armature and field currents.
The electrical and mechanical losses of the motor were determined
from electrical resistance and no load tests. These motor losses were
deducted from the total electrical output, the balance being the power
consumed by the line-shaft system in journal friction, in loose-pulley
hub friction, in belt bending and creep, in the motor belt tension idler
and in windage. All instruments used were calibrated before and
after the tests.
LINE-SHAFT EFFICIENCY 929
METHOD OF TESTING
20 The tests were divided into two duplicate series, the first, A,
with plain bearings on the line shaft; the second, B, with ball bearings
on the line shaft. The sole variable was therefore that of the line
shaft as affected by the change from plain to ball bearings.
21 In each series the effect of varying loads was determined by
changing the belt tension by approximately equal increments from 20
to 90 lb. per inch width of single belt. This was supplemented by a
test with all the belts removed except the driving belt from the
motor, leaving only the weight of shaft and pulleys for journal loads.
22 Each test lasted forty minutes of running time; a reading of
the various instruments was taken every two minutes, making a
total of 10 hr. 40 min. with a total of 960 recorded readings.
23 The loads on the line-shaft journals ranged from 126 to 662
lb. per journal; for the 2 j^-in. by 10-in. ring-oiling babbitted bearings
this gives loads ranging from 5.2 lb. to 27.3 lb. per sq. in. of projected
area. For the ball bearings used, each of which had 12 balls of ^-in.
diameter, the load per ball was from 10.5 lb. to 55.2 lb.
24 It is pertinent to mention that these bearings have so far a
record of nearly five years constant service under loads corresponding
to about three-fourths of the maximum cited and show no evidence
of wear. In that period they have been lubricated but three times, once
when put up, once for the test and once incidental to a shop moving.
RESULTS SHOWN BY THE TABLES
25 Details of the loading of the line-shaft bearings and of the
swing or countershaft idlers are given in Tables 1 and 2. Table 3
gives the averaged electrical readings for each test. Table 4 gives
electrical readings with the power in kilowatts delivered to the belt
and the percentage of saving due to the ball bearings. In the sup-
plement to Table 4 is explained the derivations of the columns in the
table in reference to the deduction of motor losses from total input.
26 Table 4, last column, shows that the saving due to changing ten
2|^-in. plain ring-oiling babbitted bearings running at 214 r.p.m.
to the ball bearings increases with increasing belt tensions from 14
per cent to 36 per cent. "With the more usual belt tensions of good
practice ranging from 44 lb. to 57 lb. per inch width of single belt
(tests 3 and 4) , the saving amounts to 36 per cent and 35 per cent.
930
LINE-SHAFT EFFICIENCY
savoq SMI
-HVag ox ONIQNOdS
-asHOO HaaiviKj iiag
00 o ii o ;^
■o" o a> o «
1;
CD
n
h-I
10
0
z
K
•<
Bi
a
H
m
n
•<
p
2
05
n
t»
<
m
' o
H
II *"*
" H
Z
<!
H
S
1 (S
H
ll «
!i
!
«o
lO
r~
^
o
t^
00
CO
O
(N
M
lO
CD
w
CO
00
•<J<
m
00
r-
C<5
CO
U5
■»
Tt"
•<)'
■<f
n
Tp
CO
CO
00
CO
o
CO
o
o
lO
OS
lO
CO
■<*<
O
lO
CO
00
o
OJ
CO
00
00
00
CO
C<1
•*
■<:(<
■*
CO
■*
CO
■*
CO
lO
lO
to
^
O)
00
(N
o
•*
o
I-l
o
U9
•<1<
<n
O
•*
N
00
(N
o>
00
CO
M
CJ
CO
■*
CO .
. CO
CO
CO
<*
C>)
*
rai
CO
lO
UJ
lO
iO
lO
lO
^
CO
(N
■*
CO
00
>o
(N
CO
>o
00
00
00
(N
l>
a
00
(N
CO
(N
(N
CO
N
■*
lO
lO
o
■^
o
CM
^
00
,-1
lO
lO
o
lO
(M
lO
<N
CO
<N
•>*
o
00
io
1
N
CM
CO
CO
lO
CO
O O "9 O
1.0 lO o o
>0 lO O l> 00 lO
<3
05 O
r~ CO
CM CM
c
CO
CO O! O -^
»-l 05 CO CO
rH 1-1 CO M
00 "O ■* '-I
OS rt rt o
(N CO <N CO
00 O 05 CO lO o
IM U5 t^ O CM iO
e^ CM IM CO C>1 C^
CO
U3
1
lO
o
CO
.-< 00 t- o
lO lO O 00
t^ CM CO O "5 O
TH O
CO CO
CM CM
o
CO
■*
m t^ r~ lO
05 CO I> 05
rt (N r^
O -H ■* CM
iO CO 00 "5
<N <N rH M
CM ^ ■>a< lO o -H
OJ >-" CO "5 C5 1-1
rH CM CM CM -< CM
■*
00
N o o o
O 00 (N W
lO ^. U5 lO >0 lO
ai to
rt CM
o
CO
CO
CM
CO
(^ to CO 00
t^ CO IN m
rt (N '^
204
212
149
205
CO -H O t^ ■* -H
"5 t^ 05 o >o r~
i-l rt r-4 OJ .-< rt
CO
-1
Tjt
t~
00
CO
<Si
t>
00
N
CO
o
">
t-
00
o
US
O
rn
■*
CO
,_,
CO
CO
■*
00
o
<M
CO
03
00
CM
CD
o
o
lO
o
1^
CM
lO
CD
lO
CM
CO
■*
CO
V
CD
CM
'^
"^
"^
"
"
'^
^
'^
'^
''
"■
"^
*^
*"*
IM
CM
04
CO
CO
lO
t-
CO
o>
CO
CO
t~
lO
US
t~
oa
■*
us
t*
CO
•-I
CO
^
lO
05
Tjt
o
o
•*
CM
CM
,-,
CO
N
CM
O
o
i;^
■*
t^
CM
00
00
00
o>
O
'"'
^^
'"*
CM
lO
t^
00
T»l
CO
CO
CO
o>
^
^
t-
U3
t»
>o
t-
US
t^
CO
^
«
00
(N
to
«o
CO
00
IN
00
CO
t^
CM
on
o
o
1 '*>
c<
■*
CO
T»l
CO
CO
1*
CO
■*
U3
•O
CO
■*
US
us
CO
o
CM
CM
soNiHvaa XV savcj
avaQ XiATHg arai
CO CD --i
O -H
OS o
1-1 CM
b- 00
"-< CO
•o^ ONiHYag
LIiN'E-SHAFT EFFICIENCY
931
»
Pi
w
H
Z
O
o
I?
o
w
en
tn
H
«
c
H
Z
<
H
HE tag
■J >^ 2 t. H
g 2 2 S g
Bj O
lO "H (O ^ o
CO CO 00 o c^
■^ «0 00 O 'H
CO W -H c^ O)
O CO -1 P
0 o o °
01 to --I o
i'^
O "5 lO U5 O
lO >o O -H
1
«
00 00 t^ 00 t>^
C4 C4 C4 M Cvl
lO lO U3 UJ U3
00 r^ ts 00
OJ N N <N
U) lO >o to
1 J
M O lO O O
O lO (N t^
■«
rj" lO •<J< lO Tf
m Tf \n Tfi
'l *J
^ ^ ^ ^ ^
■*■*•>*' Tf
i| Q
^ T^ T)< M< TJI
•*■«<■>»<•*
Z
O
00 !0 i-O 00 lO
W U5 r- M
K
>*
d ^ d -^ C3
rH d -< -<
n
5Q CO ^ O CO
CO CO CO CO
p
CO CO CO CO CO
CO CO CO CO
S
P
§
z
O O CO o o
'H CO N CO
P-I
CO
1 t^ 00 r-^ 00 t^'
00 t^ 00 t^
E4
t^ 1- t~ r- t~
I- r- t^ r~
Z
<N CJ Cl Ol C-l
C-I Cl CS f)
■<
u
3
K5 O CI lO O
"C CI CO o
§
N
•* irj ■* lo ■*
lO ■* >0 T(<
rt
O 35 O 0-. 05
05 CI 0-. C-.
' O t^ 00 t^ t-
r~ 00 00 CO
d d 2 '^ °
^ d — —
>o
o
O . - . .
M - - - -
= = c
lO
(0
lO
><
fil
lO
i-O . . . .
>J
•*----
3 3;
J
•^
a
Ol
o
«
a;
G)
a
■*
e-i . _ . ,
J
CO - - - -
Q
CO
p
z
—
z
o
H
lO
3
i
n
?s= = = ^
O)
S S 5
a •
■"*
OI
H
Ed
e^
"5 . • . .
04 - - - -
3 3 3
a
c«
'
^
3 3 3
"
55
f.
CO O CO 00 0>
MOW
«
3
a pa
o
C3» Oi -^ CO ■^
(N -H CO
^§
-. ^ ^
„
o
3 .
"xi ^
11
<5 CO ■^ CD (N
r- >o 00
S 0
"3 ■»»• t^ CO r~
CO t- ■*
> c
c
-t
< S
932
LINE-SHAFT EFFICIENCY
TABLE 3 AVERAGE ELECTRICAL READINGS FOR EACH TEST
Test No.
Date
1908
Total
Resultant
Loads on
Linebhaft
Bearings
Lbb.
Average
Resultant Total
Pressures Kw. Taken
on Idlers from Line
Lbs.
Supply
Pressure
Volts
Average
Armature
Current
Amperes
Average
Field
Current
Amperes
lA
2/4
2349.6
111.3
1.375
111.0
10.43
1.96
IB
2/6
1.302
109.6
10.1
1.79
2A
2/4
2762
194.9
1.660
109.5
12.3
1.92
2B
2/6
1.372
109.5
10.75
1.78
3A
2/4
3261.5
277.6
1.842
109.6
14.9
1.95
SB
2/6
1.616
110.0
12.0
1.82
4A
2/4
3792.6
361.2
2.051
108.6
17.1
1.82
4J5
2/6
1.653
110.0
13.25
1.81
bA
2/4
4377
444.7
2.098
110.6
17.1
1.91
&B
2/7
1.802
109.4
14.7
1.81
6A
2/4
4977
628.1
2.238
109.9
18.4
1.92
6J3
2/7
1.933
110.0
15.8
1.78
lA
2/1
2000
0
0.479
110.0
4.00
0.366
IB
2/6
0.381
103.0
3.33
0.381
8A
2/4
2070
0
1.107
111.0
8.00
1.98
8B
2/7
0.955
110.0
•
6.75
1.93
Tests TA and IB made with 1-h.p. motor. All other tests made with 10-h.p. motor. Motor,
856 r.p.m. Line shaft, 214 r.p.m
LINE-SHAFT EFFICIENCY
933
TABLE 4 ELECTRICAL READINGS, NEW POWER IN KILOWATTS. AND PERCENT-
AGE OF SAVING DUE TO BALL BEARINGS
o in
S In * a
pi S
2; o
O s
GO
3> •! ' H :
J5 o c -
1.4 1.375 111.0X1.96 = 217 1158 11
«^■
106.6
l.O' 1147 X „ =1112 586+60
110
Cl,
Q5 !>
0.466
0.402
0.660
„ „ 100.7
IB 1.302 109.6X1.79=196 1106 10 1.0 1096 X =1016 554 + 60
108.6
oA „ 105-2
2^ 1.660, 109.5X1.92 = 210 1350 15 1.2 1335 ^ ^^^ o= 1297 579+60
' 100.4 „ „ ,
2fi 1.372 109.5X1.78=195 1177 12 i 1.2 1165 X— =1079 552 + 60 , 0.467
106.3
3^ 1.842 109.6X1.95 = 214 1628 22 1.5 1606 X — =1579 585 + 60 0934
14
29
3B 1.516 110 XI. 82 = 200 1316 14 1.2; 1302 X-- — =1218 560 + 60 | 0.598|
I 1 101.8 I i , ,,
4A 2.051 108.6X1.82=198 1853 29 1.7: 1824 X— -— =1737 I 560 + 60 I 1.117
'I 106.9
4B l.essi 110 XI. 81 = 199 1454 18 l.s' 1436 X -— ^ = 1340 558 + 60 0.722
; 108.7 1
I 104.9 , ,..
110.5X1.91 = 211 1887 29 : 1.7 1858 X =1791 557 + 60 1.154
108.8
5.4
2.098
5fi
1.802
6A
2.238
6B
1.933
7A
0.479
109.4X1.81 = 198 1604 22 1.5
1582 X ^=1487 558 + 60 0.869
107.9
35
25
109.9X1.92 = 211 2027' 34 I 1.8* 1993 X— -^ = 1940 579 + 60 ! 1.301
108.1
100.4
110 X 1.78=196 1737 25 1.6 1712 X — — = 1586 552 + 60 0 9.4
108.4
92 5
110 X0.366= 40 439 48 12.0 391 X — -^ = 369 76 + 25 I 0.268|
95 4
7B 0..381 103 X0.381= 39 343 34 10.1 309 X —-^=317 t 81+25 | 0.211
' 107 3
8.-1 1.1071 111 XI. 98 = 220 887 6 1 0.8 881 X— =858 590 + 60 0.208
SB 0.9.55 110 XI. 93X212 743 5 0.7 738 X--'„ = "l-3 581 + 60 0.072
25
21
934 LINE-SHAFT EFFICIENCY
SUPPLEMENT TO TABLE 4
For 10-h.p. motor : armature resistance = 0.04 olims, and brush contact resistance = 0.05
ohms to 0.06 ohms (2.5 sq. in.) for current densities not greater than 7.5 amperes per sq. in,;
hence drop in armature and brush contact = 0.1 ohms X armature current, also loss in amiature
resistance and brush contact resistance = 0.1 ohms X (aimature current)^.
Iron loss = 5.5 amperes X e.m.f.
856
E.m.f. with 1.6 amperes in field = (94 — 0.5)-zz7, = 94.2; with 2.0 field amperes =
sou
856
(108-0.5) g52 = 108.
field current — 1.6
E.m.f. with any field current at 856 r.p.m. = 94.2 + 13.8 X
For 1-h.p. motor (tests 7A and 7B): armature resistance = 2.5 ohms, and brush contact
resistance =0.85 ohm I- sq. in. X - I , only half of brush in contact; drop in brush contact
\16 2 J
for 4 amperes (7.4) = 2 volts, and for 3.33 amperes (JB) = 1.8 volts.
Iron loss = /o.80 + 0.1 X ggld_^"ent - 0.35\ ^^^
\ 0.7 /
E.m.f. with 0.35 amperes in field = (95-0.8 X 3.35) = 89.4 volts; with 0.42 field cur-
884
rent = (109 - 3.1) — = 103 volts.
880
E.m.f. with any field current = 89.4 + 13.6 X ^^'^ i^^°*L^l°- - .
0 7
DISCUSSION OF RESULTS
27 Tests 5 and 6 with belt tensions of 70 lb. and 83 lb. per inch
width of single belt show lower savings of only 25 per cent. This
falling off is due to the fact that the pressures were too high for the
loose-pulley hub-bearing surfaces, causing excessive heating and losses.
This reduction to 25 per cent does not indicate a smaller actual saving
due to the ball bearings, but simply that the decrease was due to
improper excessive friction in the loose pulley hubs, particularly dur-
ing the " B " runs. The pressure on the smallest countershaft pulley
bearing surface during these tests, Nos. 5 and 6, rose to 124 lb. and
148 lb. per sq. in. of projected area, respectively, which are excessive
values.
28 Tests 7 and 8 were with all the belts off and the line-shaft
journals consequently sustaining only the weight of the shaft and pul-
leys and the pull of the one driving belt. The great discrepancy be-
tween a saving of 21 per cent and 65 per cent for apparently similar
conditions needs explanation. Test 7 was made with a small 1-h.p.
motor; for test 8 the same 10-h.p. motor used for the other tests was
LINE-SHAFT EFFICIENCY 935
employed. On subsequent examination it was found that the small
motor bearings were badly in need of oil and were quite hot. A no-load
reading of this motor showed 250 watts, which dropped to 100 watts
after oiling, a difference of 0.15 kw. Deducting this from the read-
ings of 0.268 and 0.211 gives 0.118 and 0.061, the latter representing
a saving of 52 per cent which compares reasonably well with test 8.
DERIVATION OF CONSTANTS FOR USE IN ESTIMATING LOSSES
29 While the conditions of loading in this series of tests certainly
include those of general practice and it may thus be safely inferred
that the savings here shown may be generally realized, it is still desir-
able to derive constants that may be applied to any set of conditions.
30 The losses incurred are: Line-shaft journal friction; counter-
shaft journal friction; belt slip and resistance to bending; belt and
pulley windage. The last two may be safely neglected as not being
a serious percentage of the total power losses under the average shop
conditions although they may become a serious percentage under
very light loads.
31 For good ball bearings the coefficient of friction is known to be
close to 0.0015. For plain bearings the coefficient of friction may be
taken at an average value of 0.03 under good conditions. For plain
countershaft bearings the coefficient of friction may also be taken at
an average value of 0.03 under good conditions.
32 Under the conditions of this test the countershaft bearings were
replaced by the hubs of loose pulleys on the swings. With the very
primitive oiling conditions and the rather high pressures the coefficient
of friction here may be safely taken as high as 0.08.
Let L = load in pounds.
d = shaft diameter in inches.
S — shaft speed in r.p.m.
/ip = 0.03 = coefficient of friction for plain ring-oiling
bearings.
/Xj = 0.08 = coefficient of friction for loose-pulley bearings,
;«b = 0 .0015 coefficient of friction for ball beajings.
kw. = power consumed in kilowatts.
0.7467rrfL5«
1 *^
]2 X 33000
= 0.000,0059 Lrfs//
and for d = 2tV in., S = 214; kw. = 0 00308 Lfi .
936 LINE-SHAFT EFFICIENCY
33 This works out for the various total loads (Table 4) of the six
tests :
LINE-SHAFT LOSSES IN KILOWATTS
Load in pounds 2350 2762 3262 3793 4377 4977
Plain Bearings, kw 0.217 0.255 0.301 0.350 0.405 0.460
Ball Bearings, kw 0.011 0.013 0.015 0.018 0.020 0.023
34 The loose pulleys on the countershafts had various diameters
and speeds, given in Fig. 2. The sum of the products of these cor-
responding diameters and speeds is 3060. For this, kw. = 0.018 LpL .
This works out for the various average loads (Table 5) of the six tests:
COUNTERSHAFT LOSSES IN KILOWATTS
Loads in pounds 111.3 194.9 277.6 361.2 444.7 528.1
Countershaft, kw 0.160 0.280 0.400 0.520 0.640 0.760
35 Adding these countershaft losses to the plain and then to the
ball bearing losses above gives
TOTAL LOSSES
Plain Bearings, kw 0.377 0.535 0.701 0.870 1.045 1.220
Ball Bearings, kw 0.171 0.293 0.415 0.538 0.660 0.783
COMPARISON OF ACTUAL AND CALCULATED LOSSES
36 In order to make convenient comparisons of these calculated
losses with those found by measurement they are tabulated and com-
pared as follows:
TOTAL LINE AND COUNTERSHAFT POWER SAVINGS COMPARED
Tests 12 3 4 5 6
Plain bearings, calculated, kw 0.377 0.635 0.701 0.870 1.045 1.220
Ball " " " 0.171 0.293 0.415 0.538 0.660 0.783
Calculated savings due to ball bearings, kw 0 . 206 0 . 242 0 . 286 0 . 332 0 . 385 0 . 437
Savings expressed as per cent 55 45 41 38 37 36
Plain bearings, measured kw 0.466 0.660 0.934 1.117 1.154 1.301
Ball " " " 0.402 0.467 0.598 0.722 0.869 0.974
Measured savings due to ball bearings kw 0.064 0.193 0.336 0.395 0.285 0.327
Savings expressed as per cent 14 29 36 35 25 25
37 A comparison of the calculated per cent of saving with the
measured y>er cent of saving as given in the preceding table shows a
\
LINE-SHAFT EFFICIENCY 937
fair correspondence in tests 3 and 4 but a considerable divergence
for tests 1, 2. 5, 6. Now 1 and 2 are for very light loads and the
difference may probably be accounted for as due to the neglected belt
resistances and windage. As these are constant and probably inde-
pendent of the load, they are a large factor for light loads and less
so for heavier loads.
38 Tests 5 and 6 showed abnormal losses in the countershafts,
accounted for by serious overheating of the loose-pulley hubs.
CONCLUSIONS FROM TESTS MADE UNDER NORMAL BELT CONDITIO^S
39 Fortunately tests 3 and 4 were made under conditions of nor-
mal belt tensions of 44 and 57 lb. per inch width of single belt and so
indicate the following:
a Savings due to the substitution of ball bearings for plain
bearings on line shafts may be safely calculated by using
0.0015 as the coefficient of ball-bearing friction, 0.03 as the
coefficient of line-shaft friction, and 0.08 as the coefficient
of countershaft friction.
6 Wlien the belts from line shaft to countershaft pull all in
one direction and nearly horizontally the saving due to the
substitution of ball bearings for plain bearings on the
line shaft may be safely taken as 35 per cent of the bear-
ing friction.
c When ball bearings are used also on the countershafts the
savings will be correspondingly greater and may amount
to 70 per cent or more of the bearing friction.
d These percentages of savings are percentages of the friction
work lost in the plain bearings; they are not percentages
of the total power transmitted. The latter percentage
will depend upon the ratio of the total power transmitted
to that absorbed in the line and countershafts.
e The power consumed in • the plain line and countershafts
varies, as is well known, from 10 to 60 per cent in different
industries and shops. The substitution of ball bearings
for plain bearings on the line shaft only, under conditions
of paragraph a will thus result in savings of total power
of 35 X 0.10 = 3.5 per cent to 35 X 0.60 = 21 per cent.
B}'' using ball bearings on the countershafts also, the sav-
ing of total power will be from 70 X 0.10 = 7 percent to
70 X 0.60 = 42 per cent.
938 DISCUSSION
EXPENDITURE REQUIRED TO EFFECT POWER SAVING
40 While power saving is of interest and desirable the man respon-
sible for the earning of dividends will want to know what it costs to
bring about such power saving and what the investment involved
will pay.
41 A reference to the bearing cost of this test will give the answer.
Ten 2TVin. by 16-in. drop ball-bearing hangers, complete cost $212 . 60
Ten 2T6-in. by 16-in drop ring-oiUng hangers, complete cost 53 . 60
Extra investment $159.00
Value of saving of 0.395 kw. at 3 cents per kw-hr. for 3000 hr. per
year $35 . 50
(Conditions of test No. 4 representing average)
This saving represents on the extra investment 22 per cent
A closer calculation, taking into account all of the elements, shows a still better
result :
First cost, plain bearing installation, $53.60
Depreciation at 20 years $2 . 68
Maintenance Oil; ^ pt. per day at 20 cents per gal 3 . 75
Labor, 2 hr. per week at 20 cents 20 . SO
Total $25.23
First cost ball bearing installation, $212.60
Depreciation at 20 years 10 . 13
4 per cent interest on first cost difference 5.15
Maintenance:
Oil, 1 gal. per year .20
Labor, 5 hr. once per year 1 .00
Total 16.48
Difference $ 8.75
Value of power saving of 0.395 kw. at 3 cents per kw-hr. for 3000 hr . . . . 35 . 50
Annual saving total $44 . 25
Annual saving as return on extra investment of $159.00 = 28 per cent
DISCUSSION
T. F. Saltfr. It has long been conceded that appreciable power
economies were to be secured through the use of ball or or roller bearings
in place of plain bearings. The following cases show the economy
obtained by the use of roller bearings.
LINE-SHAFT EFFICIENCY
2 A Pennsylvania shoe manufacturer, with an electrically-driven
shop, found himself compelled to add considerable new equipment in
departments where the motors used were already overloaded. He
concluded that new and large motors were necessary, but before tak-
ing action, he consulted engineers who after investigation recom-
mended that roller-bearing hanger boxes be purchased and the old
motor equipment retained. One department required 68 h.p., with
babbitted boxes. The application of steel roller-bearing hanger boxes
reduced the power^consumption to 50 h.p., a saving of 18 h.p., or
nearly 24.5 per cent, and enabled the old motors- to drive the new
equipment, with a small reserve for additional equipment.
-3 A Baltimore belting company had a 4 jV.-in. bearing ^which
gave a great deal of trouble through overheating. Oil bath and water
jackets were tried with more or less success. A roller bearing was
tried, proved successful, and forty additional bearings of various
sizes were installed.
4 A wire company of Worcester, Mass., equipped their entire
plant with roller bearings and have reported a 65 per cent reduction
of friction load.
5 A friction disc transmission was designed by a New Jersey cor-
poration, the requirements being that the driven shaft revolve at a
constant speed. The diiving shaft was subject to slight] variations in
speed which were to be compensated for by automatically moving the
friction wheel across the face of the friction disc. The driven shaft
was thus required to move laterally about H in., and to rotate at
500 r.p.m. Plain bearings with sight-feed lubrication could not be
used because of their resistance to lateral motion. A special ball bear-
ing was designed to permit a free radial and lateral movement of the
shaft, resulting in an extremely sensitive and satisfactory device.
6 Roller thrust bearings are widely used wherever a thrust load or
pressure parallel to the axis of a shaft is to be carried. Practically
any combination of load and speed can be provided for. Nearly three
years ago a bearing of this type was built for a Pittsburg steel com-
pany to operate under a pressure of 1,500,000 lb. at 100 r.p.m. As a
matter of fact it carried 1,477,650 lb., applied by hydraulic pressure
of 1200 lb. per sq. in. on a 32-in. piston. There was recently delivered
to the same company a set of bearings the specifications of which
required that they be capable of carrying 2,000,000 lb. or 1000 tons at
100 r.p.m.
7 These bearings have been applied with signal success on appara-
tus such as vertical hydro-electric generators, synchronous converters,
940 DISCUSSION
frequency changers, etc., and for this work are rapidly displacing the
high-pressure oil thrusts. The advantages of roller bearings are prac-
tical indestructibility, and economy of floor space (doing away with
pressure pumps, accumulator, and a mass of piping required with
pressure thrust) ; they require little attention,
8 On an installation such as a hydroelectric generating unit, it is
difficult to carry on tests which would indicate by electrical instrument
reading the efficiency of thrust bearings. This is due to a number of
losses, the values of which it is almost impossible to determine; for
instance, the loss in guide bearings, windage, variation in load on
thrust bearing occasioned by fluctuations of gate openings, etc.
Laboratory tests have enabled the manufacture! to be reasonably sure
of the possible efficiencies which could be secured. Data obtained in
this way are not as acceptable to engineers in general, however, as
results obtained through actual practice.
9 Believing that calculations could be made which would closely
indicate the efficiency of this type of bearing, tests were made in which
the rate of flow of the oil, the temperature of the oil, and the revolu-
tions per minute of the bearing, were carefully recorded. The load
was estimated and might have varied, thus affecting results. Two
machines were tested, each test lasting about a week. Readings were
taken at intervals of ten minutes.
10 The bearings tested carried an estimated load of 140,000 lb., at
a speed of 250 r.p.m. ; the temperature rise was 50 deg. fahr. ; the flow
of oil was 111 quarts (18 . 8 lb.) per min. From the data obtained the
coefficient of friction was calculated to be 0.0016 or 0.16 of 1 per cent.
11 In the tests referred to, the heat loss, due to radiation from the
oil casing of the bearing, was calculated to be 2 per cent of the total
heat generated. Another test was made later with the oil casing
jacketed with asbestos and the results showed a difference of 2.74 per
cent.
12 These figures may be somewhat low; laboratory tests indicate
that they are. I believe, however, that with a bearing of this type
designed to meet the conditions of load and speed under which it is to
operate, a coefficient of friction of less than 0.0025 can be obtained
readily.
C. A. Graves. In tests made on something over two hundred
different line shafts in various industries, I have found that a unit
termed "watts per bearing" is best suited to making comparisons.
This unit was obtained as follows:
2 Tests were made, stopping all the machines connected to the
LINE-SHAFT EFFICIENCY
941
shafting and measuring the power required to run the motor and shaft-
ing. The main motor belt was then taken off and the power required
to run the motor free was found. The hanger bearings were counted
and also the loose pulleys over which belts were passing. The dif-
ference in power, measured in watts between the shafting running free
and the motor running free, was divided by the number of hanger and
loose-pulley bearings.
3 It developed that, on the average, loose pulleys and the hanger
bearing of about the same size took approximately the same amount
of power, so that the sum of the loose pulleys and hanger bearings was
called the "bearings." These tests were tabulated, first, by class of
industry or business, and then according to the size of the shaft. For
instance, in fifty tests in machine shops, with speeds ranging from 150
to 300 r.p.m., the average power absorbed by the shaft is 49 watts per
bearing. Other tests gave results shown in the table.
Power Consumed
No. of Tests Made
Size Shaft In.
r. p. m.
Average Watts per Bear-
ing
43
1
400
27.1
21
1§
320-400
«6.8
38
2
190 400
09.1
4
2i
200-250
108
One-Inch shaft meana i in. or l-Ar In.
4 We were fortunate in having eight different shafts equipped with
roller bearings and loose pulleys. It was found that with the shafts
running from 108 to 300 r.p.m., 22 watts per bearing were required,
with roller bearings on a 2-in. shaft. Taking the author's figures of
tests, 3 A would give 5.25 watts per bearing, while 4 A would give 62.0
watts per bearing.
5 The author might have mentioned an additional saving obtained
by using ball bearings, as smaller motors may be used to drive the
shaft, thus reducing the fixed charges.
C. J. H. Woodbury. Without questioning the general conclu-
sions of the author, I wish to inquire if the three per cent coefficient of
friction referred to in Par. 31 was derived from his experiments or
from other sources. The friction of a lubricated bearing varies accord-
ing to the temperature of the bearing and the pressure upon it. Dif-
ferent oils also give different results. With light pressures, the vis-
942 DISCUSSION
cosity of the oil plays a large part, so. much so that if the film of oil is
thick, the internal resistance from the fluid friction among the par-
ticles of this oil constitutes a large element.
2 Under heavy pressures the film of oil becomes thinner, the resist-
ance due to its internal viscosity becomes diminished and the fiictional
resistance of the whole bearing approaches a direct ratio of the pressure
upon it. In other words, the coefficient of friction becomes very nearly
constant and slightly diminishes with increased pressure as long as the
lubrication is sufficient to prevent the material of the two surfaces
from coming into contact with each other, after which the frictional
coefficient increases, although it may not reach the conditions of a hot
bearing.
Walter Ferris. The coefficient of friction of railway journals is
extremely low. Without being sure of the accuracy of the statement,
I believe it is nearly always below one-half of one per cent, and
approaches one-quarter of one per cent. Under these circumstances,
granting foi the moment the correctness of the statement, the saving
of friction due to the ball and roller bearings would have to be balanced
carefully against additional complication, first cost, and delay in mak-
ing repairs.
Fred J. Miller. The author has given no description or drawings
of the bearings. The language of the paper will apply quite generally
to ball bearings, whereas I understand that the test was made with
specific ball bearings which had been in use for five years. I think we
should have all the specific information about these bearings — includ-
ing drawings — that the author is inclined to give, and a statement of
the degree of refinement necessary in the making of the bearings in
order to get these results.^
Arthur C. Jackson. An advantage of ball bearings over plain
bearings is that the speed of the shaft can be decidedly increased, per-
mitting a reduction in the weight of the shaft and the driving pulleys,
and reducing windage and other losses. The smaller driving pulleys
will give an increased arc of contact for the belt on the driven pulley.
In my experience in driving high-speed machinery, increasing the
speed of the line shaft, which can be accomplished by the use of bal'
bearings, has a distinct advantage.
^This information is given by the author in his closure and in a paper in
The Journal of the Society for May 1910.^ — Editor.
LINE-SHAFT EFFICIENCY 943
Chas. D. Parker. The value of the ball bearing or roller bearing
seems to be conceded in a general way, but its application imme-
diately biings up the question of excessive cost, so that it is hardly con-
sidered in many cases. Data of the sort given in the paper should be
highly valuable as giving confidence to engineers in recommending the
use of ball bearings on a large scale, even though the cost may be high.
The question cannot be decided by a single experiment. Several
experiments, including tests on a shaft 400 or 500 ft. in length, would
be even more valuable, especially if made on bearings that have had a
few years' service under ordinary care.
2 It might be of interest to know whether the apparently high cost
of ball and roller bearings is due to the high cost of manufacture or to
large selling expense, which we may expect to be reduced with a more
general demand for the goods.
3 With the general introduction of electric-motor drive, the belt
drive from line shafts has become somewhat old-fashioned. However,
as the motors have large factors of inefficiency, if the efficiency of the
line-shaft belt drive can be greatly improved by the use of ball bear-
ingS; it would be of interest to know to what extent this can be done.
It would probably be shown that the older method is still the more
economical method in a great many instances.
Oliver B. Zimmerman. I would like to ask Mr. Hess if he has con-
sidered the application of ball bearings to countershafts which do not
run the same proportion of time as the line shaft. What would be the
relative return on the investment in that case, as compared with the
hne shaft itself? Furthermore, would it be advisable to lengthen the
line shaft when the ball bearings are used; for instance, in group driv-
ing, would it be an advantage to use a line shaft 90 ft. or 100 ft. in
length, as compared with a group of machines driven from 60 ft. of
line shafting?
W. F. Parish, Jr. Mr. Hess's paper brings out an important point
usually overlooked in comparative tests requiring great accuracy,
namely, the influence of temperature and. relative humidity on the
power delivered, by causing variations in belt tension.
2 For comparative tests made under workshop conditions it is
advisable to have the belts made up half of cotton and half of leather,
thereby eliminating the effect of humidity, which may cause varia-
tions of 12 per cent in the power delivered.
3 An English firm five years ago purchased a cotton belt to drive a
944 DISCUSSION
dynamo, but this belt was not equal to the speed and power required
of it, so a leather belt was substituted. It was decided to use the cot-
ton belt on one of the main mill drives, but it was found to be much too
short. So a piece of leather belt was spliced in, the whole being, when
finished, half leather and half cotton. A casing was built under it, as
it was low down and in a dangerous position. The manager was
annoyed to find that this casing had been built too close to the belt, no
allowance being made for sagging.
4 The dampness greatly affected the leather belt, as the drive was
in a low part of the mill, but the casing under the patched belt was
never altered. The length of this belt never vaiies whether the
weather is damp or dry and it is the best belt drive in the mill for steady
work. Moisture has an opposite effect on leather and cotton, leather
lengthening and cotton contracting with an increase of humidity, so
that in the half-cotton and half -leather belt the weather effect is prac-
tically compensated for.
5 In tests 3 and 4, the average saving', of power by using
ball bearings instead of ring-oiling bearings is 36^ and 35 ,■ per ^cent,
respectively, which is unusually good. It would be interesting to
know what oil was used in the ring-oiling bearings during these tests
and if the oil was new or old. With a very poor oil in the ring-oiling
beaiings the saving in power may be only partially caused by the
change to ball bearings.
6 Oil and lubrication play a very important part in the economical
distribution of power. Many power tests which I have made show
that when very poor and cheap oil is used, a saving as high as 40 per
cent can be obtained simply by using a better oil. Forty-two com-
parative power tests, made in small workshops or sections of large
shops, show an average saving in power of 19 per cent, due to the use
of a good and suitable oil. By using a good oil there will be but little
increase in cost, as it can be used sparingly, so that the yearly cost for
the better oil may be even less than for the poor oil. One test on a
machine gear-driven by an electric motor showed a power saving
of 55 . 5 per cent by using a good oil instead of a poor oil and grease.
Geo. N. Van Derhoef. In the results of the tests summarized in
Par. 41 of Mr. Hess's paper, the quantity of oil required to maintain ten
2 T6-in. bearings is given as i pint a day, or 150 pints per year,
which is ^equal ^to, 18 f, gal. , There is probably no make of self-oiling
hanger on the market today that requires anything like this quantity
of oil to maintain it. Three gallons a year for ten hangers would be
iample allowance for even the poorest make.
LINE-SHAFT EFFICIENCY 945
2 The item of labor charged is two hours a week, which is also
excessive even if the enormous quantity of oil specified were used. As
a matter of fact, three or four hours a year should be ample time to
devote to the care and attention of ten 2/y-in. hanger boxes.
3 The allowance of twenty j'^ears for depreciation would seem fair
for babbitted bearings, as probably all of us know of bearings running
in daily service for a longer period. I would like to know if Mr. Hess
has any figures showing ball bearings on line-shaft service for any-
thing like this period. As I look at the matter — and I think others
will agree with me — it is not so much a matter of a lower coeflicient of
friction as it is of the "staying properties" under practical conditions,
as distinguished from a test experiment extending over a brief interval
of time.
The Author. Taking up the vaiious points raised and the ques-
tions asked during the discussion, the author wishes to reply as follows:
2 Percentage of Saving and Actual Saving. A saving of power
cannot be intelligently considered as a percentage of the entire driv-
ing power without full knowledge of the entire conditions. A
given actual saving may be one per cent or ninety-nine per cent of a
total. The saving in line-shaft journals when referred to the line-
shaft loss is one ratio, and when referred to the total power consump-
tion, is quite another ratio. So far as I am aware the literature on
the subject quite generally refers to the line-shaft friction as a per-
centage of the total power consumption. That is misleading, since
the percentage ranges from only sixteen or so in some textile mills to
seventy or more in some of the rougher machine industries. In all
piobability the actual friction loss, bearing for bearing, does not vary
in anything like so great a degree as sixteen to seventy per cent. The
thing that is of real importance is not the ratio of the saving to a given
whole, but the actual value of the actual saving.
3 Estimating Power Losses and Savings. Mr, Graves sug-
gested that the power consumption of a bearing might be stated
from experience in "watts per bearing." Such an expression would
be convenient if it could be correctly applied; but the watt loss depends
upon the coefficient of friction, the load and the surface speed. The
coefficient of friction for a given type of bearing may be said to be
fairly well known, or at least not to vary between very wide limits.
That may also be said of the load; but the surface speed is made up of
the shaft diameter, or rather the circumference, and the angular speed,
both varying between very wide limits. So general an expression is
therefore hardly possible, nor is it necessary.
946 DISCUSSION
4 For any given installation, the shaft diameter and speeds are
known; the loads are due to the dej&nitely determinable weight of the
shaft, pulleys and^belts,^and^to the belt pull, the last-named of which
should not be allowed to exceed 60 lb. perjnch] width of single belt,
while it certainly will]rarely fall below 10 lb. The coefficient of fric-
tion for plain bearings may range from 2 to 8 per cent, with 3 per cent
a very fair and general value, and i per cent [for ball bearings. A
rise to I per cent for^ball bearings would indicate a poor quality of
bearing.
5 An actual calculation, using the known constants of the installa-
tion in question, will always give closer results than the use of any
general expression, necessarily much less accuiate, such as "watts per
bearing. " In Par. 32 the expression for kilowatts is given as
Kw = 0.000,0059 Ldfi/z
or
watts = w = 0.000,000,0059 L d s /i
which may readily be converted to the convenient form
Kwy = watts per bearing for year of 300C hours
Kwy '^ 0.000,001 Ldsfi
6 Mr. Graves has found the " watts per bearing " to range from
27.1 to 108 in 106 tests of plain bearings. The measured losses of the
test cited in the paper are under average conditions of belt pull. For
the usuaLbelt load, tests 3 and 4 show for the ten plain bearings
(see table^in Par. 33)^losses_in kilowatts of 0 .350 and 0.405, and for the
ball bearings 0.018 and 0.020, or in watts per plain beaiing 30 and 35,
and for ball bearings 15 and 18.
7 Mr. Graves' four tests of a 2yVin. Une-shaft atj200 to 250
r.p.m. may be^fairly compared with the author's tests of a 2jVin.
line-shaft at 214 r.p.m; Mr. Graves' result of 108 watts per bearing, as
against the author's of 30 to 35, shows how unsafe a general wattage
figure is. Changing the coefficient of friction from the 3 per cent
found to be approximately correct for the test cited, to 10 per cent,
would raise the 30 watts per bearing to Mr. Graves' 108 watts per
bearing. In reality the tests cited by Mr. Graves are confirmatory of
the author's, since the former range from 27 to 108, proving that the
author's values of 30 to 35 for correct belt loads and 22 to 46 for
extremely light and extremely heavy loads, represent an average of
good practice.
8 Indirect ^Economies. Mi. Graves has suggested that the
mounting of line shafts on ball bearings will reduce the sizes of the
LINE-SHAFT EFFICIENCY 947
motors required to drive the shafts. While that is obvious, the eon-
sequent economy is greater than is at first apparent. A motor must
always be selected of sufficient size to perform its work safely. As the
f fictional resistance of a plain-bearing line-shaft is apt to vary between
very wide limits — 27 to 108 watts per bearing, according to Mr.
Graves' tests — the motor must necessarily be selected to cover nearly
the maximum safely. That means a rather large motor compared with
the average useful plus friction load. Not only is there thus an
unnecessary increase of fii st cost of the motor but, more seriously, the
operating cost is unduly enhanced, as it is well known that a motor
operating much below its rated capacity has low efficiency and is
wasteful of current. When, on the other hand, the line shaft is
mounted on ball bearings, the friction load is greatly reduced, its
amount is more definitely determinable beforehand, and the initially
smaller motor is used nearer to its point of maximum efficiency.
9 Mr. Jackson refeis to a possible increase in shaft speed due to
mounting the shaft on ball bearings, resulting in decreased weight of
shaft pulleys and belts and more favorable belt contacts. All of these
elements in time make for decreased bearing loads and consequently
still further increases m economy. Mr. Jackson has had under his con-
tinual observation during several years a numbei of plain and ball-
bearing line shafts of medium and high speeds, and so speaks not
merely from theoretical reasonmg, but from actual practice and obser-
vation.
10 Reduced Importance of Improper Belt or Rope Tension.
The great variations in belt tensions that may be brought about
by weather and temperature conditions, moist and dry atmosphere,
etc., have been referred to by Mr. Parish. Both leather and cotton
belts, as well as fibre ropes, are subject to considerable variations
from these conditions. Possibly fully as im.portant a factor is the
average millwright or mechanic. The properly stressed belt is
the exception. Most belts are tightened almost to the breaking
point. The work thus lost in friction in plain bearings is directly pro-
portional to a coefficient of friction ranging from 3 to 10 per cent for
those conditions; but with the low coefficient of friction of ^ to i
per cent for ball bearings a relatively enormous over-stressing of the
belt has comparatively little influence in increasing the journal
friction losses.
11 The ball bearing is a most important factor in belt economy,
since the absence of the plain bearing friction load permits the use of
slack belts and makes for greatly increased belt life. Mr. Fred. W.
948 DISCUSSION
Taylor showed the consequent economy most conclusively m his
paper, Notes on Belting.^
12 Relative Efficiency of Direct Motor Drive and Bail-Bearing
Line Shafts. Mr. Parker refers to the large factor of inefficiency
of motors and inquires concerning the possible improvement in
line-shaft belt drives due to the use of ball bearings. While in the
early days of the introduction of direct-driven tools much was
expected from the saving due to cutting out the line-shaft friction, it
soon developed that the need for using motors equal to the maximum
demand of a tool brought in greater power losses because of such
motors working on an average at points of low efficiency.
13 Unless the direct application of the motor results in greater
convenience of handling the machine to produce a greater output, the
direct drive is not justified. In that case, the mounting of counter-
shaft, loose pulleys and line shaft on ball bearings will result in very
considerable power savings. The tests made for the author by
Messrs. Dodge & Day on line shafts showed savings of 35 per cent
under average conditions; extended to the countershaft and loose pul-
leys the savings will readily amount to more than half of the total
power consumption.
14 In line with this general question Mr. Zimmerman asks
whether it would be advantageous to lengthen a group-drive line shaft
to 60 ft. to take a larger group involving a shaft length of 100 ft.
Unquestionably that will be economical so long as other considera-
tions than those of line-shaft and line-shaft-motor losses do not govern.
As to the relative losses in countershaft and line shaft, it may be said
in general that they will be fairly equal. It is true that the counter-
shaft does not run as continuously as does the line shaft, but that sim-
ply involves a transfer of the loss from the countershaft hanger to the
loose pulleys; only when the belt is actually thro"wn off does this loss
cease; if the loose-pulley diameter is decreased, as it should be to
decrease the belt tension, the loss is lessened.
15 Ball vs. Roller Bearing. Mr. Graves makes inquiry concern-
ing the relative values of ball and roller bearings and their coefficients
of friction. The coefficient of friction for good ball bearings has
already been given as close to J per cent; for roller bearings the friction
is about double, assuming always that the rollers are kept in align-
ment and that hard and true rollers rolling on true and hardened
surfaces are used. The real advantage of the ball bearing is not
iTrans.. Vol. 15, p. 204. ^
LINE-SHAFT EFFICIENCY 949
the difference in friction, but its endurance and the consequent per-
manence of the power saving. As the correct ball bearing employs
only a single row of balls it has no length ; that at once cuts out all dis-
turbances, due to deflections of shafts or housings, that seriously
affect rollers. The readiness with which the ball l^earing is housed to
retain its lubricant and to keep out injurious grit, as well as the small
space occupied, are also advantages peculiar to it alone. The coeffi-
cients of friction cited have been determined by oft-repeated tests.
They are referred to the shaft diameter so that the values are directly
comparable with those of plain journals.
16 Reasons for Ball Bearing Cost. Mr. Parker wishes to know
whether the apparently high cost of ball bearings is due to the high
cost of manufacture or to large selling expense. Concerning the
latter it may be said that the expense of selling ball bearings is not at
all high; it is, in fact, lower than in many other lines of high-grade pre-
cision machine elements. The cost resides in the absolute neces-
sity for precision, and the character of manufacture. Ball bearings
can fitly be compared only with high-grade tools of high-grade steels.
The material is a special alloy steel, relatively high in carbon, man-
ganese, chrome and silicon; this is a combination that is very refrac-
tory under the cutting tool. After hardening, rough and finish grind-
ing cannot be forced, as that spoils the integrity of the rolling surface.
Accuracy of a high degree is essential; the unit of measurement is
the ten-thousandth part of an inch. Interchangeability of a high
order is not to be secured cheaply.
17 The data showing the saving in power consumption, not in per-
centage, but in actual consumption, that Mr. Parker asks for, are
given in the body of the paper in the table in Par . 36, on lines marked
"Plain Bearings measured kw." and "Ball Bearings measured kw."
18 Ball Bearings on Railways. This use of ball bearings is out-
side of the subject matter of the paper, but as inquiry has been
made by both Mr. Ferris and Mr. Graves it may be noted that
ball bearings of the same type are in regular use for main-line railways
and electric railways, on the axles in the former and for both axles and
motors in the latter. On the Prussian-Hessian state railways the
first of these bearings are still in use, and as the result of somewhat over
400,000 kilometers' run (250,000 miles) under standard passenger
coaches, show no evidences of wear.
19 In Europe, as well as in the United States, careful comparative
measiu-ements, ex-tending over many weeks of 2-min. observations,
have shown savings in electric railway power consumption of over ten
950
DISCUSSION
per cent, with incidental decrease in motor temperature. For main-
line and electric railway' service the direct power saving is of less
importance than the ability to take advantage of coasting; this saving
may frequently rise to 37 per cent. The chief economy lies not in
power saving, but in saving of lubricant, attendance, cost of renewals
and, in electric railway operation, the keeping of the equipment more
in service, and less in the repair shop for renewing bearing linings and
rewinding armatures that worn plain bearings have allowed to sag
onto the polepieces.
20 Type of Bearing Under Discussion. The author purposely
confined the paper to a report of results of tests made for him
Fig. 1 Elevation and Cross Section op the Hbss-Bbight Ball Bearing
by Messrs. Dodge & Day, preferring to bring out the engineering
value and economic value to be expected of correctly made, correctly
selected and properly mounted ball bearings. As Mr. Miller has asked
for information concerning the specific ball bearing involved in the
test it is proper to say that it is known in the United States as the
Hess-Bright or DWF, and in Europe generally as the DWF.
21 Fig. 1 illustrates the ball bearing proper, in cross section and in
elevation. It will be seen to consist of an inner race, an outer race, a
LINE-SHAFT EFFICIENCY 951
series of balls, all of special steels hardened throughout, and a cage or
separator for the balls. The ball tracks have curvatures approximat-
ing the ball outline, the inner track very closely, the outer track
slightly less so. The contact between balls and tracks is on a plane at
right angles to the axis of the shaft, thus providing only one point of
contact of the ball with each track. The sides of the races are continu-
ous, ^vith no interruption at any point for filling in the balls; that
ensures absolutely smooth rolling of the balls and the absence of any
possible contact with the edges of fiUing openings. In lieu of side
interruptions or filling openings for the balls, assembly is by eccentric
displacement of the two races, filling in balls through the wider space
at one side, bringing the races into concentric relation, spreading the
balls evenly and retaining them in proper position by the separator.
22 As to the refinement necessary in the making of these bearings,
to which Mr. Miller kindly refers from his own observations, 1 would
say that balls must be true to shape and to size within a limit of 0 . 0001
in. The bearing bore is held within a tolerance of 0 . 0002 in. +, and
0 . 0004 in. — . The outside diameter is held within 0 . 0006 in. + , and
0 . 0012 in. — , according to fize. The width is held within 0 . 02 in. — .
Each finished bearing is gaged for trueness of rotation with reference to
the bore, and for trueness of the outer race on the ball circle. Each race
is tested for uniformity of hardness, referred to a standard, at four
points on each side, or eight per race; the sclerescope is used for this
purpose, and that in turn is occasionally checked by the Brinnell, as
well as the Turner and the Howe hardness test apparatus.
23 Lest it may appear that these refinements are not necessary,
it may be well to say that the knowledge of their necessity has been
acqu'red at great cost; also that only to the most painstaking care in
material, treatment and workmanship is the success of the ball bear-
ing due as an every-day reliable element of mechanism. A knowledge
of proper proportions for various conditions of load, speed, shock, etc.,
is, of course, also essential.
I
No. 1266
PUMP VALVES AND VALVE AREAS
By a. F. Naglb, South Bethlehem, Pa.
Member of the Society
There has grown up a custom of requiring in waterworks pump-
ing enghie specifications that the area through the valve shall
exceed the area of the plunger by a certain amount, varying from
25 to 125 per cent. The probable intent of this clause is to obtain
a low velocity through the valve and consequent low loss of head,
and it is my purpose to demonstrate that this condition is not
reahzed in practice.
2 The above form of expression, namely, proportioning the valve
area to the plunger area, is defective because (a) it fails to distinguish
between the valve-seat area and the circumferential area of the valve
at an assumed or specified lift; (6) it leads to an absurdity unless
coupled with the length of stroke and the number thereof.
3 To the first criticism it may be replied that the engine builder
interprets the clause to mean the net area through the valve seats,
but the city's engineer occasionally requires the circumferential area.
To this the builder will not seriously object, for he simply increases
the possible lift, knowing very well that " it will never go there. "
4 The second criticism can be best illustrated by the following
example: Compare two pumps, each making 25 r.p.m., one having a
plunger 6^ in. in diameter by 60-in. stroke, and the other a plunger of
13-in. diameter by 15-in. stroke. Precisely the same volume of water
passes through the two pumps, yet the rule laid down in the specifi-
cation would require for one pump four times the valve area of the
other.
5 What is the real purpose in specifying anything at all about
valve area? Evidently the same that is sought in limiting the
plunger travel per minute, and founded upon the law that in a pump-
ing engine low velocities of water are conducive to low cost of opera-
tion but proportionately great cost of construction, and conversely,
Presented at the Annual Meeting, New York, (December 1909), of The
American Socibtt of Mechanioai. Enqinbbbs.
954 PUMP VALVES AND VALVE AREAS
high velocities imply high cost of operation but lower cost of con
struction; hence, very properly, the buyer should specify the maxi-
mum velocity he will accept.
6 Briefly, it may be said that city waterworks engines are now
quite generally limited to about 250 ft. of plunger travel per min.,
although frequent attempts are made with special designs to increase
this travel to nearly twice this amount. This plunger travel with
5-to-l connecting rod entails a maximum plunger velocity of 6.67
ft. per sec, and the head due to this velocity is 0.30 lb. per sq. in. It is
desirable to speak of the maximum plunger velocity rather than the
mean, because that governs the maximum valve area to be provided.
7 The fluid losses within a pump may be divided as follows:
a Velocity head due to plunger velocity, varying from zero
to the maximum above cited. This loss may be ignored
however, since with well-rounded plunger ends and
rounded water passages, the accelerating head of the fluid
column during the first half of the stroke is conserved by
its retarding force during the second half.
6 Friction head due to surface contact. As the main parts
of a pump are comparatively large, the velocities are low;
and the lengths of contact being short, this friction-head
is equal to a velocity-head for only about 50 diameter
lengths, and becomes so small as to be negligible.
' c Velocity head through the valves. This, whatever its
amount, is a total loss because the energy of the issuing
streams is destroyed in eddies as it enters the large valve
or pump chambers. To keep this head low is the purpose
of the specification that the valve area shall exceed
the plunger area by a certain amount.
8 Let us assume that valve area means valve-seat area, and pass
on to the study of the valve. A pump valve consists essentially of
three elements: (a) a fixed seat, (6) a movable valve, (c) a spring.
The most important of these is the spring, and yet on this point all
specifications are silent. Is this because the writer of the specifica-
tion knows nothing about the subj ect? In a general way, it is obvious
that a spring may be so stiff that on the suction stroke, where only
atmospheric pressure is available, the valve will not open at all; or
it may be so light that it will nearly float in its place and will close
only with the return stroke of the plunger. Between these two
extremes, is there not an ascertainable strength of spring which will
PUMP VALVES AND VALVE AREAS
955
allow the valve to close promptly without shock and yet require for
lifting force but a small percentage of the total pressure in the pump?
So far as I am aware, this problem has not been stated and solved in
any publication, but is left for each pump builder and user.
9 The pressure of the spring per square inch of the inside seat
area seems to me to be the force that causes the rate of flow of the
water through the valve. In my experiments of 1875 (see Vol. 10
of Transactions) with a Cornish double-beat valve, this hypothesis
did not hold, that is, the velocity through the valve was from 60 to
90 per cent greater than that due to the pressure of the valve: in
other words, the valve did not rise as high as theory would demand,
0 jjin-iin- i '"• tin. ^iii. iin. ijin.
FlO. 1 DiAORAM ShOWINO VARIATIONS OF TENSION
LINE A 18 THB ACTUAL TENSION OF A 8PBINO AT TARIOOS POINTS OF LIFT. LINES B AND 0
SHOW ESTIUATEO TENSIONS AT DIFFEBENT LIFTS WITH AN INITIAL TENSION
OF 0.4 LB. AND 0.3 LB.
but I think the deviation may be attributed to the large curvature
given the upper passage. In extensive experiments recently made by
the Bethlehem Steel Company with a large flat-hinge, or flap, valve to
be used in the Baltimore sewage pumps, the hypothesis held very
well at the beginning of the lift and fell off only about 10 per cent
at full lift. These experiments also confirmed the Providence experi-
ments in that the varying lift of the valve follows closely the varying
velocities of the plunger, except as it is modified by increased weight
or spring tension.
10 I shall therefore assume that in a flat rubber pump valve held
down by a spring: (a) the velocity of the water is that due to pressure
956
PUMP VALVES AND VALVE AREAS
per square inch of the inside valve area, (6) the area of discharge is
the net circumference of the inside of the seat multiphed by the lift.
11 The well-known formula for the velocity of flow in feet per
second is
V = 8.025 VT [1]
(a) where h is the head of water in feet. As 2.31 ft. of water 1 sq. in.
in area weighs 1 lb., the formula can be changed to
V = 12.23 Vp" [2J
(6) where p is the pressure of spring per square inch of inside area
m'
Fig. 2 Section of a Standard Make of Pump Valve
FBEE IJINQTH OF SPRING IS ly^^ IN. , NO. 12 B. W. O. BPRXNO BRASS
of valve. Tables 1 and 2, computed from Formula 2, may be con-
venient in studying this subject.
12 Springs. The springs in common use vary from 0.40 to 0.60
lb. per sq. in. of inside valve area at the beginning of the lift, and as
they are comparatively short (about If in. closed), they tighten up
PUMP VALVES AND VALVE AREAS
»57
TABLE 1 CONVERSION OF VELOCITY INTO PRESSURE
Velocity Ft.
PER Sec.
Pressure
Pounds per
Sq. In.
0.107
0.168
0.242
0.329
Velocity Ft.
PER Sec.
8
9
10
11
Pressure
Pounds per
Sq. In.
0.430
0.544
0.672
0.813
Velocity Ft.
per Sec.
12
13
14
15
Presburb
Pounds per
Sq. In.
0.967
1.135
1.317
1.512
TABLE 2 CONVERSION OF PRESSURE INTO VELOCITY
1 I , I i
_ Velocity _, Velocity t „ Velocity _ Velocity
Pounds „ Pounds _ Pounds „ Pounds „
„ Ft. per „ Ft. per ! „ Ft. per „ Ft. per
Pressure ~ Pressure „ Pressure ~ Pressure _
Sec. Sec. Sec. Sec.
0.15
4.74
0.50
8.65
0.85
11.27
1.40
14.47
0.20
5.47
0.55
9.07
0.90
11.60
1.50
14.98
0.25
6.11
0.60
9.47
0.95
11.92
1.60
15.47
0.30
6.70
0.65
9.86
1.00
12.23
1.70
15.94
0.35
7.23
0.70
10.23
1.10
12.82
1.80
16.40
0.40
7.73
0.75
10.59
1.20
13.39
1.90
16.85
0.45 1
8.20 1
0.80 1
10.94 1
1.30 1
13.94 i
2.00
17.29
TABLE 3 RATIO OF PRESSURES. VE-
LOCITIES. AND LIFT OF VALVE,
LINE A. FIG. 1
TABLE 4 RATIO OF PRESSURES, VE-
LOCITIES AND LIFT OF VALVE.
LINE B, FIG. 1
I,ift Inches
Tension
Pounds per
Sq. In.
Velocity
Ft. per Sec.
^T Start
0.60
9.47
A
0.87
11.40
i
1.10
12.82
A
1.33
14.10
* i
1.55 j
15.23
Lift Inches
Tension
Pounds per
Sq. In.
At Start i
A
i!
A
A
Velocity
Ft. per Sec.
0.40
0.58
0.73
0.88
1.03
7.74
9.31
10.47
11.51
12.41
TABLE 5 RATIO O-F PRESSURES, VE-
LOCITIES ANDl LIFT OF VALVE.
LINE C, FIG. 1
Lift Inches
At Start
A
i
Te jjsion
pou sos per
f Hi. In.
Velocity
Ft. per Sec.
0.30
0.44
0.55
0.66
0.77
6.70
8.07
9.07
9 97
10.83
958 PUMP VALVES AND VALVE AREAS
rapidly as the valve rises. Fig.l, Line A, illustrates this rate of
increase taken from a new spring. The apparently needlessly stiff
springs are used (a) to provide against the relaxation sure to occur
with all bronze springs; (b) to allow for the lengthening of the spring
as the rubber valve wears away.
13 Fig, 2 shows a standard pump valve used by a prominent builder.
These valves run from 2^ in. to 3f in. inside diameter. Table 3 is
made up from Fig. 1, with the velocities computed by Formula [2].
A larger lift than ^ in. is not generally allowed for, as the valve is not
expected to rise higher or even as high as this, and considering the
increased tension of the spring one would not expect it.
14 Other Spring Tensions. In Fig. 1, Line A represents the actual
tension of a spring at various points of lift. If the same type of
spring were made of smaller wire, its varying tensions at different
lifts would be proportional to the initial tension. Lines B and C
show these estimated tensions at different lifts with an initial tension
of 0.40 lb. and 0.30 lb., respectively; Tables 4 and 5 give the tension
at these lifts, together with the velocities corresponding thereto.
15 Valve Lift. If it is true that the spring tension governs the
velocity of water through the valve, we can readily find the lift of a
valve under specified conditions. The inside net circumference
multiplied by its lift and the velocity must equal the volume of
water passing through the valve per second. By formula, P X
V„ = C X L X v^ X N, where
P = plunger area in square inches.
Vjf. = maximum velocity of plunger in feet per second;
which is 1.60 X the mean velocity.
C = net circumference of valve seat, inches.
L = lift of valve, inches.
v^ = maximum velocity of water at Lift L, found by aid
of the diagram, Fig. 1.
N = number of valves.
or
L = -^'" ■ [3]
Contractions at sharp corners and angular turns make this calcula-
tion inexact, but the method will be found exact enough for the conj-
parisons in this paper, and is the only practical method in the present
state of knowledge on this subject.
PUMP VALVES AND VALVE AREAS
959
16 Comparisons of Valve Areas and Springs. To illustrate my
views, let us take the case of a vertical triple-expansion crank and
flywheel pumping engine, having each plunger 34 in. in diameter by
60-in. stroke, making 25 r.p.ra. — practically a 25,000,000-gal. engine.
Plunger travel = 4.167 ft. per sec, and maximum plunger velocity
= 6.67 ft. per sec. Assume the pump valve to be 3f in. inside
diameter with 5 ribs, leaving a net area of 8 sq. in., and a net circum-
ference of 10.53 in. The theoretical lift of this valve, to give the
same area on the circumference as through the ribs, would be 0.76
in. Let us assume a valve-seat area equal to 150 per cent of the
plunger area and an initial spring tension of 0.60 lb., and ascertain
the number of valves, their lift, velocities at various points and loss
of efficiencies. Then let us make the same calculations for a valve-
seat area equal to the plunger area, with an initial spring tension of
(a) 0.40 lb. per sq. in. ; (6) 0.30 lb. per sq. in. The results are given in
Table 6.
TABLE 6 LOSS OF EFFICIENGIES, ETC.
Initial
Spring
Pres-
sure
Pounds
Valve
Seat Area
Per Cent
Number
OF Valves
Lift of
Valves
Inches
Velocities in Feet per Second
Loss OF
Effi-
Plunger
Valve Seat
Valve
ciency
Per Cent
1
2
3
4
5
6
7
8
0.60
0.40
0.30
150
100
100
170
114
114
i
0-6.67
0-6.67
0-6.67
0-4.44
0-6.57
0-6.57
9.47-12.82
7.74-11.51
6.70-10.37
2.45
1.97
1.50
17 Explanations. Column 4. The plunger area (908 sq.in.)
multiplied by its maximum velocity (6.67 ft. per sec.) must equal
number of valves (170 or 114) multiplied by the net circumference
(10.53 in.), its lift L, and the maximum velocity at its highest lift.
This is a trial process, but easily found after one or two trials. Taking
the first case, we would have
908 X 6.67 = 170 X 10.53 X LX 12.82, or L = 0.264 in.
Second case:
908 X 6.67 = 114 X 10.53 X L X 11.51, or L = 0.438 in.
Third case:
908 X 6.67 = 114 X 10.53 X L X 10.37, or L = 0.50 in.
960 PUMP VALVES AND VALVE AREAS
18 Column 5. While the crank velocity may ordinarily be
taken as the maximum plunger velocity with a connecting rod
five times the crank, its maximum velocity is 1.019 times that of
the crank or, maximum velocity = 5 X 3.1416 X 25 X 1.019 -=- 60 =
6.67 ft. per sec. We can also take the mean travel of the plunger
and multiply it by 1.60 to find the maximum velocity.
19 Column 6 is self-evident.
20 Column 7. These velocities are obtained from Tables 3 to
5 and Fig. 1, and were computed in the manner already described.
21 Column 8. To get an expression for the effect of strong vs.
light springs upon the economical working of a pump, I have assumed
a pump working under a total head of 80 lb. per sq. in. and computed
the 7nean pressure required to operate the valve, calling this ratio of
pressures its loss of efficiency. A careful examination of the dia-
gram of spring compression, Fig. 1, shows that at the beginning
of the lift the sprmg did not assume its full uniform resistance. It
took nearly ^ in. of motion to tighten it uniformly. I think this
is due to the fact that these single coil-wound springs are always
a little stiffer on one side than the other, thus canting the valve to
an oblique position to conform itself to the center line of resistance
of the spring. I have taken as mean pressures of the springs slightly
more than the mean of the two extreme positions, because the
times during which the different pressures prevail are not equal.
Mean pressures were taken as follows:
Case A, maximum tension = 1.10 lb. at ^-in. lift, mean pres-
sure = 0.98 lb.
Case B, maximum tension = 0.94 lb. at i^-in. lift, mean
pressure = 0.79 lb.
Case C, maximum tension = 0.72 lb. at ^-in. lift, mean pres-
sure = 0.60 lb.
As this mean pressure exists during both strokes, it must be multi-
plied by two to find its ratio to the effective head of 80 lb., operat-
ing only during one stroke. Thus the values given in Column 8 are
found.
22 Discussion and Recommendation. A study of the figures given
in Table 6 shows that the proper place to look for " loss of head " in
a pump is in the spring tension, and not in the valve-seat area. As
long as the maximum velocity through the seat does not exceed that
through the valve, it does not add to the total loss of head. The only
PUMP VALVES AND VALVE AREAS 961
reason for having a large number of pump valves and a large inside
diameter is to keep the lift down, basing judgment on the number of
reversals per minute.
23 We have seen that a 3|-in. valve, having a net area of 8 sq.
in., needs 0.76 in. lift to give the same area at the circumference as
through the seat. If there were no rib obstruction, 25 per cent of the
diameter of a circle gives the height to which a valve must lift to give
a circumferential opening equal to its area. Because of the ribs, we
need but 20 per cent of the diameter for the lift; and with the lightest
spring C, a maximum lift of ^ in. or 13 per cent of the diameter was
sufficient to discharge the required volume of water. It will be good
construction to limit the lift of a pump valve of this type to, say,
15 per cent of its internal seat diameter. The spring will not allow
it to rise to that height, but it is a safe limit for a stop.
24 The place to begin the study of the proportions of a pump is at
the spring of the valve. Make a sample spring of such diameter and
length and strength as you may think desirable, and by experiment
construct a diagram of its rate of compression, as in Fig. 1. Now
you can find the maximum velocity at an assumed lift and proceed
in the manner already pointed out. The spring would be improved,
that is, it would not tighten up so rapidly when compressed, if it
could be made somewhat longer than present practice, but this is not
practicable, as it would enlarge the valve chamber, where the valve-
cage design is used.
25 The largest and weakest castings within a pumping engine
are the valve chambers and anything that can be done to reduce
them to the minimum size permissible is good engineering. I think
the line of study I have pursued will indicate that the total valve-
seat area in this type of engine need not be more than the plunger
area. That rule, if adopted, would reduce the diameter of the valve
chambers an appreciable extent, probably 10 per cent, and this
is well worth saving.
26 The number of valves saved by the construction recommended
(about 33 per cent) is also worth while. No loss whatever would be
entailed and a part of the money saved could be expended in mak-
ing a better spring. I would make ihe spring of steel, if possible
oil-tempered, and protected against corrosion by copper electro-
plating. Then I would have all springs tested and brought to a
like tension under a rigid specification. With these improvements
I believe that a little better pumping engine than we now have
could be obtained at a little smaller cost.
962 DISCUSSION
DISCUSSION
Charles A. Hague. The practice referred to by the author,
of specifying that the area through the pump valves of waterworks
engines shall bear a certain relation expressed in percentage of plun-
ger area, is becoming less frequent, and it is to be hoped that it will
finally be disregarded altogether. The relation between the plungei
areas is merely incidental, because the valve area is a function of the
quantity of water to be handled, the important matter being the
velocity of the water through the valve seats to fill the plunger
chamber as nearly complete as possible under the conditions.
2 The total valve area, or total area of valve-seat opening, ought
to depend upon the velocity needed to pass the required quantity
of water in a given time. Some authorities advocate a velocity not
to exceed 3 ft. per sec, others 4 ft. per sec. and some as low as 2^
ft. per sec. Two factors are to be considered, as follows :
3 First, as to the^lift of the valves. The lower the pressure,
and the lower the speed of the engine, the higher the valve may lift;
on|the contrary, the higher the pressure, and the higher the speed
of the engine, the less the valve may lift, if a smooth, easy running,
economical engine is desired.
4 Second, regarding the circumferential area of the valve space,
orj the area ^of [the space around ^the edge of a^ disc valve, when
it is open or off its seat. This is [a factor f that need not be very
seriously considered, because the water, having succeeded in getting
easily through the grating formed by the^seat,|will meet with very
little resistance in]^moving out from^under the jvalve. Valves free
to lift to an unchecked height ^will often get so far away from
their seats that slamming will take place at the reversal of the plun-
ger. A pumping engine will work best when provided with sufficient
valve-seat area to keep the mean velocity of the water down to about
3 ft. per sec, the lift of the valves being so restricted that they will
return to the seats when^the plunger approaches the end of its stroke.
5 With reference to^plunger travel in conjunction with pump
valve area, mentioned or implied in^the^ paper, the vital question
is, How shall we^obtain any certain plunger travel per minute: by
a short stroke at^many revolutions per^minute, or, by a long stroke
at few revolutions per minute?
' 6 After the water is well'started through'^the pump valves, a
larger increase in speed would be permissible than is found in prac-
tice, if it were not for the reversals at the end of the strokes. The
PUMP VALVES AND VALVE AREAS 963
250-ft. per min. plunger travel mentioned in the paper would be
permissible with a 60-in. stroke at 25 r.p.m., or better with a 72-in.
stroke at 21 r.p.m. The pump valves would work in a very satis-
factory manner, the pumps would give very good hydraulic efficiency
and the engines would run smoothly. But if we should attempt to
obtain 250 ft. per min. with a 30-in. stroke at 50 r.p.m. there would
be a great reduction in economy, smoothness of running and general
efficiency.
7 The items in Par. 7 are all within the scope of mechanical effi-
ciency, and will be reasonably well taken care of, if the valve factor is
properly attended to. The most effective method for dealing with
the question of valve area is to establish a certain satisfactory area
per unit of pumpage, at some definite minimum rate of revolution as
a standard. Then, for every revolution per minute above the stand-
ard rate, add a certain per cent to the standard valve area. This
will give an engine of more revolution^; a greater proportionate
valve area than a slower machine, thus in fast engines keeping the
valves nearer to their seats than in slow ones.
8 In Par. 25, the author makes a statement, with which one feels
compelled to take issue, that, "the total valve area in this type of
engine need not be more than the plunger area.' ' As already pointed
out, there is no necessary relation between the valve and the plunger
area at all. The relation is only incidental^ or whatever it happens
to be after the proper proportions are established. A certain area
of plunger, with a certain stroke, at a given number of revolutions
per minute, sets up a certain velocity in the water through the valve
seats. A plunger of half the area, with the same stroke and at twice
the revolutions per minute, will set up the same velocity of displace-
ment, and consequently the same mathematical velocity will be re-
quired through the valve seats; although the increased frequency of
opening and closing will introduce another element for consideration,
which will call for a greater proportionate valve area, for the greater
number of revolutions per minute. In other words, a larger plunger
running slowly will require the same valve area as a smaller plunger
running faster, so' far as the calculated displacement and velocity
are concerned. The valve area in both cases depends upon the quan-
tity of water and the selected velocity through the valve seats,
regardless of the size and speed of the plungers.
9 The spring diagram and expressions are very nicely worked out,
but the differentiation is too fine for real work, and could bejfor the
most part avoided by keeping the* valves closer to their seats and
964 DISCUSSION
avoiding refinement in springs. The idea is to get away from the
laboratory engine, determine the conditions to be found in a pumping
station, and then meet those conditions as they really exist, rather
than try to adjust the working conditions to some real although
impracticable refinement in some particular factor.
10 In many pumping engines now at work, some of the details
worked out very nicely on the drawing board but failed to meet the
actual requirements. There are waterworks engines of the cage
\'alve construction, in which the ends of the valve stems, with valves
exactly like those shown in the paper, have been sawed off, the valves
being kept in place by means of wooden wedges, just because some-
one who never saw the inside of a pump after it left the shop did not
understand the requirements involved in the care and maintenance
of the machine. In one or two such cases, the cages were difficult
to remove and there was not room enough to remove the valves,
with the cages in the pump chamber, by the regular method of taking
off the spring guard.
Irving H. Reynolds. Mr. Nagle calls attention to two very
common errors which purchasers of pumping machinery fall into
when preparing specifications:
a The absurdity of specifying the ratio between plunger
and valve area without other limiting clauses.
6 Specifying an unnecessarily large amount of valve area.
Mr. Nagle suggests as a remedy for the first, specifying velocity
through the valves rather than a percentage of plunger area, and
for the second, the use of lighter springs, thus enabling the valves
to rise to their full lift and thereby reduce the number of valves
required.
2 In regard to the first, there is an increasing tendency among
engineers to specify a maximum velocity of flow through the valves
rather than their area relative to the plunger.
3 Quietness of operation rather than cost is the first considera-
tion in the design of pump valves, and the present excessive valve
areas have grown from this idea. Time is also an important element
in determining pump valve action; therefore, the number of rever-
sals or valve seatings, rather than the piston speed, is the important
factor, and consequently valves of small diameter and therefore of
relatively low lift, have displaced the large diameters in common
use a few years ago.
PUMP VALVES AND VALVE AREAS 965
4 To decrease further the lift of the valves and, therefore, per-
mit them to close quickly and quietly at high speeds, valve areas
have been increased to a point where in actual operation the valves
lift only a fraction of the theoretical height to which they should lift
to give a full opening; in other words, large valve area is provided
for the purpose of not using it.
5 If on a high-speed (high-revolution) pump the valves were
fitted with light springs, permitting them to lift to their full height
as suggested by Mr. Nagle, it is probable that the pump would be
exceedingly noisy, as the valves would be so far from their seats at
the time of plunger reversal that they would not seat until the flow
through them had reversed, and this slowness in seating would be
still further aggravated by the light springs employed. There is
no doubt, however, that in many cases the springs used are unneces-
sarily stiff and on slow-speed engines the lighter springs would be
found satisfactory.
6 In earlier practice, particularly with direct-acting pumps, the
valve area was small in proportion to the plunger and the valves
were obliged to lift nearly to their full height. In this type of pump,
as the plunger speed was relatively high to nearly the end of the
stroke, the valves became noisy if the pumps were operated at high
speed.
7 With the general introduction of the crank and flj^vheel pump
came higher rotative speeds and the necessity for larger valve area
and smaller valve opening, i. e. lower lift, until present practice has
crystallized at velocities of 3 ft. to 3i ft. per second through the
valves, and valves of between 3^ and 4 in. in diameter for ordinary
waterworks service. In general the best results would be obtained
if engineers in drawing specifications would limit the mean velocity
of water through the valves at about 3 ft. per second and the diameter
of the valves to not over 4 in.
F. W. Salmon. I prefer to make these valves somewhat different
from the one illustrated in the paper. I do not believe it is best to
use the radial ribs of the valve seat to screw it in, but that it is bet-
ter to cast small projections on the outside, as at A, Fig. 1 here-
with. This part is of such a size that an ordinary black pipe will
fit neatly when properly milled out at the end, thus making a
good socket wrench at a minimum cost.
2 I prefer to put a brass plate on the top of the rubber valve, as
shown at B (Fig. 1) , and partially to punch out and turn up little pro-
96t,
DISCUSSION
jeciions from this plate as at C (Fig. 1 and Fig. 2.) The plate pre-
vents the spring wearing into the top surface of the valve, and the
projections keep the spring properly centered.
^3 to 0 M-ebs or ribs
Fig. 1 Cross Section op Pump Valve, Showing Improvements Suggested
BY Mr. Salmon
3 Small projections should be cast on the under side of the spring
guard as shown at D (Fig. 1 and Fig. 3), the latter being the under side
of the spring guard. If the valve is ever drawn so high as to come into
Fig. 2 and Fig. 3 Showing Projections on Brass Valve Plate and
ON Spring Guard
contact with these projections it will still descend freely, not being
in the least hindered by the soft surface of the valve forming a close
contact with the smooth under-surface of the spring guard, as it
PUMP VALVES AND VALVE AREAS 967
is sometimes made. I consider that this is useful in cases of fast
running pumps, as in such machines it is particularly desirable to
have the valves seat while the crank is passing the dead center, and
so a quick closing action is required.
WiLLLAM Kent. I hope Mr. Nagle will supplement the paper by
telling us what proportion of valves and valve springs he would use
for certain conditions. The paper is now largely one of criticism,
and I would like to have the author make it a constructive paper.
Par. 24 reads "The place to begin the study of proportions of a pump
is at the spring of the valve. Make a sample spring of such diameter
and length and strength as you may think desirable, and by experi-
ment construct a diagram of its rate of compression, as in Fig. 1."
This is good advice for pump designers, but other mechanical engi-
neers are called in to confer about these points, and if Mr. Nagle
would tabulate the proportions of springs suitable for pumps, and
give the lifts at certain velocities of water, his paper would be more
useful to these engineers.
2 The author criticises the practice of specifjnng the percent-
ages of area of the valve and the pump. I see nothing very wrong
in that, provided the plunger area and the speed are also specified,
as is usually done, otherwise some of the bidders will put in a small
pump. In order to compel them to supply a pump large enough,
we limit the velocity of the plunger; and having limited the velocity
of the plunger and specified its size, we may as well say that the valve
must be so many per cent of the plunger area, as to state what the
velocity of the water mustfbe. The specification is good enough,
provided these additional items of plunger area and speed are also
specified.
Prof. R. C. Carpenter. It is quite evident to any one familiar
with hydraulics that the difficulties arising from the lessening of the
valve area are largely inherent in the spring. If a spring could be ob-
tained which would open uniformly with increase of pressure the
troubles due to certain inertia effects which are mentioned, would dis-
appear. Thi-<, however, merely points out the source of trouble and
leaves the question open as to what shall be done. '*' In substance,
defects are merely pointed out without remedies. I would suggest
that Mr. Nagle, if he can, give some of these remedies for the
troubles which he has described.
968 DISCUSSION
_jE. H. Foster. Attention should be called to the fact that this
paper refers to the valves of one type of pump. Many pumps of
other types are built, particularly those without fly wheels, to which
it is not absolutely necessary that these rules should apply. It 's
well known that a considerable pause at the end of the stroke of
the duplex pump facilitates the closing of the valves, so that these
empirical rules for lift and area must be quite different for that type
of pump.
The Author. Some new matter which has come to the attention
of the writer of this paper, is appended herewith. A careful study
of this will answer most of the points raised in the discussion, espe-
cially the point made by Mr. Reynolds and Mr. Hague, to the
effect that the maximum velocity through the valve should be limited
to 3 or 4 ft. per sec. It can be assumed that Formula [2] of the
paper and Table 1, quoted from experiments of Prof. Karl Bach
of Dresden, governing the relation of spring pressure and velocity
of flow to the coefl&cient of contraction, are correct.
2 Let us apply Formula [2] to the 3 ft. per sec. assumption. For
a lift of, say, 0.20 X diameter,
y2
p = 0.77 X — or = 0.07 lb. per sq. in.
100
of inside seat area. Such small spring pressure is out of all propor-
tion to what common practice has established, which is from 0 . 30 to
0.60 lb. at the initial and 0.75 to 1 .50 lb. at the full lift. The for-
mula for the resulting velocities is very simple. Suppose we solve for
four spring pressures of, say, 1 . 50, 1 . 25, 1 . 00 and 0 . 85 lb. at full lift,
and 40 per cent, or 0.60, 0.50, 0.40 and 0.34 lb. at the initial point,
at a lift of 0 . 20 X diameter, the formula would be
v= i.uVioox p
and the velocities for
p = 1.50 lb. V = 13.96 ft. per sec.
1.25 1b. V = 12.74 ft. per sec.
1.00 1b. V = 11.40 ft. per sec.
0.85 lb. V = 10.51 ft. per sec.
The coefiicient of contraction would be 53 per cent in each case.
PUMP VALVES AND VALVE AREAS
969
ABLE 1 PROFESSOR BACH'S EXPERIMENTS WITH A FLAT VALVE AND A FLAT
SEAT (SEE FIG. 1)
Inside diameter of valve seat d = 1.968 In. Outside diameter of valve di - 2.362. Ratio of Inside
and outside areas, 1 to 1.44. Inside area, 3.04 sq. In.
F-.1.21
-1.29 ft.
iJ - 3.08 -
- 3.11 ft.
1
M = lift of Valve, In
0.23
2.028
0.55
2.218
6.548
1.29
4.97
1.01
2.304
8.554
1.27
6.46
0.122
4.610
3.086
3.11
2.33
0.40
5.073
6.768
3.10
6.17
0.65
5.238
11.20
3.08
8 46
2
3
G = Weight of valve, lb
Q = Volume of water, lb. per
4
5
H «= Head of water, ft
ir = Velocity through seat,
ft. peraec
1.29
CaHevlations hy Nagle
= Ratio lift, to diameter. 0.12
d
G = p Weight per sq. In., lb.
persq.ln 0.666
V = Velocity due to p, ft.
per sec 9.55
V =- Velocity due to H, ft.
persec 9.12
Ratioof„- 1-04
^h
0.28
0.51
0.06
0.20
0.33
0.728
0.760
1.516
1.688
1.723
9.47
9 07
14.70
14.72
14.35
9.12
3.92
14.14
14.10
14.07
1.04
1.02
1.04
1.04
1.02
Line 7 Is obtained by dividing the weight O, given In line 2, by the area of d, or 3.04 sq. In.
m
Line 8 Is obtained by the aid of Table 2, where opposite the value of , is found the coefficient and
a
formula. For example: taking the first case of a lift of 0.23 in., or a percentage of 0.12 of the diameter
1.968 in., we find by Interpolation In Table 2, the formula, 7 = 1.17)/ 100 p, or F = 1.17 X 8.16 = 9.55
ft. per sec.
Line 9 Is obtained from the fundamental hydraulic formula V — 8.025 ]/ H, when H Is the head in
feetand F the velocity in feet per second. For example, in the first case cited we h&ve,H = 1.27 ft.
V = 8.025 )/ 1.27, or = 9.12 ft. per sec.
Line 10 is self-explanatory, and is Introduced as a check upon the work and formulae, as if correct,
It should be unity. The slight deviations are due to the various declmala not being carried farenough,
but they are carried far enough for all practical purposes.
TABLE 2 ORIGINAL DATA AND RESULTS OBTAINED
■ '--
0.05 0.10 1 0.15 0.20
1
0.25
0.30
0.35 0.40
1
0.5d
2 M =
5 p =
0.67
1.22
1
0.65 0.60 0.56
0.69 0.72 0.74
1.20 1.18 j 1.16
0.53 0.50
0.77 0.80
1.14 1.12
0.47
0.83
1.10
i 1
0.44 0.41 '0.37
0.86 0.89 0.92 ^Qg
1.08 1.06 IMy/ioOp
Line 1 — is the actual rise, or lift, of the valve, divided by Its inside seat diameter.
d
Line 2 u la the coefficient of contraction at the point of discharge with a given lift.
Line 5 p is the pressure in pounds per square Inch and Is found by dividing the weight of the valve
(In water), plus Its spring pressure In pounds by Its inside seat area In square Inches.
Line 6 z is the velocity of the issuing stream at the point of discharge In feet per second.
970
DISCUSSION
3 It is plain, therefore, that we are far from realizing four feet per
second with our present spring practice.
4 To Mr. Reynolds: The writer did not mean to lighten the
springs abnormally, in fact, 0.45 lb. to 0.50 lb. initial is probabl}'^
light enough, but if they could be made somewhat longer, so as not to
tighten up too rapidly, it would seem to be desirable.
5 To'^^Mr. Kent: The formulae given by Professor Bach are a
very great addition to our knowledge of pump-valve action. Within
the limits prescribed, we have now a safe guide for valve construction.
What it should be for other numbers of revolutions and plunger veloc-
Mol-
ii'i
u m^
J
I
Fig. 1 Apparatus used by Professor Bach for Determining
coeffcient of contraction
ities, I am not able to formulate. Prof. Herman Haeder of Duis-
burg goes into that phase of the problem, but as his theory is
not confirmed by extensive experience, I have not taken it up in
this paper.
6 In Par. 15 of the paper is given a formula for ascertaining the
lift of a pump valve, from which was omitted, as stated, the coefficient
of contraction. Not knowing the value of this coefficient with cer-
tainty, the writer hoped the information would be supplied in the dis-
cussion. The omission was not referred to, however, and he is now
able to supply it himself.
PUMP VALVES AND VALVE AREAS
971
7 In a German book on pumps and pump valves' by Professor
Haeder, the subject; is treated in an exhaustive manner. The
actual coefficients of contraction are given, with the results of impact
upon the valve, based upon experiments by Professor Bach. In what
follows reference is made only to that part which bears on the sub-
jects of flat valves and flat seats, of which the inside and outside
areas bear the ratio of 1 . 00 to 1 . 44. The notations were originally
in French, but in what follows have been transformed into English
units.^
Fia. 2 AND Fig. 3 Showing, Respectively, a Valve Open and a Valve
Closed. The Formulae for these two Positions are as Follows:
^V = Velocity in Feet per Second P= Pounds per Square inch
Valve Closed
V = 1.22 VlOOP
P = 0.67 ^
Valve Open
V = 1.04 VlOOP
P - 0-^ 100
8 Fig. 1 shows the apparatus used by Professor Bach. Table 2
(Haeder 261) gives the original data and results obtained and some
calculations of my own, the better to elucidate the subject.
9 Fig. 2 and Fig. 3 (Haeder 110 and 110a) show a valve closed and
one open, with the respective formulae for the two positions of the
valve, giving the values for velocity or pressure in the two extreme
positions. "Open" signifies a lift of one-half the diameter, which,
needless to say, is far beyond American waterworks practice.
10 Table 2 (Haeder 213) gives the values of v and p for the inter-
mediate positions of the valve, and also the value of "u", the all-
important coefficient of contraction at all positions. Use this table
to ascertain the velocity v of the water through the valve opening and
also the coefficient of contraction u at the same point.
^Pumpen und Kompressoren, Haeder, Duisburg.
* The original tables in French units can be referred to in the author'^ manu-
acript on file in the rooms of the Society. — Editor.
972 DISCUSSION
1 1 We can now say that we have a practically correct formula for
ascertaining the volume of water discharged through a flat disc pump
valve of a certain diameter, an assumed lift, and a certain tension of
spring.
12 Throughout all the following calculations a maximum lift of
valve of 0.15 d, is taken, leaving the reader to make for himself other
assumptions of lift and the consequent calculations. Various ten-
sions of springs will be taken, to illustrate the importance of giving
more attention than heretofore to the strength and length of springs.
13 Take, for examples, the same dimensions of pump and valve as
those used in Par. 16. Formula [3] (Par. 15) would now be better ex-
pressed in terms of N, the number of valves, than assuming the number
of valves and solving for the lift L. The formula would then read,
P X V
CX LXuXV^
Applying this formula to the three different strengths of springs
before used, we get the following results :
14 First, ascertaining the velocity through the valve by the aid of
Table 2, the spring tensions were as follows;
Case 1 : initial, 0 . 60 lb. ; final 1 . 55 lb. per sq. in.
Case 2: initial, 0.40 lb.; final 1.03 lb. per sq. in.
Case 3: initial, 0.30 lb.; final 0.77 lb. per sq. in.
The formula for the velocities due to these final pressures at a lift of
0.15 d, are
Case 1: F = 1.16 ^100 X 1.55, or F = 14.41 ft. per sec.
Case 2: F = 1.16 ^100 X 1.03, or F = 11.77 ft. per sec.
Case 3: F = 1.16 VlOO X 0.77, or F = 10.18 ft. per sec.
The coefficient of contraction in each ease is w = 0 . 56. Substituting
these values in Formula 4, we have,
= 127
Case 1 :
N =
10.53 X 5.625 X 0.56 X 14.44
Case 2*
N =
^^^^ - 155
3.317 X 11.77
Case 3*
N =
^^^^ - 180
3.317 X 10.18
PUMP VALVES AND VALVE ABEAS 973
15 Let us make a similar calculation for springs of the same initial
strength, but longer, so that they will tighten only one-half as much in
their nine-sixteenths lift. Then the first spring final tension becomes
1.08 lb., the second spring 0.72 lb., and the third spring 0.53 lb.; and
the velocities become
Case 1:7= 1.16^100 X 1.08 = 12.05 ft. per sec.
Case 2: V = 1.16^100 X 0.72 = 9.84 ft. per sec.
Case 3: V = 1.16\100> 0.56 = 8.68 ft. per sec.
and solving for N in formula 4, we have
r^ 1 -KT 6056 ,_„
Case 1: A'^ = = 152
3.317 X 12.05
Case2:iV = ^^ • = 186
3.317 X 9.84
Case Z:N = ^^ = 210
3.317 X 8.68
16 To calculate the loss of efficiency for these different springs let
us take the mean pressure on the springs to be the initial, plus two-
thirds of the increase, and twice this for the two strokes; and this sum
must be divided by the total pump head, say 80 lb., to obtain the loss
of eiSciency. We would then have
Case 1: [0.60 + (1.55 - 0.60) f ] X 2 -f- 80 = 3.06 per cent
Case 2: [0.40 -f (1.03 - 0.40) |] X 2 ^ 80 = 2.05 per cent
Case 3: [0.30 + (0.77 - 0.30) |] X 2 -^ 80 = 1.50 per cent
With stronger springs, we would have
Case 4: [0.60 -f (1.08 - 0.60) f ] X 2 ^ 80 = 2.30 per cent
Case 5: [0.40 -f (0.72 - 0.40) f] X 2 ^ 80 = 1.50 per cent
Case 6: [0.30 -h (0.53 - 0.30) f] X 2 ^ 80 = 1.15 per cent
Grouping these figures for better comparison, we have Table 3.
17 We have now, in Table 3, figures which enable us to study pump
valve constructions in an intelligent manner. The formute given
enable us to construct a similar table for any other assumed dimension
of plunger and its velocity, height of lift of valve, or spring tension.
18 In conclusion the writer wishes to sa}-- that now^ for the first
time in the history of the modern high-duty pumping engine, we have
a formula for designing a pump valve that is scientifically correct, and
974 DISCUSSION
one based upon hydraulic experiments carefully made by a competent
authority. The subject seems important enough to bear repetition
in grouping the previous instructions, as follows :
19 Find the area of the plunger in square inches, and the maximum
speed of the plunger in feet per second. The latter is found by mul-
tiplying the stroke in feet by the maximum number of revolutions per
minute, multiplying this result by 1 . 60, to reduce it to. its maximum
velocity (the crank velocity), and dividing by 60 to reduce it to feet
per second. This product, algebraically expressed by P X V^ in For-
mula [4], becomes the numerator of the equation.
20 Determine the size of the pump valve-seat and its net area
between the ribs, whether the valve bears on the ribs or not; that will
be the inside area of the valve against which the impinging stream
acts.
21 Decide what lift of valve you intend to have. American water-
works practice is from 0. 10 to 0.20, the diameter of the inside of the
outer seat. This lift is designated by L in Formula [4].
22 Decide what spring pressure you will have, both at the begin-
ning and at the full lift. This spring pressure is expressed in pounds
per square inch of inside valve area and usually runs from 0 . 30^to 0 . 60
lb. per square inch at the beginning of the lift, and it ought not to be
quite double this amount when the valve is full open to its stop. It
will be this final tension, plus the weight of the valve in waterj'^desig-
nated by p in Table 2, that willjbe the determining factor for thejveloc-
ity of the 'issuing stream. To illustrate, if the final pressure be 0 . 81
lb. per sq. in., with a lift of 0 . 15 d, the equation (see Table 2) for V^ =
1 . 161/81 = 10.44 ft. per sec.
23 The discharging area is the net circumference of the inside
valve diameter C, taking out the ribs whether they support the valve
or not, multiplying this by the actual lift and this product by the
coefficient of contraction u, as found also in Table 2, which, for the
lift cited, is 0 . 56.
24 Algebraically expressed, these factors become the denominator
in Formula [4],
C X L X u X v„,
PUMPS VALVES AND VALVK AREAS
975
TABLE 3 A-IN. LIFT
PLUNGER 34 IN. DIAMETER BY 5-FT. STROKE BY 25 R.P.M. MAXIMUM VELOCITY =
6.67 FT. PER. SEC. VALVES 3MN. INSIDE DIAMETER. NUMBER OF VALVES,
SPRING TENSIONS AND PUMP EFFICIENCIES
Initial aad Final
Spring Tension
Pounds
1 0.60tol.55
2 0.40tol.03
3 0.30to0.77
i I o.eotoi.os
5 ' 0.40to0.72
6 0.30to0.53
Valve
Seat
Area
Per
Cent
112
137
159
Num-
ber of
Valves
Lift
of i_
Valves
Inches Plunger
Maximum Velocities Feet per
Second
Valve
Seat
127 A
155 A
180 , A
Longer Springs
6.67 5.
6.67 4.87
6.67 4.20
Valve
134
152
A
6.67
4.98
164
186
A
6.67
4.07
185
210
A
6.67
3.60
8
9.44 - 14.44
7.71 - 11.77
6.68 - 10.18
9.44 - 12.05
7.71 - 9.84
Loss of
Efficiency
Per Cent
3.08
2.05
1.50
2.30
1.52
Column 3 Is obtained by multiplying the number of valves by the net area through the seat
> sq. in.), and finding its ratio to the plunger area.
Column 5 Is taken at A In. = (0.15 X d) in all cases.
All the other data have been already explained.
No. 1268
A REPORT ON CAST-IRON TP:ST BARS
By a. F. Nagle, Bethlehem, Pa.
Member of the Society
On machinery castings as well as on cast pipes, separate bars are
cast, to be subjected to tensile or transverse stress to the breaking-
point, these results being used as evidence of comphance with the
contract specifications. The writer has examined a large number of
such test bars for castings used in the Baltimore sewage pumps and
here reports the results of this examination and study. Perhaps the
most important conclusion is that the test bar is not to be regarded
with too much confidence as indicative of the exact strength of the
casting.
2 All transverse bars were nominally 2 in. by 1 in. by 24 in. cen-
ters. They were cast from two patterns in one mold, made in the
same kind of sand as the main casting. The flask was inclined about
30 deg. There was but one gate for the two bars, with suitable
risers. The iron for the bars was poured from a small ladle of iron
taken as near as might be from the middle of the pour of the main
casting. The breaking loads were corrected for varying dimensions
Wb(P
of the bars by the formula W = , where b and d are the actual
dimensions, W the actual breaking load, and W the corrected load
or weight. These results are used throughout this paper. The
deflections were not corrected.
3 The tensile bars. If in. by 6 in., were cast upright in the same
mold as the main casting,within three or four inches thereof, and
connected by an upper and a lower gate. The tensile bars were
turned to 1^-in. diameter and threaded, and the middle portion
reduced to 1.129-in. diameter, which is equal to 1 sq. in. area Table
1 gives the results of the chemical analysis of the several bars tested.
4 From August 5, 1907 to April 4, 1908 there were made 67 single
tensile bars and the same number of pairs of transverse bars, and the
Presented a*t the Annual Meeting, New York (December 1909), of The
American Societt of Mechanical Engineers.
978
xiBPORT ON CAST-IRON TEST BARS
TABLE 1 ANALYSIS OF CAST-IRON TEST BARS
Bars used in I. P. Bed Plate, and I. P. Frame, for Baltimore Sewage Pumps
a
K z
p
ta
Z
S z
V
o
5 s
S 0
z o
Z
O
OQ
o
a a
H
Date Cast S §
5^
S ^
0
K
00
O
(5
u
OQ
> ■<
m o
November 21, 1907 3.580
2.830
0.75
0.79
0.485
0.081
1.59
24,900
2440
0.49
November 26, 1907 3.396
1
2.736
0.66
0.38
0.459
0.124
1.91
22,000
2075
0.40
average of the latter was used in this record. From April 4 to
December 19, 1908 there were made 91 pairs of tensile bars and an
equal number of pairs of transverse bars, and each piece of the pair
is recorded, instead of the average,
5 Of these 249 tensile bars and their corresponding transverse
bars, 32 sets — 26 flat and 6 round — were rejected for defects due to
blow-holes and four tensile bars were too hard to bear threading, but
the companion piece was used in this record.
6 Of the 217 specimens here recorded, 42 are designated as
abnormal, that is, the ratio between the tensile and the transverse
bars was either considerably greater or smaller than the average.
7 By referring to Table 2, it will be seen that of the 175 specimens
of cast iron running from 20,000 to 30,000 lb. tensile strength, the
ratio of tensile to transverse loads is practically 10 to 1 and the deflec-
tion 0.45 in.
TABLE 2 COMPARISON OF CAST-IRON TEST BARS
Number op
Specimens
j^LiMir] OP Breakinq
Load of Transverse
1 Bars
Breakinq Loads
Pounds
Transverse \ Tensile
Deflec-
tion
Inches
Ratio of Tensile
to Transverse
29
36
51
43
16
2000 to 2200
2200 to 2400
2400 to 2500
2600 to 2800
2800 to 3000
2065
2289
2523
2756
2894
21,630
22,940
24,880
26,500
28,460
0.43
0.45
0.47
0.49
0,49
10.47 to 1
10.02 to 1
9.86 to 1
9.61 to 1
9.83 to 1
175
1 Averages...
2383
23.732
0.45
9.96 to 1
i
Note. — Transverse bars, rough 2 In. by 1 In. by 24 In. centers; tensile bars, turned 1.129 In.
diameter (1-sq. In. area). In this and the following tables the averages given are for the total
n umber of specimens.
REPORT ON CAST-IRON TEST BARS
979
8 Table 3 gives 25 abnormal cases where this average ratio is as
high as 12.56 to 1 with a deflection of 0.43 in.; also 17 abnormal cases
where this average ratio is as low as 7.91 to 1 with a deflection of 0.44
in. And yet the average of both normal and abnormal bars was
10.07 to 1, again very nearly 10 to 1.
TABLE 3 COMPARISON OF CAST-IRON TEST BARS
Abnormal Resdltb
Number of Limit of Breaking
Specimens ^°^° °^ Tkansvehse.
Bars
Breaking Loads
Pounds
Transverse Tensile
Deflec- Hato of Tensile
TION YQ TrANSVEKSE
INCHES
.\bove 10 to 1 ratio
10
2000 to 2200
2088
27,143
0.41
12.95 to 1
10
2200 to 2400
2294
28,530
0.43
12.44 to 1
4
2400 to 2600
2436
29,600
0.49
12.15 to 1
0
2600 to 2800
1
2800 to 3000
2890
, 34,000
0.45
11.76 to 1
25
Averages
2258
28.365
0.43
12.56 to 1
Below 10 to 1 ratio
1
2000 to 2200
2105
17,600
0.50
8.36 to 1
4
2200 to 2400
2359
18,825
0.41
7.98 to 1
7
2400 to 2600
2487
18,814
0.43
7.57 to 1
3
2600 to 2800
2656
21.230
0.45
8.00 to 1
2
2800 to 3000
2969
24,500
0.47
8.25 to 1
17
2521
19,934
0.44
7.91 to 1
9 Breaking loads, presumably alike, varied in pairs of transverse
bars, and also in pairs of tensile bars, as follows:
Out of 65 pairs of flat or transverse bars,
14, or 22 per cent, average variation 18 per cent.
17, or 26 per cent, average v^ariation 5.4 per cent.
34, or 52 oer cent, average variation less than 2 per cent.
Out of. 65 pairs of round or tensile bars,
22, or 34 per cent, average variation 15 per cent.
20, or 31 per cent, average variation 5.5 per cent.
23, or .35 per cent, average variation less than 2 per cent.
61 other pairs of flat bars, which had only one compauicm tensile bar,
varied in about the same ratios.
980
REPORT ON CAST-IRON TEST BARS
10 Two special flat bars and two special round bars, cast in one
mold, one gate and at one pour, varied as follows;
2 flat bars, 12 per cent.
2 round bars, 7 per cent.
11 In order to get some more definite information on these varia-
tions, if possible, I had a pair of transverse and a pair of tensile bars
made and cast in the same mold, and while the average ratio of tensile
to transverse strength was again nearly 10 to 1, as shown in Table 4,
the same type of bars again varied 12 per cent and 7 per cent respec-
tively as shown in Par. 10.
12 I have no satisfactory explanation for the great variations in
these test bars, and we can only accept the fact that mathematical
imiformity in strength of cast-iron bars is not found in the present
TABLE 4 COMPARISON OF CAST-IRON TEST BARS
Special, Two Sets Cast in Same Mold at Same Time
Number
of
NS
Limit of Breaking
Load of Transverse
Barb
Breaking Loads
Pounds
Deflec-
tion
Inches
Ratio of Tensile
Transverse
]
Tensile
1
1
1 __ . .
1
1
2350
2100
23,000
21,470
0.50
0.45
9.79 to 1
10.21 to 1
2
Average
2225
22,235
0.47
10.00 to 1
217
All averages
2380
23,970
0.45
10.07 to 1
state of the art. To anyone questioning the results I can only say
from my own knowledge of the circumstances that the personal equa-
tion did not enter into them.
13 Careful observation of broken bars did not show that the so-
called "skin of the metal" was of any appreciable thickness, and
the metal was remarkably homogeneous throughout. The tensile
bars being turned, the skin, if there was any, of course disappeared.
It is my opinion that the skin adds practically nothing to the strength
in either transverse or tensile bars, other causes, though obscure,
producing far greater deviations.
14 Although many castings were condemned for physical defects,
such as blow-holes, shrink-holes, sand- washes, and shifting of cores,
not a single case of cold-shut was discovered. This is in marked
contrast with the writer's experience on similar work in other foun-
REPORT ON CAST-IRON TEST BARS 981
dries. Excepting a number of steam valves, which were of iron too
soft for their purpose, but one large casting, a discharge air chamber
weighing 16,000 lb., was condemned for being of unsatisfactory iron.
In this case the iron was coarse-grained and brittle, and was
required to stand at least 23,000 to 24,000 lb. To remove all doubt
that the test bars were truly representative of the iron in the main
casting, two tensile bars were cut out of a large flange, which had
been at the bottom of the mold. These, from the most favored part of
the casting, as will be seen, stood but about 17,350 lb. , 90 per cent of
that revealed by the test bars. In this case there was a remarkable
agreement among these pairs of bars.
TABLE 5 TEST BARS FROM CONDEMNED CASTlNii
Breaking
Loads
Pounds
Deflection
INCHES
Transverse
Tensile
1,968
19,800 1
0.35
2.019
19,000
0.50
Cu
t out of flange
17,000
17,700
15 It m..y be interesting to apply these results to the formula for
the strength of cast-iron beams subjected to similar stress. The
formula commonly used (Kent, page 268) is
R= ^^^
2h(P
where R is called the modulus of rupture, or stress per square inch of
extreme fiber.
P = load at center.
I = length in inches between supports.
b and d = breadth and depth, respectively, in inches.
Making the proper substitutions, we have R = ,
^ X ■^ X 1 X 1
or R = 42,840 lb. This is not the correct figure, however, for the
extreme fiber stress, since we know this cannot exceed the tensile
strength, which we have found to be 23,732 lb.
982
REPORT ON CAST-IRON TEST BARS
16 I think it is better to use D. K. Clark's formula, given on page
Wl
507 of his Engineers' Tables, ^ etc., *S^
, where *S = extreme
1.155 6^2'
fiber stress, or tensile strength. If we use the tensile strength found
in these tests as 23,732 lb., the breaking load W would become
23,732 X L155 X 2 X 1 _ 2284 lb. As this is within 4.3 per cent of
24
the average breaking load actually found in these tests, 2383 lb.,
this formula, using the tensile strength for the extreme fiber
stress, seems to me to be the more intelligible and dispenses with the
" coefficient of rupture."
17 " Mr. Barlow found by experiment that for 1-in. square bars of
cast iron, the breaking weight in tons [2240 lb., I presume] was
expressed by the formula W= X 13.6, and Mr. Robert Stephen-
son arrived by experiment at exactly the same coefficient." (Clark,
page 561).
TABLE 6 CIRCULAR TEST BARS CAST IN VERTICAL DRY-SAND MOLD
Bab
Mark
Breaking Loads
Pounds
Deflection
Inches
Value of W
BY Formula
Original
Diameter
Transverse
Tensile
Pounds
Inches
H
H
X
3344
3344
3026
23,070
23,754
24,670
0.15
0.15
0.12
2948
3036
3153
1.305
1.306
1.300
1
2
3 j 4
5
6
18 If we should substitute the value for W found in these tests
we would have W, or ,
2240
2x1
1.064 tons = X a constant,
24
or, constant = 12.77, which is within 7 per cent of the coefficients
found by Barlow and Stephenson.
19 Since the foregoing was written I have had the opportunity to
observe two circular test bars nominally li^ in. in diameter by 15 in.
long, with 12-in. centers. These bars were cast from two patterns
in one vertical dry-sand mold and poured from a small ladle, first one
and then the other, with the results shown in Table 6.
^ Rules, Tables and Data for Mechanical Engineers, by D. K. Clark;
Blackie & Son, London.
REPORT ON CAST-IRON TEST BARS 983
20 The tensile bars were taken from the bottom end of the broken
test bar, but I do not know whether H or X was poured first. The
first tensile bar H had a small air hole, which being corrected for
added 7 per cent to its tensile strength, and this is also given in the
table. A second bar was then turned up from the immediate join-
ing piece with the result recorded first in the table. The turned bars
were 0.937 in. in diameter. Column 6 gives the original diameter.
Column 2 was found by reducing the actual breaking loads in the ratio
of the cubes of the diameters, and Column 3 was reduced to the
square inch area. Why the transverse breaking loads should vary
10 per cent and the tensile bars 4 to 7 per cent the opposite way, a
total variation of 14 to 17 per cent, I leave to the reflection of the
reader. If we apply Clark's formula for the breaking weights for circu-
lar bars, W = ' — , we find the values given in Column 5.
21 In this age of economic production, the cost of these turned ten-
sile bars is frequently objected to by the manufacturer. While blow-
holes seem to be more frequent in flat transverse bars than in round
attached tensile bars, the latter seem liable to a greater abnormal
hardness, for which I have no explanation. Some indication of the
toughness of cast iron may be seen in its deflection, which is not
revealed in a direct tensile pull. I should, therefore, be satisfied
with two or three transverse test bars 2 in. by 1 in. by 24 in. centers,
and a deflection record poured, as near as may be, from the middle of
the pour of the main casting, as giving a fair indication of the iron in
the main casting, but mathematical exactness cannot be looked for as
yet.
22 If we wish to know approximately the corresponding tensile
strength of the iron, we can multiply the breaking load of the 2 in.
by 1 in. by 24 in. flat bar by 10. If the test bar is of 1^-in, diameter by
12-in. centers, its breaking load should be multiplied by 8, to obtain the
approximate tensile strength. The general rule seems to be, that
where both flat bars agree in breaking loads, the tensile strength is 10 to
1 of the breaking load, but where they differ, the 10 to 1 ratio does not
hold. A better practice, therefore, might be to cast three round
transverse bars and accept the two that agree, if each is round, as a
fair sample of the iron, dispensing with the tensile bars. This con-
cession to the manufacturer, I believe, would entail not only no loss to
the customer's interests, but a positive gain.
984
DISCUSSION
DISCUSSION
Prof. W. B. Gregory. The writer has recently made a large num-
ber of tests of cast-iron specimens of 1-in. square cross section
and with supports 12 in. apart, a few being also broken in tension.
The results confirm the deductions of the author as to the relation-
ship between breaking loads in tension and in cross bending. The
10-1 ratio holds in these tests as in those given by the author.
Table 1 gives the results of the cross-bending tests, the load being
applied at the center.
TABLE 1 TESTS IN CROSS BENDING
Specimens 1 in. by 1 in. 12 in., between Centers, Load Applied at Cenier
i
Number '
Bbeakinq Load Lb. per Sq. In.
Deflection In.
1
2280
2250
2680
2410
2250
2370
2240
2310
2250
2470 !
2180 i
0.10
2
0.10
3
0.09
4
5
0.09
0.08
6
7
8
9 . ..
0.09
0.09
0.08
0.09
10
0.08
11
0.10
Mean
2335 i
0.09
2 From the specimens broken in cross bending, six were selected
from which were turned tension test pieces approximately i in. in
diameter at the smallest section, their length over all being 5 in.
The threads at the ends were I in. outside diameter. The test pieces
were made to fit loosely into the tension bars of the testing machine so
TABLE 2 TENSION TESTS
NXTMBEB
Breaking Load Lb. per Sq. In.
1
22,900
2
23,300
3
22,800
4
21,550
5
24,600
6
22,050
Mean
23200
REPORT ON CAST IRON TEST BARS 985
that side stresses were entirely eliminated, and the specimens were
broken in pure tension. The results are given in Table 2. The
ratio of tensile strength to load in cross bending is
23.200 ^ 9 9,
2335
This comparison can be made only on the basis of averages, as no
record was kept of the numbers of the specimens broken in cross
bending. The six tension specimens therefore represent six of the
eleven specimens broken in cross bending. Specimen No. 9 of the
cross-bending tests may be taken as fairly tj'pical. A chemical
analysis was made of this specimen with the following results:
Total carbon 4.04
Silicon 1 . 76
Phosphorus 0.562
3 The mean deflection as given by the author averaged 0.45 in.
for two sets of specimens and 0.44 in. for another set. The highest
value of deflection in any case was 0.50 in. Since the deflection varies
as the cube of the length of specimens between supports, a id
inversely as the width of section, it follows that the deflection for
specimens 2 in. by 1 in., tested with supports 24 in. apart, should
be four times the deflection bf specimens 1 in. by 1 in., for a length
between supports of 12 in. On this basis the specimens tested by
the writer should have
= 0.112 m.
4
deflection instead of 0.09 in. average as the tests showed.
4 This raises the question of what deflection ought to be specified
for specimens 1 in. square, with 12 in. between supports. Some
specifications have recently been brought to the attention of the
writer, in which the minimum deflection was placed at 0.15 in. Is
this commercial cast iron or does it call for a special mixture, expen-
sive and hard to obtain?
5 The author has mentioned that the "skin of the metal" was of
no appreciable thickness. I would like to ask if he has ever tried
the effect of rattling on specimens. The process of rattling will
remove the sand and the skin of the metal. In this connection the
results in Table 3 may be of interest.
6 The tests given in Table 4 are on specimens of the same size
as those in Table 3. The metal used was as nearly the same as the
986
DISCUSSION
TABLE 3 TESTS OF CAST IRON IN CROSS BENDING
Spbcinbns Round, li in. in Diambtbb, 12 in. between Centebs; Not Rattled
No.
Breaking Load Lb.
Deflection In.
Remarks
1
2450
3010
2670
2580
2700
2580
2620
2430
3360
2750
2990
3170
2950
2960
3080
2580
0.075
0.08
0.07
0.14
0.09
0.14
0.08
0.075
0.09
0.08
0.09
0.09
0.095
0.12
0.10
0.075
Cast In pairs on end
2
3
<( a u <> ('<
4
.< II II II i<
5
II II 11 .< 11
6
« II II II II
7
II 11 II II II
8
Cast flat
9
10
i< «
11
II II
12
<i II
13
II <i
14
II II
15
II 11
16
Cast on end
Mean
2805
0.093 !
foundry could make it and the specimens were placed in a rattler
and the sand and "skin" removed by abrasion. From these figures
it will be seen that rattling has increased the strength of the speci-
mens by 23.85 per cent. This phenomenon has been noticed by
other experimenters.
7 The statement that rattling increases the strength by about
25 per cent seems to be borne out by experiments. The increased
strength is probably due to a removal of some of the internal stresses
'in the specimens and to the fact that the particles of iron, by the
TABLE 4 TESTS OF CAST IRON IN CROSS BENDING
Specimens Like Those in Table 3, But Rattled
No.
Breaking Load Pounds
Deflection Inches
1
3750
3330
3400
3520
3640
3640
2760
3670
3060
4020
3440
3474
0.09
2
0.095
3
0.08
4
0.09
5
0.09
\i
0.10
7
8
«
10
11
0.075
0.095
0.09
0.10
0.09
Mean
0.0904
REPORT ON CAST IRON TEST BARS 987
process of tumbling the bars together, are allowed to arrange them-
selves so that they are better able to resist stresses.
8 Since the breaking load varies directly as the moment of inertia
of the cross section of the specimen about the gravity axis, we have
I, for the specimens U in. diameter = \r.r* = 0.7854 X 0.625' = 0.1203
Ig for the specimens 1 in. square = -l^ ^h^ = rV = 0.0833
Then
0.1203 , ,,
= 1.44
0.0833
Making the comparison between the unrattled round specimens and
the square ones, we have
^§25 = 12
2335
Comparing the rattled round specimens with the square ones we have
'''' = 1.487
2335
A. A. Gary. It is unfortunate that the value of the structural
study of metals and alloys, by use of the pyrometer and microscope, is
not more widely appreciated. I feel safe in saying that by such means
all the variations noted in Mr. Nagle's paper can be most satis-
factorily accounted for. The fact is now generally recognized that'
iron or steel identical in chemical composition may possess widely
differing mechanical properties which are quickly recognized by
microscopic examination.
2 Chemical analyses, as given in Table 1 of the paper, are undoubt-
edly of considerable value in the investigation of cast-iron; but without
a physical examination our knowledge of the ability of the metal to
withstand stresses and strains is very uncertain. Not only will
investigations of this kind show us the cause of the variations noted in
Mr. Nagle's paper, but they will give us the information needed to
produce a metal of great uniformity.
Prof. T. M. Phetteplace. It would be interesting to know
whether a thorough sand-blasting would have any effect., as different
results seemed to be obtained by cleaning off the skin of the material.
988 DISCUSSION
The Author. Since the paper was written 1 have had opportunity
to examine some instructive records of eleven sets (of three each) of
round test bars. The bars were li in. in diameter, rough, on 12-in.
supports, the breaking loads being corrected for actual diameters.
The deflections were not corrected.
BREAKING LOADS IN POUNDS. DEFLECTION FROM 0.12 IN, TO 0.15 IN.
1
3276
3185
3044
4400
4005
2913
3276
3306
3382
3204
3268
2
3367
3276
3162
3100
3913
3003
3185
3204
2976
3204
3124
3
3276
3534
3255
3500
3640
3115
3026
2937
3003
2912
2812
2 The three bars in each set were cast in three separate molds, No.
1, or the upper line, being cast from the first pour of the ladle, No.
2 from the middle and No. 3 from the bottom. It will be observed
that in eight of these eleven sets, the bar selected from the two
nearest in agreement came from the middle of the pour, and that
all of the extreme variations were found in either the first or last
pour. If we have only two bars they would differ as much as 22 per
cent, while if we took the two out of three nearest in agreement,
those two would not vary more than 2 per cent or 3 per cent.
3 I have had no experience with bars 1 in. by 1 in. by 12 in., but
I think that the deflectien of 0.15 in., mentioned by Prof. Gregory
would be difficult to realize in machinery castings.
No. 1269
SYMPOSIUM ON
THE EFFECT OF SUPERHEATED
STEAM ON CAST IRON AND STEEL
Three papers: Cast Iron Fittings for Superheated Steam, by Prof. Ira N
HoUis, Boston, Mass; The Effect of Superheated Steam on the Strength of Cast
Iron. Gun Iron and Steel, by Prof. Edward F. Miller, Boston, Mass.; Cast Iron
Valves and Fittings for Superheated Steam, by Arthur S. Mann, Schenectady
N. Y.
CAST-IRON FITTINGS FOR SUPERHEATED
STEAM
By Prof. Ira N. Mollis, Boston, Mass.
Member of the Society
The failure of a number of large cast-iron fittings in use with super-
heated steam has rightly created a widespread suspicion of this mate-
rial when exposed to high temperature. Yet on this subject there is
very little information of a character to justify the wholesale substi-
tution of steel castings for the ordinary heavy cast-iron fittings. The
latter have been used with success for many years at all degrees of
temperature below actual redness, and in many stations now in opera-
tion with moderate degrees of superheating (say 100 deg. falir.) cast
iron has never given the slightest trouble beyond the ordinary were
and tear.
2 The doubt as to the reliability of cast iron has seemed to spring
up with its use in long pipe lines to steam turbines where the temper-
ature has ranged from 550 deg. to GOO deg. This would lead one to
ask if the difficulty has not been in the design of the piping systems
rather than in the character of the material. Has not the cast iron
taken the brunt of a new service and has it not suffered in the estima-
tion of the engineering public because the conditions of thai service
were not fully understood?
Presented at the Boston Monthly Meeting (December 1909) of The
American Society of Mechanical Enotneerr.
990 FITTINGS FOR SUPERHEATED STEAM
3 A vast amount of experiment and investie;ation would be
required for the satisfactory reply to this quoi^tion, and this brief paper
is not intended as a reply, but rather to place before the Society a
record of some tests that may throw light on the subject. Those
tests were made for the Edison Illuminating Company of Boston for
the purpose of determining the bursting strength under hydraulic
pressure of some large fittings which were replaced with steel castings.
4 It may be well before giving the result of the tests to inquire
what is actually known about cast-iron fittings subjected to high
temperature; that is, Icnown without the possibility of controversy:
o Fittings have developed cracks and small changes of shape
after a few months of actual service.
b Fittings exposed separately to superheated steam at a tem-
perature exceeding 500 deg. fahr. have shown a perma-
nent increase of some dimensions.
c The tensile tests of pieces cut from fittings that have failed
in service indicate in some cases the possibility of perma-
nent loss of strength.
5 The remainder of the evidence in the case may be classed as
good deductions from laboratory tests of specimens previously ex-
posed to high temperature, or from some preconceived theory as to
the behavior of the constituent parts of cast iron in a rising tem-
perature.
6 One of the curious and interesting qualities of cast iron is its
permanent increase of dimensions under high temperature. This is
paralleled by the permanent set of cast-iron test pieces when subject
to very moderate tensile stresses. In both cases the cast iron appar-
ently continues to grow at a decreasing rate, at least in some dimen-
sions, when the high temperature or tensile stress is repeated.
7 How long this growth would continue is not known. Its prob-
able limit is the flow of the material under the ultimate breaking stress.
Cast iron may not be peculiar in this respect and all materials may
change their dimensions permanently under moderate stress, the
change growing with each imposition of the same stress. There is no
doubt of this where the yield point has been exceeded. It may also
be true that all materials change permanently under repeated applica-
tion of high temperature.
8 The cause of the persistent expansion under high temperature
is still veiy hazy, but two possible agencies have been mentioned in
a number of discussions:
FITTINGS FOR SUPidKHEATED STEAM 991
a A chemical, or physical, change in the relation of the iron
to the various foreign substances which fix it as cast iron.
b A molecular change due to the fact that cast iron has no
well-defined elastic limit or modulus of elasticity.
9 Both causes may be in operation at the same time, but the
theory of chemical change has far less standing than that relating
to the stresses produced by unequal expansion. While there is a
temperature at which carbon changes its relation to the iron, super-
heated steam is probably well below that point except under very
unusual conditions.
10 That the strength of cast iron is materially reduced when
exposed to superheated steam at 600 deg. is not conclusively proved.
Test specimens taken out of cast-iron fittings after one year or more
of exposure to a temperature of 550 deg. to 600 deg. have shown
a surprising irregularity of strength in the same casting, but there is
nothing to prove that new cast-iron fittings have not a great lack of
homogeneity. Irregularities exist in every casting owing to the
inability of the metal to flow when cooled below a certain
temperature. Furthermore, the strength of a test piece cast from
a given heat can rarely be taken as that of any selected part of the
fitting cast from the same heat. It is common experience to find
variations of strain in castings as well as variations of texture. Were
any large, irregular casting cut into small test pieces, the variations
of strength would probably be found to be quite as great as that
reported later on in this paper. The demonstration of the loss of
strength after, long service with superheated steam does not seem
complete, in spite of the fact that some qualities of cast iron have
shown a loss in the laboratory.
11 A very brief description of the essential features of the Edison
station at South Boston where the condition of fittings has been
investigated will help to make clear what follows. The new equip-
ment of the station is arranged in a series of complete units each con-
sisting of one vertical Curtis turbine and eight boilers set in pairs.
The main steam line extends along the rear ends of the boilers just
beneath the brick work, four 8-inch vertical steam mains connect-
ing each pair of boilers with the main line. Three of the vertical
mains discharge through gate valves into tee's, and the fourth, at
the end of the line, through a gate valve into a bend.
12 The first turbine units were provided throughout with cast-
iron fittings, which were ultimately replaced with steel fittings. No
expansion or slip joints are used. The main steam line (something
992 FITTINGS FOR SUPERHEATED STEAM
over 103 ft. long) is anchored at the turbine end and is allowed to
expand freely in a longitudinal or horizontal direction carrying the
lower ends of the vertical mains with it. The steam pressure is 175 lb.,
and the superheating generally amounts to 150 deg. fahr., all
though it is not constant. The actual temperature of the steam
varies from 500 deg. to 5(S0 deg., so that the main line is changing in
length from time to time, thus moving the lower ends of the vertical
mains back and forth. A series of variable stresses are consequently
introduced into all parts of the pipe system, probably affecting most
seriously the tees. It is this aspect of the case, namely, the effect of
varying stresses upon cast iron at high temperature, that must be
studied before a sound verdict can be reached.
13 The castings in the South Boston station were first suspected
of failure when nearly a year after the turbine plant had been in oper-
ation one of the 8 in. by 6 in. by 6 in. tees near the boiler showed signs
of deterioration, cracks appearing near the junction of the offset
with the body of the tee and in the flanges. Another fitting of the
same dimensions and location began to fail and was taken out after
fourteen months' service. Both these tees were cut up for testing
and the results were so much alike that only the second is given here
as that tee had been exposed the longest to the strain.
14 A chemical analysis gave the following: carbon, 3.47; man-
ganese, 0.10; phosphorus, 0.366; sulphur, 0.062; silicon, 1.41. The
tensile strength of six pieces taken fiom different parts of the tee was
found to be, 12646, 14295, 26080, 27270, 27440, 28280 lb. per sq. in.
There thus appears to have been considerable variation of strength
in the te? unless the first two results are errors due to some faults in
testing. Not considering them, the four other pieces do not appear
to indicate any great falling off in service. They are as near together
as would commonly be found in cast iron from the same heat. The
material was supposed to be a first-rate quality of air-furnace gun iron
which should have been good for 25,000 to 30.000 lb. per sq. in., but
no tests or analyses of the heat from which this tee was poured are
on record.
15 Four test pieces cut from a larger tee, 14 in. by 12 in. by 8 in.,
which had had about the same service as that from which the fore-
going test pieces were cut, gave a tensile strength of 23130, 23480,
23875, 24170 lb. per sq. in. Here again there was absolutely no proof
that the material had deteriorated.
16 Three test pieces were taken, for comparison, from a large
manifold which had been seven years in service with saturated steam,
FITTINGS FOR SUPERHEATED STEAM 993
and the tensile strength was found to be 16413, 16550, 17000 lb.
per sq. in. The nature of the cast iron was not known positively,
but it was bought as air-furnace gun iron.
17 It was fully recognized in the first of the foregoing tests that,
while the material might not^have suffered in service, nevertheless
parts of the casting might have been weakened by the expansion
stresses. For the purpose of testing this, two tees were removed from
the line and broken by internal hydraulic pressure, thus affording a
definite idea of the strength of the castings as a whole. A third
casting, an elbow not previously in use, was added for comparison.
The three fittings are shown in Fig. 1 in which the measured dimen-
sions and the location of the fractures are given.
18 The material was the same as that used for all the fittings of
the turbine unit, air-furnace gun iron, and the chemical constituents
were probably about the same as given for the 8-in. tee in Par. 14.
The two tees had been exposed to superheated steam of 578 deg.
fahr. and less, for fifteen months or longer, and when removed had
given no indications of weakness. A careful examination disclosed
no appreciable distortion except in the faces of the flanges which were
no longer plane surfaces. There were several high spots that could
not have existed when the flanges were faced off.
19 No. 1 was a 14-in. tee with an 8-in. offset. The openings
were closed by heavy cast-iron plates fitted to the flanges and bolted.
The pressuie was produced by a steam-driven outside-packed plunger
pump, and was measured by means of a small conical safety valve,
one-tenth of a square inch in area, and directly loaded by dead weight
applied as the pressure increased. The indications of the small
valve were constantly compared with a hydraulic gage previously
tested and calibrated at the Crosby manufactory. The fitting broke
as shown at an internal pressure of 1650 lb. per sq. in. The plates
covering the openings did not reinforce the tee to any great extent as
the bolts were smaller than the holes and the joints around the flanges
were leaking appreciably when the fracture occurred.
20y No. 2 fitting was an 8-in. by 6-in. tee. It was broken in pre-
cisely the same manner as No. 1 and gave way at an internal pressure
of 3100 lb. per sq. in.
21 No. 3 fitting was a 12-in. elbow. Its two openings were closed
with cast-iron plates and it was burst in the same way as the others.
The joints practically gave]^out at a pressure of 2000 lb. per sq. in.
although the pressure was kept on for some minutes. For the second
attempt to run the pressure up, the bolts were set up with a very
994
FITTINGS FOR SUPERHEATED STEAM
Failed [A-B]
16501b. per sq.in.
Failed [C-D-E]
3100 lb. per sq.in.
Failed [F-G]
1500 lb. per sq.in.
Pre^•iously held 2000 Ibs.J
No.3 ^
Sixteen IM bolts
17 ^i) pitch circle
Fig. 1 Showing|Points of Failure under Hydraulic Pressure of i'hree
Cast-Iron Fittings used with Superheated Steam
FITTINGS FOR SUPERHEATED STEAM 995
heavy wrench, which undoubtedly put a bending moment on the
flange. The fitting finally parted all around the root of the flange
at a pressure of 1500 lb. per sq. in.
22 Four test pieces were cut from the larger tee and broken under
tensile stress. Their dimensions were almost exactly J in. in dia-
meter by 6 in. between fillets. Two of them were broken cold and
<:ave a tensile strength of 22,150 lb. per sq. in., and two were broken
at a temperature of 590 deg. at 20,050 lb. per sq. in. The temperature
of the latter was maintained by means of a cylinder-oil bath, the oil
being placed in a large tube surrounding the test piece and kept hot
by a gas flame.
23 No information could be obtained as to the original strength
and chemical composition of the iron and it would be impossible to
prove that it had changed either in strength or in composition.
There is ground, however, for believing that it had changed, as the
8-in. tee gave as high as 28,000-lb. tensile strength in one specimen.
24 A comparison of the larger tee with results of tests reported
in the Valve World for November 1907, throws an interesting light
on the subject. The formula there published as derived from a very
large number of tests of cast-iron and ferro-steel fittings may be
taken as a basis for calculating what should have been the bursting
T
pressure of the 14-m. tee. This formula is -B = j^ X S, where
D = inside diameter of the T
T = thickness
S = tensile strength of the material multiplied by 60 per cent
B = bursting pressure in lb. per sq. in.
25 Taking the tensile strength of the cast iron when hot at 20,000
lb. per sq.in.the diameter of the tee at 14 in. and the thickness at 1|
in. the value of B is 1070 lb. per sq. in., whereas the tee actually burst
at 1650 lb. This did not seem to indicate weakness or deterioration.
26 It is interesting to inquire here what stress existed in the tee
during its service. That due to the steam pressure was small when
compared with the actual bursting pressure, but that due to expansion
may have been serious in its effect. The first tee in the main steam
line was located at 4 ft. 8f in. from the anchorage, the second 37 ft.
8f in., the third 70 ft. 8^^ in., and there was no expansion joint to case
off the pressure on the vertical mains. Taking the third tee for pur-
posco of illustration, certain suppositions can safely Ik- nuule:
996 FITTINGS FOR SUPERHEATED STEAM
a The lower flange of the vertical pipe moves in a horizontal
plane as the main pipe expands and therefore the lowest
part of the axis of the pipe moves parallel to itself.
6 The upper end of the vertical pipe is practically fixed. The
expansion of the main steam pipe thus puts an S bend in
the vertical pipes and introduces large bending moments
into both ends of it and into the tee.
27 The actual length of the pipe between its lowest flange and
the upper end is 26 ft., but the length between the upper surface of
the tee and the upper end of the pipe is about 28 ft. The linear
expansion of the main steam line is about 3 in. when heated to 578
deg. fahr. The effect of this is supposed to have been halved by
cutting the pipe short and springing the flanges into place when mak-
ing the joints. There is thus an initial deflection in the vertical pipe.
This is overcome as the pipe is heated and carried as much farther
on the other side.
28 The value of the deflection in the lower end of the pipe is then
taken at 1.5 in. The formula for the maximum deflection of a beam
fixed at one end and moved parallel to itself at the other end is
Y=
VI El
W = load in pounds or push of the horizontal pipe.
I = length in inches.
E = modulus of elasticity.
/ = moment of inertia of the pipe.
The inside diameter of the pipe is 8 in. and its thickness is 0.322 in
giving the value of / = 72.5. E is taken at 30,000,000.
29 The equation for the deflection is then
7 = 1.5 =
12 X 72.5 X 30,000,000
and W is found to be 1288 lb. Thus if the expansion of the pipe is
exactly split by cutting the main line short, the push on the lower end
of the vertical mains is 1288 lb. The point of contrary flexure in the
S bend of the pipe is about 179 in. above the junction of the offset
of the tee with its main body. The bending moment in the offset of
the tee is therefore 1288 X 179 inch-pounds and the stress set up is
Mc 1288 X 179 X 5
FITTINGS FOR SUPERHEATED STEAM 997
In this formula, 290 is the moment of inertia of the offset taken as
8 in. inside diameter and 10 in. outside diameter. The distance to
the remotest fiber from the neutral axis (or center of the offset) is
the radius 5 in.
30 While this calculation is not entirely reliable on account of the
uncertainty as to the elastic curve of the vertical pipe, nevertheles.s
it is a fair indication of the stress to be expected in this tee when the
temperature of the pipe reaches 578 deg. Furthermore, it is made
under the supposition that the expansion of the pipe was lessened
by an initial pull and that all the joints came exactly fair before set-
ting up the bolts. It is easy to imagine how serious the stresses might
have become under actual conditions of inaccurate fitting. The one
element of splitting the expansion is very uncertain. The foregoing
stress might easily have been doubled, resulting in pulling the tee
quite out of shape and in setting up internal strains certain to weaken
the material in places.
31 Under such conditions, it was, and would generally be, wise
to replace the cast-iron tees with cast steel which would yield more
readily to expansion and which would be safer at much higher tensile
stresses. The reason for the substitution ought not to be lost sight
of in such a case, if cast iron is to be judged fairly. It is made because
it is cheaper on the whole to replace the cast iron with steel rather
than to put in expansion or sHp joints. Perhaps the steel casting
is also much easier to take care of than any form of expansion joint.
The unreliability of east iron in such a service has nothing to do with
the case: it is merely that the design usually adopted for steam pip-
ing does not quite fit cast iron.
THE EFFECT OF SUPERHEATED STEAM ON
THE STRENGTH OF CAST IRON, GUN IRON
AND STEEL
By Edward F. Miller, Boston, Mass.
Member of the Society
The object of this paper is to describe some experiments made to
determine the effect of superheated steam on cast iron, gun iron and
steel. From each piece to be tested two tension specimens were made,
one to be subjected to the action of superheated steam, and one to be
used in obtaining the original strength of the piece.
2 All of the specimens were made with screwed ends in accord-
ance with the specification prepared by The American Society for
Testing Materials. The tension tests were made on a 100,000-lb.
Olsen testing machine, the specimens being screwed into spherical
holders attached to the heads of the testing machine, thus ensuring a
straight tension pull without any bending.
3 The specimens to be subjected to superheat were placed on a
wire grating suspended at the center of a 12-in. iron pipe about 3 ft.
long, supported horizontally on brackets. The ends were closed by
blank flanges. Steam was supplied by a small pipe, a flow at low
velocity being maintained at all times. The under side of the pipe
was heated by Bunsen gas burners. Thermometers, in wells reaching
down to the grating on which the specimens were placed, gave the
temperature of the steam, the pressure being read from a steam gage
on the supply pipe.
4 For the tests plotted in Fig. 1, the average gage pressure in the
superheating pipe was 93 lb. and the average temperature 660 deg.
fahr. The gas flame was extinguished at 5 p.m. and lighted again at
7 a.m. The temperature reached 660 deg. fahr. by 11 a.m. and by 5
p.m. would be as high as 700 or 720 deg. fahr. Steam was kept in
the superheater during the night. The total time these specimens
were exposed to superheated steam was 260 hours, and the exposure
to saturated steam was 460 hours. A chemical analysis of the iron
tested is given in Table 1.
5 For the tests plotted in Fig. 2 the average gage pressure was 82
lb, and the average amount of superheat about 390 deg. fahr. These
FITTINGS FOR SUPERHEATED STEAM
999
TABLE 1 CHEMICAL ANALYSIS OF CAST-IRON SPECIMENS. FIG. 1
Phos-
phorus
Total
Carbon
Graphi-
tic
CARBON
Manoan-
EBS
Silicon
Sulphur
Cast Iron Gibbt
Foundry
Gun Iron Hunt Spiu-
LER ' 0.41
3.51
3.25
3.02
2.60
2.84 '
0.37
0.24
0.38
1.88
0 54
0.05
0 09
Cast Iron Broadway
Foundry
3.34
2.26
0.09
specimens were subjected to superheated steam for 520 hours, and
to saturated steam for 920 hours. A chemical analysis of three of
the semi-steel specimens is given in Table 2. This semi-steel was
GUN
IR
ON 1
\y
k
^
^
\
^
^
J
/
*
^
,'
/
-
13
10 i 8
7
I
flee
eN
umber
y
/
/
-
/
c
^S
FT
RON
/
1
ro
idw
ay
Found
ty
j
/
j
t
''
/
('
t
1
cas;t
RON
'
Gi
.by
Fo
undry
!
X
!
J
/'
--
—
—
«
31
30
1 39
=525
6
J 27
|25
»24
O
a
3
21
20
51 65 52 15 3
Piece Number
Fig. 1 Breaking Loads of Gxjn-iron and Cast-iron Test Pieces Sub-
^ JECTED TO Action op Superheated Steam
made by adding 200 lb. of steel to 1500 lb. of cast iron. The analy-
sis of the gun-iron showed, total carbon, 3.37; graphite, 2.44; man-
ganese, 0.34; sulphur, 0.11; silicon, 1.65.
6 Four grades of steel were tested; two pieces from a bar of 65,000
to 70,000-lb. tensile strength, two from a bar of 75,000 to SO,000-lb.
1000
FITTINGS FOR SUPERHEATED STEAM
tensile strength, two each from three bars of about 90,000-lb. tensile
strength, and two from a bar of over 100,000-lb. tensile strength.
7 The composition of two of the rolled-steel pieces No. 26 and No.
27 was as follows:
Phosphohous
Total Carbon
Manganese
Sdlphub
Silicon
No. 26
0.85
0.73
0.026
0.026
No. 27
0.116
....
0.90
0.057
0.031
8 In Fig. 1 and Fig. 2 the open circles represent the ultimate
strength per sq. in. cf the original specimen while the dots on the
same ordinate gives the strength per sq. in. of the comparison speci-
men which had been subjected to the action of superheated steam.
By figuring the per cent loss in strength in each specimen and then
TABLE 2 CHEMICAL ANALYSIS OF SEMI-STEEL SPECIMENS, FIG. 2
Phobphokub
Total Cakbon
Graphite Manganese
Sulphur
Silicon
0.24
0.61
3.48
3.22
2.64
2.39 0.35
2.83 0.44
i
0.11
0.11
0.49
1.91
2.62
1
taking the average of these per cents it appears that the cast iron
from the Broadway Foundry (Fig. 1) lost 9.5 per cent; that of the
Gibby Foundry 2.4 per cent. The cast iron of Fig. 2 came from the
Waltham Foundry; here there is apparently a gain in strength of 1.8
per cent. Fig. 1 and Fig. 2 show that gun-iron loses strength, Fig.
1 showing a loss of about 3.5 per cent, and Fig. 2 about 2.1 per cent.
9 The tests on semi-steel show an average reduction of strength
due to exposure to the steam, of about 0.4 per cent, four out of six
pieces showing quite a reduction. If piece No. 154 is not considered,
the percentage reduction of strength would be much greater.
10 Of the bar steel tested that of 65,000 to 70,000 lb. tensile
strength showed a loss of 1.8 per cent due to exposure to the steam,
the 75,000 to 80,000-lb. steel a loss of 1.9 per cent, the 90,000-lb.
steel a loss' of 1.5 per cent, and the 100,000-lb. a loss of 24 per
cent.
11 While one is not justified in drawing many conclusions from
the results of as few tests as are quoted here, still it is evident from
Fig. 1 and Fig. 2 that the metals tested have suffered a loss in strength
due to their exposure to the steam. A paper bearing on this subject
Materials for the Control of Superheated Steam, by M. W. Kellogg,
FITTINGS FOR SUPERHEATED STEAM
1001
-
1
-
r-
1
1
-1
1
UJ
ei
b.
fe
«P
s
l'
^
/
CD
s
!
h
j
^
„
1
s
!
f
'^•
^
o
\
y
'
^•'''
^
j
6t
11
9>
1
00
-T
' '
Q
^-1
■-'1
g
1
'i
~--j--
J^L
1
_l
p
--•-
,^
ul
,
.'
u
^
\
,'
'
1!
.
*»,
«.
i
1;
11
I
'*
-.
t-. 1
«
1
,tl
,*-
•-'
' i ! S 1 g
rs s s ' g
ii?
J 1
_,-
--
o
spanpj pucsnc
tx
a;
Haai ajcnbg j
ad p»OT 3ai:iB3J
a
3
?:
x^
^
1
1
J3
o
oc
*■*
'»
0:
1
§
E
' '
s
3
P
z
\,
■J
s
^C
•^
K
fS
N
^
Ui
?!
k
V
^ST
H
N,
r'
1-
fH
1
,^-
'^i
.
§
in
,
.-
''
\
,
7?!
1
,
''
?s
o
'■
--
._
\
u
-"\
1
\
S
'-
S?
.- -■
ti-
^
1
\
\
\
C
•-..-
X
s
-
j
X
o
\
?^
^
-
I'
,''
V
^
■"'
■
<'\
JR
.-'
''
0)
o
1
\
H
Eh
O
05
<!
O
O
<
H
02
O
o
a
<
o
o
gpanoj putianoqx ni r(">nx spnnoj pansnotjx
CO c5 00 OT ?5
spanoj pnr^nonx,
1002 FITTINGS FOR SUPERHEATED STEAM
appeared in the 1907 Transactions of the Society. In the Valve
World, March 1908, are given the results of tests on cast iron taken
from the body of a 14-in. valve which had been in use for four years
on a main carrying steam at 200 lb. pressure and superheated to a
temperature of 590 deg. fahr. A number of test bars cut from the
body of the valve showed a loss of strength of 41 per cent when com-
pared with the strength of the original metal as determined from
coupons tested at the time the valve was made.
12 Fig. 1 in this paper formed part of the thesis of H. A. Terrill,
M. I. T. '07, and Fig. 2 part of the thesis of F. M. Heidelberg, M. I. T.
'09
CAST-IRON VALVES AND FITTINGS FOR
SUPERHEATED STEAM
By Arthur S. Mann, Schenectady, N. Y.
Member of the Society
There have been many failures of cast-iron valves and fittings in
piping systems carrying steam of high pressure and high superheat.
The ordinary extra-heavy flanged cast-iron fittings which are listed
in many manufacturers' catalogues as suitable for 200 lb. pressure and
which have to meet a close price competition, have successfully
carried a pressure perhaps as high as 150 lb. or more. No doubt the
fittings and valves can support a steady pressure of 200 lb. without
bursting, but there have been many failures when carrying super-
heated steam of lower pressure.
2 These fittings are not well suited for permanent work of
even 150-lb. pressure, and many engineers in control of such matters
in stations of a representative type prefer to design their own parts
rather than to trust the usual run of commercial extra-heavy fittings.
3 Probably on account of the advertised ability to support a high
steady pressure these extra-heavy fittings and valves have been
used in a number of instances for superheated work. After a short
time, six months or perhaps even less, cracks make their appearance ;
valves leak, seats become loose, castings grow in length and surface
cracks become so large in size and in number that the casting is
removed from the line.
4 A few repetitions of this experience seem to justify the conclu-
sion that cast iron is not fit material for high-temperature steam.
The natural substitute is steel, which is used with fair, even complete,
success in many cases.
5 It is known that cast iron will grow with repeated heatings
and coolings, often observed in the ordinary straight grate bar.
When the bar is first heated it expands and cools as it contracts; but
if the temperature has been high, the bar will increase in length.
With a second heating, a further increase takes place, followed by
many others. As a consequence the long single, straight, flat grate
1004 FITTINGS FOR SUPERHEATED STEAM
warps and proves the wisdom of McClave's rule " Keep your long lines
of metal away from the fire."
6 This subject of growth has been treated very completely by
A. E. Outerbridge in his excellent paper published in the Journal
of the Franklin Institute for February 1904. Mr. Outerbridge
heated his samples to redness or above, temperatures greatly ex-
ceeding that to which a steam-pipe fitting is subjected.
7 A rough experiment on this line was tried by the writer with two
samples, one of an ordinary cast iron and a second of a high-grade cast
iron, which has proved itself capable of carrying superheated steam
and of which a detailed analysis is given in the following pages of this
paper. The two samples were each 6 in. long and 1 in. in dia-
meter. They were placed in a banked fire over night, reaching
a dull red heat, and were allowed to cool in the air. A slight growth
as measured by micrometer was found in each piece.
8 This treatment was followed for two or three nights and the
growths were measured. There was an increase in the length of each
of the samples, the high-grade iron having increased in length slightly
more than did the ordinary iron. The experiment so far as it went
tended to show that the growth of cast iron does not necessarily unfit
it for the usual degree of superheat in power-house work.
9 Many grades of brass will crumble when heated in a forge to a
barely visible red, and are quite unfitted to support any stress at such a
temperature. But this characteristic in no way unfits very ordinary
cast brass for saturated steam work, and one should not hesitate to
use a valve of cast brass up to 3 in. in diameter for 150 lb. satu-
rated steam pressure. Three inches is not usually exceeded be-
cause large brass bodies are expensive.
10 Articles have appeared in various publications showing the
unsuitability of cast iron, tensile tests being made before and after the
use of fittings of ordinary iron. Cases of bronze seats dropping from
valves were cited and it was not difficult to prove that something
better than ordinary cast iron was needed for steam of 180 lb. pressure
and 250 deg. superheat. These failures came from two causes. In
the first place the iron itself was not of sufficiently good quality; and,
secondly, the parts were not thick enough. The static stress prob-
ably did not exceed 1000 lb. in the bodj^: but static stress is not the
important load which fittings have to support.
11 Stresses from expansion and contraction within and without
the casting and stresses from pulling up joints no doubt greatly exceed
the static load even in pipe very carefully erected. The troubles are
FITTINGS FOR SUPERHEATED STEAM 1005
aggravated by the action of the steam itself, but it is yet to be proved
that the steam or its high temporatui-e will of itself start cracks in a
properly designed fitting.
12 The ordinary commercial extra-heavy flanged tee, 8 in. inside
diameter, has a body | in. and flanges 1| in. thick. It is made of
common iron, having a tensile strength of about 18,000 lb. Such a
fitting will fail with superheated steam at 175 lb. pressure and 200
deg. superheat. Within a year the inner surfaces will have a network
of cracks, some of which will increase in depth till they extend through
the body. The flanges will crack outward from the bolt holes and the
fitting will become not only leaky but dangerous as well. The writer
has observed just such castings, an analysis of some of them being
given later in this paper. Similar effects have been experienced by a
great many steam users. The fittings are inherently weak to begin
with, so that the failures do not prove that a heavier fitting of better
iron is unsuited for superheated steam work.
13 Within the experience of the writer steel fittings have failed
with superheated steam. Out of twenty-five steel gate valves, 6, 8
and 10 in. in diameter, not more than four were fairly tight after one
year's service, the bodies themselves yielding enough to leak badly.
Some defects in the castings developed allowing steam to pass straight
through the walls, when they left the foundry. Some of these defects
were such that the fittings and valves could not be repaired. In some
cases seats were scraped in once or twice and holes were plugged up
or patched, but the material would not have been satisfactory with-
out this working over. Yet all these castings were iieavy, materially
thicker than the commercially extra-heavy cast-iron product, and had
passed a rigid inspection.
14 Fig. 1 shows a 10-in. steel fitting; the irregular line at A showing
a defect developed after use. The line does not pass clear through
the casting, and no doubt the piece was amply strong to resist rupture
even after the fault developed. Some of these fissures went 3 in.
back and were 5 in. broad. Such a large opening in a shell is objection-
able for there are blow-holes enough adjacent to it to pass steam in
large quantities. Some fittings of this kind were removed from the
line entirely, while others were plugged or patched.
15 No doubt a thoroughly sound steel casting is able to withstand
highly superheated steam. There are several connected to the sys-
tem under discussion. So far as it has been possible to observe, super-
heated steam does not of itself initiate defects and it is not supposed
that the som;id metal undergoes a change, either chemically or
1006
FITTINGS FOR SUPERHEATED STEAM
structurally. But if there is an initial defect, superheated steam is
much more active in bringing out the objectionable features of that
defect. It may well be that the material within the body, and not a
part of the actual metal, suffers through change of some sort. This
material does not add to the strength of a casting but it may serve to
stop up holes if allowed to lie undisturbed.
16 It would appear that some material better than the ordinary
steel casting was desirable for high temperature work. Such a
material is found in gun iron. Gun iron is nothing more than a high-
grade cast iron, which any first-class iron foundry can produce. In
the days of the smooth bore cannon, a few foundries discovered that
it was possible to produce an iron having a tensile strength of 30,000
lb. or more. The government specified it for its guns and it was
called gun iron. Probably a tensile strength of 30,000 lb. is not
needed in steam fittings, but iron of that quality is well adapted for
180 lb. steam with 300 deg. superheat.
Fig. 1 A 10-in. Steel Fitting, the Irregular Line at A Showing Point
OF Failure under Superheated Steam Service
17 From such observations as have been thus far possible it
appears that certain elements in the iron are liable to cause trouble
when present in excess, and perhaps the worst of these is silicon.
It is at present going too far to say that every high silicon iron will
fail and that every low silicon iron will prove successful, but there is
much evidence pointing toward the correctness of such a surmise.
In any event iron of low silicon, low phosphorus and low carbon—
in other words, gun iron — has proved successful.
18 The following analysis shows the character of a casting which
failed at 250 deg. superheat:
Silicon
2 . 40 per cent
Manganese
0.52
per cent
Sulphur
0 . 067 per cent
Total carbon
3.19
per cent
Phosphorus
0 . 94 per cent
Combined carbon
0.25
per cent
FITTINGS FOR SUPERHEATED STEAM 1007
19 A second failure developed in this iron:
Silicon
1 . 98 per cent
Manganese
0.42
per cent
Sulphur
0 . 068 per cent
Total carbon
3.31
per cent
Phosphorus
0.65 percent
Combined carbon
0.24
per eent
20 In each of these cases a sample was taken by drilling a hole
straight into the body after the part had been in service a year or
more and was in bad condition.
21 The following analysis is of an iron that has been successful in
every respect for four years under 300 deg. superheat:
SiUcon
1 . 72 per cent
Manganese
0.48
per cent
Sulphur
0.085 percent
Total carbon
2.45
per cent
Phosphorus
0.89 percent
Combined carbon
0.17
per cent
22 The latter sample is from an 8 in. valve and it is tight today,
no repairs whatever having been made upon the valve during the
four years though the bonnet was taken off once to permit internal
examination. The outer surface of the valve was covered with 85
per cent magnesia insulation, four and one-haK inches thick. The
inner surface appeared sound; a microscope revealed no cracks or
other defects. The unfinished surfaces were struck several sharp
blows with a ball-peen hammer, a hand chisel was driven straight at
the surface and some thick chips were cut off from the rough portion.
If the metal had suffered to such an extent as cast iron is supposed
to suffer, some of the defects would have made themselves manifest.
After these treatments the valve was reassembled and has continued
to perform its work properly.
23 Foundrymen are not^afraid^to^attempt^to- produce^this iron.
No difficulty whatever was encountered in securing bids for valves
made of the following mixture : .
Silicon 1 . 40 per cent to 1 . 60 per cent
Phosphorus 0 . 20 per cent to 0 . 40 per cent
Sulphur 0 . 06 per cent to 0 . 09 per cent
Manganese 0 . 45 per cent to 0 . 75 per cent
Total carbon 3 . 00 to 3 . 25 per cent
It will be noted that the percentages of silicon and phosphorous are
low.
24 There is of course a decided advantage in depending upon
chemical analysis for determining the suitabihty of fittings. A hole
can be drilled at any time in the actual fitting and a few grams of
1008 FITTINGS FOR SUPERHEATED STEAM
sample secured. Very few of us are willing to destroy a fitting to
obtain a test bar, and test coupons cast in the foundry may or may
not represent the actual piece.
25^ Superheated steam was in commercial use in Europe before
the practice had gained its present hold here. England and Germany
were using superheated steam twenty or more years ago. The writer
has not discussed this subject with engineers from abroad, but wishes
to quote briefly those who have.
26 E. D. Dickenson, of Schenectady, on a recent trip abroad
asked a great many manufacturers whether they used steel for their
superheated work and received a negative reply in each instance.
When the manufacturer was questioned in regard to his iron mixture
he shrugged his shoulders and replied that he made his iron fit his
needs, be it gas-engine cylinder or steam pipe.
27 John Primrose, in Power and the Engineer, for June 8, 1909.
states that he discussed the matter with English and German engi-
neers. In one instance a well-known German engineer, who had used
superheat for twenty-five years, was surprised that he had not learned
of the effect of superheated steam upon cast iron. The engineer
promised to investigate the matter in Germany, but he could find
nothing to bear out the contention, and could find no one who believed
that such a thing was possible.
28 It is not the author's intention to state that steel of good
quality will not do for superheated work. Some manufacturers are
putting out fittings of open-hearth steel which are doubtless good;
but any foundry can make gun iron if it will, and delay and uncer-
tainty will be decreased by its use.
29 It is clear that a pipe system for conducting steam should
be designed properly for its work. Expansion stresses must be
taken care of and water must be kept away from it. While steel
has a much higher breaking strength than cast iron, fittings that
are properly designed, properly installed and properly used may be
very well made of a good grade of cast iron.
DISCUSSION ON THE THREE PRECEDING
PAPERS
B. R. T. Collins. Last summer I ran across three valves on pipe
lines from boilers on the main steam header so located as to be sub-
jected to excessive expansion strains as described by Professor Hol-
lis. They were 10-in. extra heavy valves, with ribs running be-
tween the end flanges and also between the bonnet flange and end
flanges. These latter ribs were cracked from 1 in. to H in. deep
on all three valves. In addition one valve had a crack 1 in. deep in
one of the longitudinal ribs, and in one place on the body showed
small criss-cross cracks when examined with the miscroscope.
2 The face-to-face length of this valve was originally 18 in. but
after two and one-half years' exposure to a superheat of 150 deg.
this had increased to 18^ in. This valve was removed, broken up,
and pieces sent to Professor Miller for testing, which showed a ten-
sile strength of 11,300 lb. per sq. in. The iron was very coarse, with
crystals something like J in. across. This valve evidently was of
very poor material to start with, or else it was seriously affected by
expansion strains due to its location or to the superheat. Probably
all three of these conditions had their share in producing the result
obtained.
George A. Orrok. When we first considered the use of super-
heated steam in our power stations a few years ago there had been
developed a type of steam piping'^which most engineers considered
excellent. The piping^itself was ofsteel with VanStone flanges, the
flanges being of sufficient thicknesses to prevent buckling. The
fittings were all cast iron of a [carefully workedout pattern, much
stronger than the ordinary high-pressure^fittings. The valves were
of similar design and the whole piping system was bolted together
with steel bolts of larger size^'and] greater number than the ordinary
extra heavy standard required.
2 This piping system gave absolutely no trouble with saturated
steam. The up-keep of such a system under power station con-
ditions with 200 lb. steam pressure over a period of a number of
years was almost nothing; in fact less than $100 was spent on one
pipe line in about three years time.
1010 DISCUSSION
3 Superheated steam, however, introduced another factor, and
a very important one. From certain tests made by the General
Electric Company it was considered that this superheat might vary
over a range of more than 200 deg. and the temperature strains brought
upon the piping and the valves would be severe. It was finally
determined to make the entire pipe line of steel. The prices on
steel valves and fittings were only a little higher than if of a good
quality of cast iron of the thickness required for the high pressure
and excessive temperature strains. We adhered to the steel piping
with the VanStone joint, but made the VanStone flange of cast steel
from the cast iron pattern. The steel fittings were not as heavy as
cast iron ones of the same size but differed considerably in the detail
of design. The steel valves followed the design of the fittings and
were of various makes, both single and double wedge.
4 Our experience with the steel valves has been good and we feel
that they are giving better satisfaction than was to be expected under
the circumstances. Troubles developed from blowholes, however,
which led to an investigation of the subject about a year and a half
ago. We traced most of the blowhole difficulties to improper mould-
ing, improper gating and to over-oxidized metal in the case of Besse-
mer steel and cold metal in the case of open-hearth steel. The
valves and fittings are about equally divided between Bessemer and
open-hearth steel, all of the former being made, however, on the
baby converter by the Tropenas and Zenzes process. The manu-
facturers understand better today how to handle the work, and the
castings which we are receiving are much better than they were two
or three years ago. I believe our troubles with the steam lines resulting
from superheat are now practically over, and on one steam main in
particular we have done nothing in a year and a half. Whether or
not the valves can be shut off absolutely tight I do not know as we
have had no reason for doing this during the time.
5 After our first installation of steel fittings and valves I had
occasion to look up a number of power stations in which cast iron
fittings and valves had been installed for use with superheated steam.
In one of these stations I saw fittings which had been under the action
of superheat for approximately nine months and had been removed be-
cause of the many leaks which had developed. The castings were
supposed to have been made from the best air furnace iron, but were
swollen and bulged practically all over, the outside being covered
with fine hair cracks. None of the castings had gone to pieces bub
practically all had developed leaks, and were being replaced with
Bteel.
CAST-IRON PITTTNaS FOR SUPERHEATED STEAM 1011
6 At another station a cast iron valve had gone to pieces caus-
ing quite a little damage and many other valves and fittings had been
seriouslj' affected. There were a number of vertical engines in this
station in which superheated steam had been used. All the high-
pressure cylinders had cracked in two or three places and they were
replacing the cylinders and had so arranged their pipe line that no
more superheated steam could get to them. The fittings which I
examined, taken from the superheat line, had all undergone a growth
in size and the outside was covered with fine hair cracks and seemed
very much swollen. Analyses of the metal showed a silicon content
of from 1.88 per cent to 2.33 per cent, phosphorus about 0.7 of 1 per
cent, low manganese and almost no combined carbon. The tensile
strength of the material after its exposure to superheat was in the
case of the iron with the silicon content of 1.88 per cent about 4500
lb. per sq. in.; in the case of the silicon content of 2.33 per cent it
averaged about 8500 lb. per sq. in. We have no means of knowing
what this was when it was first made. Microphotographs of the
etched surfaces of this metal show the essentially open character of
the iron. In this particular station the superheaters have been
removed and their troubles have ceased.
7 In view of the many and excessive strains likely to come on a
pipe main with 200 lb. pressure and more or less superheat I have
not felt that we are justified in installing cast iron valves and fittings.
Even with saturated steam at the above pressure and with the
length and size of mains which we are using today in our modern
stations it seems to me that the extra expense for steel is justifiable
and might probably be saved many times over in the cost of up-keep
during the life of the station.
8 A few years ago it was the general impression that superheated
steam could not exist in the presence of water. This statement has
been made many times and no longer ago than at the Annual Meet-
ing. That this idea is fallacious is, I think, the generally accepted
belief today, and we have good evidence that it is possible in a
steam pipe carrying steam at 200 lb. pressure and 200 deg. super-
heat to have a stream of water flowing along the bottom of the
pipe. In this case the bottom of the pipe would be at a temper-
ature of possibly 380 deg. fahr., while certain other portions of the
pipe in contact with the superheated steam might have a tem-
perature between that of saturated steam and the maximum tempera-
ture of superheat.
9 Regarding the difference between European cast iron and
American cast iron, it has been my impression that the pig-iron
1012 DISCUSSION
manufacturer here is always trying to make a grade of iron which
will command a high price in the market. This iron must be an
open iron with reasonably high silicon and almost no combined car
bon, the carbon content being in the graphitic state. This iron
will sell readily. If the quality of the iron fell off, and because of
a lower silicon content more of the graphite was converted into
combined carbon, the iron would become harder — more difficult to
machine — and would not command as ready a sale. In Europe,
it is my impression that they make much harder iron and are willing
to spend the money to machine it. In America we demand an open
iron that can be machined easily.
10 If Mr. Mann continues his researches and considers his test
specimens in the light of the volumetric composition of the iron;
that is, the volume which the compounds of iron and silicon and
of iron and carbon occupy in the cast iron, in comparison with the
volume occupied by the iron itself, he may find some interesting
results.
1 1 Referring to air-furnace iron, or gun iron as it has been cal-
led, I think the great difl&culty is the fact that it is almost impossi-
ble to control the regular composition of the product. The reverber-
atory furnace, while a comparatively simple^ piece of apparatus,
is remarkably delicate, and uniform results are obtained only when
the very best of care is taken. It is a comparatively easy thing to
refine high silicon iron to some kind of refined iron, but it is a much
harder thing to get a uniform result from each heat.
W. K. Mitchell, The following notes are taken from several
years' personal experience with superheated steam and its effect on
cast-iron valves, fittings, etc.
2 Our first intimation that cast-iron fittings and valves gave
trouble under superheated steam conditions occurred about three
years ago and came in the nature of a surprise, as we had been using
superheated steam for some years previous.
3 The first case was infa railway power plant for a high-speed
electric line. The plant had been running for several months under
a fairly constant load, but owing to'a falling off in traffic it was decided
to cut^down the service to one-half or less, which made the load quite
variable. Three months after this had been done the trouble with
the fittings\nd valves began to develop. It was first found that
the valves could not be closed tight, and gaskets were giving trouble.
Then fittings began to show signs of weakness, cracks appearing on
the outer surface.
CAST-IRON FIITINQS FOR SUPERHEATED STEAM 1013
4 Fortunately these cracks never extended through. In an 8-in.
by 6-in. double tee, the metal of which was about | in. thick, the
cracks did not extend more than half way through, which' indi-
cate , that there is no advantage in very thick castings under such
conditions. The most serious of these cracks occurred at the junc-
tion of the flange and fittings, and kept growing to so alarming an
extent that several fittings were replaced. It was then noticed that
the old fittings had lengthened considerably. The original length of
some 8-in. by 6-in. double-tee fittings was 35 in., and when taken out
and cooled they measured 35f in. to 35 1 in. They had been in service
about nine months. Open hearth cast-steel fittings and valves were
substituted for those of cast iron and have been working satis-
factorily ever since.
5 About the only information we could get bearing on the cause
of this growth was from a paper by A. E. Outerbridge, read before the
mining and metallurgical section of the Franklin Institute in January
1904. Mr. Outerbridge stated that by repeated heating and cooling
of bars he had caused the metal to grow to an almost incredible
extent. He exhibited a test bar, the original dimensions of which
were 1 in. square cross-section and 14lf in. long, which had been
heated some 27 times to a temperature of about 1450 deg. fahr.,
and cooled again by various methods, some slow and some fast,
until at the end of the treatment it had grown to a length of 16^
in. and a cross-section of 1| in. square. The similarity between the
action of the fittings above referred to and the test bars which Mr.
Outerbridge exhibited caused us to investigate further along similar
lines.
6 The railway plant was designed for a steam pressure of 175
lb. per sq. in. and superheat was intended to be 150 deg. fahr. That
this temperature had been greatly exceeded, however, was made evi-
dent by the discovery of a board that had been charred by contact
with the steam trap which rested on it. This trap was connected
to a drip pipe running from one of the elbows next to the strainer on
a steam turbine and was about 10 ft. below the elbow. Investigation
showed that the trap had been so hot that its legs had burned holes
through the board until the trap was not resting on the board at all
but was suspended by the pipe.
7 The president of the company that installed the superheaters
said that while they were built to give an average of 150 deg. super-
heat, "the real question was not one of the amount of superheat but
of velocity. " This seems reasonable when one considers that if the
1014 DISCUSSION
load should fall very low, the velocity of the steam through the super-
heaters would be considerably reduced and its temperature corre-
spondingly raised. Again, a sudden increase in the demand for steam
would result in a rapid flow through the superheater, and steam at
much lower temperature, and these recurring changes of temperature
must necessarily cause rapid changes in the lines due to expansion
and contraction. We therefore concluded that in this particular
plant, at least, the damage to the fittings and valves was not caused
by the high temperature itself, but by the constantly changing tem-
peratures due to the change of load.
8 Our contention that the damage was due to variable tempera-
tures seemed to be borne out by the fact that in a cotton mill plant
installed three years previously, where the steam requirements for
pressure and superheat were higher than those mentioned (the pres-
sure being 200 lb. per sq. in., and superheat 200 to 250 deg. fahr.),
the fittings and valves were of regular cast iron, and there had been
no trouble to speak of. The load was practically constant, however,
varying not more than 15 per cent at any time.
9 On account of the discussion in several publications in the spring
of 1908 regarding the disastrous effects of superheat on cast iron, the
owners of the mill grew anxious about their piping and asked the
writer to look over the system. He found everything normal; the
fittings were tight, valves could be operated freely, and in a general
way the plant was in good condition. The first installation had been
made in 1903, and a second one in 1906, using the same class of
fittings and valves.
10 The writer suggested that measurements be taken of all the
fittings and valves in the plant and records kept of changes. The
original dimensions were determined as closely as possible from the
patterns and records of construction, and beginning with July 1908
records were kept of the dimensions of the fittings for a period of
nine months. Although the changes in the dimensions were slight,
the increase in length of certain of the fittings was such that it was
thought unsafe to continue them in use and steel fittings were sul)-
stituted throughout. Most of the valves, however, are still in service
and there have been no failures in either fittings or valves. Th(^
following will give an idea of the changes which occurred from the
dates of installation to the last date given : —
A 12-in. by 10-in. by 8-in. by 6 in. cross installed in 1903
measured 24. in. in length, and in March 1909 measured
24fiin.
CAST-mON FITTINGS FOR SUPERHEATED STEAM 1015
A 10-in. by 8-in. by 8 in. tee increased in the same time from
24 in. to 24 tfe in., and
A 12-in. by 12 in. by 8 in. by 6 in. cross, 20 in. long when
installed in 1903, was 2O/4 in. long in March 1909.
A 12-in. by 10-in. by 8-in. by 6 in. special cross, 50 in. long
when installed in 1906, had grown to 50^ in. in 1909,
also
A 10 in. by 8-in. by 8-in. tee, 24^ in. long, measured 24|i in.
in March 1909.
1 1 These facts seem to show that even at high steam temperatures,
if cast iron can be kept at a uniform temperature and not cooled off
too frequently or too rapidly, it will meet the requirements of super-
heated steam for a long period; but if the temperature is subject to
frequent changes such as occurred in the railway plant, the cast
iron will becoirie disintegrated and ultimately fail within a short
period of time.
12 In another street railway power house which had been in opera-
tion for a number of years with saturated steam at 200 lb. pressure,
cast-iron fittings, valves and pipe were used successfully. During
1906 fourteen new boilers were added, making a total of thirty-two.
The new boilers were equipped with superheaters intended to super-
heat to about 50 deg. fahr. New piping was installed similar to the
old, with cast-iron fittings and valves, and steel pipe with steel flanges.
The fittings were unusually heavy and strong. In the original instal-
lation of this piping a white metal gasket had been used, about A in.
thick, which was very satisfactory for saturated steam. These gas-
kets had a melting point of about 650 deg. fahr., and no sooner had
steam been turned in from the new boilers than the gaskets began to
melt, and in the course of a month or six weeks it became necessary to
replace every one with material that would stand the temperature of
the superheated steam. As the majority of the boilers had no super-
heaters it was hard to understand how sujficient superheat could be
generated by the new boilers to do any harm.
13 Two years later a 16-in. tee in one of the connecting pipes
between the main headers was found to be leaking. The leak becom-
ing worse, the covering was taken off and the tee was found to be
covered with small cracks or fissures similar to the cracks that had
occurred in the fittings taken out of the power house first mentioned,
except that a few of the fissures had worked through to the inside.
The tee was replaced with a new one of the same material and dimen-
sions. When the defective fitting was examined it proved to be some-
1016 DISCUSSION
thing of a curiosity. Its original dimensions were 31 in, face to face by
15| in. centre to face. It had grown on one side to 31| in. and on the
other side to 321 in. The flanges, which were originally 25 in. in diam-
eter, had grown to 25f in. and as they had been bolted to steel flanges
that had not changed under the superheat conditions they had become
dished to a depth of about | in. The original thickness of the body
of this fitting, as near as could be determined from the pattern, was
about If in., but where the surface cracks were most numerous care-
ful measuring gave a thickness of almost 2j in. Of course, there is
always the possibility of the core moving when a casting is being made,
but the thickness of this fitting was quite uniform ithroughout.
14 In this plant, the trouble did not stop with the fitting. Several
valves began to show cracks and were replaced. Then the high-pres-
sure cylinders of the engines became affected and several had to be
renewed. The engineers finally decided to take out the superheaters
and the plant is now running without superheat. This is a typical
street railway plant, subject to changes of temperature similar to the
one first mentioned. Fig. 1 shows the original dimensions of the 16-in.
tee and its dimensions after coming out of the line.
15 It is of interest to note that at the same time these superheat
boilers were installed in the railway plant, similar boilers with engines,
piping, valves and fittings of the same type and material were instal-
led in a lighting plant in the same city, where practically no trouble of
any kind had developed. This_seems to indicate that a much more
constant load is maintained
16 It is my opinion that in plants where the load is constant and
the temperature of the steam therefore constant, properly designed
piping with cast-iron fittings of good material will do the work satis-
factorily and be safe for a long time. I believe there is no advantage
in using cast-iron alloys known as semi-steel, ferro-steel or gun iron.
In the railway plant first referred to, the fittings were of cast iron from
one foundry; gate valves of semi-steel from another; and stop, check
and emergency stop valves also of semi-steel from a third. The
results in each case were practically the same. A specimen from an
8-in. by 6-in. double tee which had been in service about nine months
gave a tensile strength of 13,750 lb. per sq. in. The chemical analysis
of this piece gave the following: Carbon, 2.502; Phosphorus, 0.461;
Sulphur, 0.083; Silicon, 2.435.
17 Two test pieces from the 16-in. tee showed tensile strengths of
4970 lb. and 4340 lb. Chemical analysis: Silicon, 2.33; Siilpliur,
0.07; Phosphorus, 0.68; Manganese, 0.39; Total Carbon, 3.18.
CAST-IRON FITTINGS FOR SUPERHEATED STEAM
1017
1018
DISCUSSION
18 In Fig. 2 is shown the first fitting listed in Par. 10, taken from
the cotton mill plant where it had been in use several years under
superheated steam, the temperature of which, however, was nearly
constant. This fitting was tested under hydraulic pressure. At
first the pressure was put up to 1100 lb. per sq. in., when the gas-
kets leaked and the pressure had to be reduced to zero in order to
tighten up the gaskets. The fitting was then tested again and
broke at a pressure of 1250 lb. There were no serious defects in the
fitting. One small surface crack was of so little moment as not to
Fig. 2 12 in. by 10 in. bt 8 in. by 6 in. Extra-Heavy Cast Iron Fitting
IN Use with Superheated Steam, 1903-1908 ,^i,,
require special attention. A testpiece from the fitting showed a
tensile strength of 15,900 lb. per sq. in. The chemical analysis
was as follows :
Total carbon 3 . 05
Phosphorus 0 . 769
Sulphur 0.06
Silicon 2.07
19 It will be noted that the silicon in this specimen is lower than
in the two castings just mentioned and I think this had a good deal
to do with the case as well as the fact that the load and tempera-
ture were constant.
20 While open-hearth steel castings seem to be successfully used
under superheated steam conditions, I do not believe they will last
CAST-IRON FITTINGS FOR SUPERHEATED STEAM 1019
indefinitely because of their extreme thickness. There must be
changes taking place similar to those in cast iron, due to temperature
changes, but the ductility of open-hearth steel wll undoubtedly delay
the process of disintegration for a longer period. The material which
we recommend and use today for high-pressure superheated steam, is
wrought steel throughout, with welded nozzles instead of fittings, and
steel flanges, using bends in all cases in preference to short elbows.
John Primrose. During the past eight years the writer has been
in close touch with many plants, containing upwards of fifteen hun-
dred installations using superheated steam, and in a position where
troubles would be promptly reported to him. Almost without excep-
tion these plants use cast-iron fittings in their pipe connections. The
fact that no one of these plants has reported troubles with its
fittings is in striking contrast to the comparatively few instances
where superheated steam has been charged with being the cause of
trouble with cast-iron fittings. In order that there should be no
doubt about the absence of trouble due to superheat, letters were writ-
ten to ten concerns known to have been passing superhe ated steam
through cast-iron fittings for the past eight years, at from 100 to 150
deg. superheat, asking the following questions:
Question One. Are not the tees, elbows and valves of cast-iron
in the branch and main steam lines leading from the boilers?
Seven answered yes, two replied that some fittings and
valves were of cast iron and some of cast steel, and one
replied that while the fittings were originally of cast iron
some tees had been changed to cast steel, but stating posi-
tively that the change was not made because of any ill
effects of superheated steam.
Question Two. Are fittings of extra heavy or standard weight?
Nine replied that they used extra heavy fittings, and one
standard weight.
Question Three. What steam pressure do you ordinarily carry?
One used steam pressure of 100 lb., six used 150 lb., one
165 lb., one 185 lb. and one 200 lb.
Question Four. Have you ever noticed any injurious effect of
the superheated steam on valves or fittings? Eight
answered no, one that no trouble was experienced in fit-
tings, but that valves with cast-iron bodies and brass seats
were difl&cult to keep tight, and one reported no trouble
further than the baking of a hard deposit on inside.
1020 DISCUSSION
Question Five. Have you ever found it necessary to replace
any of these valves or fittings with cast steel? Eight
answered that no fittings or valves had been replaced on
account of superheated steam. One answered that they
had replaced no fittings, but some globe valves, and one
answered that they were replacing some fittings with cast
steel, but upon further inquiry it was found that this was
not because of the ill effects of superheat, but because the
steam mains were being changed to contain VanStone
joints and they wished to change the fittings to standard
length and deemed it advisable to use cast steel.
Question Six. Of what material are the gaskets in the steam
fine? Seven use corrugated copper or bronze, two sheet
packing, and one asbestos.
2 The chief engineer in charge of a plant in the middle west, of
some 20,000 h.p., writes that nothing has developed in any of the cast-
iron fittings to show that they are in any way affected by the use of
superheated steam. This plant has been in operation about five
years.
3 Such evidence as the foregoing proves pretty conclusively that
superheated steam does not have an injurious effect on cast iron.
There seems to be no very good reason why it should. There is noth-
ing extraordinary in the fact that several cast-iron fittings have failed
when passing superheated steam. The failures were probably due to
inferior metal, or to strains developed by expansion or contraction of the
pipe lines, as suggested by Professor Hollis and Mr. Mann. These
are much more plausible theories than that superheated steam at a
temperature of 500 deg. to 600 deg. fahr. has any effect on the metal.
In investigations by Mr. Outerbridge and Professors Rugan and Car-
penter on the growth of cast iron when repeatedly heated, their experi-
ments were started at 900 deg. C. or 1652 deg. fahr. Such instances
as the growth of grate bars, etc., are all at temperatures far exceeding
anything used in superheated steam work for power plants.
4 Samples of cast iron taken from fittings passing superheated
steam for years have been polished and micro-photographed before
and after etching, and compared with samples treated in the same way,
taken from fittings passing saturated steam. The report states that
there is no evidence of a change in the carbon conditions, or of exposure
to superheated steam, and in support of this a well known foundry-
man gives his opinion that a temperature below 900 deg. fahr. would
not produce any effect in cast-iron.
CAST-IRON FITTINGS FOR SUPERHEATED STEAM 1021
5 The tests of the famous Crane valve so often quoted are no proof
of superheated steam being responsible for the failure. Test bars from
the broken valve were compared with test bars taken from the same
heat that the valve was made from, and the valve was said to have
weakened. This is no real test, because castings from different parts
of the same heat, or, in fact, different parts of the same casting are
known to vary in strength, and it is quite likely that fittings passing
saturated steam, if compared on the same basis, would be found to
have suffered greatly from the effect of saturated steam ! It is unques-
tionably true that this valve must have been subjected to other
influences besides superheated steam. It is rather remarkable that
the body of the valve is said to have been weakened more than the
flanges — the reason given being that the metal of the body was nearer
the superheated steam. Is it not more reasonable to suppose that the
metal of the body weakened more than that of the flanges because it
was subjected to greater fatigue on account of expansion and con-
traction of the pipe?
6 The writer's experience with a great number of steel fittings used
for pressure parts of superheaters exposed to hot gases, has led him to
conclude that the metal in steel castings is anything but satisfactory
for fittings. It unquestionably has greater tensile strength than cast-
iron, which appears to be its only advantage. On the other hand, it is
difficult to get steel castings sufficiently|homogeneous to hold the pres-
sure. A large percentage of castings are "doctored " before leavingthe
foundry, but new openings frequently develop on test after machining,
and even after the castings are in place, causing the charge to be made
that the castings have not been tested before sending out. This fact is
further evidence of strains developing in service, other than those pro-
duced by internal pressure, which open up cavities or spongy places
not discovered by shop test. Steel 'castings [vary greatly in tough-
ness, as shown by the great variation of elongation on test; others are
so hard that machining is very difficult. While they can be bought
on very careful specifications to guard against these faults, there is
always the chance of porosity. The high tensile strength of the steel
is not a necessity, and cast-iron made to careful specifications is
amply strong. It machines well, is not porous, and can be relied on
to hold the pressure.
7 Care should be exercised in the design of pipe lines to guard
against straining the fittings from movement of the pipe due to expan-
sion and contraction. Where long radius bends are the means of tak-
ing up this movement, pipe of the lightest possible weight consistent
1022 DISCUSSION
with safety should be used, thereby lessening the force required to
spring the pipe. With properly connected flanges, full weight pipe is
amply strong for all ordinary working pressures, and if drawn tubing
is used, even lighter metal may be adopted. In this connection the
design and arrangement of the superheater is of great importance, and
should be such that sudden and frequent changes in the temperature
of the steam do not occur; otherwise the changes in the length of the
pipe will be more frequent, resulting in a more rapid fatigue of the
metal of the fittings.
8 A better way of taking care of expansion than with long radius
bends, is to use ball and socket expansion joints, which have the addi-
tional advantage of reducing the amount of piping.
9 The writer agrees with Professor Hollis in charging strains due
to expansion and contraction with the failure of certain fittings, and
with Mr. Mann when he charges inferior fittings with the cause of
failure in other cases and recommends the use of a good cast iron con-
taining a percentage of steel scrap for fittings passing superheated
steam. This is entirely in accord with the writer's experience.
H. S. Brown believed the troubles with cast iron would be
eliminated if the temperature could be kept constant, and further
said:
2 I think the discusssion boils down to this, that under certain
conditions steel castings will give a more satisfactory performance
than cast iron. The company with which I am connected has found it
necessary to replace a large number of cast-iron fittings with steel
castings, where superheated steam was used; and the performance
of these steel fittings, under the same conditions under which the
cast-iron fittings M^ere working, has been satisfactory.
3 In a large number of other plants where cast-iron fittings are
used with saturated steam,, the design of the piping is such that the
stresses set up on account of expansion and contraction are very
much worse than in this system; and we do not get into troubles
as we did in using superheated steam. It may be that steel fittings
will form a more practical and less expensive way of taking care of
the conditions set up by the use of superheated steam than elabor-
ate precautions in the wa}' of expansion joints and the like.
E. H. Foster. Cast iron is much too useful a metal to receive gen-
eral condemnation for steam pipe fittings, whether for superheated or
saturated steam, without very good reasons and the writer is firm in
this opinion that such reasons have not yet been advanced.
CAST-IRON FITTINGS FQli SUPERHEATED STEAM 1023
2 Having devoted the greater part of his time for the past ten
years to the study of superheated steam and the manufacture of
superheaters, the writer has eagerly followed up every report of the
failure of a steam pipe fitting, where superheated steam was used, and
it can fairly be said that no instance has yet occurred where the weak-
ness has not been readily explained by the poor quality of the iron, or
by lack of provision for expansion and contraction without straining
the metal. The many instances where cast-iron fittings are habit-
ually subjected to steam of varying degrees of superheat up to final
temperatures close to 1,000 deg. fahr. leave no doubt that good cast-
iron is equal to, if not better than, any other metal for making steam
fittings for superheated steam as well as for saturated steam, especially
in smaller sizes.
3 In the writer's experience it is as important to have regard to
the mixture of the iron to be used in cast-iron fittings as it is to secure
the proper mixture for concrete work.
4 The suggestion that better provision should be made for free
expansion and contraction of steam pipes, is, in my opinion, very
much to the point. More care applied to this feature of the design
of power plants would remove entirely from the shoulders of cast-
iron the odium of being unsuitable for carrying superheated steam.
L. B. Nutting stated that superheaters installed by his
company nine years ago, and since then in constant use, have caused
no trouble and have not changed their dimensions. These super-
heaters were made entirely of cast iron, the tubing having a smooth
bore and corrugated exterior.
2 He also reported a great many superheaters installed and
in operation delivering steam at a temperature of 1000 deg. fahr
On these the users have employed, without any distortion or without
any evidence of weakness developing, standard makes of cast iron
valves (globe valves and angle valves) under 1000 deg. final tem-
perature. But the temperature is maintained at 1000 deg., without
a variation of 25 deg. These illustrations seem to trace the cause
of the trouble with cast-iron fittings directly to widely fluctuating
temperatures.
3 In regard to Mr. Mitchell's suggestion that provision should
be made to obviate troubles from varying temperatures on cast iron
Mr. Nutting asked, why not make provision to keep the tempera-
ture constant. The art of superheater construction has advanced
to such a point that a uniform temperature should safely be counted
1024 DISCUSSION
on. The plant Mr. Primrose referred to, a 20,000-h.p. boiler plant
has a record of variation not exceeding 10 deg. either way from^the
desired amount at any time during the year, although the loads
have a fluctuation o^" from 5000 to 35,000 kw.
Andrew Lumsden.' In one case we have eight boilers of the Bab-
cock & Wilcox type, equipped with heaters, part of which were made
and installed^by the boiler company. These boilers were all con-
nected to one 12-in. main through long radius bends, valves, tees, etc.
On the superheaters installed by the boiler company there were
usually 150 deg. of superheat and on the others about 90 deg.
2 When this plant had been'^in service about two^years some of the
fittings in the main were found to leak just back of the fillets and
small cracks were discovered extending around one side of the tees.
Some long bolts were made to go the^whole length of the tees and
through the end flanges, using them to make the joints. Steel tees
were also ordered of the same dimensions as the cast-iron ones to
replace all the fittings in the main. When the old fittings were
removed, however, it was found they were from fin. to | in. longer
than when first installed. This is a turbine station and they have had
quite a little trouble with the admission valves on some of the tur-
bines, those^directly opposite the boiler carrying the 150 deg. super-
heat giving by far the^mosttrouble.
3 At another plant there are boilers of the Babcock & Wilcox type
and Curtis steam turbines, installed seven years ago, with a separately
fired superheater on which exhaustive tests were made. The tem-
perature of the steam leaving the superheater reached as high as^750
deg. The superheater was run for about six months and atjthe
end of that time all the copper gaskets in the main were destroyed and
the joints had to be remade. The^superheater was shut down^and
has not been operated since.
4 The writer visited this plant a few days ago andiound that no
large joints had been made since the superheater was shut down.
There are about ninety joints ranging from 8-in. to 12-in. diameter and
none have leaked, but three 12-in. valves were' leaking badly through
the body on the under side and about the point where the seat rings are
screwed in. These valves are laid on their sides and are on the boiler
side of the superheater and have never had superheat in them nor in
» President, Lumsden and Van Stone Co., 69-71 High Street, Boston,
Mass.
CAST-IRON FITTINGS FOR SUPERHEATED STEAM 1025
the writer's judgment can their troubles be due to expansion as par-
ticular care was taken to allow free^movement in all the piping of the
plant.
5 At another plant where they have Babcock & Wilcox boilers.
Curtis turbines, etc., they have carried 150 lb. steam pressure and 150
deg. of superheat for about six years, with cast iron fittings, etc., and
have had no trouble with the fittings. The valves have given them
some trouble with loose seat rings and by being badly cut, and some of
them have been replaced.
John C. Parker. Professor Hollis draws attention to the neces
sity for greater allowance for expansion in piping for superheated
steam. This is important and where sufficient flexibility cannot be
put into the design expansion joints should be installed. My experi-
ence accords with his statement that cast-iron fittings have been
largely and successfully used for superheated steam.
2 Six or seven years ago I was called on to furnish superheaters
with some of our boilers but could find none in the market to meet my
ideas of what a superheater should be. A design was worked out and
forty or fifty thousand horsepower of boilers have been built with
these superheaters. Two plants are above ten thousand horsepower.
In six years experience we have^had no trouble either with cast-iron
or steel fittings, valves or cylinders. I ascribe the result to the steadi-
ness of the superheat and to the fact that condensed steam in the
superheater Js not intermittently carried into the steam line.
3 In one of our first installations the men started to flood the
superheaters without my knowledge whenever the boilers were banked
and the result taught me the effect of suddenly injecting water at 360
deg. fahr. into piping and headers which had been raised to 500 deg.
fahr. Leakage started at the joints but stopped as soon as the prac-
tice was stopped.
4 I believe there is no connection between the troubles which I
have been cognizant of with some designs of superheaters and the
expansion of the steam mains. I believe the troubles have been due
solely to fluctuations in temperature and temperature shocks caused
by frequent injection of condensed steam from incorrectly designed
superheaters. I recently went into a plant where a superheaer had
been in use for about four years. It was an independently fired
superheater and/he engineer had had so much trouble with piston
rings in an engine with poppet valves designed especially for super-
heated steam, that he had cut the superheat from 600 deg. fahr. to
500 deg. fahr. and then to 400 deg. fahr.
1026 DISCUSSION
5 There is a superheater in the market that uses cast-iron to pro-
tect wrought iron tubes. I have seen some of these removed from
boilers on account of overheating and, while the cast-iron had been red
hot it had cracked less than some steam pipe fittings which had been
subjected to water jets and fluctuations under 600 deg. fahr.
6 I do not think conclusions can be drawn from Professor Miller's
experiments. It would require at least half a dozen tests of the same
sample of cast-iron at progressively increasing temperatures and
periods to obtain results of value. Ten of the tests show increased
strength of cast-iron while all the steel has lost strength. One steel
test (100,000 lb.) is unreasonable.
7 We have sixteen 800 h.p. boilers with superheaters directly over
the fire running at 175 lb. pressure and up to 170 deg. superheat with
no such trouble as Mr. Mann mentions in Par. 12.
Albert A. Gary said that after investigating a number of plants
having trouble Avitli the use of superheated steam, he had been led
to the conclusion that many if not most of their troubles have been
due to bad design in the piping arrangements.
2 Far greater care and better judgment is called for in design-
ing pipe systems for superheated steam than for similar systems
using saturated steam, as the strain due to expansion and contrac-
tion is greatly increased. The piping on each side of every offset
should be carefully considered to see that excessive stress is not thrown
upon the flanges and threads bj^ the lever which is developed there.
Several special forms of flanges which avoid the screw connection
are now used to excellent advantage, with high superheat.
3 Continued flexing on one side of the flanges of fittings, due to
the cooling and high degree of heating of the pipe system as steam is
turned on and shut off will cause ruptures not unlike those shown in
the illustrations accompanying these papers and will be apt to change
the internal structure of the metal itself.
W. E. Snyder. I may be able to touch upon certain partic-
ular phases of the papers and the discussion, in such a way as to
contribute some of the results of actual experience covering a wide
variety of conditions and several years' practice. The consideration
of this subject in the discusssions seems to have broadened to include
the effect of unequal heating of metal and also the designs of systems
of steam piping. These matters are both directly related to the
CAST-IRON FITTINGS FOR SUPERHEATED STEAM 102?
use of cast iron fittings for superheated steam, as failures may in
some cases be due to the improper design of the steam piping ; also
under some conditions to unequal heating of such irregular castings.
2 A common connection between engine and main steam pipe
is by a branch running horizontally and at right angles to the main
pipe, out directly over the high-pressure cylinder and turning down
by a bend to connect with the throttle valve on top of the cylin-
der. When this branch is long the expansion in the steam line does
not exert any harmful effect in the throttle valve, but the expan-
sion in the branch is taken up by the change of curvature of the bend
over the throttle valve, and this puts a strain directly on this valve.
In two or three instances where this kind of connection was in use,
the throttle valves were cracked immediately under the flange, and
serious accidents narrowly averted.
3 Where the branch to the engine is short the expansion in the
branch itself does not require any consideration, but the longitud-
inal movement of the steam main due to its expansion and contrac-
tions, transmits strains through the branch pipe directly to the
throttle valve and flange on the steam chest. In one instance this
resulted in a very serious accident, as the cast-iron flat top of the
steam chest was broken in by the expansion of the main steam pipe
several feet away.
4 In another installation a 24-in, cast-iron Y with an 18-in.
branch split in the fork of the Y, while under 150 lb. steam pressure.
Fortunately it was possible to take the line out of service before the
fitting exploded, but it was a very narrow escape. All the accidents
mentioned above were the direct results of the installation of sys-
tems of steam piping without proper consideration of the effects
produced by expansion and contraction. All were in systems using
saturated steam, and they emphasize the necessity of using great
care in arranging the piping that the expansion and contraction may
take place without throwing the severe strains on the cast metal
members, which are always liable to failure under such conditions.
Consideration of this feature of design is of still greater importance in
piping systems using superheated steam, on account of the higher
temperature used and the consequently greater expansion. The
avoidance of expansion strain on-castings in a system of steam piping
is of fully as great importance as is the selection of the material from
which these castings are made.
5 The effect of unequal heating of metal has })een investigated
by engineers in the French Navy (See Marine Boilers by Bcrtin &
1028 DISCUSSION
Robertson, p. 201). The theory advanced there, which seems rea-
sonable and is also confirmed by experience, is this: When one side
of a piece of metal or a boiler tube is heated to a higher temperature
than the other, the hot side tends to expand, and the expansion is
resisted by the metal on the cold side. This condition puts the metal
on the hot side in compression, and the metal on the cold side in
tension, and if the temperature difference is great enough the metal
will be strained beyond the elastic limit. When the hot side is al-
lowed to cool it is shorter than the cold side because of the strain
beyond the elastic limit which has been undergone by both sides.
This results in the piece taking a permanent set, or becoming "bow
shaped" away from the side that has been heated.
6 The bend away from the fire, of boiler tubes in some types
of boilers after they have been in service for some time, seems to be
a good example of the results of unequal heating. Another example
is the cracking of the large cast iron mud drums used in some types
of water tube boilers. Under ordinary operating conditions, in
boilers having vertical baffling, the hot gas does not come in contact
with the mud drum of the boilers until it has passed at least twice
across the tubes, and has thus been greatly reduced in'^temperature.
At times, however, holes are formed in^the front baffle, or through
the top of the bridge wall allowing the hot gases to pass directly from
the furnace to the back part of the boiler setting, where they strike
the cast-iron mud drum, heating it to a considerably higher tem-
perature on the front side than on the side away from the fire. These
conditions have resulted, in a number of instances, in the mud drum
cracking perpendicularly to its axis, causing serious accidents.
7 This matter of the unequal heating of metal is one of the most
serious with which designers of large engine cylinders have had to
contend. Features of the design of pipe fittings are very similar
to those mentioned in the design of gas-engine cylinders, the irregular
castings having flanges and other forms of construction which make
it practically impossible to avoid having the metal considerably
thicker in some places than in others. This irregularity causes
internal strains in the metal when heat is applied to one side. It
has been the experience of European designers of gas-engine cylinders
that one of the greatest difficulties they have had to overcome is this
one of distributing the metal so as to avoid the small cracks resulting
from irregular expansion, which destroy the cylinder. Where trouble
has occurred in the use of cast iron for large fittings in superheated-
steam piping, it is possible that the experience of the gas-engine
engineerstwill suggest the remedy.
CAST-IRON FITTINGS FOR SUPERHEATED STEAM 1029
8 As bearing upon the use of cast iron for superheated steam,
particularly upon the much discussed question of the possibility
of having superheated steam that is in contact with water in the
boiler, an experience of the speaker may be of interest.
9 A large furnace used for heating slabs for a plate mill was equipped
with a small vertical Cahall boiler for the purpose of utilizing part of
the waste heat. This furnace was fired vnth under-feed stokers,
using forced blast, so that it was possible to obtain a very high tem-
perature both in the furnace and in the boiler, also in the stack.
The boiler was set in the usual way for waste heat, i.e., with the large
end down. The steam pipe was connected to a flange on the top of
the boiler, this connection being made inside the conical-shaped base
of the stack which rested on top of the circular boiler setting. By
this arrangement a cast-iron elbow on this steam pipe, and about two
or three feet of pipe on each side of the elbow were located in the
hot gas directly over the upper drum of the boiler.
10 Tests on this furnace and boiler were continued for two weeks,
observations being taken every 30 minutes. Frequently the stack
temperature would rise to 1000 and 1 100 deg. f ahr. The thermometer
on a Carpenter throttling calorimeter, connected to the steam pipe
just outside the stack breeching mentioned above, ranged from 380
to 600 deg. fahr., depending on the rate of working of the furnace.
This variation at times occurred very rapidly, and at the time the
thermometer readings were high, the steam escaping from the calori-
meter was as completely invisible as though it were natural gas.
For 12 hours in succession the average superheat of the steam
was 140 deg. and during the entire time the tests were being made, the
superheat of each 12-hour period averaged 120 deg. or over; the
steam pressure being about 95 lb.
11 This boiler has been in operation under conditions similar to
the above for at least 15 years. The cast-iron elbow and flange at the
top of the boiler, although subjected to such severe service as that
described, has never given any trouble. A number of other Cahall
boilers using blast-furnace gas, with steam pipe connections made in
the same way, have been in operation for about the same length of
time, and no troubles have occurred due to the fittings. The service
under blast-furnace gas conditions are not so severe as the heating-
furnace installation described above, on account of the stack tem-
perature being somewhat lower, from 700 to 900 deg. The heating-
furnace conditions mentioned above are unusually severe and for
that reason have been described fully. It may be added that the
1030 DISCUSSION
boilers using blast-furnace gas produced superheated steam, notwith-
standing that the water level was only a short distance below the
connection to which the calorimeter was attached. (This must not
be understood as being a special feature of the Cahall boiler, as in
fact it is only incidental to its operation under these conditions.)
J. S. ScHUMAKER Called attention to the fact that in the cases
cited no difficulty had been experienced with the fittings of super-
heaters, cast iron or otherwise, but that there had been a great deal
of difficulty with steam-pipe fittings. This seems to be the result of
high temperature on one side of the fitting only. In the super-
heater itself the temperatures are balanced, to some extent at least.
Dr. D. S. Jacobus. In a large power plant that I have in mind,
where the fittings are all of cast iron, and where the superheat averages
150 deg. fahr., repeated examinations have failed to reveal any dete-
rioration. In other cases, however, where there has been less super-
heat, and even where a single boiler with superheated steam has been
connected into a common main with a number of other boilers furnish-
ing saturated steam, there has been every indication that a small
amount of superheat has had an injurious effect. It therefore seems
that a difference in the quality of the cast iron may affect the
results, and by maldng a careful study of the matter and knowing the
analysis of the cast iron there is a possibility that its action under
superheated steam may be predicted. In the meantime, we are
furnishing cast-steel fittings for all superheated steam work, as we do
not know of a single case of the failure of such fittings that can be
attributed to the action of the superheat.
2 The stresses due to expansion, as pointed out by Professor Hollis,
may tend to produce failures. In the case of fittings broken in super-
heated steam lines we have found there was a stress at the point of
rupture entirely apart from the stress produced by the steam pressure.
In the ordinary flanges the tension of the bolts produces cross strains
and the fittings give way where they would naturally fail through
this strain. We have given considerable thought to the construction
of flanges in which such cross strains are eliminated, but have not
pushed the matter forward as we have decided to eliminate all doubt
as to the safety of the fittings by employing steel castings.
3 Professor Miller's tests bear out what we have observed regard-
ing the different results to be expected from cast iron, as they show
that although there is a general falhng off in strength in one case the
CAST-IRON FITTINGS FOR SUPERHEATED STEAM 1031
cast-iron specimens did not lose in strength by being subjected to a
high degree of superheat. In connection with such tests it would be
interesting to investigate the action of superheat when the metals are
under stress.
4 Mr. Mann's conclusion that gun iron is better than cast steel
is indeed interesting, but we Avould not think of changing our present
practice of using cast steel until gun iron is thoroughly tried out in the
practical field and demonstrated all right for the work. The proper
method of determining the quality of the gun metal which is used must
also be developed by the necessarily slow process of observing the
action of the fittings in service. It would indeed be a simple matter
if bids for the fittings could be based on an analysis of the metal, and
I hope Mr. Mann may be right in this belief.
Prof. H. F. Rugan. While investigating the phenomenon of the
increase in cubic dimensions of cast iron as a result of repeated heat-
ings it became evident that the test pieces deteriorated in strength.
I am of the opinion that the influences at work producing such growth
at high temperatures are the same that cause the failure of cast iron
fittings at lower temperatures, say at from 500 deg. fahr. to 600 deg.
fahr. The effect of the higher temperatures is merely to increase the
extent of the changes, producing a maximum growth per heat.
2 Further experiments to determine the length of time required to
produce maximum growth developed the fact that a change in the tem-
perature was necessary to produce continued growth. No apparent
difference in growth was observed between pieces heated at the same
temperature for periods of 3 hours and 17 hours respectively.
3 The test pieces were heated in cast iron muffles, carefully luted
with fire clay, to protect them from contact with the furnace gases, to
a temperature of from 850 deg. cent, to 950 deg. cent.
4 Experiments were made with nine iron carbon alloys (A to I)
containing no graphite, the carbon content changing by 0.5 per cent
from 4.03 per cent to 0.15 per cent. Other constituents were low and
constant. Four bars of each alloy were cast in both sand and chill
moulds. These proved to be all white irons, the samples with low
carbon content being full of blow holes. No growth was observed in
any save the sample A which contained 4.03 per cent carbon. This
sample shrank for the first 12 heats, afterwards expanding, ultimately
becoming 6.88 per cent larger than its original volume.
5 Four alloys (J to M) were also tested. Of these J, K and L
were grey irons while M was a white iron. It was observed that M
1032 DISCUSSION
followed along lines closely approximating the action of A, shrinking
slightly during the early heats but growing after 12 to 19 heats had
been taken; ultimately becoming 6.2 per cent larger than the original.
Pieces of the bars from which the A and M test pieces were made were
inserted in the muffle to be sampled for chemical analysis after suc-
cessive heats. These analyses showed that the appearance of free
carbon (or temper carbon) coificided with those heats which produced
growth in the test pieces. Free carbon was in this way proved to be
in some way an indispensable factor in the growth of cast iron when
under heat treatment.
6 The grey irons J, K and L, grew from the start, and their pro-
gress'^indicated a close relation between their respective growth and
their silicon content.
7 To check these indications a series of alloys, with all the con-
stituents constant save silicon, having the following analyses, were
used to test the part played by silicon:
AUoy
Total Carbon
Combined
Carbon
Graphite
SI.
Mang.
Sulph.
Phoa.
N
3.98
3.98
3.79
3.76
3.79
3.38
0.64
0.68
0.30
none
none
none
1
3.34
3.30
3.49
3.76
3.79
3.38
1.07
1.79
2.96
4.20
4.83
6.14
0.25
0.23
0.25
0.27
0.30
0.30
0.01
0.01
0.01
0.01
0.01
0.01
0.013
0
0.013
P
0.012
Q
0.012
R
0.012
8
0.013
8 It will be observed that the total carbon in the series is approxi-
mately constant, that alloys N and 0 contain about the same amount
of combined carbon, that alloy P contains about half the quantity,
and that the remaining alloys contain none at all. The silicon in 0
and R is 0.2 per cent lower, inQ, 0.2 per cent higher than was desired.
The remaining constituents are satisfactorily low and constant.
9 Test pieces N to S, measuring 6 in. by about 0.88 in., were
machined from the castings. They were not taken from similar por-
tions throughout, but haphazard, some from the gate, others from a
riser either near to or at some distance from the gate. When the
growth of these alloys was investigated it became evident that the
locations from which a test piece had been cut had a considerable
influence on the rate of expansion. It was found that specimens
taken from the gate end of the casting grew more rapidly than those
taken from the top of the riser.
CAST-IRON FITTINGS FOR SUPERHEATED STEAM
1033
10 In test pieces from the same part of the bar, however, these
inequalities disappeared and a like growth was obtained in each alloy.
A slight falling off was observed in the closer-grained irons.
1 1 The results obtained are plotted in Fig. 3, the coordinates being
percentage of growth and number of heats. In this way the rate of
growth is clearly seen. In the case of samples N, O and P, curves are
plotted from the data obtained. It will be observed that the growth
20 ,30
Nuinlicr of He:its
Fig. 3 Curve showing Rate op Growth of Alloys N to S
is rapid at first, diminishes after about the seventh heat and stops at
the sixteenth heat.
12 In the case of Q and R curves are plotted in full lines from the
data obtained up to the point at which cracks appeared, viz. the
twelfth heat. Beyond this the direction of the curves can only be
guessed and this is indicated by dotted lines. The following table
summarizes the results in relation to the silicon content:
1034
DISCUSSION
Alloy
Percentage of Silicon
Percentage Growth on Heating
N
1.07
1.79
2.96
4.20
4.83
6.14
15.40
0
23.46
P
32.85
Q
43.90
R .
59.50
s
63.00
13 It is quite clear from these tests that silicon is a most important
constituent of cast iron from the standpoint of growth under repeated
heatings. If the ultimate growths and percentages of silicon are
plotted as coordinates, the curve in Fig. 4 is obtained, which shows
that, broadly speaking, the growth is proportional to the percentage
of silicon.
14 To settle the question as to the influence exerted by graphite,
and at the same time determine if iron-silicon alloys, containing little
carbon and no graphite, would grow, three alloys (T, U and V) were
experimented with, having the following analyses:
Alloy
Silicon
Carbon
Mang.
Sulp.
Phos.
T
0.65
1.10
2.71
0.17
0.18
0.19
0.17
0.19
0.20
0.045
0.049
0.051
0.017
u
0.022
V -
0.033
15 Microscopic examination showed in all three cases a solid solu-
tion of iron silicide in iron. There were no traces of graphite or any
other structural constituent.
16 Machined bars were heated fifteen times under the same con-
ditions as the previous alloys. A summary of the final values is con-
tained in the table below:
Alloy
Percentage Silicon
Percentage Change of
Volume after Fifteen
Heats
Percentage Change of
Weights after Fifteen
Heats
T
0.65
1.10
2.71
-0.025
0.000
+0.394
-0.04
U
-0.03
V
-0.02
17 It will be seen that the only alloy of the three which showed an}'
tendency to grow was V, with 2.71 per cent of silicon. The expan-
CAST-IRON FITTINGS FOR SUPERHEATED STEAM
1035
sion, however, was very slight, and compared with that of P (2.96 per
cent silicon and 3.79 per cent carbon) after the same number of heats
was almost negUgible, amounting to but 0.394 as compared with 31.35
per cent, the mean figure of P and PP.
18 Alloys K, N and P correspond closely to alloys T, U and V in
silicon content. They also contain about 3.9 per cent of carbon,
CO
«50
O
§30
a
IV)
2 20
10
^
-•
-o
1 p
''
J J
K
l'^
/
i'
•
•2 3 4 5
Percentage Silicon
Fig. 4 Curves illustrating Relation between Percentage Growth
AND Percentage Silicon
mostly in the form of graphite, as compared with a mean figure of 0.18
per cent carbon, none of which is present as graphite in the other
series. A comparison can thus be made between the changes of
volume of the two series under similar tests after fifteen heats, by
means of the following:
1
Per Cent
1!
1
1
Per Cent
Alloy
Carbon
Silicon
Change in
i Volume
i
Alloy
Carbon
Silicon
Change In
Volume
T
0.17
0.65
-0.025
K.
3.90
0.69
+ 5 40
U
0.18
1.10
0.000
|N.
3.98
1.07
+15.20
V
0.19
2.71
1 +0.394
IP.
i|
1 3.97
: 2.96
+31.35
1 9 This comparison serves to emphasize anew that free carbon, even
in the form of graphite, is one of the essential factors in the growth
of cast irons under heat treatment. The previous series of alloys,
however, N to S, brought out clearly the fact that in the constant
1036
DISCUSSION
graphite series the growth is roughly proportional to silicon present,
graphite becoming merely the agent or forming the avenues by means
of which the silicon present can be acted upon. It is clear, therefore,
that both graphite and silicon are involved in these changes of volume
after repeated heatings.
20 A sample of test piece S, known as "S |," was heated to con-
stant volume in vacuo. This resulted in a shrinkage of 0.04 per cent.
The same sample was afterwards heated in the muffle to a constant
volume, when a growth of 67.70 per cent was obtained. Fig. 5
shows curves plotted from these data. It will be seen that S | grew
rapidly during the later heats, with no cracks developed, and the sam-
ple retaining its original form throughout. S, howevei-, grew rapidly
during the earlier heats, cracks developing during the first heat, and
finally breaking in two.
o
<ii s
I 0
SB
f-r^
J
/i
"'s
/
'
/
/
1
/
/
i
/•
/
.,/
^
^'■
<'
7U
0)
a 00
o
>
I 40
o
^ 30
0)
W)
a
s 20
11
p
S 10
x'l
3,'i
.-;<'
jf--
1
A
1
/.
1
k-
1
7
/
/
/
^
^
0 10 2U 30 40 50 60
Number of Heats
10 20 30 40 50 CO
Number of Heats
Fig. 5 Curves Showing Percentage Increase in Weight and Volume and
Number of Heats : Test pieces S and S\
21 Mr. A. Wolfe, superintendent of motive power of the United
Railway and Electric Comijany, Baltimore, commenting upon the
growth and final failure of some fittings in one of their power houses
says, "The temperature was not constant, varying between that of
the temperature of saturated steam at 175 lb. per sq. in. to super-
heated steam running between 500 deg. fahr. to 550 deg. fahr. total
temperature."
22 From the experiments I have made there is considerable evi-
dence indicating that gray cast iron subjected to changing tempera-
tures from 450 deg. fahr. and up gives evidence of an oxidation of the
silicon present, forming silica in a micro-crystal form, which upon
CAST-IRON FITTINGS POH SUPERHEATED STEAM 1037
cooling causes a disintegration of the surface exposed, ranging gener-
ally along the planes formed by the graphite, changing an apparently
solid wall into one showing many cracks. It is the constant recur-
rence of these conditions, produced by the changing temperatures,
that, in time not only produces growth, but breaks down the structure
of the metal.
23 The experiments conducted by Prof. E. F. Miller were so
treated, cooling down each night to the temperature of saturated
steam. No mention is made of any growth of these specimens. One
would expect to find a marked relation between growth and loss of
strength.
24 In a comparison between gray iron samples, those subjected to
a heating and cooling treatment totaling 1000 hours would become
weaker and larger than those kept constantly at the temperature of
superheated steam for a like period. I believe that actual contact
with steam_is not a necessary condition in this experiment, neither is
it a comparative test of the metal, which in service^ has one surface
exposed to atmospheric conditions.
AUTHOR'S CLOSUREi
Prof. Ira N. Hollis. The discussion of these papers brings out
certain interesting and valuable conclusions which cannot fail to
assist in the proper use of cast iron for parts of machinery and boilers.
Previously existing differences of experience with this metal under a
high temperature are shown to be due to fundamental differences of
chemical composition or to variations in the temperature. From
this point of view, existing data, even though ^conflicting, can prob-
ably be reconciled. The following conclusions as the result of the
papers and discussions may be studied with profit in connection with
new construction.
a Cast iron varies in its behavior under high temperature,
starting from about 450 deg. fahr. In many cases it
deteriorates in structure and strength to a marked degree.
b The effect of high temperature is independent of the medium
producing it, whether superheated strain, hot gases or
solids.
c. The change of structure or deterioration is much increased
by a fluctuating temperature.
'Professor Miller ;iiid Mr. Mann did not desire to present closures.
1038 DISCUSSION
i
2 Where the temperature is constant, even though as high as 600 ■
or 700 deg. fahr., the change in cast iron is not serious enough to pro-
hibit its use, bnt where the temperature varies considerably, the metal
is certain to develop cracks and distortion that render it misuitable
for steam pipes and other parts under steam pressure.
d Cast iron of certain chemical constituents increases mate-
rially in volume when subjected to fluctuating tempera-
tures above 500 deg. fahr.
e The chemical composition of the cast iron has a material
bearing upon the change of shape and volume and upon
the development of imperfections.
3 Certain facts in this connection are well shown by Professor
Rugan's experiments. As he states, cast iron containing only com-
bined carbon does not change even under high fluctuating tempera-
tures unless the carbon begins to separate into a graphitic form. Free
carbon is one of the factors assisting in the deterioration under high
temperature, especially when associated with sflicon. The latter '
seems to be the chief cause of increase in volume. Where the free
carbon is constant, the gro\vth in dimensions is roughly proportional
to the percentage of silicon.
/ The use of cast-iron fittings for superheated strain is inad-
visable where the temperature is likely to fluctuate, but
it can be safely used where the temperature is to be con-
stant.
g Cast-iron fittings should not be placed in aiij^ parts of a
steam-pipe line where there are serious bending stresses
in addition to the stresses produced by internal pressure,
unless the combined stresses are fully allowed for, or
neutralized by expansion joints.
No. 1270
NECROLOGY
WALTER MORRISON ALLEN
Walter Morrison Allen, works manager of the Warner & Swasey
Company, died Fe})ruary 8, 1909, at his home in Cleveland, O. He
was born in Bristolvillc, O., December 14, 1866, and received his
early education in the local schools of Cherryfield, Me.
He evinced an interest in mechanics early in life and when only
sixteen years of age freqjLiently went to the nearest railway station, a
distance of thirty miles, to study and make drawings of the locomo-
tives that passed that point. In 1885, he began work as an apprentice
to the machinist's trade in the works of Warner & Swasey. He was
given special opportunities in the drafting-room, and at the comple-
tion of his term of apprenticeship was kept in this department, of
which he was made head in 1891. During the next two 3'ears, the
details of the design and construction of the 26-in. telescope of the
Naval Observatorj^ and the 40-in. telescope of the Yerkes Observa-
tory came largely under his direction.
In 1893 he had charge of the firm's exhibit at the Chicago exposition
and during the folloAnng six years was superintendent of their works.
In 1904 he was made works manager, the position which he held at
the time of his death. He had traveled much in the interests of the
company, visiting England and the Continent in 1897-1898 and again
in 1900.
Mr. Allen was a member of the Cleveland Engineering Society,
the Cleveland Chamber of Commerce, the Colonial Club and the
Automoljile Club of Cleveland.
A. KENNEDY ASHWORTH
A. Kennedy Ashworth, manager of the filter department and
traveling engineer of the Pittsburg Gage and Supply Co., died Jan-
uary 20, 1909, at his home in Grafton, Pa. He was born May 26, 1873,
in Covington, Ky., receiving his education in the schools of Pitts-
burg, Pa., and in Cook Academy; he was also a graduate in
1040 NECROLOGY
mechanical engineering of Rose Polytechnic Institute, Terre Haute,
and studied at the Western University of Pennsylvania.
He first entered the employ of the Joseph Home Company as mechan-
ical engineer, and became chief engineer of their steam and electric
plants. After about two years he became a member of the firm of
D. Ashworth & Son, consulting engineers. He associated himself with
the Pittsburg Gage and Supply Company, and was actively engaged in
engineering work after that time, except for a short period when he
was connected with the Buckeye Engine Company, establishing a Bos-
ton office. He also established offices for the Pittsburg Gage and Supply
Company in Philadelphia, New York, Chicago and other large cities.
He was a member of the Engineers' Society of Western Pennsyl-
vania, the order of Free and Accepted Masons, and the Sons of the
American Revolution.
ARCHIBALD W. BLAIR
Archibald W. Blair, Juniormember of the Society, died July 17, 1909
at the home of his father. Dr. W\ W. Blair, in Princeton, Ind., where he
was born June 4, 1865. He was a graduate of the Princeton High School
and took a special course in mechanical engineering at Purdue Uni-
versity. After serving an apprenticeship in the foundry, machine,
boiler and smith shops of the Princeton Foundry and Machine Works,
Mr. Blair went in 1890 to Cincinnati where he attained high standing
as a mechanical draftsman. At the time of his death he was cliief
draftsman of the American Laundry Machinery Company and also
instructor in the Ohio Machinists' Institute, and a member of the
American Association of Engineers. He was a man of many excellent
traits and high attainments and will be greatly missed by his asso-
ciates. Mr. Blair entered the Society in 1898.
FRANCIS H. BOYER
Francis H. Boyer died at Ms home, Somerville, Mass., February
21, 1909. He was born at Manheim, N. Y., in 1845, and at the age
of ten years went to Greensburg, Ind., where he remained until he was
18 years old. He learned the trades of millwright, carpenter and
architect. On his return to the East he entered the steamboat trans-
portation business in Brooklyn, N. Y., in which he was engaged
during the Civil War.
At the age of 23, he went to the frontier, settling at Seneca, Nemeha
NFCROLOOY 1011
Co., Kan., where hecarriedonastockandlanclbusinessforafewyears.
He then returned to Brooklyn and engaged in the refrigerator build-
ing business, and was associated with the building of the first ship
refrigerator for carr}dng beef to Europe. He eventually became super-
intendent of theDe La Vergne & Mixer Coiiij):iny, refrigerator builders,
directing the construction of machinery for brewery refrigerators.
Mr. Boyer built the first brewery refrigerator in Boston in 1884.
His work called liim to Washington, Baltimore, Newark and New
York. He settled in Boston in 1890. and was appointed master
mechanic of the John P. Squire Company. He designed the big chim-
ney at that company's plant in East Cambridge and remained with the
company until its assignment in 1900, when he went into business for
himself, with liis son as a member of the firm.
Mr. Boyer was a Manager of this Society, 1899-1902, and was
chairman of the Local Committee at the time of the meeting of the
Society in Boston in 1902. He also did important work on the Com-
mittee of the Society for Determining Standard Methods for Conduct-
ing and Reporting Steam Engine Trials. He was a member of the
Boston Society of Civil Engineers; the American Society of Refrigerat-
ing Engineers; honorary member of the National Association of
Stationary Engineers and Somerville Council, Royal Arcanum; and
president of the Somerville Board of Trade for two years.
THOMAS HALLETT BRIGGS
Thomas Hallett Briggs was born on August 21, 1870, in New
York. He received his education in the public schools of Brooklyn,
N. Y., and finished with a technical course at Cooper Union. During
his studies at Cooper Union, he was employed in the drafting room
of the Logan Iron Works of Brooklyn, with which company he
remained for fourteen years. His work there covered not only
drafting-room work, but shop inspection, and he finally became
outside representative. He became associated with the M. H. Tread-
well Company, New York, as salesman, in 1904, which position he
held until the time of his death on September 24, 1909. Mr. Briggs
was a member of the Society of Gas Engineers, and entered The
American Society of Mechanical Engineers as an Associate in 1900.
ANDRP:W JAMES CALDWELL
Andrew James Caldwell, Manager of the Society, died in New
York City, May 10, 1909. Mr. Caldwell was born in Brooklyn, N. Y.,
1042 NECROLOGY
May 1, 1858, and was graduated from the University of Maine in
1878, with the degree of B.M.E. He also took a graduate course at
Cornell University.
He fii'st entered the employ of the Delamater Iron Works, and later
that of one of the regular line steamship companies on the New York
and New Orleans route. In 1880 he became connected as drafts-
man with the firm of Henry R. Worthington, and was closely identi-
fied with the designing and development of the Worthington high-
duty pumping engine. He was soon made chief of the erecting and
testing departments, which dealt with water-works pumping machin-
ery; then assistant to the president, Charles C. Worthington; and
finally general manager of the hydrauHc works.
Soon after the formation of the International Steam Pump Com-
pany, he resigned to take a similar position with the Crane Company
of Chicago, by which he was later sent to Bridgeport, Conn., to take
charge of the new plant of the Eaton, Cole &3urnham Company, con-
trolled by them. In 1902 he resigned, accepting a position with the
National Foundry Association, and later entered the service of the
Standard Oil Company, with which he was identified at the time of
his death. i ; • .
Mr. Caldwell was elected Manager of the Society in 1906, and was
still serving at the time of his death. He was a member of the
Engineers' Club in New York, and one of the organizers and the first
president of the Brooklyn Engineers' Club.
KENTON CHICKERING
Kenton Chickering, vice-presicent of the Oil Well Supply Com-
pany, died December 9, 1908, at his residence in Oil City. He was
born in Worcester, Mass., May 16, 1847, and received his education
in the Massachusetts public schools.
In 1863 he became a dispatch bearer for General Clark of the United
States commissary depai'tment in New York City, and remained in the
government service for a time after the war. In 1870 he represented
Eaton and Cole, dealers in brass and iron goods, at Titusville, remain-
ing with the company w'hen it became Eaton, Cole and Burnham
Company, with offices at Oil City. In 1878 Mr. Chickering was made
secretary of the Oil Well Supply Company, Ltd., which was formed at
this time. This new company absorbed the Eaton, Coje and Burnham
Company and others. In 1891, when the Oil Well Supply Company
was organized in its present corporate form, Mr. Chickering was
NECROLOGY 1043
elected vice-president, the position which he held at the time of his
death.
He patented a number of useful inventions in connection with oil
well machinery, and planned the large manufacturing plant erected
bj^ the company in 1901-1902, known as the Imperial Works. He
also designed a number of special machines to increase the output and
improve the quality of product of the plant.
Mr. Chickering was very active in church, civic and fraternal organi-
zations.
GEORGE W. CORBIN
George W. Corbin was born in New Britain, Conn., March 3, 1859.
He attended the local schools and Wilbraham Academy until 18 years
of age. His first business connection was with P. & F. Corbin, who
later organized the Corbin Cabinet Lock Company, making Mr. Cor-
bin manager, and later secreatry and president. He resigned to be-
come president of the Union Manufacturing Company, and held this
position until his death, Novemember 30, 1908. He organized several
other manufacturing corporations, and took active part in muni-
cipal affairs — the savings banks, local govermnent and schools.
He was connected with several social orders, among them the
Masonic order, and numerous social clubs.
ELMER Q. EBERHARDT
Elmer G. Eberhardt, of Newark, N. J., an Associate Member of the
Societ}--, died at his home on November 21, 1908. He was born in
Newark, April 26, 1881, and was graduated from the Newark High
School in 1896. He received his technical education in Stevens Insti-
tute and Cornell University, receiving the degree of M.E, at the latter
institution in 1904.
He learned the machine trade with his father, Henry E. Eberhardt,
of the firm of Gould and Eberhardt, at Newark, and upon graduation
from Cornell University, he formed, with his father and brothers, the
firm of Eberhardt Brothers Machine Company, now the Newark
Gear Cutting Machine Company. Mr. Eberhardt was vice-president
of the company, and was engaged in the design of automatic gear-
cutting machines, in which field he invented a number of improve-
ments as well as made investigations along original lines. He was a
frequent contributor to the technical columns of the mechanical
papers. -He designed the power plant and equipment of the factory
with which he was connected.
1044 NECROLOGY
Mr. Eberhardt was elected president of the Cornell Society of Elec-
trical Engineers and vice-president of the Cornell Mechanical Societj^
He was an Associate Member of the American Institute of Electrical
Engineers, and a member of the University Club of Newark and the
Cornell Association of Northern New Jersey.
At the time immediately preceding his death, Mr. Eberhardt was
engaged, aside from his business conne ;tions, in consulting engineer-
ing work, in matters relative to gears nnd gear cutting.
DAVID HAMILTON GII.DERSLEEVE
David Hamilton Gildersleeve was l)orn in Tenafiy, N. J., August
5, 1867, and died in New York Cit}-, July 30, 1909. He was educated
at Stevens Institute of Technology, Hoboken, N. J., and graduated
in 1889 with the degree of M.E. Foj nearly ten years thereafter he
was active in gas engineering and the construction and selling of
pumps and hydraulic machinery, being associated with the United
Gas Improvement Company of Philadelphia, the John H. McGowan
Company of Cincinnati, and the Snow Steam Pump Works, New York.
For three years, during and followijig the Spanish- American War,
he served as first lieutenant. United States Corps of Engineers in Cuba,
and as assistant engineer of the Dep;i.rtment of Havana had charge
of the mechanical work executed and planned there. In 1904
he became associated with the C. W. Hunt Company, of West New
Brighton, N. Y., as sales manage^', which position he resigned in
February 1909 to become one of the partners in the shipbuilding firm
of the Waters, Gildersleeve, Colvei" Company, of West New Brighton,
Staten Island.
Mr. Gildersleeve became a member of the Society in 1908. He was
a member of the Macliinery Club, the American Gas Institute,
the New York Railroad Club, Dry Dock Association, Staten Island
Club, Royal Arcanum, Staten Island Association of Arts and Sciences,
Spanish War Veterans Association, the Chi Phi Fraternity, and was
Secretary of the Stevens Alumni Association.
THOMAS GRAY
Dr. Thomas Gray, Vice-President and Profe sor of Dynamic and
Electrical Engineering of Rose Pol3rtechnic Institute,Terre Haute,
Ind., died December 19, 1908.
He was born h\ Fifeshire, Scotland, February 2, 1850. He took
.^ cpurse in engineering at the University of Glasgow, Scotland, where
NErnoLOGY 1045
he graduated in 1878 with the degree of B.S. Later he took a four-
year cour:se in practical physics and telegraph engineering under Sir
William Thomson (Lord Kelvin).
He was engaged by the Japanese government as instructor in tele-
graph engineering at the Imperial College of Engineering at Tokio,
Japan. After this engagement he was employed by Sir William
Thomson and Professor Fleming Jenkin, engineers of the Commerical
t'able Company, to superintend the manufacture and the laying of
that Company's system of transatlantic and other cables, and had
sole charge under them, as resident engineer, of the whole of that
work. He was later chief assistant to Lord Kelvin in his engineering
work.
In 1888, he was appointed to the professorship at Rose Polytechnic
Institute, and held the position until his death.
Doctor Gray was the author of Directions for Seismclogical Obser-
vations, in the British Admiralty Manual of ' Scientific Inquiry; of
articles on telephones and telegraphs in the Encyclopaedia Britannica,
and of the Smithsonian Phj^sical Tables. He also wrote many
papers on scientific and technical subjects, and was engaged as an
expert iii electricity on the staff of the Century Dictionary.
LEWIS CLESSON QROVER
Lewis Clesson Grover died at Hartford, Conn., September 30,
1909, after a long illness. He was born November 26, 1849, at Spring-
field, Mass. After an ordinary school education there, he acted as
apprentice at the Norwalk Iron Works, Norwalk, Conn., for three
years; and remained with them on the expiration of his term for
seven additional years. He was afterwards successively connected
with the Winchester Repeating Arms Company, New Haven, Conn.,
C. W. Lacount of Norwalk, Conn., Smith & Wesson, Springfield, Mass.,
and F. C. & A. E. Rowland, New Haven, Conn.
He became general manager of the Whitney Arms Company, New
Haven, Conn., about 1880, holding this position until 1886, when he
went to Hartford as assistant superintendent of the Colt's Patent Fire
Arms Mfg. Company. He was soon promoted to the office of superin-
tendent, and later to that of general manager. In 1902 he was elected
president and a director of the company, at the same time becoming
president of the Colt's Arms Company, of New York. Because
of ill health he was finally compelled to relinquish to others the active
duties of management, and in January resigned the office of president,
1046 NECROLOGY
the same meeting making him chairman of the boards of directors of
both corporations.
Mr. Grover served as a member of the common council board of
Hartford and as park commissioner. He was a prominent Mason and
a member of the Hatchetts Reef Club. He entered this Society in
1890.
CHARLES LEWIS HILDRETH
Charles Lewis Hildreth was born October 9, 1823, at Concord, N.
H., and died suddenly on February 26, 1909, at his country home in
Westford, Mass. He received his education at a private school in
Nashua, N. H., and at the Appleton Academy in New Ipswich, N. H.
In 1845 he went to Lowell, Mass., and became an apprentice in the
Lowell Macliine Shop, which was incorporated the same year. After
a service of three years, he became a contractor on piecework. He
was identified for over fifty years with this company, being superin-
tendent for twenty-six years. During the great depression of the
iron trade in 1858 he went to Philadelphia and served as foreman in
the Industrial Works. He returned to Lowell, Mass., in 1860, and
assumed charge of the shop as general foreman, holding this position
until 1879 when he was made superintendent of the entire plant.
Mr. Hildreth was identified with the textile machinery trade during
almost the entire period of its development in America. In July
1905 he retired from business.
He was president of the Mechanics Savings Bank in Lowell, Mass.,
was closely identified with the Lowell Textile School, and while in
Lowell was an attendant at the Kirk Street Congregational Church;
in Westford he attended the Union Congregational Church.
WARREN E. HILL
Warern E. Hill was born in New York in 1835. In 1852 he entered
the service of the Allaire Iron Works in Newark, N. J., and was asso-
ciated with that company for six years. In 1858 he was appointed
superintendent in charge of the installation of the Detroit, Mich.,
Water works, which position he held until 1862, when he returned to
the East and accepted a position with the Continental Iron Works of
Brooklyn. In 1888 he was made vice-president, and in 1907 president
of this firm, the position he held at the time of his death. Mr. Hill
was the designer of the machinery and engines of the original "Mon-
itor," which defeated the "Merrimac" in Hampton Roads.
NECROLOGY 1047
His death occurred in New York, December 8, 1908. He became
a member of this Society in 1884.
ROBERT HOE
Robert Hoe, head of the firm of Robert Hoe & Company, of New-
York and London, was born in New York, March 10, 1839, and was
educated in pubHc and private schools in this city. He was grandson
of Robert Hoe of the hamlet of Hoes, Leicestershire, England, who
began the manufacture of printing machines in New York in 1803,
constructing and introducing into America the first iron and steel
machines.
Mr. Hoe at an early age entered the printing factory established by
his grandfather, and devoted his life to the improvement and develop-
ment of printing machinery. He developed the rotating-cylinder
type of press to the present double-octuple press capable of printing,
pasting, folding and delivering more than 150,000 16-page newspapers
per hour. He also invented greatly improved processes of printing in
colors, and is the author of several books on printing and binding.
Mr. Hoe always resided in New York, although his business interests
were almost as great in London, and identified himself with its interests
and prosperity. He was one of the founders of the Metropolitan
Museum of Arts, founder and first president of the Grolier Club, and a
member of the Engineers, Union League, Century, Players and
Fencers Clubs. He joined The American Society of Mechanical
Engineers in 1883.
Mr. Hoe died in London, September 22, 1909.
WILLIAM S. HUTETTE
William S. Huyette was born in Blair, Neb., November 13, 1870,
and was educated in the public schools of Detroit, Mich.
He began his shop experience in the drafting department of the
Detroit Blower Company, under the management of his father. He
was later engaged by the engineering firm of Gilbert Wilkes Company,
Detroit, leaving their employ in 1897 to open an office for the Wickes
Boiler Company, in Milwaukee, Wis.
The following year, Mr. Huyette returned to Detroit, to take charge
again of the business of the Gilbert Wilkes Company in the absence of
Mr. Wilkes, who was commi^nder of the Detroit Naval Militia during
the Spanish-American War. Upon the return of Mr. Wilkes after
1048 NECROLOGY
the war, Mr. Huyette went back to his work with the Wickes Boiler
Company, and opened their branch office in Chicago. He continued
as manager of that office until his death, January 11, 1909.
His engineering work was chiefly on boiler installations, and he also
designed and patented a gas engine, and designed and built steel sail
boats and motor boats.
He was a member of the National Association of Stationary Engi-
neers, and of the Chicago Yacht Club.
EDWARD L. JENNINGS
Edward Lobdell Jennings, whose death occurred on November 6,
1908, was born in North Wayne, Me., April 14, 1850. He received
his education in the public schools of his nati\ e town and at the Maine
Wesleyan Academy. He was apprenticed to the North Wayne Tool
Company, and in 1872 he went to Boston and entered the employ
of W. A. Wood & Co., and was their manager for several years. He
resigned his position with this company and removed to Water-
bur}'- to take the position of purchasing agent for the American Brass
Co., held for the eight years preceding his death.
Mr. Jennings was a member of the Waterbury Club, a Commandery
Mason and a member of the First Church (Congregational), Water-
bury, Conn.
EDWARD H. JONES
Edwin Horn Jones was born in Wilkes-Barre, Pa., April 15, 1844,
and died December 2, 1908. He was educated at the Old Dow Acad-
emy on South Franklin Street, and at an early age he entered the
employ of his father, Richard Jones, who then conducted the Jones
Foundry, the foundation of the present extensive Vulcan Iron Works.
He learned the iron business thoroughly, advanced to superintendent
of the works, and at the time of his father's death in 1873 became
general manager of the company and later its president, which posi-
tion he held until the time of his death.
As president of the Vulcan Iron Works he consolidated the Wyo-
ming Valley Manufacturing Company and the Pittston Iron Works
vith the original plant and later purchased the Tamaqua shops, all
of which he consolidated as the Vulcan Iron Works.
lu 1891 h(^ was made president and general manager of the Sheldon
Axle Works. In 1881 he became director of the Second National
NECROLOGY 1049
Bank, and later its vice-president. lie was interested as stoekhoider
and directoi- in a number of other industries and was an active mem-
ber of the Wilkes-Barre Board of Trade and one of its trustees. He
was a member of the Westmoreland Club and the Wyoming Valley,
Country Club, the Art Club, of Philadelphia, and the Sons of the
Revolution.
JANG LANDSING
Jang Landsing, whose death occurred July 10, 1909, was born
in Heong San, China, October 11, 1864. He was graduated from
the Worcester Polytechnic Institute in June 1887, with the degree
of B.S.
He was apprenticed to the Washburn Machine Shops for three
years, after which he became connected with the Pratt & Whitney
Company, as draftsman, leaving after two years to accept a position
as machine designer with the Brush Electric Company, of Cleveland.
About 1899 Mr. Landsing, working in cooperation with Mr. Joseph
Bijur, assisted in developing a line of special machinery for the manu-
facture of the " Bijur " type of storage battery. Mr. Landsing became
mechanical engineer and then superintendent of the General Storage
Battery Company, and was in active charge of the company's large plant
at Boonton, N. J., at the time of his death. This work afforded him
unusual opportunity to exercise his mechanical ingenuity, and many
of the company's successful machines are due to the combination of
his fine mechanical judgment and designing skill. He organized
and equipped a machine shop for the manufacture of the special
tools used by the company and also the 300-kw. power plant. Mr.
Landsing had the ability to carry out in an efficient and economical
manner work with which he was not previously familiar.
Mr. Landsing was also consulting engineer of the Chinese Legation
at Washington.
• ROBERT B. LINCOLN
Robert B. Lincoln, president of the Waters Governor Company,
Boston, Mass., died June 9, 1909, at his home in Waltham, Mass.
Mr. Lincoln began his career in the Globe Works in Boston, after-
wards serving throughout the Civil War. In 1868 he went to Cuba
as chief engineer of the Maratanza, severing this relationship to
become head draftsman at the South Boston Iron Works. In 1882
he designed the compound enghie on the Cymbria at East Boston,
Mass., and was subsequently connected with E. D. Leavitt of Cam-
1050 NECROLOGY
bridge, Mass., and later with the Portsmouth Navy Yard, where he
remained nine years. At the time of his death, he had been president
of the Waters Governor Company, for twenty-seven years, and dur-
ing his life had held many other positions of trust which were filled
with honor and fidelity.
ALEXANDER MILLER
Alexander Miller, head of the firm of Alexander Miller and Brothers
of Jersey City, N. J., died May 6, 1909, at his home in New York.
Mr. Miller was born in Aberdeen in 1857 and came to this country
at an early age. He began his engineering career at the old Dela-
mater Iron Works of New York, of which his father was superin-
tendent. Later he formed a connection with the Deeley Iron Works
and specialized in sugar-evaporating machinery. More recently he
applied with conspicuous success the experience thus gained to the
problem of the evaporation of brine in the manufacture of salt.
Three of the New York fire boats were also the product of his works.
Mr. Miller was a leading factor in the organization of the Dela-
mater Veterans Association, composed of his old associates at the
Delamater works, and until the present year was its president. He
was also a member of Scotia Lodge, F. & A. M., the Engineers
Club, New York Athletic Club and the St. Andrew's Society.
ARTHUR WARREN KENDALL PIERCE
Arthur Warren Kendall Peirce died April 13, 1909, at Driehoek,
Transvaal, South Africa, where he had been located since 1897. He
was born in West Boylston, Mass., November 19, 1873. His pre-
liminarj^ education was acquired in the Plymouth, Mass., high school,
and his technical education by home study and experience in the
engineering departments of various electrical companies.
In July 1897, he was appointed electrician, to the Knights Deep,
Ltd., a gold mining company in the Transvaal, and a year later
became consulting electrical engineer to the Consolidated Gold Fields
of South Africa, Ltd., a corporation controlling a number of mines
in the Transvaal, continuing in this position until 1906. Afte*- a
trip to England he returned to South Africa in April 1907, to accept
an appointment with the Victoria Falls Power Co., Ltd., of Germis-
ton, Transvaal, which he retained to the time of his death.
Mr. Peirce was also a member of the American Institute of Elec-
trical Engineers, the South African Association of Engineers, the
NECROLOGY 10/)!
Mechanical Engineers' Association of the Witwatersrand, and an
associate of the Chemical, Metallurgical and Mining Society of South
Africa. He made many extensive trips through the various mining
regions of America in search of information regarding hoisting from
deep mines, in which work he was professionally interested.
JASPER R. RAND
Jasper Raymond Rand, vice-president and director of the Ingersoll-
Rand Company, New York, died in Salt Lake City, Utah, March 30,
1909. He was the son of Jasper Raymond Rand, one of the founders
of the Rand Drill Company, and was born in Montclair, N. J.,
September 3, 1874.
He was graduated from Cornell University in 1898 with the degree
of Mechanical Engineer, and served in Porto Rico in the Spanish-
American war as a member of the First New York Volunteer Engineers.
During 1899-1900 he was president of the Imperial Engine Company
at Painted Post, N. Y., which position he left to take the presidency
of the Rand Drill Company. In 1905 he was elected vice-president
and director of the IngersoU-Rand Company, the position he held up
to the time of his death.
Mr. Rand was a member of the Alpha Delta Phi Fraternity, the
Spanish War Veterans, the American Institute of Mining Engineers,
the Engineers Club, the Cornell Club and the Alpha Delta Phi Club
of New York.
WIIXIAM THOMAS REED
William Thomas Reed, was born June 27, 1847, in London,
England. He received his education at the Commercial College,
Kent, and entered railway service in 1862 as a machinist ;
was apprenticed to the London, Chatham & Dover Railway, 1869-
1871; became a machinist on the Grand Trunk Railway, Canada, in
1871, afterwards serving the road as leading machinist, at Stratford,
Ont., 1875-1877, foreman erecting and other shops, at Montreal,
P. Q., 1877-1883; locomotive foreman, at Belleville, Ont., 1883-1887.
In 1887 he became master mechanic of the Western division, Ca-
nadian Pacific Railway; and from 1888 to 1894, acted as general master
mechanic, St. Paul & Manitoba Railway. From 1895 to 1898 he was
superintendent of the Chicago Great Western Railway, and from
1893 to 1901, superintendent of motive power and machinery, Sea-
board Air Line Railway, Portsmouth, Va. He was locomotive super-
1052 NECROLOGY
^ntendent for the Jamaica Government Railway, from 1902 to 1906,
at which time he was appointed locomotive superintendent of the
Gold Coast Government Railway, West Africa. He held this position
until 1908, when he received another appointment on the Jamaica
Government Railway. In 1890 he became a member of this Society.
Mr. Reed died of malarial fever, July 1, 1909, only a few hours after
leaving Kingston wharf on leave of absence, and was buried at sea.
EDWIN REYNOLDS
Edwin Reynolds, Past-President of the Society, was born in .
Mansfield, Conn., March 23, 1831, and was apprenticed at the age of
sixteen to A. P. Kenney, a local machinist. At the end of three
years apprenticeship he started on a journeyman's tour of various
shops in lower New England.
About 1857 Mr. Reynolds went West and for a time was superintend-
ent of the shops of Steadman & Company, of Aurora, 111., builders of
engines, sawmill machinery, drainage pumps, etc., for the Southern
trade. As this business practically ceased during the Civil War, he
returned to the East and was employed in several machine works
until, in 1867, having attracted the attention of George H. Corliss, he
was offered a position in the works of the Corliss Steam Engine
Company, at Providence, R. I., where he remained as general super-
intendent until 1877. His last notable work with the Corliss com-
pany was the design of a rolling-mill engine to run at 160 r.p.m. a
speed double that of previous designs. This engine was installed at
the works of the Trenton Iron Company, in 1877. In 1890 a second
fly-wheel was added and the speed increased to 180 r.p.m.
In 1877 Mr. Reynolds accepted the position of general superin-
tendent of the Edward P. Allis Company, of Milwaukee, Wis., his
first work being to place the company on a paying basis, largely
through the development of the Reynolds-Corliss engine. He later
developed the building of a varied line of machinery, including large
Corliss-engine units for pumping service, mining, air compressing,
furnace blast, street-railway work and other purposes. One note-
worthy achievement was the building, in 1888, of the first triple-
expansion pumping engine for waterworks service. This engine was
described by Professor Thurston as "doing continuously so high a
duty as to place it among the most remarkable constructions of its
class and time, and probably to make its record the highest to date
(1894)."
NECROLOGY 1053
The blowing engine de^jigned by Mr. Reynolds for the original
Joliet Steel Company and chosen from among competitive designs
submitted by leading engineers in the United States and Europe.,
embodied a radical departure from accepted practice, but after more
than 25 years of continuous experimenting the essential features of
this design have not been improved upon.
The well-known story of his sketching the design of the horizontal-
vertical four-cylinder compound engine for the Manhattan Railway
Company's power house in New York City, while on the train from
Milwaukee to New York, illustrates his marvelously quick inventive
genius.
At the time of liis death, Mr. Reynolds was consulting engineer
of the Allis-Chalmers Company, although the illness of the past three
years had prevented active duty. He took a great interest to the
last in the magnificent new works of the company at West AUis,
which had been laid out on a strictly engineering basis, after his
plans, at the time the Edward P. Allis Company became merged in
the new organization.
He was also interested in a number of other plants of Milwaukee
and vicinity, had been the first president of the National Metal
Trades Association, and bore the degree of LL.D., conferred upon
him by the University of Wisconsin, on the wall of who.-e college of
engineering his name has been carved.
Mr. Reynolds served as Vice-President of the Society from 1892-
1894, and as President for the year 1902.
His death occurred at his home in Milwaukee February 19, 1909.
KICHARI) HERMAN SOULE
Richard Herjnan Soule was born March 4, 1849, in Boston, Mass.
September 25, 1875, he entered the service of the Pemisyivania Rail-
road, svhere he remained for eight years. He held this position for
two years until promoted to the test department. Two years later,
in 1879, he was made superintendent of motive power of the Northern
Central Railway.
In 1881 and June 1882, he was superintendent of motive power of
th(; Philadelphia and Erie division of the Pennsylvania Railroad, and
in June 1882 accepted a position in the same capacity with the Pitts-
burg, Cincinnati and St. Louis Railway.
In 1883, when the West Shore Railway enterprise was carried
through, its managers secured the best talent available in the country
1054 NECBOLOGY
for their managing officers, and Mr. Soule was appointed superintend-
ent of motive power, a position which he held until the absorption of
the West Shore Line by the New York Central in 1887. From Febru-
ary 1887, to April, 1888, he was general manager of the New York,
Lake Erie and Western Railroad, and in November, 1888, he was
appointed general agent of the Union Switch and Signal Company.
He was engaged in the introduction of modern interlocking and lock-
signaling plants until 1 891 . From 1891 to 1897 he was superintendent
of motive power of the Norfolk and Western Railroad, and did much
to put the rolling-stock of the system, which was then coming into
prominence as an important coal-carrying road, on a thoroughly
sound basis.
For the next two years, Mr. Soule was in the employ of the Baldwin
Locomotive Works, spending nearly a j-ear traveling in foreign coun-
tries. He had charge of the Chicago office of this company for a year
and a half.
In 1900 he opened an office in New York as a consulting mechanical
engineer and practiced until, on account of ill health, he was forced
to retire from active business.
Mr. Soule was a member of the Master Car-Builders Association;
and author of a report on the standards of this association, which led
to a radical change in the association's practice, and to a placing of
the standards on a much higher basis. He was also a member of the
American Railway Master Mechanics Association. He was one of the
managers of this Society, from 1898 to 1901.
He was universally respected and esteemed for his many sterUng
qualities, which caused his acquaintance to be highly prized by his
associates. In all parts of the country men are found who testify to
the help given them early in life by Mr. Soule, to whom they owe
much of their later success. His memory will live long in the hearts
of those to whom he had endeared himself.
Mr. Soul's death occurred at his residence in Brookline, Mass.,
December 13, 1908.
AKVY ELROY WELLBAUM
Arvy Elroy Wellbaum was born February 12, 1881, at Brookville,
Ohio, where he attended the high school. He studied at the Ohio
Northern University, Ada, O., for one year, and received the degrre
of M.E. in 1902 from Ohio State University. During the sunmier
vacations he was in the employ of the C. & G. Cooper Company,
Mt. Vernon, O., and Piatt Iron Works Co., Dayton, 0.
NECROLOGY 1055
In 1902 he became draftsman for the Morgan Engineering Co.,
Alliance, O. He became connected with the Foos Manufacturing
Co., Springfield, O., in 1903, as designing draftsman, and in 1905
he accepted a similar position with the Foos Gas Engine Co., Spring-
field, O. For three years he was instructor of mechanical drawing
and machine design in the Young Men's Christian Association of
Springfield, O. Up to the time of his death, August 31, 1908, he was
associated with The Hydraulic Press Company, Mt. Gilead, 0..
having had charge of the engineering department.
GEORGE W. WEST
George Washinglon West died at his home in Middletown, N. Y.,
December 24, 1908. He was born April 3, 1847, at Troy, N. Y., and
received his early education in the public schools of that city. In
1865 he entered the service of the New York Central & Hudson
River Railroad at Schenectady, as machinist, and was later made
foreman and master mechanic, leaving this position to accept a similar
one with the West Shore.
In 1886, he entered the employ of the New York, Lake Erie and
Western, now the Erie, as master mechanic of the Mahoning division,
was later transferred to the main shops at Meadville, Pa., and in 1888
to the Eastern division. From 1891 until the time of his death he
held the position of superintendent of motive power of the New York
Ontario and Western.
Mr. West was a past-president of the American Railway Master
Mechanics Association, a member and past-president of the New York
Railroad Club, and past-president of the Central Railway Club. He
was a member of the Masonic order and the order of Elks. The
George W. West Association of Engineers at Carbondale was named
for him. He was also a director of the First National Bank of Middle-
town, president of the Ontario and Western Savings and Loan Associa-
tion, a member of the Middletown Club and a member of the Board
of Water Commissioners.
ALFBED R. WOLFF
Alfred R. Wolff, who died at his home in New York on January
7, 1909, was born in Hoboken, March 15, 1859. He entered the Stevens
Institute of Technology with its class of 1876, when less than fourteen
years of age. He nevertheless easily carried the studies of the four
1056 NECROLOGY
years course, and was recognized as one of the leading students of
a strong class.
His graduating thesis on windmills, which contributed original
experimental data to the theory of the subject, was published through
several numbers of the Engineering and Mining Journal of 1876,
with favorable editorial comment. This thesis, supplemented by a
compendium of modern American windmills, with tabular statements
of their power and relative economy in practice, was published in
1885 by Wiley & Sons, in book form, under the title, The Wind Mill
as a Prime Mover, and remains the only book on this subject.
After graduation, Mr. Wolff entered the office of the late C. E.
Emory, then consulting steam engineer of New York and also con-
sulting engineer to the U. S. Revenue Marine Service.
About 1880, Mr. Wolff decided to build up a practice as consulting
engineer in New York. For about eight years his work consisted in
the miscellaneous commissions of the steam expert. During this
time he wrote several articles, among them a paper on The Value
of the Study of the Mechanical Theory of Heat, presented at the first
meeting of The American Society of Mechanical Engineers, of which
he was a charter member; a series of editorial articles on steam
and energy questions, which appeared in The American Engineer,
with whose staff Mr. Wolff was connected for several years, and a
supplement to Robert Briggs' essay on Steam Heating, published
in the Van Nostrand Science Series.
In 1888 Mr. Wolff was engaged to assist the architect of the New
York Freundschaft Club to complete its heating and ventilating
plant. This engagement proved to be his opportunity to secure a
lucrative specialty. At this time the architect depended for the
design of the heating and ventilating plant upon the largely gratui-
tous plans and specifications of the prospective contractor and con-
sequently the heating and ventilating requirements of a building were
liable to be sacrificed in undue proportion to their importance.
Obviously there was field for a middleman as the authorized agent
of the architect.
In establishing himself in this field, Mr. Wolff encountered many
difficulties which he overcame so successfully that at the end of six
years he had referred to him more problems of heating and ventilat-
ing than he could execute. It was during this time that he wrote
and published the pamphlet entitled The Ventilation of Buildings,
an outline of the elements of physics, chemistry, and mechanics
involved in the design of a heating and ventilating plant.
He introduced, from the German practice, in 1893, the " heat-unit
NECROLOGY 1057
system," under which the radiator surfaces for direct heating in a
building are systematically calculated from the heat lost by the
various thicknesses of walls and proportion of window surface. This
was a substitute for the crude American rule, previously in vogue,
allowing a square foot of radiator surface per various cubic feet of
room contents, depending entirely on the judgment of the engineer.
He made popular the use of the combined plenum and exhaust
system operating with "tempered" air, for ventilation, supplemented
by direct radiators to supply the loss of heat by walls and windows'
as a substitute for the "switch-damper" method of heating and
ventilating formerly prevaiUng in the metropolitan district.
He introduced the thermostat in high-class residence work in 1893;
and in 1902, he stimulated Johnson to apply this mechanism to the
automatic control of humidity. This automatic "humidostat" was
first successfully applied in the Carnegie residence.
Mr. Wolff introduced the use of the cheese-cloth filters, for strain-
ing the dust out of the air drawn into a building for indirect heating,
into the metropolitan district in 1894, for use in the New York Life
Building. He installed a heating and ventilating plant in the Board
Room of the New York Stock Exchange in which the problems of
constant temperature and constant hydrometric conditions were
successfully met. The refrigerating feature of this plant is the only
example of the artificial control of summer heat in an office building,
and as such it is a unique monument to Mr. Wolff's ability.
Mr. Wolff took part in the organizations of the Ethical Culture
Society, especially in the maintenance of their charitable schemes. He
was also an alumni trustee of the Stevens Institute of Technology
from 1893 to 1896 and a permanent trustee after 1900.
Among a great many important buildings in which Mr. Wolff
installed heating and ventilating plants are the following: Century
Club; Waldorf-Astoria Hotel; Carnegie Music Hall; Lakewood Hotel,
Lakewood, N. J.; United Charities Building; University of the City
of New York; New York Herald Building; C. Vanderbilt residence;
Teachers College; J. J. Astor residence; St. Regis Hotel; the Lying-
in Hospital; Princeton Library; Brooklyn Institute; Columbia Uni-
versity; Sherry's Hotel; Delmonico's Hotel; University Club; Hotel
Martinique; American Museum of Natural History; Cornell ]\Iedical
College; Carnegie residence; Library of J. P. Morgan, Esq.; Hispanic
Museum; Evening Post Building; Plaza Hotel.
1 Mr. Wolff illustrated the application of this combination together with the
"heat vinit system," in a lecture before the Franklin Institute in 1894, which was
published in pamphlet form under the title The Heating of Large Buildings.
INDEX TO VOLUME 31
NOTE
1 Names of authors and dlsoussors and headings of groups of similar subjects are In CAPS and
SMALL CAPS.
2 The straight title of each paper la In Italics. The cross Index of a title la in CAPS and lower
case.
3 The Society Is not responsible as a body for the statements of fact or opinion In Its papers and
discissions.
Allen, C. M. Tests on a Venty.ri Meter for Boiler Feed 589
Automatic Feeders for Handling Material in Bulk, C. Kemble Baldwin . . . 161
Under-Gate Feeder; Lifting-Gate Feeder; Screw-Conveyor Feeder;
Roll Feeder; Rotary-Paddle Feeder; Revolving-Plate Feeder; Apron-
Conveyor Feeder; Swinging-Plate Feeder; Plunger Feeder; Recipro-
cating-Plate Feeder; Shaking Feeder.
Discussion
T. A. Bennett, 169; Closure 170
Baldwin, C. Kemble. Automatic Feeders for Handling Material in Bulk 161
Barth, Carl G. The Transmission of Power by Leather Belting 29
Beams, Stresses in Reinforced Concrete, G. Lanza and L. S. Smith. . . . 511
Belt Conveyor, A Unique, E. C. Soper 151
Belting, The Transmission of Power by Leather, Carl G. Barth 29
Best Form of Longitudinal Joint for Boilers, The, F. W. Dean 823
Early Forms of Joints; Defects of the One-Sided Butt Joint; New Form
of Butt Joint.
Discussion
R. P. Bolton, 826; E. D. Meier, 826; A. M. Greene, Jr., 827;
W. A. Jones, 827; S. F. Jeter, 829; Closure 830
Bibbins, J. R. Cooling Towers for Steam and Gas Power Plants 725
Bituminous Gas Producers 877
Bituminous Gas Producers, J. R. Bibbins 877
Essential Requirements ; Description of Power Plant ; Schedule of Tests ;
Discussion of Results; Operating Results; General Conclusions.
Discussion
G.M.S.Tait,894; R. H. Fernald, 894; W. B. Chapman, 898;
H.M. Latham, 899; H.H.SuPLEE, 900; E. N. Trump, 900;
BE. F. Smith, 901 ; G. D. Conlee, 901 ; Closure 901
Boiler
Tests on a Venturi Meier for Boiler Feed, C. M. Allen 589
Tan Bark as a Boiler Fuel, D. M. Myers 685
An Experience with Leaky Vertical Fire-Tube Boilers, F. W. Dean 799
The Best Form of Longitudinal Joint for Boilers, F. W. Dean 823
1060 INDEX
Bucyrus Locomotive Pile Driver, The, Walter Ferris 905
Discussion
A. F. Robinson, 919; L. J. Hotchkiss, 919; Closure 921
Caine, W. p. Governing Rolling Mill Engines 783
Carpenter, R. C. High-Pressure Fire-Service Pumps of Manhattan Bor-
ough, New York ' 437
Cast-iron Fittings for Superheated Steam, I. N. Hollis 989
Manner of Failure; Possible Reasons for Failure; Difficulties at South
Boston; Chemical Analj^ses and Tensile Tests; Fittings that Failed;
Service Stresses.
Discussion, see Superheated Steam.
Cast-Iron Test Bars, A Report on, A. F. Nagle 975
Cast-Iron Valves and Fittings for Superheated Steam, Arthur S. Mann. . 1003
Growth of Fittings ; Instances of Failure ; Experience with Steel Fittings ;
Analysis of Specimens; Practice Abroad.
Discussion, see Superheated Steam
Celebration, Hodson-Fulton 873
Concrete Beams, Stresses in Reinforced, G. Lanza and L. S. Smith 511
Conveyors
A Unique Belt Conveyor, E. C. Soper 151
Automatic Feeders for Handling Material in Bulk, C. K. Baldwin. . 161
Cooling Towers for Steam and Gas Poiver Plants, J. R. Bibbins 725
Present Field; Representative Installations; Special Phases of Cool-
ing Towei Operation; Elements of Design; Present Types; Lath Mat
Construction; The Evaporative Cooler: Standards of Design; Booster
Type of Tower; Conclusions.
Discussion
Geo. J. FoRAN, 758; W. D. Ennis, 763; H. E. Longwell, 764;
B. H. Coffey, 767; C. G. de Laval, 770; E. D. Dreyfus,
776; T. C. McBride, 778; Closure 779
Cylinders in Single-Acting Engines, Offsetting, T. M. Phetteplace 223
Darling, Philip G. Safety Valve Capacity 109
Dean, F. W.
An Experience with Leaky Vertical Fire-Tube Boilers 799
The Best Form of Longitudinal Joint for Boilers 823
Design of Curved Machine Members under Eccentric Load, Walter Rau-
TENSTRAUCH 559
Usual Analysis for Stresses; Andrews' and Pearson's Investigation;
Values of Constant for Hooks; Goodman's Formula; Analysis of Punch
Frame.
Discussion
G. Lanza, 566; C. R. Gabriel, 568; G. R. Henderson, 569;
Wm. H. Burr, 570; A. L. Campbell, 571; F. I. Ellis, 572;
E. J. Loring, 573; C. E. Houghton, 579; H. Gansslen,
580; J. S. Myers, 580; Closure 584
Dynamometer, A New Transmission, W. H. Kenerson 171
INDEX lOGl
Effect of Superheated Steam on the Strength of Cast Iron, Gun Iron and Steel,
The, Edward F. Miller 998
Method of Testing Specimens; Results of Tests.
Discussion, see Superheated Steam. ^
EflBciency, Mechanical and Economic, Line-Shaft, Henry Hess 923
Efficiency Tests of Steam-Turbine Nozzles, F. H. Sibley and T. S. Kemble.. 617
Theory of Nozzles; Methods Suggested; Piston Method; Flexible Tube
Apparatus; Forms of Nozzle Tested; Flow Tests; Spring Calibrations;
Search-Tube Tests ; Reaction Tests; Calculation for Efficiency; Results
and Conclusions.
Discussion
J. A. MoYER, 643 ; C. C. Thomas, 645 ; S. L. Kneass, 645 ; Closure 647
Electric Gas Meter, An, C. C. Thomas 655
Characteristics of the Meter; Description of Meter for Gas or Air; Auto-
graphic Records; Operation of the Meter; Calibration of the Meter;
Accuracy ; Energy Required to Operate ; Description of Meter for Steam ;
Theory and Method of Obtaining Standard Results.
Discussion
L. S. Marks, 676; W. D. Ennis, 678; E. D. Dreyfus, 679;. A. R.
Dodge, 679; Closure 679
Engineer in the U. S. Navy, The, Geo. W. Melville 253
Engines, Governing Rolling Mill, W. P. Caine 783
Engines, Single-Acting, Offsetting Cylinders in,T. M. Phetteplace 223
Experience unth Leaky Vertical Fire-Tube Boilers, An, F. W. Dean 799
Description of Boilers; Trouble Experienced; Cause of Trouble;
Remedy.
Discussion
R. P. Bolton, 804, 813; Wm. Kent, 809, 813; J. C. Parker, 809;
O. C. WooLSON, 810; A. A. Gary, 810; L. P. Brecken-
ridge, 812; A. M. Greene, Jr., 813; E. D. Meier, 813;
D. M. Myers, 815; A. Bement, 815; Closure 817
Feeders, Automatic, for Handling Material in Bulk, C. K. Baldwin 161
Ferris, Walter. The Bucyrus Locomotive Pile Driver 905
Fittings for Superheated Steam, Cast-iron, T. N. Hollis 989
Discussion, see Superheated Steam.
Fittings for Superheated Steam, Cast-iron Valves and, A. S. Mann! 1003
Discussion, see Superheated Steam.
Fuel, Tan Bark as a Boiler, D. M. Myers 685
Fire-Service Pumps of Manhattan Borough, New York, High Pressure,
R. C. Carpenter 437
Garland, C. M. Testing Suction Gas Producers with a Koerling Ejector. . 831
Gas Meter, An Electric, C. C. Thomas 655
Gas Power, Marine Producer, C. L. Straub 185
Gas Power Plant, Operation of a Small Producer, C. W. Obert 209
Gas Producers
Marine Producer Gas Power, C. L. Straub 185
Operation of a Small Producer Gas Power Plant, C. W. Obert 209
1062 INDEX
Testing Suction Gas Producers with a Koerting Ejector C. M.
Garland and A. P. Kratz 831
Bituminous Gas Producers, J. R. Bibbins 877
Gephardt, G. F. The Pilot Tube as a Steam Meter 601
Governing Rolling Mill Engines, W. P. Caine 783
Two Methods of Rolling; Power Required; Details of Opeiations; De-
scription of Governor and Indicator ; Discussion of Autograph Records.
Discussion
H. C. Ord, 792; James Tribe, 794; E. W. Yearsley, 796;
Closure 797
Handling Material in Bulk, Automatic Feeders for, C. Kemble Baldwin. 161
Heck, R. C. H. Some Properties of Steam 345
Hess, Henry. Line-Shaft Efficiency, Mechanical and Economic 923
High-Pressure Fire-Service Pumps oj Manhattan Borough, New York,
R. C. Carpenter, 437
Source of Water Supply; Water Required foi Fire Purposes; Motive
Power; Distribution System; Supply Piping; Pumping Stations;
Equipment; Tests; Conclusions.
Discussion
Geo. F. Sever, 462, 485; Wm. M. White, 464; Geo. L. Fowler,
465; J. H. NoRRis, 470; J. R. Bibbins, 471; J. J. Brown,
473; Geo. A. Orrok, 474; Frederick Ray, 475; H. Y.
Haden, 476; T. J. Gannon, 476, 485; H. B. Machen, 478;
R. H. Rice, 480; C. A. Hague, 481 ; A. C. Paulsmeier, 485;
W. B. Gregory, 486; C. B. Rearick, 487; H. E. Longwell,
488; W. M. Fleming, 488; H. S. Baker, 493, 504; E. E. Wall,
501, 505; H. C Henley, 502, 505; Edw. Flad, 504; H. Wade
Hibbard, 504; W. H. Reeves, 505; E. L. Ohle, 506;
Closure 506
HoLLis, I. N. Cast-Iron Fittings for Superheated Steam 989
Httdson-Fulton Celebration 373
Ivens, Edmund M. Tests upon Compressed-Air Pumping Syste7ns of Oil
Wells 311
Joint for Boilers, The Best Form of Longitudinal, F. W. Dean 823
Kemble, T. S. Efficiency Tests of Steam-Turbine Nozzles 617
Kenerson, Wm. H. A Neiv Transmission Dynamometer 171
Kingsbury, Albert. Polishing Metals for Examination with the Micro-
scope 181
Koerting Ejector, Testing Suction Gas Producers with a, C. M.
Garland and A. P. Kratz 831
Kratz, A. P. Testing Suction Gas Producers with a Koerting Ejector 831
Lanza, Gaetano. Stresses in Reinforced Concrete Beams 511
Leather Belting, The Transmission of Power by, Carl G. Barth 29
Line-Shaft Efficiency, Mechanical and Economic, Henry Hess 923
Plan of Tests; Method of Testing; Results; Discussion of Results; Deri-
vation of Constants; Comparison of Actual and Calculated Losses;
Conclusions.
INDEX 1063
Discussion
T. F. Salter, 938; C. A. Graves, 940; C. J. H. Woodbury, 941;
Walter Ferris, 942; F. J. Miller, 942; A. C. Jackson, 942;
C. D. Parker, 943; O. B. Zimmerman, 943; W. F. Parish,
Jr., 943; G. N. Van Deruoef, 944; Closure 945
Locomotive Pile Driver, The Bucyrus, Walter Ferris 905
Locomotives, Safety Valves for, Frederic M. Whyte 105
Machine Members under Eccentric Load, Design of Curved, Walter
Rautenstraxjch 559
Mann, A. S. Cast-iron Valves and Fittings for Superheated Steam 1003
Marine Producer Gas Power, C. L. Stratjb 185
List of Marine Producer Installations; Two Types of Stationary Pro-
ducers; Adaption to Marine Service; Description of Installation on
Great Lake Steamer ; Compaiison of Steam and Gas Equipments.
Discussion
C. L. Straub, 200; Geo. Dinkbl, 202; H. Penton, 202; I. E.
MouLTROP, 204; H. M. Wilson, 205; E. T. Adams, 205;
Closure 205
Meetings op the Society
Annual Meeting 381, 386
Program; Committees; Account of Meeting; Opening Session;
Lecture; Business Meeting; Professional Sessions; Gas Power
Section; Excursions; Entertainment Features.
Boston, Apiil 16 13
Boston, June 11 15
Boston, October 20 382
Boston, November 17 384
Boston, December 17 385
Conservation, Joint Meeting on 6
John Fritz Medal Award 9
Address of Dean W. F. M. Goss; Address of Prof. F. R. Hut-
ton; Address of Robert W. Hunt; Address of Frank J.
Sprague.
New York, January 12 5
New York, February 23 5
New York, March 9 6
New York, October 12 381
New York, November 9 383
St. Louis, April 10 8
St. Louis, May 15 14
St. Louis, October 16 382
St. Louis, November 13 384
St. Louis, December 11 384
Washington Meeting 16
Program; Account of Meeting; Business Meeting; Profes-
sional Sessions; Lecture; Gas Power Section; Presenta-
tion of Portrait of Rear-Admiral Melville; Entertainment;
Elections to Membership.
1064 INDEX
Melville, Geo. W.
The Engineer in the U . S. Navy 253
Presentation of Portrait of 253
Metals, Polishing for Examination with the Microscope, Albert Kings-
bury 181
Meter
An Electric Gas Meter, C. C. Thomas 655
The Pilot Tube as a Steam Meter, G. F. Gebhardt 601
Tests on a Venturi Meter for Boiler Feed, C. M. Allen 589
Miscroscope, Polishing Metals for Examination with the, Albert Kings-
bury 181
Miller, E. F. The Effect of Superheated Steam on the Strength oj Cast-
iron, Gun Iron and Steel 998
Myers, D. M. Tan Bark as a Boiler Fuel 685
Nagle, a. F.
Pump Valves and Valve Areas 953
A Report on Cast-iron Test Bars 977
Necrology
Allen, Walter Morrison 1039
AsHWORTH, A. Kennedy 1039
Blair, Archibald W 1040
Boyer, Francis H 1040
Briggs, Thomas Hallett 1041
Caldwell, Andrew James 1041
Chickering, Kenton 1042
CoRBiN, George VV 1043
Eberhardt, Elmer G 1043
Gildersleeve, David Hamilton 1044
Gray, Thomas 1044
Grover, Lewis Clesson 1045
HiLDRETH, Charles Lewis 1046
Hill, Warren E 1046
Hoe, Robert 1047
HuYETTE, William S 1047
Jennings, Edward L 1048
Jones, Edward H 1048
Landsing, Jang 1049
Lincoln, Robert B 1049
Miller, Alexander 1050
Pierce, Arthur W. K 1050
Rand, Jasper R ] 051
Reed, William Thomas 1051
Reynolds, Edwin 1052
Soxjle, Richard Herman 1053
Wellb AUM, Arvy Elroy 1054
West, George W 1055
Wolff, Alfred R 1055
INDEX 1065
New Departure in Flexible Staybolts, A, H. V. Wille 359
Failure of Former Types ; Formulae for Flexure ; Description of New Bolt ;
Comparative Tests; Calculation of Expansion in Fire-Box.
Discussion
Wm. Elmer, 365; W. E. Hall, 365; Alfred Lovell, 368; F. J,
Cole, 369; Closure 370
New Transmission Dynamometer, A, Wm. H. Kenerson 171
Description; Operation; Important Features.
Discussion
A. F. M.\suRY 178
New York, High-Pressure Fire-Service Pumps of Manhattan Borough,
R. C. Carpenter 437
Nozzles, Efficiency Tests of Steam-Turbine, F. H. Sibley and T. S.
Kemble 617
Obert, C. W. Operation of a Small Producer Gas Power Plant 209
Offsetting Cylinders in Single- Acting Engines, Thurston M. Phetteplace 223
Length of Stroke; Side Pressure of Piston on Cylinder Walls; Inertia
Forces: Vibration and Balance; Conclusions.
Discussion
W. H. Herschel, 248; J. H. Norris, 252; Closure 252
Oil Wells, Tests Upon Compressed-Air Pumping Systems of, E. M. Ivens 311
Operation of a Small Producer Gas Power Plant, C. W. Obert 209
Description of Plant; Operating Results.
Discussion
J. A. Holmes, 219 ; J. H. Norris, 220; W. A. Bole ; . . . . 220
Orrok, George A. Small Steam Turbines 263
Peabody, C. H. The Specific Voluine of Saturated Steam 333
Phetteplace, Thurston M. Offsetting Cylinders in Single-Acting
Engines 223
Pile Driver, The Bucyrus Locomotive, Walter Ferris 905
Pilot Tube as a Steam Meter, The, G. F. Gebhardt 601
Types of Steam Meters; Various Uses of Bitot Tube; Description of
Author's Design; Equations foi Velocity of Flow; Operating Methods;
Results Obtained.
Discussion
W. B. Gregory, 614; Walter Ferris, 614; A. R. Dodge, 615;
Closure 616
Polishing Metals for Examination with the Microscope, Albert Kingsbury 181
Desirable Features ; Preliminary Trials; Method Adopted; Treatment of
Samples.
Portrait of Geo. W. Melville, Presentation of
The Engineer in the U. S. Navy, Geo. W. Melville 253
Presentation of Portrait, W. M. McFarland 258
Acceptance of Portrait, C. D. Walcott 261
Powder Plants
Cooling Towers for Steam and Gas-Power Plants, J. R. Bibbins. ... 725
Operation of a Small Producer Gas-Power Plant, C. W. Obert 209
1066 INDEX
Profession of Engineering, The, Presidential Address, Jesse M. Smith . . 429
Pumping Systems of Oil Wells, Compressed Air, Tests upon, E. M. Ivens 311
Pumps of Manhattan Borough, New York, High-Pressure Fire-Service,
R. C. Carpenter 437
Pump Valves and Valve Areas, A. F. Nagle 953
Need for Change in Specifications; Losses in a Pump; Variations in
Spring Tension; Comparison of Valve Areas and Springs; Recommenda-
tions.
Discussion
C. A. Hague, 962; I. H. Reynolds, 964; F. W. Salmon, 965;
Wm. Kent, 967; R. C. Carpenter, 967; E. H. Foster,
968;Closure 968
Rautenstrauch, Walter, Design of Curved Machine Members under
Eccentric Load 559
Reports
Annual Reports of Council and Committees 407
Finance Committee ^ 413
House Committee 418
Library Committee 419
Meetings Committee 421
Membership Committee 423
Publication Committee 424
Research Committee 426
Report on Cast-iron Test Bars, A, A. F. Nagle 977
Unreliability of Test-Bars; Method of Casting; Results of Tests;
Comparison by Means of Cast-Iion Beam Formula; Relation of Trans-
verse to Tensile Strength.
Discussion
W. B. Gregory, 984; A. A. Cary, 987; T. M. Phetteplace,
987; Closure 988
Rolling Mill Engines, Governing, W. P. Caine 783
Safety Valves for Locomotives, Frederic M. Whyte 105
Methods of Determining Size of Valve; Essentials of a Safety Valve;
Relation of Valve Capacity to Steam Generating Capacity.
Safety Valve Capacity, Philip G. Darling 109
Factors Determining the Area; Description of Testing Apparatus;
Results of Tests; Massachusetts Rule; Philadelphia Rule; Tests to
Determine CoeflBcient of Flow ; Constants for Various Types of Boilers.
Discussion on Safety Valves
L. D. LovEKiN, 129, 139; A. C. Ashton, 130, 146; A. B. Carhart,
131, 143; E. A. May, 132; F. J. Cole, 133; C. E. Lucre, 135;
J. M. Smith, 136; G. P. Robinson, 136; H. C. McCarty, 137 ;
M. W. Sew ALL, 137, 146; A. A. Cary, 138; F. L. Du Bosque,
138; N. B. Payne, 140; H. O. Pond, 142; F. L. Pryor, 142;
E. F. Miller, 144; G. H. Musgrave, 145; A. F. Nagle, 146;
J. J. AuLL, 147; P. G. Darling 148
Sibley, F. H. Efficiency Tests of Steam-Turbine Nozzles 617
INDEX 1067
Small Steam Turbines, Geo. A. Orrok 263
Horsepower Sold; General Characteristics; De Laval Turbine ; Terry ;
Stuitevant; Bliss; Dake; Curtis; Keir; Description of Details; Opera-
tion; New Turbines; Steam Economy; Field for Small Turbines.
Discussion
C. B. Rearick, 287, 297; W. D. Forbes, 288; R. H. Rice; 288,
303, 307; R. C. Carpenter, 291; H. Y. Haden, 292; F. D.
Herbert, 293; W. E. Snyder, 294; F. H. Ball, 296; C. A.
Howard, 29C; W. J. A. London, 296, 306; F. B. Dowst, 298;
C. B. Edwards, 300; V. F. Holmes, 300; J. S. Schumaker,
302; C. A. Read, 302; LN.Hollis, 302; E. F.Miller, 303;
J. T. Hawkins, 303; C. H. Manning, 304; C. P. Crissey,
305; J. H. LiBBEY, 307; Closure 309
Smith, Jesse M.
Biography 3
Presidential Address, The Profession of Engineering 429
Smith, L. S. Stresses in Reinforced Concrete Beams 511
Some Properties of Steam, R. C. H. Heck 345
Pressure-Temperature Relation; Specific Heat of Water; Diffeient
Heat Units.
Discussion
Sanford a. Moss, 355; G. A. Goodenoxjgh, 355; Closure 357
SoPER, E. C. a Unique Belt Conveyor 151
Specific Volume of Saturated Steam, The, C. H. Peabody 333
Thermodynamic Equation; Total Heat of Steam; Heat of the Liquid;
Comparison of Barnes', Dieterici's and Regnault's Results ; Holborn and
Henning's Pressures of Saturated Steam; Comparison of Experimental
and Computed Values.
Discussion
W. D. Ennis, 342; Closure 343
Staybolts, A New Departure in Flexible, H. V. Wille 359
Steam
The Pilot Tube as a Steam Meter, G. F. Gebhardt 601
The Specific Volume of Saturated Steam, C. H. Peabody 333
Some Properties of Steam, R. C. H. Heck 345
Steam Turbines
Efficiency Tests of Steam-Turbine Nozzles, F. H. Sibley and T. S.
Kemble 617
Small Steam Turbines, Geo. A. Orrok 263
Straus, C. L. Marine Producer Gas Power 185
Stresses in Reinforced Concrete Beams, Gaetano Lanza and L.S.Smith 511
Previous Observations; Theories Employed in Making Calculations;
Notations and Analyses of Formulae; Details of Beams Tested; Com-
parison of Results; Conclusions.
Discussion
Chas. T. Main, 522; S. E. Thompson, 522; F. S. Hinds, 525; C.
M. Spofford, 526; J. R. Worcester, 527; Geo. F. Swain,
530; H. F. Bryant, 531; H. E. Sawtell, 532; E. P. Good-
rich, 536; Walter Rautenstrauch, 539; B. H. Davis,
540; C. B. Grady, 542; Frank B. Gilbreth, 543; W. H.
Burr, 543; J. C. Ostrup, 545; E. L. Heidenreich, 549;
C. E. Houghton. 549; W. W. Christie, 550; Closure 551
1068 INDEX
Superheated Steam on Cast Iron and Steel, The Effect of, I. N. Hollis,
E. F. Miller, A. S. Mann 989
Discussion
B. R. T. Collins, 1009; Geo. A. Orrok, 1009; W. K. Mitchell,
1012; John Primrose, 1019; H. S. Brown, 1022; E. H.
Foster, 1022; L. B. Nutting, 1023; A. Lumsden, 1024; J. C.
Parker, 1025; A.A. Cary, 1026;W. E. Snyder, 1026; J. S.
ScHUMAKER, 1030; D. S. Jacobus, 1030; H. F. Rugan,
1031; Closure 103
Tan Bark as a Boiler Fuel, D. M. Myers 685
Ten Physical Conditions ; Calorimeter Tests ; Effects of Leaching; Mois-
ture; Tan Bark Compared to Coal; Chemical Analysis; Evaporative
Tests ; Effect of Pressing and Burning With Coal ; Tan Presses ; Effect of
Small Combustion Space; Draft and Grate Surface; Summary.
Discussion
A. A. Gary, 712; Wm. Kent, 716; F. R. Hutton, 717; Closure 720
Test Bars, A Report on Cast-Iron, A. F. Nagle 977
Testing Suction Gas Producers with a Koerting Ejector, C. M. Garland and
A. P. Kratz 831
Description of Apparatus; Method of Starting Test; Taking Tempera-
tures; Sampling Gas; Measuring Steam and Gas; Method of Calculating
Results; Results of Trials; Items Used in Computations; Guide Sheet
of Formulae.
Discussion
R. H. Fernald, 862, 867; G. M. S. Tait, 866; H. H. Suplee,
866; L. B. Lent, 866; H. F. Smith, 867; W. B. Chapman,
867; E. N. Trump, 868; Closure 868
Tests wpon Compressed- Air Pumping Systems of Oil Wells, EdMUND M.
Ivens , 311
Terms; Description of Systems; Observations; Summary of Results;
Conclusions.
Discussion
F. A. Halsey, 330; S. A. Moss, 330; J. G. Callan, 331 ; Closure 332
Tests on a Venturi Meter for Boiler Feed, C. M. Allen 589
Requirements of a Hot-Water Meter; Description of Apparatus;
Determination of Coefficients; Difficulties Encountered; Results.
Discussion
F. N. Connet, 594, 597; Clemens Herschel, 596; S. A. Moss,
596; Geo. A. Orrok, 597; Closm-e 598
Thomas, C. C. An Electric Gas Meter 655
Transmission Dynamometer, A New, Wm. H. Kenerson 171
INDEX 1069
Transmission of Power by Leather Belting, The, Carl G. Barth 29
Relation of Pulling Power to Tensions at AH Speeds; Problems and So-
lutions; Slide Rule for Solution of Belt Problems; Means of Securing
and Maintaining Definite Tensions; Elastic Properties of Belting; Law
of Variation in Two Tensions of Long Horizontal Belt; Testing Formula
by Lewis' Experiments; Belt Creep; Effect of Centrifugal Force;
Formula for Pulling Power of Vertical Belts; Pulling Power in Terms of
Tight and Slack Tension.
Discussion
H. R. TowNE, 64; W. Lewis, 66; W. D. Hamerstadt, 67;
F. W. Taylor, 71, 80; Chas. Robbins, 74; G. N. Van
Derhoep, 75; W. C. Allen, 77; D. V. Merrick, 81;
F.A.Waldron,83;A.A.Cary, 84; A.F.Naqle, 85;W.W.
Bird, 88; C. H. Benjamin, 89; H. K. Hathaway, 91; W. S.
Aldrich, 96 ; Closui e 99
Turbines, Small Steam, Geo. A. Orrok 263
Unique Belt Conveyor, A,E. C. Soper 151
Description; Operation; Calculation of Work Done; Power Tests;
Initial and Operating Costs.
Discussion
T. A. Bennett, 158; H. Emerson, 159; F. .T. Miller, 159;
Closure 160
Valves
Safely Valves for Locomotives, F. M. Whyte 105
Safety Valve Capacity, P. G. Darling 109
Pump Valves and Valve Areas, A. F. Nagle 953
Venturi Meter for Boiler Feed, Tests on a, C. M. Allen 589
Whyte, Frederic M. Safety Valves for Locomotives 105
WiLLE, H. V. A Neiv Departure in Flexible Staybolts 359
BINDING SECT.
'^'10 14 t^BO
TJ American Society of
1 Ilechanical ::]nGineers
A7 Transactions
V.3I
Engin .
^Q^ ^<^^^^
PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET
UNIVERSITY OF TORONTO LIBRARY