Skip to main content

Full text of "Steel rails; their history, properties, strength and manufacture, with notes on the principles of rolling stock and track design"

See other formats


Google 



This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project 

to make the world's books discoverable online. 

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 

to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 

are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover. 

Marks, notations and other maiginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 

publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing tliis resource, we liave taken steps to 
prevent abuse by commercial parties, including placing technical restrictions on automated querying. 
We also ask that you: 

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain fivm automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attributionTht GoogXt "watermark" you see on each file is essential for in forming people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other 
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner 
anywhere in the world. Copyright infringement liabili^ can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web 

at |http: //books .google .com/I 



>, Google 



I 



(V\<s.G.W,^ Pouttttson 




.,GoogIe 



— . ( 



BMEW" 
Umy 
TF 
3.Si 
.347 






/heui^ A. /U.e/ie.^ 



>, Google 



>, Google 



STEEL BAILS 



THEIR HISTORY, PROPERTIES, STRENGTH 
AND MANUFACTURE 

WITH NOTES ON THE 

PEINCIPLES OF ROLLING STOCK AND TRACK DESIGN 



WILLIAM Hr SELLEW 

PBtNClPAL ASSISTANT ENOINSBR, HtCHlOAIf CENTRAL RAILBOAD 



sei ILLVSTRATIONS-SS FOLDING PLATES 



NEW YORK 

D. VAN NOSTRAND COMPANY 

Twenty-five Park Place 

LONDON 

CONSTABLE & COMPANY, Ltd. 
1913 



>, Google 



COPTOIGHT, 1913 

D. VAN NOSTRAND COMPANY 



>, Google 



MRS. C W. PATTERSON 



PREFACE 

In this work the author has endeavored to systematize the knowledge in 
existence upon the subject, and to present in a concise yet clear form the most 
important features of the problem. 

The fu^t chapter treats of the development of the present design of section 
with a comparison of the American rails with those in use on English Railways 
and on the Continent. 

In chapters two to five, incluave, the external forces acting on the rail 
and the corresponding stresses they produce in the rail are discussed. The 
necessity and desire for information on this subject are widespread. While a 
considerable amount of general mformation is to be found scattered through 
the technical press and in the proceedings of the various Railway Associations 
and En^neering Societies, yet very little has been published dealing broadly 
with the principles of design of the rail in reference to the rolling stock and 
track structure. 

In recent years much thought has been given to the manufacture of rail 
steel, and investigators, it would seem, have turned their attention more to an 
examination of the various defects found in the process of manufacture than 
to the study of the duty of the rail. 

Quite early the question of the intemdty of pressure existing between the 
wheel and tiie rail began to recave attention, but it was not until later that the 
bending stresses in the rail were investigated. Purely theoretical contributions 
to tiie latter subject were made by Zimmerman in 1888. The first practical 
investigations of the bending stress In the rail were apparently those made 
by the United States Government in 1894 by measuring the strains in the rail 
under the static load of the locomotive wheels. These were followed by 
Dr. P. H. Dudley's stremmatograph experiments for measuring the effect 
of dynamic loads. In the time elapsed since the pubhcation of these investi- 
gations hardly anything has been done to further elucidate this problem. 

The sixth chapter deals with the detail of manufacture of the rail. The 
differait stages in the process are described and the influence of each upon the 
finished product is pointed out. It would be outside the limits of the present 



,y Google 



iv PREFACE 

work to attempt a complete treatise on the manufacture of steel; the discussion 
concerns itself, therefore, chiefly with the practical results obtEoned rather than 
with theoretical considerations. 

In the last chapter are given Rail Specifications representing the best 
modem practice in this country and abroad. The forms recommended by the 
American Railway Engineering Association for reports and record of the rail 
are added for the sake of completeness in an appendix. 

The greater part of the bibliography of rail specifications ^ven in article 
39 was prepared for the present work by the Secretary of the American Soci- 
ety of Civil Engineers. The other shorter bibliographies appended to the 
discusMons of several of the subjects were compiled by Mr. McClelland, 
Technology librarian of the Carnegie Library of Pittsburgh. These, which 
are intended to supplement certain parts of the text, are not exhaustive but 
are thought to contain most of the important articles since 1906 which come 
within their scope. A bibliography for the years 1870-1906 with chronologi- 
cal arrangement appears in the Transactions of the American Institute of 
Mining Engineers, Vol. 37, pp. 617-627. 

A comprehenave bibliography of steel manufacture would be so extensive 
as to be unwieldy. Exhaustive bibliographies on this subject appear in the 
various volumes of the Journal of the Iron and Steel Institute, and a good 
selective bibliography of iron and steel manufacture appears in Bradley 
Stoughton's "Metallurgy of Iron and Steel." For the average reader who 
desdres a more detailed discus^on of the processes of manufacture of steel, 
Harbord's and Hall's excellent volume on the Metallurgy of Steel will gen- 
erally be found sufficient. It is believed that these references, together with 
the information contained in the footnotes throughout the book, will permit 
a thorough examination of any of the subjects to be made. 

The work is essentially a compilation. The author has, however, in ev«y 
case endeavored to give credit where anything has been drawn from an outside 
source, and if he has been remiss in this respect it has been unintentional. In 
the discus^on of the granular structure of steel he has been much indebted to 
the work of Mr. J. W. Mellor, from whose writings a conaderable part of 
article 26 has been taken. 

The publications of the " Railway Age Gazette " have been freely quoted 
from and the author wishes to express his appreciation of the courteous permis- 
don given for the use of this material and for the many quotations taken from 
other sources, especially those from articles which have appeared in the Proceed- 
ings of the American Rzulway Engineering Association. 



,y Google 



He has much pleasure in expressing his indebtedness to the following gentle- 
mcm who have given assistance in reviang manuscript or proofs of the parta named: 
Professor Gaetano Lanza, mathematical discussions of Chapters II to V inclusive; 
Dr. A. B. Rerce, checking the author's calculations in these chapters; Professor 
W. F. M. Goss, wheel pressures; Dr. Hermann von Schrenk, forestry; Professor 
W. K. Hatt, strength of tie timber; Dr. P. H. Dudley, stremmatograph tests; 
Mr. Harry D. Tiemann, impact; Mr. James E. Howard, repeated stress; Mr. A. 
L. Colby, manufacture and spedfications; Mr. Robert W. Hunt, influence of 
det^ of manufacture; Mr. Bradley Stoughton, effect of temperature during rolling, 
and Mr. E. T. Howson for examination of the proofs heSore finally going to press. 

WILLIAM H. SELLEW. 

Detkoit, MicmGAN, July, 1912. 



,y Google 



>, Google 



CONTENTS 

Chapteb I. — Development op the Present Section p*oe 

1. Early sections 1 

2. Present sections 14 

Chapter II. — Pbessurb of the Wheel on the Rail 

3. Speeds of modern locomotives 21 

4 Weights of modern locomotivea 29 

5. Effect of excess balance and angularity of main rod 35 

6. Effect of irregularities in the track 45 

7. Effect of rocking of the engine 49 

8. ESect of flat spots in the wheels 54 

9. Impact tests 62 

10. The dynamic augment of the wheel load 69 

11. Electric locomotives 74 

12. Cars 81 

Chapter III. — Supports of the Rail 

13. The tie 90 

14. Bearing of the rail on the tie 122 

15. Fastening of the rail to the tie 138 

16. Strength of the tie 153 

17. Bearing on the ballast 179 

18. Bearing on the subgrade 180 

19. Supporting power of the tie 188 

Chapter IV. — Stresses in the Rail 

20. Stress at point of contact of the wheel with the rail 193 

21. Proposed solutions of the bending stress in the rail 210 

22. Tests to determine the bending stress in the rail 218 

23. Calculation of the bending and shearing stress in the rail 239 

24. Effect of the joint 259 

Chapter V. — Strength of the Rail 

25. Influence of stress and stridn on the strength of the rail 270 

26. E^ect of low temperatures on the strength of the rMi 284 

27. Physical tests of the strength of the rail 288 

28. The strength of the rail and proper weight for various conditions of loadii^ . 310 



,y Google 



Viii CONTENTS 

Chapteh VI. — Influence op Detail of Mandfactuhb p*ob 

29. Chemical composition 326 

30. Extraction of the iron from its ore 344 

31 . Conversion of the steel 366 

32. The Ingot 395 

33. Influence of mechanical work 420 

Chapteh VII, — Rail Specifications 

34. Comparison of American specifications 463 

35. Specifications (New York Central Lines) for basic open-hearth rails 478 

36. British standard specifications of bull head railway rails 484 

37. British standard specifications of flat bottom railway rails 488 

38. Specifications for street railway rails 491 

39. Bibliography of rail specifications 494 

Appendix 

Reports and records 501 

Index 525 



,y Google 



LIST OF ILLUSTRATIONS 

Fio. Faob 

1. ComparisoD of Rail Failures between Different Sections of Bessemer St«el Rails, from Oct. 31, 

1908, to April 30, 1909 9 

2. Increase in Axle Londa, 1885-1907 15 

3. Classification of Locomotives 29 

4. Progress in Locomotive Building 30 

5. Decapod Locomotive of 1903 and American Type of 1857 31 

6. Rail Pressures. Eight-wbeel Engines 35 

7. " " Ten-wheel Engines. (Light Weights.) 36 

8. " " 442 (Atlantic) Type Engines 40 

9. " " 462 (Pacific) Type Engines 41 

10. " " 460 (Ten-wheel) Type Engines. (Heavy Weights.) 42 

11. " " 260 {M(«ul) Type Engines 43 

12. " " 280 (Consolidation) Type Engines 44 

13. Damaging Effect of Badly Balanced Locomotive 39 

14. Profile of Rail from CuSDot's Track Erperimenta 46 

15. Rail Profile taken with a Railroad Automatic Track Inspector Machine 47 

16. "Valley" or Local Depression in Track Profile 47 

17. Summit between Two Depressions of Track Profile 47 

18. Locomotive Driving Wheel Springs 48 

19. Deflection of Ixicomotive Springs 49 

20. Recording Device and Cab Controlling Mechanism for Testing Driving Wheel Springs 50 

21. Recording Device in Place on Driving Wheel Spring 50 

22. GeneraJ Arrangement of Apparatus for Testing Driving Wheel Springs SI 

23. Main Stylus used in Driving Wheel Spring Tests 51 

24. Streas-atrain Diagram. Locomotive Driving Wheel Springs 53 

25. Flat Spot in Wheel 55 

26. Lrregularity in the Roundness of Preaent-day Chilled Car Wheels 58 

27. Apparatus for Measuring the Effect of a Flat Spot 60 

28. Diagram of Testa on Freight Car with Flat Wheels 61 

29. Wire Tests 63 

30. Deformation of Bridge Members under Passing Trains 64 

31. Dynamic Wheel Loads of Typical Passenger Steam Locomotives 72 

32. " " " " " Freight Steam Locomotives 73 

33. Detroit River Tunnel Company's Locomotive 75 

34. Pennsylvania Electric Locomotive in Use in the New York Tunnek 76 

35. Details Pennsylvania Electric Locomotive 77 

36. Typical Load Diagrams for Electric Locomotives 78 

37. Box Car 81 

38. Flat Car 82 

39. Gondola Car 82 

40. Coke Car 83 

41. Stock Car 83 

42. Vestibuled Coach 83 

43. Twelve-section Sleeping Car 84 

44. Steel Combination Passenger and Baggage Car 84 

45. Vestibuted Dining Car 84 



,y Google 



X LIST OF ILLUSTRATIONS 

Pra. Pio, 

46. Baggaf^ Car 84 

47. Typical Load Diagrama tor Cars gg 

48. Typical Dynamic Load Diagrams for Motor Cars 85 

49. 70-foot MoKeen Motor Car 86 

60. 70-foot General Electric Gaa Electric Motor Car 86 

51. Electric Railway Cats 87 

fi2. Electric Railway Cars 88 

53. Typical Load Diagrams tor Electric Railway Cars 89 

54. Carnegie Steel Tie 90 

65. Carnegie St«el Ties on the Bessemer and Lake Erie Railroad 91 

56. Effect of Three Derailmenta on Steel Ties 91 

57. Steel Tie after Four Yeare Service 92 

58. Carnegie Steel Tie with Wedge Faatener 93 

59. Hill Faatening on Carnegie Steel Tie 93 

60. Hansen Steel Tie 94 

61. Universal Metallic Tie on Pennsylvania Linea 95 

62. Snyder Steel Tie 95 

63. Buhrer Combined Steel and Wood Tie on L.S. and M.S. Ry 96 

64. Mexican Railway Steel Tie 97 

65. Buhrer Concrete Tie 98 

66. Bottom Surface of Buhrer Concrete Tie 99 

67. Section of Track on Chicago and Alton R.R., showing Kimball Tie 99 

68. Kimball Tie put in Track on N. Y. C. A.St. L. R.R., July, 1904 100 

69. Kimball Tie, showing Spiking Plugs 100 

70. Percival Concrete Tie 101 

71. Sarada Tie 102 

72. Adriatic Railway Tie 102 

73. Riegler Concrete Tie 103 

74. " " " Appearance in the Track 103 

75. Forest Regions of the United States 107 

76. Hunnewell Plantation 109 

77. Farlington Forest 110 

78. SUndard Prussian Ties of Baltic Pme 117 

79. Standard Oak and Beech Ties on the French Eastero Railway 117 

80. Distribution of Pressure from Tie Plate 118 

81. Half Round Tie Proposed by the Forest Service 118 

82. Spacing of Halt-round Ties 118 

83. Pole Tie 119 

84. Extreme Form of HaF-round Tie 120 

85. Test on McKee Tie Plate 122 

86. Wear of Tie under Tie Plate 123 

87. Loblolly Pine Tie. Section of Tie under Rail Bearing 124 

88. " " " Section- from Middle 124 

89. Belgian State Railways, lOS-pound Rail and Tie Plate 125 

90. " " " 115-pound Rail and Tie Plate...: 126 

91. Kingdom ot WOrttemberg State Railways, Tie Plate 127 

92. Bavarian State Railways, Joint Hook Plate 128 

93. Kingdom of Saxony State Railroad, Joint Hook Plate 129 

94. Elsaas-Lothringen State Railways, Tie Plate 130 

95. Prussian State Railways, Tie Plate 131 

96. Bavarian State Railways, Intermediate Wedge Plate 132 

97. Wooden Tie Plate on French Eastern 132 

98. Plain Bearing Plates, German Experiments on Tie Plates 134 

99. Hook Plates, German Experiments on Tie Plates 135 

100. Hook Pkttes with Clips, German Exoerimcnts on Tie Plates 135 

101. Group 1, German Experiments on Tie Plates 136 



,y Google 



LIST OF ILLUSTRATIONS xi 

102. Group 2, German Ezperimente on Tie Plates 136 

103. Group 3, German ExperimeQla on Tie Platea 137 

10*. Short Leaf Pine Tie, after 2 Years' Service, cut through Spike Holes 138 

105. Croea Section through the Spike Holes of Short Leaf Pine Tie 139 

106. Common Spike 140 

107. Common Screw Spike 140 

108. Screvr Spike used by Grand Duchy of Baden State Railnaya 140 

109. Early French Screw Spikes 141 

110. Machine Preparing Ties for Screw Spikes 142 

HI. Showing AppUcation of Screw Spikes on A. T. & S. Fe R.R 142 

112. French Railways — Rail Fastenings 143 

113. German Railways — Rail Fastenings 145 

114. English and Scotch Railways — Rail Fastenings 146 

115. Screw Spike deduced from European Practice 147 

116. French Screw Spike 140 

117. Wooden Tie Plug used on French Railways 149 

1 18. Collet Trenail 150 

1 19. Cross Section of Pine Tie through Dowel 151 

120. Three Ties of Baltic Pine on the Prussian State Railways 151 

121. Comparative Resistance to Vertical Pressure of Screw Spikes in Pine Ties 152 

122. " " " " " " " " "BeechTies 152 

123. Control Plan — Creosote Tie Teats 161 

124. Tie Plate Forms used in Tests at Purdue University 170 

125. Elastic Curve of Tie, 7 feet 10.4 mches long 172 

126. " " " " 8 feet 10.3 mches long 172 

127. Wood and Composite Ties used in Cufinot's Experiments 173 

128. Measuring Apparatus for Ties under Static Load 174 

129. " " " " " Dynamic Load 176 

130. Results of M. Cuenot's Tests on Ties 176 

131. Strain Diagram of Entire Tie 177 

132. " " " Tie between Rails 178 

133. " " " Tie outside of Rails 178 

134. Bsllaet Eiqteriments — Schubert. Six inches of sand and 6 inches of gravel 181 

J35. " " " Six inches of sand and S inches of stone 181 

136. " " " Stone with thin layer of sand 182 

137. " " " Stone resting on clay subgrade 182 

138. Effect erf Overloading the Subgrade 183 

139. Pennsylvania Track Testing Apparatus 183 

140. Distribution of Pressure to Subgrade 186 

141. Bell's Apparatus for Measuring Depression of the Track 191 

142. Reaction of Tie 191 

143. Compreasion Modulus — Condition of Free Ftow 193 

144. " " Partially Restricted Flow 193 

145. " " Rffltricted Flow 193 

146. Area of Ctutact between Wheel and Rail 195 

147. Relation between Areas of Contact and Load on Wheel 196 

148. Tire Wear, Ten-whee! Engines 198 

149. " " Eight-wheel Engines 199 

150. Two Pieces of a Worn 100-pound Rail after Testing 204 

151. Reciprocating Machine for Testing Flow of Metal in Head of RaU 206 

152. Section of 70-U>. Bessemer Rafl Tested for Flow of Head 207 

153. D'stribution of Tie Pressure under Rail 213 

154. "Claw I" Engine with 75 per cent Impact 214 

155. Track Depression under "Class I" Loading 214 

156. "Class K"Engme with 75 per cent Impact 215 

167. Track Depression under "Class K" Loading 210 



,y Google 



xii LIST OF ILLUSTRATIONS 

Fio. Piia 

158. Bending Moment of Rail placed on Ties 218 

159. Railroad Track Experiments, Boetoo and Albany R.R 219 

160. " " " Photograph <^ Leveling Instrument for Measuring the Depros- 

eion of the Track 220 

161. " " " Photograph of Micrometer for Det«3itiiniiig the Fibre Stress in 

the Baae of the Rail 220 

162. " " " C.B.&Q.R.R 222 

163. Advance Wave Determinations 223 

164. Movement of Rails Laid Alongside of Track 224 

165. Railroad Track Experiments, View showing Micrometer for Measuring Strains in Rails, in 

Position on Base of Rail under Driving Wheel 226 

166. " " " Pennsylvania R.R. Depression In Ballast 234 

167. " " " " " Stress m RaU 236 

168. Stremmatograph Teats at 19 and 40 m.p.h 236 

169. " " " Slow Speeds 237 

170. Wheel Loads for Different Spacing of Drivers 240 

171. Rail Diagram for Wheel Spacing of 60 Inches 242 

172. " " " " " " 70, 80 and 90 Inches 245 

173. Distribution of Horizontal Stress in Rail 248 

174. Shearing Stress of Point Distant y' from Neutral Axis 249 

175. Shearing Stress in lOO-poimd A. S. C. E. Rail 260 

176. Lines of Principal Stress in Beam 261 

177. Diagram of Pieces tested for Sag of Rnil Head and Bending of Web 254 

178. Method of Stationary Testa for Sag of Rail Head and Bending of Web 254 

179. Sag of Rail Head in Stationary Tests .' 255 

ISO. " " " " " ' Rolling Tests 257 

181. Rails after Rolling Test with Load of 90,000 Pounds 268 

182. Shearing Stress in 100-pound A. S. C. E. Rail and Splice Bar 262 

183. 100 per cent Joint 263 

184. Joint showing Uneconomical Distributlan of Metal 263 

185. " " Economical Distri'iution of Metid 263 

186. Diagram of Watertown Arsenal Tests on 100-pound Joints 266 

187. Pure Swedish Iron 270 

188. Pure Copper 270 

189. Copper-bismuth Alloy 271 

190. Iron with 1.8 per cent Carbon 271 

191. Cleavage Planes with Crj^tala arranged Symmetrically 273 

192. " " " " " in an Irregular Manner 273 

193. Iron Strained beyond the Elastic Limit 273 

194. Lead Strained beyond the Elastic Limit 273 

195. Cross Section of Unstrained Metal 274 

196. Cross Section of Metal after being Stressed 274 

197. Slip Bands 275 

198. Polished Surface with Small Cracks 275 

199. " " " Large Cracks 276 

200. Behavior of 0.55 Carbon Steel under Repeated Alternate Stresses 278 

201. Behavior of 0.82 Carbon Steel under Repeated Alternate Stresses 279 

202. Comparison of the Behavior of DifTerent Grades of Steel under Repeated jUtcmate Stresses 281 

203. Number of Repetitions before Rupture in Endurance Tests of Materials 283 

204. Standard Drop Testing Machine 290 

205. Diagram of Tests with Drop Testing Machines of Old and New Design 291 

206. Relation of Work Done in Bending Rail in Drop an-i Static Testa 294 

207. Time-deflection Curve, Masaless Beam, within the Elastic Limit 296 

208. " " Beam Stressed beyond the Elastic Lunit 298 

209. Seleroscope 298 

210. Scleroacope Tests on Open-hearth Rail 299 



,y Google 



LIST OF ILLUSTRATIONS xUl 

PiO. Paob 

211. Scleroacope Testa on Bessemer Rail 299 

212. " " " New Titanium Rail 300 

213. Amsler-LaFFon lostrument for Measuring Hardness 301 

214. Machine for Testing Hail Wear at Pennsylvania Steel Company 304 

215. Diagramof Round Teat Pieces; Tensile Testa on Rail Steel 304 

218. Diagram of Flat Test Pieces; Tensile Testa on Rail Steel 308 

217. Location and Numbers of Test Pieces used in Waterhouse's Teats 309 

218. Effect of Repeated Loads on Beams 311 

219. Rffliatanco of Sub-grade to Pressure of the Track 314 

230. Prices of Iron and Besaemer Steel Rails, 1855-1910. 325 

221. Comparative Wear of Rails of Similar Chemical Composition 327 

222. Tenacity of Iron-Carbon Alloys 329 

223. loSuencoof the Proportion of Nickel and Varying Heat-Treatment upon the Tensile Strength 

of Nickel Sleel 336 

224. Influence of the Proportion of Nickel and Varying Heat-Treatment upon the Ductility of 

Nickel Steel 337 

225. Influence of the Proportion of Manganese on the Tensile Strength of Manganese Steel 337 

226. Influence of the Proportion of Manganese on the DuctiUty of Manganese Steel 338 

227. Tensile Strength and DuctiUty of Carbon Steel and of Manganese Steel 339 

22S. Elasticity and Ductility of Carbon Steel and of Manganese Steel 340 

229. Ore Roasters, Norway Furnace, 1883 345 

230. Open Fit Mine on Mesaba Range, Mountain Iron Mine, near Hibbing, Minnesota 346 

231. View of the West Cut, looking North, Biwabik Mine 347 

232. Steel Ore Dock at Two Harbors, Minn 348 

233. Steamer " Augustus B. Wolvin " 349 

234. Great Northern Railway Ore Dock at Allouei Bay, Superior, Wis 350 

235. The "Wolvin"; a Typical Lake Steamer for the Transportation of Ore 361 

236. Ten-ton Bucket of Unloader in Hold of the "Wolvin" 352 

237. General View of Ore Unloader with Steamer at the Dock 362 

233. Brown Hoist Unloader Unloading Cargo of Ore 353 

239. Blast Furnace with Stoves and Buildings 354 

240. Ground Plan, Showing the General Arrangement of Blast Furnace No. 4, Built at the Hacelton 

Plant of the Republic Iron and Steel Co 355 

241. Sectional View of Hazelton Blast Furnace No. 4 356 

212. Top Rigging of Blast Furnace 357 

243. The ISO-ton Furnaces, Hot Stoves, and Gas-cleaning Plant in Course of Erection at Gary. . 358 

244. The WhitweU Hot-blast Stove 359 

245. Julian Kennedy Stove 360 

246. Isabella Furnace, Cam^e Steel Company 361 

247. Operation of Isabella Furnace on Dry Blast 362 

248. 300-ton Mixer 364 

249. Ten-foot Iron Cupola, Maryland Steel Company 365 

250. Early Experiments of Blowing Air through Bath 367 

251. Bessemer Steel Works, Johnstown, Pa 368 

252. American 5-ton Bessemer Plant. Plan 369 

253. " " " " Section 369 

254. Arrangement of Converters at Maryland Steel Company 369 

255. 18-ton Converter, Maryland Steel Company 370 

256. Typical 16-ton Bessemer Converter 371 

257. Charging Bessemer Converter 372 

258. Bessemer Converter in Full Blast 373 

259. Modem Open-hearth Furnace 376 

260. Open-hearth Plant 377 

261. Wellman Tilting Open-hearth Furnace 378 

262. Pouring Steel into I.Adle nt Open-hearth Furnace 379 

263. Chai^ng Platform of the Open-hearth Furnaces at Gary 381 



,y Google 



xiv LIST OF ILLUSTRATIONS 

Fro. Pau 

264. Hiroult Eleotric Furnace 384 

265. Stassano Electric Furnace 385 

266. Roechling-Rodenhauser Furnace 386 

267. DetaUs (rf Casting Ladle SSQ 

268. Crushed Head 392 

269. Crushed Head 394 

270. Teeming Ingots at Beasemer Converter 396 

271. " " " Open-hearth Furnace 396 

272. Stripping the Mold from Ingots 397 

273. Soaking Pits — Gary 397 

274. Soaking Pits 398 

275. Formntion of Pipe in Ingot 399 

276. Section of Ingot, containing Cavity of 128 cubic inches 400 

277. Bloom from an Ingot of Same Heat and of Same Size as Fig. 276, showing Reduction of Cavity 400 

278. Structure A. — Brinell's Testa 403 

279. " B.— Brinell's Testa 403 

280. " C. — Brinell's Tests 403 

281. " D. — Brinell's Testa 403 

282. " E. — Brinell'a Testa 403 

283. " O. — Brinell's Teats 403 

284. " H. — Brinell's Tests 403 

286. Ordinary Steel Ingot and Titanium Steel Ingot 406 

286. Sulphur in Ordinary Steel 407 

287. " " Titanium Steel 407 

288. Pbosphorua in Ordinary Steel 408 

289. " " Titanium Steel 408 

290. Carbon in Ordinary Steel 409 

291. " " Titanium Steel 409 

292. Influence of Conditiona of Casting as shown by Wax Ingots. (Figs. 1-6.) 412 

293. " " " " " " " " " " (Figs. 7-13.) 413 

294. lUingworth's Preea for Compressing Steel Ingots 414 

295. Williams' Abdominal Liquid Compression of Solidifying Steel Ingots 414 

296. Whitwortb's Hydraulic Press for the Compression of Steel Ingots 4lfi 

297. Hannet's Liquid Compression by Wire Drawing 416 

298. Steel Entering the RoUa 421 

299. Croas Section of 8 by 8-inch Rail Bloom 422 

300. Rail from Early Paaa in Roughing Rolls 422 

301. Same Rail as shown in Fig. 300 after Further Reduction 423 

302. Finished Rail from Same Ingot as Bloom and Pieces from Roughing Rolls 423 

303. Cooling Curve of Solid Copper 425 

304. CooUng Curve of Water 425 

305. Recaleacence 425 

306. Cooling Curve of Iron 426 

307. Coohng and Heating Curves of Steel 426 

308. CooUng of "Solid Steel" 427 

309. The Influence of the Finishing Temperature on the Size of Grain. 428 

310. Influence of Finishini; Temperature on the Size of Grain of Steel of 0.50 per cent Carbon , , . 429 

311. Diagram of Results of Experiments on Rolling at Different Temperatures 430 

312. Raa "B" near Surface 431 

313. " "A" near Suriace 431 

314. " "B" Center of Head 431 

315. " "A" Center of Head 431 

316. Top View at Top of Head, 70-lb. Rail 433 

317. " " " Center of Head, 70-lb. Rail 432 

318. Side View at Top of Head, 70-lb, Rail 432 

319. " " " Center of Head, 70-lb. RaU 432 



,y Google 



LIST OF ILLUSTRATIONS XV 

320. Tranflverae View at Top of Head, 70-lb. Rail 433 

321. " " " Center <rf Head, 70-lb. Rail 433 

322. Pieces (or Mieroecopic Views shown in Figs. 316-321 433 

323. Rail Mill, Algoma Steel Company 439 

324. Housing tor 28-inch Three-high Mill 440 

323. Rolls used in Three-high Rail Mills 440 

326. Three-high Rolls in the RaU Mill at Gary 441 

327. Pass Diagram, Rail Mill, Illinois Steel Company, South Works 442 

328. Rail Mill, Illinois Steel Company, South Works 443 

329. Saw Runs ot American Rail Mills 446 

330. Head Sweep 445 

331. Cold Straightening Press, Maryland Steel Company 446 

332. Value of V/E for Tables XCVI, XCVII and XCVIIi 447 

333. Diagram of C<«ging Rolls, Table* XCVI, XCVII and XCVIII 448 

334. Bar Sections of Passes 14^23 (Tablre XCVI, XCVII and XCVIII) 449 

335. Sections in which only "Direct Pressure" occurs in the Process of Rolling 456 

336. Illustration of "Indirect Pressure" 456 

337. Effect of Inclination of Inner Surface of the Rail Flsi^ on Energy required in Rolling. . . . 457 

338. Work done in Accelerating the Rotating Masses in Reveising Mill 4£9 

339. Recent Rail Sections 460 

340. Shrinkage Allowed in American Specifications in 1909 475 

341. Test Pieces C and D, British Standard Specifications of Rails 487 

342. M. W. 401. — Report of Chemical and Physical Examination 603 

343. M. W. 402. — Certificate of Inspection 604 

344. M. W. 403. — Report of Shipment 606 

345. M. W. 404, — Report of RaU Failures in Main Tracks 506 

346. M. W. 405. — Superintendent's Monthly Report of Rail Failures in Main Tracks 508 

347. M. W. 406. — Annual Statement of Steel Rails Existing in Main Tracks 510 

348. M. W. 407. — Laboratory E;tamination of Special Rails 5H 

349. Standard Locations of Borings for Chemical Analyses and Standard Tensile Test Pieces 512 

350. M. W. 408. -- Summary of Steel-rail Failures for One Year Compared with the Same Period 

of Previous Year 513 

361. M. W. 409. — Summary of Steel-rail Failures for a Period of Years 614 

362. M. W. 410. — Comparative Number of Failures of Steel Rails of Different Section or Pattern, 

Rolled by Different Steel Companies 515 

353. M. W. 411. — Poaition in Ingot of Steel Rails which Failed 616 

354. M. W. 412. — Cover Page for Forms M. W. 408, 409, 410 and 411 517 

355. M. W. 413. — Location Diagram One inch equals one mile 618 

366. M. W. 414. — " " Two inches equal one mile 619 

367. M. W. 415. — Diagram showing Lines of Wear 520 

358. M, W. 416. — Record of Comparative Wear of Special Rail 521 

359. M. W. 417. — Cover Page for Forms M. W. 413, 414, 415 and 416 622 

360. Defective Rail Sheet 623 

361. Diagrwn of Rail Failures, Har 



,y Google 



>, Google 



LIST OF PLATES 

(All plates except V and VI are in the bade of the book.) 
(Plates V and VI are between pages 11 and 12.) 

I. Standuij Rail Sections of the American Society of Civil Engineers. 

II. Rail Sections used b^ore the Adoption of the A. S. C. E. Standard Sections in 1893. 

III. Standard Wheel Sections showing Coning of Wheel. 

IV. Rail Sections used during the Period between the Adoption of the A. 8. C. E. Standard 

Sections in 1893 and tiie Recommendation of the New Standard Sections by the 
American Railway Association in 1907, 
V. Examples of Defective Rails; Broken Rails, Fbw of Metal and Crushed Head. 
VI. " " " " SpUt Head, Split Web and Broken Base. 

VII. Propoeed Standard Rail Section of the American Railway Association, Series "A." 
yjjj ., ,, .. » .. ., ■■ .< .■ g^^ .^3 » 

IX. Standard "P.S. " Rail Section of the Pennsylvania Railroad System. 
X. Rail Sections of the Vignole Type. 
XI. Rail Sections used on German Railways. 
XII. Midland Railway, Permanent Way. 
Xin. L. & N. W. R. Details of Permanent Way. 
XIV. British Standard Bull Head Railway Rails. 
XV. " " Flat Bottom Railway Rails. 

XVI. Rail Sections for Street Railways, Tram Girder Rails and High Tee Rails. 
XVII. " " " " " Standard Girder Sections of the American Electric 

Railway Engineering Association, 
XVIII. British Standard Tramway Rails. 
XIX. Deflection of Driving Wheel Spring, Consolidation Engine No, 1064, Boston & Maine 
Railroad. 
XX. Passenger Locomotive Diagrams, 
XXI. Freight Locomotive Diagrams. 
XXII. Examples of American Tie Plat«e. 

XXIII, Rail Diagram of Love. 

XXIV. Examples of American Rail Joints, 
XXV. Joints Tested at the Watertown Arsenal. 

XXVI. Dynamic Wheel Loads for Various Rails and Axle Spacing. 

XXVn. Bending Momento in Different Weights of Rail Corresponding to Loading on Plate 
XXVI. 

XXVIII. Weight of Rail for Various Conditions of Loading and Classes of Track, 

XXIX, Plan of Gary Steel Plant. 

XXX. Reversing Cogging Mill. 

XXXI. American Three-high Mill, with 36-inch Rolls. 

XXXII. Power required to Roll Rails about 35.5 kg. per meter. 

XXXIII. Form M, W. 418, Am. Ry, Eng. Assn. 



,y Google 



>, Google 



STEEL RAILS 



CHAPTER I 
development of the present section 

1. Early Sections 

Apparently the use of steel rails was first resorted to on account of the 
poor quality of the iron tails of later manufacture. The wear of these iron rails 
took the form of crushing or lamination, which destroyed the running surface 
of the rwl and rendered it unfit for use. An iron rail when manufactured, 
even in the best way, was Uttle more than a bundle of rods; and the top slab 
under the heavy pounding of the locomotive had a tendency to spread side- 
ways and become laminated. A steel rail, on the contrary, was rolled from a 
solid ingot and for that reason was much more durable.* Iron, in the matter 
of wear, exhibited very great irregularity, some rails showing signs of distress 
within a year or two of being laid down, while others afforded very satts- 
factory results. 

t As an illustration of the latter assertion we can instance the experience 
on the main line of the North-Eastern Railway on certain sections of its system 
which may be taken as fair samples of the others. On that extending between 
Newcastle and Berwick, 66.8 miles of double way, the iron rails laid down in 
1847 wdghed 65 pounds per yard. Renewals commenced in 1855 and terminated 
in 1867. In these the weight was increased to 82 pounds per yard. The maxi- 
mum duration of the 65-pound rails was 21 years and the minimum 8 years, 
the average being 12.8 years. 

Mr. T. E. Harrison stated in 1867 that on 700 miles of permanent way of 
the North-Eastem Railway the average duration of the last complete set of 
rmls was found to be about 15.5 years; and some which were laid down in 1834 
were still in use. 

* See The Manufacture and Wear of Rails by C. P. Sandberg, Minutee of Proceedinga of the 
Institution of Qvil Engineers, Vol. XXVII, Session 1867-8, and R, Price Williams's Paper "on the 
Maintenance of Permanent Way," ibid., Vol. XXV, p, 353. 

f Principles of the Manufacture of Iron and Steel. I. Lowthian Bell. London, 18S4. 



>, Google 



2 STEEL RAILS 

The statements just submitted do not afford any proper criterion of the 
resasting powers of iron rails; for this can only be determined by the comparative 
weights of the engines, the amount of traffic, and the speed of the trains which 
have passed over them. According to Mr. R. Price Williams the ava-age life of 
an iron rail, on the most heavily worked portions of the rmlways in the United 
Kingdom in the year 1878, may roundly be taken at about 17^ millions of tons. 

There is nothing speculative in the assertion that iron rails made b^ore 
the complete discontinuance of refining were, gen«tiUy speaking, longer lived 
than those of later manufacture. No doubt in the later days of iron rails the 
permanait way was much more severely taxed than was formerly the case. 
The engines were more ponderous, the traffic was heavier, and the speed greater; 
but the experience of the North-Eastern Rjulway at all events indicates that 
rails of iron have occasionally been made to give very satisfactory results. 
Whether this be due to their having been made from refined metal, or whether 
indeed they were so made, we have unfortunately little means of proving. It 
is significant to note that during the twenty years preceding 1868 the price of 
iron rails had been gradually reduced to one-third of their original cost, and that 
this reduction was accompanied by the production of an inferior quality of rail. 

* In America several of the railway companies began to use steel rails 
as far back as 1864. In that year the Chicago and Northwestern, the Phila- 
delphia, Wilmington and Baltimore, and the Old Colony and Newport each laid 
portions of track with this metal. In the following year the Boston and Albany, 
the Boston and Providence, the Connecticut River Railroad, the Chicago, Rock 
Island and Pacific, and the Chicago and Alton each began the use of steel. 

In September, 1869, a commission appointed to ascertain the extent to 
which steel rails had been tried in the United States ascert^ned that of fifty- 
seven r^ways then in operation, from which reports had been obtained, twenty- 
six had made use of steel in weights varying from 100 to 15,000 tons, the whole 
bulk reported as in use being 49,800 tons, equal to about 518 miles of tracks. 
This, however, did not by any means represent the total weight of steel rails 
laid throughout the States. The Commission already referred to was, indeed, 
particular in calling attention to the fact that on the first of January, 1870, 
there were at least 100,000 tons of steel rails laid down in America, and 10,000 
tons of steel-headed rails besides. Of this quantity the largest bulk had been 
supplied by England, and almost entirely by the Atlas, Barrow and Dowlais 
Works, although several thousand tons had been contributed by three estab- 
lishments in Germany. 

• steel — Ite Hiatory, Manufacture, Propertiea, and Uses. J. 8. Jeans. London, 1880. 



,y Google 



DEVELOPMENT OF THE PRESENT SECTION 3 

In France the Paris, Lyons, and Mediterranean Railway Comp)any decided, 
so early as 1867, to use only steel rails in relaying its permanent way on 860 
kilometers of the Paris and Marseilles line, where more than 10,000 trains ran 
over each line of way yearly, at speeds which might reach 90 kilometers per hour. 

In Austria steel rails were used as early as 1859 on the Northern Railway, 
connecting Vienna with Cracow. They were, however, of puddled steel, manu- 
factured at the works of the Archduke Albrecht, at Carlshutte. 

In 1861 one German mile (about 4.8 miles) of the main line was laid with 
steel rails by way of experiment. So much satisfaction was afforded by this 
trial that in 1865 it was resolved to reconstruct in steel the whole of the main- 
line permanent way, and by the close of that year thirty-five English miles had 
been laid. Previous to this, however, steel had been largely used in Austria 
for railway crossings — so much so, indeed, that at the close of 1864 there were, 
on the Northwn line, 468 steel crossings as compared with 977 of iron. 

In Russia several lots of steel rails were liud down previous to 1872, but 
the use of that metal cannot be said to have received a thorough impulse until, 
in the year named, the Russian Imperial Administration approved the con- 
struction with steel r^ls of the railways from Wjasma to Tula, Rjask, and 
Jetetz, sud from Morschunsk to Siezran, a total length of 1200 kilometers. 
Previous to this time steel rails had been laid experimentally on the Nicolas 
Railway, where they were found to answer so well that in 1872 about 70,000 
tons of steel rails were ordered for Russia, chiefly from Creusot. 

In England, before the end of 1861, steel rails had been laid down on the 
Caledonian, Lancashire, and Yorkshire, London, Brighton and South Coast, 
and Rhymney railways, as well as on the London and North- Western. 

The first Bessemer steel rails made in America were rolled at the North 
Chicago Rolling Mill on the 24th of May, 1865, from hammered blooms made 
at the Wyandotte Rolling Mill from ingots of steel made at experimental Steel 
Works at Wyandotte, Mich. The experimental Steel Works at Wyandotte 
were erected in 1864, and were the first works started in the country for conduct- 
ing the pneumatic or Bessemer process. The rolls upon which the blooms were 
rolled at the North Chicago Rolling Mill were those which had been in use for 
rolling iron rails, and, though the reduction was much too rapid for steel, the 
rails came out sound and well shaped. The first steel rals rolled in the United 
States upon order, in the way of regular business, were rolled by the Cambria 
Iron Company at Johnstown, Pa., in August, 1867,* from ingots made at the 

* See paper on the Development of the American Rail sod Tiack by J. Elfreth Watkina, Trans. 
Am. Soc. of Civil Eagra., April, 1890, Vol. XXII, p. 228. 



,y Google 



4 STEEL RAILS 

works of the Pam^lvania Steel Company, at Harrisburg, Pa., rails were rolled by 
the Spuytrai Duyvil Rollirig Mill Company, at Spuyten Ehiyvil, N. Y., early in 
September of that year, from in^ts made at the Bessemer Steel Works, at Troy, 
N. Y., then owned by Winslow and Griswold, but these were on expmmental 
ordas, and not r^ular ones from any railway company.* 

Before, however, the American steel works had produced any Bessemer 
rails, or, indeed, before any such works had been executed in this country, the 
Fennsylvama Railway Company had imported from England a lot of about 
150 tons. This was towards the close of the year 1863. Some little delay took 
place in slotting the rails to recme the track fastenings; they were not laid 
down until the early part of 1866, when they were placed on sidings in the yards 
at Altoona and Pittsburg, where they would be subjected to considerable use. 
As the rails appeared very brittle, it was not deemed expedient to place them 
in the main track where they would be passed over by trains at high rates of 
speed. None of them, however, were broken in the track, and as they exhibited 
little or no appearance of wear, other steel rals were ordered in 1867, of a quality 
combining more toughness with a sufficient degree of hardness, and experiments 
were continued to test the relativ? merits of the several descriptions of rail. 

About 1864 the Erie Railway Company ordered from John Brown and 
Company, of Sheffield, England, 1000 tons of Bessemer steel rails at 25 £ per ton. 

The American Commission of 1869 concluded, as the result of comparing 
reports obt^ned from twenty-six railways then using steel rails: (1) That ex- 
tremes of temperature do not injuriously affect steel rails. The Grand Trunk 
Railway reports them as not injured by a temperature of 30 degrees below zero 
of Fahrenheit, and no other road appears to find them unable to stand a cold 
winter. (2) That the durability of steel rails far exceeds that of the best iron 
rails. The Erie Railway reports their stee! rails as having outworn thirteen 
sets of iron rals, and as showing scarcely any sign of wear. The Philadelphia, 
Wilmington and Baltimore reports them as having outworn seventeen iron r^ls, 
and as showing little wear. The Chicago and Northwestern say that steel rails 
have outworn fifteen iron rails, and show no perceptible wear. 

In 1874 a small committee of American expertsf conducted a very careful 
and elaborate inquiry into the form, endurance, and manufacture of rails. In 

• private nomnmnication from Mr. Robt. W. Hunt. 

t Appointed Jaouaiy 8, 1873, by the American Society of Civil Engineera, to determine " the 
beat form of standard rut sections of the United States; the proportion which the weight of rails should 
bear to the nuudmum loada carried on a single pmr of wheels of iocomotives or can; the beet methode 
of isanufacturiDg and testing rails; the eadurance, or, as it is called, the ' life ' of rails; the causes, of 
the breaking of rails and the most effective way of preventiog it, and the experieuce of railways in America 



>, Google 



DEVELOPMENT OF THE PRESENT SECTION 5 

Speaking of the comparative value of steel and Iron rails this committee stated, 
that "while steel rails as we get them are tolerably imiform in quality, iron 
varies so much that no comparison can be made except of particular qualities 
or of averages of qualities widely different. We can as yet do little more than 
give the results of our own experience. In so doing we shall not only compare 
steel and iron, but also the effects of some different circumstances on the duration 
of both. It seems probable that the best iron, if homogeneous and the head 
of uniform hardness, so as to wear off evenly like steel, would, with machinery 
of moderate weight, wear a third or even half as long as steel. The chairman 
has found that his 62-pound iron rail, after carrying about 14,000,000 tons 
gross load, has worn off only about 25 per cent more than the steel rails on the 
same track and under the same circumstances. Probably it will not wear so well 
when the top crust is worn tiirough. But owing to want of homogeneousness 
and uniformity the iron scales, splinters, laminates, or somehow dismtegrates 
or mashes in spots before it wears out." 

Ashbel Welsh, the chairman of this committee, subsequently presented a 
final report, giving particulars of the behavior of the steel rails, 53 poimds per 
yard. Bdieving that a very thin stem and a very thin base would possess 
sufficient strength, he designed a pattern in which as much metal as posable 
should be placed in the head, and as little as possible anywhere else. The height 
was 4 inches, the width of base 4 inches, the head fully 2f inches wide and li 
inches deep; radius of sectional curvature of the head 12 inches, stem t'j inch 
thick, base A inch thick at the edges; angle of base and of under sides of head 
14 degrees, length of rails 30 feet, weight 53 pomids per yard. 

The rails were rolled by John Brown and Company, and were laid in 1867 
and 1868 at places exposed to very heavy traffic, on the railway between Phila- 
delphia and New York, where iron rails had lasted only four months. In straight 
portions of the fine, after having carried a gross weight of about 50,000,000 
tons, mostly at high speeds, the heads had been worn down^ inch, having lost 
in weight about 6 pounds per yard. In some sharp curves the sides of the heads 
were so much worn that the rails were taken up in June, 1876. 

The early steel rails were naturally made to the existing iron pattern. 
These were generally pear-headed in order to prevent the side of the head from 
breaking down, and were therefore not adapted to fishing. In 1866, as we 
have seen, Mr. Ashbel Welsh designed a section differing but slightly from the 

in the use of steel rails." See paper on the Form, Weight, Manufacture and Lite o! Rails. A 
Report by Ashbel Welsh, C. E.; M. N. Forney, M. E.; 0. Chanute, C. E.; and I, M. St. John, C. E. 
TraiM. Am. Soc. of Civil Engra., Vol. HI, p. 87. 



,y Google 



6 STEEL RAILS 

modern rail,* and in 1874 Mr. Chanute, chief engineer of the Erie Railway, 
investigated to determine the proper contour of the head by observing 
the contom- of the rails worn down by the action of the wheels. The width 
and shape of the head having been provided for, the rail was considered 
as a beam, and as much metal as possible was taken from the web and flange 
to deepen it. 

Witii the older sections the connections at the joints were very unsatis- 
factory, the design preventing the fishplate from supporting the head. If the 
plate could bear against horizontal surfaces, it would not be forced out laterally 
by the loads, but the rail could not be properly filled by rolling and the play 
would rapidly increase and could not be taken up. Mr. Chanute experimented 
to determine the correct angle of the under side of the head to hold the fishplate 
and found that with an angle above 15 degrees the plate was loosened by 
stretching of the bolts. This relieved the pressure and friction of the plate 
agMHst the nuts and allowed them to turn. He therefore adopted the angle of 
15 degrees under the head, and to avoid unnecessary metal in the flange he 
made its angle 12 degrees. 

The adoption of an improved section was very slow, and as late as 1881 
119 patterns of steel rails of 27 different weights per yard were regularly manu- 
factured, and 180 older patterns were still in use, making a total of nearly 300 
different patterns. This great variety of sections in use required the mills 
to keep a large number of different rolls in stock, and finally to standardize 
the design of the rail the present A. S. C. E. section, shown in Plate I, was 
presented to the society on August 2, 1893. These sections met with favor, 
and were adopted by many railroads, so that in a few years about two-thirds 
of the output of the rail mills conformed to this design. 

The gradual evolution of the present design of rml is shown in Plate II. 
The earlier rjuls show the pear shape of the old iron rails, followed by the rails 
where the section was more adapted to fishing and having a better distribution 
of the metal to afford a stiffer rdl. 

The question of cylindrical tires and flat top rals was one on which there 
existed for a long time a great deal of difference of opinion among railroad engi- 
neer In the early days of railroading the wheels were generally coned to a ratio 
of 1 in 20, and after the organization of tiie Master Car Builders' Association this 
ratio was adopted as the standard. This particular ratio apparently grew out of 

* Robert L. Stevens ia 1830 designed a " T " section of iron raU for the Camden and Amboy 
Railroad, and is generally coneidered to have been the inventor of the flat-footed rail. See Trans. 
Am. Soc. of Civil Engrs., Vol. IV, p. 236, and ibid.. Vol. XXII, pp. 209, 216. 



,y Google 



DEVELOPMENT QF THE PRESENT SECTION 7 

the prevaQing practice at the car wheel foundries, and not from any theoretical 
conaderation of the relation of the wheel to a curve. 

The a^tation for the cyhndrical wheel grew out of efforts to measure the area 
of craitact between the wheel and the rail, to determine the intensity of pressure 
on the metal, and led the Master Car Builders' Association, in 1886, to diange 
thdr standard wheel section and reduce the coning ratio from 1 in 20 to 1 in 38, 
which was about the last draft that would allow free withdrawal from the mold. 

This section is shown on Plate III which also shows the cylindrical wheels 
conadered by the Association at this period. The section thus recommended and 
adopted by the Association passed into general use on the r^ways of the country. 
It was noticed, however, that the change was followed by a large increase in the 
number of broken and sharp flanges, and after u£dng the section for over twenty 
years it was restored to the former ratio of 1 in 20 as shown by the 1910 wheel 
given on the plate. 

The rails of heavier section manufactured within the last few years are not 
giving the service that should be expected of them. The fault may lie in im- 
proper methods of manufacture or in the design of the rail itself, which, while 
suitable for the conditions existing nineteen years ago, may be unfitted for 
the heavy wheel loads of to-day. 

It has been claimed that the old committee of the American Society of 
Civil Engineers did not properly appreciate the importance of low finishing 
temperature in designing their rails, and that the sections recommended in its 
report in 1893 do not permit of a low enough finishing temperature in rolling 
owing to the wide, thin flanges. 

As a matter of fact this was one of the jwints which received most careful 
consideration, not only by discussion between the members of the committee, 
but also in consultation with rail manufacturers. But a peculiarity of the 
situation comes from the fact, that at that time, what we now consider sections 
of necessary weight were then not in general use. The committee was in- 
structed to devise sections from 100 pounds per yard down, decreasing by 5 
pounds, but 80-pound sections were then regarded as the heaviest likely to 
be extenavely used. Only one railroad at that time had heavier sections, and 
that was the Philadelphia and Reading, which had a few 90-pound rails in use. 
The New York Central had put in 80-pound rails, and perhaps they had a few 
heavier ones, but their standard was 80-pound. The Delaware and Hudson had 
adopted the 80-pound r^l, also the Michigan Central. 

The question was to devise a section which the committee considered a 
good one and which could be easily rolled. Unfortunately the sections be- 



,y Google 





1--^, -^ 
















1 


































uiuom 
















i^' 


■- iS5— 
















1 










































A 




««*. 


1^ 


—zi— 

■OHlYlOi 














jw» 


s 


^ 
(q 

^ 














sjj n« 














;U OWH 
















—yg— 

•OH IViOi 
















1 












































'!. 


sia/ivrj 
















3Sra 


i 
















3 
































\ 




UDOiia 


i'' 


















1 


ssvavrj 
















tlM 










-N 




















ililM *""" 


fs 


S3Vmi¥J 
H 1¥101 


















8 

i 


S 
















«* 
































M 


HI 


•JWUf 


i^' 


— -J— - 

ssvmirj 
■oinvioi 






























































—Stl" 

■BM 11/ Wl 
















a 


s 














3 «« 














J 0»,H 














A 




'"818" 

SlUflVVJ 

■OH IVWi 














; 




1 














i 












'^ 




















^Mllllii 


»iiiom 


^1^ 


SI 


















s 
1 
































UliM 

3 




















—or- 

S3ttmi¥J 
















1 














































"ill 




III 


"est- 
















1 


s 










































^Ijll wwove 


set" 

■OK 1¥101 
















1 






























"i^r 
















£<, 


—9S--^ 
ssumiti 
















1 


s 














1 "- 


■^^S 


■ON l¥i01 














^niiliii" x3io»T^ 




Sf" 

S3Hmi¥3 
■ON lYlQl 
















1 
















m 


















1 xiiiiiiiuiiiiiaii 




"WOM 



>, Google 



1 


-—a — 














%\ »ra 


1 


S 












































— f— 

ssy/ivrj 

■OH 7V101 
















































s 
















lis 


















1 


0) 














si Bsn 












^^ 


^m ^"^, 














l§ 


"It IS' 

•Off 7W0i 














Wk 3sn 


^ 








































DMBI 




4o 


—3S— 
■oil itioi 


















i 














































■-■BSl" 
















1 














*|] »« 














■ ■, -1 "-■" 














sH w«w» 


1 


aivmivj 














" - 




1 




















































—S-- 


















1 














































^1. 


"Sir- 

ssiimivi 
















1 


to 
























■i^= 


i^ 


^g zz. 
















■OH If 101 
















1 










_ 




^l]il'll]lll,ll "« 


— 











"H 


imij 


mnous 


1 


-—g — 


















S 


<4 
































av3H 


OMIYIOI 














5 


Kinase 




















1 




■OH 1*101 






























3k 


ariH 


















ll« 


—£t£- 

■OH JWOi 














"^ 




i 
















' 
































^llljll 


mnom 


—819" 

ssamiyj 














:H 3S" 


S 








































3 




.,.«. 




—gs— 

■on 1 If 101 














,„. 


1 


1 












e^^ 


SlipiM s^« 










£p 




II310IIB 
















J^ 


—11— 


-J 












......a """ 


1 


■ON 1*101 














j|B 



















''tis 
■ =^ i § 






•III ° ^ 

rill sl 
ttii i!.! 



>, Google 



10 STEEL RAlia 

yond 80 pounds were matters of compromise, and as they progressed arith- 
metically less satisfactory results were obtained as the weight increased. 

It has been the invariable experience in chan^g from a light to a heavy 
section, in any class of rolled steel, that difficulties have been met and modifi- 
cations have been made in the methods of rolling, in order to get as good structure 
in the heavia* sections as was formerly obtained in the lighter sections. In 
ordinary sections other than r^ls it was a comparatively easy matter to over- 
come the trouble and get a good structure; but the thin flange of the rail, and 
the higher carbons called for in the heavier sections, further complicated 
matters. 

The greatest need at this time is for reliable statistical information taken 
from properly kept records. The Committee on Rail of the American Railway 
Engineering Association have been engaged for several years in collecting 
statistics of defective rails on American roads. The classification adopted by 
the committee is as follows: 

1. Broken Rail. 

2. Damaged. 

3. Flow of Metal. 

4. Crushed Head. 

5. Split Head. 

6. Split Web. 

7. Broken Base. 

It is the intention of the classification that all rails which broke in service, or 
which have a straight crack working from top to bottom or from bottom 
to top, which would very quickly result in a broken rail, should be classified as 
"broken rails," regardless of internal defects. All of the other defective rails 
which are removed, not being "broken" or damaged on account of wrecks, 
broken wheels or similar causes, are to be classified under one of the other 
heads, from 3 to 7, both inclusive. 

Fig. 1 shows a comparison of rail failures between different sections. (See 
Plates I, IV, VII and IX for description of sections.) The most striking char- 
acteristic of the diagram is the comparatively lai^e number of head failures of 
85 N, 85, both on tangent, 36.1 failures per 10,000 tons, and on curve 27.7 
f^Iures per 10,000 tons laid. The legend on the diagram explains that this is a 
Chicago, Burlington and Quincy section. It can hardly be said that the carbon 
is excessively high, although pretty high, imless it is badly segregated, the chem- 
ical constituents bang: 



,y Google 



DEVELOPMENT OF THE PRESENT SECTION 11 

Carbon 48 to .58 

Phosphorus 10 

Manganese .80 to 1.10 

Silicon 20 

Section 852, 85-pound, also a Chicago, Burlington and Quincy section, 
has the same composition, but the failures are not so numerous. 

Carbon 58 

Phosphorus 10 

Manganese 80 to 1.10 

Silicon 20 

The next most numerous head failures are in the A. S. C. E. 90-pound on 
tangent, 15J failures per 10,000 tons, and on curve 12J failures per 10,000 tons 
laid. The P. R. R. 100-pound head fmlures on curve are 12.8 per 10,000 tons 
laid, while the head failures on tangent are small, and the head failures of the 
New York Cenljal 80-pound are about as large, both on tangent and curve, 
11.4 per 10,000 tons laid. The A. S. C. E. 80-pound on tangent, the P. R. R. 
85-pound on tangent, and the A. S. C. E. 85-pound on tangent and curve have 
had the same number of head failures as the New York Central 80-pound. 
The A. S. C. E. 100-pound on tangent and curve comes next, and then 852, 
85-pound on curve, while the rest were all less than 5 failures per 10,000 tons laid. 

The breakages are most numerous in 85 N, 85-pound on tangent, 9.6 per 
10,000 tons l^d, and next of 852, 85-pound on tangent, with the A. S. C. E. 90- 
pound on tangent and the Dudley 80-pound on curve, both the same, following 
closely. Next comes the New York Central 80-pound and 852 85-pound on 
curve and the Boston and Maine 75-pound on tangent, all the same, and then 
A. S. C. E. 100 and 85-pound on tangent. The breakages of the others are less 
than 4 per 10,000 tons laid. 

It will be observed that the breakages of so-called stiff sections are more 
numerous than those of the lower sections with the heavier head. The carbon 
is generally higher in the C. B. & Q. sections than in the A. S. C. E. and P, R. R. 
sections. The web and base fmlures are less than 4 per 10,000 tons laid. 

Plates V and VI present photographs of typical rail failures collected by 
the committee. 

Nearly nine million of tons of Bessemer steel r^ls, from seven different 
mills and varying in weight from 100 pounds to 75 pounds, were reported in 
the tracks of the American railroads on October 31, 1910. This corresponded 
to 21,503,803 rails, and for the twelve-month period from October 31, 1909, 



.yGoogle 



m 



>, Google 



1 iS 

■5 

I 



>, Google 



>, Google 



t^ go: 

■Is 

■8 

I 



>, Google 



12 STEEL RAILS 

to October 31, 1910, there were 30,086 failures or one defective rail for every 
714 rails laid in the track. 

It will be interesting to turn to the conditions of twenty years ago. The 
following table * shows the rail failures on one of the American railways during 
the years 1884 to 1888 inclusive. 

In track, June 1, 1884 121,685 tons 

In track, January 1, 1889 162,526 tons 

Removed from track, 1884 to 1888 inclusive, on accoimt of: 

Broken 1,2931 tons 

Bruised 1,435} tons 

Split 1,353} tons 

Worn out 28} tons 

No fault 35^ tons 

Total 4,147 tons 

The record is deceptive in some respects without an explanation. Many 
of the breaks in the older rails were caused by punching bolt holes. The record 
of the bruised or battered rails, which constituted the lai^est item, would have 
been still greater, except for the fact that long pieces of track laid with soft or 
so-called " pewter" raiU showed up in such bad shape that they were taken up 
and the rails sent to branches or used in sidings after having been in use only a 
few months on the main line. 

The road received the last of these soft rails in 1884, and the record given 
below of the rails received in the following four years is very good. All of the 
rails purchased in this period weighed 65 pounds per yard, but after 1888 an 
80-f)ound section was adopted as standard. It will be observed that there 
were no failures from bruising in the harder rails received after 1884. 



Ye»r Roiled. 


1S8S 


188S 


1887 


1888 


To«l. 


Total number of rails received during year 


23,208 

39 


13 


52 

23,166 

0.22% 

0.06% 

21 


29,171 

5 


12 


17 

29,154 

0.06% 
0.02% 


41,678 

43 


16 
2 

61 

41,617 

0.14%, 

0.07% 

24 


30,366 

i 



1 
2 

7 

30,359 

0.02% 

0.02% 
7 


124,423 


Bruised .. 

















Nnmber of rails in the track, April 1st, 1889 

Detective rails: 


124,286 















• Cylindrical Wheels and Flat Topped Rails for RaUwayB, D. J. Whittemore, Trans. Am. Soc. 
of Civil EngTB.. Vol. XXI, 1889, pp. 185, 186. 



,y Google 



DEVELOPMENT OF T3E PRESENT SECTION 13 

There appears to have been a critical period occurring about every twenty 
years in the history of the rail. When the iron rails replaced the old strap 
irons and other early forms of track construction, they proved very satisfactory 
for the light wheel loads of the day. The wheel loads, however, were rapidly 
increased and soon demanded heavier sections and a metal better able to reast 
wear at the running surface of the ral. It was claimed that the metal in the 
larger sections was poorer than that found in the early iron rails. Various ex- 
periments were tried with rails having steel heads, but it was not until the in- 
vention of the Bessemer process for making steel, which enabled a stronger and 
more uniform rail to be made, that the difficulty was successfully met. 

The use of steel in place of iron for rails commenced about 1865 and enabled 
heavier wheel loads to be used with safety. The early steel rails were generally 
made of mild steel which, while suitable for the loads of the early seventies, was 
found to be too soft for the heavier equipment of the next decade. The situ- 
ation was unfortunately complicated by the experiments on the Pennsylvania 
Railroad which showed, or seemed to show, that low carbon steel I'ails were to 
be preferred to those made from steel of greater hardness, and for several years 
following 1881 the rails were made too soft, and, while there was not a return to 
the serious difficulties of the time of the iron rails, the condition of affairs was 
far from satisfactory. 

Relief was found by increasing the hardening constituents in the steel, 
but with the constant increase in the weight of engines and cars, as well as the 
greater density of traffic incident upon the growth of the industrial resources of 
the country, the situation again reached an acute stage about 1905 when the 
failures of rails became so numerous as to cause the gravest concern on the 
part of those in charge of the operation of the roads. 

The failures as before were principally a question of wear rather than of break- 
age. It appeared that each increase in section produced a rail that wore out more 
rapidly than tiie lighter section which preceded it. This condition was further 
accented by tiie form of the American Sodety of Civil En^eers' sections with 
their thin bases, which turned black in the rolls while the heads were still hot, and 
the fact that the larger sections took more time to cool and so underwent a i}artial 
annealmg, making the metal more readily abraded. 

Three principal reasons were advanced as to the probable cause of the 
poor service of these latter rails. It was claimed that the wheel loads in use in 
this country were exceeding the limits of strength ot the steel in the rail and, 
without resorting to extraordinary methods of manufacture and consequently 
greatly increased cost, the rails could not be made to carry the loads imposed 



,y Google 



14 STEEL RAllS 

Upon them with a proper degree of safety. The standard sections then in use 
were those of the American Society of Civil Engineers and this design of rail, in 
the heavier sections then demanded, was stated to be an impracticable one to roll. 

The manufacturers of rails proposed these explanations as the real reason 
which accounted for the failures of the rails in service. The railways, on the 
other hand, while admitting that the metal of the rails would not stand the 
heavy wheel loads, claimed that this was due to the fact that the steel was of 
poorer quality than that obtainable in rails of earlier make, and that sufficient 
care was not being given to the details of manufacture in the various processes 
at the mills. The increase in the number of r^ f^lures of the type de^gnated 
as "crushed heads" and "split heads" the manufacturers claimed was caused by 
the metal breaking down under the excessive pressure of the heavy wheel loads, 
and the railways contended that they were due to some defect in the structure of 
the individual r^ls. 

No one at all conversant with the situation will attempt to msdntain that 
the subject is not a pressing one. The making of steel nuls for use under 
high-speed passenger trains is something more than a mere commercial propo- 
ffltion. Both the producer and the consumer have great responsibilities in 
the matter, and neither can lay them aside nor shift them upon the other. 

2. Present Sections 

Realizing the impMiance of the question, the American Rmlway Asso- 
ciation appointed a special committee on Standard Rail and Wheel Sections. 
This committee, through a subcommittee on which the manufacturers were 
represented, devoted a large amount of time and attention to the matter of 
sections and spedfications for steel rails and presented a preliminary report to 
the association, October 1, 1907. 

While the A. S. C. E. section was apparently well adapted for the hgjit- 
weight rails of 65 pounds and 75 pounds in use when it was designed, the 
increase in weight on railway wheels (see Fig. 2) necessitated a heavier rail, 
and the manufacturers of rails claimed that it was difficult to make such rails 
of the A. S. C. E. section, due to the thin edge of the base. 

Accompanying the report of the committee were two series of proposed 
standard ral sections: Series "A " designed to meet the requirements of those 
who advocate a rail with thin head and a high moment of inertia, and series "B" 
to meet the requirements of those who think that there should be a narrow, 
deep head, with the moment of inertia a secondary matter. These sections are 
shown in Plates VII and VIII. 



,y Google 



DEVELOPMENT OF THE PRESENT SECTION 



15 



The one known as Series "A" is characterized by a shallow head, wide base, 
thin flanges, and a greats height of section than Series " B." It is appar- 
ently advocated by those who think that more of the duty of the track should 
be borae by the rail and less by the other elements. It is obvious that the 
stronger the rail, as a beam or girder, the more the strains are distributed, and 
the less need, therefore, for exacting attention to the other features of track 
nmntenance. Its advocates think that the distribution of metal between head, 
web, and foot, is such that the rolling difficulties, and especially the question 
of finishing temperatiu^s, can be met with better success. 























y 


r- 










/ 












/ 




,.. 




/ 












/ 










^^ 




















"* 






























'jm ihamm HKivosad 










1 





tk^K lias 

Fio. 2. - 



' Increase in Axle Loads, 1 



oosam 

ol907. 



























noM 




















^ 


' 


1 

1 






^^ 


■ 


' 







^ 






$ 


^ 


■^ 








5 


^ 








































ilmftmit»a«uc»flxlli. 








\ 1 



(Railway and Engineering Review — Wille.) 



It is entirely problematic whether this section will prove die best of the 
two under consideration, and especially whether the transference of more of the 
duty to the rail will result in ultimate track economy. Those who oppose this 
section fear that the shallow head is an element of weakness. According to 
thdr view, with such steel as it is at present possible to get in rails, the pounding 
of the heavy traffic will lead to such crushing and splitting of the heads, owing 
to internal physical defects in the metal, that the section will prove a failure, 
especially on roads with heavy wheel loads and dense traffic. 

Series "B" is modified to meet this latter view. The distribution of metal 
is beheved, as in the "A" section, successfully to meet the manufacturers' criti- 
cism, the head and foot in the 100-pound rail having slightly over forty per 
cent each of the metal, and the web the balance. This section is weaker as a 
^rder than the "A" section, and it would appear that for lighter rails the 



,y Google 



16 ST££L RAIIS 

section "A" should preferably be used to obtain the greatest stiffness. Where 
the wheel loads are sufficiently large to require the heavy head of the "B" 
section, the design of the rail should, however, approach more nearly the latter 
section, even at the expense of lack of stiffness which may be compensated for 
by strengthening the track structure or increasing the weight of rail used. 

There is no good reason why the same characteristics of design should be 
carried out for an entire series. With the excessive loads borne by the heavier 
rails more attention must undoubtedly be given to the effect of the concen- 
trated pressure at the point of contact of the wheel and the distribution of 
this force to the base of the rail. The bending stress while of equal impor- 
tance can be reduced by strengthening the track structure as a whole. Hence 
under the most severe conditions a section should be used in which ample 
proviaon has been made for the former stresses and the question of the bend- 
ing stress, while not lost sight of, becomes of secondary importance. As the 
section decreases in weight the importance of the stiffness of the rail increases, 
imtil in the lighter sections, supporting small wheel loads, a very much higher 
relative moment of inertia is to be desired than in the heavier rails of the same 
series. 

The sections "A" and "B" have been proposed as "recommended prac- 
tice" by the American R^lway Association, and have been referred to the 
American Railway Engineering Association to study and accumulate data and 
make a report after the sections have been sufficiently tried in service to enable 
an opinion to be formed as to their respective merits. 

The American Railway Association Committee, in its report of October 1, 
1907, submitted a statement of cardinal principles which should govern the 
design of a series of rail sections, as follows: 

(o) There should be such a distribution of metal between the head and 
the base as to insure the best control of temperature in the manufacture 
of the rail. 

(6) The percentage of metal in the base of the rail should preferably be 
equal to or slightly greater than that in the head, and the extremities of the 
flanges should be sufficiently thick to permit the entire section to be rolled at 
low temperatures. The internal stresses and the extent of cold straightening 
will be reduced by this means to a minimum, and at the same time the texture 
of the section will be made approximately homogeneous. 

(c) The sections should be so proportioned as to possess as great an amount 
of stiffness and strength as may he consistent with securing the best conditions 
of manufacture and the best service. 



,y Google 



DEVELOPMENT OF THE PRESENT SECTION 17 

(d) The following limitations as to dimension details of the sections are 
conadered advisable for the various weights per yard: 
I. The width of base to be J inch less than the height. 
II. The fishing an^es to be not less than 13 degrees and not greater than 15 



III. The thickness of the base to be greater than in the existing sections of 

cmresponding weight. 

IV. The thickness of the web to be no less than in the existing A. S. C. E. 

sections of corresponding weight. 
V. A fixed percentage of distribution of metal in head, web, and base for 

the entire serira of sections need not be adhered to, but each section 

in a series can be conadered by itself. 
VI. The radii of the under comer of head and top and bottom comers of base 

to be as small as practicable with the colder conditions of rolling. 
VII. The radii of the fillets connecting the web with head and base to be 

as great as possible, for reinforcing purposes, conastent with securing 

tile necessary area for bearing surface imder the head for the t<^ of 

the spUce bar. 
VIII. The sides of the head should be vertical, or nearly so. 
IX. The radii of the top comers of the head should not be lees than | inch 

so long as the wheels continue undo* the present standard of the 

Master Car Builders' Association. 

The principles (a), (6), and (c), above enumerated, appear to cover the 
proper design of T-rail sections. The (d) Kmitations as to dimension details 
should be approached tentatively rather than regarded as a cardinal principle. 

Since October, 1907, a large tonnage has been rolled of rails substantially in 
accordance with the new sections, both series "A" and "B." It has been dem- 
onstrated that these sections can be finished in the mill at a lower temperature 
than the A. S. C. E. sections,* and therefore a finer grained and better wearing 
rail should be secured with the new section. However, great care must be ex- 
ercised at the mills to see that rails are actually rolled at lower temperatures. 
The 90-pound series " A " is now used on a majority of the Western prairie 
roads, and the " B " section is used on the group of coal roads in Maryland 
and Vir^nia. On account of the heavier head found in the " B " section, it 
seems to be preferred by the crooked roads of the East, especially those in the 

* This refers to the temperature of the bead; uo part of the oew eections is finisbed ae cold u 
Uie thin bases of the A. S. C. E. raib. 



.yGoogle 



18 STEEL RAILS 

mountains of Pennsylvania, Vir^nia and Maryland; while on the prairie roads, 
where little curvature is found, the series " A " rail with the lighter head finds 
more general use. 

On Jime 5, 1907, a joint committee of the Pennsylvania R^lroad system — 
Mechanical and Civil Enpneers east and west of Pittsburg — was appointed 
to study the rail question, and on September 20, 1907, their labors resulted in 
the designs for 85-pound and 100-pound rail sections shown in Plate IX. 

It will be noted that the sides of the head are vertical in the 85-pound 
section and sloping in the 100-pound section. This difference is not intentional, 
but arises from the method of constructing the sections. The bearing surface 
underneath the head for supporting the shoulder of the splice bar was con- 
ddered of great imjKirtance, and the committee was not warranted, in the light 
of past experience, in reducing this surface. This was therefore fixed at not 
less than the easting bearing surface. 

As the same equipment is run over the 85-pound and 100-pound sections, 
there was no good reason why the width of the head on top should not be the 
same in each case, and, after studying the contour of the wheel tread, this width 
was fixed at 2J inches. The result was sloping sides in one case and vertical 
sides in the other. 

This section, known as the "P. S." section, is a step farther away from the 
"A" section. It has a still heavier head, a narrower base, and thicker flanges 
than the "B" section. The radius of the web is smaller, thus producing 
more of a buttress where the head and web join. The experience of the 
Pennsylvania system seems to be that with their heavy wheel loads and 
dense traffic, and with the grade of steel that it is now possible to get in 
rails, more rails fail from crushing and disintegration of the head, apparently 
due to the pounding of the traffic, than from any other one cause, and 
accordingly in this section the maximum effort has been made to strengthen 
the rjul in its weakest point. The distribution of the metal is satisfactory, 
and the strength of the rail as a girder or beam is practically the same as the 
"B" section. 

In Europe a T-rail section is used. The Vignole rail used extensively 
abroad was invented in England in 1836 by Mr. Charles Vignoles. Plate X 
shows the type of Vignole rail used on the French Eastern Railway. It is their 
intention eventually to modify the section by decreasing the height of the head 
a little and increasing its width. The weight of the section is 45 kilos per 
meter or about 91 pounds per yard. The maximum axle load is about 40,000 
pounds. 



,y Google 



DEVELOPMENT OF THE PRESENT SECTION 19 

Plate X shows the rail used on the Paris, Lyon and Mediterranean, weigh- 
ing 48 kilos per meter, or 96.7 pounds per yard. The maximum axle load on 
this road is, for passenger service, 38,000 pounds, and, for freight service, 35,000 
pounds. 

Plate XI illustrates types of German rails. The figures show that the 
corner radius of the head in all the rails is -^ inch, as prescribed by the "Tech- 
nics Conventions of the Union," which also recommends a width of head of at 
least 2J inches with a minimum radius of the top of the head of 7J inches. On 
the majority of the German rails the latter radius is from 7| to 8| inches. An 
inclination of ^ is given to the rails, which is obt^ed by the use of wedge 
tie plates. 

Plate X shows the Vignole rail recently used on the Egyptian State Rail- 
ways, of 46 kilos, or 92.7 pounds.* 

In England the idea seems to prevail that a T-rail track is undesirable, 
and a double-headed, or bull-headed, rail is generally used on the En^ish rail- 
ways. Plate XII shows the construction of the permanent way of the Midland 
Railway of England, and Plate XIII the permanent way of the London and 
North Western Railway of England. On the latter road the British standard 
bull-headed rail is used, as shown on Plate XIV. Plate XV shows the British 
Standard flat bottom rail. 

For street railway work either a T-r^l or grooved-rail section is employed, 
as illustrated on Plates XVI and XVII, which show the sections recommended 
by the American Electric Railway Engineering Association. This association 
has only taken up the study of ^^er and high T-raU sections, and the sections of 
tram rails given on Plate XVI, while representative of good practice, are not the 
standard sections of the association. Plate XVIII shows the British Standard 
tramway rails. 

While the T-rail is generally recognized in this coimtry as having superior 
merits for street railway purposes, it is necessary, however, on heavily traveled 
streets to use the grooved rail. The use of a number of different rail sections 
in street railway work is open to the same objections as are found in steam 
railway practice, and the tendency is toward the adoption of standard sections. 

In determining the proper form for a rail, the subject should be considered 
from two points: First, and most important, the duty required. Second, and 
about equally important, the influence of detail of manufacture upon the char- 
acter of the finished product. It is perfectly proper that all the stresses to 

* This rail has been replaced by another weighiog flfi Ibe. per yard, according to British Standard 
sections, which waa iaiS from the year 1911. 



>, Google 



20 STEEL RAII^ 

which it mil be subject should be con^dered and calculated, but its ability to 
resist them will depend quite as much upon the character of the metal as upon 
the form of section. 

In considering the duty, we have first to examine the external forces acting 
upon the rsul, which conast of the pressure exerted by the wheel on the rail and 
the supporting forces represented by the ties. When these are known, the 
stress induced in Uie r^ can be calculated for diffo^nt sections. 



,y Google 



CHAPTER 11 
pre55ure op the wheel on the rail 

3. Speeds of Modern Locomotives 

The eight-wheel or American engine was formeriy the favorite type for 
fast passenger service. The arrangement of this engine provides a four-wheel 
leading truck and four-coupled driving wheels and afforded ample starting 
capadty for the trains of moderate weight used at that time. 

The Atlantic type is the result of the demand for large heating surface and 
grate area in combination with large driving wheels in an effort to meet condi- 
tions which could not be met successfully by the preceding American-type 
engine. The Atlantic-type engine combines a four-wheel leading truck and 
four-coupled driving wheels witii trailing wheels. 

The increase in the weight of the train due to heavier equipment and longer 
trans has resulted in the use of the Pacific locomotive with six-coupled wheels 
in place of the Atiantic type with four-coupled wheels. The latter engine is 
better suited to high-speed service than the former, but it cannot accelerate 
heavy trains to running speed nor maintain speed on grades as well as the 
Paci6c. The internal friction of the Pacific en^ne is much greater than that 
of the Atiantic and it reaches its speed limit sooner, and in fact these powerful 
engines have not been able to show any material increase in the speed of our 
fast trains. 

A train * was recently made up for test purposes which was intended to 
i^resent modem express equipment which could be hauled at high speed on 
level track. The ax cars weighed 350 tons, and the Pacific locomotive 194 tons, 
total 544 tons. The Pacific locomotive, which was selected for its good record 
on that line, was not able to accelerate the train to more than a fraction above 
60 miles per hour on a straight level ti^ack where atmospheric conditions were 
normal. 

On sevwal railways in the West it was for a time thought that it would be 
necessary to electrify the mountain divisions in order to attain speeds which 
would carry the large volume of traffic over the grades and avoid congestion 
and blockade. The Mallet locomotives have overcome this difficulty, and their 

* Railway Age Gazette, January 28, 1910. 



>, Google 



22 STEEL EAILS 

remarkable perfonnance has for the time rendered the electric locomotive on 
mountain lines, where there are no long tunnels, unnecessary. 

On account of the good results obtained by the use of the Mallet compound 
locomotives it will prove interesting to conader the question of adopting these 
machines for fast sendee. The principal advantage to be derived from the use 
of the Mallet type appears to lie in its ability to develop enormous force at the 
draw-bar, but it will be observed that these forces are only possible af compar- 
atively low speeds. 

At speed,* whatever the type may be, it is the boilra- and not the adheaon 
that limits the output of power. The moment the speed is increased by any 
considerable amount, high draw-bar forces become impossible and the wheel 
arrangement peculiar to the Mallet type unnecessary. The assumption even 
of a moderate speed will permit wheel arrangements, now common, to absorb 
the full power of the largest boilers now considered practicable. For ex^nple, 
assume that a locomotive is used which is to have sufficient boiler capacity to 
permit 2000 h.p. to be developed in compound cylinders at all practicable 
speeds. Such a locomotive would require a boiler having in the neighborhood 
of 5000 feet of heating surface which, if fired with coal, would need to be supplied 
with 6000 or 7000 pounds per hour. The draw-bar force equivalent to 2000 h.p. 
for several different speeds is as follows: 

At 1 mile an hour, the tractive force will be 750,000 lbs. 



5 miles " " 


" 160,000 


n 


" 100,000 


10 " 


" 76,000 


20 " 


" 87,600 


SO " 


" 26,000 


60 " 


" 16,000 



Assuming the driving axle of the proposed locomotive to carry a load of 
50,000 pounds, and assuming the adhesion to be 25 per cent, each driving axle 
will serve to develop 12,500 pounds tractive force. A Mallet compound having 
eight axles would be capable of developing a maximiun tractive force of 100,000 
pounds, which force is equivalent to the development of 2000 h.p. in the cylinder 
at a speed of 7| miles per hour. At speeds lower than this the adhesion derived 
from eight axles will not permit the cylinders to develop this rated power, and 
for speeds higher than this the full adhesion of eight axles will not be necessary 
to the development of 2000 h.p. 

• Railway Age Gaaett«, April 22, 1910. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 23 

Table I shows fast and unusual runs in the last three decades.* The 
foregoing table is severely condensed.! The time in every case is from the 
beginning to the end of the run, including stops. In speeds alone, for moderate 
distances, there has been little change since 1895. For example, the engines of 
the Atlantic City Railway make substantially the same time as was made over 
the same line ten years ago; but with the larger boilers and fireboxes now used, 
heavier trains are hauled without loss of speed. The Empire State Express of 
the New York Central, which for years was limited to four cars, now usually 
has five cars, and still makes its trip of 440 miles at the scheduled speed of 

* Locomotive IKctioDaiy, 1909 Edition, Chicago. 

t March I, 1901. — The record of 107.9 miles an hour ia pven by an officer of the road, 
The grade waa descending, mostly at 30 feet per mile. 

March 24, 1902. — Thia run was made on a descending grade, which for some of the way 
was as much as 32 feet per mile. 

June 21, 1902. — This nm is notable by reason of the rising grade. Altoona is 861 i feet 
higher than Harrisburg. 

June 19, 1903. — This ran was made without a stop, but there were two engines. The 
weight of the train was 1 ,008,000 pounds. There are a nmnber of long ascending grades in the line. 

August 8, 1903. — On this and the later run between the same places there were, of course, 
many changes of en^nes. The record gives no data concerning the sizes of the engines, but most 
or all of them were of the most powerful types made in the United States at that time. 

June 9, 1904. — On thia run engines were changed at Bristol. The dimensions given are 
those of the ei^ine used on the second stage of the journey. A car was left at Bristol and the 
weight pven is the average weight for the whole journey. The first engine had drivers 6 feet 
8 inches in diameter, four-coupled; cylinders, 18 X 26 inches. The train making thia run was the 
r^ular mail train scheduled to run regularly, without a stop, from Plymouth to London in 4 hours 
25 minutes. 

July 20, 1904. — This is the best record which has been made over this line. The run of 
June 19, 1906, was made with one more car. 

1905. — Eighteen hours between New York and Chicago is the regular schedule time of one 
daily train each way over the New York Central lines, and one over the Pennsylvania, the latter 
beii^ about 60 miles shorter. There is no published record of less time through, though on many 
occasions the trains of both roads have made up much lost time. The run of November 3, 1905, 
is an example of what has been done in such cases. In this ran the number of cars was three, 
except over portions of the road where a dining car was added, making four. 

October 23, 1905. — This ran and that of May 5, 1906, were not undertaken with a view 
of making the highest possible speed, and each of the divisions over which these trains traveled 
has been traversed no doubt in shorter time; but these transcontinental records are notable for 
the long distances covered, even though the time be not the very highest of which the engines are 
capable. Both of these rans were made by special trains throughout, except that in the ran of 
May, 1906, the ran east of Buffalo was that of the regular Empire State Express. 

June 19, 1906. — On this ran a distance of 12 miles was traversed in 8 minutes (90 m.p.h.). 



,y Google 



li;ll m M s.s: sHI Isjl i m Si %iim I 

■IS8SS SSSS SSSS SS^SS S8S8 S8SS S SS3£ 38S~ SS"8""SS8S , 



ii: 



1 






I 



inMlNiiiniiii 



Inn Hi 



:g : : : ;» : :S :8 3 



|: In: 11 



mm 3 



Nil 



Ml 



11^ 



1 Hi 11 1 



ESfS : :% XX KS 



i :|iii 



il;l ; Nil HlHilii 



InilMl 



IsP. =i;s ik-. »;ii is^s sigl i ^- •^- ri-=|Si 



i It 



iiiiiiJiiii 



aluloi B;jSd IkJJ Jzud eTaaia; Qizdz 12 ft:Bz3 tCJZ zzzja;uj< 



i| I iisissseiisiiiiiii 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 



liilMii: liiiLliiJJtiiMM 



ilin i 



h 



i: 



mm 



Jl-i 



lililll ill II I 



; :S :3E2 :XS S 



liiiillilHiill 



X ;3 : !S ;33 f^i i: 



jii|| i|ii:ii:|:i;i|iii:i 



illi:ini! 



!:i:!!MH;? 



ilMMiii 



lIslH III; mas |K.?sfc5 ium .Jl i:H 



^ S « 33S 



iiH UH iMHI; MM;;::: ; MM; ;;M ;:;; 




m 



iiiiiiii 



iiiiiiiijiiii 



Jill '^^^ 






siiiisiiKsei :iiis!ii 



>, Google 



n 



H 

H 

I 



immmmmiJl 

iiiMijMiiiNiMNiM : ii : I 



III Mil 



X 3 : :!3S 



III ii Nil i 



:li i 



m mm m 



i 



\i \ I 



l!lM:|rni|ii:|;iJ| \ l\ i i 



i^mw^mmimii. g h . i i 



838S 88S8 SSSS S3S 888 



E? 8 3 S ;R3 t; 3 

' ~ii E E£a = i. 



i^if^ssSSm&nH a is s 5 



I ill ill i 



I 



iililiii! i i j 



I 



m§w 



^y. 
i^ 



Mr 



mm i z i I 



i 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 27 

53.3 miles an hour and with remarkable punctuality. Some of the records above 
80 miles an hour lack the elements necessary to make them entirely credible. 

It will be observed that the notable performance of October 24, 1895, 
which made a number of records that stood unsurpassed for years, was set aside 
by another performance equally remarkable just ten years later — October 24, 
1905. 

Table II sbov« the best records for ^ven distances. They are classified by 
speeds alone, no account bnng taken of the modifying effects of load or grade, 
or ^ze of en^e. It will be borne in mind that these records are all those of 
steam locomotives. For distances under 15 miles, the electric locomotive 
which was tried on the Berlin-Zossen line in Germany in 1903 made speeds over 
130 miles an hour which have not yet been equated by any steam locomotive. 

TABLE n. — BEST RECOEDED SPEEDS OF STEAM LOCOMOTIVES 

(LoeooutiTft DiotioiuTy) 





DiattUHW, HUo. 


R«ta,Mil»pBrHoiu. 


Data. 


ROMl. 


1 


3,255 


45,60 


May 5, 1906 


Various. 


2 


2.246 


SO.OO 


July 9, 1906 
Feb. 15, 1897 


A. T. & S. F. 


3 


1,025 


64.27 


C. B. tQ. 


4 


717 


66.00 


Nov. 3, 1905 


Penn. 


5 


625 


69.63 


June 13, 1906 


L, 8. & M. 8. 


6 


257 


74.55 


Oct. 24, 1905 


Penn. 


7 


131 


77.81 


Oct. 24.1905 


P«m. 


8 


55.5 


78.26 


May 14, 1905 


Atlantio City. 


9 


50 


79-00 


June 8, 1905 


Penn. 


10 


15 


98 66 


Mar. 24, 1902 


C. B. & Q. 

S. F. & W. 


11 


4.8 


107,90 


Mar, 1, 1901 



Considering passenger service alone, the crowning achievement of the 
locomotive designers of the past twenty yeara has been not speed alone, nor 
speed and power combined, but speed, power, and reliability. The Penn- 
sylvania trains running dmly between Jersey City and Chicago, 905 miles, at 
50.9 miles an hour, were on time at destination during the year ending June 11, 
1906, 328 times out of 365, or 89.8 per cent of the trips of the year, west bound, 
and 85.2 p^ cent of the trips east bound. Of the 37 late arrivals at Chicago 
14 were not over 10 minutes late. The New York Central reported for its 
gdmilar trains a somewhat less favorable record; but the Central fast trains 
travel at a higher speed, the distance being greater, and the trains were often 
made up of five, ^, or seven cars for a part of the distance. 

From a table published in the Railroad Gazette of January 12, 1906, page 34, 
pving the speeds of a large number of regular scheduled trains between London 
and other English cities, the examples shown in Table III are selected: 



,y Google 



STEEL RA1I5 
TABLE III. — REGULAR ENGLISH EXPRESS TRAINS, 1905 





Rulny. 


UihB. 


*n«?-" 


London to — 




118 
194 
246 
201 
125 
126 
165 
162 














65.7 










Midland 


56 a 



























On the two important long lines of Great Britwn, the West Coast and the 
East Coast routes to Scotland, the best schedules in effect in 1909 were as 
follows: London & North Western and Caledonian, London to Gla^ow (mid- 
night tr^), 401.5 miles; 8 hours; rate, 50.2 miles an hour. Great Northern, 
North Eastern and North British, London to Aberdeen (day train), 523.5 miles; 
11 hours, 7 minutes; rate, 47.1 miles an hour. 

Table IV below shows the best performances of American railroads. The 
fast trains between New York and Philadelphia, which for years were notable 
as the fastest trains in America, are now outclassed by the New York-Chic^o 
tnuns. The Pennsylvania's Chicago trmn is regularly scheduled from Jersey 
City to North Philadelphia, 84 miles, in 83 minutes. 

TABLE IV. — SCHEDULED SPEEDS OF FAST REGULAR TRAINS ON 
AMERICAN RAILROADS, AUGUST, 1906 

(Looomociva Dietioo«ry) 



Jersey City 
New York . 
New York. 
Washington 
Jersey City 



Chicago 
BufTalo. 
Boston. 
Jersey City 
Waehingtc 

Camden . . 
Atlantic 
City 



Southern Pacific, Unicm Pacific, Chi- 
cago & North Western 

Atchison, Tqpeka & Santa F6 

New York Central & Hudson River 
and Lake Shore & Mich. Southern. . 

Pennsylvania 

New York Central 

New York, New Haven * Hartford. . . 

Pennsylvania. 

Central of New Jersey, Philadelphia & 
Reading, Baltimore & Ohio . . 

West JerHev & Seashore (Pennsylvania; 

Atlantic City (Philadelphia &R«ading) 



ISflO 
17:46 
8:15 
6fl0 
4:46 

4:48 
0:52 
0:50 



34.2 
63. S 



47.1 

68.1 
66.6 



At the International Railway Congress held at Berne, Switzerland, July 
4 to July 16, 1910,* Mr. Blum expressed the opinion that the primary reason for 

* Track Strengtbeoing for Increased Weight of Locomotives and Speed of TralnB. Bulletin d 
the international Railway Congress. London and Brusseb, 1910, Vol. XXIV, p. 2497. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



29 



the strengthemng of tracks was not the increase in the ^)eed of the wheel loads 
but the greater increase in the traffic, particularly fraght traffic. The weakest 
part of the track, the rail joint, was stated to be most fatigued not by the fast 
trains but by the slower running trains. Si>eeds of 130.5 miles an hour had 
actually been attained on a line having a weaker superstructure than that now 
used for the express lines of the Prusaan State Railways vnthout any danger 
to the track, and but little anxiety need be felt if higher speeds were used on 
the existing track. That was the opinion of the majority of the railways from 
which he had eUcited opinions. 

In calculating the dfect of the wheel pressure on the track a speed of 60 
mites per hour will be taken for passenger service and 40 miles per hour for 
freight service. As is seen from the preceding tables these speeds are exceeded 
in special runs, but from the evidence of tests we may conclude that, at the 
high^ speeds, while the dynamic force of the wheel increases the track possesses 
at the same Ume a greater resisting power. This subject will receive further 
consideration in the following chapters. 



4. Weights of Modern Locomotives 



two j^ GO 



^ooo 



0440 .<aoo on 



0660 -dnnn ooo 



^oooo oooo- 



2440 ^nOn on 

2662 ^.,000 OOOr. - 

2882 ^^onoo nnnnnA. 

080 ^ OOOO BWMEI 

240 ^qOO *c°UP 

260 ^»000 -~">- 

280 ^nnoon 

2100 



^nOOOOO 



440 ^n nOO 



460 ^^nOOn 



A ncir 



onno 



OOOn 



082 ^oonon 



Annri„„ «,»,.. ot»«j» 


^no„„„ .. . .. 


^nnn„„„ .. . .. 




^„nno„ »«.,■ 


A„nnnn^ 


IKAOO 


^nOnoo„ 


eoUBLIB 


^r.nrinc>r>„ 







^nnn„„ 






^nOOOnn 






^„nnnn„„ 






^„on„„„ 






z^nnno„„„ 






^n^nOn 


TLANT10 


^„^onn„ 


»«!.Fre 


^n„on„n 


CI>UK€D MU.LI CHDn 


^n„onn„„ 




A„„nn„„„ 






Pio. 3. — GkHsiiicatioii of Locomotives (Whyte'a SyBtem). 

The classification given in Pig. 3 will be adopted for the different types of 
engines under discussion. This locomotive classification is based on the repre- 
sentation by numerals of the number and arrangement of the wheels, commenc- 



>, Google 



30 STEEL RAILS 

ing at the front. Thus, 260 means a Mogul, and 460 a ten-wheel en^ne; the 
cipher denoting no trailing truck is used. 

The total weight is expressed in 1000 of pounds. Thus, an Atlantic loco- 
motive weighing 176,000 pounds would be clasmfied as a 442-176 type. U the 
engine is compound, the letter C should be substituted for the dash; thus, 442C 
176 type. If tanks are used in place of a separate tender, the letter T should 
be used in place of the dash. Thus, a double-end suburban locomotive with 
two-wheeled leading truck, ax drivers, and six-wheeled rear truck, wei^ng 
214,000 pounds, would be a 266T 214 type. 

The locomotives shown in Fig. 4 give an idea of the progress in engine 
building from 1832 to 1911. The three engines shown in the illustration were 



t^ 



Fig. 4. — Progress in Locomotive Building. 

built by the Baldwin Locomotive Works. The engine in the upper figure of the 
illustration was the Old Ironsides, built in 1832, and weighed in working order 
about four and a half tons. The engine built in 1876 weighed 103,000 pounds 
exclusive of the tender and had 22,250 pounds on each driving-wheel axle, while 
the modem locomotive appearing in the lower figure weighs 462,450 pounds 
exclusive of the tender and has axle loads exceeding 50,000 pounds. Fig. 5 
shows a fxarther comparison of early and modem locomotives. 

The following tables illustrate the extent of development that has taken 
place in the Pennsylvania locomotives from 1850 to 1910.* 

• Paper on ScientiBc Management ot American Railways by S. M. Felton, at the Coogress of 
Technology held at Boston, Mass., on April 10, 1911, under the auspices of the Massachusetts Insti- 
tute of Technology. 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 







UiO 


1810 


Passenger Locomotives. 


Weight 
Weight 
Weight 
Weight, 


D each driving axte (aF^rox.) 


7,500 
15,000 

30,000 
45,000 


Lbs. 

69,500 

178,500 








272,000 






Freight Locomotives. 


Weight 
Weight, 


n each driving axle (approx.) 


13.000 
26,000 
19,000 
45,000 


54,100 








240,945 







Since 1910 the wdght of the Pacific type 
' engine shown in the table has been 
increased from 272,000 pounds to 292,000 
pounds by the introduction of superheaters and 
mechanical stokers, in the case of some engines. 
The Pennsylvania R^h^iad also have running 
a Pacific type engine, built in 1912, weighing 
317,000 pounds which has axle loads of about 
66,000 or 67,000 pounds, and an expmmental 
Atlantic type engine with 68,800 pounds on an 
aile. 

In Tables V to XI are given wei^ts and 
axle loads of modem engines. The passenger 
en^nes are noticeable for their trailing trucks 
and four-wheel leading trucks, while the freight 
engines commonly use a two-wheel leading truck 
and no trailer, although the Mikado type (282) 
is designed with the trailing truck. This engine 
is a comparatively new development, and is not 
shown in the tables. It is rapidly coming into 
favor for heavy freight service in which the 
Consolidation locomotive was formerly em- 
ployed. 

The weight of the ej^ne can, of course, 
be determined accurately. The maximum rail pressure of a driving 
when the locomotive is running is, however, not at all indicated by the 
load of the wheel on the rail. 



wheel 
static 



,y Google 



32 



STEEL RAILS 



The dynamic augment* is due to several causes: First, the effect of the 
"excess balance" necessary to counteract the reciprocating parts is to cause 
an. impressed load. Second, the angularity of the main rod causes an increase 
of pressure on the m^n wheel. Third, consideration of impact or imperfections 
existing in the rolling stock or roadway. Fourth, the rocldng of the en^ne on 
its 'springs. 



TABLE v. — AXLE LOADS AND TOTAL WEIGHTS OF ATLANTIC TYPE (442) 
LOCOMOTIVES 

(Rom data [gnu*h«d by the American Locomolive CompsDy) 



Name of railroad 

Worka 

Weight on leading truck, pounds, , 
Weight on driving wheels, pounds. 
Weight on trailing truck, pounds.. 
Weight, total of engine, pounds. . . 

Weight of tender, pounds 

Wheel baae, driving 

Wheel base, total of engine 



Wabaah 


Oregon Short Line 


Southern 


Rock Island 






Richmond 


Schenectady 


42,000 


47,500 


42,000 


49,000 












4S,000 


39,500 




191,000 


195,000 


196,000 


202,000 














7' 6" 




30' 11!' 


27' 7' 


29' 6' 


30' 10' 



-AXLE LOADS AND TOTAL WEIGHTS OF PACIFIC TYPE (462) 
LOCOMOTIVES 

(Prom data Inniiihad by the Ameriiiaa lAODmolive CompaBj) 



Name of railroad 

Works 

Weight OD leading truck, pounds. 
Weight on driving wheels, pound) 
Weight on trailing truck, pounds. 
Weight, total of engine, pounds. . 

Weight of tender, pounds. 

Wheel base, driving 

Wheel base, total of engine 



M. & St. P. 




Brooks 


Schenectady 






167,300 


162,000 


41,300 


44,5!)0 




256.000 


152,600 


145.900 








34' 81' 



N. Y. C. 
.Schenectady 
48,000 
170,500 
4g,000 
266.500 
164,500 
14' 0' 
36' 6' 



Pa. Lines 
Pittsburg 

48,000 
176,500 

45,500 
270,000 
144.000 
13' 10' 
35' 21' 



* See Rail Pressures of Locomotive Driving Wheels. Barnes. Trans. Am. Soc. Mech. EngrB., 
Vol. XVI, 1895, pp. 249-289. 



,y Google 



PRESSURE OF "raE WHEEL ON THE RAIL 



TABLE VII— AXLE LOADS AND TOTAL WEIGHTS OF PRAIRIE TYPE (292) 
LOCOMOTIVES 

(From dfttA fdmUhed by the Anurioafl Loconwtlva Company) 



Name of railroad 


C..B. &Q, 
Brooks 

158,000 
25,500 
33,500 
217,000 
148,500 
13' 41' 
30' 81' 


P. L. W. of P. 

Schenectady 
163,000 

28,500 

40,500 
230,000 
140,000 
14' 0' 
34' 3' 


P. R. R. 

Schenectady 
167,000 
27,000 
40,500 
234,500 
139,600 
14' 0- 
34' 3' 


L. 8. & M. S. 


Weight on driving wheels, pounds 

Weight on leading truck, pounds 

Weight on trailing truck, pounds 

Weight, total of engine, pounds 


170,000 
28,000 
47.000 

245,000 








34' 3' 







- AXLE LOADS AND TOTAL WEIGHTS OF TEN-WHEEL TYPE (460) 
LOCOMOTIVES 

( Fram dkU (uniubad by the Americu LooomotiTe Company) 
FREIOHf LOCOMOTIVES 





C. * N. W. 

Schenectady 

135,000 

46,000 
181,000 
143.500 
14' 10' 
25' 10' 


D. N. W. & P. 

Schenectady 
143,000 
47,000 
190,000 
149,000 
14' 10' 
25' 9' 


St. L. A S. F. 
Brooks 
142,000 
60,000 
192,000 
127,000 
14' 10' 
25' 9' 


C. P R. 






Weight on driving wheels, pounds. . 
Weight on leading truck, pounds. . . 
Weight, total of engine, pounds — 


143,000 
52 000 
195,000 






Wheel base, total of engine 


29' 1- 



PASSENGER LOCOMOTIVES 



Name of railroad 

Works ■■ ■ 

Weight on driving wheels, pounds. 
Weight on leading truck, pounds. . 
Weight, total of engine, pounds. . . 

Weight of tender, pounds 

Wheel base, driving 

Wheel base, total of engine 



Ore. R. R. A Nav. Co. 
Brooks 

181,000 
44,000 
205,000 
164,000 
13' 10' 



N. Y. C. 

Schenectady 

158,000 

51,000 

209.000 



_ 



D, N.W.&P. 

Schenectady 
161.500 
49,500 
211,000 
142,000 
14' 10' 
25' 9' 



D., L. & W. 
Schenectady 

170,000 
48,000 
218,000 
135,500 
14' 4' 
25' 6' 



,y Google 



STEEL RAILS 



-AXLE LOADS ANt) TOTAL WEIGHTS OF MOGUL TYPE (260) 
LOCOMOTIVES 

(Pram data bniuhed by the AmertcaB Looamotive Compuy) 



Name of railroad 

Works 

Weight on driving wheels, pounds.. 
Weight on leading truck, pounds, . , 
Weight, total of engine, pounds. . . ■ 

Weight of tender, pounds 

Wheel base, driving 

Wheel base, total of engine 



K.CM.&O. 
Pittsburg 

14S,000 
25,000 
173,000 
146,000 
14' 2' 
23' 4' 



So. Pacific 
Brooks 
150,000 
27,000 
177,000 
139,000 
16' 2' 
24' 0' 



D., L. 4 W. 

Schenectady 
156,500 
21,000 
177,500 
124,000 



15' 



' 10' 



Vandal i a 
Schenectady 
169,000 
28,000 
187,000 
144,000 
14' 9' 
23' 10' 



TABLE X. — AXLE LOADS AND TOTAL WEIGHTS OF CONSOLIDATION TYPE (280) 
LOCOMOTIVES 

(From dBla (uniiBhed by th> Amerioui LfKODHHive Company) 

Name of railroad 

Works 

Weight on driving wheels, pounds. 
Weight on leading truck, pounds. . 
Weight, total of engine, [Kiunds. . . 

Weight of tender, pounds 

Wheel base, driving. 
Wheel base, total o7 < 



if engme . . 



Mich. Cen. 


C, I. AS. 


Union R. R. 


D. &H. 


Schenectady 


Brooks 


Pittsburg 


Schenectady 










26,500 




25,000 


29,000 


241,000 


241,500 


251,000 


252,000 


161,000 


148,000 


136,000 


151,000 






15' 7' 


17' 0' 


26' 5' 


26' 5* 


24' 4' 


25' 11' 



- AXLE LOADS AND TOTAL WEIGHTS OF FOUR-CYLINDER ARTIC- 
ULATED COMPOUND LOCOMOTIVES 

(Pnini daU (mnuhcd by tfae Amarkaii Locomotive Compuiy) 



Type 

Name of railroad 

Works 

Weight on driving wheels, poundt 
Weight on leading truck, pounds 
Weight on trailing truck, pounds 

Weight, tutal of engine 

Weight of tender. 
Wheel base, drivi 



Whei 



base, total <^ engine. 




445,000 

167,000 
9'& 14' 9' 
40' 2' 



.)gle 



pressure of the wheel on the rail 35 

5. Effect of Excess Balance and Ahgularity of Main Rod 




















p 


X 








p 


\ 




x^ 


<^ 

















AOMi. pen HH. 









fr 


V 








1 


A 














\, 


/ 






50C» 


V 


^/ 










flo m- 


°en Hn. 






■LEFT onivens 



Fia. S. — Rail Pressures. Eight-wheel EngineB. (Am. Ry. M. M. Assn.) 

The effect of the excess balance and the angularity of the main rod can 

be accurately calculated, and is shown on the diagrams given in Figs. 6 to 12.* 

The calculation for the diagrams given on Figs. 6 and 7 were all made from 



* FigB. 6 and 7 are taken from Proceedings Am. Ry. M. Mech. Assn., 1 
froiQ data kindly furnished by the American Locomotive Co. 



Figs. 8 to 12 a 



,y Google 




STEEL RAII£ 





/- 


N 


/- 


\«.,i 






























— 




^ 






lOOOO 













•o at> WD sao 






'if\ 




r 


^ 


t 


\ 


■eooo 










\\ 


i 


"^ 


^ 






/ 








M 











30000 










.5000 




V 


^ 


^ 


lOOOO 


^*i= 


.«r* 


ns: — 





^^ 






^^ 



iV 



5^ 



iwiT onivem^ ,Lerr mivcrs 

©©©a© 




Fia. 7. — Rail Preaaures. Ten-wheel Engines (Li^t WeigbU). (Am. Ry. M. 

data obtained from eight- and ten-wheel engines on the C. M. & St. P. Ry. 
The following are the principal dimensions and weights of each * 

' Report of Committee on the Wear of Driviog-Whed Tires, Proceedings Am. Ry. M. Mech. 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 









Cylinders 

Steam pressure 

Diameter driving centers 

Driving wheel base 

Length of main rod 

Diameter piston rod 

Weight of reciprocating psrta, & 
Weight on drivers 



16 by 24 inches 

160 pounds 

S6 mches 

8 feet, 6 inche 

7 feet, 21 inches 

2] inches 

480 pounds 

54,000 pounds 



19 by 26 inches 

150 pounds 
56 Inches 

io feet 

3i inches 

739 pounds 

84,000 pounds 



The piston, piston rod, crosshead, and front end of the main rod are taken 
as reciprocating parts, the back end of nuun rod as a revolving weight, in all 
calculations which follow. 

The weights of the ends of the rods were found by supporting each end 
at the center of the box or bearing, and resting them alternately on scales. 

The eight-wheel engines had the entire weight of the reciprocating parts 
balanced, by adding one-half this weight in each driving wheel to the weight 
necessary to balance the revolving parts when weighed at the crank pin. The 
ten-wheel engines were not counterbalanced alike, but all agreed in having 
the forward and back wheels overbalanced; that is, with a heavier counter- 
balance than that required to balance the revolving parts only; while the main 
wheels of thirty-five of the fifty-three en^nes from which measurements were 
taken were underbalanced for the revolving parts alone, and all of them under- 
balanced according to the rule of adding to the weight necessary to balance 
the revolving parts two-thirds of the weight of the reciprocating parts, divided 
equally between the driving wheels. 

The counterbalance in the wheels of each of these engines was carefully 
weighed by resting the journals of each pair of drivers on level straight edges, 
pladng the crank horizontally, and hanging on the crank pin a sufficient weight 
to just balance the counterbalance opposite. From this weight the weight of 
the revolving parts attached to that pin was subtracted, the remainder bdng 
the amount of overbalance weighed at the crank pin. If the weight of the 
revolving parts exceeded the weight so found, qt course the wheel was under- 
balanced by the amount of such excess. 

The actual average condition of the coimterbalimce in the wheels of the 
fifty-three ten-wheelers was as follows; 

Average overbalance weighed at the crank pin above that required to 
balance revolving parts only: 

Front wheel 271 pounds overbalance 

Main wheel 80 pounds underbalance 

Back wheel 237 pounds overbalance. 



,y Google 



38 STEEL RAII^ 

These weights* are used in the calculations for the ten-wheel engines 
plotted on Fig. 7. 

The following formulae have been used in calculating the forces in action: 

NOTATION 

P = Pressure of one driving wheel on rail. 
W = Weight of each wheel on rail, engme at rest. 
C = Centrifugal force of the excess weight in the counterbalance over that 

required to balance the revolving parts. 
A = Horizontal acceleratmg (or when negative retarding) force of the recip- 
rocating parts. 
Pi - Pressure against crosshead pin from steam in cylinder. 

= Angle of the crank with the horizontal. 

N = Ratio of length of main rod to length of crank. 

Hence, P ^W - Csin o + ^ ,^'7^^ ■ (1) 

V sm*a 
But, 

w = Weight of the excess in the counterbalance over that required to balance 

the revolving parts. 
t = Velocity of the center of gravity of the overbalance, 
r = Radius of the center of gravity of the overbalance. 
w' = Weight of the reciprocating parts. 
v' = Velocity of the crank pin. 

1 = Length of the crank. 

g •= The acceleration of gravity, 32.16. 

Hence, C = ^'- (2) 

gr "• ' 

A = ^'coso. (3) 

Or, by substitutmg in (1) the values of C and A given in (2) and (3), 

P=W~^sina-i-^- g' ^ (4) 

V sin' o 

The above formulse include the centrifugal force of the overbalance in the 
drivers, the effect of the acceleration and retardation of the reciprocating parts, 
and the angularity of the main rod. Formula (3), for the acceleration of the 
reciprocating parts, assumes that they move as they would were the main rod 

• These weights are the equivalent weights at & distance from the center equal to the crank length, 
and oot the actual counterbalance w^ghta used. 



>, Google 



PRESSURE OF THE WHEEL OK THE RAIL 39 

infinitely long, but the error this produces is too small to affect the accuracy of 
the results, while the formulaa are much simplified. 

The left-hand ends of the diagrams correspond to the position of the engine 
when the right crank is on the forward center, positive rotation being that 
produced by running the engine forward. 

The pressures upon the piston used in the calculation for Figs. 6 and 7 
were obtained from actual indicator cards taken at these speeds, and with a 
point of cut-off found by the examination of a large number of cards to be the 
usual point at which an engine is worked at the speed taken. 

The points of cut-off used are: 

Eight-wheel engine, just starting, 22 inches; 40 miles per hour, 6 inches; 
60 miles per hour, 6 inches. 

Ten-wheel engine, just starting, 22 inches; 10 miles per hour, 13 inches; 
20 miles per hour, 11 inches; 30 miles per hour, 8 inches; 40 miles per hour, 
6 inches; 60 miles per hour, 5| inches. 

Curves for just starting, ten and twenty miles per hour, show that the total 
pressure of the main driver on the rail is always greater at these speeds and 
cut-offs than the actual weight of driver on the rail when the engine is at rest. 



Fio. 13. — Ditmaguig Effect of Badly Balanced Locomotive. 

This is due to the angularity of the main rod always causing an increase of 
pressure on the main wheel. There is, of course, a corresponding upward 
pressure on the guide, reducing the weight on the truck. 

Figs. 8 to 12 are for heavier engines and are calculated from some of the 
largest en^es that have been built of each type. 

Yig. 13 shows the damaging effect upon the track of a badly balanced 
locomotive. 



.yGoogle 



STEEL RAILS 






^ 


S 


^ 


>T 




/ 


\ 


/ 


s 


80000 


-^=^. 


tt Ml. «i| 


nrs^ — ' 


=5S. 











,P10 




^ 


-^ 


y 




80O00 






W HB. 


•=;&. 



- — I — .l)f' 



E 



^xJ 






// 


..•>,\ 




^-1 


-^t 




^^ 




•-.,_ 






_ 









A 




b 


3 




^ 


laeooo 

IIOOOO 


V 




/^WM 






^ 




^ — 














60000 



', Workini 

Google 



Fio. 8. — Ron PwMurea. 442 (Atlantic) Type Engines: Cylinders 211' X 28', Wheels 79", Working 
FreflBure 180 lbs. (Am. Locomotive Co.) 



PRESSURE OF THE WHEEL ON THE RAIL 




STARTING 





^ 


\ 


/^ 


^' 








L= 















OO ISO >Tr. w^ 








-^0- 
^ 








y 






\ 
























ERHB. 









^\ 


N 


/^ 


^. 






H 


-•=- 


P^ 








PER HR. 











^^ 


M^O. 


ooooo 




N 


/ 


\ 




-■^-z-J 


__v 




R»K_V 






























A 


aioooo 








^ 










eo»i,pn 

JOM.PH 


leooo* 














v> 




40NPK 






\^ 










AM-m 






l«OO00 



©©©o© 



FiQ. 9. — RaU Prewuree. 462 (Pacific) Type Enpnes: Cylmd«s 22* X 28*, Wheels 79', Working 
Preasure 200 lbs. (Am. Locomotive Co.) 



>, Google 



STEEL RAILS 






; 


N, 


> 


^.. 


•oo 


/ 


\ 


/ 


\ 










'4?iu 


soooo 




ie"^i:"p 



-J^r^^. 









^ 


^,. 






■^ 


/ 




SOOOO 


U-+--^ 


-«B «a 


R ' 











K-« 






^ 


/,<.. 


--Ss 




"T— - 




/ 


...N 



























A 










n 










^ 






^ 










V / 




""■■" 






\/ 










\^ 


■•»"■'•' 







DO© ©OOO© 

lera 22' X 26', 

Google 



FiQ. la — Rail Pressures. 460 (Ten-wheel) Type Engines (Heavy Weights) : Cylindera 22' X 26', 
Wheels 89', Working Pressure 200 lbs. (Am. Locomotive Co.) 



PRESSURE OF THE WHEEL ON THE RAIL 



,0000 


RIGHT SIDE 


A 


N 


^ 


^ 




/ 


\ 


/ 


\ 















/' 


^ 


/■ 


N 




/ 


\ 


/ 


\ 





























A 


~\ 




/ 


\ 


/ 


\ 


80000 


--=-- 


^__„ 





















r 


\ 








//' 


w 




^ — 


—— =^ 














V 


^y 













/; 


■n \ 








// 


\\ 


...J. 


t^— 






: 


00000 
40000 


v^ 


— / 








^QO© ©©00© 



Fto. 11. — RaU Preasuree. 260 (Mogul) Type EnginMi Cylindere 21' X 28', Wheels dT, Wcaking 
Freesure 200 Ibe. (Am. Locomotive Co.) 



>, Google 



STEEL RAILS 





«.""" s 


£• ^ - 




r 


k 




-N;? 




/ 


\ 




\ 


■oooo 




«■-■> , 








' 




r 


V 




/ 


\ 


/,*tto 


-k 


aoott) 






_jft»- 










40000 






"^ 


:a 






V 


:.-A 


s 


„, — 


y 


\ 


\v 


'y 








■■OMT 


SIDI 











^ 








/ 


^^ 


asoooo 






/ 


^ \ 




=. 


1 








A 




V 

^ 






— ' 




"— ' 












aioooo 
aooooo 




■^ 




"— 




\^ 








\/ 


L„ 




<eoooo 




/ 








u 


— ■ 





X 32', Wheel 

,y Google 



FiQ. 12. — RaaPrMsuree. 280 (Consolidation) Type EnRJnea: Cylindera 23' X 32', WheelB 63*, 
Woriting Pressure 200 lbs. (Am. Lonomotive Co.) 



pressure of the wheel on the rail 45 

6. Effect of Irregularities in the Track 

Fig. 14 shows the exaggerated profile of the rail observed by M. Cuenot 
in his track experiments. 

Figs. 15 and 16 show the rail profile taken with a Railroad Automatic 
Track Inspector machine. These diagrams show the unloaded profile of the 
rail, or the permanent set left in it by the passage of the trains. Evidently 
the loaded profile will be below the unloaded line, and both profiles will probably 
show the same general features, as indicated by the approximate loaded position 
of the rail shown by the dotted line in Fig. 16. 

The wheel as it passes over the curved surface of the rail shown in the 

figure is constrained to move in a curved path whose radius is about 5000 feet, 

Mo* 
and the pressure of the wheel on the rail is the centrifugal force, C = -j^ > 

ti 

directed away from the center of curvature. For 30,000-pound wheel loads, 

u.^ 

J 82.16 

For 60 miles per hour, 

5280 X 60 

60 X 60 ■ 

and R = 5000. 

„, , „ Mr" 30,000x88x88 ,.._ , 

Therefore C-^. 32.16x5000 - '^^ pounds, 

which is the excess wheel pressure caused by the irregularity in the track shown 
by the figure. To be on the safe side, it would seem desirable to increase this 
amount. If, however, 4000 pounds be taken to represent the excess wheel 
pressure, due to this cause, an ample factor of safety will apparently have been 
provided. 

It will be seen from the above that the increase of wheel pressure, due to 
any change in the grade line, will be so small as to be negligible. It is good 
practice to change from one rate of grade to another with a vertical curve, 
changing the grade at each 100-foot station by 0.1 feet; this would give a radius 
for the vertical curve of 50,000 feet, and a corresponding value for C of about 
150 pounds. 

Let us now consider the path of the wheel when passing over the summit 
between two of the depressions shown in the track profile. Fig. 15. When tie 
wheel is in the act of leaving the valley, or depression, its path lies in a direction 
away from the surface of the rail before it. It is, however, under the influence 
of two forces, — neglecting for the moment the action of the springs. First, its 



i feet per second, 



,y Google 



STEEL RAIIS 




>, Google 



PRESSURE OF THE WHEEL ON THE RAH, 47 

momentum, acting along a line of direction tang«it to the vertical curve of the 
rail; and, second, the force of gravity. The trajectory of the wheel acting under 



^V^VvWi^/V-v^ JL 



la. — Rnil Fn^le taken with a Railroad Automatic Track Inspector Machine. 



^^ 






















\ 












^ 






SURFACE OF TRACK, 


J 




■•^^ — .^_ 




^ 







-,^ 






- 






""--» "" 


^.o";!©-- '>AfH 0|F WHEEL 



Fia. 16. — " Valley " or Local Depresuon in Track Profile. 

these forces will be a parabola with its axis vertical. The greatest height of 
ascent, tf, and the horizontal range, x, are pven by the following equations: 
y =h sin* a, 
I = 2 fe sin 2 a. 
h being the ideal height due to the velocity, we therefore have for a speed 
of 60 miles per hour, 

- »^ = 2fffe, 

2ff 



" 2 X 32.16 



= 121 feet. 



Fig. 17 is derived from the 
same record as that from which 
the diagram of Fig. 15 has been 
taken and shows a summit be- 
tween two depressions in the 
profile of the track. We see from 
the figure that the value of o is 

0°14'. Substituting these values Fia-l? —Summitbetween Two DepreesionBotTrack Profile. 

in the expressions for x and y, there results for the greatest height of ascent 0.002 
feet, or 0.024 inches, and for the horizontal range, 1.97 feet. 




,y Google 



STEEL RAITfl 
SPR1N6,H0.I 




Fio. 18. — Locomotive Driving Wheel SjviDgB. 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 



49 



Fig. 17 shows that this curve practically coincides with the profile of the 
rail. It is hardly conceivable, therefore, that the wheel can leave the rail when 
passing from one depression to another, as the action of the springs, as well 
as the re^ience of the rail, which would tend to prevent this, are neglected in 
the preceding discussion. 

7. Effect op Rocking of the Engine 

The pressure caused by the rocking of the engine on its springs can best 
be determined by observing the amount the springs deflect under their load. 

By referring to Plates XX and XXI, it will be seen that the wear of the 
guides of the driving boxes will give a means of telling how much the springs 
deflect. The maximum amount of wear 
13 probably about one inch. Turning to 
Figs. 18 and 19, which show the springs 
used for the locomotive drivers, we see 
that the depression of the spring one inch 
corresponds to a range of pressure of 
about 8000 pounds. However, as the 
rocking of the en^^ne causes at times a 
less pressure as well as a greater on the 
springs, one-half of this amount, or 4000 ^''- i9-i>«fl«=ti«'«fL«>«»<'«ve Springs, 
pounds, should be taken as the pressure which will cause the spring to deflect 
an amount equal to that obtained under service conditions. ." 

A careful series of experiments have been made by Messrs. Goes and 
Howard * to determine the live load on locomotive driving springs under actual 
running conditions. 

The apparatus consists of three distinct parts; (1) a recording device, 
which fits on the spring band or saddle; (2) a spanner bar or beam, which is 
fastened to each end of the spring link hangers and is connected to the record- 
ing apparatus ; (3) a battery box, which is in the cab with rheostat, switches, keys, 
clock, and all the necessary controlling mechanism. See Figs. 20, 21, and 22. 

The recording apparatus (F^. 20) is in a box, which is bolted to a steel 
plate (1) by four bolts; this plate in turn is bolted to a U-shaped band (2) which 
is fastened to the spring band by four hardened steel set-screws. The record 
is made on metallic-faced paper, 4 inches wide and about 750 feet long. This 
paper is wound up upon a detachable drum (3) and travels across a curved brass 
guide plate (4) under two guide rolls (5) on to the main drum (6). 

* Thesia 1906 at the Massachusetts Institute of Technology, under the supervision of Prof. Lansa. 







/ 


^ 




s 

i 


/ 






i 


/y 









>, Google 



STEEL RAILS 



Tig, 20. — Recording Device and Cab Controlling Mechanism for Testing Drivii^ 
Wheel Springs. (Coea and Howard.) 



Fia. 21. — Recording Device in Place on Driving Wheel Spring. (Goes and Howard.) 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 51 

The main drum is driven by a motor (m) behind the curved plate, throu^ 
a worm and wheel drive (w). The main stylus is on a steel bar (6) machined 
to fit two steel boxes (7) and is free to slide up and down. Considerable trouble 



Fto. 22. — General AmmgemeDt or Apparatus for Testing Driving Wheel Spring. 
(Goes and Howard.) 

was encountered with the stylus on account of the excessive vibration and 
jarring, and finally the type shown in Fig. 23 was designed, which gave entire 
satisfaction, and with which the whole apparatus was equipped. 

This is so constructed as to make the stylus spring always work in tension, 
which is better than using the spring in compression. The spring is suffi- 
ciently long to be sensitive and -still not be thrown from the plate when the 
engine strikes a curve, a trouble characteristic of all former instruments. Be- 
sides the main stylus there are three others 

of identical construction. First, the zero /^^^^S™^""*^' styuus 

stylus (8), which draws a straight line across ' ^ ^^ ' 'j^ ^ ^^ 
the roll and to which all deflections are 
referred; second a stylus (9) which is on a 
magnet that is operated by a Morse key in Fia.23.— MainStyiususedinDrivingWheei 

J.I. 1. .Li.' J i. 1 /irt^ u- L • SpringTeste. (Cora and Howard.) 

the cab; third, a stylus {10) which is on a 

magnet and is operated automatically by a clock in the cab. 

The spanner bar (11) is shown in Fig. 21 and needs no description except 
its method of fastening to the spring link hangers and its mode of operation. 
It is fastened to the hangers by means of two blocks, which are slotted and fit 
over the ends of the hangers, these blocks being held on to the hangers by four 




.yGOOgl'^ 



52 STEEL RAII£ 

hardened steel set-screws. The spanner bar (11) is connected to the stylus 
bar (6) by means of a short link (Z). Thus, whatever relative movement is 
given the recording apparatus by the spring is transmitted as a vertical line on 
the paper by means of the stylus bar (b) and the spanner bar (11). Hence, 
since the paper is being driven horizontally by the motor we have a wavy line 
^ving a complete record of every movement made by the spring, and by means 
of the records made by the key and the times recorded by the clock we can 
account for most of the deflections due to frogs, switches, curves, crossings, 
brake applications, and bridges. 

The cab apparatus (Fig. 20) consists of a suitable box containing a portable 
storage battery and six dry batteries. The storage battery pves 5 amperes 
for 8 hours at a pressure of 6.6 bolts. This runs the motor. The six dry 
batteries operate the clock and the key. On top of the box is a key (K), which 
is connected by means of flexible lamp cord, fastened to the running board, to 
the magnetic stylus (9). By means of the key the operator can record by code 
any observation that may be necessary in working up the records. On tiie side 
of the box is fastened a clock, which automatically records 15-second intervals 
on the paper by a magnetic stylus (10). The motor is kept at the proper speed 
by a rheostat fastened to the top of the box. 

The first successful run was made on engine 1064, consolidated type 2-8-0, 
with 36-inch springs, 17 leaves, 4 full-length leaves 4 inches by f inch. This run 
was made on tiie Fitchburg Division of the Boston and Maine Railroad from 
Boston to Ayer Junction on February 17, 1906. The spring tested was the 
second (counting from the cylinder) on the left side. From this run was 
obtained a maximum deflection of 0.34 inch. (See Plate XIX.) 

A second run was made March 3, 1906, over the same route and on the 
same engine to see if the same deflections were obtained. The curves obtained 
by this latter run were practically identical with the test of February 17, 1906. 
(See Figs. A and B, Plate XIX.) Figs. C D, E, and F present curves taken 
at other points of the track. 

The spring from engine 1064 was taken out and sent to the Engineering 
Laboratory of the Institute and tested on the 100,000-pound Olsen Machine. 
The results of this test are plotted and shown on Big. 24. Two tests 
were made, one with rollers under the knife-edges and one without. The set 
had been measured on the engine and the ends of the leaves had also been 
marked. The spring was then placed in the testing machine and the loads 
applied, corresponding micrometer readings of the deflections being taken until 
the spring had been loaded down to the set as measured on the engine. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



DEFLeCTlON AS OBTAINED 
FROM RECORDING MACHINE 
MINIMUM MAXIMUM 




DEFLECTION IN INCHES 
Diagnun-Looomotive Driving Wheel Springs. (Coee and Howud.) 



■yGoogle 



54 STEEL RAILS 

This load was 14,200 pounds, or the static load on the enpne, the dead 
load for which the spring was designed being 14,144 pounds. Next the maxi- 
mum deflection, as recorded by the spring apparatus, which was 0.34 inch, 
was put in the spring, the load corresponding being 3500 pounds gradually 
applied, making a total load on the spring of 17,700 pounds. 

The excess load appUed to the engine when running should be classed as 
a suddenly applied load. If we consider the relation between the load slowly 
applied to the spring in the testing machine and that suddenly applied when 
tile engine is in service, we see that in the first case the load gradually increases 
from 14,200 pounds up to 17,700. The average load acting through 0.34 inch 
is only 15,950 pounds and the total work done on the spring amounts to: 
15,950 X 0.34 = 5423 inch-pounds. 

In the case of the suddenly apphed load under service conditions it wiU 
be observed that owing to the load reaching nearly its full intensity before the 
spring deflects, the load producing the deflection in this case would be obtained 
by dividing 5423 inch-pounds by the deflection,* 

or ^ = 15,950 pounds. 

This amounts to 12.3 per cent more than the static load of the engine 
on the spring (14,200 pounds) and represents the dynamic augment of the 
spring-bome weight of the locomotive. 

It will be noticed that the dynamic augment as determined by these experi- 
ments is considerable in excess of the figure arrived at by observing the wear of 
the guides of the driving boxes. In the latter case the average deflection corre- 
sponded to a load in the testing machine of 4000 pounds. Taking half of this or 
2000 pounds as the dynamic load producing the same deflection of the spring, 
we find a dynamic augment to the 25,000-pound static wheel load used of 
08 per cent. This may very probably be accounted for by the fact that the 
maximum deflection obtained is so infrequent as to cause no perceptible wear. 

8. Effect of Flat Spots in the Wheels 
We have now taken account of all the forces exerted by the wheel on the rail 

except the impact caused by the existence of flat spots in the tread of the wheel. 
The violence of the blow upon the track, delivered at every rotation of 

the flat wheel, is a matter of common observation ; but the amount of its force 

and the damage done thereby are very hard to determine. 

* See Applied Mechanics, Gaetano Lauia, 1S95, page 246. 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 



55 



A theoretical discussion of the force of the impact is not likely to lead to 
any practical results because of the indefiniteness of the shape of the spot, due 
to the rounding of the comers by wear, also to the lack of knowledge of the 
effect of the springs and the resilience of the track. 

The kinetic energy of the impact is represented by the expression i Mt?, 
where Af represents tiie mass and r the velocity. In order to show a loss of 
energy there must be a change of velocity, but any perceptible change in the 
horizontal velocity of a moving car, due to the impact of the flat spot, is quite 
inconceivable. There may, however, be a change in the vertical velocity of 
the load as the flat spot comes over tiie rail. 

Professor Hancock, of Purdue University, has made a very careful study 
of the mathematical relations existing between the speed, impact, and length 
of spot.* 

Following Professor Hancock's analyas, let 
A, in Pig. 25, be the caiter of a car wheel D 
inches in diameter, revolving as shown by the 
arrow, and CP be a flat spot L inches long just 
beginning its contact with the rsdl. The whole 
wheel is turning about the point C, and will so ' 
turn until P reaches R and the blow is struck on 
the rail. At this lattra* instant A will have reached 
A' and will be moving downward with a velocity 
represented by the line he. If the velocity of A', 
which is practically the same as that of the train, 
is assimied as d feet per second, then 




(k: = usinfl = if 



CB' 



If we regard the mass of the wheel and its load as concentrated at A and call 
the total weight W pounds, the kinetic energy of the mass just before the rail 
is struck will be: 

This formula will give for the energy of impact of a flat spot 2.5 inches long 
in a wheel 33 inches in diameter, carrying a load of 20,000 pounds when the 

• Paper read before the Indians Eogineerbg Society, January, 1908. See also diacussioD by 
L. S. Spilabury, presented by H. H. Vaughan in the American Engineei aad Railroad Journal, December, 



>, Google 



56 STEEL RAILS 

train is traveling 60 miles per hour, 13,800 foot-pounds. At this speed it would 
seem, however, that the results obtained by the formula would be open to 
question. In the derivation of the formula it is assumed that the wheel turns 
about C until P reaches R. This assumption only holds true for speeds from 
zero up to about five miles per hour; * at speeds greater than five miles per hour 
the point C will tend to leave the rail, and the whole wheel will revolve for an 
instant entirely clear of the rail. 

The above discussion neglects the effect of the springs, which will be to 
increase the acceleration caused by gravity, and the resilience of the rail, which 
will cause it to rise to meet the flat spot. 

It is very questionable whether, on account of the very small time interval 
required for the wheel to pass the length of the flat spot,t there is an appreciable 
increase in the stress in the rail, except at the point of contact of the wheel with 
the rail. 

To increase the load on the rail a change in the vertical velocity of the 
load must be made; but at high speeds, when the effect of the flat spot is most 
detrimental, the time required to go the length of the flat spot is so small that 
the acceleration of the wheel and its load, even when augmented by the action 
of the springs, is so small as to be negligible. The real danger seems to lie 
in the metal of the running surface of the head of the rail; the metal here is 
under a high state of compression (see Figs. 146 and 147), which is momentarily 
relieved by tiie passage of the flat spot and then applied suddenly. 

When the flat spot is long enough so that the surface of the flat spot is 
brought in contact with the rail, a sensible change in the vertical movement 
of the load results and the load on the rail is increased. This is well shown by 
the following example given by Jlr. L. R. Clausen,J of the Chicago, Milwaukee 
& St. Paul Railway; 

"Some time in the year 1900 we had an engine with a flat spot on rear 
right-hand driver 32 inches long and aV inch deep, which broke about 27 rails 
during one week's time (85-pound rails, not to exceed one or two years old) ". 

This flat spot was not apparent to the eye and was only detected by cen- 
tering the wheel and then measuring around it with a gauge. 

■ E. E. Stetson, Rdlroad Age Gazette, December 4, 1908. 

t The present allowable length of flat spots in car wheels is 2) in. This rule was adopted by the 
Master Car Builders' AaBocisition in 187S. In 1909 the question of reducing the limit for freight wheels 
to less than 2i inches was considered by committees of the Master Car Builders' Aaaociation and the 
Ainerjcan Railway Eogineering Association, but it was not then considered advisable to make any 
change in the rule. 

X Proceedings Am. Ry. Eng. & M. of W. Asan., 1B09, Vol. 10, Part 2, p. IISS. R«portonFlat 
Spots OD Car Wheela. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



57 



In the Railway Age Gazette of March 16, 1910, was reported 200 85-pound 
rails broken in 14 miles by a fiat spot which had grown to a length of 6 inches, and 
a maximum depth of f inch. In the extreme cold weather experienced in the months 
of January and February, 1912, many tires fmled by shelling out, and the following 
examples, taken from the same authority, are representative of the conditions ex- 
isting on lines in the Northern parts of the countiy during this period. 



January, 1912. 
January 7, 1912. 
January l^i 1^12. 



January 20, 1912. 
January 24, 1912. 
February, 1912. 



Minnesota. 
Savanna, III. 
South Dakota. 



New York State. 
New York State. 
Ohio. 






Flat spot 4 

steel tire i 
Flat spot 5i 

wheel inp ....... .. 

Two steel wheels with Hat spots, 500 rails. 

on different trucks of a dining 



IS. long I 
passenge 

la. long 01 



9 80-lb. rails in 3 miles. 



steel-tired 



160 rails. 



Flat wheel on a fast train. 
Flat wheel on an observation ca 
Shelled-oiit steel-tired wheel; 

end of run the flat spot was 9 ti 

long. 
Flat spot on steel-tired wheel u 

der a baggage car. 



Nearly 100 rails. 

15 rails. 

960 rails in 200 miles. 



50 rails in 70 miles. 



It is generally known by those familiar with the manufacture and use of 
chilled car wheels that only a very small percentage of them are evenly chilled. 
This, apart from weakening the wheel, also produces a lack of roundness tend- 
ing to cause poimding on the rail. The following information upon tests on 
the roundness of tread of chilled car wheels has been furnished by Mr. S. K. 
Dickerson, Assistant Superintendent of Motive Power, and Mr. H. E. Smith, 
Engineer of Tests, of the Lake Shore and Michigan Southern Railway Company.* 

To make these tests six pairs of wheels cast by different founders were 
selected. An axle with a wheel pressed on each end was placed in a lathe and 
the centers were firmly pressed. The wheels were then hand-turned. This 
done, the tread was divided into eight sections, each the same distance from 
the flat edge, and a specially constructed micrometer used to discover any 
variations in the roundness. All the testing was done with great care and 
predsion. 

The tests are illustrated in Fig. 26. The dotted line in each diagram is a 
circle through that point on the tread having the smallest radius, and is assumed 
as the datum line. In plotting the diagram the variations from this datum 
line have been multipKed by five in order to emphasize the irregularity of the 

• Proceedings Am. Soc. for Teat. Materials, 1910, Vol. X, p. 307. Unevenly Chilled and Untrue 
Car Wheels by Thomas D. Wert. 



,y Google 



STEEL RAII£ 





Fig. 26. —Irregularity in the Roundness of Present-day ChUled Cm Wheels. (The Iron Age.) 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 






Fia. 26. — Continued. 



,y Google 



60 STEEL RAIIS 

tread. It is to be understood, however, that the figures given are the actual 
variations in the radii of the wheels from the datum circle. 

Owing to the present imperfect state of oiu- knowledge on this subject it 
would seem desirable to determine experimentally the exact effect of the blow 
delivered by a flat wheel on the rail. 

Professor Benjamin* has designed an apparatus for such tests, which is shown 
in Fig. 27. The apparatus shown in the figure will permit of the continuous 



L.OAD 




Fio.27. — Apparatus for Measuring the Effect of aFlatSpot. (BenjamiD.) 

Operation of one wheel upon one section of rail indefinitely and permit at the 
same time measurements of the effects of the blow. The truck is so supported 
that one wheel turns freely upon an idle pulley, while the other wheel on the 
same axle rests on a section of steel r^l and in turning drives the latter by 
friction. The section of rail is bent to a circle, lying in a horizontal plane, and 
is firmly riveted and bolted to a supporting web, which is then fastened to a 
central hub of cast iron or steel. This hub turns freely on a vertical mandrel 
and is supported by a thrust bearing underneath. The rail and its attachments 
thus turn in a horizontal plane under the rotating car wheel. The portion of 
the rail immediately under the wheel is supported by friction rollers, which turn 

* Pt^Kr presented at Meeting of Western Railway Club, November 17, 1908. See also dis- 
cussion of Professor Benjamin's paper by H. H. Vaughan, American Engineer and Railroad Journal, 
December, 1006, and a further article by Professor Benjamin in the Railway Age Qaxette, June 28, 
1912, p. 1613. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



61 



freely in a steel box or yoke. This latter forma a portion of the main casting 
supporting the hub of the rail, and this casting is bolted to a wooden pier so 
as to have a certain amount of elastidty. On the lower side of this casting and 
directly beneath the point of contact between the wheel and the raU is a hard- 
ened steel hammer, or ball, resting on a strip of soft metal. The soft metal is 
supported on a heavy anvil of cast iron and is fed slowly breath the hammer 
by friction rollers. 

The truck being loaded with the dedred amount of pig iron or other material, 
the wheels and their axles are rotated by means of a variable speed motor, and 
the energy of a blow delivered by a flat spot on the wheel is measured by the 
indentations of the strip of soft metal underneath the hammer. The amount 
of energy due to any ^ven indentation can be readily measured by produdng a 
amilar indentation under a drop press. The curving of the rail in a horizontal 
direction is not sufficient to interfere with the action of the wheel and the energy 
of the blow is transmitted directly to the soft metal. 

A subcommittee of the American Railway Engineering Association have 
made an attempt to measure 
the force of the blow caused 
by flat wheels under working 
conditions. 

For this pmi)ose an 
80,000-poxmd capacity car 
was equipped with regis- 
tering devices to measure 
the compression of the car 
springs and a pair of wheels 
with flat spots was placed 
in one of the trucks. The 
poation of the flat wheels, 
springs, and the recording 
device is shown in Fig. 28. 
The recording device con- 
sisted of an apparatus for 
measuring the maximum 
deflection of the springs. 
The springs were calibrated, 
and it was found that a load of 32,500 pounds applied to a nest of four springs 
produced a compreaaon of one inch. 




POSITIONS OF RECOBD1NG DEVICES 



1 


















£ 






















































jft 


s — ' 


' i 


i — 




; 


1 


; — 


i 



Pia . 2S. — Diagram of Testa on Freight Car with Flat 
Wheele. (Am. Ry. Eng. Asm.) 



>, Google 



ffil STEEL BAILS 

The car was loaded with splice bars to its full capacity, the l(»d h&ng 
imiformly distributed. The train was then run for a short distance, brought 
to a stop, and the maximum deflection of the springs noted. Several different 
tests were made in this manner at different rates of speed, the results of which 
are shown by the diagram in Fig. 28. 

The diagram shows quite uniformly for all of the tests a greater deflection 
of about a sixteenth of an inch for the trucks with flat wheels, corresponding 
to an increase in pressure on these trucks of from 1000 to 2000 pounds. The 
results would appearto indicate, then, that the flat wheels, either by increasing 
the oscillation of the car or for other reasons, cause an increase of pressure on 
the track. 

On account of the small number of tests made and the fact that they were 
confined to one end of a car the results should not be regarded as conclusive. 
However, assuming that the effect of the flat spot in the wheel is to cause addi- 
tional rocking of the vehicle, as the experiments would appear to indicate, it 
will be noted from article 7 that this force is already taken account of in the 
consideration of the excess pressure caused by the rocking of the locomotive 
on its springs. 

9. Impact Tests 

Professor Goss * experimented to determine the effect of the counter- 
balance pressure with the Purdue Test Locomotive. This engine is mounted 
with its drivers resting upon wheels of approximately the same diameter with 
the drivers, and when the drivers are turned by the engine the supporting wheels 
roll in contact with them, and the engine as a whole remains stationary. The 
en^ne was in complete horizontal balance and was counterbalanced heavier 
than it would be in ordinary road service, the main wheel being about 0.4 per 
cent and the rear wheel about 54.0 per cent more heavily balanced than is 
the usual practice. 

A common annealed iron wire 0.037 inch in diameter was used and run 
under the drivers. Fig. 29 shows the effect of the drivers on the wire. 

Wire I shows slight variation. 

Wire II shows a jump of the wheel just after the counterbalance Irft the 
highest point, the lifting being retarded, probably due to inertia of the mass 
to be lifted. 

• An Experimental Study of the Effect of the Count«rbal&DCe in Locomotive Drive-wheeb 
upon the Preeaure between Wheel and Rail. — Gobs. Tnms. Am. Soc. of Meeh. Engra., Vol. XVI, 
1895, p. 305. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



63 



Wires III, IV, and V show the more marked lifting effects, due to mcreased 
speeds. 

A light nick from a sharp chisel was made across the face of the wheel to 
serve as a rrference mark, which left a clean-cut projection upon the wire. It 
was found at high speeds that the single nick across the face of the wheel leaves 
two projections upon the wire, usually about one-eighth of an inch apart. The 
contact between the wheel and the track would evidently appear, therefore, 
not to be continuous, but a succession of exceedingly rapid impacts. Professor 
Goss derived the following conclusions from his experiments; 




Fia. 29. — Wire Testa — Professor Goss. 

(a) When a wheel is lifted through the action of its counterbalance, its 
rise is comparatively slow and its descent rapid. The maximum lift occurs 
after the counterbalance has passed its highest point. 

(6) The rocking of the engine on its springs may assist or opi>ose the action 
of the counterbalance in lifting the wheel. It therefore constitutes a serious 
obstacle in the way of any study of the precise movement of the wheel. 

(c) The contact of the moving wheel with the track is not continuous, even 
for those portions of the revolution whai the pressure is greatest, but a rapid 
succession of impacts. 



,y Google 



«4 



STEEL RAJI5 



The question of impact has received a great deal of attention from bridge 
engineers. Recent work in this direction consists of a large number of experi- 
ments that were made on the Baltimore and Ohio Railroad by a subcommittee 
of the American Railway Endearing Association to determine the amomit of 
impact in bridge members. The records of these tests were, unfortunately, 
burned in the Baltimore fire before they were put in permanent form. The 
experiments were continued by the Association under the direction of Professor 
Tumeaure. 

The importance of this subject was widely realized, and most of the large 

railroads in the country con- 
tributed towards a fund which 
enabled the work to be carried 
on with a great degree of 
thoroughness. 

Professor Tumeaure's ap- 
paratus consisted of an auto- 
graphic extensometer for 
recording the deformation of 
bridge members and multi- 
plying that deformation by 
some factor like eighty or 
ninety and recording it on a 
moving strip of paper. On 
Fig. 30 are shown some of the 
records taken by one of these 
machines in 1906. It will be 
seen from an examination of 



y^^^''■^*'^v-vVV-" 




Fia. 30. — Defoimation of Bridge Membere tmder Paaaing 

Troine. (Am. Ry, Eng. AesQ.) 



the records that the effect of the different wheels can be readily traced in 
the diagrams. 

During the season of 1907 * further work was done by the committee. 
The instruments used consisted of a deflectometer and eight extensometa^. 
The deflectometer is the instrument described in Transactions A. S. C. E., 
Vol. XLI, June, 1899, p. 411. (Some Experiments on Bridges under moving 
Train-loads.) The instrument is itself attached to tiie bridge, while the con- 
nection with the groimd below is made by means of a wire attached to a 
heavy weight resting upon the ground. The deflection is multiplied by two and 



* Proceedings Am. Ry. Eng. and M. of W. Asen., 
A Structures. Impact Tests. 



Report of Committee on Iron and 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 



66 



recorded on a moving strip of paper. The extensometers are clamped to various 
members of a structure and record the extenaona or compressions of the mem- 
bers over a length of about four feet The ratio of multiplication of the exten- 
someters is about 50. 

Test trains were made up of a selected type of locomotive, followed genwally 
by a sufficient number of loaded cars to cover tiie span. Longer trains were 
not used, for the reason that it was desired to secure speeds as high as possible, 
and also because many observations under the freight tr^s of the reguW 
traffic showed that at the speeds practically attmnable the impact effect was 
much less than with the shorter test train. 

In carrying out the tests the train was headed in the more favorable direc- 
tion for speed, and was moved back and forth over the structure at various 
rates of speed. Such speeds were selected as fdrly to cover the range from 
about 20 miles per hour to the maximum attainable. A few movements were 
made at from 10 to 20 miles per hour. Little difference was noted in the results 
at various speeds below 15 miles per hour, and in general the results at 10 miles 
per hour may be considered as practically equal to static stresses. 

The speed of the tran was determined by the use of stop watches and 
signals by observers stationed at the ends of a 500-foot base line. The loco- 
motive was generally working when crossing the span, but in some cases was 
not. Differences in this respect caused no noticeable differences in results, 
so far as the field observers were able to judge, although this point was con- 
adered mainly with respect to the higher speeds. Each test was given a serial 
number, and all records obtained for that test were given tiie same number. 

TABLE XII. -CALIBRATION TABLE, IMPACT TESTS ON BRIEM3ES. 



■n™.,.,.' 


Unlti^i^tio.. 


V>h» oil inch OrdiDBU, 


L^^l^i. 


Deflection 
1 
2 
3 


2 
55.4 






11,300 


0,fti2 


48.4 
63.7 
61.7 
54,2 
64,5 
64.3 


12,900 
11,600 
12,100 
11,500 
11,500 
11,600 


0.547 
0.613 
0.623 
0-703 
0.683 
0.602 



Table XII contains calibration data of the instruments and the approximate 
speed of movement of the record paper. Table XIII contains detaled data of 
one (rf the tests. This sheet represents tiie work of one day on one structure. 



,y Google 



STEEL RAILS 



^ 



s I 



mt 



i 
ri 



SX5S55S3SSSSSSSS33SSgi;S3£RSSSSSS3S^SS 



SXS3SSSS3SSS££$SiSS;;iSiESK33SSS33SSSSSS 



£8 SSSSS S*SS£mmSS2 



S, SMSBSSS SSS^SStSSSBaSStiUSnSSUSSStiSS 



:-a3SSSS2 : : SSSSaaS. 



t;?S$S! :S3S3S3SiSS«S3SS3S3SSS?S : '3399S : 



$s;:ssseRiSSK£SS3;:SSssSi3SSS3rS;:n;:s3sss 






;S:i3Si3sS!£l3^S£S^^i5S£iS^3Sa333S333S5S3S 



SS :SSSSSSS : 



3 ^ss^sisessssssssersf^KSSESs : issess^scsss 



;S!3!3SKsa;;s»s»S8ss;si;ssssassss!i3Ss;i;iSSSS$ 






S3§3§SS?2§§S^?s52HS§SSSB§3'S2I3S^r"235 



"- SS88S»SaSSSSS5:R3aS5SSSS5£ = 2=aSSE3SSS 



I g..,.g., 



,„g„„„,g-.,.|- g-, 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 67 

and with a particular locomotive and train. At the top of the dieet are given 
the weights of test train, and a diagram of the bridge, showing the general 
location of the instruments by number. The table then gives, first, the number 
of the test, then the speed in miles per hour, then the position of the counter- 
balance, as will be described more fully later. Then follow the data relating 
to the deflection and measurements of the various extensometers from Nos. 1 
to 8. The location of each instrument is given immediately below the number. 

In the column headed "Maximum" is given generally the value of the 
marimimi ordinate of the diagram resulting from the test, expressed in hun- 
dredths of inches. An exception to this is where there was obviously considerable 
instrumental vibration, in which case an effort was made in scaling the diagrams 
to eliminate this instrumental ^ect. This could be done in many cases quite 
satisfactorily, but not in other cases, and all such records are open to more or 
less doubt. 

In the column marked "Amplitude" is pven the amplitude of the vibra- 
tions in the diagrams where such vibration is apparently due to the structure 
and is not instrumental vibration. Vibrations may be due in general to three 
causes: (a) vibrations of the structure as a whole, as shown most clearly in the 
deflection diagrams and the chord or flange stresses; (&) vibrations of individual 
members, especially eyebars, and (c) instrumental vibrations. 

Other marked variations in the diagrams occur in such members as stringers, 
floor beams, and hip verticals. It can only be stated at present that it is generally 
possible to distinguish instrumental vibrations from others because of their 
much greater rapidity. It may be said, however, that in most of the digrams 
obtained the instrumental vibrations are not serious. 

In the column marked "Peak" are given the measured ordinates to the 
highest points of the curves, including instrumental vibration. The excess 
of this value over the "Maximum" shows the extent of the instrumental vibra- 
tion as estimated. In many cases this is small, but in many cases also it is 
large. In some cases it is so excesave that no attempt has been made to measure 
the records. The deflectometer gave no trouble in this respect. 

In the column headed "Remarks" are noted various remarks by the use of 
letters: "I" signifies instrumental vibration. 

Returning to column three, headed "Counterbalance": In the later tests 
the portion of the counterbalance was determined with reference to some panel 
point of the bridge. This was done by inserting an index made of }-inch steel 
into the rim of one of the drivers, exactly opposite the counterbalance. Then, 
alongmde the rail was placed a 2- by 4-inch strip, on which was placed a ridge 



,y Google 



68 STEEL RAILS 

of clay or putty at such a height as to be indented by the index as it passed 
along. The posiUon of the indentation is noted in feet north or south of the point 
of reference, which point of reference is shown on the truss diagram by the 
letter "C." 

During the seasons of 1908 and 1909 the experiments were continued, 
and very complete data gathered of a mmiber of bridges under different con- 
ditions of loading.* 

The experiments obtained in this series of tests indicate that with track 
and rolling stock in good condition the main cause of impact Is the unbalanced 
condition of the drivers of the ordinary locomotives. The great importance of 
unbalanced drivers is well brought out by a comparison of the results of com- 
parative tests with the ordinary locomotive and the balanced compound and 
electric locomotive. The impact caused by balanced compound and electric 
locomotives was very small. 

While it is interesting to study the effect of the impact in the different 
members of a bridge in a consideration of the stresses in the rail, it is doubtful 
whether the two cases are sufficiently alike to afford anything more than a 
general comparison. 

In considering the impact stresses in the rail the effect of the inertia of the 
track must be borne in mind. 

If we examine what takes place in the track under the impact of the wheel 
it is seen that there is a force F acting between the wheel and the rail. When 
impact occurs this force is increased by a force F' which produces a change in 
the relative velocities of the wheel and the track, but on account of the nature 
of the track the change in its velocity is almost imposdble to determine. 

The average value of this force F' is exactly equal to the change in momen- 
tum produced by its action divided by the time required to produce this change 

or f'.™(^), 

where m = mass, 

V = velocity, 
t = time. 

It will be observed that the interval of time ( - (' during which the impact 
acts is very small, and is not sufficient to allow the force F' to depress the 
track at high speeds, with the result that the force F' is overcome mainly by 
the resistance of the rail to compresaon. The heavier and harder the rail, there- 

• ProceedmgH Am. Ry. Eng. and M. of W. Aaan., 1011, Vol. 12, Part 3, p. 12. Report of Sub- 
■XHtunittee on Impact. 



>, Google 



PRESSURE OF THE WHEEL ON THE RAIL 69 

fore, tiie greater is this force F', other things being equal. Indeed it is quite 
pos^ble that it may at times exceed the crushing strength of the r^ on its upper 
surface. 

A diviaon of the forces acting on the wheel into those which produce local 
str^dns in the running surface of the head and the forces which tend to set up bending 
stresses in the rail is difficult to determine; probably the effect in producing bend- 
ing stresses of all the forces except the static load on the wheel and the centrifugal 
force described in Article 6 are condda^bly lessened owing to the inertia of the 
track. 

This especially applies to the pressure exerted by the springs of the loco- 
motive when the engine is rocking. If the full effect of the compression of the 
springs was realized the pressure might well be twice what has been taken. 

This was illustrated in the impact tests on bridges when the excess pressure, 
caused by the counterbalance, tended to produce well-defined strains, and at or 
near the critical speeds would set up vibrations in the bridge itself, while the 
impact from the rocking of the locomotive on its springs apparently caused a 
much less serious effect. 

10. The Dynamic Augment of the Wheel Load 
While we have examined the various causes of the increase in wheel pressure 
when the locomotive is running, it is still necessary to consider the effect of the 
velocity of the wheel load. 

Assuming the track perfectly smooth, the wheel without imperfections and 
all of the rotating parts perfectly balanced, the effect of a load moving over the 
track at a high rate of speed depends wholly upon the vertical ciurature of 
the track and the effect which this curvature has upon the path over which 
the center of gravity of the load travels. 

In the case of a bridge, if we assume the track orig^ally straight and 
absolutely ri^d, the amount of impact or centrifugal force resulting from the 
deflection of the structure can be approximately determined on theoretical 
grounds. Such an analyms has been made by Dr. H. Zimmermann for the 
case of a single rolling load, and a formula which is very closely approximate 
to his exact formula is as follows: 

f _P_J , 

gi' _o 

in which F = centrifugal force, P = weight of rolling load, e = velocity in feet 
per second, d = deflection of structure, and I = span length. If, for example, 



,y Google 



70 



STEEL RAILS 



d = jiinr of span length and u = 90 feet per second (about 60 mUes per hour), 

we have 

F = P , 

0.595i-3 

a fonnula which is practically exact for spans greater than 15 feet. For a 

25-foot span this gives 8.7 per cent impact, and for a 50-foot span 3.7 per cent. 

For a 100-foot span the value would be 1.7 per cent. 

With the yielding supports under the rail the center of gravity of the load 
tends to travel along a straight line (assuming the track to be perfectly uniform). 
It seems therefore probable tiiat the dynamic stress in the r^l is not increased by 
any appreciable amount by the velocity of the wheel load, and the maximum 
pressure to be used in calculating the stress in the rail can be taken to be made 
up of the static wheel load, the excess pressure due to the counterbalance and 
angularity of the m^ rod, the pressure due to the wheel passing over irregu- 
larities in the surface of the track, and the pressure caused by the rocking of 
the engine on its springs. 

The imposed pressure due to the angularity of the main rod and the e 
balance is given in Table XIV. 

TABLE.XIV.— MAXIMUM DYNAMIC PRESSURE, IN POUNDS, DUE TO 

ANGULARITY OF MAIN ROD AND EXCESS BALANCE, 

FOR SPEEDS UP TO 60 MILES PER HOUR 



Typ., 


W«i(hlo«A].l», 


Eic^ Prenure Dua to Auularity of 

MsiB Rod iuhI Exwn GhbUt- 

baluce. (One Sida Only.) 


S. 


Front or B«k 
Driv.™. 


Driver. 


^^C"" 


4-4-2 
4-«-2 
4-B-O 
2-6-0 
2-8-0 


5S,000 
60,000 

65,000 
64.000 
54,000 


52.000 
56,000 
52.000 
53,000 
60,000 


12.000 
10,000 
11,000 

15,000 
14,000 


0,000 

7,000 
10,000 
11,000 

9,000" 



This pressure is not a direct function of the wheel load and should be ex- 
pressed in pounds for each class of locomotive. 

The excess pressure for' the freight locomotive, types 2-6-0 and 2-8-0, is 
noticeably greater than is that of the passenger locomotive for the high speeds 
^ven in Table XIV. Inasmuch as the engines in freight service are not called 
upon for as high speed as in the passenger service, the excess pressure at 40 miles 
per hour may be taken as the maximum for the engines of this class. The 
greatest pressure occurs with the Mogul locomotive, type 2-6-0, and is 10,000 
pounds for tJie main driver and 5000 pounds for the front and rear drivers. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



71 



The extra pressure due to irregularities in the track is dependent on the 
weight on the wheel and may be expressed in per cent of the wheel load, and is 
4000 



30,000 



= 13 per cent of the weight on the driver. 



The pressure due to the rocking of the engine on its springs is likewise a 
function of the wheel load, and is 13 per cent of the load on the driver. 

Dr. P. H. Dudley has observed that the drawbar pull and the tender on 
stiff rails can become important factors in distributing the df ect of the expended 
tractive power to a longer portion of the track than that occupied by the driving 
wheel base. 

The distribution of the expended tractive dfort through the drawbar 
pull may be extended in 6-inch, 100-pound rails to all of the wheels under the 
locomotive. In the 80-pound rail this reduces to about two-thirds of the length 
of the total wheel base of the locomotive. 

Two stremmatograph tests were made for the purp(ffie of tracing the dis- 
tribution of the stresses due to the expended tractive power of the locomotive 
when drawing its train. The first test was made with the engine light, backing 
over the stremmatograph, and then in a few moments coming forward. The 
second test was made with the same engine drawing its train. It indicated 
that the unit fiber stresses are increased under all of the wheels of the loco- 
motive, and also the bending moments. 

The effect of the expended tractive effort through the drawbar pull and 
the stiff rails was distributed for the entire length of the wheel base of the loco- 
motive, instead of being restricted to that of the driving wheels. The total 
wheel effects under the light locomotive for one rail were 1,368,706 inch-pounds, 
and 1,864,363 inch-pounds when drawing the train, an increase of nearly 500,000 
inch-pounds, or 1,000,000 inch-pounds, practically, for both rails. 

It is evident that this increase in pressure on the rail is allowed for by the 
dynamic augment to the apring-bome load of the locomotive, and we may now 
prepare Table XV, giving the dynamic augment to be added to the static 
wheel load. 

TABLE XVft. — DYNAMIC AUGMENT TO BE ADDED TO THE STATIC 
WHEEL LOAD WHEN THE LATTER IS 25,000 POUNDS OR OVER 

Note. — The dyaamic augment 
as ^ven id pounds in columns 1 or 
2 should be added to the per cent 
of the whed load pven in column 3, 
to ^ve the total dynamic augment 
for the drivers. Column 3 only 
should be used for the truck wbeds. 



Cten. 


DyBUDicAixniM.t. | 


Pouid*. 


P««,t. 


dSS.. 


FnMt ud Back 
DriTBrt. 


1 


1 


I 




12,000 
10,000 


9,000 
5.000 


26 
26 





>, Google 



STEEL RAII5 

TABLE XVb. — DYNAMIC AUGMENT TO BE ADDED TO 
THE STATIC WHEEL LOAD WHEN THE LATTER 
IS LESS THAN 25,000 POUNDS 



Claa. 


Dyumic Aogmat. 


Uain Driven. 


FiwtudBuk 
Driven. 


TnukWba^. 


Passenger 


POTMBt 

75 
66 


Ptrcmt. 

60 
45 


Paro«t 
26 

26 








$ 9 3 

Atlantic Type, 442-190. 



A£) 



tw X " ' .1. "' 

_557t 




Pacific Type, 462-225. 






Prairie Type, 262-234, 

Fia. 31. — Dynamic Wheel Loads of Typical Paaeenger Steam 
Locomotives. 

See Plate XX for static loads and Table XV for relation between 
static and dynamic loading. 

Note. — Total wei^t on drivers assumed to be equally divided 
between the driving axles; if the main wheel is more heavily loaded the 
dynamic pressure will be increased accordin^y. 



By referring to 
Figs. 6 and 7 it is seen 
that the excess pres- 
sure, caused by the 
angularity of the main 
rod and counterbal- 
ance, is less in amount 
for these lighter en- 
gines. Table XVb 
shows the dynamic 
augment of engines 
having wheel loads less 
than 26,000 pounds. 

We may now pro- 
ceed to construct typi- 
cal load diagrams for 
the diflETent classes of 
locomotives. Plates 
XX and XXI give the 
principal dimensions of 
each type of engine 
under discussion, from 
which, with the ^d of 
Table XV, the load 
diagrams given in Fig. 
31 and Fig. 32 have 
been prepared. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



-®i 



e 

€ 



OOOZ? H 




- I Oi 



§4! 

m 



1 I' 
j J 



» > 8 5 3 
•.-■a Se • 

lis S| 



1 I1J|!| 
1^ Ult- 

^ ^ -2 e 5 P 2 

•3 a 2 j|i^ 



4' 



^ ill -I 

-5 i ^ g * 5 
S 'C s o. a ^ 

' ! itiM 

M '-S 5 S O 

S Is-ll 
s ^ a 1 5, 1 

£ '''t.i 

« *f^5 I a 
1 1. Ill 



>, Google 



74 



STEEL ILUL9 



11. Electric Locomotives 

* In dealing with the electric locomotive the question of the excess balance 
necessary to counteract the reciprocating parts can be entirely neglected, and these 
machines, when properly constructed, would appear to have a more favorable 
action on the track than is the case with a steam engine of the same capadty. 

The low center of gravity possessed by the earlier locomotives of this type 
imposed, under certain conditions, a very severe duty on the rail, f In order 
to bring out the facts experimentally, the Pennsylvania Railroad Company, 
who were about to design locomotives for their tunnel entrance into New York 
City, constructed a special test track with apparatus for measuring side pres- 
sures upon the rail; they built sample locomotives of different designs and 
instituted a series of tests of electric and steam locomotives to determine their 
relative riding qualities at speed. 

It was foxmd that all types of locomotives were practically steady at speeds 
under 40 miles per hour, but that above this speed marked differences appeared; 
that the steadiest riding machines were those with high center of gravily and 
with long and unsymmetrical wheel base. In other words, that the nearer 
steam-locomotive design is approached in wheel arrangement, distribution of 
wei^t, height of center of gravity, and ratio of spring-borne to under-spring 
weight, the less the side pressures registered on the rail head. 



TABLE XVL- 



ELECTRIC AND STEAM LOCOMOTIVES, COMPARISON OF WEIGHTS 
AND CENTERS OF GRAVITY. (Gibbs.) 



Tola] weighl oT loconio 
nuuiing ordor [a pounds . . 
H^ht, ceatcr (cniviiy, com; 



HflighU of cmtv of ^rai 



T41T 



: V. Tunnel, 
332,000 



N.Y..N.H.4H. 



I3S,0flO 



Table XVI presents a comparison of weights and centers of gravity of 
modem electric locomotives and steam locomotives. Fig. 33 shows the Detroit 
River Tunnel Company's locomotive. 

• See the Rwiroad Age Gazette, Vol, XLVII, 1909, pp. 271,319, 537, 881, and the RaUway Age 
Gazette, Vol. XLVII, 1910, p, 829, for deacriptions of electric locomotives given in this article, 

t Electric Traction by George Gibbe, report presented before the International Railway Con- 
gresa, July, 1910. See also a very complete article " The Electrification of Railways " by George 
Westinghoiise. Appendix No. 2. Data on Electric Locomotives of American Design, pp. 970-079. 
Trans. Am. Soc. of Mecb. Engrs., 1910, Vol. 32. 



,y Google 



PRESSUEE OP THE WHEEL ON THE RAIL 



>, Google 



76 STEEL RAILS 

Fig. 34 illustrates the type of the Pennsylvania Electric locomotives which 
are used for handling the Pennsylvania Ralroad trdns into the New "York 

station. 

This locomotive incorpo- 
rates many novel features in 
electric-locomotive design, and 
^ is the result of several years' 
■g cooperative development be- 

tween the Pennsylvania Rail- 
^ road Company and the 
I" Westinghouse Electric and 
J Manufacturing Company. It 

is distinctively a high-powered 

1 machine, built for high speed 
^ operation. 

% In wheel arrangement, 

ft weight distribution, trucks and 
^ general character of the run- 
^ ning gear, it is the practical 
J equivalent of two American 
a type locomotives coupled per- 
1 manently back to back. 
I The connecting rods are 

J all rotating links between rota- 
■S ting elements, and are thus 
I perfectly coimterbalanced for 
4 all speeds. 

J The employment of this 

B transmission permits the 
(^ mounting of the motors upon 
J the frame, secures their spring 
a support, and, in common with 
the rest of the locomotive, the 
center of gravity at approxi- 
mately the same height above 
the rails, found de^rable in 
the best high-speed steam 
experience. The same freedom of motion in the wheels and axles that is 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



77 



characteristic of the present steam locomotive is also obviously secured. It 
will be seen from Fig. 35 that the locomotive is an articulated machine and 
that each half carries its own motor and has fom- driving wheels, 68 inches 




,y Google 



re STEEL RAILS 

in diameter, and one four-wheel swing bolster swivd truck with 36-mch wheels. 
In these locomotives the variable pressure of the unbalanced piston of the steam 
locomotive is replaced by the constant torque and constant rotating ^ort of 
the drive wheels, and the pull upon the drawbar is thereby constant and 
uniform. It might to the casual observer appear that by this arrangement of 
driving a return has been made to steam locomotive practice as regards \ 
counterbalancing difficulties, but it will, upon examination, be seen that nothing 
of the kind is true. There are no questions of unbalanced reciprocating weights 
involved, and all weights are revolving ones and directly counterbalanced. 

In Table XVII is given the general characteristics of the electric locomo- 
tives of this country. 

In determining the dynamic augment of the wheel load in the case of the 
electric locomotive, the effect of the counterbalance, which plays such an im- 
portant part in the pressure of the driver of the steam locomotive, can be entirely 
neglected. The other causes remain approximately the same, and by referring 
to Table XV (column 3), it is seen that the dynamic augment amounts to 
26 per cent of the whed load. Fig. 36 gives the load diagrams of electric loco- 
motives based upon this assumption. 

.B.aO.R.Rv1S95 



Ci) (i) CD C D 



) am am a 

N.YN.H.aiH.R.R.1908 



D= Dynamic Load per Wheel, poundB. 
rS=Statio Load per Wheel, pounds. 
D-1.26S. 



'OCD^ 



oooo oo oo oooo 

oooo oo oo oooo 

oooio om cm o^ioo 

too 'HO iDO oo tOOrjiO 

a tn a in a ir ou> oinoc) 

PENNSYLVANIA R:R. 1909 

oo oo oo go oo oo oooo 

CO oo oo oo oo oo oo GO 

mo mo oo oo oo oo mo mQ 

_— — rt (M n fj row in fa — — —J 

Ou> Otf) Qin a in o V) QUI QVlQiO' 



Fig. 36. — Typical Load Diagruns Tot Electric Locomotives. 



>, Google 



PKESSHRE OF THE WHEEL ON THE RAIL 



1, s=sa--a 


iiiij . 






iP| ' 


1 |8 S£H=*HS 




1 1 Ip;^ ;::::::!: ifll = 




IPl » 


II Hi l;f|lljll 

lliilliiiii 


111 1 

m 1 

pill J 



>, Google 



i 



m 



I 



li-i 



I- 



:::::! i m 



II 
II 



f 



¥ 



ft. 



H 



in 



!"■ 



19: 



!=■ 



I 



liiii, 



ip: 



y ' 



3 ft-j:a«"-s 



I 



m 



Mil 



I 



>, Google 



pre^ure of the wheel on the rail 81 

12. Cars 

In dealing with the pressure of the wheels of cars it will be noted that the 
dynamic load is much less than is the case with the heavy steam locomotive. 
Here the only dynamic effect caused by the rapidly moving wheel is that of 
the rocking of the car on its springs, the irregularities of the track and lack of 
roundness of the wheel. 

Fig. 28 gives a convenient means for estimating the excess pressure caused 
by the rocking of the car. It will be noted from the figure that the maximum 
dynamic pressure corresponds to a depression of the nest of springs of about i\ 
inch at a speed of 50 miles per hour, or, as the springs are capable of resisting a 
slowly applied load of 32,500 pounds for a depression of 1 inch, the dynamic 
load at 50 miles per hour for the two wheels supported by the nest of springs 
is probably about 5000 pounds, or 2500 pounds for each wheel greater than the 
static load. 



Fia. 37. — Box Car. Capacity 100,000 lbs. Weight 48,000 lbs. Length 40 ft. 

An allowance of 10,000 pounds per wheel would appear to be ample to 
cover higher speeds obtained and heavier cars than that used in the test to which 
this figure refers, as well as any increase in pressure caused by irregularities in 
equipment. With Ughter loads the dynamic augment may be decreased, and for 
wheel loads less than 15,000 pounds it has been taken as 0.7 of the static load. 

Figs. 37 to 41 * illustrate types of freight cars and Figs. 42 to 46 * passen- 
ger cars. It will be seen that the freight cars use a four-wheel truck, each 
wheel of which sustains about 20,000 pounds for 100,000-pound capacity cars 
when loaded to their full capacity. The minimum spacing of wheels used under 
100,000-pound capacity cars is 5 feet 6 inches. Cars in passenger service as 
seal frran (iie figures use four-wheel trucks and also six-wheel trucks. The wheel 

* Figs. 37 to 46 have been taken from the Car Buildera Dictionaiy. 



>, Google 



82 STEEL RAILS 

spacing of the four-wheel trucks is 8 feet and for those having six wheds, 
5 feet 3 inches. 

The increase in weight in Amraican passenger equipment in recent years 
has been very great. Steel coaches with wood inside finish used by one of 
the Chicago lines weigh 139,000 pounds. New buffet library cars for smother 



FiQ. 38. — Flat Car, Presged SWel Underframe. Capacity 100,000 Jbs. Wei^t 38,000 lbs. 

western line weigh 153,000 pounds. In the design for new sleepers for the 
Pennsylvania it is difficult to limit the weight to 160,000 pounds, and 75 
to 80 tons may be taken as the weight of modern sleeping cars. The 70-foot 
steel coaches, which are being built in large numbers, are so heavy that it is ■ 
necessary to use six-wheel trucks under them, and these alone weigh over 
40,000 pounds for two trucks.* 



Fia. 39. — Gondola Car. Wooden Body. Pressed Steel Underframe. Capacity 100,000 lbs. 
Weight 44,500 lbs. 

The proposed use of 80-ton freight cars with four-wheel trucks for 
special service f suggests some comparison of these loads with those imposed 
by the engine drivers. The proposed 80-ton car will weigh about 50,000 
pounds and the total weight of the loaded car will be 210,000 pounds, which, 
when carried by eight wheels, produces a static load on the rail of 26,250 pounds. 

Weights on drivers have always greatly exceeded the load on smaller car 
wheels, and the reason for this seems to have been the greater strength of 
the larger wheels. Owing to the work locomotive tires receive in rolling they 

• R^lway Age Gazette, January 28, 1910, t Ibid,, December 22, 1911. 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 



Fio. 40.~CokeCar, AIISt«eI. Capacity 100,000 lbs. Weight 47,600 lbs. 



Fig. 41. — Stock Car, Steel Underframe. Capacity 100,000 lbs. Weight 47,400 lbs. 



Pio. 42. — Vestibuled Cooch, AU Steel. 



,y Google 



STEEL RAILS 



Fia. 43. — Twelve-section Sleeping Car. 



Fio. 44. — Steel Combination Passenger sad Baggage Cor. 



Fia. 45. — Veatibuled Dining Car. 



Fia. 46. — Baggie Car, AU St«el. 

Dignzed by Google 



PRESSURE OF THE WHEEL ON THE RAIL 85 

would appear better able to carry heavy loads than the smaller car wheels. 
While the loads on car wheels in themselves are not necessarily detrimental 
to the rail, it has seemed desirable to provide for the possible affect of a de- 
fective whed by taking a relatively high dynamic augment of the wheel load. 

Freight Cars. 

[^ I ) Weight of Car Empty 60,000 

Weight of Load 100,000 

OO OO 10%0v»load 10,000 

O O O O — : 

O O O O Total 160,000 

O O O O 

O 0) a in Paeeenger Cars, 

4-wheel trucks. 



lOi 



Weight of Car Empty 120,000 

Weight of Load 20,000 

Total 140,000 



O O O O 

O O O O 

O O O. O Poasenger Cars. 

<D <0 (0 (0 

ftj — ft — 6-wheel trucks. 

o in □ (0 t,b». 

Weight of Car Empty 160,000 

L 5'-3'X5'-S'j Weight of Load 20,000 

Total 180,000 



(hCbG) 



337 



o o o o o 

DOOOOO ^" ^y"*""'" '°**^ P*' viheel, pounde 

01 n m fi in m 5— Static load per wheel, pounds 
w-w-N- D = S+ 10,000 

o <n am o in 

Fia. 47. — Typical Load Diagrams for Cars. 

Fig. 47 show^ typical load diagrams for passenger equipment and freight cars. 
Fig. 48 shows the dynamic wheel diagrams of motor cars used on steam 
roads, and F^gs. 49 and 50 illustrate these cars. 

1 , y-o't L sa -a* J ^ C-B* J 

It Is 88 g| 

.1 ■■ '■ ?; 

. O" on o« an 

[, 7-0", I, ^4-6" ^. ,1, 7-0" I 

g g g B D- Dynamic load per wheel, pounds %% gg 

5« 5« S = Static load per wheel, pounds «£ 55 

"Y ~7 D = S + 10,000 where S is more than 15,000 ^j ^i 

an on Z) ■= 1.7S where <Ei is less than 15,000 qm qm 

Via. 48. — Typical Dynamic Load Diagrams for Motor Cats. 



>, Google 



STEEL RAILS 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 87 

In Figs. 51 and 52 are given examples of the cars in use on electric rail- 
ways, both for city service and on the longer runs of the intenirban lines. 



58-ft. " Single-end " Smoking and Express, Mail and Baggage Car. 
Seating capacity, 38. Weight of car body, a£out 34,000 ibe. 

Wheel base of trucks, 6'6'. Weight of trucks, 21,460 Ibe. 

Weight complete, about 38) tons. 



62-ft. Buffet Observution Parlor Car. 
Seating capacity, 35. Weight of car body, about 44,000 lbs. 

Wheel base ot trucks, 6'6'. Wpight of trucks, 21,460 lbs. 

Weight complete, about 44 tons. 



51-ft. "Double-end" Special Parlor Car with Smoking Room. 
Seatbg capacity, chairs 26. Weight trf car body, about 29,000 lbs. 

Seating capacity, seats 52. Weight of trucks, 20,322 lbs. 

Wheel base ot trucks, 6'6', Weight complete, abait 39 tons. 

Fio. 61. — Electric Railw^ Cara (Niles Car and Mfg. Co.). 



,y Google 



STEEL RAILS 



51-[t. "Single-end," Two-compartment, Fast Interuiban Car. 
Seating capacity, 52. Weight of car body, about 26,500 lbs. 

Wheel base of trucks, 6'6'. Weight of trucks, 18,800 Ibe. 

Weight complete, about 32 tons. 



48-ft. Cent«r-vcstibule, Arch-roof, Steel Prepayment Car. 
Seating capacity, 54. Weight of car body, 22,000 lbs. 

Wheel base of Irucka, 6'4'. Weight of trucks, 16,132 lbs. 

Weight complete, about 26 tons. 



42-ft. Double-truck, "Single-end," Pay-AB-You-Enter Car, 
Seating capacity, about 45. Weight of car body, about 16,000 lbs. 

Wheel base of Irucks, 5'0'. Weight of trucks, about 12,000 Ibe. 

Fio. 52. — Electric Railway Care {Nilee Car and Mfg. Co.). 



,y Google 



PRESSURE OF THE WHEEL ON THE RAIL 89 

Fig. 53 presents the typical dynamic load diagrams of this class of equipment. 
The dynamic augment has been taken the same for these cars and the motor 
cars as was used for the freight and passenger cars on steam roads. 

Weight o( car and equipment 43,000 lbs. I a'-o' j , - I bV. J 

We^t o( paseenger load 15,000 lbs. JT f J-. ~ P^^-U 

(100 's 160 n«.) CD iCD v---"-"'^"-- (DiCD 

Total 58,0001bs. || i| || || 

Equipped with two motora weighing 34,000 lbs. aa •>•> na «o 

each, both motors on axle of the rear truck, q-. q 



Intentrbon Passenger Cars. 

Wright ot car body 35,000 lbs. 

Weight of trucks 24,000 lbs. 

Weight of motor equipment 19.000 lbs. 

Weight of passenger load 11,250 lbs. 

(75 ® 150 lbs.) 

Total 89,250 lbs. 

Equipped with four motors hung on four axle»i, bo that the load is evenly distributed. 

L— ?-Q1J i. ?-o" j 

F 1. ~| £o1p" IZ^ 

CD i CD '"""-"°-°" CD I G) 



Interurban Freight Cars. 

Weight of car and equipment , . .62,320 lbs. 

Weight of load (rated capacity) 50,000 lbs. 

plus 10 per cent overload 5,000 lbs . 

Total 117,320 lbs. 

Equipped with four motors, one motor to each axle. 

— Typical Dynamic Load Diagrams for Electric Railway Cars. 
D = Dynamic load per wheel, pounds 
S = Static load per wheel, pounds 
D~3 + 10,000 where S is more than 15,000 
D = 1.7S where 5 is less than 16,000 



,y Google 



CHAPTER III 
supports of the rail 

13. The Tie 

The most common material used for the tie is wood. Some have sug- 
gested (and this suggestion is made with increasing frequency) that ties should 

be made out of materials other 
than wood. Granite ties were 
among the earliest substitutes 
offered ; they were used for some 
time in Dublin, Ireland, and 
on the old Boston and Lowell 
Rmlroad in Massachusetts. For 
some fifty years various forms 
of metal ties have been sug- 
gested, and a large number of 
steel ties have been tried in 
various countries. In recent 
years concrete ties have been 
made. 

The following examples 
present in a general way what 

Fio. 54. — Carnegie Steel Tie. l u ^.i. j. j i i.- 

has been attempted as a substi- 
tute for the wooden tie. The list is far from complete, and must necessarily 
remain so, as new forms of metal and concrete ties are constantly being developed. 
The subject has been very fully reported upon by the committee on ties of the 
American Railway Engineering Association.* 

Steel ties have been used quite extensively on the Union Railroad and on the 
Bessemer and Lake Erie Railroad. The total number of steel ties on these two 
roads is over one million or enough to lay 300 miles of track. There are a large 
number of steel ties of the Gamete type (Fig. 54) in use throughout the country. 

• See Report of Committee on Ties, Proceedings Am. By. Eng. Assn., 1909, 1910, 1911, cuid 
1912, pp. 343-370. 



,y Google 



SUPPORTS OF THE RAIL 



Fig. 55 shows the insulated tie in use on the Bessemer and Lake Erie Rail- 
road. The insulated tie is made by placing a piece of fiber on the steel tie and 



Fio. 55. — Carnegie Steel Ties on the Bessemer and Lake £rie Railroad. 

(Am. Ry. Eng. Asan.) 



Fia. 56. —Effect of Three D^-ailments on Steel Ties. (Am. B.y. Eng. Assn.) 



,y Google 



92 STEEL RAILS 

then firmly riveting a steel plate over the top of the fiber for the rail to rest on. 
This is intended to stop all wear on the fiber. 

Fig. 56 shows the effect of three derailments on the steel ties, which, in 
this case, merely bent down the upper flange of the ties, and in no way injured 
their usefulness as a tie. 



Fia. 57. — Steel Tie after Four Yeara Service. (Am. Ky. Eng. Asm.) 

Tig. 57 shows a steel tie taken from the track that had been in service four 
years. Very little rust was found on the web of the tie and the bottom flange 
of this tie showed very little corrosion. There were no signs whatever of the 
tie failing in any respect. The cutting of the slots or holes in the web of the tie, 
as shown in the figure, has been abandoned, as it was found that with slag or 
stone ballast the holes with the web turned out were not necessary in order to 
keep the track from sliding ^deways. 



,y Google 



SUPPORTS OF THE RAIL 93 

The tie was smooth on the upper face where the base of the r^ rests and 
diowed very little, if any, wear. Providing the wear in years to come is no 
greater in proportion than it has been during the past four years, the tie would 
be good for 25 or 30 years. 



Fia. S8. — Carnegie Steel Tie with Wedge FtuteneT. 



Fig. 69. — Hill FaBtening on Camegie Steel Tie. (Am. Ry. Eng. Aoan.) 



,y Google 



04 STEEL RAII5 

I^g. 58 shows the Carnegie steel tie with the wedge fastener. 

Carnegie Steel Tie with Hill Fastening. — (fig. 59.) Approximately 
100,000 of these ties have been installed in yard tracks at the Duquesne Plant 
of the Carnegie Steel Company. 



?_ 



Fta. 60. — Uanacn Steel Tie. (Am. Ry. Eng. Asan.) 

Hansen Tie. — (Fig. 60.) Ji^ve hundred of these ties were placed in the 
track, July, 1905, near Emsworth, Pa., on the Pennsylvania Lines West of Pitts- 
burg. A great deal of trouble was experienced with the insulation, also from 
the ties sliding transversely and lon^tudinally through the stone ballast, and 
the ties were in consequence removed from the main track in November, 1905, 
and placed in a passing siding. 

Universal Metallic Tie. ^ (Fig. 61.) The figure shows these ties in the 
Pennsylvania Lines tracks near Emsworth, Pa. The design is the trough type, 
being a 6- by 8-inch by 8-foot steel channel. Holes are cut in the web of the 
channel on each side of the rail, and this metal is bent up vertically on each 



,y Google 



SUPPORTS OF THE RAIL 96 

side of a wooden block which fits in the channel under the rml. Clamps, fitting 
over the base of the rail and extending down vertically outside these bent-up 
portions of the channel, 
bind the block, rail and 
tie together. The clamp 
on the gauge side of the 
rail extends tJirough the 
hole in the -base of the 
channels about 4 inches 
into the ballast, giving 
an additional bond with 
the roadbed. A bolt, 
with a tapering head at 
one end and with a taper- 
ing washer at the other 
end, holds the connection 
tight. An insulating 
fiber is inserted between 
the rail and the clamp. 
The weight of this tie is 
175 pounds. 

Fia. 61. — UniverBal Metallic Tie on Pemuylvania Lines. 

{Riulway Age Gazette.) 



Fig. 62. — Snyder Steel Tie. (Am. Ry. Eng. Assn.) 

Dignzed by Google 



96 STEEL RAILS 

Snyder Steel Tie. — (Fig. 62.) The illustration shows these ties in the 
Conemaugh yards of the Pennsylvania Railroad. There is also about one 
mile of these ties in use at Deny, Pa., on the same road, none of them being in 
the main tracks. The standard type of the Snyder tie consists of a steel shell 
1% inch thick, 8 feet long, 7 inches wide, 7 inches deep, and with the bottom 
open. The interior of the shell is filled with a mixture of asphalt and crushed 
stone. In 20 of the ties the mastic had disintegrated and fallen out of the ends of 
the ties after four years service. With this exception the Snyder tie has given very 
satisfactory service in the trade of the Penn^lvania Raibtjad at Conemaugh and 
Dary. 



Fio. 63. — Buhrer Combined Steel and Wood Tie on L. S. & M. S. Ry. 

(Am. Ry. Eng. Assd.) 

Buhrer Steel and Wood Tie. — (Fig. 63.) The figure shows the fourth or 
freight track of the Lake Shore and Michigan Southern Railroad, east of Toledo, 
tied with the Buhrer combined steel and wood tie. Early in 1907 the Carnegie 
steel ties on the Lake Shore and Michigan Southern Railroad were removed 
from the high-Speed track. To care for the insulation the top flange of the tie 
was cut off and two wooden blocks bolted to the web of the tie for spiking strips 
and for the rail to rest on. These strips also rest on the bottom flange of the 
steel tie. 

Mexican Railway Tie. — (Fig. 64.) Practically the whole of the Mexican 
Ralway system of 361 miles is laid with these ties. These ties weigh about 125 



,y Google 



SUPPORTS OF THE RAIL 



«7 



pounds, and cost $2.25. The ties are apparently pving excellent srariee. The 
axle load on this road, however, is not heavy on the hght grades, and on the 
mountain grades, where axle loads as high as 50,000 pounds are employed, the 
speed is slow. 



A. O HA LP LCNOTH 1 




I AT CCHTER 



CROSS SCCT10N 







UOMOlTUOINAl. SECTION Of Tit 




Fig. 64. — Mexican Railway Steel Tie. (Am. Ry. Ei^. Assn.) 

Buhrer Concrete Tie. — (Fig. 65.) About 600 of these ties were used on 
the Pennsylvania Lines west of Pittsburg during 1903 and 1904 in stone ballast. 
Nearly 500 were subjected to heavy and high-speed traffic and the balance to 
medium traffic. The ties failed under traffic, the concrete breaking and crumb- 
ling off from the refinforcement. The ties were removed from time to time and 
by December, 1906, all had been removed on account of breaking. 



,y Google 



STEEL RAILS 




>, Google 



SUPPORTS OF THE RAIL 99 

Fig. 66 shows the bottom or bearing surface of this tie, which illustrates 
how the concrete is left out at the center to provide against side motion. 



Fia. 66. — Bottom Surface ot Buhrer Concrete Tie, (Am. Ry. Eng. Aaan.) 



Fio. 67. — Section of Track on Chicago and Alton R. R. showing Kimball Tie. 

Kimball Concrete Tie. — (Fig. 67.) This figure shows the Kimball tie in 
the track of the Chicago and Alton Raihoad, near Lockport, HI., during October, 



,y Google 



STEEL RAILS 



FiQ. 68. — KimbaU Tie put in Track on N. Y. C. & St. L. R. R., July, 1904. 
Photograph token January, 190S. 



- Kimball Tie Showing Spiking Plugs. (Am. Ry. Eng. Aaan.) 



>, Google 



SUPPORTS OF THE RAIL 101 

1905. The track is kept in good condition, but several of the ties have developed 
cracks which may be due to improper application. Fig. 68 presents further ex- 
amples of the Kimball tie, and Fig. 69 shows a tie in good condition, taken 
from the track. In this tie the spikes entered the spiking plug and the con- 
crete was not damaged, as was found to be the case in most of the ties which 



Fia. 70. — Percival Concrete Tie. (Am. Ry. Eng, Assn.) 

developed cracks. Ties were not rusted to any extent in the center of the track 
between concrete ends. In 1912 there were about sixty Kimball ties in succesrful 
xise in the track of the Chicago and Alton Railroad, the ties having been installed 
in 1905. 

Percival Concrete Tie. — (Fig. 70.) The figure shows ties which were 
used on the Pittsburg and Lake Erie Railroad for about two years. They were 



,y Google 



102 



STEEL RAIia 



removed from the track in 1908 for the reason that the ties failed. The figure 
illustrates very plainly how and where these ties failed. 

The Sarada and Adriatic Railway ties, ^ven in Hgs. 71 and 72, illustrate 
concrete ties used on the continent, and Fig. 127 shows the combined wood and 
metal ties used in France. 



Fio, 71. — Saiwia Tie. (Concrete Review.) 

Sarada Tie. — (Fig. 71.) 3.9 inches in center and 5.9 inches under rails 
by 9^ inches by 8 feet long. Reinforcement, 4 sheets of expanded metal, set 
vertically and connected transversely by iron wires. The rail fastening bolts 
enter from below and are held in tubular castings embedded in the ties. Wei^t 
about 310 pounds. 



V 


• 
• 




^ 


• 
• 


\j 






r\ 


/i_. 




K|; ir^ ^i:^ 



PiQ. 72. — Adriatic Railway Tie. (Concrete Review.) 

Adriatic Railway Tie. — (Fig. 72.) Reinforcement, 29 rods having a total 
area of 3 square inches. The rail is fastened by bolts passing through the tie 
and inserted from below. The beveled rail seat is in accordance with European 
practice. Weight about 286 pounds. 

Riegler Concrete Tie. — (Fig. 73.) Some of these ties have been in ser^ 
vice on the high-speed tracks of the Pennsylvania lines for several years 
without showing signs of deterioration. The ties have a large bearing sm^ace 
and fifteen are used for a 33-foot rail, instead of eighteen standard wooden ties. 



,y Google 



SUPPORTS OF THE RAIL 



103 



The ties have rounded sides, which assist in distributing the downward thrust 
over a short distance on each side of the tie, and the reaction assists in holding 
the tie from slewing, all the ties remaining as first placed at right angles to 



-i:r^ 




r^r-j^ 


s; 


H ^ 


-^_ — 1 — p^ 


1 


"i 


h-ijr-..-.... ::„'!zzn'SvE^^3 " 






[Z^:::::;ij;zzx°: 






J. ■- 




"" 


■""™" 


H. ^// S // S 


j 



Pto. 73. — Riegler Concrete Tie. 



Fio. 74. — Riegler Concrete Tie, Appearance in the Track. (Railroad Age Gazette.) 

the track. The weight of the complete tie is about 850 pounds. It is proposed 
to cast a ring in the end to which a short rope can be attached for hauling into 
the track. 

Table XVIII presents a summary of the service tests on concrete ties. 



,y Google 



STEEL RAILS 



Ei 

m 



I £ i 7 



II 



iri 1 



lllli 



i i |I3 I ; 



5 I ll-g I I's 
s s III i >> 



j § : g|i I ill 



I II p I ; 13 



IIS 



:l S II i 



1 1 iJiliii lljl nil 



i« 



'4 .1 ^ la 



■ -'4 J 
i a* 1 



r is 
I'll 



d I Sll ? 



1 I 



« s 



^ 9 ^ 

I I I ^ 

^ ^ ,s sj 



>, Google 



SUPPORTS OF THE RAIL 



11 

HI 

lii! 



I 
I 



ii 

11 



ilB 



m :HSIHJ II 111:1 



I 



I I Is iiili 



Mii ^s 



4 



; 3jsgt!== 2- .; s 8 ssgE s 



I 



ii 



u 



lllllll 



ii 



I 



1 1 i I «- 

ilii 

Jill III 




I 
Jl 



ti i 



I i I 



>, Google 



106 STEEL RAILS 

Probably no form of reinforced concrete tie has been made which is suitable 
for heavy and high-speed traffic. The real field of usefulness for the concrete 
tie appears to lie in its application in places where speed is slow and where 
conditions are especially adverse to the life of wood or metal. 

The steel tie seems much more promising, but the fact remains that most 
of the rmlroads in this country to-day are using wood, and, so far as the author is 
able to judge from present tendencies, are likely to continue to do so for some time. 

The question of a future timber supply for wood ties is a very important 
one. The railroads are rapidly exhausting the available timber near their lines 
and not on>y is the tie becoming deai-er, but in many instances it is found im- 
possible to obtain a sufficient supply to meet the jumual requirements of the road. 

The experience as set forth in a paper read at the American Forest Congress 
by Mr. L. E. Johnson, President of the Norfolk and Western Railway, is typical 
of most roads.* 

"Ori^nally the country passed through by the raiht)ad was well timbered. 
The first extensive depletion of timber land was on the fu^t hundred miles 
adjacent to the seaboard, where the original timber was cypress and Virginia 
or loblolly pine. 

"Up to the year 1888 the road used a great many cypress ties, but such 
timber is no longer procurable. The second growth of Virginia loblolly pine 
in this district is very knotty, and, further, it is not suitable for crossties until 
it is treated to improve its lasting quaUties. 

"All the balance of the road is in territory where both white oak and 
chestnut oak is indigenous, and up to quite recentiy all the crossties that have 
been needed have been obtained within moderate hauling distance from the 
railroad line. 

"The average requirements in oak ties per year for renewals are 310 per 
mile, aggregating, in round numbers, 800,000 ties per year for the entire road. 
At prevailing prices 800,000 ties cost per annum about $315,000, which is shown 
to be about 15 per cent over the cost of a like numba- ten years ago. This 
total is far below what some railroads less fortunately atuated must pay for 
a like number." 

The general distribution and character of the original forests! of the 
United States are shown by Fig. 75. A glance at this discloses that five groups 

* Proceedings of The American Forest Congress, Washington, 1905, p. 265, 
t The Tlniber Supply of the United States, Kellogg. Forest Service, Circular 97. OrigiDal 
Forests, R. S. Kellogg, Vol. 2, pp. 179, 180. Report of the National Conservation OommieBkHi, Feb- 
ruary, 1909. 



,y Google 



SUPPORIS OP THE HAH 



'I 



I 



I! 

II 



of states embrace the natural timbered areas of the coimtry, — the Northeastern 
states, the Southern states, the Lake states, the Rocky Mountain states and 
the Pacific states. Of these, the two groups last mentioned are occupied by- 
forests in which practically all the timber-producing trees are coniferous, the 



,y Google 



108 



STEEL RAIIS 



first three of both conifers and hardwoods. The earliest atteck was upon 
the white pine of the Northeast, the original stand of which is almost 
entirely cut out. 

The Northeastern states reached their relative maximum in 1870 and the 
Lake states in 1890. The Southern states are undoubtedly near their maximum 
to-day, and the time of ascendency of the Pacific states is rapidly approaching. 
There will be no more shifting after the Pacific states take first place, since 
there is no new region of virgin timber to turn to. 

The percentage of the total lumber cut, furnished by the principal regions 
ance 1850, according to census figures, is given in Table XIX. 



TABLE XIX.- 



■ GEOGRAPHIC DISTRIBUTION OF TOTAL 
LUMBER PRODUCT 



Y«r. 


Nen- Enelud 


SlHtea. 

6 4 

13,6 
24.4 
33 4 

36,3 
27,4 


Smitbern 

~P^«n7 
13.8 
16,5 
9.4 
11.9 
15,9 
25.2 


Pacific 

Pra«Bt. 
3,9 
6.2 
38 
3.5 
7.3 
9,6 




P««iit. 
54.5 
36,2 
36.8 
24.8 
18.4 
16.0 















It is evident that at the present rate of consumption the available supply 
of the present timber used for ties, especially white oak and yellow pine, will 
be exhausted to a serious degree before many years, and that the railroads must 
consider the question of what course they are to pursue in the future. 

Under these conditions there are obviously two courses: First, the reduction 
of the amount consumed, which can be done by the substitution of other 
materials for wood and by the use of preservative methods for prolonging the 
life of the tie, which, by increasing its durability, will diminish the annual 
requirements for renewals; second, by the adoption of forestry methods, 
having for their purpose the proper care and management of the forests still 
remaining and the cultivation of new tree plantations. 

The question of forest preservation and perpetuation is beginning to receive 
attention in this country through the several State Bureaus of Forestry which 
have been established, and attention is given to forest preservation by these 
as well as by the National Government. 

It has been found that the most important need for most of the railroads 
at this time is definite technical information. It is not sufficient to know that 



,y Google 



SUPPORTS OF THE RAIL 109 

timber supplies are being exhausted, but one should also know exactly what 
these supplies are, and what the rate of exhaustion is, and what the probable 
rate of regrowth is in any particular region upon which that particular road is 
depending. 

The need of such investigation is being universally felt, and has manifested 
itself in very striking form, as diown by the two meetings of the governors of 



Fw. 76. — Hunnewell Plantation (Catalpa). {Bureau of Forestry, Bulletin 37.) 
Average diaracter 3.85 ins., 21st year. 

the various states, called by the President in May and December, 1908, in 
Washington, D. C. 

Many tree species * in the United States are adapted to a certain degree 
at least for the production of crossties. Notwithstanding this, in making the 
majority of railroad plantations only two species have been used. These two 
species are catalpa and black locust. 

CataljM t has been planted for a great many years on a great variety of soils 

• ProceedJnKs Am. Ry. Enjt. and M, of W. Assn., 1908, Vol. 9, p. 715. 
t Practical Arboriculture, J. P. Brown. 



>, Google 



110 STEEL RAII5 

and throi^out a wide range of territory, and although many plantations have 
reached the age of twenty-five years or more,* so far as known, the trees in none 
of the plantations have reached a aze suitable for crossties (FHgs. 76 and 77). 
The black locust, although it is a rapid grower and thrives on a variety of soils, 
is so subject to the attacks of insects that trees seldom reach a sufficient dze 



FIQ. 77. — Farlington Forest (Catalpa). (Bureau of Forestry, Bulletin 37.) 
Avamge diameter 4.39 ins., 21st year. 

to make a crosstie. Trees which do reach this size are usually so weakened by 
numerous cavities made by the boring of the insects that the wood cannot be 
used with safety. 

Table XX shows that of the total number of trees planted, the locusts 
predominate, with the catalpa second; the results to date favor the former, 
although it is perhaps too early fairly to estimate the ultimate value of any of 
the plantations now under cultivation. 

• The Hardy Catalpa, Bureau of Forestry, Bulletin No. 37. The Farlington Forest, p. 15. The 
Hunnewell Plantation, p. 26. 



>, Google 



SUPPORTS OF THE RAIL 



I 



i!l Piii 

II I if III 



I 



5 i 



i' I ■ ■ 



I 



i i i 



11 

nil li I 



i iiiiiiiii 



i i 



i jji i I lull 



I II 111 



I I M i 



>, Google 



STEEL RAILS 



1 

1 


i i |ti 

JIIUii! lit. 


1 1 


i 
1 


1 

s 


3 




1 g 




s 
1 

1' 
i 






1 ^ 




t 


S 
« 




s s 




] 


a 1 

3 1 C 


1 I 




J 




1 § 




1 


1 11 1 


1 i S 


! 


1 ■!. 


1 i 1 


1 

1 


1 : 


^ i 


ill 





I 

% 
pt 
ii 

II! 
.11 

i-r 

ill! 

lilt 
111 

Ml 

iii 



>, Google 



SUPPORTS OF THE RAIL 113 

Tree planting as such by railway companies has not been a very successful 
matter, and it is generally felt that the planting should be regarded as supple- 
mentary to other methods for securing a tie supply, particularly to the manage- 
ment of forest lands. 

There are, without question, large areas of timber in the South which can 
be obtMned at a reasonable cost at the present time, and it seems to be very 
much more advisable to buy forest regions, or where cut-over lands are pur- 
chased, to encourage the growth of natural forest trees, rather than to go into 
extensive experiments for the planting of new trees. 

Forest planting in some cases may be desirable when a railroad has waste 
land for which it has no particular use. It is a good object lesson to the farmers, 
and if the plantations are successful they will net a fair return on the invest- 
ment and furnish a limited supply of tie and timber for the future. 

It should be observed, however, that it would not be practicable for the 
individual roads to plant enough trees to supply their timber requirements, 
and further the critical period of scarcity and high prices would come before 
any of the trees so planted would reach maturity. 

The information assembled by the Committee on Ties of the American 
Railway Engineering Association, in 1910 (Table XX), shows what has been 
done by the railroads in the way of tree planting; the situation is very little 
changed at the present time, and, in the opinion of those best able to judge, 
relief from this source is very uncertain. 

If the railroads wish to provide agmnst future scarcity and excessive 
prices with any degree of certainty it will be necessary for them to actively 
engage in forestry operations, having for their purpose the management of 
mature timberlands and the cultivation and reforestation of the cut-over lands 
within the forest area. This is an individual problem with every road, but, 
generally speaking, it is the only sound policy which will provide for the future 
requirements fifteen or twenty years hence. 

Some of the railroads have now undertaken to preserve the timberlands 
which they acquired through land grants or otherwise. The Southern Pacific 
in northern California and southern Oregon still have quite large areas of good 
timber from which they can cut mature trees. The Northern Pacific has been 
cooperating with the government for some years with a view to finding how 
best to handle their western holdings, and provide a source of tie supply at the 
eastern end of then- lines. In the East, the Delaware and Hudson have put 
about one hundred thousand acres in the Adirondacks xmder management.* 

• Timber Supply io Relation to Wood PreaervaUon, E. A, Sterling. Proceedings, American 
Wood Preserver'B Awooiation, 1911, pp. 140-144. 



>, Google 



lU STEEL RAII^ 

While the great desideratum is the obtaining of a permanent source of 
supply of tie timber, the economic side of the problem must as well be 
considered. 

The application of actuarial methods to forestry is, despite the obvious 
difficulties about the assessments of the differait factors used in making calcula- 
tions, the only correct way of estimating the finandal position of timber crops 
as a commercial investment. 

The most profitable rotation is what should, both in theory and in practice 
recrave most consideration in the management of a forest. It is found by mak- 
ing various calculations, each as if for a single crop, in accordance with Faust- 
mann's formula, and ascertaining that particular rotation which shows the 
greatest profit by indicating the maximum productivity or largest capital value 
of land and growing stock. 

Faustmann's formula is as follows:* 

A n+(r.xi.Op--)-Kr,xi.op-^)-i- ■ ■ • (r.xi.oy-')-(Cxi.Op-) g 

1.0P--1 O.Op* 

where A = The productivity of the woodlands (as estimated by the net 

value of the timba- crop, etc.) ; 
F, = The net income, free from cost of harvesting, yielded by the 
mature fall at the year (n); 
T„ T^ . . . T^= The net income, free from cost of harvesting, yielded by 
the thinnings at the years a, b, . . . q; 
p = The percentage or rate of interest which the woodlands are 
supposed to yield annually on the investment represented 
by their capital value; 
C = The cost of forming the crop originally, or of regenerating 

.or replanting the area on the fall of the mature crop; 
g = The annual outlay for general charges (supervision, protec- 
tion, taxes, etc.). 

After detamining the most profitable period of rotation, the amount of 
land required to produce a given amount of ties annually can be found. 

The cost per acre that can be paid for the land is determined as follows: 
The average annual charge, at present prices, for different kinds of ties may be 
taken as about 12.8 cents. 

* The Foreeta, Nisbet, Vol. U, p. 239, London, MCMV; cmd Eoonomics of Foreetry, Femow, 
Hew York, 1902. 



,y Google 



SUPPORTS OF THE RAIL 116 

This may be arrived at by the following relation. The discounted present 
value of an annual rental or return r obtainable for n years in all, the rate of 
interest being p, is expressed by the formula: 

^_ ra.Op--l) C(1.0p-xO.Op) 

LOp-xO.Op 1.0p"-l 

In the case of a white oak tie, 

C = % .90, cost of tie in the track, and r = $ .14 annual charge. 

n = 8 years, life of tie; 

p - rate of interest, 5 per c^it. 

The table given below shows the annual charge for different kinds of 
wood. 

White oak 14.0 cents annual charge 

Heart pine 12.5 cents annual charge 

Red odk, untreated 12.7 cents annual charge 

Miscellaneous 12.0 c^ts annual chaise 

Assuming the life of a treated tie produced by the forest to be 12 years, the 
value of such a tie can then be expressed by the formula 
„ r(1.0p--l) 
1.0p"x0.0p' 
where C - value of the tie; 

r = return or annual charge, 12.8 cents, obtainable 12 years in all; 
p = rate of interest, 5 p^ cmt. 

Substituting these values in the formula, there results for the value of the 
tie $1.13. From this there must be deducted: 

Cutting $0.10 

Handling 0.05 

Treatment 0.30 

Transportation 0.20 

Putting in track 0.15 

Total $0.80 

which leaves for the stump value of the tie $0.33. 

The amount of ties produced by the forest will depend upon the kind of 
trees grown and the location of the tract. An annual yield of three ties per 
acre should be expected under careful management in most cases of moderately 
rapid growing trees. This will bring in a return per acre of: 



,y Google 



lie STEEL RAILS 

S - 3 ties at $0.33, leas management and taxes $0.30 - $0.69, and the 
investment per acre wiiich wili give a five per cent return will be: 

The wasteful methods employed in cutting ties in the past have called forth 
many protests and suggestions as to how this waste might be checked. The 
Forest Service states in this connection: 

" The suggestions made for economy in the cutting of ties have been largely 
in the direction of preventing wasteful cutting. The manner in which they 
have been cut from trees has been largely determined by the ease and rapidity 
with which ties could be made, and by the knowledge that certain portions of a 
log were more serviceable for tie purposes than others. 

" Ties were usually made out of heart wood, using the best and only the 
straight, live trees. No attention was paid to the waste mcurred by cutting 
off all the sapwood top section, by leaving dead trees, etc. But with the intro- 
duction of treated ties certain new developments in tie making have taken 
place. Treated ties allow the use of sapwood, of sawed dead timber, and of 
sawed ties, consequently tie forms which were altogether impracticable under 
the old methods are now within the field of possibility, and must be considered 
on their merits." 

In view of this Dr. von Schrenk has proposed a form of half-round tie vAich 
has be«i used extensively abroad (Figs. 78 and 79). The following description of 
the proposed form is taken from his excellent paper on Cross-Tie F(»ins and Rail 
Fastenings, t 

This form of tie is probably a more economical tie tiian the present rectangular 
tie used in this country, and, on account of its proved merits, should properly be 
considered as a posable substitute for the present form. 

If we consider the manner in which the load is distributed from the base 
of a rail resting on a 5-inch plate, which in turn rests on a tie 8 inches broad, 
we shall find that the lines of force acting from such a tie plate are distributed 
on the ballast as indicated in Fig. 80. 

* The cost per acre that can be paid for the forest laud ie baeed upon the annual chai^ of un- 
treated ties as representing the average outlay by the railroads for thia material at the preset time- 
The use of treated ties would probably reduce the aooual chai^ per tie, but at the same time it must 
be borne in mind that owing to the rapidly increasing cost of the timber from wliich the tie is made, 
the annual charge for a treated tie will probably rise as high as the present figure for a natural tie, 
before auffident time has dapsed for the treated ties to affect the general average. 

t CroBs-Tie Porma and Rail Fastenings, Von Sohronk; Bureau of Fwestry, Bulletin No. fiO. 



>, Google 



SUPPORTS OF THE RAIL 



- Standard Prussian Ties of Baltic Pine. (Bureau of Foraetry, Bulletin No. SO.) 



- Standard Oak and Beech Ties on the French Eastern Railway. 
(Bureau of Forestry, Bulletin No. 50.) 



>, Google 



STEEL RAILS 



Keeping in mind the dedrability of an increased bearing surface on the 
ballast, it is suggested that a type of tie with a top-bearing surface of about 





Distribution of Preaaure from Tie FUto Distribution of Preasure from Tie Plate Id Balf- 

in Ordinary Tie. round Tie. 

Fia. 80. — Distribution of Pressure from Tie Plate. 



6 inches and a base-bearing surface of 
anywhere from 8 to 12 inches will not 
only give a suffident bearing surface 
for the rail, but will also give a much 
more stable track. Such a tie is shown 
m Fig. 81. 

Fig. 82 shows the 7 by 8-mdi tie 
and tie with 6-inch top and 12-inch 
base, spaced as closely as is consistent 
with the proper use of the shovel or 
other tool employed to tamp the tie. 



- Half-round Tie Proposed by the 

Forest Service. 




Fm. 82. — Spacing of Half-round Ties. 

The comparative showing of rectangular 7 by 8-inch and 7 by 9-lnch ties 
and of ties with 6-inch top and 12-inch base, spaced respectively at 11 and 10^ 
inches, as shown in Fig. 82, is given in Table XXI. 



,y Google 



SUPPORTS OF THE RAIL 1 

TABLE XXI. — COMPARISON OF RECTANGULAR AND HALF-ROUND TIES 

(Bunu of Fonatry. Bullalin No. SO) 







^'^nTfc 




IXSbKba,. 


TX>iDch». 


it^tlSi. 


DUtaDce between bearing centers, on both top and base of 

tie, inches 

Increase in distance between bearing centers by use of ties 


19 

3.5 
3,242 
426 
2,161 

17,290 

5,238 


20 

2,5 
3.168 
352 
2,376 

19,008 

3.520 


22.5 












2,816 


Bearing surface on ballast per mile, with 8-feet length, 


Gain in bearing surface by use of tie of the new form, square 









According to this table the number of ties of the new form required per 
mile is 352 less than with the 7 by 9-inch tie, and 426 less with the 7 by 8-inch 
tie, while the amount of bearing surface obtained is greater by 3,520 square feet 
than that obtained by the 7 by 9-inch tie, — an increase in bearing surface of 
over one-sixth. At the same time there would seem at first sight to be a con- 
adderable saving from the smaller number of ties, but in reality there is little 
difference in expense because of the larger number of feet, board measure, in 
the new tie. 

It now becomes necessary to conader 
the changed tie form from a lumber 
standpoint. Ties are now being cut from 
trees of all diameters from 9 inches 
upward. If cut but one from a cross 
section, they are usually termed pole 
ties. Most of these are rounded at the 
edge and squared on two sides (Fig. 83), 
with a required bearing surface of 6 to 
8 inches. Pole ties are now cut from 
trees as large as 17 inches in diameter. Most of them are hewn, and in the 
hewing much of the outer portion of the tree is wasted. In larger trees also a 
great deal of timber is wasted, even when ties are split in the most economical 
fashion. In the majority of instances no wane is admitted for a first-class tie, 
so that logs less than 10 inches in diameter will not make ties of this class. This 
means that a great many tops are now left in the woods because they are too 
small. By adopting the half-round tie suggested above (and here emphasis 



Fia. 83. — Pole Tie. 



,y Google 



120 STEEL RAUfi 

ought to be laid upon the fact that ties cut according to this shape will all be 
treated) it will be possible to utilize a great many logs which now do not make 
ties, and also to cut a good many more ties out of the same amount of timber 
than under the present specifications. 

The cutting of ties of this new form will be essentially a sawmill proposi- 
tion. Where now there is a great deal of waste in hewing, if the log were sawed, 
it would mean the obtaining of several 
boards on the side. The number of 
boards to be sawed from a tree 16 inches 
in diameter, making two ties, will depend 
largely upon the value of the timber from 
which the ties are made. For instance, 
it will pay to make as many boards as 
possible out of a 16-inch, two-tie log of 
red oak or gum, while with timber like 
loblolly pine, the lumber of which has a 
low value, it will at present not pay to 
cut oS many boards. In the case of such 
T; , „ „ , ^. timber an extreme form of the half-round 

Fig, 84. — Extreme Form of Half-round Tie, 

tie will be applicable (Fig. 84). 

The influence which the new tie form will have upon the size of trees cut 
for tie purposes ought to be a marked one. It certainly would discourage the 
cutting of pole ties to a very considerable extent. It would not pay to make a 
tie out of a small tree when by leaving it for a few years two ties could be made 
from the same tree. In other words, the present policy of cutting trees 11 or 
12 inches in diameter would be found less profitable than cutting trees 16 or 
17 inches in diameter. 

There is probably no other branch of the lumber industry in which so 
many small trees are annually destroyed and the possible regrowth of forests 
retarded to such an extent as in the manufacture of ties. The practice of sawing 
ties from logs is going to be more and more prevalent as the old feeling that a 
sawed tie is not worth having disappears. This feeling is abeady rapidly dis- 
appearing. It certainly will disappear entirely when railroad men realize that 
with a chemically treated tie it makes no difference whether it be sawed or 
hewn. With increasing permanency in the source of supply, it will pay more 
and more to put up small sawmills, which will saw ties and such lumber as may 
incidentally come to them. This will be particularly true in regions where 
there are rapidly growing tree species, such, for instance, as loblolly pine. The 



,y Google 



SUPPORTS OF THE RAIL 



121 



cutting of these trees will, moreover, make possible the use of large quantities 
of timber which now is practically wasted and from which the lumberman has 
no return. This is particularly true of tops. 

As the rail should be designed to have sufficient stiffness to enable it to 
distribute the load over a number of ties, allowing only such a proportion of 
the wheel load to come on each tie as can be safely carried, it will be necessary 
to determine the safe load that it will be proper to put on the tie. As a mean 
representing the average general practice, we may take in the following discus- 
sion a 7 by 8-inch by 8-foot 6-inch tie and a 7 by 9-inch by 8-foot 6-inch tie 
(see Table XXII). It would seem desirable also to consider the strength of 
the half-round tie. 

TABLE XXII. — SIZE OF TIES AND SPACING 

(Am. Ry. Enj, Aa«.) 



Southern 

Penn. R.R 

L. &N 

B,&0 

S. & W 

P.4R 

Peim. (S. W. Sys.)... 

Lehigh Valley 

N..C. &.St. L 

D. AH. Co 

-A... B. 4 A 

Cent, of N. J 

B..R.& P 

C..C.&0 

A. C. L 

Penn. (N. W. Sys.),. 

D., L. A W 

Fla. East Coast 

C.,C.,C.&St.L.... 

Hocking Valley 

L. S. AM. S 

Erie 

Long laUnd 

South. Pacific 

Union Pacific 

S. A. L 

X. Y., N. H. AH... 

0. of Ga 

G., H. A S. A 

Georgia 

M. AO 

Norfolk Southern., . 
N. Y. C. AH. R.... 

Great Northern 

a P., L. A. A S. L.. 

Nortliem Pacific 

D. A R. G 

C.,B. &Q 



7X7 and 9X8 

7X7 and 9X8 

7X7 and 9x8 

7X7 and 9X8 

7X7 and 9X8 

7x7 and 9X8 

7X7 and 9X8 

, 7x7 and 9X8 

7X7 and 9X8 

7X7 and 9x8 

7X7 and 9X8 

7X7 and 9X8 

7x7 and 9X8 

7X7 and 9X8 

7X7 and 9X8 

7X7 and 9x8 

7X7 and 9X8 

7X7 and 9X8 

7X7 and 9X8 

7X7 and 9X8 

7X7 and 9X8 

7X7 and 9X8 

7X7 and 9X8 

7X9X8 

7X9X8 

7X9X8 

7X9X8 

7X9X8 

7X9X8 

7X9X8 

7X9X8 

7X9X8 

7X9X8 

7X8X8 

7X8X8 

7X8X8 

7X8X8 

6X8X8 



2816 
2816 
2S16 
2816 



3040 
2720 
2720 



3164 
2816 
3200 



2900 
3200 
32P0 



C.,R. I. A P 

St. L. AS. F. 

Grand Trunk 

M..K. A T 

CoL A Sou 

Maine Central 

C. &E. I 

C, I. A L 

El. P. &S.-W 

St. L., B. 4 M 

Ft. W. AD. C 

C. A N.-W 

C, M. A P. S 

C, M. A St. P 

C.I. AS 

St. L. S. W 

M. ASt. L 

S. A. A A. P 

Rutland 

Mo. AN. Ark 

S, Fe, P, AP 

L. E. A W 

G. R. AI 

W. ft L. E 

N. W. Pac 

Mo. Pac 

B. AM 

K.C., M. A O 

Tcnn. Cent. 

C. 0. w 

C, H. A D 

M. C 

Bangor A Aroostook , 

N. Y.,0. AW 

M., J. A K.C 

est. P., MAO... 
D..S. S, A A 



6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8x8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6X8X8 
6x8x8 
6X8x8 
6X8X8 
6X6X8 
6X9X8 
7X9X9 
7X7X8 
7X7X8 



3200 
3200 
3200 



3000 
3000 
2092 
2992 



3564 
2880 
3120 



,y Google 



122 



STEEL RAILS 



14. Bearing of the Rail on the Tie 
The genera] tendency at the present time is more and more towards the use of 
tie plates. With the introduction of the treated tie it is necessary to adopt some 
means to protect the wood from wear at the rail bearing on account of the longer 
life of the tie. 

The objections which have been made to tie plates were, first of all, that 
they buckled severely. This, however, has taken place only when the plates 

were too thin, and the 
following record of tests 
made of a prominent 
make of tie plate show 
that the present plates 
have ample strength to 
resist buckling (Table 
XXIII and Fig. 85). 

Most plates have 
been made with the idea 
of being anchored to the 
tie so as to prevent the 
communication of the 
motion of the rail to the plate. As a result, we have a large number of different 
forms of plates, provided with prongs, spines, or flanges on the bottom, which 
are pressed into the tie either by table xxiii.-test of McKEe tie plate 
the weight of the passing load or 
before the rail is laid (see Plate 
XXII). 

The chief objection which is 
made to plates at this time, par- 
ticularly in connection with the use 
of softer woods, is that not only do 
they not aid in preventing the wear 
of fibers, but they actually assist 
the rail to wear. This is well illus- 
trated in Fig. 86, showing a tie plate 
which has been in position on a loblolly pine tie for about four years. 

The constant rocking motion of the rail, which had become very marked 
as the spikes were pulled from the soft wood, had transmitted itself to the tie 
plate, and when a load passed over the rail the tie plate moved back and forth 



Fig. 86. — Teat on McKee Tie Plate. 



LouLiuAiiplwd. T* 


«No 1 


Twit No, I, 


PtHudii. DeB« 


Ion InobB. 




250 


000 


,000 


4,000 


022 


022 


8,000 


029 


030 


12,000 


035 


,038 


16,000 


042 


,056 


20.000 


054 


070 


24,000 


066 


-081 


28,000 


081 


.100 


32,000 


101 


.129 


36,000 


122 


,165 


40,000 


154 


,210 


44,000 


188 


.248 


48,000 


221 


.285 



>, Google 



SUPPORTS OF THE RAIL 123 

in unison with the rail. It was not long before the soft fibers of the loblolly 
pine suffered under this treatment, and in the course of time so great did the 
abrasion and crushing of the fibers by the plate become that a considerable hole 
was made under the plate, in which water gathered. The plate gradually sank 



Fia. 86. — Wear of Tie under Tie Plate, 
The upper illustration shows a Loblolly Pine Tie treated with Zinc Chloride, aft^ four years' service 

in Texas. 
The lower iUuBlration shows a longitudinal section through the spike hole of a Western Yellow Pine 
Tie after several years' service in Texas. 
(Bureau of Forestry, Bulletin No. 50.) 

down into this hole, as shown in the illustration. When the tie was removed 
it had disappeared in the wood, and the base of the rail was resting on the outer 
edges of the tie beyond the plate. 

This tie had been treated with zinc chloride. The water which gathered 
under the tie plate leached out the salt, and as a result decay started on both 
sides of the plate, as the illustration shows. The tie had to be removed, al- 
though the rest of it was perfectly sound (Figs. 87 and 88). 



,y Google 



STEEL RAILS 



Fig, 87. — Section of Tie under Rail Bearing showing Wear and Decay. 



Fio, 88. — Section from Middle of Same Tie showing Entire Soundneas. 

LOBLOLLY PINE TIE TREATED WITH ZINC CHLORIDE, AFTER FOUR 
YEARS' SERVICE IN TEXAS. 
{Bureau of Forestry, Bulletin No. SO.) 



,y Google 



SUPPORTS OF THE RAIL 125 

The types of plates used in Europe are without exception flat plates.* Figs. 
) to 96 represent plates med by several European roads at the present time. 




Plan of 

Fig. 89. — Belgian State Railways, 105-Ib. RaU and Tie Plate. (Am. Ry. Eng. Assn.) 



' See "The Question ot Screw Fastenings to Secure Rails to Ties." W. C Cuflhing, Proceedings 
Am. Ry. Eng. and M. ot W. Assn., 1909, Vol. 10, Part 2. 



,y Google 



STEEL HAILS 



s 


Hi 




,0^^^ 


s 


r 


j( 


1^- 


t ';" 






1 ■ 


' % 


/ l.«- ,|o. 


Efl u^- \. ; . 


r\ 




k - 


-^ : I 


\ 1 


^ ji ■ 


)r^ 


I .Ik 


lA 


'■^ 




-d^ 




A/ 1 


\ ^ 


_^ 




r^ 


» 


■-y 


A'l 


'^A 




1 


^\ 


\-k 


-iH^ 


S^— cnzi 


-.■1'J1L'J,J-.J-._.| 


1 ' 


1 i 


^: 


j '§ 




1 i-/ 


\ 1 1 




Fio. 90. — Belgian State Roilwaya, 115-U>. Ran and Tie Ftat«. (Am. Ry. Eng. Aam.) 



,y Google 



StrPPOBTS OF THE RAIL 




InTermadiote MooK Plote. 
Fia. 91. — Kingdom of WOrttemberg State Railways, Tie Plate. (Am. Ry. Eng. Ama.) 



,y Google 



STEEL RAILS 




92. — Bavari&n State Railways, Joint Hook Plate. (Am. Ry. Eng. Aaan.) 



>, Google 



SUPPORTS OF THE RAIL 




Joint HooK Plan. 
Fia. 93. — Kingdom of Saxony State Railroad, Joint Hook Plate. (Am. Ry. Eng. Aaso.) 



>, Google 



STEEL RAILS 




Lothringen State Railways, Tie Plate. (Am. Ry. Eng. j 



>, Google 



SUPPORTS OF THE RAIL 




SeCTION THROUGH JOINT TIC 



] c 


n 




\ 


\ 






•nS 






^ 






h. 




1.89^ 










»J 


'".I "" 




p 


fij 




\ 




«30- 


^ 


\ 








rs 


- 




11 




11.42" 




■i^ 












1 





State Railways, Tie Plate. (Am. Ry. Eng. Assd.) 



,y Google 



STEEL RAILS 




Intarmediatfi Wedge Pkite 
Flo. 9S. — Bavarian State Railways, Itrtermediate Wedge Plate. (Am. Ry. Ei^. Asan.) 

The general tendency on the Continent has been toward adopting more and 
more rigidly flat plates, with firm fastenings. The almost universal adoption 
of this principle is very striking at the 
present day. 

On the French Eastern* the rail 
rests on the tie without metallic plates, 
except on very sharp curves (of 984.25 
feet radius and under). Plates of poplar 
or felt are placed under the rail, solely to 
protect the wood against the mechanical 
action of the base. These plates are 
compressed before being used, so that 
they will not be further compressed 
under the pressure of the rail. The 
plates are furnished 0.28 inch thick, and 
Fia. 97. — Wooden Tie Plate on French Eaatem. the Compression brings them to 0.16 inch. 

* Bureau of Foreetiy, Bulletin No. 50, von Scbrenk. 



>, Google 



SUPPORTS OF THE RAIL 133 

The ties are adzed at the treating plant so that a place is left for this flat 
wooden shim. When the track is laid, the shim is placed in position (fig. 97) 
and screw spikes are screwed into the tie. Their pressure holds the plate firmly 
between the base of the rail and the tie. In Fig. 97 the wooden tie plate is 
represented by the thin unshaded portion between the rail and the tie. It is 
exactiy the width of the rail. In the course of time the motion of the rail wears 
out this shim, and a new one is substituted by giving the screw spike one or 
two upward turns. A new plate is then shoved in endwise and the screw is 
fastened. The length of life of one of the wooden shims on the main-line tracks, 
such as that of the French Eastern from Paris to Strassburg, is about one and 
one-half to two years. 

Dr. von Schrenk gives the theory upon which this wooden plate is used as 
follows: The principal function of the plate has been s^d to consist of prevent- 
ing the wear of the fibers of the tie immediately under the rail base. This wear 
consists in the actual breakage of the wood fibers under a grinding and tearing 
action rather than in crushing them. 

In considering the function of the tie plate we have three bodies to deal 
with; the tie, the tie plate, and the rail. Motion might conceivably take place 
either between the rail and the tie plate or between the tie plate and the tie. 
When a metal tie plate is used on the hardwood tie, and is successfully anchored 
in it, the tie plate and the tie act as one body, over which the rail moves back 
and forth. As soon as the tie plate loses its holding power, however, the chances 
are that when the rail moves across the tie the tie plate will oscillate back and 
forth in unison with the rail. This results in breaking the wood fibers under- 
neath the plate. Where a wooden plate is used, it adheres so closely to the 
wood that when the rail moves across the tie the wooden plate and the 
wooden tie are liable to act as one, even though the tie plate is not anchored 
to the tie. 

The Forest Service tests have not shown results favorable to wooden tie plates. 
While the tests have not been veiy thorough, they have been thought to throw 
much doubt on the efiiciency of this form of plate. 

• For some years the question of a satisfactory fastening between rails and 
soft-wood ties has been a subject of continuous experiments on the Prussian 
Government railroads. The first investigations followed the general use of 
plain bearing plates, 7J by 6J inches (rolled steel) in size, shown in Fig. 98 a. 

* FaBtening ot Rails to Soft Wood (Kne) TieB, Organ fur die Fortachritt« des Eiaeabahnweseiis, 
May 15, 1908, et seq. TranBlation appears id Vol. 10, Part 2, Proceedings Am. Ry. Eng. 4 M. of W. 
Assn.. p. 1333. 



,y Google 



134 



STEEL RAIIS 



It was soon discovered that on soft-wood ties, the small adhedon between 
the spike and the wood permitted the spikes to pull out to a more or less extent, 
and the loose rail, under the sudden applications of load, would quickly batter 
down the wood. Besides, the pressure on the tie not being uniform would 
produce a kind of convex wear in the wood, as illustrated in Fig. 98 6. 

Some improvement was obtained by the use of screw spikes, without, how- 
ever, entirely overcoming the abnormal wear and the consequent looseness 




Fia. 9S. — Plain Bearing Flatee, German Experiments on Tie Plates. 



and inefRdency of the structure (Figs. 98 e and d). The hook plate, shown 
in Fig. 99 a, was next tried. The hook, which was made to hold the outade 
flange of the rail, necessitated a plate somewhat longer on the outaide than on 
the gage side, resulting in an uneven distribution of pressure on the wood and 
a condition as shown in Fig. 99 5. The bending of the screw spikes observed 
in thii case was at first thought to be due to the lack of support of the head of 
the spikes on the far ade from the rail flange, and, to remedy this supposed 
defect, rail clips wwe introduced, giving the head of spikes a full support all 



,y Google 



SUPPORTS OF THE RAIL 



135 



around (Fig. 100 a). This arrangement proved to be much better than any 
previous one, but still did not produce a satisfactory fastening. Fig. 100 6 indi- 
cates clearly the manner of f^lure of these plates. 




Fig. 99. — Hook Plates, German ExperimeatB on Tie Plates. 



It is evident that the plate hook, being rigid and incapable of producing 
any actual pressure against the rail flange, would cause all the stresses to be 
carried against the screw fastening, and as soon as this would wear in any of 
its parts the rail would become loose under the hook, and the shocks would 

a 




Fia. 100. — Hook Plates with Clips, German Experiments on Tie Plates. 



begin their destructive work. Also, the direct pressure under the head of the 
spikes would tend to pull these out of the tie, and constitute another element 
of weakness to the general construction. 

The impossibility of fastening the rail with the same amount of holding 
power on both ades, besides the drawbacks enumerated above, led to the intro- 
duction of an entirely different system of fastening. 

The first set, or Group 1, of plates are shown in Fig. 101. These plates had 
a bearing on the ties of 90 square mches as against 80 square inches in the 
largest previous plate, and were fastened to the ties by means of four screw 
spikes absolutely independent of the rail fastemng. The rail was In its turn 



,y Google 



136 



STEEL RAILS 



fastened to the plate by means of two bolts and clips, these bdng independent 
of the tie fastening. 

The clips were made so as to be capable of adjusting the gage of the track 
by being reversible, and also of such a shape as to take up and transmit hori- 
zontal forces at the base of the rail to the shoulders provided for in the tie 
plates. In this manner the upward forces would be resisted by all the screw 
spikes, and dmilarly all the horizontal forces would be taken care of. Spring 




I2?f6 

it 320.'::"- — » 




oil IMMo 


I 


■:^ 4c 


§■ 


o|i II 1 ]o 


1 




Fia. 101. — Group I, German Experimenta on Tie Plates. 

washers were provided under the head of the screw spikes and the rail-fastening 
nut. 

Eighty-three of these plates were put in service in 1898 and removed from 
the tracks, together with the ties, in 1907 for examination. The tie wear 
was found to be very slight and very uniform under the base of the plate, 
varying from a minimum of 0.14 millimeter (1^5 inch) to a maximum of 0.19 
millimeter (^J^ inch), except in a few cases where spikes had become loose 
and caused an increased as well as an irregular wear. 

- 250"- 




FiG. 102. — Group 2, GermaD Experiments on Tie Plates. 



Sand, however, was found between the ties and plates, and this might 
have caused even this slight wear. The rail seat in the tie plate had worn to 
about the same extent (maximum iJj inch). No other sign of deterioration 



The second series of tests. Group 2 (Fig. 102), was carried on under con- 
ditions similar to those for Group 1. The main diiference between these plates 



,y Google 



SUPPORTS OF THE RAIL 137 

and the plates of Group 1 lies in the size of the bearing surface over tie, being 
about 70 square inches for Group 2 plates as against 90 square inches for Group 1 
plates. 

On removal of the plates, it was found that the wear of tie developed the 
same uniform wear as in the previous group. The slight wear gave the impres- 
sion of being purely compression, there being no indication whatever of side dis- 
placement. As a matter of fact, after the screw spikes had been removed the 
plate had to be knocked off with a hammer. Sand was found under the edge 
of only a few plates. The screw spikes used on these plates were 4| inches by 
f inch and had a deeper thread than those of Group 1. 

In spite of the smaller bearing of this plate, as compared to Group 1, the 
amount of tie wear was actually smaller, and the fastening generally more 
satisfactory. 

Center of Raik , 




4% 4M" 

Fia. 103. — Group 3, German Experiments on Tie Plates. 

Condition of test in Group 3 (Fig. 103) was similar to the previous tests in 
Groups 1 and 2. The mmn difference in this case consisted in the sloped top 
of the tie plate, which gave the rail a desired amount of inclination toward the 
gage side. This arrangement brought the center of rail closer to the outer edge 
of the tie plate by about 10 millimeters (f inch). Screw spikes used were similar 
to those in Group 2, but with a somewhat better grip, the holes having been 
drilled smaller. To increase the rigidity of the fastening, double spring washers 
were employed on all the screw spikes. 

The tie wear was smaller than in any previous instance. The gage of the 
track was measured frequently and found to remain practically unchanged. 
The spring washers, which had shown some falures when used singly, were 
found in this test to have their original elasticity unimpaired. This design of 
tie plate, however, failed in a few instances, as shown at "a," Fig. 103, which 
would seem to indicate that a greater stress was carried against this point than 
in the other arrangements. 



,y Google 



138 STEEL RAILS 

It is clearly evident from the behavior of plates of Groups 2 and 3 that the 
wear of ties is not at all directly proportional to the extent of the bearing surface 
<rf the tie plate, but depends more upon the rigidity of the fastening. In the 
case under consideration, the most important point developed is the necessity 
of rigidly fastening the tie plates to the ties in order to preserve the life of the tie. 

15, Fastening of the Rail to the Tie 
In this country the ordinary nail spike is generally used for fastening a rail 
to a wooden tie. The most important objections to the spike are: first, in 



Fio. 104. — Short Leaf Pine Tie, after 2 Years' Service, cut through Spike Holes. 
{Bureau of Foreetry, Bulletin No. 50.) 

the soft-wood tie the spike does not hold with sufficient firmness to keep the 
rail securely to the tie; second, in driving the spike into the softer woods the 
fibers are broken to an unusual extent (Fig. 104). As a result they do not 
withstand lateral pressure of the rail, and consequently the spike hole is rapidly 
increased to such an extent that the spike no longer holds. Water collects in 
the enlarged hole and decay sets in (Fig. 105). 

Table XXIV * compares the holding force of a common spike (Fig. 106), 
weight 165 spikes to 100 pounds, with that of the common screw spike (Fig. 107), 
similar to those used on the French and other continental railroads, weight 
85 spikes to 100 pounds. 

• Holding Force of Railroad Spikea in Wooden Ties, Forest Service, Circular 46. 



>, Google 



SUPPORTS OF THE RAIL 



- Crofls Section through the Spike Holes of Short Leaf Pine Tie, treated with Zinc Chloride, 
Texas (Bureau of Forestry, Bulletin No. 50.) 



TABLE XXIV. 


—HOLDING FORCE OF COMMON AND 

(FoTBBi Sen-ice. Citoilar 18) 


SCREW SPIKES 




oT-i^ 


Condition or Wood. 


Force R^jeited w Pull Spike. 


^^ of Spik.. 


a™^. 


«.^»™, 


Minimum. 


White oak: 
Common spike 


5 
5 


Partially seasoned 


Foui>d>. 
6,950 
13,026 


Pound*. 

7,870 
14,940 


Poonds. 

6,160 










5 

8 








Oak (probably red); 

Common spike 

Screw spike 


4,342 

11,240 

2.61 


5,300 
13,530 




.do 












28 
2» 


S.„o»d 






Loblolly pine: 
Common spike 


3,670 

7,748 
2.11 


6,000 
14,680 


2,320 
4,170 










12 

U 








Hardy CBtalpa: 
Common spike 


3,224 
8.261 
2.56 


4,000 
9,440 








Ratio 








11 
11 


Green 






Common catalpa: 

Common spike 


2,887 
6.939 
2 42 


4,500 
8,340 
















4 
5 








Cbeatnut: 
Common spike 


2.980 
9,418 
3-15 


3,220 
11,150 


2,600 

7,470 



















>, Google 




STEEL RAILS 



i-%^ 




m 



Fia. 106. — Common Spike. 



Fia. 107, — Common Screw Spike, 

Tables XXV and XXVI are taken from "Studies of the Stability of Rail- 
way Tracks," by Jules Michel,* and ^ve the holding power of hook and screw 
spikes. 

T.4BLE XXV. — PULLING FORCE NECESSARY TO PULL OUT FOR 0.20 INCH IN A 
HOOK SPIKE AND A SCREW SPIKE BURIED 4.13 INCHES IN THE WOOD 







P.L.U. Hook Spike. 


P.L.M. Screw Bpike. 




PoUDds 

992 
1,598 


Pound. 














6,732 
3,968 
5,688 


9,480 






American cypreM... 




6,063 



Figs. 108 and 109 present examples of early screw fastenings. 



Fio. 108. — Screw Spike used by Grand Duchy of Baden Slat« Railways (1860). 
* Revue G<Sn6raIe des Chemim de Fer, July, 1884, and June, 1893. 



,y Google 



SUPPORTS OF THE RAIL 
NORTHERN 




Fio. 109. — Early French Screw Spife 



TABLE XXVI. — FORCES NECESSARY FOR EXTRACTING BY 0.20-INCH SCREW 
SPIKES 0.79 INCH AND 0.91 INCH IN DIAMETER WITH THREADS OF 0.39-INCH 
AND a59-lNCH PITCH, SUNK 4.14 INCHES IN WOOD OF VARIOUS SPECIES AND 
AGES 

(Jules Hichft) 



Dmtaof 




".■sf 


n?." i ^^- 


Beech. 


0»k. 


TS 


Rwnork.. 




iBDhH. 


iKtUB. 


Pound.. 1 Pounds. 


PouiKls. 


PcHiids. 








Ne 


w Wood Cdrrbntlv Employ 


ED IN Tr 


CKS 






0.79 

0.T9 
0.79 

079 
79 
0.79 
91 


039 

0,39 
39 

0.49 
0.59 
0.49 
0.49 


5,733 




9,481 
10,143 


9.923 

10,5&4 
11,576 

12.844 
12.458 


0.59 

0,59 

0.55 

0.55 
0.5S 








July. 1884. 




\ 


Rolled screw spikes, 

24 trials. 


















10 trials. 




7.640 


11,378 






1889 


13.010 


12,348 


67 




1 










Wood Having Been 9 Ybari 


IN Trac 








0,79 

0.79 


0.3« 
0.49 






10,143 

11,576 


0.55 
0.55 






















Fig. 110 shows a machine used on the Atchison, Topeka and Santa Fe 
for preparing ties for screw spikes. Wooden dowels as shown in Fig. Ill are 
screwed into the ties.* Table XXVTI gives the cost of equipping a mile of track 
with screw spikes, the estimate being based on work actually done on a section 
of track five miles in length on the Illinois division of the railway. 

■ Railroad Age Gazette, December 24, 1909. 



,y Google 



STEEL RAILS 



Fia. 110. — Machine Preparing Ties tor Screw Spikes. (Railroad Age Gazette.) 




Hole Tapped Plug in Place Dreased for Tie Plate Completed RaU 

Faatening 

Fig. 111. — Showing Application of Screw Spike on A. T. & 8. Fe R. R. (RaUroad Age Gacette.) 

TABLE XXVIL — ONE MILE OF TRACK WITH SCREW SPIKES AND DOWELS 

12,000 spikes at 2.7 centa each S 324 

fi,000 tie plates at 21 cents each i;!60 

Boring ties for, and driving, 24,000 dowels, at 1 cent each 240 

24,000 wooden dowela at IS cents each 360 

Driving screw spikes (per mile) 150 

Total *2,334 

ONE MILE WITH CUT SPIKES 

12,000 spikes 8 127 

6,000 tie plates at 21 cents each 1,260 

Driving cut spikes (per mile) 150 

Total 11,637 

* Apparently the French railways were about the first in Europe to begin 
the use of the screw spike (tirrfond) as a rail fastening, and it is to-day uni- 
versally employed by the large systems (Fig. 112, Table XXVIII). 

• For a very full discuswon of the subject, see " The Question of Screw Fastenings to Secure Rails 
to Tiaa," W. C. Cuehing, Proceedings Am, Ry. Eng. & M. of W. Assn., 1909, Vol. 10, Part 2, p. 1456. 



,y Google 



SUPPORTS OF THE RAIL 





Flo. 112. — French Railways — Roil Fastenings. (Am. Ry. Eng. Assn.) 



>, Google 



STEEL SAILS 



TABLE XXVIII. 


— FRENCH RAILWAYS — RAIL 

(Am. Ry. Ea«. A«m.) 


FASTENINGS 




Ralliray. 


Nam'bHUd 
PoaJtion of 


lO;^^, 


SafwraukM. 


NumtMrol 


JoinU. 


Hhot. 


Hi]«a 


ii 


1.. 




III 


III 


a™„., 


Bolu. 


fD«IV«iLyon« 

i la HediUcnnte 

D'Ontana 


3083 
ZJSD 


liaaidsandl 
oauida 

ZiDHdeaodl 
2 inrida and 1 

S3 

Soataide 


T. ■ 
•B.H. 

T. 

■B.H.4T. 
Inohain 

T. 

T. 

•B.H. 


11. 02 


G.tl 

S.S2 

».73 
4.73 

.1. 


0.87 
O.W 

O.SI 


o.u 
on 


o.» 
o.n 

o.« 

o.u 

0.31 


lltoHpatW 

17p«-3».87- 
UpatM' 

HtoltpvM' 


Square «^ 


' 


D*L'OiMrt 


SquaraiiBil 












D.L-6U. 





in o«k or bMch, tl 



TABLE XXIX. — GERMAN RAILWAYS — RAIL FASTENINGS 





Kind of Tie 
Plata. 


as.* 

HcivK Spikn 
ofxie.'' 




Screw Spikes. 1 


loiDU. 


Nam* of Stat* 
RmUwsy. 


P 


1=1 


llll 

3I |°" 


ill 


tL-s 


'sr 


No. 
B^Ita. 




Iiarudeaad2 

l?li)k"»^ik« 
]ia»de«id 

C» book an 

1 inaids and 1 
outaida 

2oiiUideaBd 
linrids 


T. 
T. 

?:' 
Si.. 

»1T. 


Hirt 


07» ' OSB 


0.3» 

O.SB 
iuiSl 

0J9 




Su^HdKl 

Suapa«l«l 

aiumdad 

3<|uareBnd 
S,up«d«t 


J 




Hook ouUide 
Hook plate. 
Hook inaide 

Hook pEaie. 

".'i-ffi'-"- 

Hc»k plats. 


pikaat 
87 


O.M 

M 

Bava 
O.M 










BavarUoO) 

BmdeiWl 

Pn^anO) 


and 
HIS 

aUHSDC 


and 
fl.M 

B.S 
5»l 




lands 
• 



12) l-'untHl ties aim 

(3) In tunnels uw'caw 
" Hook apikea an no 



jflively- U» voodea ties on bridE« wit 



tl with Midlaad Ry., Enfland. 



>, Google 



SUPPORTS OF THE RAIL 



145 



The German r^lways did not adopt this style of fastening as early or as 
generally as those of France, and the use of the hook spike is quite widespread. 
In 1899, the general employment of the screw spike on all Knes of the system 
was prescribed for the Prussian Government Railways (Fig. 113, Table XXIX). 





Fta. 113. — German Railways — Rail Fastenings- {Am. Ry. Bng. Ahso.) 

Same design of Screw Spike used by the Wiirttcmberg, Bavarian and Baden State Railways a 

shown for the Saxon. 



>, Google 



STEEL RAII5 




2 1 

I r 



^ /n— r- 



■ f punoj vqat J«uj»3 



«d 



II 






nr 



i r 

The common hook spike used in the United States has been oftrai severely 
condemned by writers in the technical press, and the readers have been usually 
led to infer that it is employed everywhere in Europe, which is seen from the 
above not to be the case. Indeed, the screw spike in Great Briton is almost 
as rare as it is in the United States, at least on the large systems, tiie only one 



,y Google 



SUPPORTS OF THE RAIL 



147 



makiiig use of it being the London and North Western, and that only partially 
(Fig. 114, Table XXX and Plate XIII). 

All of the large r^lway systems in Great Britain use the double-head rail, 
held in position in large cast-iron chairs by "wedges, and consequently the fasten- 
ings are for securing the chairs to the ties. For the purpose of fastening the 
chairs to the ties, the almost universal plan is to use two iron or steel spikes and 
two wooden trenails. The spikes are not pointed, and are driven into previously 
bored holes. Instead of the trenjuls, the London and North Western Railway 
makes use of two screw spikes, which resemble those of the Belgian State 
Railways. 



TABLE 


XXX. — ENGLISH AND SCOTCH RAILWAYS — 

(An. Ry. Ei«. Ann.) 


RAIL FASTENINGS 




Number aiHlKiiirtol 
FnsIMiiiEa par Chair. 


SpikB. 


Tre«ib. 


RuLwiy. 


Ijwgth Diam. 1 Diam. 


ToUl 
L«,tb. 


L»,th 
TopC™ 


■a. Is- a: 


ilii 


•2 unkat and 2 tmsilg 

■2iipik»udZlt«B.il> 

*2>pili«iud2trui«llK 

12 Iwisud ipik«ud 2Ir«iii1i 

11-inch bofoithrou,h.i« 

12 ipikH vd 2 «nw> 


i ! 1 i 1 


lacha. 


T 


iHhu. l.ich«. , iDchu. 

11 !l 1 li' 


'i* 




"1 


1 


-J— 

2A 


ii 


ri 


I«d«*Xo«h 


« 


11 


tl 


















1 



* Hu ihort milHES ol chairx secured by two j-lnch bolte throu 
t IlJiutiBWd by Tu. I ol Kw. 114. 
1 IIIoiimMd by Fit. 2 o' Fig. IK. 
i Illonnlsd by Put. a of Fin. IK. 
i IllulntRl by FU. 4 of Fig. 114. 

On th* Forth Bridie the North Britiih Ry. asm flat'bottoni 
■ The twined epika are to be nbuidtHied lor plain onea. 




DiomBtec of Sale iwjuinid fa- thia kth II 

— Screw Spike deduced from European Practice. [Cushing.) 



>, Google 



148 



STEEL RAILS 



Fig. 115 shows a deagn of screw spike deduced by Mr. Gushing from 
European practice. The form of the thread seems to have httle influence upon 
the holding power of the screw spike. Table XXXI gives the resistance for 
threads of right-angle form and those of isosceles triangular form. 

TABLE XXXI 

RESISTANCE OF SCREW SPIKES HAVING DIFFERENT THREADS 
IJuIn Uiohel — Revua G^irfislB dte Cbemiu da F«. Juh. 18MJ 



DidMuioM ol Scraw Splku. 


1hw»I« Tbnad. 






Screw Spikes 0.79 inch id diameter. New tie. 


Pitch of 0.39 inch 
Pitch of 0.59 inch 
Pitch of 0.49 inch 


11,887 pounds 
12,458 pounds 
13,010 pounds 


12.039 pounds 
12,513 pounds 
13,561 pounds 


Average of 4 tri&is. 
Average of 4 trials. 
Average of 4 trials. 


Screw Spikes 0.91 inch in diameter. 


Pitch of 0.49 inch 


13,424 pounds 1 13,424 pounds 


New tie. 


Screw Spike 0.79 inch in diameter. 


Pitch of 0.39 inch 
Pitch of 0.49 inch 


10,253 pounds 


0,923 pounds 
11,576 pounds 


Ties 9 years in service. 
Ties 9 years in service. 


Screw Spikes 0.91 inch in diameter. 



Pitch of 0.49 inch 



11,246 pounds 



5 pounds 



Ties 9 years in service. 



The proof that the screw spike is not a thoroughly efficient rail fastening 
lies in the devices which have been invented to assist it in its work, —the square 
plug, the Collet trenail, the Thiollier helical lining, and the Lakhovsky screw 
and case.* 

The mmn objection to the Collet trenail is its size; it is illustrated in Figs. 
116, 117, and 118 with a screw spike and the wooden plug commonly used on 
French rmlways for repairing old holes. The difference in size is large, the 
Collet trenail being 1^ inches in diameter outside the threads. This cuts away 
a considerable portion of the critical part of a tie, and is considered by many 
engineers to weaken the tie too much. The plug is only about an inch square. 
Nevertheless this screw dowel is largely used in Germany. 

The Collet ti-enail has been tested from its inception by the Chemins de 
Fer de I'Est, but the square plug illustrated in Fig. 117 is preferred. The wooden 
screw, often made of elm, cannot be put in place without removing the tie from 

• Lakhovsky trenail. Revue Gen&ale des Chemins de Fer. Paris, 1909, Vol. XXXIII, 
pp. 324-327. 



,y Google 



SUPPORTS OF THE RAIL 



Cnernm' de Fe.r 



V 



Chemin de Fer 
de L'Est 



"^ScrewSpiKe 



Square Wooden Plug 
used for repairing old holes. 



Pi«. 116. — French Screw Spike. 
(Am. lly. Eng. Aaan.) 



Fiu. 117. — Wooden Tic Plug used on French Railways. 
(Am. Ky. Eng. Asan.) 



>, Google 



150 STEEL RAILS 

the track, and it frequently splits. The ties on the Chemins de Fer de I'Est 
are principally oak and beech. Pigs. 119 and 120 illustrate pine ties with 
dowels in place. 

The diagrams of Figs. 121 and 
122 give the comparative resistance 
to vertical" pressure of screw spikes 
with and without dowels. 

The Thiollier steel helical hning 
is being experimented with as a sub- 
stitute for the Collet trenail, and the 
Lakhovsky screw and steel casing 
(Bulletin of the International Railway 
Congress, March, 1907) are considered 
worth trying by the Chemins de Fer 
de I'Est, de I'Etat, and de Paris a 
Orleans. 

From its greater holding power, 
the verdict of the engineers of the 
French, Belgian, and German railways 
is that the screw spike is superior to 
the hook spike, because they consider 
it very important to hold the rail fast 
to the tie. 

On the other hand, the Briti^ 
railways do not seem to find the screw 
spike necessary for their large and 
heavy chars, and they use creosoted 
ties, as well as the Continental lines; 
but the holes for their spikes are bored 
in advance. 

According to our present knowl- 
p 1 75 - M edge, the amount of bearing surface 

_ - . the tie plate has upon the tie is ap- 

iio. 118. — CoUetTrenaU. (Am. Ry. Eng. Aaan,) parcntly not the determining factor 
in providing against wear. The 
question of securing the plate firmly to the tie is fully as important as the size 
of the plate used, and in selecting a proper unit stress for the bearing on the 
tie it is evident, therefore, that the area of the bearing surface cannot be eon- 



,y Google 



SUPPORTS OF THE RAIL 



sid^ied without taking account of the kind of fastening employed to hold the 
plate to the tie. 



Pia. 1 19. — Croaa Sectioa of IHne Tie through Dowel. (Bureau of Forestry, Bulletin No. 50.) 



Fia. 120. — Three Ties of Baltic Pine on the PniMian State Railways, Berlin, showing the manner in 
wliicb screw dowels appear in the tie when ready to be shipped. (Bureau of Forestry, Bulletin No. 60.) 



>, Google 



STEEL EAII^ 



'"'"^TIES 


"" -_::_::::::: :::::::::nq^;qiTi:.: 


















g Jill U ^ J 


















tt 6000 a' !>' ^tr 


















■••-['' .Ui'JJii,i.''liiL!F>JillliiJiiiili 



— Comparative Reustance to Vertical Presaure of Screw Spikes io Pine Ties, Old and New, 
with and without E>owele. (Bureau of Forestry, Bulletin No. 50.) 



ou, . :. BEECH TIES 


,„,. i~z~~z 




„„,i 1 1 1 1 ' rn 


Z:::-:::::::::-±-V:/:-;::-: 




































,.„, 1 / 


'71-'--' — ^ ^t^"" "^ 











Oi .04 .06 . 

VGRTICAL OSPLACeMENT- INCHES 

- Comparative ResiHtance to Vertical Pressure of Screw Spikes in Beech Ties, Old and New, 
with and without Doweb. (Bureau of Forestry, Bulletin No. 50,) 



,y Google 



SUPPORTS OF THE RAIL 153 

In the case of a white-oak tie, where the spike holds well and the life of the 
tie is comparatively short, the ordinary working stress of the timber to resist 
crushing at right angles to the grain may probably be safely taken in propor- 
tioning the strength of the tie. With soft woods, however, which offer less 
resistance to the spike pulling loose, and which, when treated, possess long life, 
the ordinary working stress of the wood has little application to the bearing under 
the tie plate unless some means are used to secure the plate firmly to the tie. 

As will be seen in the discussion of the Supporting Power of the Tie (Article 
19), one of the weakest points in the support of the rail lies at the bearing of 
the tie plates on soft-wood ties, even when the normal crushing value of the 
wood is taken as is done in the calculations. It is thus of considerable impor- 
tance that with a soft-wood tie a more secure fastening than the ordinary spike 
be used to hold the tie plate firmly to the tie. 

With the increase in density of traffic there has developed a growing 
tendency for the rail to creep or move in the direction in which the traffic 
moves. On account of the joint ties being spiked through slotted holes in the 
joint, these ties move with the rail, with the result that correct spacing of the 
adjacent ties is not mdntained. 

To overcome this difficulty there have been devised numerous devices 
for anchoring the rails to the ties. These are generally fastened to the base 
of the rail and bear against the side of the tie; when employed in sufficient 
numbers they are fairly efficient in preventing the movement of the rail.* 

16. Strength of the Tie 
Assuming the tie to be in good condition jmd free from decay, we have 
now to determine the strength of the wood of which it is composed. Let us 
first examine the kinds of woods used in the United States. 

• Some recent literature on this subject is as followa: 

KuNZE, W. — Das schienenwandera, ursache uDd abhilfc. 2,500 w. 111. 1909. (In Glasere 
annalen fUr geweriw und baiiwesen, Vol. 65, p. 122.) 

Considers cause o( creeping in rails and devices for itit prevention. 

ScHLCsSEL, L. — On the working loose of screws when used as rail faatenings, 21 p. III. 
1907. <ln Bulletin of the International Railway ConRress, Vol. 21, p. 3-) 

Concludes that we<lge fastenings should be substituted for screw fastenings. 

Tex. K. dex, — Creeping of raib in the direction of the trains. 800 w. lU. 1911. (In 
Bulletin of the International Railway Congress, Vol, 25, p. 292.) 

Use of rai'l anchors. 2,000 w. III. 1911. (In Railway Apt Gaiette, Vol. 51, p, 125.) 

Considers tendencies in the creeping of rails and forma of anchors most successful in over- 
coming it. 

WiBTH, Alfred. — Die schienenwanderung und ihre verhlitung. 10,000 w. 1909. (In Zeit- 
pchrift des Osterreichischen Ingenieur — und Archileklen — Vereines, Vol. 61, p- 317, 333.) 

Discussion of rail creeping at some leogtb, considering theory and prevention by rail-faateuiag 
devices. 



>, Google 



154 



STEEL RAILS 



The following statements are based on the number of ties bought rather 
than on the number actually used. For all practical purposes, however, the 
two are identical, because the purchases in twelve months are an accurate index 
of consumption for a corresponding period. 

Table XXXII shows the number and value of the different kinds of ties 
purchased by the steam and street railroad lines in the United States in 1906, 
and contrasts the purchases of steam railroad companies in 1905 and 1906. 

TABLE XXXII. — NUMBER AND VALUE OF TIES PURCHASED BY STEAM AND 
STREET RAILROADS IX THE UNITED STATES IN 1905 AND 1900 

(Forest Ssrvice, CircsUr lU) 





Suain 






SMtm nilr«d>. i»0«. 


S«e. railed.. I««.- 




Numbw. 


Value. 




Number. 


VelM. 


Avw 


Naoibet. 


V»lu«. 


Aver- 
se 


Oaks 

South'mpincet 

Cedar 

Douglaa fir. . . 
Cheatnut.... 

Cypress 

Western pine. 
Tamarack., . 

Hemlock 

Redwood 

Lodgepole pine 
White pine . 
All others 


34.677,304 

18,351,037 
6,962,827 
3,633,276 
4,717,604 
3,483.746 

3,060,082 

1,713,000 
590,852 

(t) 

(t) 
791,409 


(19,072,517 

7.707,436 
3,033,644 
1,198,981 
2,264,450 
1,149,636 


t0.S5 

.42 
.44 
.33 
,48 
.33 


41,532,629 

17,538,090 

6.416.867 

6,706,222 

4,646,763 

4,988,685 

3,909,500 

2,430,236 

2,037.002 

725.346 

553.838 

258.030 

1,734,517 


(21,256,518 

8,905.009 
3,044.446 

2,782,967 

2,132,984 

1,813,500 

1,673,359 

837,217 

576,806 

248,814 

210,458 

76,833 

661,601 


(0.61 
51 

.47 
41 

.46 

43 

.34 
.28 
.34 
.38 
.30 


3,825,245 

1,303,120 

666,575 

542,340 

1,942,212 

115,911 

60,105 

146,623 

21,196 

523,283 

900 

115,357 

93,550 


12,021,534 

662,736 
265,670 
227,426 
862,958 
48,635 
24,668 
52,344 
6,072 
287.328 
360 
74,219 
64,643 


(0 63 
.61 
.40 
.42 
.44 
.42 


1,101,630 
565,320 
118,170 


.36 
.33 
.20 


.36 
.29 
.66 






■fll 


343,662 


.43 


09 


Total, , , 


77,981,227 


S36,585,446 


10.47 


93,477,625 


J44,220,532 


SO. 47 


9,356,417 


$4,598,692 


SO 49 



The purchases of ties reported by the steam railroad companies in 1906 
exceeded those of 1905 by more than 15,000,000. Nearly one-half of this 
excess was oak. The purchases of cedar ties showed a decrease of about one- 
half million, due possibly to the sharp demand for cedar poles, which operated 
against the production of ties. Douglas fir ties nearly doubled in quantity, 
and both cypress and hemlock increased by a large percentage, but tamarack 
purchases fell off more than one-fifth and chestnut about 1.5 per cent. 

Oak, the chief wood used for ties, furnishes more than 44 per cent, nearly 
one-half of the whole number, while the southern pines, which rank second, 
contribute about one-sixth. Douglas fir and cedar, the next two, with approxi- 
mately equal quantities, supply less than one-fifteenth apiece. Chestnut, 
cypress, western pine, tamarack, hemlock, and redwood are all of importance, 
but no one of them furnishes more than a small proportion. 



,y Google 



SUPPORTS OF THE RAIL 155 

Table XXXIII shows, by kinds, the number and cost of the cross-ties pur- 
chased, by stBMn and electric railroads in the United States in 1907. 

Table XXXIV gives a comparative statement showing the niunber of 
cross-ties purchased by the steam and electric raitoads during the years 1910, 

1909, 1908, and 1907. 

Of the total purchases of cross-ties during 1910, 139,596,000, or 94.2 per 
cent, were made by steam railroads, while electric railroads purchased 8,635,000, 
or 5.8 per cent. The steady increase in the number of cross-ties reported as 
purchased for new track is noteworthy. The total for this purpose in 1910 was 
22,255,000, as against 16,437,000 in 1909, 7,431,000 in 1908, and 23,557,000 in 
1907; the total for 1910 exceeding that for 1909 by 35.4 per cent, for 1908 by 
199.5 per cent, and nearly equaling that for 1907, the largest ever recorded. 
Largely as a logical result of the greater demand for cross-ties during 1910, the 
average cost per tie at point of purchase advanced to 51 cents, the same figure 
reached in 1907, as compared with 49 cents in 1909 and 50 cents in 1908. 

In 1910, as in preceding years, oak was the principal kind of wood used for 
cross-ties. The number of oak cross-ties formed 46.1 per cent of the total for 

1910, as compared with 46.2 per cent in 1909, 42.8 per cent in 1908, and 40.2 
per cent in 1907. 

A substantial increase in 1910 over 1909 is shown in the number of southern 
pine cross-ties reported; the increase in the cut from this species over 1909 
being 22.8 per cent, as against an increase of 20 per cent in the total number of 
cross-ties reported from all woods. Douglas fir also showed for 1910 over the 
preceding year a larger increase, namely, 28.2 per cent, than the increase in the 
total purchase from all woods. On the other hand, chestnut, cedar, and cypress, 
with increases over 1909 of 17.1 per cent, 7.8 per cent, and 17.6 per cent, 
respectively, were bought in relatively smaller quantities. 

While tJie bulk of the cross-ties were cut from the six woods mentioned dur- 
ing each of the four years and while combined they contributed 85.5 per cent of 
the total in 1910, 85.3 per cent in 1909, 86.5 per cent in 1908, and 87.2 per cent 
in 1907, a remarkable and significant showing in connection with the figures for 
1910 is noted with reference to certdn woods which hitherto have been utilized 
as cross-tie material to only a very limited extent. The increase in the number 
of cross-ties over 1909, reported as cut from elm, was 451.7 per cent; gum, 328.8 
percent; birch, 323.3 percent; spruce, 121.5 per cent; and mesquite, 114.9 per 
cent. A very large percentage of the cross-ties cut from these woods were given 
some preservative treatment, thus increaang then- life to or beyond that of 
imtreated cross-ties made from the more commonly used or standard cross-tie 



,y Google 



156 STEEL RAILS 

TABLE XXXIII. — CROSS-TIES PURCHASED BY STEAM AND ELECTRIC ROADS OF 
THE UNITED STATES IN 1907 ^ 

IBiuBu o[ Chs Cnnis, Fomt ProdueU Mo. S) 













Kind. 




Hewed. 


Siwed. 


Number. 


ToldCMl. 




Number. 


Total Coet. 




Number. 


Toial Cow, 




Total... 


153,699,620 


$78,958,695 


(0.51 


112,309,246 


$56,522,768 


tO-50 


31,776,434 


117,020,882 


10.54 


Oake 

Southern pines 
Douglas fir... 

Cedar 

Chestnut 

Western' pine!! 

Hemlock...!! 

Redwood 

Lodgepole pine 
White pine. . . . 
Another 


61,757,418 
34.216,081 
14,524,266 
8.953,2fti 
7,851,325 

6,778,944 
5,019,247 
4,562,190 
2,366,459 

2,030,952 
666,916 
474,455 

4,499,132 


32.9S5,122 
18,434,198 
6,818,869 
4,473,960 
3,772,048 

3,099,430 

2.515.798 

2,254,617 

807,241 

1,198.497 
332,984 
193,606 

2,072,316 


0.53 
0.54 
0.47 

0.50 
0.48 

o.« 

0.50 
40 
0.34 

0.59 
0.50 
0.41 
0.46 


51,169,478 
25,629,749 
1,436,258 
7,941,1,^2 
4,922,831 

5,695,640 
3,206,754 
4,144,127 
2,283,675 

884.552 

666.916 

289,624 

4,038,490 


26,774,251 
13,100,589 
590,754 
3,987,035 
2,337,697 

2,552,381 

1.575,457 

2.083,646 

770,969 

507,154 

332,984 

106,528 

1,802,323 


0.52 
0.51 
0.41 
0.50 
047 

0.45 
49 

0.50 
0.34 

0.57 
0.50 
0.37 
0.44 


6,929,572 

7,415,686 

12,366,640 

396,891 

889,420 

884,915 

1,626.330 

340.618 

79,256 

406,519 


4,033,150 

4,669,060 

5,884,822 

190,322 

426,523 

453,058 
835,895 
137,481 
34,796 

224,525 


0.58 

o!48 
0,48 
0.48 

0.61 
0.61 
0,40 
0.44 

0.55 


131,671 
308,916 


53,041 
178,209 


40 
0.58 



TABLE XXXIV. — CROSS-TIES PURCHASED BY STEAM AND ELECTRIC ROADS OF 
THE UNITED STATES DURING THE YEARS 1910, 1909, 1908, AND 1907 

IBuieau of tbe Ceuua) 



Kind ot Wood. 



Oak 

Soul hem pine... 

ch^imi '.'.:'-'.: 

Cedar 

C>-pre« 

Tumerack .... 

Ile-ntock ,,.'.'! 

Redo'ood 

Gum 

Beech 

gJS" ■:;:::;: 
tepi^e-::; 

Birch 

{fB^vfaii^lTiShit 



3,7-3.000 
3.m.<K0 



4,328,000 I 2,007.000 



>, Google 



SUPPORTS OF THE RAIL 157 

woods. The growing scarcity of these last-mentioned woods, however, tends 
to increase their cost and accounts largely for the introduction of substitutes 
cut from cheaper species. The drift in this direction is clearly brought out by 
a comparison of the figures relating to treated cross-ties during the past four 
years. In 1907 the number of cro&s-ties reported as having been given some 
preservative treatment was 19,856,000; in 1908, 23,776,000; in 1909, 22,033,00; 
and in 1910, 30,544,000; the number for 1910 showing an increase over that for 
the preceding year of 8,511,000, or nearly 39 per cent. 

The question of tie preservation is becoming more and more important as 
the donand for tie matmai Increases and the traffic requirements become more 
eicacting. So long as plenty of white-oak ties could be secured, the necessity 
for tie preservation was not felt; but with the constantly increasing use of pine 
and other less decay-resistant woods, it has become a vital economic question. 
The r^Iroad companies have met the problem by establishing treating plants 
in various parts of the United States and by laying experimental tracks with 
treated ties to determine the efficiency of the several preservatives xmder vary- 
ing conditions.* 

Table XXXV, prepared by the Forest S«Tice,t gives the results of an 
elaborate series of tests upon the strength of treated and untreated pine ties. 

In outlining the plan for these tests two divisions were made, dealing 
respectively with the rflect on the strength of timber of the preliminary proc- 
esses of steaming, superheating, vacuum, etc., commonly employed in the 
preservation of wood, and the effect of the preserving materials themselves. 
The tests were confined to sapwood, and were made on small pieces taken from 
the tie, and also on full-sized ties. 

The effect of the preliminary processes was detrarained on both green and 
seasoned timber. Both green and seasoned timber were also used in deter- 
mining the effect of preservatives. The preservative fluids included only 
creosote t and zinc chloride. 

The material for the experiments was railroad ties II feet long. One 
8-foot section of each tie was put through the particular treatment, and the 
untreated section, 3 feet long, was used for control test pieces. 

Prom each tie 12 pieces were taken, 4 from the control section and 8 from 
the treated section. All of these pieces were 2 inches by 2 inches in cross section 
and 36 inches long, with one side parallel to the direction of the annual rings 

• Eicperiments with Railw;iy Cross-ties, Forest Service, Circular 146. 

t Experimente on the Streogth o( Treated Timber, Forest Service, Circular 39, by W. K. Halt. 

X The treatment was essentially the "Ruepinf;" procesa, although this name is not used in the 

circular. 



>, Google 



STEEL RAII5 



TABLE XXXV. — EXPERIMENTS ON THE STRENGTH OF TREATED TIMBER 



!) THE STKENQTH OF GREEN 
2 isobea: Bir4lri«d Mare Usta] 





CyliidxCooditkHL 


Slnngtli. 


.?& 


UoiRiin, 


GravLli- 

(dry). 




SMuniiic. 


II 


Sutio. 


^l. 


1 


1 




Trntmnt. 


1 


1 


1 


1 


if 

■ 2 


1 


t 

i 








■" 


r 


6^ 


w 


* 


- 








Lba. 




Lbe. 


Per 


Ftt 


Pw 


Per 








sq. u. 




























UntreaMd oood -100 puonC 








(* 


•a) 


257 




92.8 93-l| 93.81 93.2 


7 5 


1,5 


13. ^ 


13.4 


l.fW 


^Md 


Steam, at variotu presaures 


V 


40 


2fi7 




99.8 104. 3I102.OI1O2.O 
94.6 99.0107.7:100.4 


7.5 
B.O 


7.0 
fi.O 


12. V 


12.2 
12.4 


.553 
525 


.571 




(4 


50 


•m 




94.5 96.4|103.5l 93.1 






13 7 


in, 5 






Creosote, injected at 160°F, 
























under a pressure of 100 






























pounds per square inch. . . 


4 


■JO 


2flh 


25 4 


81. « 


79 1 


102 4 


w.t: 


1,5 


i.O 


13.2 




,.1.V 




Zinc chloride; 






























2.5 per cent solution. . . 


4 


20 


?4fl 




fi7 4 


92 f 


113, fl 


m.r, 






12. H 


\H.7 


.V-U 




3.5 per cent solution... 


4 


20 


24fi 




97,4 


9.1,2 


92.7 


%.i 


7,5 


7,n 


13.1 


13.5 


5;«. 


,.'vW 


5.0 per cent aolutioa. . . 


4 


a) 


?4fi 




99 « 


96 7 


7H,9 


tl !■ 


% 5 


>,0 


12 a 


13 4 


511 




10.0 per cent solution. . . 


4 


20 


255 




100,1 


100.4 


74.8 


91.8 


8.5 


8.5 


13,2 


13.5 


,682 


.612 



,1. CSAUCTEUBnca ai 



I AVKBAQB BTSEHdTH 



.•DUEo Untbeaiid Wood 



Moisture per cent 13,4 

Weight per cubic foot (dry) pounds 33,6 

RiDRS per inch 7 

Modulus <rf elasticity pounds per square inch 1,893,000 

Bending strength at elastic limit " " " 6,998 

Bending strength at rupture " " " 13,198 

Compression strength parallel to grain " " " 6,555 

Compression strength at right angles to grain " " " 834 

Sbe&iiag etreogtb radial to grain " " " l,li5 



,, Google 



SUPPORTS OF THE RAIL 



TABLE XXXV. — Cmtinued 



InlJ-riwd tiea: M«ao»d. trwMd, ud rawioDsd bataia Ustal 





Cylii»letCo«litiou. 


Slnaith (Malic). 


SinkoPnitiix. 


J 








B«Hlld( 


Co. 


1. 


Pull Soik*. 


1 










« 


1 


























1 


1 


H 


ii 


S 

P 


1! 

< 


'1 

■< 


1 


1 


1 


i 






LtM. 


•F. 
















^ 






^^. 
















rit. 










Untnxsd wood- 100 iw 










A 


10 


237 


M.2 


79.3 


91 1 


89.9 


118.51110.7 


4.9 38.0 




A 


ai 


2SH 


«:i.7 


78-4 


99.1 


90.4 


103.6109.4 


6.2 37 3 


Steam, at Tortous pressures 


■ A 


30 


W4 


S7 H 


Ki.4 


92.7 


88.0 


100. l' 96.5 


5.3 37.9 




A 


40 










80.4 


93.0 77, t 






A 


50 


MS 


m.1 


fiO.fi 


74.4 


68-0 


RO.^ 


70. ;i 


4.8 36.1 




■f 








81.9 


87.1 




97. t 


93.7 




St«ain, for various periods 




20 


258 


93.7 


78.4 


99.1 


90.4 


103 ( 


109 4 


5.2 37.3 


' f 


20 




S7.5 


78. a 


92.0 


86. 1 


8;i.f 


79. r 


4.6: 36.7 






















4.6 36.7 


Zinc chloride. 2.5 per cent boIu- 
























tion 


< 






74.7 


05.1 












































257 


69.5 


61.2 


flO.l 


63-6 


68.2 


68.1 


4.6 













Phthi 



[. Cbiracteubttcs un AT>B4aa E 



I Ukt 



*TBD Wood 



Moisture .per cent (approximate) 

Weight per cubic foot (air-seasoned) pounds 

RinKs per inch 

Modulus of elasticity pounds per square i 

Bending strength at elastic limit 

Bending strength at rupture, 

Compression strength parallel to grain 

Compression strength at right angles to grain (rail-bearing) 

Spike pulling — common spike 

Spike pulling — screw spike 



,568.000 
3,429 
6,458 
4,452 



,y Google 



TABLE XXXV. — CorUmued 

EFFECT OF STEAU AND CREOSOTE ON THE STRENGTH OF SEASONED LOBLOLLY PINE 

ISpBeimmi, 1 by I iDebrn t«««i immadBUly riur trmtnmill 



CylindM CoBdUiou. 



Steam, at vftrious presBuree 

Creosote, iDJected at 150° F. 
under a, pressure of 100 
pounds per square inch. . . 

Soaking, wood previously 
treated with creosote in- 
jected at 150° F. under a 
pressure of 100 pouods per 
stiuare inch 



19.20 



ScBgoiisd wood - 1 N par nut 
92 6106.41124,01107,' 
70, 9i 5S,ll 72, 9I 66.; 

Stsamed wood - 100 jxtcait 

97,11102.21110,71103,; 

SaaaosKl vood - 100 permit 

81,01 7S'4| 83,0] 80,' 

Sooksd iiDtr«Bt«d wood - 



160,4 



6.0 



78,1 



Panicii. Cbabactebibiicb imd AvaBAQB Stkbnotbs or tbd Uhtbeatbd Seauhbd 

Moisture per ci 

Weight per cubic foot (dry) poui 

RinsB per inch 

Modulus of elasticity pounds per square inch 1 

Bending strength at elastic limit " " 

Bending strength at rupture " " 

Compression strength parallel to n^ia " " 

Compression strength at right an^ee to grain " " 

Shearing strength radial to grain " " 

TABLE XXXV. — Conduded 
EFFECT OF CRE080TINQ WITHOUT STEAM ON THE STRENGTH OF SEASONED LOBLOLLY PINE 







ihySuebrntteetonBd. 


mud, >nd 


•rlnch. 


mtaUl 








„.„,. 


Stnuftb (atktid). 


Rin*. 


Moiature 
Control, 


Weight (BirHwmdl. 


Tn 


«,T..- 

ol Ruptun. 




Comrol. 


Tn»tBd, 


COBttol. 


Tronted. 




Percent, 

Untreate 
100 pe 
117 


Paroeol. 

J wood- 
r cent. 


5 


7 


Percent. 

16,1 


34,79 


'-^r 











AVEBAQB STBEHmHg OF THE URTaEHTBD WooD 

Moisture per cent 

Weight per cubic toot (dry) pounds 

Rines per inch 

Modulus of elasticity pounds per square inch I, 

Bending strength at elastic limit " " " 

Bending strength at rupture " " " 

Compression strength parallel to grain " " " 

Compression strength at right angles to grain " " " 

Shearing strength radial to grain " " 



>, Google 



SUPPORTS OF THE RAIL 



161 



and the other at right angles to it. After the bending tests had been made 
on these pieces, smaller pieces, 2 Inches by 2 inches in cross section and 4 inches 
long, wesre cut from their ends and used for the compression and shearing tests. 
In any tie the test pieces were taken out according to Fig. 123, variation being 
allowed only to secure clear pieces. 

The test pieces from each tie were marked consecutively from 1 to 12. 
The untreated pieces, marked 1 and 2, were used for control-impact tests, and 
those marked 3 and 4 for control-static tests. The treated pieces, marked 5 
and 6, were used for impact tests; those marked 7 and 8 for static tests. The 











coNTnot. 


_._! 




/ 






/ 


.r 


9 


^ 




2 




2 


6 


10 




3 






3 


7 


'I 





9 






IZ 




i& 






f 







Flo. 123. — Control Plan — Creoeote Tie 'l%HtB. 

treated pieces, marked from 9 to 12, were similarly tested, but were resoaked, 
if necessary, to bring them back to the degree of moisture found in the control 
pieces. Ordinarily the steaming process did not decrease the moisture content 
of the wood, in which case tests on resoaked pieces were not required. 

In addition to Uie tests on small pieces, the strength of full-sized ties in 
bending and in compression, both parallel and at right angles to grain, was 
obt^ned, as well as the capacity of the wood to hold a spike. The ties used 
were 8 feet long. The entire tie was treated and afterwards tested in full size. 
In the bending tests under a static load, the ties were supported on a span of 
80 inches and loaded at the third points of the span. 

Short sections of tiie ties were used for tests to determine the resistance 
gainst compression parallel to grain, against compres^on at right angles to 
grain (which is similar to that produced on a tie by the base of a rail), and 
against the force withdrawing a spike. In the tests of compression at right 
angles to grain, the width of the tool equaled that of the base of an 80-pound 
A. S. C. E. rail. The force necessary to cause the yielding of the wood was 
measured. Both screw spikes and common spikes were driven into the tie, 
and tiie force necessary to pull them out directly along their length was meas- 
ured. Any common spike was driven but once, since it was found that the resist- 
ance aganst pulling diminished when the spike was redriven into new wood. 

The weight of the tie before treating, after treating, and at the time of 
test was determined. The physical characteristics of the wood, such as per 



,y Google 



162 STEEL RAILS 

cent of sap, rate of growth, shakes, knots, and moisture content, were also 
recorded. 

Impact tests were made on certain of the full-sized ties. In general, it 
was found that the influraice of the various factors may be determined by both 
static and impact tests. 

The results of these tests form a body ctf evidence from which tie fol- 
lowing general conclusions may be drawn; 

(1) A high degree of steaming is injurious to wood in strength and spike- 
holding power. The degree of steaming at which pronounced harm results 
will depend upon the quaUty of the wood and its degree of seasoning, and upon 
the pr^sure (temperature) of steam and the duration of its application. For 
loblolly pine the hmit of safety is certainly 30 pounds for 4 hours, or 20 pounds 
for 6 hours. 

(2) The presence of zinc chloride will not weaken wood under static load- 
ing, although the indications are that the wood becomes brittle under impact 
when treated with solutions above 3.5 per cent concentration. 

(3) A light treatment with creosote will not weaken wood of itself. Smce, 
apparently, it is present only in the opening of the cells, and does not 
get into the cell walls, its action can only be to retard the seasoning of 
the wood. 

The Committee on Wood Preservation of the Amaican Railway Engi- 
neering Association in its report at the March, 1910, Convention of the Asso- 
dation presented the following conclusions based on the best data available 
at the time on the strength of treated timber: 

(o) High steaming will diminish the strength rapidly. 

(6) Treating with strong solution of zinc chloride will render the timber 
brittle, perhaps because of free acid in the solution. 

(c) Creosote is inert. 

(d) Seasoned timber treated with light doses of creosote is as strong as 
the original timber. 

Tables XXXVII and XXXVIII give the results of tests of the Forest Ser- 
vice on a number of woods, and Table XXXIX shows the unit stresses recom- 
mended by the Committee on Wooden Bridges and Trestles of the American 
Railway En^neering Association. 

The great variation in strength, which is noticeable in timber of the same 
species, makes it necessary to accept with caution the result of a limited number 
of tests representing the average of the species. One of the most troublesome 
factors influencing the strength of wood is the amount of moisture in it. 



,y Google 



SUPPORTS OP THE RAIL 



TABLE XXXVI. — ACXX)TTNT OF TEST MATERIAL USED IN TABLE XXXVII 






LMBlitlM aad NumtMr ol TiMB han EwA. 



StraRlearptH 



(P&Mraii'lUMJ 

^racapiBa 

Bald eyproi. 

fTuadiam diatichiim. ) 



Cow oik '. 

IQiiamu miohnn 
Rtdotk 

(Qnereai mbrs.) 



V^w^t* 



.^usreuaduiUIB.) 
Shuhatk hickory.... 

iHkoruoviu.) 
Uoekanut hkkory. . 

(HicoriailbLl 
Wtur hickory. 

Nitmsf bickoiy . 
Paan hiek^ 
I^i^t iiiekoiy. . . . 



nhilauh 



la, ouut pliiii (B); DpFuds it); hill dlnrict IS): 

jii, unduuuu Dpi&adj it). South C&roliBA. enaat 

fibim ny.Uimmapi. low cout pjaiii (3); Louuimii. 



iSm S^ 



•t (S). 



[S)lT*iaj,law 
l«): a*or|ia, npludi (I); Sooth 

le): 



imn, uplflndi (4); HuBOUri, low 

(tuvBSt low billy nplimdi (ll);T«^^ innouuD ivi. 

ima. moiuUiBoiia plataui (R}i low ooaat plaui (HI; 

lunm. laval Hood pliuB (SI; Oaondk, hvd ca«)l 

in it): South Cirolis, low sout plamTl). 

— 1- -'— .uplaKla (S); — •■ ■'- "' •"■■ ' 




Alabsms, TmwBn Vallay (I]; 
battoni t3), 

■'-" — - ■" Vuliev H). ^.™»™. 

lippi. low jtlaiD r4j. 
. ... Vall^ [SJ; ArkaoBU. 



Akbama. 



Alabuu, ToBMM* Vallay (IJ. 
UiHiannri, low pUia (4). 

Atabanu. TeaaeaMe Vallay (1) 
bottom (3): UisabHwi low plain (1). 

Alabama. Tmaute Vallay <S); Arkaaaaa 
bottom (3|;IIL»iL«ippi, lowplaio (3). 



* TbMt two (bodd probably bo elaassd u ) 
. itobr — Th« vahMi [or ipanfle iravity b«ra ( 



luthani red tmk. Thoy were oollMtad belon tbo diminclioB wu finally di 

VKB rater to "dry" wood o( «Bl malarial, i.e.. wood cootaiBin* variabla an 
hai thanfora Bot ban Uikoa iato ukoudC. but more cirslul eiperimggU in 
•r caot u H> imall that it may be ntalaotad lor pmctual purpoaai. 



>, Google 



STEEL RAILS 



In Table XXXVII all values except those for the Southern pines have been 
referred to 12 per cent moisture, which may be said to be the lightest average 
moisture content of seasoned wood. 



TABLE XXXVIL 

SUUMARY OF 



- RESULTS OF TESTS IN BENDING — AT RUPTURE 

3TS ON THIRTY-TWO SPECIES OF AllERICAN WOODS 
lii-iMOB o( Fortuity. Ciregiar IS) 
IPoonda par aquan iKhl 









1! Pk mm Moiiture. 



?tMoSc.. 
Rodokk.'.' 



Willow oak 

8|Mniahoak 

Shubvk biekory. . . 
MockonDt hickory . . 

Watw hickory 

BittanMt hickory ... 
NutDME biekory. ... 

FaanAckory 

nnut hickory 

wGiUBlm.. 

whiiBish. .'..!!!!!!; 

SnetcufD 



13.30) 
20.700 
18.000 
19, HD 



1S.700 
IS.HO 

20400 

19.700 
17,300 
19.300 
15.«00 
18,100 

13.W0 
1740) 
H.200 
16.000 



7J0O 

s.m> 



oatatial DOI reduced [or i 



>, Google 



SUPPORTO OF THE RAIL 



TABLE XXXVII. — Con/inaed —RESULTS OF TESTS IN BENDING— AT RELATIVE 
ELASTIC LIMIT 

SUMMARY OF MECBANtCAL TESTS ON THIRTY-TWO SPECIES OF AMERICAN WOODS 

(DivMioD dI FonatiT. Ciiculu IS) 

(Pooadi pv sqiura iachl 



No. 


SpOdM. 


Nnmbw 
oT 


HilhHI 

Toil 


l-awto, 




10 Per 
tSji-i. 


T«su. 


'Mt 


oCAv«n*«. 


(Avorunol 
all Teaul. 


•^rS-s^^ 


330 

uo 

130 

1 
1 

1 
1 

s 

30 

I 


ii.m 

.000 


1.W0 

3,100 
1.100 

IS 

!.9oa 
i.tco 

s 

sjoo 

iS 

4.S0O 

s]ioo 

s 

sisoo 
siaoo 

if 


11.100 

lolsoo 

,s 

11.200 

g.too 

7.390 

S 

l.BOO 
3, NO 

uoo 

1.400 

1 

1,800 
4.000 

s 

(.400 

is 

10.100 


1 

.MO 
.900 

■ESS 

.WO 
1400 

;«oo 

.100 

.uo 

i 

.aoo 

.SCO 
.300 

z 

.200 

.200 


3 

7.700 

1 

•!200 
9,400 
S,10O 

ill 

«>00 

8.900 


P«<*ot. 

il 

4« 


P«rc«t. 
SS 

89 

M 

sa 

M 

i 

1 
1 

1 

i 

TO 
B2 


















Red-oed Id 12 Psr 






















000 

700 
700 






^^t"^' 


.B 










^^i:::-: 








jss, 












sir-r- ; 


1! 


(too 
uo 

no 


as 
















^1: 


Sffl 

S.73D.OOD 



































1 "dry" 1 






TABLE XXXVII.— Conduded—RESULTS OF TESTS IN COMPRESSION, ACROSS 
GRAIN,* AND SHEARING WITH GRAIN 

BDlfUARY OF MECHANICAL TESTS ON THIRTY-TWO SPECIES OF AMERICAN WOODS 

( Diviiioo of Foronry. areolar 15) 

IPounda per aquare inch] 



Sbaarutc 






RednoKi to It Fv cwt 



Sbnnliaf^^V.'!!!! 
Loblolly pine 

Reduced Ui 12 Per cent 



WUle cedar 

! Doocluipnicet. .. 
I Wbiteoak 

Overcupoak 

PoetiHk 



need to 12 Per cent 
rtqre. — Condudtd. 

oak 

Souiheta red oak 

Black oak 

WHlaroak 

Wiltowoak 

SWbark hle'lniry 

While hickory 

NdlDw hicitor7 -, 

Pecu hickory 

Piiiuit hickory 

While elm 

- ■irelm 



Ltarial not radncad lor m 



jOc:)g-te 



STEEL RAIIB 



TABLE XXXVin. — STRENGTH VALUES FOR STRUCTURAL TIMBERS 

(Font Sirvi», Qnski 189) 
BENDING TESTS ON GREEN MATERIAL 





SiM. 


J 

■3 


■3 


1 

i 
1 


F. S. at E. L. 


H.ofR. 


H.o(E. 


CUcalitsd 


SpHiia. 


1 




sl 




4 




ll 






J_ 


1 
1 

Z 


1 


1 


i 


II 




II 


i 




i». 


Ins. 








Lbs, 




u». 




IK? 




Lb«, 




LoDslcftf piitC- , ^ ,-,..,.. . 


12X12 
10X16 


138 

168 


i 

4 


28.6 
26.8 


9.7 

16. 7 


4099 
4193 


0.83 
,85 


6710 
6453 


0.74 
.71 


1^ 

1626 


0,99 
1,05 


261 
306 


0.86 




l.Ol 




8X16 


156 


7 


28.4 


14.6 


3147 


.64 


5439 


.60 


1368 


,89 


390 


1,29 




6X16 


132 


1 


40.3 


21.8 


4120 


.83 


6460 


.71 


1190 


.77 


378 


1.25 




ax 10 


ISO 


1 


31.0 


6.2 


3680 


.72 


6500 


,72 


1412 


.92 


175 


,58 




6X 8 


180 


2 


27.0 


8.2 


3735 


,75 


5745 


.63 


1282 


,83 


121 


,40 




2X 2 


30 


15 


33.9 


U.l 


4950 


1,00 


9070 


1,00 


1540 


1,00 


303 


1.00 


Douglas fir 


8X16 
5x 8 


180 
180 


191 

84 


31.5 
30.1 


u.o 

10.8 


3968 
3693 


,761 5983 
,71, 5178 


-72 
.63 


1517 
1533 


,95 
,96 


269 
172 


.81 




.52 




2x12 


180 


27 


35.7 


20.3 


3721 


,71: 5276 


.64 


1642 


1,03 


256 


.77 




2X10 


180 


26 


32.9 


21.6 


3160 


,60, 4699 


.57 


1593 


1,00 


189 


.57 




2X 8 


180 


29 


33.6 


17.6 


3593 


,69 5352 


.65 


1607 


1-01 


171 


.51 




2X 2 


24 


568 


30.4 


11.6 


5227 


1.00 8280 


1,00 


1597 


1.00 


333 


1.00 


Douglas fir (fire-killed} . . 


8x16 


180 


30 


36.8 


10-9 


3503 


.80 4994 


,64 


1531 


.94 


330 


1.19 




2x12 


180 


32 


34.2 17.7 


3489 


,80 


5085 


.66 


1624 


.99 


347 






2X10 


180 


32 


38.9.18,1 


3851 


,88 


5359 


.61 


1716 


1.05 


216 


;78 




2x 8 


180 


31 


37.0.15.7 


3403 


,78 


5305 




1676 


1.02 


169 


.61 




2x 2 


30 


290 


33.2.17.2 


4360 


1,00 


7752 


I'.Ct 


1636 


1.00 


277 


1.00 


Short leaf pine 


8X16 
8X14 


180 
180 


12 
12 


39.5i!2.1 
45.812.7 


3185 
3234 


.73 
.74 


5407 
5781 


-70 

.75 


1438 
1494 


1.03 
1.07 


362 
338 


1.40 




1,31 




8X12 


ISO 


24 


52.2 


11 8 


3265 


,75 


5503 


.71 


1480 


1,06 


277 


1,07 




5x 8 


180 


24 


47.8 


11-5 


3519 


.81 


5732 ; ,74 


1485 


1.06 


186 


.72 




2X 2 


30 


254 


51.7 


13-6 


4360 


1.00 


7710 ,1,00 


1395 


1,00 


258 


1.00 


IVsHtern larch < -,-..-.,-, ^ 


8X16 
8X12 


180 
180 


32 
30 


51.0 

50.3 


25,3 

23,2 


3276 

3376 


,77 
.79 


4632 
5286 


.64 
-73 


1272 
1331 


.97 
1.02 


298 
254 


l.U 




,94 




5X 8 


180 


14 :56.025.6 


3628 


.83 


5331 


.74 


1432 


1,09 


169 


.63 




2X 2 




189 


46,2 


26,2 


4274 


1,00 


7251 


1.00 


1310 


1 00 


269 


1.00 


Loblolly pine 


8X16 
5X12 


180 
180 


17 

94 


45.8 
60.9 


6.1 

5,9 


3094 
3030 


.75 
.74 


5394 

5028 


.69 
.64 


1406 
1383 


.98 
,96 


383 
221 


1.44 




.83 




2X 2 


30 


44 


70.9 


5,4 


4100 


1,00 


7870 


1.00 


1440 


1.00 


265 


1.00 


Tamarack 


BxI2 

4X10 


162 
162 


15 
15 


57.6 
43.5 


16.6 
11,4 


2914 
2712 


,73 4500 
.70, 4611 


.66 
.68 


1202 
1238 


1,05 
1,08 


255 
209 


l.Il 




.91 




2X 2 


30 


82 


38,814,0 


3875 


1,00 6820 


1.00 


1141 


1-00 


229 


1.00 


Western hemlock 


8X16 


180 


39 


42.5 


15,6 


3516 


,80 5296 


.73 


1445 


t.Ol 


261 


.92 




2X 2 


28 


52 


51.8 


12,1 


4406 


1,00 7294 1,00 


1428 


1,00 


284 


1.00 


Redwood 


8X16 
6x11 


180 

180 


14 
14 


86.5 

87-3 


19,9 
17,8 


3734 

3787 


-79 4492 
80 4461 


,64 

.64 


1016 
1068 


,96 
1.00 


300 
224 


1.21 




,90 






180 


14 


79.8 


16.7 


4412 


.93 5279 


,76 1324 


1,25 


199 


.80 






180 


13 


86,1 


23,7 


3506 


,74| 4364 




947 




255 


1.03 




2X12 


180 


12 


709 


18,6 


3100 


,65' 3753 


:54 


1052 


!99 


187 


.75 




2X10 


180 


13 


55.8 


20,0 


3285 


,69 4079 


,58 


1107 


1.04 


169 


.68 




2x 8 


ISO 


13 


63.821,5 


2989 


.63: 4063 


.58 


1141 


1.08 


134 


.64 




2X 2 


28 


157 75.519.1 


4760 




1.00 1061 


1.00 


248 


1.00 


Norway pine 


6X12 

4X12 


162 
162 


15 50 312.5 


2305 
2648 


,82 
.94 


3572 ■ 69 1 987 


!:S 


201 
238 


:.i7 




18 


47,914 7 


4107 ,79 


1266 






4x10 


162 


16 


45,7,13,3 


2674 




4206 .81 


1306 


1.36 


198 


;!i5 




2X 2 30 


133 


32,311,4 


2808 




5173 'I. 00 


960 


.00 


172 


1.00 


Red spruce 


2X10 


144 


14 


32,5l21.9 


2394 




3566 


,60 


1180 


.02 


181 


.80 




2X 2 


26 


60 


37,321.3 


3627 




5900 


1.00 


1157 


,00 


227 


1.00 


White spruce 


2X10 
2X 2 


144 

26 


16 

83 


40.71 9 3 


2239 
3090 


;i 


3288 
5,8. 


,63 
1.00 


1081 
998 


,08 


166 
199 


.83 




.3 


10.2 


1,00 



,, Google 



SUPPORTS OF THE RAIL 



167 



TABLE XXXVIII. — CoJi(maed 

COMPRESSION AND SHEAR TB9T3 ON QREEN MATERIAL 







pnin. 








*-. 


! 

1 


1 

■3 

1 


1 
£ 


1 

6^ 


1 
1 

SI 


1 


i 
J 


J 


1 

■5 


■5 
1 

1 






In.. 

4X4 


46 


26,3 
34.7 
30.7 
309 
29,8 
34.8 
37.9 
41,2 
43,5 
51.4 


Lb.. 

3480 


loon 

Lbs, 


Lb., 
4800 
4400 
3500 
3490 
4030 
3290 
3430 
3436 
3423 
3570 


4X4 


IB, 

4 


22 


25,3 


568 


44 


... 








Douglas fir 


6X6 515 
5X6170 
2X2'902 
6X6;I08 


2780 
2720 
3500 
2620 


1181 
2123 
1925 

1801 


4X8 


16 


259 


30.3 


570 


531 


2S,7 


765 




















Douglas fir (fire-killed) . , 


6X8 


16 


24 


33.7 


368 


77 


35.8 


631 




6X6 
5X8 
2X2 


95 
281 


2514 
2241 


1565 
1529 


5x8 
6X8 
5x8 
5X5 
2X2 
fivR 


16 

14 
12 

8 
3 
16 
12 
6 
4 
8 
8 


12 
12 
24 
24 
277 
22 
20 
53 
30 
16 


37,7 

42,8 
63,0 
47,0 
48,5 
43,6 
40.2 
52,8 
50.4 
67.2 
44.6 


361 
366 
325 
344 
400 
417 
416 
478 
472 
392 
646 


179 


47.0 


704 














































Western larch 


6x6 
2X2 


107 
491 


49.1 

50.6 


2675 
3026 


1575 

1545 


3510 


179 


40.7 


700 












4X6 

4X4 
8X4 
4X4 




























Loblolly pine 


8X8 
4x8 
2x2 
6X7 
4X7 
2X2 
6X6 
2X2 
6X6 
2X2 


14 

IS 
53 
4 
6 
165 
82 
131 
34 
143 


63.4 


1560 


365 

691 


2140 
3560 
3240 
3032 
3360 
3190 
3355 
3392 
3882 
3980 


121 


83,2 


630 




74.0 












-riv, 


1432 
1334 












24 


39.2 


fm 








































Western hemlock 


46.6 
66.6 
B3.6 
72,1 


2905 
2938 
3194 
3490 


1G17 
1737 
1240 
1222 


6x4 


6 


30 


48.7 


434 


54 


05.7, 630 


Redwood. 


6X8 

6x6 
6X7 
6X3 
6x2 
6x2 
6X2 
2X2 


16 
12 
9 
14 
12 
10 
8 
2 


13 
14 
13 
13 
12 
11 
12 
186 


86.7 
83.0 
74.7 
75.6 
66.5 
55.0 
56.7 
75.5 


473 
424 
477 
411 
430 
423 
396 
560 


148 


84.2! 742 






















































1 






































6X7 
4X7 
2X2 

2X2 
2X2 


5 

8 
178 
58 
84 


29.0 1928 
28 4 2164 

26.81 

35 4' 


905 
1063 


2404 
2652 

2504 
2750 
2370 


20 
































,.,- 




2x2 


2 
2 


43 

46 


31,8 
50. 


310 
270 


30 ^32, 61 758 




61 




40 58,0 651 















>, Google 



168 STEEL RAIIS 

TABLE XXXIX. — UNIT STRESSES FOR STRUCTURAL TIMBER RECOMMENDED BY 
THE COMMITTEE ON WOODEN BRIDGES AND TRESTLES 

AH. RY. ENQ. ASSN. 
IFoanda ps Sqnira Iiehl 





fiiKlip.. 


Bb«rinc. 




» 


Kiodol 
Tlmbor. 


1E''W 


^.'S!.'" 


DnlSbnriii 


■«- 


"sa." 


For 
ISDums. 


&™St««' 
15 Dianu. 




<s 


n 


1 


fl 

-<5 


a 


1 

■^5 


n 


II 


^1 


1 


n 


s! 


D04«l«fir .... 


SIM 


IIOI 


i.sio.aoo 


eoo 


170 


270 


1,0 


» 


.,. 


3800 


1200 


«0 


■"(-ife) 


10 


LMcWpiH... 


uoo 


,m 


1.010.000 


720 


180 


300 


120 


320 


2«0 


mo 


.300 


«80 


.»(-*&) 


10 


BboKkBtpiH... 


««0D 


1100 


1.480,000 


710 


170 


330 


130 


340 


170 


3400 


1100 


830 


"«(-„'^) 


10 


VhiUpiM. ... 


4400 


two 


i.i3aooo 


400 


100 


180 


70 


2*0 


150 


3000 


,««, 


750 


'-(-«) 


10 


SpniM 


(Wl 


lonn 


1310.000 


imn 


110 


rn 


TO 


170 


ISO 


■^■m 


lion 


830 


uoofi ' •) 




Norway plw. . . 


4200 


«» 


1.1».000 


m 


130 


250 


100 




ISO 


20O0* 


800 


6C0 


»(-») 




Tapunck 


4«00 


too 


1.120.000 


«70 


170 


280 


lOO 




230 


3200' 


1000 


760 


■"(>-ife) 








llflO 




Ml 


ton 




100 








i?ro 


too 


,»(,_'> 






























•<"(' ') 




Bald<9prwi. - 


4800 


too 


,.,»«» 


500 


110 






340 


170 


3000 


1100 


830 


"•(-.fe) 








urn 










470 


™ 


WOO 


win 




«»(l-ifel 




Whiwo^ 




1100 


I.IMUJOO S40 


a. 


tm 


,1. 


020 


450 


3500 


.3.0 


B80 


.»(-sft) 


,. 


Not<..-T 
br imp&ot. 


.^^ 


Pvlii 
tatrt) 


lyur-dry 


..r- 


DOOpd 


liOBOl 


timbi 


<ud 


l«>Db 


beg;! 


witbc 


tbinbob 
•itipcnu. 


ng tb« LIvc-hMid 


«-. 



The difference between green and seasoned wood may amount to as much 
as 50 per cent as shown by Table XL. The influence of seasoning consists in 
(1) brining by means of shrinkage about 10 per cent more fibers into the same 
square inch of cross section than are contained in the wet wood; (2) shrinking 
the cell-wall itself by about 50 per cent of its cross section and thus hardening 
it, just as a cowskin becomes thinner and hardens by drying. 

Table XL applies only to small, clear pieces of wood seasoned under spedal 
conditions with great care. The Forest Service has found * that a comparison of 
the results of tests on seasoned material with those from tests on green matmal 
shows that, without exception, the strength of 2 by 2 inch specunens is increased 
by lowering the moisture content, but that increase in strength of other dzes is 

* strength Values (or Structural Timbers. McGarvey Cline. Forest Service, Circular 189, 
Jan. 25, 1912. 



,y Google 



SUPPORTS OF THE RAIL 



169 



mucb mcve erratic. Some specimens, in fact, ^ow an apparent loss in strength 
due to seasoning. In the li^t of these facts it is not safe to base working stresses 
on results secured from any but green material. 

TABLE XL.— REDUCTION FACTORS FOR STRESS AT ELASTIC LIMIT IN BEND- 
ING OF LONGLEAF PINE 

(ForuC Serviee. BullHiii 70) 





from- 


To- 






















Houtora 












Uoiit 




at. 






























3 


*. 


«. 


a. 


10. 


11. 


14, 


IB. 


18. 


». 


12. 


Ji, 


M.' 


2 


1 


1. 13 


1.31 


1.53 


1.75 


1.99 


2,20 


2-39 


2.54 


2.70 


2.85 


2.99 


3,14 


4 




1 


1.15 


1.35 


1.54 


1.75 


1.94 


2.11 


2.25 




2.52 


2.64 


2,77 


6 


>67 


,867 


1 


1.17 


1.34 


1-52 


1,69 


1.S3 


1.95 


2-07 


2.18 


2.29 


2,41 


8 


.656 


.742 


.S56 


1 


1.15 


1.30 


1.44 


1.57 


1.67 


1,77 


1.87 


1.96 


2.06 


10 


.572 


.MS 


.746 


.882 


I 


1.13 


1.26 


1.37 


1.45 


1.54 


1.63 


1.71 


1.80 


12 


.503 


.570 


.657 


.768 


.881 


I 


l.Il 


1.20 


1.28 


1,36 


1.44 


1.51 


1.58 


14 


.455 


.515 


.5M 


.694 


.795 


.904 


1 


1.09 


1.16 


1.23 


1.30 


1.36 


1,43 


16 


.419 


.475 


.547 


.639 


.733 


.832 


.921 


1 


1.07 


1.13 


1.19 


1.26 


1,32 


IS 




.445 


.513 


.600 




.781 


.865 




1 


1,06 


1.12 


1.17 


1.24 


20 


.'370 


.420 


.485 


.565 


^648 


.735 


,814 


^88^ 


.941 


1 


1.05 


l.U 


1.16 


22 


.351 


.397 


.45S 


.535 


.614 


.697 


,771 


.838 


,&93 


.948 


1 


1.05 


1.10 


24 


.335 


.379 


.437 


.510 


.585 


.665 


.735 


.799 


.851 


.904 


.954 


I 


1.05 


•28 


.318 


.361 


.415 


.486 


.556 


,633 


,700 


.760 


,810 


.860 


.908 


.951 


1 



A recent instructive series of tests* have been conducted by W. K. Hatt in 
the Laboratory for Testing Materials of Purdue University, in cooperation 
with the Wood Preservation Committee of the American Railway Engineering 
Assodation and with the following organizations: 

Big Four Rsulroad Company. 

Illinois Central Rmlroad Company. 

American Creosoting Company. 

Ayer and Lord Tie Company. 

Atchison, Topeka and Santa Fe Railway Company. 

Forest Service, U. S. Department of Agriculture. 
The principal results of these tests are shown in Table XLI. One of the main 
determinations of the tests was of the re^tance of the ties to the direct pressure 
of the rail. It was shown that the various treatments had not weakened the ties, 
except in the case of ties newly tr^ted with crude oil. 

The tie was put through a planer, so that one surface was true. The 
othw surface was adzed at the place of bearing to provide a true bearing for 
the plate representing the bottom of the rwl or the tie plate. 

* Fourth Progreea Report of Tests on Treated Ties, Proceedings Am. Ry. Eng. & M. of W. 
Asaa., 1910, Vol. II, Part 2. 



,y Google 



STEEL RAILS 



TABLE XLI. — BEARING STRENGTH OF TIES UNDER THE RAIL 

(3m FracMdiiica Am. Ry. Eat. & U. ol W. Am., VdL 11. Pmrt I. pp. 8U-4) 





Natunl. 


TraiWd, 


R<in.rk>. 




Nc.ol 
Tert.. 


c™|^« 


No. or 

Tarts. 


Ott- 


Red oak 


100 

70 
30 
20 
29 
249 


Kfflo'""' 
612 
640 
688 
830 


100 
70 
30 
20 
29 

2i9 


Ll».p«<q. ». 

lOSO 

591 
618 
725 
791 




Loblolly pine 


1. Crude oil, 539. 
1. Crude oil, 373. 


Longteaf pine 


Red gum 

Total 


1. Crude oil, 655. 



Tests were made upon 581 half-ties to determine the relation of the crush- 
ing strength of the ties with and without the tie plates. 

The fiber stress per unit of area of wood under the tie plate at the elastic limit 
in the case of the oak are less than those under the rail alone except for plate C 
(see F^g. 124). Of course, the total load is greater. This is accounted for by 
the perceptible springing of the tie plates, thus producing a non-uniform pres- 



i?^'j%>- 




Fio. 124.— Tie Plate Fonna i 
Teste at Purdue University. 

f sz. ^'*^ Stress in Tie Plates at Elaatie 

|*7» Limit: A37,500 lbs. peraq. in.; B68,800 

fln. per sq. in.; C 59,800 lbs. per eq. in. 



,y Google 



SUPPORTS OF THE RAH. 



171 



sure on the wood under the tie plate. The loads, therefore, carried with the Md 
of a tie plate, while larger, are not increased in the same ratio as the increase 
of bearing surface. 

In the loblolly-pine ties and in plate C on red oak, no perceptible spring- 
'mg of the tie piate was observed within the elastic limit of the timber, the load 
being increased in practically the same ratio as the surface. 

Tie plate A (see Fig. 124), t'a inch thick, was permanently bent at the 
edge of the r^l bearing when the test was carried to J-inch compression on oak 
ties. The yielding was confined almost entirely to the edge of the rail bearing. 

Tie plate B, J inch thick at edge of the rail, was not permanently bent 
by the same test. It, however, springs as much, or more, than plate A, but 
the sprin^ng was more uniform. Plate B is harder metal, and this would seem 
to be an advantage in this test. 

The three tie plates were tested under flexure to determine the quahty 
of the metal. The results are shown in Fig. 124. 

In the calculations of the strength of the tie, if we take the strength of 
the wood as shown by Table XLII, the result will be not far from correct. 
The working stress at the r^I bearing given in the table refers to the allowable 
stress under the tie plate. 

TABLE XLII 
WORKING STRESSES FOR TIE TIMBER 



Kind of Wood. 



Compreauoii M 
RaiIB«ariB(. 



Oak 

Longle&f pine. . . 
Inferior woods. , 






Examining the safe load the tie in the track will carry, we have to con- 
sider two sources of possible failure of the tie: 

1. The compression of the fibers under the tie plate. 

2. The rupture of the tie due to too great a bending moment in the tie. 
A 6 by 9-inch tie plate gives an area of 54 square inches. Referrmg to 

Table XLII, we find the permisdble load on the tie plate to be 27,000 pounds 
for oak ties, 17,500 pounds for longleaf yellow pine, and about 11,000 to 
13,000 pounds for the inferior woods. 

Let us now consider the bending moment in the tie. 



,y Google 



172 STEEL RAILS 

If the tie were completely ri^d, there would result a uniform distribution 
of the pressure on the ballast. This is, however, never realized, and there is 
an unequal distribution of the pressure. 

The tie should be considered as a continuous beam, supporting a vertical 
load at two points and resting on material which is, within certain limits, com- 
pressible. Exactly what takes place in the ballast under the loaded tie is of 
the greatest importance in determining the bending of the tie. 

* M. Couard found that the vertical displacements of cross-ties hardly 
reach three millimeters (J inch), and that they are not proportional to the 
weights supported. He has concluded from his experiments that "the cross- 
ties fixed to the raW remain, at certain points, suspended above the ballast, 
and that right at the rail there is formed under even the best tamped cross- 
ties some depressions of ballast on the edges of which the cross-tie is supported; 
that under the passage of a wheel even lightly loaded the cross-ties come in 
contact with the ballast and deflect to the depth of the depressions." 

Shwedler, Hoffman, Schwald, Riese, and Zimmerman, from the theoretical 
researches of Winckler, have derived the elastic curve of the tie represented 



Fio. 125. — Elastic Curve of Tie, 7 feet 10.4 Fio. 126. — Elastic Curve of Tie, 8 feet 10.3 
iochea long. (After Winckler.) inchea loDg. (After Winckler.) 

by Fig. 125 or Pig. 126, according as the cross-tie was 2 m. 40 (7 feet 10.4 inches) 
or 2 m. 70 (8 feet 10.3 inches) long. 

Very careful experiments have been made by M. Cuenot on the relative 
action of the lie and the ballast.f The following record of his tests is taken 
from Mr. W. C. Cushing's translation of his work: 

"The rails employed were of the type used on the Paris, Lyons and Medi- 
terranean, either the P. M. type, of a weight of 39 kilograms per running meter 
(78.6 pounds per yard), or the P. L. M.-A. type, of a weight of 34^ kilograms 
per running meter (69.5 pounds per yard). 

"All my experiments, during nearly three years, have been made, first, on 
a side track, then on track No. 2 of the hne from Mouchard to Bourg, trav- 
ersed by the express and fast tr^ns, comparatively with oak cross-ties employed 
on the P. L. M. system, and with compoate cross-ties (wood and steel). (See 
Fig. 127.) Finally, a special track for experiments was laid at the Bourg 

• Revue des Chemina de Fer, July, 1897. 

t Deformations of Railroad Tracks and the Means of Remedying Them. G. CuSnot, 1007, 
New York. 



,y Google 



SUPPORTS OF THE RAIL 173 

station, and there was tested, at the same time as the two types of cross-lies 
mentioned, the metallic cross-tie in use on the State System. 

"The wooden cross-ties were oak, creosoted, and of the following dimen- 
sions: 

Length 2 m. 60 (8 ft 6.36 in.) 

Width m. 22 to m. 25 (8.66 in. to 9.84 in.) 

Depth m. 14 to m. 15 (5.51 in. to 5.91 in.) 

"The composite cross-tie was composed of a metallic skeleton in the form 
of an inverted trough, provided in the interior with two symmetrica] blocks 
of wood solidly fixed, and leaving between them an empty central space." 

4»Hr . 



FiQ. 127. — Wood and Composite Ties uaed in Cuenot's Experiments. 

The measuring apparatus for experiments in the static state was as follows: 
"There were first placed in the surface of the wood cross-ties, screws with 
square heads distributed over their whole length and giving 15 or 16 fixed 
points, which were to serve as bench marks for the determination of the defor- 
mation. A rigid steel rule in the form of a T (Fig. 128) presented, right at the 
points, whose spacing was the same for all cross-ties, vertical rods terminated 
by a notch, in which was brought, while resting on the screw with square head, 
a gage in the form of an inclined plane, whose divisions were calculated in a 
manner to correspond with a tenth of a millimeter. The inclination of the 
inclined plane had been so chosen that the interval between two divisions was 
at least of 2 centimeters (fVo inch), which allowed estimating the tenth of a 
miUimeter with exactness. 



,y Google 



174 STEEL RAII5 

"The rule was fixed in an uncliangeable manner to two stakes of strong 
dimensions, buried in the embanlanent about 1 m. 10 (3.61 feet), in order to 
eliminate the influence of the load on the supports of the rule. When the 
rule was in place, an observer introduced the wedge-shaped gage in the notch, 
while mainlining it horizontally on the head of the screw, and stopped it at 
the moment when it commenced to become wedged; he then made a first read- 




FiG. 128. — Meaauring Apparatus for Ties under Static Load. (Cuenot.) 

ing on ail the pdnts of reference, proceeding, for example, from left to right, 
then a second, proceeding in the reverse direction, from right to left. The 
readings made were recorded and the mean taken, which thus gave the actual 
position of the cross-tie. 

"The vehicle, which served to load the cross-tie conadered, was brought 
up, always taking care to place the same wheels at the same spot with reference 
to the piece submitted to the test; it was allowed to remain during about 10 
minutes, and the readings were recommenced. Two successive readings were 
made, and the mean of them was taken. The diffCTence between the inscribed 
means gave the deformation of a cross-tie under the load considered. 

"The measuring apparatus (see Fig. 129) for the experiments in the 
dynamic state was essentially composed of a stylus arranged in a stable manner 
at the face of a plate of smoked glass and fixed on the points of the cross-tie 
under observation. The black smoke deposited on the glass plate, which was 
displaced at the same lime and by the same amount as the points, was re- 
moved by the point of the stylus; the height of the part removed gave the value 
of the deflection, or of the raising, of a cross-tie at the points conadered. The 



,y Google 



SUPPORTS OF THE RAIL 



175 



reading of this heigbt was made by means of a magnifying glass nearly to the 
tenth of a millimeter. 

" The stylus, with a fiat point of tempered steel, was mounted on a very 
flexible spring, which could be approached to or removed from the glass plate 
at will, with the aid of a thumbscrew. The glass plate was fixed by screws 
on one of the faces of a cross-tie, then smoked in a flame of a candle at the 
moment when it was desu^ to put it in service. The thumbscrew passed 
through an iron rod and simply 
rested on the spring, which, left 
free, moved back and forth on 
the rod fixed by means of two 
bolts on a stake deeply buried in 
the soil. 

"In order to make an ob- 
servation, the screw is pressed 
i^ainst the spring until the point 
of the stylus comes in contact 
with the blackened plate. In 
this position a light blow is given 
to it, which makes it oscillate 
and defines a horizontal trace 
of 2 or 3 millimeters' (iH to 
-^ inch) length on the black 
smoke, which forms the reference 
mark. 

"At this moment one can 
either place the vehicle on the 
CTOSs-tie, or allow trains at speed 
to pass over it. The height of the part of the glass plate rubbed off by 
the point of the stylus gives, above the reference mark, the values of the 
depression, and below, the uplift, of the cross-tie. The latter is always 
inferior to the former; for the flexure Is important in comparison with the move- 
ment of uplift of this piece under the influence of loads at a distance. The 
successive influence of each of the axles cannot be noted, but it is solely a maxi- 
mum indication which is produced." 

The results of the dynamic tests were always slightly less than those 
obtained from the static tests. Fig. 130 shows the general results from the 
large number of experiments made. 





POSITION OF 3 REFERENCE MARK3 R 
THE LENGTH OF A CROSS TIE , IN PLAN. 

Fio. 129. — Measuring Apparatus for Ties under 
Dynamic Load. (Cugnot.) 



>, Google 



STEEL RAILS 



W^';-..H^JIJi.,Y^^ WOOD TIE ^^^T^^^^i^W^^!^^. 

&55' ^^T^ — *] 

r TAMPED With TJMBINO BAB 




3P0NTWK0US 




COMPOSITE TIE 




Fio. 130. — Results of M. Cuenol'a Tesla on Ties, 

The following conclusions have been drawn by M. Cufinot from his ex- 
periments: 

" (a) The long ties, 8 feet 6.36 inches to 7 feet 6.6 inches, take, under the 
load, the form of a basin with the bottom slightly raised in the centra*. 

" (6) The short ties, 7 feet 0.6 inch to 6.5 feet, are deformed according to 
a curve, convex or otherwise, and inclined toward the extremity. 



,yGoog[e 



SUPPORTS OF THE RAIL 177 

" (c) The ties between 7 feet 0.6 inch and 7 feet 2.64 inches are lowered 
parallel with themselves without sensible curvature. 

"(d) The unsymmetrical tamping raised the curve towards the center; a 
very feeble lack of symmetry reacts very clearly in this direction. 

"(e) It is possible, by increasing the rigidity of a cross-tie, notably by 
concentrating the material about the supports, to reduce its anking to the 
quantity which is intended as a limit, and its flexure in such measure as one 
would wish. 

"(J) The permanent sinking of the ballast is variable according to the 
case, but the elastic sinking, the only one there is reason to consider, is, so to 
speak, constant, whatever be the length and type of the cross-tie adopted. The 
deformation is slowly produced and augments with time." 

It is here seen that a tie 8 feet 6 inches long, which is the usual length of 
tie employed in this country, under proper conditions of tamping, will assume 



1_^ 



Fia. 131. — Strain Diagram of Entire Tie. 

the loaded portion shown in Fig. 131. In the figure, the loads W at A and 
C represent the load at each r^I. 

Considerations of the tamping under the tie will not admit of any exact 
mathematical formula for the distribution of the pressure of the ballast. If, 
however, as a working hypothesis, we assume that the pressure of the ballast 
is uniformly distributed between the rails and that the pressure is similarly 
uniformly distributed, but of greater intensity from the rail to the end of the 
tie, and, further, that the tangent to the elastic curve of the loaded tie is hori- 
zontal under the points of support of the rail, then approximate formulse can 
be readily derived for the maximum bending moment in the tie and the greatest 
intensity of pressure of the ballast in terms of the load on the tie at either rail 
bearing. 

This assumption, while not taking into account all the conditions of the 
loading (rf the ballast nor giving the exact distribution of pressure under the 
tie, will, nevertheless, when applied to an 8-foot 6-inch wood tie, afford a means 
of determining the maximum bending moment and the greatest intensity of 
pressure with sufficient accuracy for our purpose, and very possibly as exactly 
as the present state of our knowledge of the subject warants. 



,y Google 



178 STEEL RAII5 

Where it is desired to investigate the action of ties in a more tlxorough 
manner, the calculations may be proceeded with in the same manner as those for 
the case where the rail acts as a continuous girder in Article 23. It should be 
borne in mind, however, that the coefficient of the ballast or the ratio of pressure 
to sinking is not the same for all parts of the tie, but with proper conditions of 
tamping is considerably greater xmder the tie in the region adjacent to the rail 
bearing. 

Keferring to Fig. 132, the moments at the supports A and C are equal 
and each ^ W'U, where W is the total load uniformly distributed over 



Fia. 132. — Strain Diagram of Tie between Rails. 

the span U. The bending moment at the center of the span B is a** W'U. 
Therefore, the maximum bending moment occurs at the supports and is 
M, = I'a W'U. 

The free part of the tie outside the rail acts as a cantilever; the maximum 
bending moment occurs at C and is M, = i W"U', where W" is the total 
load uniformly distributed over the span L" (Fig. 133). 



Fig. 133. — Strain Diagram of Tie outside of Rails. 

Conadering the tie as a whole, Fig. 131, we have, from the principle of the 
continuous girder, M^ = -^.j W'U = ^ W"U', but for an effective length of the 
tie of 100 inches, L' = 60 inches and L" = 20 inches; 
therefore ^\ W' 00 = ^ W" 20 

W' - 2 W". 
But the reaction at any support is equal to the algebraic sum of the shear 
to the right and left of the support, and 

W =J,,+ J,, 
W = ^W' + W" 
W = ^W' + ^W' 
W =W' =2 W", 



where 



W = the load at either rail b«uTng, 

J^i - the shear immediately to the left of C, 

J„ = the shear immediately to the right of C. 



,y Google 



SUPPOETS OF THE RAIL 



Substituting the value of W in the equation for the maximum bending 


moment. 




M. = M. - M„ 
M. = A W X 60, 
M. - 6 W, 


where 




M^ = maximum bending moment, 
M^ = bending moment at A, 
M. - " " " C. 


Theextremeflberstress, /-^-^, or W - £ • 


For 


a rectangular 


tie l-^"^*-'A', 
"'- » - 12 ■ 2 - 6 ' 


and 




" 80 



Turning to Table XLII, we find the allowable extreme fiber stress in bend- 
ing 1000 pounds per square inch for oak, and 750 pounds per square inch for 
the inferior woods. We can, therefore, prepare Table XLIII, showing the safe 
load that the tie can bear and not exceed a proper bending stress. 

TABLE XLIII. — ALLOWABLE LOAD ON TIE AS DETERMINED 
FROM EXTREME FIBER STRESS IN BENDING 



KiDj<orW(»ii 


SiieoITi., 


Allowable Load 




InchH. 

7X8 

7X9 
•Half round 

7X8 

7X9 
•Half round 


Poyads. 

13,100 
14,700 
15,000 
g,800 
11,000 
11,300 















17. Bearing on the Ballast 

Considering now the bearing power of the ballast on which the tie rests, 
the maximum loading on the ballast under the tie per linear inch of the tie, 
from the preceding article, is 

W" _ W 
20 2 X 20' 



,y Google 



180 



STEEL RAILS 



To express W in terms of the bearing power per square foot of the ballast 
(p), and the width in inches of the base (6), we have the allowable load per 
linear inch of the tie equal to 

hp _ W 
144 2 X 20' 

^ = &- 

For bearing on gravel or broken stx)ne, not well confined, three tons pra* 
square foot is as much as should be allowed. 

We may now prepare Table XLIV, showing the safe load that can be 
applied at each rail bearing as determined by the proper load on the ballast. 



Width oIBa»o( Tie. 


Beuini of the Tie. 


InchB. 
8 

9 

•12 


Pound.. 

13,500 

15,000 
20,000 



18. Bearing on the Subgrade 

Before assuming a proper bearing under the tie, an examination must be 
made of the distribution of the load to the subgrade. The following experi- 
ments have been made in Germany to determine the distribution of force upon 
the subgrade.* 

An experimental box, 37 inches long, 20 inches high, and 6 inches wide, was 
filled with a layer of clay 8 inches high at the bottom, on top of which was placed 
a layer of sand 6 inches high, and then a layer of gravel 6 inches high, upon 
which a tie was laid. This tie was tamped with the ordinary tamping pick and 
then subjected to a load of 57 pounds per square inch, or 8200 pounds per 
square foot, by which the rail level was depressed. By the use of an eccentric 
the loading was alternately lifted from the tie and again returned, thus imitating 
the process of passing a loaded wheel over the track. As soon as the tie had 
settled 1.2 inches, which was registered upon an attached sliding plate, the 
tie was again raised and tamped. From time to time photog aphic views 
and observations as to the stage or condition of the experiment were taken 

• Gloaer's Annalen fur Gewerbe und Bauwesen. May, 1899. (EHrector Schubert.) Transla- 
tion appeara in Proceedings Am. Ry. Eng. & M. of W. Assn., Vol. 7, p. 105. 



,y Google 



SUPPORTS OF THE RAIL 181 

by removing the front wall of the experimental box. After the eleventh tamp- 
ing the experiment was considered as completed, and the section shown in 
Fig. 134 was taken. 

From this view we can eaaly see how a short depression, measuring about 
12 inches to 14 inches wide, has been formed in the clay, with an upward swelling 



Fto. 134. — Ballast Experiment — Schubert. Six inches of sand and 6 inchea of gravel. 

on each side. The pressure transmitted from the tie has accordingly distributed 
itself over this small width when the depth below the bottom of the tie was 
12 inches. 

In a subsequent experiment, broken stone was used in place of gravel; 
otherwise the procedure was the same. From a photograph of the section 
after the fifth tamping (see Fig. 135) a depresaon in the clay extending nearly 



Fia. 135. — Ballast Experiments — Schubert. Six inches of sand and 6 inches of stone. 

over the entire width of the experimental box (27^ inches to 29j inches wide) 
is noticeable. The distribution of the force is consequently double that of 
the previous experiment. 

Still more favorable appears this distribution when the height of the stone 
ballast is increased. In doing this, it is judidous to retmn a thin layer of sand 
so as to prevent the larger pieces of broken stone from entering into tiie clay. 



,y Google 



182 STEEL RAILS 

As will appear from the section shown in Fig. 136, a depres^on in the clay has 
not taken place, and only a few of the broken stones have gone through the 
sand to the clay. In emptying the box only a very unimportant depresaon 
was noticeable. 



FlO. 136. — BallsHt Experimenta — Schub«rt. Stone with thin layer ot Bond. 

Finally, the behavior of a foundation layer was investigated, and after 
the fourth tamping the section shown in Fig. 137 was taken. The stones of 
the foundation layer have penetrated the clay rather deep, and not only those 
in the center, but also stones on the sides, from which we can conclude that 



Fio. 137. — Ballast Experiments — Schubert. Stone resting on clay subgrade. 

the force transmitted through the tie has distributed itself nearly over the 
entire width of the box. 

Hence, the most favorable distribution of forces is accomplished by the 
use of ballast of broken stone, with or without a foundation layer. The latter 
is, however, not suitable m a yielding subgrade, inasmuch as the stones pene- 
trate into the grade, and the yielding soil will swell into the spaces, thus making 
the drainage ineffective. 

The ^ect of overloading the subgrade is very clearly shown in Pig. 138. 



,y Google 



SUPPORTS OF THE RAIL 183 

* The road department of the Pennsylvania RaUroad has installed an 
interesting piece of apparatus on the grounds of the South Altoona foundry 
to test the bearing qualities of different kinds of roadway and ballast. The 



Fio. 138. — Effect of Overloading the Subgrade. (Am. Ry. Eog. Aaen.) 

particular ballast or subgrade to be tested is placed in three heavy boxes tiiat 
extend across the track and have sufficient depth to serve the purpose. The 
track crosses this on a level and extends out on either side, terminating in a short 



Fig. 139. — Pennayivania Track Testing Apparatus. (Railway Age Gazette.) 

and sharp incline. A four-wheel car on this track is loaded with pig metal to 
obtain any desired weight on the wheels. This car is also equipped with electric 
motors. A shed built across the track carries an overhead rail, from which a 
motor current is obtained, and a contact shoe is on the car. Fig. 139 illus- 
trates the apparatus. 

• Railway Age Gaiette, June 11, 1909. and July 21, 1911. 



,y Google 



184 STEEL RAILS 

When current is turned on, the car moves out to the end of the conductor 
rail, and here, as the contact is broken, the power of the motor is shut off. The 
car runs on until stopped by the adverse grade, and meanwhile a trip reverses 
the current connections to the motor. Stopped by the grade, the car runs back, 
beneath the current rail, when its motor drives it to the other end, where the 
movement is again reversed. In this way the car is made to travel back and 
forth automatically over the track until the desired results are obtained, the 
number of trips being automatically registered upon a counter. 

These tests are the most extensive of the land ever conducted in this 
country. It was felt that while the data obtained by Mr. Schubert were very 
instructive yet more valuable data could be obtained from a series of experi- 
ments if made in a manner more nearly approaching actual track conditions.* 

The track was 109 feet in length, built of new P. R. R. standard 
85-pound rail with 7-inch by 9-inch by 8§-foot ties spaced 25i inches center 
to center. It being impracticable to run the car faster than about five 
miles per hour, at which speed any effect upon the track, due to impact 
alone, would be negligible, a weight of 75,000 pounds per axle was chosen 
for the experimental truck. 

A series of five tests has been completed; the first one beginning on Sept. 
2, 1908, and the last one ending on Aug. 2, 1910. Table XLV gives general 
data of the tests. Water was applied by sprinkling the boxes to observe the 
effect of moisture on the ballast; the amount applied in each test is shown in 
the table by inches of rainfall. 

In test No. 1 the line of demarkation between the bottom of the ballast 
and the roadbed material was not straight. The test showed conclusively that 
a depth of 8 inches of trap-rock ballast, when laid on the usual roadbed ma- 
terial, was not sufficient to distribute the weight carried by the ties uniformly 
over the roadbed. 

The results in the third box showed, however, that if 12 inches of per- 
meable material, such as cinder, were used beneath the 8 inches of ballast, the 
distribution of the weight over the roadbed material was much better. 

The results of the first test led to the second test to determine how a depth 
of 12 inches, 18 inches, and 24 inches of trap rock under the ties would behave. 
In test No. 2 the dividing line between the ballast and the loam was quite 
straight in box No. 3, but in boxes Nos. 1 and 2 there existed some depressions 
in the line especially under the rail. 

• Experiments to Determine the Neiiessary Depth ot Stone Ballast. Report of the Geoeral 
Manager's Committee Penosylvania Railroad. Proceediogs Am. Ry. EDg. Assn., 1912, Vol. 13. 



>, Google 



SUPPORTS OF THE RAIL 



185 



A study of the sections in test No. 3 showed that the loam was more evenly 
depressed in box No. 3 than in the other boxes where stone had been substi- 
tuted for part of the cinder during the test. 

Test No. 4 showed that the gravel and slag distributed the pressure upon 
the loam with about the same efficiency. 

Test No. 5 was made to determine whether a combination of rock and 
cinder would prove as satisfactory as the rock alone. It was found, however, 
that the tine between the ballast and the loam in box No. 3 was not as good 
as in box No. 3 of Test No. 2. 

TABLE XLV. ~ SUMMARY OF ROADBED TESTS AT ALTOONA. 



TwtNo. 


■ 


= 


' 




• 


let P.rl. 


Ind Pan. 


lit Part. 


lad Part. 


let Fait. 


2nd Pari. 


T 


KK 


Apr. IB. IB09 
Jime IS. 1909. 


lely 20.IBm. 


it.V.i'^ 


Oct. IV, IWD 
Nov. 17, 1B0» 


Nov. 17. IKOe 
Dec.J. 1»0». 


May le. I BIO 
lui»24.IB10 


Jue 24, igio 
Aof. 2, 1910. 


BisMo. I 

BoiNo-S 

No. Ol RMDd 

BoiNo.8 
RsiBlell 


B'tnvniek 
27'Hiidy 

kWD 

K-'SS" 

u- bid elm- 
SI.«00 

f 


Ii; trap rock 

M;tnpTDak 

4fl,fl32 


24; cinder 
24; cinder 


trap rock 
L'Bchuged 


I2'«iidy 
4S,H1 

f 


"Lis .2 

40,080 

1^ 


24' cinder 
IJ-audy 

ie,2io 


10' cinder re- 
moved end 

S3.0M 

I- 



For computing the bearing on subgrade we are furnished a method * by 
Mr. Thomas H. Johnson, who has made a study of Director Schubert's report 
with a view to deriving a formula which would show the thickness of ballast 
necessary to produce an equal distribution of the axle loads on the surface 
of the roadbed underneath the ballast. 

Referring to fig. 140, the following formuke are su^ested by Mr. Johnson : 
For gravel, ab = x = b' + ^d'. 
For stone, ah = x = b' + d'. 

The relatively small arcs will approxinoate to parabolas and may be con- 
ffldered as such. 

• Proceedings Am. Ry. Eng. & M. of W. Asm., 1906, Vol. 7. 



,y Google 



186 



STEEL RAIIS 



The intensity of pressures is proportional to the ordinates of the carve. 

Areas of parabolic segment = Ixy; hence, mean ordinate = f y, or mean 
pressure = | maximum pressure, or maximum pressure equal to | mean pressure. 

Pressure at 6 =0; hence, to obtain an approximately uniform distribution 
over the surface of roadbed, the tie spacing S must be such that the curves 
overlap and have a common ordinate, y' = hv- This will occur when db = {cb; 
or eb = \ ab; OT mo = i mn.* 

We should obviously aim to space the tie so that the area of distribution 
of adjacent ties will overlap and give approximately an equal distribution of 
the axle loads on the surface of the roadbed underneath the ballast. 




Fia. 140. — Distribution ot Pressure to Subgrade. (Johnson.) 

With a tie spacing of 23 inches centre to centre of ties, by applying Mr. 
Johnson's formula, we find that it will be necessary to use 45 inches of gravel 
ballast and 22 inches of stone ballast under the tie to obtain equal distribution 
on the subgrade. 

Tie spacing, iS = 23 inches = | x. 

For gravel, S = i{b' +h d'), 
or S = J 6' + I d', 

N'=5-f6', 
and d' = I (S - 1 6') = I (23 inches - 1 8 inches) = 45S mches. 

For stone, S = H&' + d'), 
or S = f &' + 2 d', 

and d' = j (S - f 6') = 4 (23 inches - f 8 inches) = 22J inches. 

* Approximate; to be exact, db « 0.29 eb and mo = 0.71 mn. 



>, Google 



SUPPORTS OF THE RAIL 187 

It will be seen that unless an excessive depth of ballast is used, a uniform 
distribution of pressure on the subgrade will not be obtained. However, if the 
maximum pressure on the subgrade does not exceed its allowable bearing po^er, 
the fact that it is not uniformly distributed will not necessarily prove detrimental. 

From the above, we see that the maximum pressure = | mean pressure; 
W 
but from article 17 the mean pressure is ^rp ' and the maximum pressure is, 

therefore, 

2 ^40 a: 27 1' 
Substituting the value of x for stone and gravel ballast, we have the maxi- 
mum pressures: 

Gravel ballast, ^.^^^L, (i, 

w 

stone baUart, ^ri^Tr,- f^) 

We may take the depth d' for gravel ballast as 18 inches and for stone as 
12 inches. Equations (1) and (2) will, therefore, reduce to: 

Gmvel ballast, 27(5*^9) ■ <'' 

w 

stone ballast, 27{b' + 12) ' <^^ 

Equations (3) and (4) are, then, the expressions for the maximum pressure 
on the subgrade per square inch, in t^ms of the load on the tie and the width 
of the base of the tie. 

For bearing on clay foundation, subject to frost and usually made ground, 
1 to 1^ tons per square foot is good practice. Therefore, putting equations (3) 
and (4) equal to the bearing power of the subgrade, we can obtain the value of W. 
3000 W 



Gravel ballast, 
Stone ballast, 



144 27 (6' +9) 
3000 W 



144 27 (6' + 12) 
From which 

Gravel ballast, W = 563 (6' + 9), 
Stone ballast, W = 563 (6' + 12), 

NoTB. — Frofeesor Talbot is now engaged on teeta at the Univeraity of riinoiB having for their 
purpose the detemunation of the distributioD of pressure in gravd. These teeta are not complete, but 
the evidence produced so far appears to indicate that the pressure under the center of the tie ae shown 
in Fig. 140 is less than that at the edges, due to a distinct arching effect of the material under the tie. 
A very great difference in the distributing power of the sand was noted under different conditions of 
dampneas. These tests when finished will doubtless furnish information of T«lue in reference to the 
distribution of the rail pressure to the subgrade- 



>, Google 



188 



STEEL RAIIS 



where W is the safe load in pounds applied to the tie at the rsul bearing and 
b' is the width of the tie at its base in inches. 

Table XL VI shows the value of W for the different ties under conaderation. 

TABLE XLVL — ALLOWABLE LOAD APPLIED TO TIE AT RAIL BEARING 
AS DETERMINED FROM BEARING ON SUBGRADE 



Width of Ti. 


B.,1^. 


^SJ? 


Kind. 


Depth batow 
Tie- 


8 
9 

12 
8 
9 

12 


Gravel 

Gravel 
Gravel 

Stone 
Stone 
Stone 


is 
18 
18 

12 
12 
12 


Poimde. 

9,600 

10,100 
11,800 
11,300 
11,800 
13,500 



19. Supporting Power of the Tie 
Table XLVII assembles the information given in the previous tables. 

TABLE XLVII. — BEARING POWER OF TIES IN THE TRACK 



7X1«. lailochei 
% Feet inchee. 
(Halt Round.) 



Allowable load, io pounds applied at bearing of 

rail on tie. as determined by: 
Bearing of tie plate. 

Oak 

Longleal yellow pine 

Inferior woods 

Bending of tie, 

Oak 

Inferior woods 

Bearing on ballast 

Bearing on grade, 

IS-inch gravel ballast 

12-incb stone ballast 



27,000 

17,500 

1 11,000 

(13,000 



27,000 
17,500 
11,000 
13,000 



It would apparently seem that the weakest part of the substructure of the 
track lies in the bearing on the subgrade. In some cases of a very weak sub- 
grade, as in the muskeg swamps of Canada, it has been found necessary to 
resort to unusual methods of track construction in order to maintain the track 
in proper condition. Mr. D. MaePherson reported at the January, 1912, 
meeting of the Canadian Society of Civil Engineers the use of 12-foot ties in a 



,y Google 



SUPPORTS OF THE RAIL 189 

stretch of track over muskeg, the resulting cheapening in cost of maintenance 
apparently fully warranted the extra expense of the large ties. 

If we consider the effect of the dynamic load, it will be noted from the 
discussion in the previous articles that the sinking of the tie in the ballast undra- 
the action of the dynamic load is little, if any, greater than under the static 
load, although the dynamic load is from 50 to 75 per cent greater in amount 
than the static load. 

As the calculations of the strength of the track must be made for the greatest 
load put upon it, which is the dynamic load, it would seem desirable to increase 
somewhat the safe bearing values ^ven in the table as determined by the bend- 
ing of the tie and bearing on the ballast and subgrade. We are not warranted, 
however, in assuming a like increase in strength at the bearing of the tie plate 
under the action of the dynamic load, as the effect of the moving loads is, in 
this case, to reduce the strength of the wood. 

Examining Table XLVII as it applies to dynamic loading, it is seen that 
a bearing value of 14,000 pounds, or 7 tons, on half the tie can probably be 
taken with safety except in the case of the bearing of the plate on the soft-wood 
tie. The use of a soft wood, as cedw or loblolly pine, for ties under heavy 
traffic, with the customary form of plate and fastening in use in this country, 
is to be discouraged, and the general tendency at the present time is to use a 
wood better adapted to resist mechanical wear under these conditions. 

The rail in the track acts as a continuous girder, resting upon yielding 
supports. Evidently, therefore, not only must the allowable safe load on each 
lie be determined, but the yielding of the tie under the pressure of the r^l must 
as well be considered. 

The relation of the bearing power of the tie to the amount it is depressed 
in the ballast is not thoroughly understood. 

The German engineers, Weber, Winckler, and Zimmerman, have advanced 
the theory that the pressure, P, of the ballast per unit of surface of the cross 
tie which it supports is, at each point, in direct ratio with the anldng, Y, of 
the latter; or P= CY when C is a coefficient depending upon the character 
(rf the ballast. The researches of these engineers may be summed up as follows: 

(a) The results of experiments are stated to agree quite closely witii the 
supposition that the pressure on the unit of surface is in direct proportion with 
the measure of the sinking. 

* (6) With a subsoil supposed to be good, the magnitude of the coefficient 
of ballast has been found: for gravel ballast (without metalled bed) C= 3; for 

* Piu kilograms per gquare ceatimeter; Y in centimetera. 



>, Google 



190 



STEEL RAU^ 



gravel ballast (with metalled bed) C = 8; for ballast of small stones and sccHise 

(c) The anking observed under a load in motion, at speeds varying from 
40 to 60 kilometers (24.85 to 37.28 miles) per horn', was not much greater than 
the sinking observed under the same load in a state of repose. 

Here again the fact that we are dealing with a dynamic load must be borne 
in mind, and at high speeds, when the dynamic augment of the wheel load is 
greatest, the bearing value of the tie corresponding to a given depression in 
the ballast is largely increased. 

The amount the tie is depressed in the track may be judged from the follow- 
ing evidence. 

Dr. P. H. Dudley gives from 0.2 inch to 0.4 inch as tiie amount the 
general running surface of the rail is below the trackman's surface. Director 
Schubert states that a wooden tie is depressed from 0.3 inch to 0.4 inch 
brfore it reaches a solid bearing. M. Couard observed that the maximum de- 
presdon of the tie was about 0.12 inch, and states that the amount of depres^on 
is not proportional to the load. 

Fig. 130 shows M. Cuenot's tests in which a depresMon is left under the 
tie of about 0.04 inch and the loaded tie is depressed about 0.12 inch. 

* Fig. 141 illustrates an apparatus used by Bell for measuring and recording 
the deflection of the rmls at various speeds. The following were the results 
obtained by tiie passage of a trsun in which the weights were: 





Tdh. 


Cwt. 




46 
33 

22 


12 
18 

71 


Total weight of six carriages. 


HpaKlofTnua. 




UilM p« Hour. 

4.2 
14.3 
26.7 
40.4 
57.1 
65.2 


0.25 
0.25 
0.27 
0.25 
0.33 
0.30 



The depression of the tie in the ballast is very erratic. Table LIX shows 
that in the tests made by the United States government on the depression of 



* The Development of the Manufacture and Use of KailB L 
;. of Civil Engra., Vol. CXLII, April, igOO, p. 133. 



Great Britain, Bell. Proceeding 



>, Google 



SUPPORTS OP THE RAIL 191 

rails the mean depresaon, under the drivers of an eagine having axle loads of 

44,000 pounds, was as follows: 

60-pound rail 073-inch deflection 

70-pound rail 138-inch deflection 

85-pound rail 233-inch deflection 

All of these depressions were obtained in gravel ballast with static wheel loads 

and give results the reverse of what might have been expected. 




Fio. 141. — Bell's Apparatus for Measuring Deprt 



Fig. 142 shows the relation of the depression to the pressure on the tie. 
The dotted hnes give Zimmerman's coefficients 3 and 8 and the dash line that 
suggested by Freeman's discussion in Article 21. These curves are strjught 

LOAD PER LINEAR INCH UNDER 0^4E LINC OF RAILS 
(TIES SPACED aOIN) TONS (2000 LBS) 



S P 
g o 



LOAD ON ONE HALF OF TIE TONS (2000 LBS> 
Fio. 142. — B«action of Tie. 

lines plotted through the origin, this appears to be Freeman's assumption, but 
in the case of Zimmerman's analysis, owing to different parts of the tie depress- 
ing unequally, some variation should probably be made from a straight line. 



,y Google 



192 STEEL RAILS 

It is qjiite evident that under the tie at the rail there is formed a depression 
of ballast, that even under a comparatively hght pressure the tie deflects to 
the depth of this depression, and that from this moment only is the relation of 
the deflection to the load of importance. 

From Mr. Love's analysis of the Government rail experiments (Article 23) 
we are furnished with a means of determining the relation between the pressure 
and deflection after the tie comes to a bearing in the ballast. The points plotted 
in Fig. 142 are obtained from Mr. Love's diagrams and represent the reaction 
of the tie referred to the depression measured from the highest point in the 
elastic curve of the rml between two drivers. The tie is assumed to come to a 
sohd be^ng at 0.20-inch depression below the trackman's surface. 

Bearing in nund that these points are obtained from a static !oad and 
that as far as the stresses in the rail are concerned the depth the tie depresses 
before it comes to a solid bearing is of comparatively small importance, we may 
construct the curve of pressures shown by the full line in Fig. 142. 

It is very probable that what really takes place is shown by the full line in 
the figure. The rail deflects under light pressures in some cases to as much as 
0.20 inch and the tie comes in contact with a compact bed of ballast and the 
pressure from this point rises very rapidly in proportion to the deflection. In 
general it was found from the government tests that the ties in the center of the 
span between the drivers on light rail were supporting very hght loads although 
in some cases they were considerably depressed in the ballast (see Plate XXIII) 
and for this reason it appeared that a better knowledge of the action of the 
ballast would be gained by referring the depresdon to the highest point in the 
rail between the wheels rather than to the trackman's surface. In the figure 
the pressure on the tie under the highest point of the rail between two drivers 
is plotted with a deflection of 0.20 inch below the trackman's surface and all the 
other deflections in the span referred to this. 

Above the limits of the experiments, the curve is flattened to provide for 
more rapid sinking caused by the increased pressure. 



,y Google 



CHAPTER IV 
stresses in the rail 

20. Stress at Point of Contact of the Wheel wfth the Rail 

Passing from an examination of the external forces acting ux>on the rail 
to a consideration of the resulting stresses they produce in the material of the 
rail, let us first examine the stress at the point of contact between the wheel 
and the rail. 

The essence of the wheel is that its theoretical bearing surface shall be a 
mathematical line or point, affording no area of bearing surface whatever. 
In practice this is not strictly the case, owing to the elastic compressibility of 
the surface, but the bearing surface is always very small, nor can it be increased 
to advantage by making either the wheel or bearing surface more compressible. 
To such bearing surfaces the ordinary compression moduli of the textbooks have 
no application, as they are derived from experiments upon prisms which have 
the same bearing surface as the greatest section, or nearly so. 





Fio. 143. FiQ. 144. Fia. 145. 

Condition of Free Flow. Partially Restricted Flow. Restricted Flow. 

Compression Moduli. (After Johnson.) 

* When a plain cylindrical column is subjected to a uniform compression 
stress over its entire cross section, as in Fig. 143, it may be said to be in a con- 
dition of "free flow," since it is free to spread in all directions throughout the 
length of the column. In Fig. 144 the material is compressed uniformly over 
a small area, as with a die. Here there is a flowing of the metal laterally, and 

• Paper contributed by Profeesor Johnson to the EngineerB' Club of St. Louk, December, 1892. 



,y Google 



194 STEEL RAII5 

then vertically, iinding escape around the edges of the die. This is a condition 
of confined or restricted flow, and evidently the elastic limit here will be much 
higher than with the simple column. 

In Fig. 145 the surface is compressed by a cylinder, the greatest distortion 
being at the middle of the area of contact. When this metal is forced to move, 
or flow, it can find escape only out around the limits of the compressed area. 
But at these limits the metal is very little compressed, and, hence, must be 
moved from the center. The confined ring of metal inside the limits of external 
flow is now much wider, and, hence, the resistance to flow much greater, so that 
this condition will be found to have a higher elastic limit stress than that shown 
in Fig. 144, and very much above the ordinary " elastic limit in compression " 
which is found for the free-flow condition of Fig. 143. 

A careful set of experiments was made by Professors Crandall and Marston* 
to determine the elastic limits of steel rollers on steel plates. In these experi- 
ments eleven rollers were employed, from one inch to 16 inches in diameter, 
with pressures varying from 1000 to 14,000 pounds. Their results showed 
that the elastic limit load with soft steel rollers on steel plates per linear inch is 

P - 880 D, 
where P = load in pounds per linear inch of roller, 

D = diameter of roller in inches. 

Professor Johnson t experimented to determine the area of contact between 
locomotive and car wheels and rails. Sections of wheels were mounted in a 
100,000-pound Riehle testing machine and short sections of rail were placed 
in the machine so that the wheel treads rested upon them in a normal position. 
They were then loaded with 5000-pound increments from 5000 to 60,000 
pounds, the area of contact being measured after each loading. These actual 
areas of contact are given two thirds actual size in Fig. 146, and in Fig. 147 the 
areas are plotted with the area of contact as abscissa and the loads as ordinates. 

Professor Johnson states that no permanent distortion was noted upon either 
rails or wheels at the contact surface up to the 60,000-pound limit. It is seen 
from Fig. 147 that these areas plot practically upon a straight line through the 
origin, indicating that the area is directly proportional to the load. This being 
true, it must follow that the load divided by the area of contact, or the average 
stress per square inch over the area of contact, is a constant for all loads. This 
constant is something over 80,000 pounds per square inch. 

" Friction Rollers by C. L. Crandall and A. Marston, Trans. Am. 8oc. of Civil Engrs., August, 
1894, Vol. XXXII, pp. 99-129. 

t DiscuBsion of Crandall and Maraton'a paper on Friction Rollers. Trans. Am. Soc. of Civil 
Engra., September, 1894, Vol. XXXII, pp. 270-273. 



,y Google 



STHESSES IN THE RAIL 



* Mr. Fowler, in his experiments on the relation of the load on the wheel 
to the area of the spot, found: 



L«doiiWb»l. 


An ot Spot. 


Pound.. 


Sqiun Inch. 


6,000 


0.11 


10,000 


0.12 


11,500 


0,13 


14,500 


0,15 


16,500 


0-17 


17,500 


0.18 


19,000 




26,000 


0.20 



The following conclusions are drawn by Mr. Fowler from tests on the 
contact areas between wheels and rails: f That the average pressure on the 
metal in wheel and rail is within the safe limits at low loads, but at a load of 
20,000 pounds the elastic limit is reached and permanent set begins in the rail; 






DIRECTION ALONG RAIL 



oooobOOOn 



20000 30000 40O00 



WooooOOOO 



9000 tOOOO 20000 SOOQO 40000 9000O lbs. 

Fig. 148. — Area ot Contact between Wheel and Kail. (Johnson.) 

that the accumulated pressure at the center of the contact area Ls excessive 
with comparatively small loads, and is only prevented from doing injury by 
the support of the surrounding metal; that the effect of difference in diameter 
in wheels under the same load is insignificant and only appreciable when the 
difference is great; that a hard and unyielding cast-iron wheel damages the 
rail more than a steel wheel, and the wear of the rail will be greater with cast- 
iron than with steel wheels. 

* ProceedingB Kttabun; Railway Club, November, 1907. 

t Bulletin of the International Railway Congress, London and Brussels, 1008, pp. 651-663. 
G. L. Fowler, Contact Areas between Wheels and Rails. See also G. L. Fowler, The Car Wheel 1907, 
p. 161. Giving the results of a series of investigations made for the Shoen Steel Wheel Co. 



>, Google 



196 



STEEL RAIIS 



* Hbnigsberg describes a method proposed to be applied to measure the 
actual forces between the wheel and the rail. This is based on the fact that 
polished surfaces of iron or steel show peculiar markings — sometimes known 
as Luder's Lines — on the limit of elasticity being exceeded. As the limit of 
elasticity can be artificially raised to any value between the primitive elastic 
limit and the breaking strength, this gives a means of making standard test 
I_OAD ON WHEEL IN POUNDS 




Fia. 147. — Relation between Ateae of Contact and Load on Wheel. (Att«r Johnmn.) 

pieces; since when lines appear it may be concluded that the artificially raised 
limit has been exceeded. 

If a wheel passing over two calibrated pieces of metal causes the lines to 
appear on one and not on the other, it may be concluded that the actual stress 
caused by the wheel lies between the elastic limits of the two standard pieces. 

Tire wear would seem to indicate that the elastic limit of the metal was 
exceeded or too closely approached. The investigation of a committee of the 
Master Mechanics Association in 1895, on the wear of locomotive tires, has thrown 
interesting li^t on this subject. 

* Measurement oT Forces between Rail and Wheel. 0. Hdnigaberg, Organ fQr die Fortachritte 
des Eisenbahnwesens, Wiesbaden, 1901, pp. 109-160. 



,y Google 



STRESSES IN THE RAIL 197 

* Fig. 148 shows a diagram of the average wear of the tires of the 
fifty-three ten-wheel engines for which the calculations plotted in Fig. 7 
were made, and Fig. 149 shows the same data of the eight-wheel engines 
shown in Fig. 6. 

The lower diagrams in Figs. 148 and 149 show the ratio of the rotative 
force to the weight on the rail, which we may call the " coefficient of slip." 
Since the coefficient of slip is the rotative force at the rail divided by the total 
weight of the drivers on the rail, it is evident that as this coefficient increases 
the tendency of the drivers to slip increases, and when it just equals the coeffi- 
cient of friction between the tire and rail the engine is on the point of slipping. 

The committee of the Master Mechanics Association in its report says: 

An examination of the tire wear shown in Figs. 148 and 149 shows no dis- 
tinct relation between the worn spots and the curves of maximum pressure 
of the wheels as given in Figs. 6 and 7. A very clear relation can, however, 
be traced between the worn spots and the parts of the wheel where the greatest 
coefficient of slip is combined with the heaviest wheel pressure. 

Local peculiarities of the tire, such as soft spots in it, as well as flat spots 
caused by slight sliding, aflfect the final contour of the worn tire, and it is only 
by taking the average wear at the same point on a large number of tires that 
the irregularities due to general conditions show themselves with the necessary 
clearness. 

Referring to the diagram of the average wear of the tire of the fifty-three 
ten-wheel freight engines shown in fig. 148, first we will consider the wear of 
the front and back tires only, as these wheels were overbalanced, the main 
wheels being underbalanced, and, on account of the effect of the angularity of 
the main rod, subject to quite different conditions from the others. 

Directing our attention to the wheels on the right side of the engine, an 
inspection of the figure shows quite uniformly, in both right forward and back 
tires, two locations of maximum wear, one beginning at about 160° and attain- 
ing its maximum at 220° or 230°, the other becoming pronounced at about 
10° or 20" and attaining its maximum at about 50°. It will also be noticed that 
both of these low spots are connected from 220° to 50° in the direction of rotation 
by a portion of the tire much more worn than that portion from 50° to 220°. 

To xmderstand the cause of this irregular wear, it is necessary to bear in 
mind that there are at least two ways in which driving wheels are slipped: first, 
when the slipping is slightiy but distinctiy noticeable, extending through but 
a small portion of the revolution; second, when the hold on the r^ is entirely 

* Proceedings Am. B.y. M. Mech. Aagn., Vol. 28. 



>, Google 




WEAR ON TIRES 



^■^LErr DBivei^"^ ZZ. ^^ 




COEFFICIENT OF SuP 30 MILES PER HOUR 



COEFFICIENT OF SUP 40 MILES PER HOUR 



COEFFICIENT or SLIP eO MILES PER HOUR 
Fia. 148. —Tire Wear, Ten-wheel Engioes. (Am. Ry. M. Mecli .\s8n.) 



,y Google 



STRESSES IN THE RAIL 



FROWT TIRE 



BACK TIRE !^ 



TROMT .TIBE % 



BACK TIRE 



■SA/EAR OF TIRES 




COEFFICIENT OF SLIP ENGINE JUS" 



^ 



^ 



COEFFICIENT OF SUP 40 MILES PER MOUH 



COEFFICIEMT or SLIP 60 MILES PER MOUR 

Fw. 149. — Tire Wear, Eight-wheel Eagines. (Am. Ey. M. Mech. Asaa.) 



,y Google 



200 STEEL RAILS 

broken, and the wheels slip through a number of revolutions, usually turning 
with considerable velocity. 

The first case, of slipping through but a small part of a revolution, occurs 
almost without exception on heavy pulls at slow speed, often being seen when 
an en^ne is pulling hard on a hill with just enough sand being used to avoid 
serious slipping, but not enough to prevent a slight slip at points where the 
rotative force is the greatest. The beginning of slip must occur under these 
conditions at or near the maximum of the coefficient of slip. Referring to 
Fig. 148, we find a maximum value of the coefficient of slip at 40° to 50°, and 
130° to 140° with engine just starting. At 20 miles per hour, the maxima are 
at 40° and 130°, and at this speed the tendency to slip at 100° is also almost as 
great as at the other points. The figure shows a small spot following 100° on 
the front tire, but none is seen on the back. The diagrams on Fig. 7 indicate 
the cause, as the pressure of these wheels upon the rail at 100° is almost at a 
minimum and is much less than at 140° to 160°. 

It is also noticeable that the amount of wear following 160° is greater than 
that following 40° or 50°, for the same reason. This variation in pressure upon 
the Tsil increases rapidly with the speed, and Fig. 7 shows very clearly that 
following 40° the pressure of the front and back wheels on the right side 
decreases very rapidly, while the reverse is the case following 160°. 

The same conditions as to pressure on the rail occur, for the left-hand front 
and back wheels, just 90° back of those on the right side, and irregularities of 
wear produced by the drivers slipping through a number of revolutions at consid- 
erable velocity should occur on the left wheels at points 90° back of the corre- 
sponding point on the right wheels: 90° back of 40° is 310°, and 90° back of 220° 
to 230° is 130° and 140°. Fig. 148 shows the greatest depth of wear of tires of 
the left front and back wheels to be abnost exactly at these points. There is also 
a small spot worn at 40°, due to the slipping at slow speeds when the influence 
of the counterbalance is nil. 

The irregularities of wear of the main wheels follow the same law as those 
of the front and back wheels, but the conditions are considerably modified 
by the difference in pressures caused by the influence of the angularity of the 
main rod, and to a less degree from these wheels being under- instead of over- 
balanced. 

The spots caused by the slight slipping at slow speeds at about 40° and 
130° should be found in these wheels as in the front and back wheels, unless 
the accompanying condition of necessary pressure is absent. Fig. 7 shows from 
16,500 to 17,000 pounds at 40° on the right main wheel, and from 12,700 to 



,y Google 



STRESSES IN THE RAIL 201 

17,500 pounds on the left wheel at the same point, indicating greater wear on 
the right than on the left tire at this point, which the diagram, Fig. 148, shows. 
The wear at 130° is found in these wheels, but, owing principally to the influence 
of the angularity of the main rod and partly to the wheels being underbalanced, 
the conditions of pressure following 130° on the right main wheel are very dif- 
ferent from those of the right front and back wheels. Fig. 7 shows that the 
pressure on this wheel is always rapidly decreasing following 130°, instead of 
increasing, and, consequently, the worn spot at this point extends but a short 
distance in the direction of rotation. Not so, however, with the left main tire. 
Here the pressure is always increasing following this point, and the figure shows 
the great elongation of this spot in the direction of rotation, extending it as 
far as 210°, while that on the right tire extends only to 165°. 

There still remains to be explained why the heavy spot on the main tire 
should slightly precede the point of the maximum coefficient of slip at 130°, 
and why that on the left wheel still farther precedes this point and, in general, 
is greater than on liie right. An inspection of the diagram on Fig. 7 shows that 
the pressure of the right main wheel on the rail is always greater preceding than 
following the 130° point. P^g. 148 also shows that the coefficient of slip is high 
as early as 110° after a speed of 10 miles per hour is attained, and increases 
but slightly to its maximum at about 130°. Any slipping occurring between 
110° and 130° will, on account of the pressure, cause a serious spot at this point 
on the main wheels, which the diagram shows. 

Fig. 148 shows the worn spot under consideration on the left tire, not 
only elongated in the direction of rotation, which is explained by the difference 
in pressure in this direction, but also in the opposite direction, extending beyond 
the 80° point. This is doubtless due to the slight slip caused by the main rod 
passing the forward center and suddenly thrusting this wheel back an amount 
equal to the lost motion in the bearing shoe and wedge. The same thing occurs, 
of course, on the right wheel, and the sharp, but slight, wear following the 350° 
point shows it quite clearly. On the left wheel, however, this wear is imme- 
diately followed by the more serious one due to the approach of the maximum 
point of coefficient of slip from 110° to 130°, and becoming merged into it, both 
are increased. 

The upper diagram on Fig. 149 shows the wear of the tires on the en^ne 
for which the calculations are plotted on Flig. 6 for the eight-wheel engine. This 
shows in a general way the same characteristics of the average wear for the 
fifty-three ten-wheelers, shown on Figs. 7 and 148, but is, undoubtedly, affected 
to a considerable extent by unknowable local conditions. Here the front wheels. 



,y Google 



202 STEEL RAILS 

of course, correspond most nearly to the main wheels on the ten-wheeler, and 
here, as there, the left main tire shows the most serious irregularity of wear. 

The committee presented the following concluaons as a result of its investi- 
gation, which it should be borne in mind was in connection with much lighto* wheel 
loads than obtain at the present time. 

" There is no doubt locomotive tires wear without slipping, and there should 
be, and probably is, a portion of the irregular wear due to the pulverizing or 
crushing action being greater under heavy than light loads. 

" An experiment was made by removing all the overbalance in the counter- 
balance of an engine, when the irregularities of wear in the main wheel were 
almost exactly duplicated in location and to a remarkable degree in magnitude. 
This, together with similar experiments attended by the same general results, 
leads us to believe that the irregularities of wear of the tire are almost wholly 
caused by abrasion from slipping, and that the pulverizing of the steel from 
pressure alone is of secondary importance." 

With the smaller wheels under the cars and locomotive tenders the con- 
ditions are quite different.* The smaller steel wheel or tires do not render as 
satisfactory service under heavy loads and high speed as the larger locomotive 
tires, principally for the reason that the manufacturer is not able to put into 
the small tire sufficient mechanical work to obt^n uniform physical properties 
for the full circumference of the tire, and in the service portions of the tread 
thickness. 

The experience with 36-inch steel-tired wheels under locomotive tenders 
illustrates this point. The steel tender wheels have failed by shelling out, and 
portions of the metal of inferior physical structure on one-third of the circum- 
ference have worn so as to make an eccentric tire which has caused such severe 
impacts on the rail as to require the removal of the wheels after a short service. 
The average load on these wheels is 18,000 pounds, and the maximum static 
load 20,400 pounds; many of them ^ve only six months' service, or 30,000 
miles, after a first or second turning. 

In a paper read at the last International Railway Congress, published in 
the Bulletin of October, 1911, Dr. P. H. Dudley has proposed a new method of 
measuring the tonnage service of rails and wheels. He explains that the ton- 
nage supported by a ^ven portion of the bearing surface of a rail due to a pass- 
ing wheel is the total load multiplied by the number of wheels passing over it. 
The tonnage sustained by the metal in the treads of the wheel is the total wheel 
load multiplied by the number of revolutions, and this tonnage accumulates 

• Railway Age Gazette, December 22, 1911. 



>, Google 



STRESSES IN THE RAIL 203 

more rapidly than that of the nul. It is greater also as the diameter is less 
and the number of revolutions larger, so that the tomiage service of 36-inch 
tender wheels is much greater than that of the 75- or 80-inch drivers, though 
the loads on the latter may be much larger. 

The pressure and movement of heavy loads on the rail causes a cold rolling 
to take place on the head of the rail, which tends to expand the metal and, if 
the rail were free to move, would cause it to assume a curved form with the 
head on the convex surface.* As the rail cannot bend, this cold-rolled metal 
is subjected to a compression stress. A tensile component would be expected 
in the vicinity of and next below the part which was affected by compression 
strains, and this has led to a theory of the cause of the oval silvery spots or 
transverse fissures in the head of the rail observed by Mr. Howard.f These 
r^ fissures which resanble the smooth surfaces of a progressive fracture have 
so far been largely confined to one steel cwnpany. The theory advanced as to 
their probable cause should not be regarded as final until further substantiated, 
and many eng^eers feel that it is not the true explanation. No other adequate 
reason has been offered as yet to explain their formation, although a careful inves- 
tigaHon is bemg made which will doubtless throw further hght upon the subject. 

Mr. Howard J states that: " The flow of the metal of the head, appar- 
ent to the eye and witnessed very generally in portions of the track, may 
be taken as evidence of exhausted ductility of the metal. The ability of 
the steel to elongate, as found in the primitive state of the rail before going 
into service, is lost by reason of its development, and the rail, at first tough 
and capable of being bent, is now brittle and will bend only to a lunited extent 
before rupture. 

" The brittleness is due to the flow of the metal at or immediately below 
the running surface of the head. The structural continuity has not been de- 
stroyed, as may be shown upon annealing the metal, which effects a restoration 
in its ability to elongate. A rail from service will not bend well with the head 
on the tension side, ance the surface metal has been subjected to cold flow in 
advance of its being worn away by abrasion." 

* Mr. Howard found that in the case of a rail, exposed to this action, on cutting ofi the head froin 
the web, the former sprung into a curved shape with the running surface on the convex side. The 
deflection at the middle of the length of the piece, 5 feet loDg, was 0.20 inch. It is probable, however, 
that some of this curvature was caused by the strains set up in the rail when cooUng. 

t Appendix to Report by the Interstate Commerce Commission on Accident to a Lehigh Valley 
Railroad Train at Manchester, N. J., on August 25, 1911. See also Broken Lehigh Valley Rail, Iron 
Age, Vol. 88, Part 2, p. 800. 

t Some Causes Which Tend toward the Fracture of Steel RmIs. James E. Howard, Journal 
AsoodstJon of En^neering Societies, July, 1908. 



>, Google 



204 STEEL RAILS 

Removing the surface meta], in the planer, restores the bending quali- 
ties of the rail, but in this case it is necessary to plane away the metal from the 
sides as well as from the top of the head, that is, as far down as the cold flow 
has taken place. 

The difference in the bending qualities of the same rail according to 
the head being on the tension or compression side is shown by Fig. 150. The 



Fig. 150. — Two Piepps of a Worn 100-Ib. Rail after Testing. The 
upper piece, with head on compresaion side, bent 21 degrees with- 
out rupture. The lower piece, with heiul on l«nsioa side, bent 
4i degrees and then ruptured. (Howard.) 

upper piece of rail in the figure was bent with the head in compression, while 
the lower one had the head on the tension side of the bend. 

Rails of this series of tests have ruptured with a deflection of only 3 to 5 
degrees when the head was in tension, but remained unruptured when bent 
through an angle of 20 degrees or more with the base in tension. After an- 
nealing these old rails, of exhausted toughness, the bending qualities were 
restored, after which the rail could be bent in either direction through about 
the same number of degrees without fracture. 



,y Google 



STRESSES IN THE RAIL 



205 



The effect of the exhausted metal in the head is well illustrated by 
Table XLVIII, which presents some of the results of Kirkaldy's tests 
on rails. * Kirkaldy states that " the rail appears to have been gradually 
hardened under the action of the traffic, more especially on the immediate 
sldn or surface, until the steel thereon cracked under the upward flexion of the 
rail in the regions just over the chairs, or the minute cracks may have been 
induced by the severe action of the brakes on trains, so to speak, tearing up 
or disintegrating the surface of the steel." 



TABLE XLVIII. - 


BENDI 


NG TESTS ON 


WORN 


RAILS. 


(Kirkaldy.) 




(Dislanoebelw 


ennuHBrU 


Ifeet, lou 


dmppliHln 


™wc) 










Dime 


aaaa. 


Strw. 


Deflection 




























Dfptl.. 


Wrfj. 


El«lic. 


Ummat.. 


f^ 


UltiDUtft 






PouodB. 


locbn 


Inch. 


Pounds. 


Founds. 


Inch. 


Incl™. 




Worn rait; SO Iba. nuin li». 23 

>«an' »rvicc. 
Same. t«Md iBV«n»d. 


jn,» 




.« 


i MJOO 


M.780 
H.2M 


O.M 




1» 


RamovHi uuraeked. 
Snapped, 


iima. UaUd iDvartsd. 




*.» 


.t» 


{ «:SS 


'S^ 


o!28 




t 


Removsd uncracksd. 


n'orn rail; M lb*., in iwd 1 

yonnwilbbaavy WBK. 
Sune. taud igverUd. 




S.4S 


.t» 


( 47,800 


88.710 
88.130 


0.24 







RetnoVBd uncracked. 
Ren.D'.-ed buckling. 


Won.™[;a!lb...lO)-«n.»r- 




.» 


.H 


1 40,000 


73.130 


0.31 







Snapped. 


Wom rail; sipoaKi lo breks 


j «.o 


4.80 


M 


1 34.400 


81.440 


0.84 















1 31.500 


«,*00 






38 


Snnppwl; tli(h( flnir. 







The slipping of the driving wheel of the locomotive when starting a tr^ 
may cause roughness of the metal of the rail, accompanied by intense heating 
of the immediate surface metal of the head. In addition to the loss in ductility 
of the steel by reason of its flow under the wheel pressures, the metal at the 
running surface is hardened through this action of the wheel. Showers of 
sparks attend instances of this kind, from which the high temperature acquired 
by the particles of the steel may be judged of. There follows also a sudden 
reduction in temperature through conductivity of the cold metal below, which 
has an effect similar to quenching steel from high temperatures in water or 
other quenching liquids, and there results a surface hardening of the metal. 
During this period of hardening the surface metal is placed in a state of intense 
tension, relief from which is obtained by the development of cracks in the steel. 

A very interesting experiment is reported by Wickhorst t on the flow of 
metal in the rail head under the wheel load. The test was made to determine, 

* Kirkaldy on Effects of Wear upon St«el RaJIa, Appendix II. Min. of Proceedingn of the Inst. 
of avil Engre., Vol. CXXXVI, January. 1899, p. 166, 

t Flow of RmI Head under Wheel Loads. M. H, Wickhorst, Proceedings Am. Ry. Eng. & M. 
of W. Awn.. 1911, Vol. 12, Part 2, p. 535. 



,y Google 



206 STEEL RAIIB 

if possible, what change ia made in the microstructure of the head of the rail 
by the rolling over the rail of heavy wheel loads. At the same time, measure- 
ments were made of the spread of the head and the width of the bearing produced. 
The test was made on a new 70-pound Bessemer rail with a reciprocating 
machine in which a piece of rail is moved back and forth under a wheel to which 
a load can be applied by means of levers. A diagram of the machine is shown 
in Fig. 151, from which it is seen that the rail is fastened to a steel bloom 





- 


— 


-^. f— ^^ hi 1 










5TEEL BLOOM / \ 




« 






n— 


-8(1 


















STHL BLOOM 


I 




























Fig. 151. — Reciprocating Machine for Testing Flow of Metal in Head of Rail. 
(Am. Ry. £ng. Asan.) 

which runs on rollers running on another steel bloom that forms the bed of the 
machine. The rail bed is connected by means of a connecting rod to the bed 
plate of a planer, which furnishes the power to run the rail machine. The 
weights applied to the weight hanger are multiplied 600 times, as applied to 
the axle of the wheel. The piece of rail tested was 12 inches long, which was 
set up between two other similar pieces, which acted as end pieces onto which 
the wheel could roll when leaving the piece under test. The piece tested had the 
sides of the head planed vertical to a width of head of 2 inches. This width was 
used partly to accentuate the test and partly to do away with the rounded 
comer, so as to allow of measuring the width closer to the top of the head, and 
the sides were made vertical so the measurements could be made satisfactorily 
with a micrometer along the whole depth of the head. The section of the 



,y Google 



STRESSES IN THE RAIL 



207 



ori^nal rail and as tested are shown in Fig. 152. Before testing, the width 
of the head was determined at the top, at the bottom and halfway between, 
both at the middle of the piece tested and at one end, by means of a micrometer. 
To determine the sag of the head, two prick-punch marks were put on each dde 
of the rail at the middle of its length, one on the side of the head near its bottom 
and the other on the top side of the base, about f inch from the web. In order 
to have a vertical side on which to prick-punch the mark, the base was gouged 
at the deared place. The distance between the marks was measured in .01 




Fig. 152. — Sectioa of 70-lb. Besaemer Rail Tested for Flow of Head. 
(Am. Ry. Eng. Asbd.) 

inch by means of a fine-pointed toolmaker's dividers and a steel scale reading 
to .02 inch. 

The test was started with a load of 30,000 pounds applied to the wheel, 
using 1000 double strokes or 2000 rollings of the wheel over the rail under 
test. The bearing assumed a width of .64 inch. The only effect on the width 
of head was to spread the top of the head .002 inch, and the load was, there- 
fore, increased at once to 60,000 pounds and the teat continued until the head 
seemed to no longer spread as measured with the micrometer. The width of 
the head of the rail and the width of the wheel bearing on the rail, at various 
stages of the test under the load of 60,000 pounds, are shown in Table XLIX. 
The spread of the top part of the head and the width of bearing in this table 



,y Google 



208 



STEEL RAII£ 



includes also the effect of the preliminary rolling of 2000 wheel applications 
with 30,000 pounds. The head did not show any sag throughout the test. 

TABLE XLIX.— ROLLING TESTS ON RAIL HEAD WITH LOAD OF 60,000 POUNDS 







SprwdofHsBd. 














RoUiiVB. 


Buriiif. 
















TopolHBKi. 


MlddUiot 


H^. 


Top of H«d. 


HgdJeC 


Bottom or 






.002 


.000 


.000 .004 






2,000 


.92 


.009 


,003 




.006 


-002 


.000 


4,000 


.84 


,010 


.004 


-001 


.006 


.002 


.000 














.003 




32.142 


1.04 


,013 


.006 


.001 


.008 


.003 


.000 



It should be remarked that the wheel was beveled some, and the bearing 
was, therefore, on one side of the top of the head, remaining throughout about 
.2 inch from one side and increasing in width toward the other side. 

A microscopic examination was made of the rail at the top of the head and 
at the center of the head both before and after rolling. While the micro-photo- 
graphs obtained indicated a slight stratification of the grains in a longitudinal 
direction, there was little, if any, difference between the specimens before and 
after rolUng. The material tested was good ductile material of medium hard- 
ness for rail steel, and as the maximum lateral stretch at the top of the head 
was only .013 inch, much difference in the microstructure could hardly be 
expected. 

It is evident that the metal in the bead of the ral must have a high elastic 
limit to successfully meet the severe conditions of modern service. This fact 
was clearly brought out several years ago by one of the writers in connection 
with some service tests on annealed rails on the Philadelphia and Reading Rail- 
way. An account of this investigation appears in the Proceedings of the New 
York Railroad Club, December, 1906. Eleven 90-pound rails were sawed into 
halves, and one half of each rail was annealed. The carbon content averaged 
0.54 and the manganese 1.06 per cent. After 88 million tons traffic it was 
found that the annealed rails averaged 31.9 per cent more wear and they also 
showed a greater tendency to crush and splinter, but it was found on test that 
the elastic limit had been reduced over 10 per cent. The annealed rails, in spite 
of their finer structure and consequent greater toughness, did not wear so well 
on account of the lower elastic hmit. 

In Article 9 attention was called to the dfect of the inertia of the track on the 
stresses produced by impact, and in Artide 8 it was shown how the lack of round- 



,y Google 



STRESSES IN THE RAIL 209 

ness of the vrbed may cause excessive strains in the running surface of the head. 
These factors make it much more difficult to control this stress than that produod 
by bending; in the lattra- the forces acting on the r^l can be determmed witiiin 
closa* limits and the remedy is easier to apply. 

The rail is in fact called upon to perform two quite distinct functions, one of 
which is to resist the strwn produced at the area of contact between the wheel and 
the rml and the other to reast the bending stress; the latter can be reduced by 
increasing the moment of inertia of the section or strengthening the track structure, 
but the former is in a measure independent of the form of the r^ and requires a 
change in the characta" of the material of which the rail is cranposed. 

The defect known as "roaring rails" is caused by an imperfect surface or 
corrugations in the head of the rail. These corrugations are confined almost 
exclxisively to the rails used on electric roads and few problems, confronting the 
maintenance of way engineer of such roads, have attracted the attention and study 
being ^ven to this trouble. 

There are many conflicting opinions as to the cause of the phenomena, 
none of which appear adequate to properly explain it.* 

The corrugations of rails in recent years have increased rapidly in number; 
once they start they rapidly grow worse and it is important to remove them as 
soon as the indentations appear. This is generally accomphshed by means of a 
rail grinding device which consists either of a carborundum block rubbing over 
the rail or an emery wheel which grinds the rail to a true surface. 

The cost of removing corrugations in rails varies from a few cents to 60 
cents per foot of rail, depending on the depth of the waves; fortunately 
after the corrugations have been removed there is little probabiUty of their 
ever returning. 

* Some of the recent literature on this subject ie aa follows: 

Andrews, J. H. M. — ^Some notes on rail corrugation. 1500 w. 1910. (In Electric Rail- 
way Journal, Vol. 36, p. 370.) 

Outlines prominent causes of corrugation and presents a few notes and coDclusiooa. 

BusBB, A. — Rail coiTugfttJon. 1500 w. 1910. (In Electrician, Vol. 65, p. 930.) 

Observations from many points indicate strongly that corrugation is due primarily to defects in 
tbe rail metal, resulting from the rolling. 

Panton, Joseph A. — Rail corrugation. 26 p. HI. 1907. (In Journal of the Institution of 
Electrical Enpneers, Vol. 30, p. 3.) 

Concludes, in summary, that corrugations are caused, directly or indirectly, by lateral play in 
weak trucks, the weakness being intensified by vmsymmetrically driven axles. 

WiLSOK, C. A. Carns. — Rail corrugation. 3000 w. III. 1908. (In Engineering, VoL 86, 
p. 90.) 

Aims to show conditions under which corrugations are produced. 



>, Google 



ST££L RAII£ 



21. Proposed Solutions of the Bending Stress in the Rail 

Following the path of the forces as they pass through the rul and are 
distributed to the ties, we find very complex and unstable conditions. The 
rail is supported on a series of yielding supports. These supports, through 
their unequal yielding, bring about distributions of stress in the rail that are 
difficult to calculate. 

Before proceeding further with the discussion of this subject, let us turn 
to some of the methods advanced for the proper solution of the problem. 

* Mr. 0. E. Selby approaches it in the following manner: " Examining 
first the bending stress in the rail, we have 50,000-pound axle loads on supports 
20 inches apart. For these conditions the American Railway Engineering and 
Maintenance of Way Association specifications for steel bridges, paragraph 5, 
call for 100 per cent impact, making the stresses equivalent to those from a 
100,000-pound axle load, or a 50,000-pound wheel load. 

" For a simple beam the bending moment in one rail would be 250,000 
inch-pounds. For a continuous beam with rigid supports, it would be two- 
thirds that, and for a continuous beam with partially yielding supports, three- 
fourths of the bending moment for a simple beam is reasonable, giving 187,500 
inch-pounds. If we consider the wheel placed over a tie which yields enough 
to carry one-fourth of the load to each adjacent tie, the resulting moment in 
the rail is the same. The section modulus of an 80-pound A. S. C. E. rail is 
10.0, making the extreme fiber stress 18,750 pounds per square inch. For a 
100-pound rail the unit stress is reduced to 12,800. 

" Pasdng to the load on the tie, we encounter an element which must 
vary between rather wide limits with the stiffness of the rail and yielding of 
the supports. With a simple beam and load placed midway between the sup- 
ports, the reaction on each support would be one-half the load. The theory 
of the continuous girder would make the reactions about 55 per cent to 67 per 
cent, depending on whether the load is placed over a support or midway between. 

" The yielding of supports would undoubtedly reduce these percentages. 
Bridge spedfications usually consider the load equally distributed among three 
ties, but bridge ties are spaced usually 14 inches between centers, so that, if 
the load going to one tie is proportioned to the tie spacing, the amount for 
20-inch spacing would be 20 -^ 14 X 1 -^ 3 = 20 -=- 42 = 47.6 per cent. There- 

* 0. E. Selby, Bridge Engineer, C, C, C. & St. L. Ry. A Study of the Streesee Existing in Track 
Supentructuree and Rational Design Baaed Thereon. Proceedings Am. Ry. Eng. & M. of W. Assn., 
1907, Vol. 8. 



,y Google 



STRESSES IN THE RAIL 



211 



fore, the assumption that the maximum load on a tie is half the axle load seems 
a proper one. 

" Modem specifications call for an E-60 loading, which contains 60,000- 
pound axle loads, spaced 5 feet between centers. A tie spacing equal to one-half 
the wheel spacing would load the ^rder (r^) at the quarter points and produce 
moments (in a simple beam) equal to those produced by a uniform load. 

" A more practicable tie spacing, one-third of the wheel spacing, would 
similarly produce moments ^ part, or only 1.4 per cent greater than those 
from a uniform load, so that Lf we design a rail for a uniform load equal to the 
wheel load divided by the wheel spacing, the result would be very nearly correct. 
The wheel load with impact is 60,000 pounds, or 1000 pounds per linear inch 
of rail. For a continuous beam (indefinite number of spans and loads) the 
maximum movement is 1 + 12 x WL^ = 1 -^ 12 x 1000 x 60* = 300,000 
inch-pounds." 

The following has been presented by Mr. Bland: * 

" The rail acting as a beam under passage of wheel loads is in a condition 
of ' restrained ' ends, and the maximum moment from a wheel load Q is ^ven 
by M = ± i QI, ' Q ' bemg concentrated load and ' I ' being span center 
to center of ties. The moment alternates from positive to negative, and alter- 
nates equally. The dynamic augment to static wheel load is taken 60 per cent 
for a speed of 75 miles per hour. 

" Assume a static axle load of 60,000 pounds, giving static wheel load of 
30,000 pounds. The dynamic augment for 75 miles per hour is 60 per cent, 
making a dynamic wheel load of 48,000 pounds." 

TABLE L. — RAIL STRESSES. (Bland.) 



Rail W«itht. 


^^ri^' 


m'SSiiS-z.' 


Dyumio Uomnt- 


RnultlDE Van 


P«ihU per Yard. 
60 

70 
85 
100 


iBCllH. 

24 
24 
22 
22 


6.70 
8.30 
11,30 
15.00 


lMh-p«md*. 
144.000 
144.000 
132.000 
132,000 


Lbs. per Sq. In. 
21,500 
17,350 
11,680 
8,800 



The following investigation of the stresses in the rail was suggested by 
Mr. F. B. Freeman, in a paper presented to the New York Central Lines Main- 
tenance of Way Committee entitled " Investigation of Stresses in Track Super- 
structure." This paper is of considerable interest, as the results obtained by 



* J. C. Bland, Engineer of Bridges, Penn. lines West of Pittsburg, 
Heavy Roil tor Eristing Heavy Engines, 1907. 



Q the Capadty of Modem 



>, Google 



212 STEEL RAIIS 

the methods proposed are compared with the stresses actually g^ven by test 
of the rails mider the moving trains. An abstract of the paper follows: 

By means of plotted curves recently published showing resultant wheel 
loads in terms of the unit static loads at various speeds, Dr. P. H. Dudley shows 
as a result of his experiments that with smooth wheels rolling without accelera- 
tion, the impact approaches 50 per cent as a limit at 100 miles per hour. A 
curve of resultant wheel loads at various speeds with acceleration shows 100 per 
cent impact as the limit at 100 miles per hour. At speeds of 50 to 70 miles p^ 
hour the impact appears to be about 75 per cent when the train is under 
acceleration. 

By acceleration it is meant that the locomotive is exerting its maximum 
tractive effort at that speed. 

From his strenunatograph tests, Dr. Dudley finds that the maximum 
extreme fiber stresses in the 100-pound rail under the present Class " I " loco- 
motive (Atlantic type) sometimes run as high as 22,000 pounds, while with the 
80-pound rail stresses as high as 28,000 pounds are not uncommon. One thing, 
however, must be borne in mind: though the extreme fiber stress in steel rails 
may seem high as shown by test, these maximum stresses are of very short dura- 
tion, lasting but a small fraction of a second and often reversed immediately. 
Within certain Hmits the stress seems to vary directly as a function of the si)eed; 
hence the greater the speed, the greater the stress, but the shorter its duration. 

The present steel rails have an elastic limit between 50,000 pounds and 
60,000 pounds and an ultimate strength of 110,000 pounds to 120,000 pounds 
as shown by test. 

If we investigate the stresses which are supposed to exist under the Class 
" I " locomotives and then compare the resultant stresses with those of tests, 
we may arrive at a conclusion regarding the trustworthiness of our assumption. 

As the Atlantic type of locomotive (Class " I ") cannot draw an ordinary 
train at a greater speed than 70 miles per hour with acceleration, we are justi- 
fied in using 75 per cent impact in investigating the existing stresses under 
Class " I." 

The Pacific type (Class " K "} and the electric locomotive (Class " T ") 
each may haul tr^ns at speeds approaching 100 miles per hour,* and the former 
will be investigated with 100 per cent impact. 

As the wheels roll along the rail there is a general depression of the track, 
local depressions being greater under the heaviest concentrations, and fairly 
uniform where the wheel loads are equal and evenly spaced. According to 

* This statement seems open to question; compare with Article 3. 



>, Google 



STRESSES IN THE RAIL 213 

Dr. Dudley, this general depression varies from .10 inch to .20 inch, varying 
with the stiffness of the rail, elasticity of subgrade, and tamping of ballast. 
Due to the deflection of the rail, there will ne a greater depression under the 
wheels and tiie lesser depression about midway between them. The tie pressure 
may be assumed proportional to the tie depression, and from the tie pressure 
the stress in the rail may be approximated. 

In considering the deflection of the rail, we may consider the rail as a con- 
tinuous beam being supported by the wheels and having varying concentrated 
loads applied by the ties. These concentrations decrease toward the center 
of the span and give an effect practically similar to that of the uniform load, 
equivalent to the sum of the tie pressure (Fig. 153). 





'^mfmrffntffffT 



Fia. 153. — Distribution of Tie PrMsure under Rail. (Freeman.) 



In order to approximate the rail deflection, there will be no great error in 
considering the load of the drivers and second wheel of the first truck as uni- 
formly distributed from a point between the two wheels of the forward truck 
to a point midway between the rear driver and the trailer; likewise that the load 
of the trailer be distributed from a point midway between the rear driver and 
the trailer to a point midway between the trailer and the first wheel of the tender. 

The load of the drivers of the present Class " I " locomotive, including 
impact, is 96,000 pounds, which will be distributed over approximately 19 feet, 
giving an equivalent uniform load of 6400 pounds per lineal foot. Under the 
trailer the equivalent uniform load will be 4000 pounds per lineal foot, and 
under the front wheel about 3000 pounds per lineal foot (Fig. 154). 

Under the drivers we have a span of 7 feet and a uniform load of 6400 
pounds per lineal foot. With the 100-pound rail this loading would cause a 
deflection of .047 inch and with the 80-pound rail a deflection of .08 inch. 

The average uniform load between the rear driver and the trailer being 
about 5000 pounds per lineal foot, the deflection of the 100-pound rail becomes 
.127 inch, while the 80-pound rail deflects .22 inch. 

It will be seen by a glance at Fig. 155 that the tie reactions vary with the 
stiffness of the rail, being much more uniform with the heavier rail. 



,y Google 



STEEL RAILS 






SrOO lbs. PER LIN. FOOT 



6400tb& P. L.F. 



Fig. 154. — Class "I" Engine with 75 per cent Impact. (Freeman.) 

o OOo 




Fio, 155. — Track Depression under Class "I" Loading. (Freeman.) 

Plotting the most probable deflection of the 100-pound rail (shown in solid 
line) and the 80-pound rail (shown in broken line), and then taking the tie re- 
actions as proportional to the ordinates to the curves of deflection, we may 
expect the following depressions and reactions: 

"I" LOADING 



Ti* 


IDO-pouDd lUil. 


SO-pouDd lUil. 


DeprBBHOD. 


Bearim. 




Baarinc. 




iKb. 

0.10 
0.15 
0.19 
0,15 
0.10 
0.10 
0.17 
22 
0.22 
0.18 
0,18 
0.22 
0-22 


Paa«]e. 

5.300 
8,100 
10.200 
8,100 

5.300 
5,300 
9,400 
11.800 
11.800 
9.700 
9.700 
11,800 
11.800 


iMh. 

0.04 
0,18 
23 
O.IS 
004 
0.04 
17 
02 
024 
17 
17 
24 
0.24 


PodikIi. 

2,200 
10,000 
12,600 
10,000 
2,200 
2.200 
9.500 
13.400 
13.400 
9,500 
9.500 
13.400 
13,400 




c 




e;;."*": 






H . 






K . 









,, Google 



STRESSES IN THE RAIL 



215 



If we use these tie reactions and assume with Winckler that the moments 
in a continuous girder over yielding supports are three-fourths those of a 
simple beam, we may compute the bending moments and the extreme fiber 
stress in the rail. 

Taking moments under the rear driver, we find an extreme fiber stress of 
19,430 pounds with the 100-pound rail and 28,950 pounds with the 80-pound 
rail. Under the trailer we find 17,750 pounds in the 100-pound rail and 19,800 
pounds in the 80-pound rail. 

These stresses given for Class " I" are practically those found from the 
stremmatograph tests. 

According to Dr. Dudley, when the Class "I" type engine strikes a piece 
of soft or rough track, it begins pitching or rocking about its center of gravity, 
located between the two drivers. This rocking increases the pressure on the 
trailer sometimes as much as 50 per cent. This tends to increase the stresses 
" -under the trailer by that amount. In the Class " K " type, this difficulty is 
obviated to a great extent by the increased length of wheel base. 

We may now compare Class " K " with Class "I," using 75 per cent impact 
and considering the two engines working under similar conditions. 

Following the same analyas as in the investigation of Class "I," we may 
assume the weight of the drivers and the second wheel of the first truck as 
uniformly distributed over a length of about 28 feet of rail, giving a uniform 
load of 6500 pounds per lineal foot. Under the trailer we find a uniform load 
of 3800 pounds per lineal foot, and under the tender a load of 3700 pounds per 
lineal foot (Fig. 156). 




tIJ 



i-.<L^ ^ ^-1^ ^i-l 



4^^^4) 



87O0 lbs. PER LIN. FOOT 



3aOOItMi!L.F. 



ftSOOtt^CR L 



Pio. 156. — Class "K" Engine with 75 per cent Impact. (Freeman.) 

Computing the deflections of the rail, we find .049 inch for the 100-pound 
rail and .083 inch for the 80-pound rail on the 7-foot span under the drivers; 
on the 10-foot 11-inch span between the rear driver and the trailer a deflection 
of .24 inch is shown by the 100-pound rail and .42 inch by the 80-pound ral. 



,y Google 



STEEL RAILS 



Plotting the curve of probable deflections for the 100-pound and 80-pound 
rail and using a general average depression under the drivers, proportional to 
the uniform load under the drivers, we may find the tie depressions (Fig. 157). 




FiQ. 157. — Track Depression under Class "K" Loading. (Freei 



The depression under a uniform load of 6500 pounds is taken at .20 inch. 
Upon this basis we get the following tie depressions and reactions: ' 

TABLE LIL — TIE DEPRESSIONS AND REACTIONS, CLASS "K'' LOADING 

(FREEMAX) 





l»|»UBd Riiil. 


«v^.«aiuii. 




Ti», 


D^,^ion. 


BwriBg. 


DapTMBon. 


BaiiDg. 




1.75 P„«,. 


l-imPSTMBt. 




I-7SPer(»t. 


I-lOOPercsrt. 




0.05 
0.05 
14 
0.20 
0.20 
0.14 
0.05 
0,05 
0.13 
0.23 
0,23 
0.18 
0.18 

0^18 
18 


3,000 

3,000 
7,000 
10,000 
10.000 
7,000 
3,000 
3.100 
7,900 
14,000 
14,000 
11,000 
11,000 
14,000 
14,000 
11,000 
11,000 


3,400 
3:400 

8,000 
11,400 
11,400 
8.000 
3,400 
3,500 
9,000 
16,000 
16,000 
12,600 
12,600 
16,000 
16,000 
12,600 
12,600 


0.03 
0,03 
0.11 
0.25 
0.25 
0.13 
03 
0.04 
0.14 
0.24 
0.24 
0.16 
0-16 
0.24 
0.24 
0.16 
0.16 


1,560 

1,560 
5,700 
13,000 
13,000 
6,800 
1.100 
1,300 
9,100 
15,600 
15,600 
10,400 
10,400 
15,600 
15,600 
10,400 
10,400 
























.3 




I 


1,600 


L 

M 


17,800 
17,800 
11,900 
11,000 




P 


17,800 
11,900 







Taking moments under the rear driver, we find an extreme fiber stress of 
22,000 pounds in the 100-pound rail and 32,150 pounds in the 80-pound rail. 
These stresses may be expected up to speeds of 60 miles per hour. At speeds 
of 90 to 100 miles per hour the impact will approach 100 per cent and the stress 
in the 100-pound rail may run as high as 25,200 pounds and up to 36,600 pounds 
in the 80-pound rail. These are stresses which may be expected in ordinary 



,y Google 



STRESSES IN THE RAIL 217 

main-line track, at maximum speed. On a soft piece of track they may run 
up 25 per cent higher. 

Zimmerman's analyas of the stresses in the rail * is based upon the stiffness 
of the rail and tie and on the compressibility of the ballast. For the bending 
moment in the rail he derives the following equation: 



87+7 



G^ = 



in which 



" 4-r + lO 4 4 7+10"' 

Ma = Maximum bending moment in the rail. 
Too = Bending moment of a simple beam loaded in 

the middle of the span a with a load G, 
G = Wheel load, 

a = Distance center to center of ties, 
B 



"" V4E/ 

E = Modulus of elasticity of the rail, 

/ = Moment of inertia of the rail, 

6 = The width of the tie, 

I = One-half the length of the tie, 

C = Coefficient of the ballast, 

2 = An auxiliary value depending on the form of the tie. 

The coefficient of ballast represents the pressure in kilograms per square 

centimeter of the ballast which causes a depression of one centimeter. The 

coefficient 3 corresponds with simple gravel and the coefficient 8 with gravel 

on a bed of dry stone or on rocky soil. 

M 
Fig. 158 shows that the ratio — ^ increases with y, that is to say, with 

the stiffness of the rml and the flexibility of the tie. 

Table LI II presents calculations for several German railroads by the 
aid of Zimmerman's formulae. The table is taken from an article on " The 
Track Superstructure of German Railways " by M. Blum in the Revue Gen^rale 

■ Calculation of the Superstructure, Berllu, 1888. 



,y Google 



218 STEEL RAILS 

des Chemins de Fer, No. 5, November, 1908, a translation of which is pven 
in Proceedings American Railway Engineering and Maintenance of Way 
Association, Vol. 11, Part 2, 1910. 



Pio. 158. — Bending Moment of Rail placed on Ties. (Zi 

In questions of such moment we cannot rely on mathematical analyses for 
conclusions. We can deduce general results from specific experiments by their 
Md, but it is somewhat unsafe to attempt any generalization on its evidence 
alone. 

22. Tests to Determine the Bending Stress in the Rail 

* Experiments were made in the track of the Boston and Albany Railroad 
in 1889. The experiments consist in measuring the depression of the rail at 
different places along the length when loaded, and in measuring the extension 
or compression of the metal at the upi)er surface of the base of the rail, near 
the edge of the outside flange. 

For measuring the depression a row of stakes was driven alongside the 
track, three feet away, and the relative level of points on the base of the rail 
and nails in the tops of the stakes was ascertained by means of a sensitive 
spirit level. 

For ascertaining the strains, gauged lengths of 5 inches each were estab- 
lished and defined by center punch marks on the base of the rail at places 
over the ties and midway between them, and the amount of extension or 
compression, as the case might be, was measured on these gauged lengths. 

The rails were 72 pounds per yard, 4^ inches high and 4J inches width of 
base. The results of the experiments are shown in Fig. 159. 

■ Houae Executive DocumeiitH, Ist Session, 51st CoDgress, 1889-90, Vol. 25, Tests on Metala. 



>, Google 



u u 


Bu*. 


P^. 


sl-- 


i. 

■^■r 

H 


Ib.*: 


20.16 


1.23 


20.16 


1.23 


28.96 


1.12 


29.01 


1.12 


29.01 


1.31 


23.62 


1.11 


14.40 


1.82 


10.^ 


2.47 


14.88 


2.35 


9.84 


4.14 


27,60 


l..*8 



Columns 26 and i 



>, Google 



>, Google 



STRESSES IN THE BAIL 



219 



• Experiments were made by the Ordnance Department, U. S. Army, during 
the month of October, 1893, on the track of the Chicago, Bm-lington and Quincy 
RaikY)ad, at Hawthorne, 111. 

The experiments consisted of measuring the depression of the rails under 
the weights of different classes of locomotives, and the fiber stresses developed 
in the base of the rail. 

For the purpose of observing the depression of the rails, bench marks were 
established on a row of stakes driven alongside the rml, 31 inches distant from 



TENOen WITH COAL , 



)ftU6'.r-4^4'-o»j*--'e'-4."'--4*— a"-*' — 4" — e'-o' 



C fi g) 






DEPRESSION 
OF RAIL 



STFUUNS IN J 
BASE OF rail' 
5" LENGTH 



Fia. 159, — Railroad Track Experimenta, Boston and Albany R. R. 

it. A beam carrying a micrometer and an astronomical level bubble were used 
in observing the depression of the rail (see Fig. 160), first measuring the height, 
using points on the outer flange, when the rail was unloaded, and repeating the 
observations when the engine was standing on the track. 

It was found that the roadbed in the vicinity of the locomotives was sensibly 
depressed and that the bench marks were within the influence of that depression. 
It was possible to detect a depression of the roadbed as far as 91 inches from the 
locomotive at the side of the track. 

A correction for the depression of the bench marks was obtained by means 
of a cantilever supported 10 feet from the track, and the total depression of 

• House Executive Documents, 3rd Session, 63rd CkmgreM, 1894-95, Vol. 30, Teste of Metals, etc. 



,y Google 



STEEL RAILS 



Fia. 160. — Railroad Track Experimeats. 
Pholof(raph of Leveling Instrument (or Measuring the Depression at the Traclc. 



Fra. 161. — Railroad Track Ejcperiments. 
Photograph of Micrometer for Determining the Fibre Stress in the Base of the Rail. 



>, Google 



STRESSES IN THE RAIL 221 

points on the rails was also determined with reference to the cantilevers in some 
of the experiments instead of using stakes. 

The fiber stresses were determined in the base of the rail by measuring 
the elongation or compression of the metal on a gauged length of 5 inches, 
established on the top surface of the outer flange, observing the strains when 
the wheels were directly over or when spanning the gauged length (see Fig. 161). 

The observed strains were then computed for the stresses per square- inch, 
assuming a modulus of elasticity of 30,000,000 pounds per square inch and 
that the fibers in the base were strained proportionally to their distance from 
the neutral axis of the rail; the computed stresses referring to the outside fibers 
most remote from the neutral axis. 

It will be observed that the strains and the computed stresses refer to 
a gauged length of 5 inches, and, consequently, the maximum stresses may be 
somewhat greater than those shown, considering the maximum bending moment 
to be directly under the point of application of the load. Some of the results 
are graphically shown in Fig. 162. 

The moment of inertia of the 66-pound rail tested was 19.127 and the 
19.127 
2.24 

Fig. 162A shows the depression of one rail its entire length and the ends 
of contiguous rails, the locomotive occupying one position thereon as shown 
with reference to the rail and ties. 

Fig. 162B shows the curve of depression under another type of locomotive. 
This enginehad no leading truck nor tender, buthadatwo-wheeled trailing truck. 

In the position it occupied during the test, the greatest depression of the 
rail occurred under the forward drivers, the rail presenting a sharp acclivity 
before the engine, and beyond the joint the contiguous rail rose slightly above 
the normal level. 

In the diagram, Fig. 162C, are shown the fiber stresses as measured on 
the base of the rail at station 14|, midway between ties Nos. 14 and 15. 

Advance wave determinations were made on the 66-pound rail on cinder 
ballast (8 inches under the tie) with the same class engine as shown in Fig. 162A, 
the engine weighing 125,000 pounds. With the locomotive slowly approaching, 
an upward movement of the rail began when the leading truck wheel was about 
15 feet away; the wave increased while the locomotive continued to advance, 
reaching a maximum of .0037 inch when the truck wheel was about 85 feet away. 
Then followed a sudden depression, and the height of the rail was reduced to 
the normal level when the truck wheel was about 7§ feet away. 



,y Google 



222 STEEL RAILS 

The position of the locomotive when the upward motion of the wave first 
reached the station could be identified with considerable precision, but, owing 
to an appreciable interval of time being necessary for the level bubble of the 



TEMOER »0OOIb>. 



30ZaOiba. 3Z600IIM, ZSTOOtb*. IBSOOIbB. 




SAME CHtJINE A& SHOWN 



O O O Q 



noo 



5aMeTRACK 3TRISSCS LBa. 



OH A OAU&GD lENQTK .OF 5 INCHES. 




Fia. 162. — Railroad Track Experimenta. C. B. & Q. R. R. 

measuring instrument to stop and reverse the direction of its movement, the 
position of the crest of the wave, as well as the time when the height of the rail 
was returned to its normal level, could not be so well defined. 

The wave length was probably somewhat less than the observations showed. 
The abruptness with which the direction of the wave motion was changed and 



,y Google 



STRESSES IN THE RAIL 




FiQ. 163. — Advance Wave Determinations. (Cuenot.) 

the rail returned to its normal level, after which, of course, it was depressed 
below the normal, was a very striking feature of the observations.* 

• Fig. 163 aljowa the advanre wave observed by M. Cu^ot. It was found that when the first 
wheel of the engine is about 20 feet (foia a tie the upward moveroent commences and reaches a maxi- 
mum at about 10 fec^. 



>, Google 



224 



STEEL RAILS 



The observations of the depression of the roadbed made in these experi- 
ments are of importance. On cinder ballast that part of the roadbed in which 
the stakes were driven (31 inches from the track) was depressed a maximum 
of .049 inch and on gravel ballast the maximum was .036 inch. Wooden stakes 

and iron bolts were driven 
different depths into the 
roadbed with similar re- 
sults; in fact, the few 
observations which were 
made showed the longer 
stakes" to have been quite 
as much depressed as the 
shorter ones, which did 
not penetrate the cinder 
ballast. 

Following out the 
depression of the road bed 
in a lateral direction, 
on cinder ballast, when 
the middle driver of the 
engine was abreast the 
place of observation, 
there was a measurable 
depression at a distance 
of 91 inches from therail. 
The recovery in the 
depression of theroadbed 
was not complete imme- 
diately upon the removal 
of the engine from that 

Fio. 164. — Movement of Rails Laid Alongside of Track. vicinity. The principal 

The right-hand rail lying by the near telegraph pole moved 40 feet. pJ^J.^ Qf ^Jjg recovery at 
The trail it left may be traced from a point near the angle bar io 

the foreground. (Iiaihy>ad Age Gazette, Dec. 17, 1909.} Once tOOk plaCC; the 

remaining portion of the 
depression, however, was very sluggish in returning. The length of time 
required to effect complete resilience was not determined. One observation, 
however, made nine minutes after the load was removed from the vicinity, 
showed the resilience then incomplete. 



,y Google 



STRESSES IN THE RAIL 225 

JRg. 164 shows an exaggerated case of the wave motion and depresaon of 
the track. The road runs on an embankment about five feet above the level 
of a wet meadow. The wave motion of the track and embankment is so great 
that rmls lying by the side of the track move along apparently of their own 
accord at the rate of nearly a foot a day. This movement was undoubtedly 
due to the undulatory movement of the track and entire fill and probably some 
reaction of the fill itself against the track. 

Further tests were made in 1894 and 1896 on the tracks of the Pennsylvania 
Railroad and the Boston and Albany by the Government. * These experiments 
comprise observations on the fiber stresses developed in rails in the track, the 
depression of the rails, and the slope or inclination of the rails caused by the 
weight of the different wheels of the locomotive. The results show some phe- 
nomena displayed by rails in service under the static conditions of loading or 
when a locomotive passes slowly over the track. 

The series were made chiefly on the track of the Pennsylvania Railroad, 
where exceptional opportunities existed for examining roadbed, embracing a 
wide variety of conditions of weight of rails and different kinds of ballast and 
its behavior under heavy types of freight and passenger locomotives. 

The tests were made during the early part of the month of November, 1894, 
on track in the condition it was found in service. 

The experiments on the Boston and Albany Railroad were made with track 
on frozen gravel ballast, in the month of February, 1895. 

The fiber stress tests were made by means of a micrometer mounted on 
the upper side of the flange of the base of the rail, at a place midway adjacent 
ties. The instrument covered a gauged length of 5 inches. The micrometer 
was adjusted in position, and then the several wheels of the locomotive were 
successively brought over the gauged length, or until the same was midway 
adjacent wheels. 

The instrument was read when the locomotive was at each of these posi- 
tions. It was found practicable to make the micrometer observations without 
arresting the locomotive in all cases, taking the readings as the locomotive 
passed slowly over the rail. In this manner the strains developed were measured, 
an elongation of the metal showing tensile stress, and a contraction in the gauged 
length showing compressive stress. 

The measured strains were reduced to stresses per square inch, assuming 
the modulus of elasticity of the steel to be 30,000,000 pounds per square inch, 
and correcting the observed strains in order to obtain the maximum flber stresses, 

* House Documents, Vol. 46, 54th Congress, 1st Session, 1895-06. No. 54, Tests of Met&ls. 



,y Google 



226 STEEL RAII5 

on the further assumption that the strains were proportional to their distance 
from the neutral axis of the rail. 

Fig. 165 shows the micrometer in position on the base of the rail, under 
the driving wheel of a locomotive. 

The depression of the rails was measured by means of a sensitive level 
bubble, mounted on a rod, carrying at one end a screw micrometer, which 



Fia. 165. — Railroad Track Experimeata. 

View showing Micrometer for Measuring Strains in Raiia, in PodtioB on Base of Rail undo' 

Driving Wheel. 

rested on a stake driven in the roadbed 30 inches from the rail; the other end 
of the rod rested upon the base of the rail. The depression of the track was 
thus measured with reference to the top of the stake used as a bench mark. 
In this series it was necessary to arrest the movement of the locomotive at 
each observation. 

The slope tests, or inclination of the rails, were made by means of a sensitive 
level bubble mounted on a frame 12 inches long. At one end of the frame there 



,y Google 



STRESSES IN THE RAIL 



227 



was a fixed supporting rod having a conical point; at the other end there was a 
screw micrometer, the contact end of which was also made with a conical point. 

In the use of this instrument, two center punch marks, 12 inches apart, 
were made on the base of the r^l. The conical points of the instrument entered 
these center punch marks and furnished definite contact points with the rail. 
The instrument was then leveled and the changes in slope, when the rail was 
affected by the locomotive, were measured from this initial adjustment of the 
level bubble. 

Fig. 165 shows the slope instrument resting on the second tie to the right 
of the fiber-stress micrometer. The rails examined ranged in weight from 60 to 
100 pounds per yard, and were supported on oak ties resting on cinder, gravel, 
and stone ballast, in the case of the Pennsylvania Railroad. 

On the Boston and Albany Railroad, yellow pine ties, with shoulder tie 
plates, were used, the roadbed being ballasted with gravel, which was in a 
frozen condition at the time of the tests. 

TABLE UV. — RAILROAD TRACK EXPERIMENTS. GENERAL DIMENSIONS 
OF RAILS 

GOVERNMENT RAIL TESTS 
(HouH Documnta. Vol. it, Mth Caii)[r«a, 1st SsBion. 18M-H. No. M, T«M ol HeCaJa) 



P^tt. 


n»,h, 1 Widlhot 1 Widlhof Thicki.™ 1 Moment q( 


^z^ ^^'x^Srx^^'^ 


Pound,. 


Inches. liifhB<. Inches. Incb. ■ / 


«-^ 1 ii.r„" ! r^-: 


60 
70 
85 
100 
95 




4k ' 21 
? 1 It 


i u.222 
{ 18.055 
iJ 26.374 
1 38.957 
1 1 32.28 


6,693 
8.282 
10-853 
14.812 
13.563 


2.125 
2,32 
2.57 

2,87 
2.63 


2,125 
2.18 
2.43 

2.63 
2.38 



The genera! dimensions of the rails are given in Table LIV. What was 
then considered a heavy type of freight and passenger locomotive was em- 
ployed, the weights of which are recorded in Table LV. 

Referring to the tests on the Pennsylvania Railroad, the tensile fiber stresses 
developed under the weight of the driving wheels ranged from 2810 to 19,540 
pounds per square inch, and the compresdon stresses, when the gauged length 
was between wheels, reached 7880 pounds per square inch. These values 
belonged to the rails in their ordinary condition of service. A tie was removed 
from the track, laid with 100-pound rwl, which increased the distance between 
centers of ties to 52 inches, and here the maximum tensile stress developed 
was 18,970 pounds per square inch, against 9840 pounds per square inch for 
another rail of the same section resling on ties 26 inches apart. 



,y Google 



228 



STEEL RAILS 



A splice bar on a 70-pound rail was strained 22,140 pounds per square 
inch, tension, and 8300 pounds per square inch, compression stress, by the 
driver of passenger engine No. 809. 

TABLE LV. — WEIGHTS OF LOCOMOTIVES 

GOVERNMENT RAIL TEST9 





. 4S, Mth CUMinH, lit StMion. lS»-9e. No. M. T«t. ol H«ul>} 






Totd. 


EDd». 


TendK. 


Weiih. p«r Whwi. 


Pilot. Driven. 


Wheel. Pounds. 


TOIB. 


Pa^aenger No. 809, 
Clwa PK. 

Paaaenger No. 1515, 
Class T. 

Freight No. 557, 
B. & A. R.R. 


POUDdl. 

197 050 

222,500 
188,600 

199,700 


Po«nd>. 

39,750 
50,300 
11,000 

40,700 


PoBDde. 

87,300 

95,200 

800 

75,000 


Pounds. 
70,000 

77,000 

63,800 

84,000 


Pilot 

Driver, first... 
Driver, second 
Tender 

Pilot 

Driver, first. .. 
Driver, second. 
Tender 


9,937 
21,750 
21,900 

8,750 

12,575 
24,250 
23,350 
12,833 

5,500 
13,250 
13.750 
15,650 
14,250 

7,975 

10,175 

18,750 
18,750 


4.968 
10.875 
10 950 

4.375 

6.287 
12.125 
11,675 

6.416 


Driver, first. , . 
Driver, second. 
Driver, third , . 
Driver, fourth.. 
Tender 

Pilot 


6.625 
6.875 
7,825 
7.125 
3,987 


Driver, first. . . 
Driver, second. 
Tender 


9 375 
9.376 




First truck 

Second truck... 


9,250 
11,750 


4.625 
5.875 



Table LVI shows the maximum tensile fiber stress caused by the wheels 
of the pilot, engine, and tender on the different rails and kinds of ballast, also 
the maximum compres^on stresses developed in each experiment. The place 
of observation in these experiments was between ties and about one-quarter of 
the length of the rail from the end. 

From the irregular manner in which the stresses were developed in the 
different weights of rail, it is evident that the peculiar condition of the track 
at individual rails has an important influence on the magnitude of the fiber 



The lightest section of rail examined, 60 pounds per yard, resting on ties 
on gravel ballast, gave exceptionally low fiber stresses, and it will be seen that 
this rail was depressed a correspondingly small amount. 

So much variation is found in the stresses as to practically obscure the 
relative strength of the different weights of rails, and it seems necessary to 
compare the extreme sections to show a well-defined difference in the maximum 



,y Google 



STRESSES IN THE RAIL 229 

On account of the peculiar conditions influencing the behavior of the 
individual rails, the relative values of the different kinds of ballast are less 
conspicuously shown in the fiber-stress experiments than in the series on the 
depression of the rails. 

TABLE LVL — MAXIMUM FIBER STRESSES IN BASE OF RAIL 

GOVERNMENT RAIL TEST8 
(HouH Dooumeota. Vol. U, 5ith ContnH. lit Sa«kia. t8M-M. No. H, Taats ot HMab) 



"¥iid 



Gravel!! 

Stone... 

Gravel.. 



Passenger No. 809. . 
" ightNo. 557... 






Spite. - 
Cinder 
Cinder 

Gravel 

Gravel 

Stone 

Stone, tie removed. . . . 

Frozen gravel, rail No. 1 
Frozen gravel, rail No. 2 



Freight No. 557. 

Passenger No. 80 
Passenger No. 8C 
Freight No. 557. 
Passenger No. 80tf. 
Freight No. 567 . . 
Passeoger No. 809. 
Passenger No. 809. 
Passenger No. 809.. 
Passenger No. 1515. 
Freight No. 557,.., 
Passenger No. 809. . 
Passenger No. 1515 
Freight No. 557. . 
Passenger No. 809. 
Freight No. 557... 
Passenger No. 8O0 
Passenger No. 809, 
Freight No. 557 , . 
Passenger No. 209 
Passenger No. 209. 



BT Stma HT Square 



5,730 
3,580 
10,750 
9,310 



10,640 
3,510 
6,870 



16,060 

17,170 
18,620 
13,7S0 



8,430 
9,920 
11,450 



UompM- 



3,490 

1,400 
4,290 



2,180 
8,300 
3,580 
4,300 
4,300 



4,300 4,300 

5,020 3,580 

5,620 4,220 

8,430 2,110 

4,220 2,810 

6,870 3,050 

6,870 '7,630 



The relative effect of the several wheels of the locomotives are shown 
with greater precision than some other features of the test, inasmuch as in this 
comparison the action of all wheels are referred to the same point on the rail. 
Table LVII shows the tensile stresses developed per ton weight on the dif- 
ferent wheels of each locomotive on the several rails. !From these results it 
appears that the stresses are generally greatest under the outside wheels. 

An examination of the results shows, as an extreme case, that the pilot 
wheels of freight engine No. 557 on a 60-pound rail, with stone ballast, gave 
a fiber stress of 4058 pounds per square inch per ton on the wheel, whereas 
the first driver of the engine, per ton, strained the rail only 1685 pounds per 
square inch. In this instance the total stress per square inch was the same 
under these two wheels, namely, 11,160 pounds, although the weight on the 
drivers was more than twice that on the pilot wheel. 



,y Google 



STEEL BAILS 



TABLE LVII. — TENSILE FIBER STRESSES IN BASES OF RAILS PER TON 
WEIGHT ON THE DIFFERENT WHEELS 



PASSENGER LOCOMOTIVES 





LocomDtive 


Bmllut. 


Ten. 


le Fibv StHH (in P«md» per Tox Weigbc 


on Wh«I< of 


Rail 

"VS."/.- 


Pilot. 


DriYBf. 


T«d»r. 




I 


« 


1 


1 


1 


1 


> 


i 
















































No. 800 


Cinder 


1440 


721 


949 


916 


1147 








85 


No. 1515 


Cinder 


ttll 


Sll 


94.'; 


1043 


vm 


782 


1116 




60 


No. 809 
No. 809 
No. 809 


Gravel. .. 


1244 
1805 
2164 


1805 
866 


1073 
1585 
lOM 


949 
1700 
1112 


471 
1262 
1474 


629 
1735 
1474 


629 


629 








86 


Gravel 


1147 ' 1474 


85 


No. 1515 


Gravel 


14S1 


911 


1359 


1466 




lllfi 


1451 


60 


No. 809 


Stone 


raS7 


2387 


1798 


1784 


1595 


1913 


1595 


2279 






Stone 
















1808 




No. 809 
No. 809 
No. 809 




1441 
1272 
2122 


866 
566 
1415 


989 
905 
1422 


785 
899 
1732 


983 
M2 
1(107 


818 
642 
1607 


8ia 

642 

12a.'; 




100 




1285 


100 




1927 


70 


No. 809 


Bridge... 


1463 




1672 




2327 


1995 


1995 


2494 






















2110 


96D 


No. 209 


Frozen grave , rail No. 1 


im 


900 


7ai 


977 


659 


973 


909 


116B 






Froien grave , rail No, 1 




1050 






S26 




1038 






No. 209 


Frozen grave , rail No. 2 


1.=MKI 


1351 


iiaq 


1221 


Wl 


1485 


«09 


909 






Frozen gravel, rail No. 2 


1351 


10.W 






1319 


1485 










Frozen gravel, rail No. 2 


l.STKI 


1050 


814 


1139 


1155 


11. ^i 


909 


1038 




No. 209 


Frozen gravel, rail No. 2 


1500 


1050 


895 


1025 


1155 


IIW 


1169 


1038 


95L 


No. 209 


Frozen gravel, rail No. 2 


















519 






Frozen gravel, rail No. 2 


lO.'MI 


149 


408 


651 


XU 


-164 


129 


260 




No. 209 


Frozen gravel, rail No. 2 


751 


600 


651 


733 


659 


495 


519 


650 



FREIGHT LOCOMOTIVES 





Locomotive, 


Ballmie. 


Teasile Fiber SCtoh (in Ponnds) per Too Weifbt OB Wheels of 


i*. 


PUot. 


nrn^. 


Teiid«. 




.1 


1 

1041 

999 
1»05 
1146 
2335 
1046 
1041 

511 


■ 


t 


1 


s 


» 


* 




No. 557 
No. 557 

No. 557 
No. 557 
No. 557 
No. 557 
No. 557 
No. 557 




1302 
1247 

2760 
2604 
4058 
2353 
1564 
1276 


540 
518 

1250 
974 

1685 
761 
865 
424 


366 

878 
1323 
824 

1872 
1011 
732 
718 


1408 

1060 
2512 

1408 
2056 
1615 
1408 


898 

860 
1558 

720 
2450 
1264 
1078 

705 


8S8 
617 
1384 

720 
1926 
1084 

880 


720 
860 

1038 
359 

2275 
903 
898 
351 


1259 


60 










85 
GO 


Gravel 


720 

2450 
1623 
12.'i9 
1058 






85 
100 


Stone 









>, Google 



IN THE RAIL 



231 



Throughout this and earlier series of track experiments the same tendency 
has been found, the outside wheels exerting the most severe action on the rails 
in proportion to the weight which they carry. 

The maximum and minimum tensile stresses per ton on the different wheels 
are shown in Table LVIII. 



TABLE LVIII. - 



MAXIMUM AND MINIMUM TENSILE STRESSES PER TON ON THE 
DIFFERENT WHEELS 









Gravel.. 

Stone 

Stone — 
Ciader.. 



No. 809. 
Freight No. 557... 
Passenger No. 809. 
Freight No. 557.. . 
PasBcnger No. 809. 
Paasenge 



Ginder. . 

Cinder . 
Gravel. , 



Pasaenger No. 1515. 
Freight No. 557,..- 
Passenger No. 809. . 
Passenger No. 1515. 
Freight No. 557, , 
Stone i Passenger No, 809, , 

Stone ! Freight No. 557,, 

Stone I Passenger No. 809. , 

Stone, tie removed' Passenger No. 809, . 
Stone Freight No. 537. . . 



Gravel, froien* . . , Passenger No. i 
Gravel, frozen* . . . Passenger No. ' 
Gravel, frozent . . . Passenger No. 



Gravel, frozen t , 
Gravel, (rozenf . 
Gravel, frozent . 
Gr&vel. frozent , 



209. 



Passenger No. i 
Passenger No. i 

Passenger No, 2 
Passenger No. S 



lat pilot . . 

Pilot 

Pilot 

Pilot 

4 th lender , 
4 th tender , 

Pilot 

1st pilot. . , 

Pilot 

4th tender , 

1st pilot, , , 

Ist tender.. 
4th driver.. 
Ist pilot 
1st pilot 
Pilot,., 
lat pilot 



Pilot,., 
1st pilot 
Ist pilot 
Pilot,.. 
1st pilot 
1st pilot 
lat pilot 

2Qd tender 
1st pilot . . 
lat pilot. , . 
Ist pilot. . . 



Pouiids. 
1244 
1247 



2353 

2494 
2789 
1441 



2164 
1481 
2604 
1441 

1564 
1272 
2122 
1276 
1199 



I4S5 
1500 
1500 



lat tender... 
1st driver. ,. 
3rd tender. . 
Ist driver. . . 
2nd driver. . 
Ist tender. , 
3rd tender. . 
1st. 2nd, 3rd 

tender 
1st driver. .. 
2nd driver . 
2nd tender.. 
2nd, 3rd 

2nd pilot. , . . 
1st driver,., 
2nd pilot, .. 
1st tender, . 
3rd tender , , 
2nd. 3rd 

tender 
3rd driver , , 

2nd pilot 

3rd driver. , 
3rd driver . . 
lat driver. . . 
Ist tender. , 
3rd, 4th 

tender 
1st driver, . , 
1st driver,., 
1st driver, , . 
2nd, 3rd 

tender 
2nd tender . 
&d tender. . 



1262 
1038 
1152 



Illustrative of the influence which the condition of the roadbed has on 
the fiber stresses, the 60-pound rail on gravel ballast showed 1247 pounds 
per square inch stress per ton under the pilot wheel of the engine, whereas. 



,y Google 



232 STEEL RAIIS 

with the same weight of rail on stone ballast, the same wheel gave 4058 pounds 
per square inch. 

The fiber stress experiments on the Boston and Albany Railroad were 
made on rails 95 pounds per yard, on frozen gravel ballast, and observations 
were taken at sevwal points along the length of the rails. The observations 
on rail No. 1 were made with the rail in the condition in which it was found 
in the track. There was some looseness between the tie plates and the rail 
and ties, which, in nul No. 2, was diminished as far as possible by redriving 
the spikes and by the use of a number of additional ones. This is the only 
instance in which spikes were redriven before testing. Rail No. 1 was ex^nined 
at two, and No. 2 was examined at seven, places along its length. 

The tensile fiber stresses at the first end of rail No. 2 were higher than 
those developed at the middle and near the second end of the rail. In this 
rail, as the tensile stresses diminished at the second end, the compressive stresses 
increased. At a space 33 inches from the end of the rail, the compressive stress 
in the base reached 7630 pounds per square inch when this space was midway 
the drivers. The same stress was also shown when the space was between the 
tender trucks. 

Concerning the relation between the fiber stress developed and the total 
depression of the rail, the evidence generally favors the deduction that di- 
minished depresfflon will be accompanied by diminished fiber stress. 

The depression of the rails examined on the Pennsylvania Railroad shows, 
with the 60-pound riuls, the least depression on the gravel ballast, the order 
of rigidity being gravel, stone, and cinder ballast. With the 70-pound sections, 
the order of rigidity is gravel, cinder, and stone ballast. Under the 85-pound 
nuk, the stone ballast gave greater rigidity than the gravel. No test for de- 
pression was made with cinder ballast under the 85-pound rails, and only stone 
ballast was used under the 100-pound rails. 

Table LIX states the mean depression of the driving wheels, and also 
the mean depression of all the other wheels of the locomotive in each experi- 
ment. There is in the table a column of differences which states the excess 
of depression of the drivers over that of the other wheels. The column of 
differences is useful in showing the additional depression of the rails under 
the weights of the driving wheels after they have been loaded by the other 
wheels. 

Under the 60- and 70-pound sections, the gravel ballast gave the greatest 
rigidity under the drivers, as well as under the other wheels, and in the column 
of differences the excess of depression was least for this kind of ballast 



,y Google 



STRESSES IN THE RAIL 233 

The total depression with 85-pound rails was less for the stone than for 
the gravel ballast, although the excess of depression under the drivers was 
practically the same in the two cases. 

The depression of the rails on frozen gravel ballast, in which there was 
no visible movement of the ties, would seem to represent about the attainable 
Kmit of ri^dity in track on wooden ties. 

The fact that 60-pound rail on gravel compares favorably with the heavier 
section on the frozen ballast indicates that this light section of rail was in a 
condition approaching rigidity. 

In the slope tests, the approach of the locomotive was felt for a distance 
of 12 to 15 feet in front of the first wheel. The first observed movement was 

TABLE LIX. — DEPRESSION OF RAILS— MEAN DEPRESSION UNDER DRIVING 
WHEELS AND MEAN DEPRESSION UNDER PILOT AND TENDER WHEELS 



B Doctmifliila. Vol. 



GOVERNMENT RAIL TE8TS 
U, Mth Oyagrtm. l9t Saei'on, ISM-m. No. H, Tests at Metali] 



M 


Ballait. 


Locomotive. 


Dcivm. 


?ia.r 


DiffonBoe. 


60 






.229 
.073 
.162 
.230 


luh. 

,154 
042 
122 
157 


laeh. 










60 
























70 


Stone 




.277 1 .207 
.233 1 .184 
.144 .097 
. 168 , 116 




















100 










Gravel, rroieo, rail No. 1 

















an upward one, the inclination of the rail sloping in a direction from the loco- 
motive. This was followed by a reversal in the direction of the inclination, 
which then sloped toward the locomotive. As the several wheels successively 
passed over the place of observation, the inclination of the slope reached a 
maximum and was reversed in direction, these motions being repeated under 
each wheel with some modifications, according to the condition of the track. 
After the locomotive had passed over the place of observation the incKnation 
gradually diminished, and eventually the rail practically resumed its original 
level. A very critical examination led to the conclusion that each passage of 
a locomotive left the rail in a slightly different state than it before occupied, 
and that some sluggishness of recovery in the ballast had an infiuence on these 
minute displacements. 



,y Google 



234 STEEL EAIia 

Figs. 166 and 167 show graphically the results of the tests for depression 
and stress in difFerent kinds of ballast and weights of rails- 






■FT^^TjrrTT'i 



gj^ i g> Vi 



307-SOIb> +«OOJbB rtseooibs- -reNDER 70000lb« 

LOCOMOTIVE Na809 CLASS PK. 



■trackmanS suRFa:^ o 







Sftj*I^E£r-*BiLL AST 



'^/m^ig0^ '^0^ i/ m,. 




Fia. 166. — Railroad Track Eitperimenta, Pennaylvania R. R. DepreaaioD in Ballast. 



>, Google 



STRESSES IN THE RAIL 



■ I ! 



Am 



i'-sF _,Li_ r-r — Ue lo- 



A ( bd ) 6 



TENDER 7O00O 
STRESSES (N BASE OP RML - OAAVEL. BALLAST. 




STRESSES IN 



OF RAIL - STONE BALLAST. 




Fig. 167. — Railroad Track Eiqjerimente, PennsylvE 



>, Google 



236 STEEL RAILS 

All of the experiments just described have been made with the static load 
of the engine. In 1897, Dr. P. H. Dudley commenced a smes of interesting 
tests to determine the effect of the dynamic load of the engine. 

* In Figs. 168 and 169 are shown the results of tests made by Dr. Dudley 
with the stremmatograph. The principle of the stremmatograph is to record 
on a moving strip the molecular compression or elongation of the metal in a 
given length of the base of the rail, induced by the str^ns produced by each 

9T!!CAR ATBCAR aOBCAH SWCAR ICCAR TENDER ENeiNE. 

«v^ nm fw. — cao-QQQ — mn nan — coa qqq — Qno.a.n "" ^X-^ -^ "^- 



Fia. 168. — Stremmah^aph Tests at 19 and 40 m.p.h. 

wheel of the moving train. These records can be measured by filar-micrometers 
under a microscope, and then from the modulus of elasticity of the steel we may 
compute the stresses which produce the g^ven compression or elongation per 
square inch of the extreme fiber in the base of the rail. 

The object of the stremmatograph is to convert rails of any section and 
wei^tj of any system of permanent way construction, into testing machines 
in the track and show how much they are strained, due to the wheel loads and 

• Dr. p. H. Dudley, The Railroad Gazette, May 20 and October 21. 1S!I8, Stresses in Rails 
under Moving Loads. Vol. Ill, Proceedings Am. Society tor Testing Materials, 1903. Stremmato- 
BT^h Testa, by P. H. Dudley. Vol. IV, ibid., 1904, Bending Momenta in Rails, by P. H, Dudley. 



,y Google 



STRESSES IN THE RAIL 



237 



spacing of any type of locomotives and cars moving over the rails at the dif- 
ferent speeds of service. In principle it is the same as the device for measuring 
the strain in bridge members, described in Article 9. 




o oc x ^. , 




COM 
TENS 



Record N?5 
.on. IOO-RA.L 


COMP. „ 1\ A N 


TENSION 
sooo 


'V 




ivy 


sooo Record N?4 


COMP. n M N 


TENSION 






















20000 

































Fio. 169. — Stremmatograph Tests at Slow Speeds. 

Record No. 1, Fig. 168, is taken on the New York Central and Hudson 
River Railroad tracks. The instrument was applied on the outside rail of a 
3-degree curve at the Grand Central Terminal, over which nearly all of the 
heavy trains from the terminal pass outward; the tonnage was from 20,000 to 



,y Google 



?38 STEEL RAILS 

25,000 per day, and there was more looseness in the track than generally found out 
on the main line. The following is the data of the test. 



Date, 




June 28, 1898. 


Weight of rail, 




100 pounds per yard. 


Height of rail. 




6 inches. 


Ballast, 




Stone. 


Ties, 




Oak with tie plates spaced 24 inches, 
center to center. 


Speed, miles per 


hour, 


19. 


Temperature, 




90° F. 


Locomotive and tender, 


202,000 pounds. 


First car. 




96,000 pounds. 


Second car. 




86,200 pounds. 


Third car. 




82,000 pounds. 


Fourth ear. 




94,950 pounds. 



Record No. 1, Fig. 169, shows the extreme fiber stresses in the same track 
and under the same engine in a test of December 29, 1897, temperature 23° F., 
when the engine was moving at a speed of 10 miles per hour. 

When the December test was made, the rail was not only firm on the ties 
but was under some tension, due to the low temperature. In this test the 
small stresses under the front truck wheel show at once that the rail was very 
firmly supported and was not loose on the ties. The rail being under tension 
before loading, the actual stresses of tension in the rail were higher than indi- 
cated, and those of compression lower. 

When the June tests were made, the track had not been tamped or sur- 
faced by the trackmen since the preceding October. Over 16,000,000 tons had 
been carried over the rails since that date, or 12,000,000 tons since the previous 
test. 

Since the December test, air brakes had been applied to the engine trucks, 
which adds more to her weight. While the engine truck carried more weight 
in the June than in the December test, the increased stresses under the front 
truck wheels and the much larger stresses of compression show that the rail 
and ties were much looser in the June test. 

Record No. 2, Fig. 168, was taken at West Albany (N. Y. C. & H. R. R. R.) 
September 30, 1897. The engine was drawing five Wagner palace cars at a 
speed of 40 miles per hour; 80-pound rail, 5 inches high; ties spaced 25 inches, 
center to center. 



,y Google 



STRESSES IN THE RAIL 289 

Records Nos.'2 and 3, Fig. 169, show tests made at 2 miles per hour 
and 10 miles per hour. The total weight of the locomotive was 96 tons; the 
engine 60 tons, with 15,500 poxmds on pony truck and 104,500 pounds on three 
pdrs of drivers. The tender weighed 72,000 pounds. The instrument was 
placed on the outside rail on a 3-degree curve and down grade of 10 feet per 
mile. The rail was 80-pound section, 5i inches high. The ties were yellow 
pine, 7 by 9 inches, spaced 25 inches, center to center. Gravel ballast, the 
track being in good condition. 30,000,000 pounds was taken as the modulus 
of elasticity of the steel. 

The two ties between which the stremmatograph was attached to the rail 
were very firm in the ballast, and to the eye did not seem to depress as much 
as those on either side; therefore the compression sirens were probably 
higher than on ties all practically depresang alike in the ballast. 

Records Nos. 4 and 5, Pig. 169, were taken on 65-pound and 100-pound 
rails, respectively. The 65-pound rails were of steel with an elastic limit of 
60,000 pounds, while on the 100-pound rails it was 65,000 pounds. The loco- 
motive was a switching engine, having 125,000 pounds upon drivers. The 
instrument was placed between ties spaced 30 inches, center to center, having 
tie plates. 

Dr. Dudley states that tests with his stremmatograph show that the bend- 
ing moments in 80-pound rails under wheel loads used in 1905 may be as high 
as 300,000 to 350,000 inch-pounds, indicating a unit fiber stress in the base of 
the r^ of as much as 30,000 or 35,000 pounds on worn 5-mch 80-pound sections. 
With the 65-pound rail, stresses were frequentiy found as high as 40,000 to 45,000 
poimds.* 

23. Calculation of the Bending and Shearing Stress in the Rail 
Examining, first, the moment and shear in the rail between two pairs 
of driving wheels, we see that if the rail were completely rig^d there would 
result a uniform distribution of the wheel pressure to the ties, and if w equal 
the pressure of one wheel divided by the wheel spacing, I, the resulting moment 
and shear in the rail would then be 

Maximum moment, ^ = To "'^*' 

Maximum shear, J = -^ • 

From the discussion on the support of the rail, we can take .35 ton as 
a maximxmi value for w, and 150 inch-tons will be taken, for the present, as 

" Private communication, June 7, 1912. 



>, Google 



240 



STEEL RAII5 



/ 






/ 
/ 

/ 
/ 

/ 




\ \ 


/ / 




\ 
\ 

\ 



SPACING OF DRIVERS INCHES 



the maximum allowable moment, the following calculations being based on 
100-pound A. S. C. E. rail. From the above we may constract the dotted 
line shown in Fig. 170. 

Considering, now, the effect of the 

8 /^v unequal distribution of the wheel pres- 

'' sure to the ties: Fig. 142 shows the 

relation of the reaction of the tie to the 

depression in the ballast. We can, 

therefore, by constructing the elastic 

curve of the rail to give a shear curve 

corresponding to the support of the 

tie, given in Fig. 142, obtain the true 

"so Mastic curve of the rail, from which 

the moment and shear diagrams can 

Fio. 170. — Wheel Loads for Different Spacing be developed. 

of Drivera. t. ■ ■ .1 i -n 1 

By this method we will assume the 
probable curve of the rail and from this curve deduce the moment and shear 
curves. The shear curve will give the reactions under the r^l, and by comparing 
these with the pressure exerted by the ties for the given depression, as shown 
by the curve of the rml, the correctnras of the latter can be checked. After a few 
trials the actual curve taken by the rail for any span and wheel load can be drawn, 
and the maximum bending moment and pressure on the ballast ascertained. 

Mr. C. E. Love of the University of Michigan has made a very interesting 
analysis of the results of the government tests of 1889, 1894, and 1895, described 
in Article 22. As it is proposed to use the method adopted by Mr. Love a re- 
view of one of the cases he has worked out will prove instructive and, on account 
of the careful measurements made in the test selected, the accuracy of the 
analysis can be shown before proceeding with its application to the general 
case under consideration. 

In Mr. Love's discussion the elastic curve of the rail, plotted from the careful 
measurements given in these tests, is taken as the original curve, and the curve of 
slopes, bending moment diagram, and shear diagram derived by the following 

relations: Elastic curve, i> = /(i), or ^. 

z 

Curve of slopes, C=f {%), or -^ . 

Moment diagram, M = E7/" (x), or EI^-i> 
aar 

Shear diagram, F = Elf" (x). or EI^ ■ 



,y Google 



1 space = = 



STRESSES IN THE RAIL 241 

Plate XXIII shows one of the digrams worked out by Mr. Love. This 
plate is based on diagram No. 5 of the 1894 experiments.* The rail used was 
30 feet long, with tie centers at points 1, 2, 3 - - - - 18 (diagram A) and 
loaded as shown, one horizontal space representing 2 inches. 

In dia^;ram B, points on the elastic curve were plotted from the observed 
deflections and a smooth curve drawn through them, one vertical space 
representing a deflection of .002 mch. This curve was then divided into 
segments 6 inches (three spaces) in length along the horizontal axis and the 
slope of each segment measured as if it had been straight. In determining the 
tangent of each slope angle, the measurements were made in sixtieths of an 
inch. The base, of course, was constant, three spaces equaling jj inch of 
actual distance on the paper, so that, calling the altitude dv, the actual slope was 
tan = yi' where dv was measured in sixtieths of an inch. In drawing the slope 
curve, diagram C, however, one space corresponded to an altitude of sV inch. 
For example, the extreme left-hand segment, diagram C, is three spaces below 
the axis. This makes the value of one space in diagram C, 

2 1 J_^ 1_ 

'1000*2*14 14,000' 

The bending moment was constructed by plotting points directly from 
diagram C, by taking the ordinates of the bending moments equal to the corre- 
sponding breaks in the slope curve. The ordinates of the points in diagram D 
are three times the actual slope of the curve in diagram C. The moment of 
inertia (/) of the rail is 19.127, and a value of the modulus of elasticity (E) is 
taken as 30,000,000. 

Hence, since M = El-r^, we have for the value of one space in diagram D, 

1 space = T77^ -14*30,000,000 x 19.127 = 6800 mch-pounds. 
14,U0u o c 

The construction of diagram E is now very simple. The actual slope of a 

segment of the moment curve was measured and the ordinate of the shear curve 

taken as 10 spaces for each unit of the measured slope. Hence, for diagram E, 

1 space = 6800- —'I = 340 pounds. 

The bending moment at any point can now be scaled directly from dia- 
gram D, and the change in shear under any tie in diagram E is the reaction at 
the tie. 

The total load on this rail is 56,500 pounds. After examining the re- 

* House Executive Documents, 3rd Seeaian, 53rd Congress, 1894-05, Vol. SO, Tests of Metala, etc 



,y Google 



STEEL RAIIS 



actions at ties 1, 2, and 3, it is reasonable to suppose that 3000 pounds of the 
load at No. 1 is supported by the rail to the left, thus leaving a net load of 53,500 

pounds. The total up- 
ward reaction is 52,300 
pounds. 

The error at No. 
Sj is 100 pounds, i.e., 
the actual load is 16,- 
000 pounds, while as 
read from diagram E 
it is 16,100 pounds; 
at No. 10§ the er- 
ror is + 300 pounds; 
at No. 13J, +300 
pounds. 

Other experiments 
were performed on this 
rail, in which the fiber 
stresses were meas- 
ured. The maximum 
bending moment here 
is +98,500 inch- 
pounds. In the other 
experiments it was 
117,000 inch-pounds, 
caused by an engine 
10 per cent heavier, 
with the load at the 
middle of a 24-inch 
span. The minimum 
here is - 24,000 inch- 
pounds; in the other 
experiments it was 
- 38,500 inch-pounds. 
Fig. 171 shows the 
construction for 100- 
pound A. S. C. E. rail 
with spans of 60 inches, between centers of driving wheels. 















1 

o 


0" 7 


o" s 


o" 




> 








OM' 












3 
O 




















o 








2,26: 












s 












:;: 


:■■■- 
















■...;i 


■; 








> 








































UJ 










.,.. 














(A 










m 












^ 


















111 
> 






oooo 


3IIUU 
























3 


S 
















WWn 








£ 


•0 
U 




















..... 




z 


i 




















(A 

3 


















g 


? 




2000( 


1 LBS 


s — 










a 














■ ■ ■ 






!)i 















i'-': 





>, Google 



IN THE RAIL 243 

The construction of Fig. 171 is similar to that of Plate XXIII. The scale 
to which the elastic curve is drawn is, 0.002 inch equals one space vertically, 
and 2 inches equals one space horizontally. In deriving the slope curve, 
the actual ordinates of the elastic curve are measured in fiftieths of an inch, 
and one vertical space on the slope curve is taken to represent one-fiftieth of 
an inch of the elastic curve ordinate. The base is taken in all the curves as 
two spaces. 

The value of one vertical space of the slope curve is, then, 
2 1 1 1 
1000'50'(A=H)*2' 

2 1 50 1 
1000'50'l0'2' 

2 1 1^ 1 
1000 ■ 10 '2 10,000" 

For example, the ordinate of the elastic curve at 2 inches to the right of the 

5 5 2 5 10 

wheel is about rrr inch ; the actual slope is, therefore, tan = ?^ -h r;; or =;;-=- — = .5, 
oO 50 10 50 50 

as shown by the slope curve between and 2 inches. The ordinate of the 

12 5 
elastic curve at 4 inches is about -=^ or the rise from 2 inches to 4 inches is 
50 

inches. 

The ordinates on the moment curve are taken the same number of spaces 
as the corresponding breaks in the ciu-ve of slopes and consequently, on account 
of the base being 2 spaces, represent twice the actual slope of the slope curve. 
The value of one vertical space in the moment curve is, then, 

10,000 2 2 

ITjL- -l-i. 30,000,000 X 43.8 - 32,860 inch-pounds. 

The ordinates of the shear curve are five times those of the moment curve, 
and the value of one vertical space in the shear curve is 
1 1 1 
*5*2'2' 

The most simple way to construct the diagram is to approximate the 
elastic curve of the rail and then follow out the operations, as shown by Table 
LX, correcting the elastic curve from column 11 of the table and readjusting 
the calculations, if necessary, to the corrected elastic curve values. 



,y Google 



STEEL EAIIfi 



TABLE 


LX.— CALCULATIONS OF RAIL DIAGRAMS FOR 60-INCH 


WHEEL SPACING 




Dtprawoa 
track- 


Pounds. 


sbMt. 


Moxi 


«.. 


Stop., 


Elmti«eiirv» 
in A inch. 


Inch. 


Pou«(.. 

4 


s 


1 


OrdiMle, 

7 




0^.. 


IXlu. 


Ordi- 




0.30 




20,080 


12.2 


2.44 


6,1 
4.9 














2800 


4,9 








17,280 


10.5 


2.10 


4.9 


4.9 








2760 


2.8 


2,8 








14,520 




1.76 


7.7 


7.7 








2720 


1,0 


1.0 








11,800 


7,2 


1.44 


8 7 


8.7 








2680 


- .4 


- ,4 






0.29 


0,120 




i.i2 


S.3 


S.3 






2640 


-1.5 


-1,5 


32.1 






8,480 




.78 


6.8 


6,8 








2610 


-2.3 


-2.3 








3.870 


2.4 


.48 


4,5 


4.5 








2590 


-2,0 


-2.9 






0.2Si 


1.280 


6,8 


.16 


1,6 



1,6 







1280 


-3,1 


-3,1 


45 















Col. 8 la round (rom col, 6 by dividiOK the fieutm given in col. S by 5. 

In Fig. 172 are given diagrams for spans of 70 inches, 80 inches, and 90 
inches between centers of drivers. 

The maximum bending moment can be read directly from the moment 
curve and the wheel load is twice the load supported on half the span or twice 
the total reaction shown by the shear curve of the diagrams. In comparing the 
shear curve with the curve of pressures of the ties it would seem desirable to 
assume the r^ to be supported by a distributed load in place of a series of loads 
concentrated at the ties, the effect will be practically the same and the calcula- 
tions much simphfied. 

The results of static and dynamic tests of the stress in the rail both indi- 
cate that the negative bending moment between the drivers is very much smaller 
than the positive bending moment produced in the rail under the drivers. The 
dynamic tests appear to show a ratio of about 1 to 4 or a compressive stress in 
the base of the rail only one-fourth as great as the tendle stress, under normal 
conditions. In poorly tamped track the compression stress seems to increase. 

If the ral were uniformly supported between the wheels the compresaon 
stress would be one-half the tension stress under the wheels of a set of drivers. 

Such a condition of the pressure exerted by the tie would be represented by 
a horizontal line in Fig. 142. The diagrams given in Figs. 171 and 172 show that, 
as the spacing of the drivers increases, the negative bending moment or the 
compression stress in the base of the r^ decreases in relation to the tension 



,y Google 



STRESSES IN THE RAIL 





















































































































































































































y 


' 








































/ 


' 










































/ 














ntif 




















y 








/ 










" 
























/ 










/ 










i 






run 






^ 


- 








/ 










/ 












o 










/ 












/ 






















\ 








y 












y 
















































s 










r 


N 








«i 










^^ 










/ 




\ 














\ 






= 








/ 




\ 








1 






\ 












\ 






ol 








1 






^ 






1 






\ 














\ 




\ 




























































































ioot 


>I>L* 


s, 












\ 












\ 












% 








\ 














\ 












v 










5 












^~~ 


^ 























^^ 





















































^ 
















































UM 


LBS. 


\ 












\ 










t» 


\ 












g 










\ 












\ 












V. 










a. 












N 


K 












^ 












^ 






i 





















































































































































































Fia. 172. — Rail Diagram for Wheel Spacing of 70, 80, and 90 incbea, one-balf size of original diagram 



>, Google 



246 8T££L RAIIS 

stress; for drivers spaced 60 inches apart it is practically one-half the tensile 
stress, but when the spacing is increased to 90 inches it is not much more than 
one-third. 

An examination of Fig. 142 makes this clear and shows that with the 
greater deflection obtained in the 90-inch span the ties in the center of the 
span support relatively much less of the load. With lighter rail the deflection 
would be still further increased and a greater ratio of the tension to the 
compression stress obtained. 

The full line, shown in Fig. 170, shows the true allowable wheel loads given 
by the diagrams of Figs. 171 and 172. Up to spans of 80 inches the wheel load 
is limited by the safe bearing power of the tie and is obviously less than that 
obtained from the assumption that there is a uniform distribution of the wheel 
pressure to the ties upon which the dotted line of Fig. 170 is based. After the 
spacing of the drivers exceeds 80 inches the wheel load is limited by the bending 
moment in the rail; here the bending moment is greater for a uniformly dis- 
tributed load and, consequently, the dotted line in the figure falls below the 
full line. 

Turning our attention to the allowable wheel load as determined by the 
conditions at the front and rear drivers. Figs. 162 and 163 show that there is 
a wave motion of the rml ahead of the engine and the rail rises slightly above 
the trackman's surface. This lack of pressure on the rail at the outside wheels 
causes these wheels to exert a more severe action on the rail. This is clearly 
shown by Table LVIII and in Fig. 162, where the outside drivers, although 
carrying less weight, gave practically the same stress as the middle driver. 
Records Nos. 2 and 3 of Fig. 169 illustrate the same tendency. 

The rail diagrams, just worked out, show that in increaang the wheel 
spacing from 80 to 90 inches the permissible wheel load fell from 24 to 22 tons. 
From the diagram for the 90-inch wheel spacing it is seen that the ties in the 
middle of the span afford little support, and while it is somewhat problematic 
what load will be carried by the rail ahead of the front driver, we will not be 
very far wrong if we assume it to carry 8 tons with no leading truck, 9 tons 
with a two-wheel leading truck, and 10 tons with a four-wheel leading truck. 

On accoimt of the load of the tender wheels and the effect of the draw- 
bar pull, we may reasonably take 10 tons where a trailing truck is used and 
9 tons where there is no trailer, as the load carried by the rail back of the rear 
drivers. 

Table LXI may now be prepared showing the allowable dynamic wheel 
load under different conditions of wheel spacing. 



,y Google 



STRESSES IN THE RAIL 



ProDt wheel, 

No leading; truck 

Two-wheel leading truck. . 
iFour-wheel leading truck. . 

Back wheel, 

No trailing truck 

Trailing truck 



36,000 
38,000 
40,000 



38,000 
40,000 
42,000 



40,000 

42,000 
44,000 



Referring to Figs. 31 and 32, of typical load diagrams of en^nes, it will be 
seen that with the exception of the articulated engine there is had a very satis- 
factory agreement between Table LXI and the diagrams. 

We have now to consider the stresses in the rail caused by the bending 
moment and shear derived in Pigs. 171 and 172. The maximum bending 
moment in these figures is 300,000 inch-pounds, and the maximum shear is 
24,000 pounds. 

It is beyond the scope of the present work to enter into the discussion 
of mathematical investigations of continuous web strains, and in order to form 
some conception of the nature of stresses in the continuous rail we shall view 
the matter in the simplest manner possible.* 

In the rail under the wheel it is evident that, by virtue of the bending 
stress, that part of the rail above the neutral axis is subject to compression, 
and that below to tension, both of which stresses attain maximum values at 
the outermost fibers of the rail, and decrease to zero at the neutral axis. This 
intensity of the stress at any point is at once obtained from the well-known 
equation of flexure: 



M 



y = f. 



(a) 



where M is the bending moment, / the moment of inertia of the section of the 
rail, y the distance of the point from the neutral axis, and / the intensity 
of the stress at that point. 

Table LXII gives the extreme fiber stress in the base due to bending in 
different sections of 100-pound r^, caused by a bending moment of 300,000 
inch-pounds. The hi^ moment of inertia in the Series "A" of the American 
R^way Association would give a slightly different elastic curve for this rail 
than is shown for the A. S. C. E. section in Figs. 171 and 172, with the result 

• See Plate Girder ConstructioD, leami Hiroi, New York. 



>, Google 



STEEL RAILS 



that the bending moment would be increased and the unit of load supporting 
the rail decreased. 

- TABLE LXn. — EXTREME FIBER STRESS DUE TO A 
BENDING MOMENT OF 300,000 INCH-POUNDS 



Section. 


WaiRbt. 


Ertrme Fiber 


A. S. C, E 

Am. Ry. Assn., Series "A" 

Am. Ry. Assn.. SerieB "B" 


100 
100 
100 


18,800' 
16,900 
19,100 



B£EES: 



- Diatribulion of HmizoDtal Stress ir 



The bending moment M decreases as we proceed toward the center of the 
space between the wheels, and with it evidently the intensity / of the hori- 
zontal stress also; so that / varies not only in vertical directions on both sides 
of the neutral axis, but also in the direction of the length of the rail. 

^ Let XX and x'x', in Fig. 173, be 

/~\ ^v y two sections of a rail, very close to 

each other, and NN the neutral axis. 
The variation of the value of / 
in both sections may be represented 
by triangles with apices in the 
neutral axis, and the variation in 
the longitudinal direction between 
these two sections by the difference of the areas of two triangles, as shown 
shaded in the figure. This increase of horizontal stress from one section to 
another produces at each longitudinal layer a force tending to slide it past the 
layer next above it, and is transmitted undiminished toward the neutral axis, 
where this shearing force, which has been increasing at every layer, attains its 
maximum intenaty. 

This stress is called the longitudinal shear, and can be at once obtiuned 
from equation (a). Thus, let/' be the corresponding value of /in section x'x'; 
and let M and M' be the bending moment in the two sections xx and x'x' re- 
spectively, and a an infinitely small cross area, distant y from the neutral axis. 
The total horizontal stresses acting in that part of the section lying between 
the extreme fiber distant h from the neutral axis and the layer y'y' distant y' 
from the axis in xx and x'x' are respectively: 



Xfa. 



tfa. 



>, Google 



STRESSES IN THE RAIL 249 

The longitudinal shear in the layer y'y' between the two sections is, there- 
fore, equal to 

Substituting in the expression the values of / and /' given by equation 
(a), we obtain, 

%lo.- 2,fa= = ^ya. 

Since the area on which this horizontal shear is acting is equal to 6 a:e, 
when b is the breadth of the cross section at the layer y'y' and Ax the distance 
between x and x', we obtwn for the intensity of the shear, 

Thus at every point in the rail there are two shearing actions taking place 
at the same time, one the longitudinal shear and the other the vertical shear. 
Imagine abed, Fig. 174, to be an infinitely small ^ t^.^x.b. ^ 



•/ 



portion of the side of a rail at a point distant y' 
from the neutral axis. Suppose the side of this •'■' 
area element to be Ax and Ay, and the breadth of the 
beam at the point to be 6. There are then found two Fio-m.—ShearingStresaof Point 
shearing stresses on this element, one vertical and Distant y' tram Neutral Axis, 
the other horizontal. These two shears form two pairs of couples acting around 
the body, as shown by the arrows. Let t^ represent the intensity of the horizontal 
shear at this point and t^ that of the vertical. The amount of the horizontal 
shear is equal to (, Ax 6; that of the vertical shear is Hkewise equal to i. Ay b. 
In order that the body be in equilibrium, the moment of these couples 
must be equal, i.e., t,Axb Ay = t^Ayb Az. Consequently, (^ = (, (which is 
always the case), showing that at every point in the rail the intensities of the 
vertical and horizontal shears are equal, and we will hereafter designate them 
with the common letter (. The value of t^ has already been deduced in equation 
(b), namely: 

. M' — Af V* , , 

M' — M 
But the value of = 5,, where S, is the total vertical shear at the sec- 
Ax 

- M 

in equation (c) , there results: 



>, Google 



250 



STEEL RAIIS 



or the intensity of the shearing stress at any point in the rail is equal to the 
total shearing force on the entire cross section multiplied by the statical moment 
of the area of the section outside the longitudinal plane of shear in question 
about its axis in the neutral plane, divided by the product of the moment of 
inertia of the entire section into the breadth of the section at that point. 

Fig. 175 shows the intensity of the shearing stress in a 100-pound r^l, 
the total vertical shear at the section being 24,000 pounds. 




SHEAR 
Fia. 175. — Shearing Streaa in 100-pound A. S. C. E. Rajl. 

There still remains to be considered the horizontal force /, whose value 
is given in equation (a), tending either to compress together or pull asunder 
the two faces ac and bd (Fig. 174), according as it is on the upper or lower side 
of the neutral axis. 

At the neutral axis where / = 0, (, and t* are then the only stresses, and 
we know from mechanics that the resultant action of two equal shears at right 
angles to each other, exactly as t^ and t^ are, is equivalent to that of two equal 
and opposite stresses at right angles to each other, called the principal stresses 
and making an angle of 45° with the shearing stresses. But at a distance each 



,y Google 



STRESSES IN THE RAIL 251 

side of the neutral axis the third stress, /, now comes in, which evidently ^ves 
a new direction to tiie line of resultant stress, turning the axis of principal stresses 
toward itself more and more as its intensity increases. 

Fig. 176 represents the appearance which the lines of principal stresses 
thus obtained present in a beam loaded in the middle and supported at each 
end. The lines of maximum tension are shown dotted and cut the lines of 
compression always at right angles. Both lines cross the neutral axis at an 
inclination of 45° and run almost parallel to it in the middle of the beam in the 
neighborhood of extreme fibers. 




Fia. 176. — Lines of Principal Stress in Beam. 

Now comes the question how the web should be proportioned to resist 
such stresses: The greatest intensity of the vertical shearing stress on the verti- 
cal section of the rail, shown in Pig. 175, is about"10,000 pounds per square inch. 
In modem bridge practice a shearing stress is allowed in web plates of 10,000 
pounds per square inch, which gives a satisfactory thickness of the web for the 
rail shown in the figure. But as has already been explained, the action of the 
shearing stresses at the neutral axis is equivalent to compression and tension 
at right angles to each other and of equal intensity, making an angle of 45° with 
the axis, and the web is still in danger of failing by flexure under this com- 



Consequently, the web with its thickness as already proportioned for 

shearing must now be examined for its strength as a column. We will probably 

be not far from correct if the length of the column is taken as A sec 45°, h being 

the vertical distance between the top of the flange and bottom of the head of 

the rail. Then, for the 100-pound A. S. C. E. rail, h sec 45° equals 4.4 inches, 

and the load is 

p 

~ = p = 10,000 pounds per square inch. 

This amount is correct for the bending stress caused -by the load which is 
central over the rail head. The wheel load is, however, rarely applied exactly 
in line with the vertical axis of the rail, and the additional couple due to the 
eccentricity causes a torsion in the rail. To make a correct analysis would 



,y Google 



252 STEEL RAILS 

be very cbmplicated and decidedly uncertain on account of the lack of experi- 
mental evidence. 

No column formula can be made to apply exactly to the web of the rail. 
If we apply the formula for a column with an eccentric loading of .6 of an inch, 
the resulting stress amounts to over 50,000 pounds per square inch. It is 
doubtful, however, whether the stress introduced by this torsion can be com- 
bined with those due to bending in this manner. 

It will be observed that, even were this large stress correct, there is a 
tensile stress acting at right angles to the compression stress and tending to 
hold the strip in its true plane. Just what the restraining influence of this 
tensile stress is cannot be determined theoretically, but the following experi- 
ments show it to be of importance. 

During I9I0 teats were made at the Maryland Steel Company's plant at 
Sparrows Point, Md., for the purpose of determining what effect the eccentric 
loading of the wheel had on the head and web of the rail.* The tests were made 
with a 200,000-pound test machine by canting a piece of rail 18 inches long 
and applying the load at the edge by means of a block with a radius of 16^ 
inches, to represent a car wheel, where it came in contact with the rail. Other 
tests were made with a reciprocating machine representing a loaded wheel 
rolling back and forth on the edge of the canted rail. 

For the tests a rail was taken from stock and six pieces each 18 inches 
long were cut from it for test in the stationary test machine and six similar 
pieces were used for test in the reciprocating machine. In order to have the 
material as imiform as possible throughout the section and in the different 
pieces, a " C " rail was selected, that is, the third rail from the top of the rail 
bar. The rail was a 90-pound A. R. A. type B section and the pieces were planed 
down to thicknesses of head at the side of i inch, § inch, f inch, | inch, J inch, 
and 1 inch, two pieces of each thickness, one for each kind of test. In each 
case the brand side of the head of the rail, which was the bottom side as rolled, 
was planed vertical to a width of 1^^ inches from the center line. Fig. 177 
shows the dimensions of the section used and also gives diagrams of the pieces 
tested. The essential dimensions of the head of the pieces tested, as indicated 
by letters A, B, and C, on Fig. 177, measured as shown in Table LXIII. 

Two samples were taken with a J-inch drill for analysis from a section 
near the middle of the length of the rail, one close to the upper comer and the 
other at the junction of the head and the web. The results of the analyses are 
shown in Table LXIV. 

• strength of Rwl Head, M. W. Wiokhoret, Proceedinga Am. Ry. Eng. & M. of W. Asbd., 
1911, Vol. 12, Part 2, p. 618. 



,y Google 



STRESSES IN THE RAIL 

These results show the material to be very uniform. 

TABLE LXIII. — DIMENSIONS OF HEADS AS TESTED FOR 
STRENGTH OF RAIL HEAD 



TMNumtMn. 


ThkkDtaofHwd. 


widih. 


E^. 


^T^ 


SdatoCatH, 




1.02 
.91 
76 
,62 
.51 
.38 


1.26 
1.15 
1.02 

.86 
.75 
.64 


1.15 
1.18 
1.17 
1.16 
1.16 
1.17 




5 and 6. 




9 and 10... . 







TABLE LXIV. - ANALYSES 
FOR STRENGTH OF 


OF RAILS TESTED 
RAIL HEAD 




Comer of 


HairMd Wsb. 




.538 
.070 
.050 
,81 
.103 
10 
None 
None 


.523 
.070 
.055 

.81 
.103 

.18 

None 

None 






Manganese 




NXrv.'v:;::::::::::::::::': 

Chromium 



Tensile tests were also made of pieces cut from near the middle of the rail, 
two pieces J-inch diameter and 2-inch gauge length, for longitudinal test from 
the center of the head, and two pieces i-inch diameter and 1-inch gauge length 
for transverse test across the center of the head. The yield point in the 2-inch 
pieces was determined by means of a Capp's multiplying dividers. The results 
of the tests are shown in Table LXV. 

TABLE LXV. — TENSILE TESTS OF RAILS TESTED FOR 
STRENGTH OF RAIL HEAD 





Ywld Pouit 
S<i«.™l«h). 


rPouiidB per 
SqMTBlnqh). 


Eloiioti«. 


ReduciioB at 


Longitudinal a 

2-incli gauge length b 

Average 

Transverse a 

1-inch gauge tengtli 6 

Average 


51,000 
52,700 
51,850 


111,600 
111,500 
111,550 

110,200 
111,200 
110,700 


16 

16.5 

16,3 

6 
7 
6,5 


29 
29 
29 

7 
9 

8 









These results show material of good ductihty longitudinally and the stretch 
crosswise of the head shows up well for a transverse test. 



,y Google 



STEEL KAILS 



-i — gy— ^ 




- Diagram of Pieces tested for Sag of Rail Head and Bending of Web. 
(Am. Ry, Eng. Assn.) 




Fia. 178. — MeOiod of Sutionary Teats for Sag of Rail Head and Bending of W(i>. 
(Am, Ry. Eng. Asen.) 

The arrangement used for making the stationary tests is shown in Fig. 
178, and is intended to represent a 33-inch car wheel resting on the edge of 
the top of the rail. The head is thus tested as a cantilever, the load tending 
to sag the head locally and to also bend the web. 



,y Google 



STRESSES IN THE RAIL 255 

The load was applied in increments of 10,000 pounds up to 60,000 pounds 
and then in increments of 20,000 pounds up to 200,000 pounds, the capacity 
of the test machine. The sag of the head was determined by measuring the 
distance by means of dividers, between prick-punch marks placed on the side 
of the head near the bottom and on the base, as indicated in Fig. 178, the 
load being on while taking the reading. The marks on the base were placed 
about one inch from the web, by gouging some of the metal so as to have a 
vertical surface on which to prick-punch the mark. The amount that the 
opposite side of the head elevated, or the " lift," was determined in a similar 
manner. The results of these tests are shown in Table LXVI. 



TABLE LXVI. - 



■ STATIONARY TESTS IN TEST MACHINE OF STRENGTH 
OF RAIL HEAD 

Sac ud Lift is Inehli 



La«l, 


|-iHtlH«d. 


t-iuh H«d. 


l-iocb 


H»d. 


l-lBChHdul. 1 


iBCh 


Haul. 


i-i«b 


Hwd. 


at. 


Lirt. 


St. 


1^. 


Sw. 


Li,t. 


Sai. 


LUt. a 


X. 


Lin. 


Sw. 


UU. 




,00 
.02 
,05 
,06 
,08 
.09 


,00 
.00 
.00 
.01 
.01 
.02 


.00 
.01 
.02 
.03 
.04 
.05 
.06 
.07 
.08 
.00 
.10 
.11 
.13 


.00 
-00 
.01 
.01 
.01 
.02 
.02 
.02 
.02 
.02 
,02 
.02 
.02 


.00 
.01 
.02 
.04 
.06 
.07 
.10 
.11 
.12 
.13 
.14 
.15 
.16 


.00 
.00 
,00 
.00 
.01 
.01 
.02 
.03 
.03 
.03 
.04 
04 
,04 


,00 
.00 
.01 
.01 
.02 
,03 
.04 
.05 
.07 
.08 
.08 
-09 
.10 








.00 
.00 
.01 
.02 
.02 
.03 
.04 
.04 
.05 
.05 
.06 
.07 
.08 






.01 
.01 
.02 
.02 
,02 
.03 
.03 
.03 
,03 
,03 
.03 
-03 


01 
01 
02 
03 
04 
05 
06 
07 
08 
08 
09 
10 


.02 
.02 
.02 
,03 
.03 
.04 
.04 
.04 
,04 
,04 
.04 
.03 




















80,000. . . . 


.02 








02 












.15 
.17 
,20 


.03 
.03 
.05 






03 






200,000 


.02 











These results 
are plotted in Fig. 
179, in which the 
load is plotted 
against the sag of 
headfor each thick- 
ness of bead tested. 
The f-inch head, 
according to these 
curves, gave a 
greater sag than 

the 4-inch head. P>«- l^— sag of lUa Head in stationary Tests. (Am. Ry. Eng. Assn.) 

Although this is according to the measurements obtained, it would seem to 
be in error, due, perhaps, partly to errors of measurement, but probably also 




,y Google 



256 



STEEL RAILS 



due to some condition which cannot be accounted for, as, for instance, 
application of the load. 

The curves show that a load of 10,000 pounds does not sag the head with 
the load applied to the edge of the top side, with any thickness down to f inch, 
and probably neither does a load of 20,000 pounds, although, as the load was 
on when the measurement was taken, we cannot say how much of the sag was 
elastic and how much permanent. A load of 30,000 pounds seems to cause a 
permanent sag with the f-inch head, but not much, if any, with the heads of 
greater thickness. 

It is interesting to note in this connection that the web seemed to stand 
the load of 200,000 pounds successfully. 

Tests were also made with the reciprocating machine, shown in Fig. 151, 
in which a piece of rail is moved back and forth under a wheel to which a 
load can be applied by means of a system of levers. The rail is fastened to 
a steel bloom which runs on rollers running on another steel bloom that forms 
the bed of the machine. The rail bed is connected by means of a connecting 
rod to the bed plate of a planer, which furnished the power to run the rail machine. 
The weights attached to the weight hanger are multiplied 600 times as applied 
to the axle of the wheel, and in these tests weights of 50, 100, and 150 pounds 
were used, so that the wheel loads were 30,000, 60,000, and 90,000 pounds re- 
spectively. A piece of rail 18 inches long was tilted on an inclined plane of 
1 in 10, as in the other tests, and the wheel, loaded with 30,000 pounds, was 
run back and forth over a length of about 10 inches for 100 double strokes, 
which made 200 movements of the loaded wheel over the rail. The sag of the 
head and the width of the bearing taken by the wheel were then measured with 
no load; the load was then increased to 60,000 pounds and the wheel again 
run on the rail as before. The- measurements were ag^n taken and a final test 
made with a load of 90,000 pounds. The results of these tests proved to be 
interesting and fairly definite, and are shown in Table LXVII. 

TABLE LXVIL — RESULTS OF ROLLING TESTS OP STRENGTH OF RAIL HEAD 



ThickHwolHMd, 


'•Viz-'- 


Widtl^oI^B«riBg. 


JO.0O0 


60,000 
Pounda. 


90.000 
Poimds. 


pS 


^^. 


I^i^. 




.05 
,01 

.00 


.13 
.08 
,05 
.03 
.01 
00 


.17 
,13 
.13 

.10 

.08 
05 


,56 
.34 
31 


1.20 
1 00 


1.56 
1.42 
1.40 
1,28 
1.16 
1.00 








,32 .70 

.32 1 ,5S 




.00 

.00 















>, Google 



< 

Ul 

*.i<5 

i 


s. 




ROLLtNe TESTS 


s 




^ 


90,o>o 


LBS. 






s 


V 


60,000 


LBi* 


V 




:k 


^^3etooo 


tea^ 


-^ 






THICKNESS SF HEAD 



-Sag of Rail Head in Railing Tests. 
(Am. Ry. Eng. Assn.) 



STRESSES IN THE RAIL 267 

The results showing the relation of thickness of head and sag of head und^ 
loads of 30,000, 60,000, and 90,000 pounds are plotted in Fig. 180. It will be 
noted that 30,000 pounds 
produces no sag when the 
head is f inch or overin thick- 
ness. With 60,000 pounds 
the head must be 1 inch 
or over in thickness. All 
the samples tested sagged 
under 90,000 pounds, but 
by extending the curve 
it seems probable that a 
head If inches thick would 
hold up a rolling load of 
90,000 pounds when con- 
centrated at the edge. 

After the rails were tested they were cut in two and their sections are 
shown in comparison with the original rail section in Pig. 181. 

The rolling load of 90,000 pounds applied at the edge of the head pro- 
duced very little or no bending of the web with the section used, which was a 
90-pound A. R. A. type B, with a thickness of about ^ inch at the middle. 

While there seems little liability of the web failing as a column, the height 
of the ral in reference to the stability of the outer rail on curves must be 
considered. 

* Mr. E. E. Stetson found that in many cases the resultant of the hori- 
zontal force and wheel pressure on curves falls entirely outside the base of the 
rail in the 100-pound section. 

In 1907 the Pennsylvania Ralroad Company prepared a piece of track 
on the West Jersey and Seashore R^road, on a 1-degree curve, near Franklin- 
ville, N. J., with special cast-steel ties and measuring apparatus, for the pur- 
pose of comparing the effects of lateral horizontal forces on the outer rail of the 
curve generated by different classes of electric locomotives and standard steam 
locomotives. 

The force exerted by the locomotive was communicated to steel plates by 
means of hardened steel spheres of small diameter, and the effect of the force 



• A Study of Rail Pressures and Stresses in Track Produced by DiEFerent Types of Steam Loco- 
motives when Rounding Various Degree Curves at Different Speeds. E. E. Stetaon, Proceedings Am. 
Ry. Eng. 4 M. of W. Assn., 1B09, Vol. 10, Part 2, p. 1432. 



,y Google 



258 



STEEL RAILS 



was to cause the spheres to make a more or less deep impression in the steel 
plates. 

By means of laboratory tests, it was determined what forces in pounds 
were required to produce various known depths of the impression of the steel 
balls in the plates, and after this calibration had been made it was possible to 
transform the depths of the impressions of the balls in the plates into pounds 




ROLLING- TEST- LOAD SdowLsa 

Fia. ISl. — Kails after RoUiog Teat with Load of dO.OOO Pounds. (Am. Ry. Eng. Asm.) 

of pressure. Table LXVIII-A gives the results of some of these tests for steam 
locomotives, in order that they may be compared with the computations made 
by Mr. Stetson, which are shown in Table LXVIII-B for the same degree of 
curve and for one locomotive of the same class as used in the tests on the 
Atlantic City Une. Mr. Stetson's calculations are for speeds of 60 miles per 
hour, which are from twenty to thirty miles per hour less than the actual 
tests, but the weights of the locomotives are heavier. The lower speeds should 
give smaller pressures, while, on the other hand, the heavier locomotives should 
give higher pressures. 



,y Google 



STRESSES IN THE RAIL 

TABLE LXVIII. — HORIZONTAL PRESSURES EXERTED BY STEAM LOCO- 
MOTIVES AGAINST RAIL ON CURVES 

(Am. Ry. £n(. Ahl) 



TABIfA. 


- RESULTS OF TEST3 MADE IN 1907, AT FRANKLINVILLE. N. J., ON WEST JERSEY 




ft SEASHORE RAILROAD ON 


A l-DEGREE 


2URVE 




S^«. 


Siwed in Mil» 
per Hour. 


SpMdCorrs- 
5uf»nlev>tj^. 


Typ«o(Looo. 

moH™, 


Conditicn of 
RaU. 


l>m«u«. 


S^ 












iBCh 




9 


88.4 


70 


B 


Dry 


0.204 


10,600 




87.7 


70 


B 




0.246 


13,000 


17 


92.3 


70 


B 


Dry 


0.222 


11,500 




90.5 


70 


B 


Dry 


0,216 


11,200 


19 


85.03 


70 


B 


Dry 


0.199 


10,300 






70 




Dry 


0.193 


10,100 




75.5 


70 


B 




0,162 


8,500 


22 


80.7 


70 


B 


Wet 


0,217 


11,200 




81.29 






Wet 


0.217 


11,200 


111 


85-30 


70 


B 


Dry 


0,179 


9,500 


118 


80-3 


70 


B 




0,181 


9,500 




83.9 


70 


B 


Wet 


0.188 


10,000 


120 


83-9 


70 


B 


Wet 


0.199 


10,300 


11 


81.3 


70 


D 


Dry 


0.165 


8,700 


12 


83.5 


70 


D 


Dry 


0.134 


7,000 



Class B ia BD Atlantic type locomotive, total weight— 176,600 pounde, height center of gre,vity 
above base— 73 inchee. 

Class D is ao Americao type locomotive, total weight= 138,000 pounds, height center of gravity 
above base=65 inches. 

TABLE B. - RESULTS OF COMPUTATIONS MADE BY E. E. STETSON FOR A l-DBQREE CCRVE 



Sp^gjjile. 


S^ 


TypsolLocc 
motive. 


^SS.'- 


RenHlu- 


60 


60 


Class B 


11,500 


Qass B is an Atlantic type, total weight= 


70 


60 


Class B 


12,950 


183,150 pounds, height center of gravity 
above rail taken as TQ^S. 


60 


60 


Class A 


11,120 


Class A is Pacific type, total weight 270,100 


70 


60 


Class A 


12,830 


pounds, height center of gravity above rail 


60 


60 


Class C 


13,180 


Class C 19 consolidation type, total weight 


70 


60 


Class C 


14,700 


238,200 pounds, height center of gravity above 



24. Effect op the Joint 

The preceding discussion is based on the assumption that the joint affords 
100 per cent efficiency. 

If we examine the functions the joint performs in carrying the load 
from one rail to the other, we see that the splice bars, by fitting tightly to the 
inclined surfaces of the head and base of the rail, are able by their frictdon to 
transmit large horizontal strains from one rail to the next. The proportion 
of the bending moment of the rail transmitted to the splice bar by this i 
is important in determining the correct proportions of the joint. 



,y Google 



260 



STEEL RAII5 



To determine the friction of the bar the following tests were made at the 
Watertown Arsenal in 1904.* There was first made a series of track observa- 
tions on the Boston and Albany Railroad at Faneuil station, near Boston, to 
determine the resistance of nuts on bolts of splice bars as found in the track 
against further tightening. 

Tests were made with a wrench 33 inches long, the resistance against tighten- 
ing being shown by the force required at the end of the wrench to turn the nuts 
forward. The average of 60 observations was 52 pounds on a 33-inch wrench. 

Tests were then made at the Arsenal on the frictional resistance of two 
6-hole splice bars on two sections of 6-inch 100-pound rail. Spring nuts 
were used under the nuts, J-inch bolts, 10 threads per inch, length of wrench 
used 33 inches. The results of the tests are shown in Table LXIX. 

TABLE LXIX. — FRICTIONAL RESISTANCE OF SPLICE BARS 



™"--Ti^5.r"'"""™'' 


Prictioiul Ren 


«.«,». i«n,. 


InitiKl 

(PMind.). 


ContiDuoiu 

33,800 
44,700 
65,500 

66,500 
28,600 


50 


37.500 
46.900 

72,800 
72,800 
31,000 




no — 1 bolt 





The maximimi pull applied to five of the bolts in the third test, 85 pounds 
on a 33-inch wrench, was the limit of strength of the bolts. This pull on the 
wrench caused a permanent elongation of about .06 inch to .10 inch on each 
of the five bolts. The sixth bolt resisted a pull of 110 pounds on the wrench 
without material elongation. 

After making observations on the frictional resistance in these tests, the 
first test, with bolts tightened to 50 pounds' pull, was repeated. 

The spKce bars were now used on one piece of rail, using four bolts, the 
nuts of which were tightened with a pull of 50 pounds on a 33-inch wrench. 
The initial resistance was 50,900 pounds and movement continued under 
31,200 pounds. 

Tests with four bolts in one piece of rail, with 50 pounds* pull on the 
wrench, were repeated with an initial resistance of 59,200 pounds. The 
movement continued under 41,600 pounds. 

* House Documents, Vol. 78, No. 291, 58th CongreBS, 3rd Seaaion, 1904-05, T«ete of MeUls. 



,y Google 



STRESSES IN THE RAIL 2ftl 

The tension on the bolts was reduced during test, and after the last observa- 
tions were made the nuts could be further tightened with a pull of 30 pounds. 
Each nut could be turned up 90 degrees before again attaining a resistance of 
50 pounds on the wrench. 

One-half joint was again made up with four bolts and 50 pounds* pull on 
the wrench. The initial resistance was 66,500 pounds. The slipping of the 
angle bars occurred with a series of throbs, immediately followed in each 
instance by a reduction in the load on the bars. The succeeding throbs took 
place under gradually diminishing loads, following to 49,300 pounds at the 
fourth throb. When removed from the testing machine, the nuts could be 
turned on with an average pull of 35 pounds on the wrench. 

* The experiments carried out by Messrs. R^sal, Foutzen, and Menard on 
the longitudinal slipping of rails connected by fishplates were of four descrip- 
tions, viz. : 

(1) On new rails with new fishplates and bolt holes; 

(2) On the same rails lubricated with mineral oil; 

(3) On old rails with worn bolt holes; and 

(4) On the same with the addition of a thin layer of sand between the 

surfaces of contact. 

It was found that the old rails gave the best results and required a pressure 
of 18 tons to effect any appreciable movement, whereas the new rails were 
least satisfactory, particularly after oihng. As to the experiment with sand, 
it was found that, owing to the reduction of the surfaces of contact caused by 
the sand, the shpping was about the same as in the case of new rails. 

Dr. P. H. Dudley found that a well-fitted splice bar for a 5-inch rail re- 
quired over 4000 pounds per linear inch of one-half of the length of the bar to 
overcome the friction in the rail ends, and for 90-pound and 100-pound 6-inch 
T3XI, 4500 and 4800 pounds respectively 

We are probably not warranted in taking the frictional resistance of the 
joint at more than 40,000 pounds; nor can the friction between the rail and the 
spUce bar be well increased by the use of special joints, without at the same 
time increasing to an imdesirable extent the stresses in the rail, caused by 
sudden changes in temperature. 

It will be seen that this frictional resistance may cause an initial tensile 
stress of about 4000 pounds per square inch in the 100-pound rail at times of 
a sudden fall in temperature. 

* Revue G^nirale dea Cheauns de Fer, Paris, 1908, Vol. 31, pp. 8-14. 



,y Google 



262 STEEL RAIIS 

The tendon set up in rails of lighter section in falling temperatures, before 
they render in the splice bars, is considered by Dr. P. H. Dudley to be 
important and indirectly responsible for a large number of the cracked or broken 
rails which occur during falling temperatures. Records and dates of broken 
rails taken by Dr. Dudley for a number of years, when compared with the 
dates of decided falling temperatures, were found to practically coindde, but 
as soon as the temperature would rise, reheving the rails from tension or 
I putting them in compression, the breakages would cease, except in cases of a 
development of a check which commenced in a falling temperature. 




SHEAR 
Fia. 182, — Shearing Streas in 100-poimd A. 8, C. E. Rail and Splice Bar. Tolal Shear, 24,000 Pounds. 



If we consider the eifect of the frictional resistance between the splice bar 
and the rail, it is apparent that the bar shown in Fig. 182 will act as an 
integral part of the rail until the longitudinal shear at the surfaces of contact 
of tiie rail and the bar exceeds the resistance caused by friction on these 
surfaces. This resistance for a 20-inch splice bar may be taken as 4000 
pounds per linear inch for the entire joint, or 1400 pounds per square inch for 



,y Google 



STRESSES IN THE RAIL 263 

the upper surface of contact, and 500 pounds per square inch for the lower 
surface of contact. 

It is seen from the figure that the surface friction is sufficient to carry a 
total shear at the section of 24,000 pounds, and by referring to the rail diagrams 
given in Figs. 171 and 172 it would appear that the maximum bending moment 
in the ral would be transmitted to the splice bars without slipping. However, 
between the two rails the splice bars must carry the entire moment, and unless 
the section of the bar is increased at the middle of the joint there results an 
excessive deflection at this point. 






183. — 100 pa- 


FiQ. 184. —Joint showing 


Fig. 185. — Joint showing 


cent Joint. 








tion of Metal. 


ot Metal. 



To overcome this source of weakness in the joint, the form shown in Fig. 183 
has been found to embody most of the essential elements demanded by the 
extra reinforcement needed at the center of the joint. This section is only 
used at the middle of the bar and the section shown in Fig. 185 is used for the 
rest of the bar. 

It will be seen that the added metal is distributed in such a way as 
to still keep the vertical axis within the vertical surface that is gripped by 
the bolts. The sectional area and moment of inertia of the reinforcement 
shown in Fig. 183 can readily be adjusted to match the stiffness of the nul 
that is to be sphced, whereas, with the space limitations of Fig. 185, it is 
not jMJssible to get a higher relative jwrcentage of strength than, say, 40, 
as compared with the rml. 

The splice bare, shown in Figs. 184 and 185, show the greater stiffness that 
can be obtfuned by means of a proper distribution of the metal. Taking the 



,y Google 



264 STEEL RAILS 

stiffness of the rail as 100 per cent, the relative stiffness of the bars, shown in 
Figs. 184 and 185, is 29.1 per cent and 37.3 per cent respectively.* 

As far back as 1876 quite full experiments were made with a modified 
type of reinforcement shown in Fig. E, Plate XXIV, on the Swedish Government 
Raih-oads, and in the German handbooks of somewhat later dates quite a variety 
of sections are found of this same general shape. The reinforcing vertical 
fiange occupied a plane at some distance from the axis of the rail, which 
causes the vertical axis of the splice bar to assume a position outside the ver- 
tical surface that is gripped by the bolts, and in consequence the resisting 
stresses in the flange itself must cause an outwardly rotating action, tending 
to strip the threads of the bolts. 

Plate XXIV shows types of joints used in this coimtry. t Fig- A on this 
plate illustrates the common type of angle bar. The variations from this 
section, as applied to 80- and 85-pound rail, are in many directions. A com- 
paratively frequent one is the thickening of the vertical web to I inch. Another 
tendency is to put more metal into the upper part of the web near the under 
Mde of the rail head. An extreme development of this latter practice is shown 
in the Pennsylvania's angle bar for use with its new section of rail. Fig. B 
shows this section. The horizontal extension of the lower flange of the bar 
is another direction in which the angle-bar section is frequently modified. 

There are six patented joints which are now in service in sufficient numbers 
to merit consideration. They may be divided into two classes: those with deep 
girder flanges, namely, the Hundred per cent, the Duquesne, and the Bonzano, 
Figs. C, D, and E; and those which are base-supporting, as the Continuous, the 
Weber, and the Wolhaupter, Figs. F, G, and H. 

The Rail Committee of the American Railway Engineering Association 
have recently made a series of interesting tests on rail joints at the Watertown 
Arsenal, t 

(1) Three joints of each kind were furnished, of which two w^e used for 
testing and the third joint was reserved for future use if needed. 

(2) All joints were full-bolted. Several of the joints flrst tested had various 
sized openings between the rail ends. After the test of the first three joints, 
all other joints were changed so that the opening between the ends of the rails 
was as close to three-eighths of an inch as possible. The span between supports 
in the testing machine was 30 inches. 

* Railroad Age Gaiette, April 9, 1909, p. 804. 

t Railroad Age Gazette, March 19, 1909. 

t Bulletin No. 123, May, 1910, Am, Ry. Eng. & M. of W. Assn. 



,y Google 



STRESSES IN THE RAIL 



266 



(3) One joint was tested with the load first applied to the base, in incre- 
ments of 2000 pounds, until the limit of 32,000 pounds was reached, and then 
the joint was reversed and the load applied on the head imtil the joint failed 
or the limit of the machine was reached. 

(4) The second joint was tested by first applying the load on the head and 
then reversing it, applying the load on the base, until the limit was reached. 

(5) With the exception of the joints furnished by the Cambria Steel Com- 
pany and Mr. A. Morrison, the joints were selected from material which had 



I5OOO0 
















/ 


/ 




























)^ 


/ 




























^ 


D/ 


X 












__ 


^- 










/ 




Y^ 












i- 




^ 


---' 








y. 


P> 












kT 




,-— 


-'^ 


^^ 


-^ 


uj 




/ 


/ 










A- 


^ 






^ 




■^ 








// 








^ 


^-' 




' 




'^ 










- 




/ 






-^ 






/ 




/ 


^ 










s 


/ 


b- 


^ 







=g 


fH 




^ 














a 


/ 


f/. 


-!^ 




" 1 






















50000 


jj 
































II 
































III 
































W 

































> 











s 








9 1 


O 


1 


— 


3 1 


4 1.1 



DEFLECTION.IN INCHF*^ 
Fia. 186. — Diagram of Watertown Arsenal Teste on lOO-poiuM^ Jointa. 
(Am. Ry, Eng. Aasn.) 

been furnished by the manufacturers to the railroad companies in the regular 
routine of business, and therefore fairly represent the material ordinarily fur- 
nished by the manufacturers. 

Figs. A to F (Plate XXV) show some of the joints tested; the results of 
the tests on these joints are presented in Fig. 186. The material in the different 
splice bars varies so widely that it is difficult to judge of the value of the differ- 
ent designs. The excellent results obtained with the Dudley joint {Fig. D) is 
probably due to the high strength of the metal as compared to the other joints 
tested. 



,y Google 



266 STEEL RAII^ 

The rolling mills are reluctant to make splices of higher carbon, Bessemer 
process, than .10 to .20 per cent. Some railroads have specified as high as 
.63 per cent, and with good results both as to manufacture and experience. 
The mill, however, suffers by such a high standard. One mill claims to have 
broken one-third of the total quantity of bars in the straightening process, 
and it also broke many of the punches. 

Splices made from steel of .50 carbon appear to give much better results, 
as might be expected, than the softer steel bars. It is necessary to hot-punch 
the higher carbon steel, and when this is done there is no difficulty in properly 
manufacturing them. The Cambria Steel Company are rolhng bars of this 
grade of steel which are hot finished and oil tempered. 

* The economic advantage of high-carbon steel, hot finished, is that, with 
the expenditure of about 10 per cent more than the cost of soft steel, a joint 
is ^ven a carrying capacity that can be equaled only by the addition of double 
the quantity of metal of soft steel and at the additional cost of 100 per cent. 
This latter joint will cost 100 per cent more for freight, while there is no addi- 
tional cost for freight in the former. 

The oil treatment of steel is a natural sequence of the use of high carbon, 
and its advantages are about equal to those of high carbon over soft steel. 
This, however, varies with the section of the bar and hardness of the steel. In 
economy, oil-treated steel is as much in advance of high-carbon steel as is the 
latter over soft steel. 

Intimately connected with the rail-joint problem is the question of the 
length of the rail. In a recent bulletin (August, 1909) of the International 
Railway Congress, the practice in English-speaking coimtries is very fully dis- 
cussed and abstracts from this report are given below. 

In Great Britain and Ireland the railways have been gradually increaang 
the length of rails, with a view to reducing the number of joints. Some rail- 
ways still use rails only 30 feet long, and a few iise 60-foot rails, but a large 
number have 45-foot rails, and it appears that this may be taken as a standard 
for the near future. The principal reasons ^ven for limiting the length, as 
given by the engineers of different railways, may be summarized as follows: 
(1) Difficulty of straightening rails at the mills; (2) cost of manufacture; (3) 
difficulties of transportation; (4) expansion and contraction; (5) unloading and 
handling on the track. 

So far the use of long rails does not appear to have called for the 
adoption of any special arrangements other than proper proportioning 

* Railway Age Gazette, March 16, 1010, daily edition. 



>, Google 



STRESSES IN THE RAIL 267 

of the bolt holes and the play at the joints, and the strengthening of 
the maintenance gangs (by consolidating neighboring gangs or otherwise) 
while handling long rails. , The temperature varies from 0° F. in winter to a 
maximum of 130° F. in the sim in summer. To allow free play for expansion 
during extreme heat, it is the practice of most en^neers to ease the joints by 
slackening the nuts. 

In the United States the standard length on a number of railways is 33 feet, 
and the reasons given for limiting the length are, in general, similar to those 
noted above. Experiments have been made with rails of greater length, but 
on the whole these have not been satisfactory, although the opinions expressed 
by some of the railways give 40 feet, 45 feet, 50 feet, and 60 feet as admissible 
lengths. The range of temperature is 100° F. in some parts of the country, 
while in others it is 180° F. In some cases the nuts are slackened in the early 
part of the summer. 

In the roads of other countries investigated which include 17 railways 
in South America, India, South Africa, Australia, and Canada, the limiting 
length of rails varies from 30 to 40 feet. The range of temperature is from 
100° to 155° in India, and 160° in South Africa and Australia. 

Inquiry was made as to whether any railways contemplated welding the 
rails at the joints. All the replies were in the negative, and the general opinion 
was that continuous rails would be unsafe, on account of the temperature 
changes. It is well known, of course, that welding the joint is common practice 
in street railway work, but in such cases the rails are protected by the paving, 
so that only a small portion is exposed. 

The following very interesting report from the Pennsylvania Lines is quoted 
in the bulletin: 

" In 1897 a continuous rail, 1050 feet long, made up of 35 80-pound 30-foot 
rails joined by angle bars with drilled holes and machine-turned bolts (no pro- 
vision being made for expansion and contraction), was laid in the eastbound 
m^n track, near New Brighton, Pa. The ends were held by bent rails bedded 
in concrete, so placed as to bear against the ties. Special long and wide angle 
bars were used at the ends, fastened to the anchor ties with lag screws. The 
track was a tangent with stone ballast. 

" The rail crept and kinked out of line badly. An examination made in 
August, 1900, after three years' service, showed that the entire rail crept in 
the direction of traffic (eastward). At the west anchorage, the vertical holding 
rals had cut into the cross-ties forming the anchorage framework, while at the 
east anchorage there was a space of If inches between the vertical rails and the 



,y Google 



268 STEEL RAILS 

framework. All of the spikes were bent eastward, and both slots and spikes 
were badly cut. The bolts were all slightly sprung. The alignment at the 
joints was verj- bad." * 

The conclusions presented in the bulletin are ^ven below: In Great 
Britain and Ireland, the lengthening of rails and the consequent reduction of 
the number of joints has been steadily proceeding at an increasing rate during 
the past 60 years. In 1840-50, the normal length of rolled iron rails was 
from 15 to 18 feet. The length of these iron rails increased at the rate of about 
3 feet in each decade until 1870-80, when steel rails from 24 to 30 feet long 
were brought into general use. Since then the decennial increase has been about 
4J feet, and at the present time rails 36 feet and 45 feet long are in general use, 
while two railways have adopted 60 feet in length. 

In 1904, when the Engineering Standards Committee issued the British 
Standard sections, it recommended the adoption of the following as the normal 
lengths of rails: 30, 36, 45, or 60 feet. In other countries embraced by the 
report, the length of rals has been steadily and uninterruptedly increasing, 
but within narrower limits than in Great Britain and Ireland. 

The conclusions to be drawn from the numerous replies and remarks by 
engineers throughout the English-speaking coimtries are that there is a maxi- 
mum length of rail somewhere between 33 feet and 60 feet, which should not 
be exceeded; and continuous or welded joints over a long length of railway are 
impracticable and dangerous. 

BIBLIOGRAPHY 

Bldu. — Beport ... on the quealion of rail joinla. (All countrieB except France, Belgium, 
Italy, Spain, Portugal, Austria-Hungary, Rumania, Bulgaria, Servia, Turkey, Egypt, and countries 
using the EngliBh language.) 48 p. 111. 1910. (In Bulletin of the International Railway Con- 
gress, Vol. 24, Part 1, p. 1701.) 

Bouchard, H. ■ — Note aur le joint asymfitrique. 4000 w. 111. 1909. (In Revue g&i^rale des 
chemins de fer. Vol. 32, p. 9.) 

Describes tJieory, construction, and results with favorable r^l joint. 

Chatgad. — Report ... on the question of rail jointjj (France, Belgium, Italy, Spain, and 

* The author several years ago had experience with a continuous rail designed by the late 
Mr. Torrey. 

The rail was about the length and of the same weight as in the Pennsylvania test. It was 
l^d on a branch line of the Michigan Central Railroad on a tangent and ballasted with a good 
bed of gravel. Provision was made at intervals of about 500 feet tlwiughout the test track for 
the expan^on of the contiimous rail by means of special cxpan^on joints, the rail being anchored 
midway between these joints. 

When first installed the riding; qualities of the track were exceptionally good, which may, 
however, have been accounted for by the unusual care that had been taken in constructing it. 
After several years the test track appeared to ride about aa well and to require the same amount 
of attention as the other track on the division. 



>, Google 



STRESSES IN THE RAIL 269 

Portu^l). 38 p. 111. 1910. (In BulleUn ot the InternaUonal Railway Congrees, Vol. 24, Part 1, 
p. 1427.) 

Edeutein, Leon. — Prevention of play between rail and fishplate. 2000 w. III. 1908. 
(In Bulletin of the International Railway CongrcBs, Vol. 22, Part 1, p. 436.) 

Shows faults that develop with wear, and gives Bu^eetiona for prolonging life of both raiU 
and fishplates . 

GoDFERNADX, R, — Note about rail-joint.H. 24 p. III. 1911. (In Bulletin of the Inter- 
national Railway Congress, Vol. 25, p. 14S0.) 

Reviews development of rail joints and different forms used. 

Haaruann, a. — Der scbienenstoas. 5000 w. III. 1911. (In Stahl und eisen, Vol. 31, Part 
1, p. 49.) 

Discusses types of rail jointu and probable future practice in track construction. 

Improvements and experiments in rail joinia. 2500 w. 1910. (In Engineering News, Vol. 64, 
p. 281.) 

Abstracts information from reports to Int^^mational Railway Congress on practice in the United 
States, Great Britain, France, Belgium, and Austria-Hungary. 

International Railway Conobebh. — iGeneral discussion on ndl joints.] 38 p. 1911. (In 
Bulletin of the International Railway Congress, Vol. 25, p. 405.) 

Kramer, Friedrich. ^ Report ... on the question of rail-joints (Austria, Hungary, Rumania, 
Bulgaria, Servia, Turkey, and Egypt). 82 p. III. 1910. (In BuUetin of the International Rail- 
way Congress, Vol. 24, Part 1, p. 1967.) • 

Jones, Cyril Walter Lloyd, — Design of fishplate rail joints. 7000 w. Rt. 1910. (In 
Minutes of Proceedings of the Institution of Civil Engineers, Vol. 182, p. 282.) 

Pellarin. — Etude des joints des rails. 27 p. 111. 1908. (In Annales des ponts et chauBS^, 
M^moirfe, series 8, Vol. 33. p. 98.) 

Report of commission for investigation of rail joints in Bel^um, Italy, Switaerlwid, and Holland. 

Ross, Alexander. — Report ... on the question ot rail joints. (Countries using the English 
language.) 44 p. III. 1909. (In Bulletin of the International Railway Congress, Vol. 23, Fart 2, 
p. 689.) 



,y Google 



CHAPTER V 
strenotm of the rail 

25. Influence of Stress and Strain on the Strength of the Rail 

In determining the safety of any structure, not only the amount of stress 
induced in the different members by the load must be found, but its character 
and the effect it may produce on the material of which the staiictm-e is composed 
must be considered. 

A series of rapidly repeated stresses will, under certain conditions, affect 
the breaking strength of the metal. The cause of this loss of strength under 
the influence of repeated stress is a much-mooted question among engineers, 
and it is of interest to examine the problem in detail.* 



Fio. 187. — Pure Swedish Iron. (Mellor.) Fig. 188. — Pure Copper, (Araolii.) 

The junction of the crystalline grains of pure iron, shown in Fig. 187, and 
of pure copper, shown in Fig. 188, are typical of pure metals; but when impuri- 
ties ai'e present the crystals of the pure metal, in the act of crystallizing, reject 
the impurities which collect at the crystal boundaries. The particles of pure 
metal slowly migrate and coalesce together, so as to form little islands sur- 
rounded by the impurity; accordingly, in the solidified mass, we find the crystals 
of pure metal enveloped by a film of the metal associated with the foreign sub- 

* The author is indebted for the discussion on pp. 270-276 to Mr. J. W. MeUor's nt)rk. The Crys- 
tallizaUon of Iron and Steel, Longmans, Green & Co., London, 1905. 



>, Google 



STRENGTH OF THE RAIL 271 

stance. This investing membrane separates the crystals of pure metal one 
from the other. Obviously, the mechanical and physical properties of the alloy 
will depend upon the character of the film. 

The mass of pure metal, for example, may be quite ductile like gold, while 
the mass of metal with the impurity may be quite brittle, as Arnold * found 
to be the case with an alloy of gold with .2 per cent of bismuth; and copper 
containing .5 per cent bismuth. A representation of the latter alloy is shown 
in Fig. 189. When such a metal is fractured, the hne of fracture follows the 
junction of the grains. 

Stead calls this ailment intergranular or intercrystalline weakness (inter,- 
between). We have had examples. Arnold's work on the influence of bismuth 



Fio. 189. — Copper-bismuth Alloy. FiG. 190, — Iron with 1.8 per cent Carbon. 

(Arnold.) (F. PoppleweQ.) 

on copper and on gold. One per cent of sulphur arranged as a mesh of iron 
sulphide will entirely destroy the ductility of the iron, reducing the ultimate 
stress from 20 to 2 tons per square inch. 

The network of cementite which envelops the crystal grains of steel cen- 
tring over 1 per cent of carbon are the principal lines of weakness. The 
metal when fractured generally breaks through the center of this brittle envelope. 
The coefficient of contraction of the cementite cell walls is greater than of the 
cell contents. Pearlite cells, for example, bound together by thick cementite 
walls (Fig. 190), are liable to rupture, because the coefficient of contraction of 
the cementite cell walls is greater than the cell contents. The mass is, in con- 
sequence, very feebly held together, and a sudden blow will easily fracture the 
metal, t 

■ J. O. Arnold and J. Jefferson, Engineering, 61, 177, 1896. 
t J. 0. Arnold, MetallographiBt, 5, 267, 1902. 



,y Google 



272 STEEL RAIIS 

Intergranular weakness resembles the weakness of a brick building with 
faulty mortar. 

There is another type of intergranular weakness which is due to imper- 
fect union of the crystal grains. This is particularly marked in phosphorous 
steels. The crystal grains, on cooling, contract unequally and tend either 
to draw the grains away from each other or to leave the mass in a state of 
unnatural tension. The fracture then follows the granular junctions. Thick 
plates and bars are frequently brittle because comparatively little work has 
been done on them. The crystals are not interlocked one with another, as in 
steel which has been well worked. 

Intergranular weakness may, therefore, be of two kinds: 

(1) Brittle envelope surrounding the crystal grains; 

(2) Imperfect union of the crystal grains. 

Stead has pointed out another type of weakness in sheet steels which has 
to do with the crystals themselves, without reference to the imion of one crystal 
with another. It is a kind of intracrystalhne weakness (intra, within). 

It is characteristic of some crystals to break more readily in some directions 
more than in others. This property of crystals is called cleavage. The direc- 
tions in which the crystal splits are called cleavage planes. If a bar of iron 
could be cut from a single crystal, that bar would have three lines of weakness 
in the direction of the three cleavage planes; while if the bar were built up of 
a number of crystals whose cleavage planes were all in the same direction, that 
bar would be more readily broken in the direction of its cleavage planes, neglect- 
ing for the moment intergranular weakness. On the other hand, if the cleavage 
planes of the adjacent crystals are inclined at considerable angles to one another 
the bar would be less liable to break than one in which the crystals were arranged 
symmetrically. Figs. 191 and 192 will make this clear. The dotted hues ab, 
Fig. 191, represent the cleavage planes across a sheet of iron when the crystals 
are arranged symmetrically, while in Fig. 192 the crystals are arranged m an 
irregular maimer. The cleavage planes of Fig. 191 run along parallel lines, 
and the sheet would, therefore, be more liable to rupture than the sheet shown 
in Fig. 192, where the lines of weakness are not in the same direction, and this 
in spite of the fact that Fig. 191 has a finer grain. 

Other things being equal, a fine-grained structure is stronger and tougher 
than a coarse-grained piece. Figs. 191 and 192 show that this order of things 
may be reversed. Fortunately, the crystals of one steel do not generally grow 
symmetrically. 



,y Google 



STRENGTH OF THE RAIL 





Pia. 191. — Cleavage Planes with Crystals a 
ranged Symmetrically. (J. W. Mellor.) 



Flo. 192. — Cleavage Planes with Crystals arranged 
ID an Irregular Manoer. (J. W. MellorO 



Fio. 193. — Iron strained beyond the Elastic Limit. 
{J. A. Ewing and W. Hoeenhain.) 



Fig. 194. — Lead Btreioed beyond the Elastic 
Limit. (J. A. Ewbg and W. Roeenhain.) 



Examining now what takes place in the metal under repeated stress, Ewing 
and Rosenhain * found if a metal is strained past its " yielding point " — elastic 
limit — the faces of the crystal grains (Fig. 193) show fine black lines, which 
increase in number as the strain increases. Lines appear on certain crystals 
nearly transverse to the pull, as the strain increases lines appear upon other 

• J. A. Ewing and W. Rpeenhain, Phil. Trans., 193, 353, 1899; 195, 279, 1900. J. A. Ewing 
and J. C. W. Humfrey, Ibid., 200, 241, 1902. W. Rosenhain, Journal Iron and Steel Inst., 67, i, 335, 
1904. F. Osmond and C. Fi^moat and G. Cartaud, Revue de M£tallurgie, 1, i, 1901. 



,y Google 



274 STEEL RAILS 

grains. Intersecting lines then make their appearance on some of the grains. 
Such a strained surface is shown in Fig. 194. 

The lines are apparently not actual cracks in the surface, but rather slips 
along the cleavage planes of the crystal. They are called slip bands, or 
slip lines. 

Let AB (Fig. 195) represent a cross section through a polished surface 
of metal. Let C be the junction between two contiguous grains, A and B. 
When the metal is pulled in the direction of the arrows, a number of slips are 
developed along the cleavage planes, a,h,c,d . . , and the surface now presents 



// / '-v 



Fia. 195. — CroBs Section of Unstrained Metal. Fio. 196. — Cross SectioD of Metal after being 

(J. W. MeUor.) Streased. (J. W. MeUor.) 

the appearance shown in Fig. 196. With still greater strains the slip bands 
develop into actual cracks, and rupture takes place. Hence it follows that 
under progressively augmented strain, rupture takes place, not at the crystal 
boundaries, but through the crystals themselves. 

Ewing and Humfrey have subjected Swedish iron, with a breaking stress 
of 23.6 tons per square inch, to a series of compresaon and tension stresses, 
9 tons in magnitude, repeated 400 times per minute. On examination it was 
found that fine slip bands appeared in a few crystals after a few, say, 5000 
reversals of stress; with a greater number, say, 40,000, the slip bands increase 
in number, and those which first appeared broaden and develop into small 
cracks, as shown in Fig. 197. 

If the specimen be repolished, so as to clear off the slip bands, the cracks 
alone become visible, as at A (Fig. 198). The crack, or flaw, gradually creeps 
across the specimen when the number of alternations is still further increased, 
as shown in Fig. 199. Finally the specimen breaks. 

Ewing and Humfrey state: " Whatever the selective action of the stress 
is due to, the experiments demonstrate that in repeated reversals of stress 
certain crystals are attacked, and yield by slipping, as in other cases of non- 
elastic strain. Then, as the reversals proceed, the surfaces upon which the 
slipping has occurred continue to be surfaces of weakness. The parts of the 
crystal lying on the two sides of each such surface continue to slide back and 
forth over one another. 



,y Google 



STRENGTH OF THE RAIL 



Fio. 197. — SUp Bands. (J. W. Ewing and J. C. W. Humfrey.) 




Fio. 198. —Polished Surface with Small Cracks. (J. W. Ewing and J. C. W. Humfrey.) 



,y Google 



276 STEEL RAILS 

" The effect of this repeated sliding or grinding is seen at the polished 
surface of the specimen by the production of a burr or rough and jagged irregu- 
lar edge, broadening the slip band, and su^esting the accumulation of debris. 
Within the crystal this repeated grinding tends to destroy the cohesion of the 
metal across the surface of the slip, and in certain cases this develops into a crack. 

" Once the crack is formed, it quickly grows in a well-known manner, 
by tearing at the edges, in consequence of the concentration of stress which 



Fia. 199. — Polished Surface with Large Cracks. (J. W, Ewing and J. C. W. Humfrey.) 

results from lack of continuity. The experiments throw light on the known 
fact that fracture by repeated reversals or alternations of stress resembles 
fracture resulting from ' creeping flaw ' in its abruptness, and in the absence 
of local drawing out, or other deformation of shape." * 

The rupture of steel is not caused by the gradual growth of the crystal- 
line structure of the metal under the influence of shocks and vibrations. The 
breaking down is due to fatigue. When fatigued, the metal breaks more readily. 
Again, when subjected to sudden shock, the metal has no time to " flow." The 

• P, Kreuzpointer, Joum. Franklin Inat., 1S3, 233, 1902. J. A. Ening, Nature, 70, 187, 1904. 
P. Breuil, Suppl. Joum. I. & S. Inet., 1904. 



,y Google 



STRENGTH OF THE RAIL 277 

slipping of the crystal planes, or the plasticity of the metal, has no time to come 
into play.. The metal, in consequence, appears to be abnormally brittle. 

* The experiments made by Wohler, from 1859 to 1870, were the first that 
indicated the laws which govern the rupture of metals under repeated applica- 
tion of stress. For instance, he found that the rupture of a bar of wrought 
iron by tension was caused in the following ways: 

By 800 applications of 52,800 pounds per square inch. 

By 107,000 applications of 48,400 pounds per square inch. 

By 450,000 applications of 39,000 pounds per square inch. 

By 10,140,000 applications of 35,000 pounds per square inch. 
Here it is seen that the breaking unit stress decreases as the number of 
applications increases. It was further observed that a bar could be strained 
from up to a stress near its elastic limit an enormous number of times without 
rupture, and it was also found that a bar could be ruptured by a stress less than 
its elastic limit under a large number of repetitions of stress which alternated 
from tension to compresdon and back again. 

t The apparatus used by Wohler and his successor, Spangenbei^,t was 
of four kinds: 

(1) To produce rupture by repeated load; 

(2) For repeated bending, in one direction, of prismatic rods; 

(3) For experiments on loaded rods under constant bending stress; 

(4) For torsion by repeated stress. 

The amount of the imposed stress was determined by breaking several 
rods of like material, ascertaining the breaking load, and taking some fraction 
of this for the intermittent load. 

From the results of these experiments of Wohler, extending ova- eleven 
years, the following law was deduced: 

" Wohler's Law: Rupture of material may be caused by repeated vibra- 
tions, none of which attain the absolute breaking limit. The differences of 
the limiting strains are sufficient for the rupture of the material." 

The work of Wohler and Spangenberg has proved what was long before 
supposed to be the fact: that the permanence and safety of any iron or steel 
structure depends not simply on the greatest magnitude of the load to be sus- 
taned, but on the frequency of its application and the range of variation of 
its amount. 

• A. Wohler, En^neeringy 1871 ; Zeitschritt flir Bouwesen, 1870, p. 83, Berlin. 
t Iron and Steel, Materials ot Engineering, Part 2, Thuraton, 1909, p. 618, 
i Zeitachrift fQr Bouwesen, 1874, p. 485, Berlin. 



,y Google 



278 STEEL RAILS 

Prof. L. Spangenberg resumed the line of experiments at the point of its 
discontinuance by Wohler, and his results tend to confirm the law of the latter. 
Spangenberg directed his attention to other metals than iron and steel, and 
also endeavored, by inspection of the surfaces of fracture, and by his hypoth- 
esis as to the molecular constitution of metals, to explain the phenomena of 
fracture. Among the several observations noted in his " Fatigue of Metals " 
is the important fact that when subjected to often-repeated transverse stress 
fracture of iron took place only on the tension side of the bar and extended 
only to the neutral axis. From this he inferred that the working strength of 
wrought iron is less than its elastic resistance. 

Fowler states, in this connection, that a steel rail tested for transverse 
strength in a machine will, as a rule, bend many inches, and fail by distortion of 
the head under the compressive stress. In actual work hundreds of such rails 
break, but it is the tensile and not the compressive stress which causes the 
failure, and there is no distortion of the head, as in the testing machine. 

Reynolds and Smith * extended Wohler's conclusions to steel bars tested 
under direct tension and compression and at a rapid rate of alternation. The 
work of Stanton and Bairstow,t published in 1906, while less concordant than 
that of Reynolds and Smith, confirms their general conclusions; it extends the 
conclusions of Ewing and Humfrey to notched specimens tested for endurance 
under direct tension and compression, and it clearly points out the advisability 
of testing a material for endurance in approximately the form in which it is 
to be used in practice. 

French engineers, commenting upon the work of Wohler, Spangenberg, 
Weyrauch,J and Launhardt, consider that the result is simply to base upon the 
ultimate strength a deduced limit of working stress which corresponds closely 
to the elastic limit, and generally urge the use of a reasonable factor of safety 
related to the limit of elasticity.! 

Figs. 200, 201, and 202 present three diagrams on the behavior of steels 
under repeated alternate stresses, illustrating some of the tests which have 
been made at the Watertown Arsenal laboratory.jl 

On Fig. 200 the heavy vertical lines represent the number of loads which 
were applied to a number of steel bars of .55 per cent carbon, and which 

• Phil. Trans. Royal Society of London, A-Voi. 199, p. 265, 1902. 

t Proceedings Inst, of Civil Engra., Vol. 166, p. 78, 1906. 

t Various MeUiods of Determining Dimensions. Dr. J. Weyrauch; translated by G. R, Bodmer. 
Proc. Inst. C. E., 1882-83, Vol. LXXI. 

3 R£sum« de la Soci^t^ des Ingfnieura avils, 1882. 

II Notee on the Endurance of Steels under Repeated Alternate Streea. Howard, Proceedings 
Am Society for Testing Materials, 1907, Vol. VIL 



,y Google 



STRENGTH OF THE RAIL 



279 



caused rupture of the metal. Be^nning with the highest fiber stress, 60,000 
pounds pw square inch, the progressive gain in endurance of the steel as the 
loads were successively reduced will be noted, as indicated by the lengths of 
the different lines. The lowest fiber stress experimented with did not end in 



REPEATED ALTERNATe STRESSES 
o' '.55 CAHBON STEEL. 



TENSrUE TEST 
Q-ASnC LIMIT 59^0 LBS Ftn SOfCH 
TENSILE STRENGTH 111.200 
ELONGATION 12 PER CENT 

CONTRACTION 33,5 



REPEATED ALTERNATE STRESSES 
.82 CARBON STEEL., 

TENSILE TEST 
CLASTIC LIMIT S4000LBS,PCnSO.INCH 
TENSILE STRENGTH H2JB0Q - ■ - - 
ELONGATION 7 PEfl CENT 

CONTRACTION It8.~ 



gi 



I I I 

40,000 45000 5OO0O SSXHJO ' 60.000 
3DflOO 3^000 4apD0 4EpOO MpjO SOjOOO TIBRE STRESSES, LBS. PER SOUARE INCH 

FIBRE 'STRESSES. LBS. pen SQUARE INCH. Ir^NOT RUPTURED 

Fio. 200. — Behavior of 0.55 Carbon Steel under Repeated Fia. 201. — Behavior of 0.82 Carbon Steel under Repeated 
Alternate Streases. (Am. Soc. tor Testing Materials. — Alternate Stresses. (Am. Soc. (or Teeting Materials. — 

Howard.) Howard.) 

rupture of the shaft; after 76 million repetitions, in round numbers, under a 
load of 30,000 pounds per square inch, the fiber stress was incrrased to 60,000 
pounds per square inch, which higher load caused ruptiu-e after about 8000 
rotations. The enormous gam in endurance of the steel, under 30,000 pounds 
fiber stress, over its behavior with the higher loads, would be represented by 
a vertical line about 28 feet in height, according to the scale of this diagram. 
The results of the tendle tests of this grade of steel ^e entered on the diagram, 



,y Google 



280 STEEL RAILS 

from which it may be seen that the several fiber stresses were, with one exception, 
below the tensile elastic limit of the metal. 

On Fig. 201 similar Imes represent the behavior of specimens containing 
.82 per cent carbon. The behavior of this grade resembles that of the previous 
diagram, and similar to other steels belonging to this series of experiments. 
The endurance under corresponding loads is seen to be greater than displayed 
on the preceding diagram. After making 202 million rotations the test of the 
shaft loaded with 40,000 pounds was temporarily discontinued, the steel being 
unruptured. A line drawn to scale to represent the endurance imder this load 
would be about 73 feet in height. 

On Fig. 202 appear curves representing the relative endurance of each 
of the six grades of steel used in this series of experiments. Their endurance 
under the higher fiber stresses only are shown, loads which caused compara- 
tively early rupture of the steel in most of the tests. 

It may be remarked that the fiber stresses experimented with were gener- 
ally below the tensile elastic limits of the steels. The greater endurance of the 
steels of .73 and .82 per cent carbon in comparison with either the higher or the 
lower carbons is an interesting feature of the tests. 

Elastic properties only are displayed by steels, prior to rupture when rup- 
ture is caused by a large number of alternate stresses of tension and compression ; 
no appreciable display of ductility, as shown by elongation and contraction of 
area, need precede rupture, in any grade of steel, following the application of 
stresses of this kind. If the fiber stresses somewhat exceed the tensile elastic 
Kmit, a hmited display of elongation, other than elastic, may occur, but rupture 
caused by loads which are in the vicinity of or below the tensile elastic hmit 
is not attended with an appreciable display of ductility. 

While tests by repeated alternate stresses are characterized by the absence 
of ductility, as witnessed in tests by tension, there may be elastic movements 
of the metal aggregating considerable distances. The a^regate exten^on of 
the most strained fiber of the .82 per cent carbon steel which has successfully 
endured 202 miUion repetitions amounts to nearly 5 miles per linear inch of 
specimen, a distance quite incomparable to the permanent extension of the 
metal in the tensile test. 

* Little attention seems to have been pven to the possibility of finding 
a relation connecting the stress used in endurance tests with the number of 
repetitions required for rupture. Spangenberg, Reynolds, and Smith, and 

'The Exponential Lan of Endurance Tests, 0. H. Basquin, Proceedin(p of the American Society 
for TeBtiog Materials, IBIO, Vol. X, p. 835. 



,y Google 



STRENGTH OF THE RAIL 



281 



Stanton and Bairstow, have shown stress-repetition curves drawn to ordinary 
Cartesian coordinates. 

Logarithmic coordinates present a distinct advantage in the study of 
simple exponential curves, because these curves become straght lines for these 



REPEATED ALTERNATE STRESSES. 




GRADE .I7C. .340. .55C. 730. .B2C- I09C. 

E.L. OlflOO 54,000 59.000 64.000 64flOO 77.000 

Fia. 202. — Compariaon of the Behavior of DifFerent Grades of Steel under Repeated Alternate Stresies 
(Am. Soc. for Testing Materials. — Howard.) 

coordinates and their equations may be written at once. Fig. 203 shows 
stress-repetition curves for nineteen sets of endurance tests, made by five dif- 
ferent observers. The names of the experimenters, the kind of test, and the 
material are ^ven in Table LXX. 



,y Google 



STEEL BAILS 



TABLE LXX. — EXPERIMENTS ON REPEATED STRESS 

(BuqniiO 



Curv. 

LMWr. 




Kind of 
Test. 


Uauriil. 


rrboiiu«bi| 


E>po«t 






h 

b 


Wrought-iron axles, Phoeoix Co . . . 
Wrought-iron axles, Phoenix Co. . , , 


217 
109 
103 

94 
130 

36 

1000 
920 
320 
310 
90 
97 
115 
250 
102 
66 
110 
150 
135 


0.12 
















D' 


1 


Cast steei, Borsig 

Homogeneous iron, P. C. & Co 


-0.11 






OS 






Cast iron, locomotive cylinder. . 

Krupp'B spring steel, hardened. . . 

Krupp's axle steel 




H 




-0.20 














t; 






















Beni 


nmmRober 






p 

Q 


Reynolds & Smith ' ' 


Hardsteel 


-0.15 


R 






s 


F. B 




Unannealed steel 












Steel B, grooved 











ic. 



K bsA round shoulden near grip; L has aqnare Hhoutden. 
OrdiDaUa are " tlalf Ruicc at Stn»," innud of muiinum ttnte. 
.3S peir cmt ; yield point. 39.000 jxHodB par aquaie iuh. 



Kinds of TesU : 

(a) Bending in one direction only (+ and 0). 

(b) Rotating under bending load (+ and — ). 

(c) Tension only (+ and 0). 

(d) Bending hack and forth (+ and — ). 

(e) Tension and smaller compression (+ and ■ 



The curve A is represented by the equation 
5 = 217,000 R-°-'\ 
which has the form 

S = CR", 

in which 5 is the maximum stress used in each test and R is the number of repe- 
titions of this stress required for rupture. The coefficient, 217,000, was found 
by extending the line A to the left until it intersected the vertical line R = 1 
(i.e., 10°), and the stress at this intersection was read off the logarithmic scale as 
217,000 pounds per square inch. The coefficient is the stress given by the curve 
for a single repetition. All the coefficients given in Table LXX in the column 
marked C were found in the same way. The value of the exponent, -0.12, 
was found by measuring the angle (130°) which this line makes with the hori- 
zontal axis and then taking one-tenth of its natural tangent. The factor " one- 
tenth " comes in because, in Fig. 203, the scale used along the vertical axis 



,y Google 



STRENGTH OF THE RAIL 



283 



in plotting the stresses is ten times the scale used along the horizontal axis in 
plotting the repetitions. In the same way the exponent for each curve of 
Fig. 203 has been found and is listed in the table under the column marked n. 

In looking over the curves, Fig. 203, it is evident that in many cases the 
straight line represents the results of endurance tests very accurately through- 
out a considerable range of stress. One is also impressed with the approximate 




Fio. 203. — Number of Repetitions before Ruptui 
(Am. Soc. for Testing Materiala. - 



Endui 



parallehsm of many of these lines. Curves B, C, D, E, F, G, M, N, and S 
represent tests made in much the same way, — by rotating a specimen under 
bending load. The curve for hard steel, tested by Baker * in much the same 
way, has a steeper slope; the same is also true of the grooved specimens tested 
by Foppl-t Curves H, J, K, and L are approximately parallel and represent 
tests in one direction only; i.e., the stresses are not reversed. They have about 

• Some notes on the Working Stress on Iron and Steel, Trans. Am. Soc. of Mech. Engre., 1887, 
Vol. VIII, p, 157. 

t Mitteilungen aus dem Mecb. Engre., Vol. 130, 1909. 



,y Google 



284 STEEL RAlLii 

double the slope of the other curves mentioned. Why curve A, on wrought 
iron, does not fall into this class is not clear. 

" Small changes in temperature occur when a bar of metal is stressed 
within the elastic limit, it becoming cooler under tension and warmer under 
compresMon. The measurements of these changes made by Turner, in 1902, 
have shown that these changes in temperatures continue at a uniform rate up 
to about three-fifths of the elastic limit, and that then a marked change occurs, 
the bar under tension then beginning to grow warmer, while the temperature 
of the bar under compression increases at a more rapid rate. 

It thus appears for stresses higher than above three-fifths of the elastic 
limit, at least, energy is converted into heat under repeated applications; prob- 
ably this occurs also at lower stresses when repeated stresses range from ten- 
sion into compression in a bar, or when a beam is subjected to alternating flexure. 

In the case of the rail the bending stress alternates about in the proportion 
of 4 to 1, and it is very probable that by taking a unit stress less than half tiie 
elastic limit, we may safely ignore the effect of fatigue on the metal of the rail 
produced by this stress. The disturbed metal at the running surface of die head 
which is frequently conspicuous ui old rails is evidence of the elastic limit of the 
metal bemg exceeded rather than the ^ect of repeated stress beknv this limit. 
Generally speaking the effect of repeated stress is not to produce distortion of fiie 
metal, and prior to rupture elastic properties only are displayed. 

26. Effect of Low Temperatures on the Strength of the Rail 
Very complete investigations were made to determine the effect of 
changes of temperature in modifying the physical properties of iron and steel by 
Styffe and Sandbcrg.f The conciuaon of Styffe is that the strength of iron 
and steel is not diminished by cold. Aiding from these experiments, investi- 
gators have assumed that the cause of the frequent breakage of rails in cold 
weather, and of articles made of iron and steel, is unequal expansion and con- 
traction and the rigidity of supports, where, as in the case with rails, frost may 
very greatly affect them. 

X Sandberg, while admitting the care and the accuracy which distinguished 
this extensive series of experiments, still doubted whether the reasons just 

• Mechanics of Materials, Merriman, p. 354, New York, 1905; and Trans. Am. Soc. of Civil 
Engrs., Vol. XXVIII, 1902, p. 27, Thermo Electric Measurements of Stress, Turner. 

t The Elasticity, Extensibility, and Tensile Strength of Iron and Steel, by Knut Slyffe, trans- 
lated from the Swedish, with an original appendix by Chriater P. Sandberg; with a preface by John 
Percy, M.D.. F.RS., London, 1869. (Sandberg's investigations appear in the appendix.) 

\ Iron and Steel; Materials of En^neering, Tburaton, 1009 p. 556. 



,y Google 



STRENGTH OF THE RAIL 285 

given were the sole reasons why metals should more readily break in cold than 
in hot weather, and, having obtained the consent of the State Railway Adminis- 
tration, he conducted a series of experiments in the summer and winter of 1867, 
at Stockholm, to determine whether, with equal rigidity of supports, iron rails 
would yield with equal readiness to blows at the two extremes of temperature. 

The rails experimented upon were each cut in halves, and one piece was 
tested in cold and the other in warm weather, at temperatures of 10 degrees and 
84 degrees Fahr. respectively. The supports at the end of the rails were granite 
blocks placed four feet apart, and resting on the smoothly leveled surface of the 
granite rock. They were broken by a heavy drop weighing 9 cwt. 
Sandberg's conclusions from twenty experiments are thus given: 

" (1) That for such ux)n as is usually employed for rails in the three principal 
rail-making countries (Wales, France, and Belgium), the breaking strain, as 
tested by sudden blows or shocks, is cozisiderably influenced by cold, such iron 
exhibiting at 10 degrees Fahr. only from one-third to one-fourth of the strength 
which it possesses at 84 degrees Fahr. 

" (2) That the ductility and flexibility of such u-on is also much affected 
by cold; rails broken at 10 degrees Fahr. showing, on an average, a perma- 
nent deflection of less than one inch, whilst the other halves of the same 
rails, broken at 84 degrees Fahr., showed a set of more than 4 inches before 
fracture. 

" (3) That at summer heat the strength of Aberdare rails was 20 per cent 
greater than that of the Creus6t rails, but that in winter the latter were 20 per 
cent stronger than the former." 

Sandberg suggests that this considerable lack of toughness at low tem- 
peratures may be due to the " cold-shortness " produced by the presence of 
phosphorus. 

Jouraffsky, of St. Petersburg, has reported* the results of tests of rails 
made for the Russian government, which supplement the preceding in a very 
valuable manner. It was found that by placing pieces of rail from 6 feet to 
8 feet long in a mixture of ice and salt, the temperature of the rail could be 
lowered in a very short space of time, during warm weather, 36 degrees Fahr. 
below freezing point. 

A special commission, Messrs. Erakoff, Beck, Guerhard, Nicolia, and Feo- 
dossieff, was appointed to carry out a series of tests on this plan. Pieces of 
rail 6 feet long were taken in pairs, one of which was tested at the 
natural temperature, the others being placed in a box filled with a mixture of 

• Communicated to the LondoD Meeting of the Iron and Steel Institute, 1879. 



>, Google 



286 STEEL RAIia 

two parts of broken ice and one part of salt, and, after being cooled to a 
temperature of from +3 degrees to -6 degrees Fahr., which occurred in half 
an hour, they were all submitted to the same tests. Altogether, 86 samples 
were tested, and these were, for the sake of comparison, divided into two 
groups, viz.: (1) Rails which broke under the test; and (2) tbUb which stood 
the test. 

The results indicated that the brittleness of the steel increases very much 
at low temperature if it contains more than a moderate amount of phosphorus, 
silicon, and carbon. The total of the three elements in the rails which broke 
under the test averages 0.54 per cent, and in those which stood the same test 
0.41 per cent, the first average (0.54 per cent) varying from 0.44 to 0.67 per 
cent, and the second average {0.41 per cent) varying from 0.37 to 0.55 per cent. 

The tests on steel at different temperatures made at the Watertown Arsenal 
in 1888,* showed within the range of 0.37 to 0.71 per cent carbon a slight mcrease 
in the elastic limit and traisile strength at d^:rees Fahr. above that at summer 
heat, accompanied by very little change in the elongation or contraction of area. 

Dr. P. H. Dudley recommends the use of basic open-hearth rails of 0.62 
to 0.75 carbon with phosphorus under 0.04 to insure a more uniform range of 
toughness and ductility of metal where exposed to low temperatures than has 
been obtained in plain Bessemer with 0.50 carbon and 0.10 phosphorus. 

Tests of Bessemer heats were made in which one-tenth of one per cent of 
metallic titanium was added to the steel and the carbon increased from 0.50 to 
a range of 0.60 or 0.70, the metal having higher physical properties and tough- 
ness at the same time. The manganese and silicon are lowered slightly to 
prevent shrinkage cavities in the ingots. 

Tests were made under the drop comparing the ordinary Bessemer rails 
and this grade of steel cooled to zero temperatures, and some specimens were 
also cooled to 22 degrees Fahr. below zero. The tests of the Bessemer steel 
with the one-tenth of one per cent of metallic titanium withstood a drop of 
2000 pounds, falling sixteen feet without breaking, while the plain Bessemer 
would fail at the low temperatures at about one-half that height. 

Dr. Dudley! conaders that basic open-hearth rails (C = 0.68; Mn = 
0.86; Si = 0.10; P = 0.012) which show ductility ranging from 15 to 20 per cent 
under the drop test are proper for use under high-speed trains where the tem- 
perature in winter falls to 20 degrees Fahr. below zero. 

• Tests ot Metala, 1888, House Doc. No. 45, 50th Congress, 2nd Sesa., p. 505. 
t Dudley on Ductility Teete on Rails. ProceedingB American Society for Testing Materials, 
Vol. X, 1910, p. 229. 



,y Google 



STRENGTH OF THE RAIL 287 

The question of obtaining a proper amount of ductility in rails, used in cold 
climates, is a very important one. The experience of the railroads in the 
Northern and Western States, during the very cold weather of the winter of 
1911-1912, was more severe than ever before reported. The effects were so 
great that rails which had heretofore been quite free from breakages were broken 
in considerable quantities by the wheels of the passing trains. The record of 
rails of large ductility or tenacity and toughness, on the other hand, showed a 
much greater freedom from breakages. 

With the recently adopted standard drop-testing machine (see Article 27) 
the ductility of the rail can be measured with much more accuracy than was 
possible in the machine which preceded it. With the older machines the re- 
bound of a 2,000'pound hammer was as large as 12 to 14 inches, while in the 
present machine it is confined to 3 or 4 inches. 

The maximum elongation per inch can be obtained by stamping the base, 
head, or edge of the base of the butt, as the case may be, before testing with a 
spacing bar of six inches, directly under the point of impact for either a sin^e 
blow or for two or more required to exhaust the ductility of the metal. These 
six inches include about two-thirds of the metal affected by the impact. 

The elongation In the base of a 6-inch, 100-pound rail, with a moment of 
inertia of 48.5, under a single blow for an 18-foot fall in the present drop-testing 
machine, will be from 6 to 7 per cent for a steel of 0.50 carbon, 0.10 phosphorus, 
and 1.00 manganese. The elongation of the metal under the drop-testing 
machine compares favorably with that obtained by static loads in the tension 
machine. The tendency is an increase of possibly one or more per cent, owing 
to the fact that the base of the rail in stretching does not neck as in the case of 
a tenale specimen. 

To raise the mean ductility, it has been found necessary to reduce the 
average percentage of carbon in heavy sections to 0.50, when the phosphorus 
content is 0.10, for rails which are to be used where the temperatures fall below 
zero. The Bessemer rails, owing to the greater content of phosphorus, oxides, 
and nitrogen, show a greater tendency to irregular low ranges of ductility than 
the basic open-hearth rails. The use of ferro-titanium in Bessemer steel, to 
take up a large percentage of the oxides and also a part of the nitrogen, makes 
it possible to increase the carbon content without any sacrifice to the ductility.* 

Probably some of the failures of rails in cold weather can be attributed to the 
effect, alr^y noted in article 24, of the contraction of the metal, which may set 

• The Bubject of ductility in rail steel has been reported upon very fully by Dr. P. H. Dudley in 
p^jera presented to the American Society tor Testing Materials, eee Proceedinga, Vol. X, 1910, pp. 
^3-232, &nd Vol. XI, 1911, pp. 454-461. 



,y Google 



288 STEEL EAILS 

up tensile stresses of some magnitude in the rails before the ends render in the 
splice bars. 

The effect of frost, in heaving the track where the ballast or subgrade con- 
tains much moisture, is to cause an irregular surface and, on account of some 
of the ties rising or "heaving" above the others, often produces larger bending 
moments in the rail than would be expected if the ties were free to adjust them- 
selves to the elastic curve of the rail. In the case of well-drained track, on stone 
ballast, where heaving is absent, smaller bending moments will be found in the 
frozen track than when in its natural condition. 

27. Physical Tests of the Strength of the Rail 
The impact hammer or drop test, introduced by Sandberg and Styffe, 
in 1868, is most generally used in this country in testing the strength of the rail. 
From the very prominent place given drop tests in rail specifications,* 
it might be seen that the behavior of the rail under the drop test is generally 
regarded as valuable information as to its character. As a matter of fact, 
however, engineers differ widely as to the advisability of accepting this test as 
an index to the reliability of the rail, on account of the great variation in the 
results obt^ned. The test shows, it is true, whether the piece being tested is 
brittle or not, and by observation of the permanent set, whether the steel 

Note. For an account of the state of knowledge relating to impact teate and (or a bibliography of 
literature, reference is made to the following: 

American Sectjon, International Aaaociation for Testing Materials: Bulletin No. S, October, 
1899. Report of Comtnitteo on Present State of Knowledge Concerning Impact Teats. W. K. 
Hatt and Edgar Marburg. 

Bibliography on Impact Tests and Impact Testing Machines. Proceedings American Sodety 
for Testing Materials, Volume II, page 2S3. W. K. Hatt and Edgar Marburg. 

The Resistance of Metals under Imp^t. Mansfield Merriman. Proceeding American 
Association for the Advancement of Science, Volume 43, ISd4. 

Theoty of Impact and its Application to Testing Materials. H. D. Tiemann. Journal of 
the Franklin Institute, October and November, 1909. 

International Association for Testing Materials, Vth Congress, Copenhagen, 1909, Impact 
tests papers. III,, llli. III., III., III., III>, III,, III>. 

Elongation and Ductility Tests of Rail Sections under the Manufacturers' Standard Drop- 
Testing Machine. P. H. Dudley, Proceedings American Society for Testing Materials, Vol. X (1910). 
p. 223. 

The Same. Iron Trade Review, Vol. 47, p. 410. 

Nouvelle Mothode d'essai des Rails, Ch. Fremont GSnie Civil, Vol. 59 (1911), p. 7, 26, 48, 72 

The same. Railway Age Cinzette, Vol. 51, p. 1176. 

New Types of Impact Testing Machines for Determining Fragility of Metab. T. Y. Olsen. 
Proceedings American Society for Testing Materials, Vol. XI (1911), p. 815. 

* Rome Results Showing the Behavior of Roils under the Drop Test, and Proposed New Form 
of Standard Drop Testing-Machine. S. S. Martin. Proceedings Am. Soc. tor Teat. Materials, 1908, 
Vol. VIII. 



,y Google 



STRENGTH Of THE RAIL 289 

is soft or hard. Recent work in Germany and France points to the conclusion 
that some form of impact test is fowid necessary to detect faults of structure 
that are not evidenced by the static test. 

Average specifications, which a majority of the railroads have in recent 
years used as a standard, contain the following clause as to drop test: 

One test shall be made on a piece of rail, not less than 4 feet, nor more than 6 feet, selected 
from each blow of steel. The test piece shall be taken from the top of the ingot. The ruls shall 
be placed head upwards on the supports, and the various sections shall be subjected to the follow- 
ing impact testa under free falling weight: 

70 to 79 pound rail, 18-foot drop. 

80 to 89 pound rail, 20-foot drop. 

90 to 100 pound r^l, 22-foot drop. 

If any ml breaks, when subjected to the drop test, two additional testa may be made of 
other rails from the same blow of steel, also taken from the top of the ingot, and if either of these 
latter rails fail, all the rails of that blow which they represent will be rejected; but if both these 
additional test pieces meet the requirements, all the rails of the blow which they represent will 



The drop-testing machine shall have a tup of 2000 pounds weight, the striking face of which 
shall have a radius of not more than 5 inches, and the test rail shall be placed head upward on 
solid supports 3 feet apart. The anvil block shall weigh at least 20,000 pounds, and the supports 
shall be part of or firmly secured to the anvil. The report of the drop test shall state the at- 
mospheric temperature at the time the test was made. 

These specifications, while used by the railroads, had to be modified ac- 
cording to the character of the drop-testing machines at the different mills. 
Thus, we find machines answering closely the following descriptions: 

1. A drop-test machine consisting of some concrete and loose stone, supporting a number 
of 12 by 12 inch oak ties, 12 feet long, on which is placed an oak block 18 inches by 18 inches by 
11 feet. On the oak block are two steel plates 1 by 18 inches by 7 feet, which become the bearings 
for the rail supports. These supports weigh 1300 pounds. 

2. A drop-test machine con^sting of a wooden foundation 4 feet deep and 10 by 10 feet, 
on vhich were placed two blooms, probably 8 by 10 inches by 10 feet. On the blooms arc placed 
the rul supports. 

3. A drop-test machine consisting of a concrete or stone foundation, on which rests a 20,000- 
pound an^il, to which the rail supports are securely fastened. 

Up to within the last few years no rail mill has been equipped with a drop- 
testing machine for rails that was built on thoroughly scientific principles, owing 
to the lack of proper foundations or proper anvil, as well as many other essential 
details. Further, no two rail mills had machines built on even comparatively 
the same lines. Consequently, any exact determination of the loss of energy 
of the falling weight which is dissipated by the machine would have had no 
general application, and the results obtained from testing rails in the drop- 
testing machines of any two mills were not comparable. 



,y Google 



5W0 STEEL RAILS 

The Manufacturers' Committee, recognizing these defects, prepared speci- 
fications and plans of a proposed standard drop-testing machine that will ^ve 
satisfactory and comparable results, and the Rail Committee of the American 
Railway Engineering Association has, with ceri:ain small modifications and 
additions, approved these plans and specifications. 

Specifications for Drop-Testing Machine 

<3es Fli. 204.) 

1. The tnachiDe shall be arranged to allow a 2000-poimd tup to fall freely at least 25 feet 
on the center of a rail resting on supporte that can be adjusted to epans varying from 3 feet to 
4 feet 6 inches. 

2. The anvil shall be a solid casting, weighing, with the attachmente that move with it, 
20,000 pounds. It shall be free to move vertically independently of the lead columns and shall 
be supported on 20 sprinp known as the standard " C " spring, without center coil, as employed 
by the Master Car Builders' Association (their figure 5614). This spring has a free length of 
$t inches, an outside diameter of 6iV inches, and is made froma bar having a diameter of lA 
inches. These spring are to be arranged in groups of five at each corner of the anvil and are to 
be held in place by hubs raised on the top of the base plate and by circular pockets on the under 
side of the anvil. Anvil to be guided in its vertical movement by removable finished wearing 
strips, these strips to t>e suitably attached to the finished edges of the column base. 

3. The base plate shall be of cast iron or cast steel, 8 inches thick in the area covered by 
the anvil. It shall be firmly secured to the substructure by four bolts 2 inches in diameter. 

4. The substructure shall consist of a timber grillage resting on a masonry foundation. The 
grillage shall project 9 inches beyond the ends of the base plate, and clear the columns at the side. 
It shall consist of one course of 12 inches by 12 inches sound oak or southern yellow pine, pref- 
erably creosoted, laid close and well bolted together. The masonry, preferably concrete, shall 
not be less than 5 feet deep below the grillage and be suitably supported on the subsoil. 

5. The pedestals for supporting the test rail shall be substantial castings, and the surface of 
the anvil between these pedestals shall be formed to receive a wooden block to absorb shock under 
broken pieces. The rail supports shall be removable pieces of steel securely held in the pedestals 
having an upper cylindrical bearing surface with a radius of 5 inches. The pedestals shall be 
adjustable to spans, varying from 3 feet minimum to 4 feet 6 inches maximum l>etween centers. 
They shall be securely held together and so fixed to the anvil as to insure that the center of span 
shall always coincide with the center between leads. 

6. The leads shall be firmly connected to the base plate and well braced. They shall be 
long enough to provide the prescribed free fall of the tup. They shall be provided with a con- 
venient ladder and a plainly marked gauge, divided into one-foot intervals. The zero of this gauge 
shall be 5| inches above the top of the rail support, and the specified height of drop shall be 
measured from this zero, irrespective of the height of the rail Ixing tested. One of the guides 
shall have a removable section, 6 feet long at the bottom, so that the tup or tripping block can 
be readily removed. 

7. The tup shall weigh, with the accessories that drop with it, 2000 pounds. The striking 
die shall be steel having a cylindrical striking face with radius of 5 inches and a length of 12 inches. 
The guide grooves shall have finished surfaces. The tripping head shall allow a grip of the tongs 
that will release at the exact height for which the tripping device is set and that will be safe from 
accidental release while the test piece is being shifted. 

8. The tongs and tripping device shall be arranged to release the tup automatically only; 
no manual releasing will be allowed. The tripping device shall be easily adjustable at one-foot 
intervals. 



>, Google 



Mil 



1-i 
iP 

I MS 
|I|S 

Ijfi 



iiii 

kn 

Ifil 1 



I* 



>, Google 



>, Google 



STRENGTH OF THE RAIL 291 

A diagram showing graphically the relation between results of tests as 
between the old and new methods, furnished by Mr. Thomas H. Johnson, 
is given in Fig. 205. 

These tests were made on rails of the same grade of steel, viz., carbon, 
about .50, and manganese, .90 to 1.00. All of the tests were made with a 



Fio. 205. — Diagram of Testa with Drop-Testing Macbioea of Old and New Design. (Johnson. ) 

2000-pound tup, with a radius of striking face 5 inches, and span between 
centers of supports 3 feet. Table LXXI gives the result of these tests. 

Since 1908 a number of machines of this design have been installed at 
various mills with satisfactory results. The introduction of a standard drop- 
testing machine has been of such benefit to the manufacturers and consumers 
of this country that the International Association for Testing Materials has pub- 
lished a description and cuts of the American Standard drop-testing machine in 
French, German, and English.* The seventh London Resolution of the Council 
was as follows: 

"That a standard drop-testing machine for rails be adopted in each country, as has already 
been done in the United States, in order to make tests comparative." 

• Proceedings, Vol. 11, No. 4, May 20, 1911. p. 237. 



,y Google 



292 



STEEL RAILS 



If we consider what occurs in dealing the blow on the rail with the falling 
weight, it will be seen that the work utilized by the rail in order to take a cor- 
responding set is not the potential energy of the weight at the moment when 
the weight touches the bar. The energy of the falling weight serves to deform 
the rail, to compress the supports, the body of the drop weight, the anvil, and 
the ground upon which the anvil is placed. 

TABLE LXXL- 





New Btudud MubiM. 


Old-Styl* Uuhi», 


HHghtol 
















Drop. 


P. R.R, 


P. S. 


P. R.B. 


P. 8. 


P. R.R. 


P.S, 


P. R.R. P.S. 




M-P«md 


SS-Pouad 


lOftPoimd 


lOO-Pwod 


il-PoBBd 


BS-Pomid 


loo-PouHi iaa-p«iDd 




SM. 


Set. 


Set. 


Set, 


Set. 


Set. 


Set. 


SM. 


F»t. 


iMbn. 


iKbo. 


lKb«. 


llKbH. 


Inebea. 


bobM. 


Iiicb«. licb«>. 


.5 


.02 


.02 


.02 


,02 


-0 


.0 


.0 





1 


.08 


.08 


.08 


.06 


.031 


.0 


.031 





1.5 


,16 


,17 


,14 


.15 


.093 


.062 


.062 


031 


2 


.25 


.22 


.19 


,19 


,126 


.093 


.125 


093 


3 


.38 


.37 


.28 


.25 


.250 


.171 


.218 


109 


4 


.49 


,47 


.40 


.38 


.343 


-250 


.281 


250 


5 


.60 


.53 


.50 


.45 


,406 


,343 


.375 


313 


6 


.76 


.66 


.64 


.62 


.513 


.468 


.437 


375 


7 


.88 


,76 


.74 


.70 


,700 


.600 


,650 


375 


8 


.90 


.90 


.86 


,81 


,800 


,700 


,700 


600 


9 


1.12 


1.01 


.98 


.90 


.900 


,800 


,750 


700 


10 


1.20 


1.13 


1,06 


1.02 


, 1.000 


,850 


,800 


800 



Breuil* has shown that for the same amount of actual work the bending 
curve of the impact test is the same as the slow-bending test, and Hattt has 
observed that there is little difference In the total elongation and unit rupture 
work whether the bar is ruptured in ten minutes or in from one to two one- 
hundredths of a second4 

* RelatioD between the Effect of StresMs elowlf applied and Stressea euddealy applied in th« 
Case of Iron and Steel. P. Breuil. 1904. 

t Teuaile Impact Teate of Metak. Hatt. Proceedings Am. Soc. for Test. Materia 1904, 
Vol. IV. 

t As further evidence we may quote the follawiiig opioiona (Report on Impact Tcflta of Metals. 
Official report by G. Charpy, Montlugon, International Association for Testing Materials, 5th Con* 
gress, Copenhagen, 1909. McGraw-Hill Book Company, New York): 

Captain Duguet writes: " The effect {of the duration of the stress) is very marked, especially 
during the period of great deformations; that would in itself suffice to render any too detailed 
investigation of the phenomena illusory. But we must not exaggerate the importance of this 
. point. The extreme deformations are, in the case of soft steel, submitted to bending, very sen- 
sibly the same, whether they be produced slowly by hydraulic pressure or bj' the impact of a tup." 

Id his treatise on material testing, Professor Martens points out that, according to the 
experiments of Kick, the velocity of fall has only an insignificant influence on the magnitude of 
the deformation in impact bending tests. As regards tensile strength tests. Professor Martens 
concludes: " From the impact tension tests so far conducted in the Charlottenburg Laboratory, 
I have acquired the conviction that the deformations are produced exactly as by slow tension. 



>, Google 



STRENGTH Of THE RAIL 2»3 

Fig. 206 shows the amount of energy dissipated in 90-pound A. R. A. type 
" B " Bessemer rails when tested in the drop-testing machine.* 

The weight of the tup was 2000 pounds and the distance between supports 
three feet in both the dynamic and static tests. The anvil in the drop test 
weighed 10 tons, spring supported. Calculating the work done on the rail in 
the static test from the load deflection diagram, and in the drop test from 
the height of drop and weight of tup, it would appear from the lower diagram 
of the figure that about two-thirds of the energy of the falling tup is utilized 
to deflect this rail. 

The difficulty of comparing the values of the stresses in impact tests with 
those which occur in static tests (i.e., where the momentum of the load does not 
factor) hes in the difficulty of accurately determining the value of the force 
acting between the hammer and the specimen in the former. Comparisons of 
the total work required to rupture a specimen, or to produce a ^ven deflection, 
are comparatively simple. 

The manner in which impact stresses are related to so-called static stresses 
requires careful theoretical con^deration before it can be clearly comprehended. 
The author is indebted to the work of Mr. H. D. Tiemann f for the following 
presentation of the subject. 

The rail tested by impact is in reality in the nature of a cushion between 
the two impacting bodies, namely, the tup A and the anvil B, and the anvil B 
must be of such proportions that its relative velocity v^ to that of the com- 
mon center of gravity of itself and the tup, V, may be disregarded, as other- 
wise a correction must necessarily be made, which not only complicates the 
subject, but on account of the nature of the foundation of the anvil is almost 
impossible to apply. 

In tension teats by means of Beveral blows, we find often that the elongation waa greater than 
in the slow-tension testa." 

M. Lebasteur (Annales des Ponts et Chaussees, 1890) had likewise arrived at the following 
cone lupous; 

" 1 . The elongations observed in the fracture of bars under high-drop impact testa are nearly 
identical with the elongations of similar bara under slow tension. 

" 2. The appearances of the broken sections are absolutely the same in the two cases. 

" 3. The total intensities of the blows necessary to break bars under high-drop impacts are 
proportional to the dynamic resistance to rupture (as determined from the area of the curve of 
slow tensile stress);" and further on M. Lebasteur added that "the dynamic re^tance to 
rupture measures the strengths of metals equally well for dynamic as for static stresses." 

• Report, M. H. Wickhorst to Rail Committee of Am. Ry. Eng. & M. of W. Assn., Proceedings, 
Vol. 12, Part 2, p. 389-394. 

t The Theory of Impact and its Applicatjon to Testing Materials, Journal of the Franklin In- 
stitute, October and November, 1909. 



>, Google 



STEEL RAILS 




soNood avoi 



s 










\ 


N, 


Z 

-1 








\ 












z^ 


s 








si 


^ 















\ 


\ 




1 


A 


O 


0. 



a: 


V 


1 




\ 


A 















0° f 



«l 



o 

SNOJ. iOOJ 



O ul O H) 

ia3J dOUQ JO 1H013H 



>, Google 



STRENGTH OF THE RAIL 296 

If we consider what occurs at impact, it is seen that at the beginning of 
contact a mutual repellant force F begins to come into play between A and B, 
which produces a change in the relative velocities (»„ - V) and (v^ - V), where 
r, = the velocity of A, 
V, = " " " B, 

V = " " " the mutual center of gravity of A 
and B along a line joining their 
respective centers of gravity. 
This force F, starting from at first contact and increasing to a maximum 
when both relative velocities of A and B are brought to 0, then when rebound 
begins again decreasing until it becomes at departure, is counterbalanced 
directly by the local compression of the material of both bodies at the point 
of contact. 

This force is made up of two parts, one being the elastic resistance of the 
body to compression or deformation, /, and the other of the force necessary to 
produce the local acceleration of the particles compressed, /t . 
F = / + /,. 
At any instant F is exactly equal to the change in momentum produced 
by its action divided by the time required to produce this change; 

_ TO (P - Pi) 



Or more exactly. 


„ mdv ^ dh 


where 




m= mass, 
8 = space, 
( = time, 
v= velocity 
o = acceleration. 



Consequently, if the time-velocity curve can be determined, the force F can 
be calculated for any instant. 

Let us examine first the relations between the various quantities of an 
impact test graphically, and then proceed to develop formulae of the time-space 
curve and of the interrelations of the values. 

Consider, first, a rail lying horizontally on the anvil B and supported 
freely at its ends, and let it be struck at the center by a falUng tup A. For 
simplicity let us consider the r^ as massless as well as weightless. 



,y Google 



296 



STEEL RAILS 



Take the velocities as relative to the anvil, as explained above, or assume 
the mass of the anvil so great that its motion may be taken as zero. Let 
the motion of the center of gravity of the tup A be plotted as in Fig. 207, with 
space as ordinates and time as abscissae. If the tup start at some point J, and 
fall freely the distance H before striking the rail, the curve JB will be a parab- 
ola, or if its velocity be uniform its motion would give the straight line 
JiB. Impact with the rail begins at B, and, assuming the rail as massless, 
the only resistance offered to the momentum of the tup will be the bending 
stress of the rail, which will be the force F. The motion curve then becomes 




Fia. 207. — Tinie.deflection Curve, Masalees Beam, withiD the Elastic limit. 

(Journal of the Franklin Institute. — 



BC. The resisting force, starting with at B, increases in proportion to the 
deflection of the rail until the maximum value is reached at C. (In this case, 
let the elastic limit of the rail be not exceeded.) It is this force F which 
overcomes the momentum of the tup A by producing a negative accelera- 
tion until the momentum is reduced to at C Rebound then begins. If the 
rail is perfectly elastic, the force F continuing to act will restore exactly the 
same amount of momentum to A, and in the same manner, but in the opposite 
direction that it had at the beginning of contact. The curve will be CD. At 
D departure takes place (the rail being considered as massless). 

The total mutual repellent force F acting on the rail, between the tup A 
and the anvil B, is at any instant equal to 

where 5 is the space traversed by A and v is its velocity at the instant under 



,y Google 



STRENGTH OF THE RAIL 297 

con^deration. If 5 is in feet, t in seconds, and A = (weight in pounds -^ accel- 
eration of gravity in feet), then F is given in pounds, by the last formula. 

If the force of gravity is to be considered, as well as the initial velocity of 
A at impact, then this formula should be written: 
F~W =^A = -A 

where W„ is the weight in pounds of the tup A. This could be avoided by 
having the tup move horizontally instead of falling vertically. 

Examining Fig. 207, it will be noted that the acceleration of the tup from 
J to B is equal to g and is produced by the uniform force W^. From B to 
D it is produced by the force W^- F and becomes negative as soon as the 
value of F begins to exceed W„. This is at the point of reverse curvature, since 
the acceleration 

is zero, and evidently occurs at the point b or the deflection which would be 
produced fay the static load W„. In horizontal motion the change would occur 
at contact, B. It should be remembered that while H', is a constant force, 
Fisa. variable, ranging from at £, to a maximum at C, and again to at D. 
Whenever 

becomes a maximum the force F becomes a maximum, and this evidently occurs 
at the sharpest part of the curve, which in this particular case is at C. 

The value of F at any instant may, therefore, be determined from the curve 

F = ^,XA + W,. 
Or 

If the force F of the impulse becomes sufficient to cause complete f^ure 
of the specimen, the conditions are those shown in Fig. 208. 

The first part of the curve JB is the same as before. The velocity or 
momentum of the tup is, in this case, not entirely overcome by the resistance 
F of the raU, so that at failure the tup retains a portion of its velocity as 
indicated by the tangent line DE,. If the tup works vertically in free fall, 
instead of horizontally, then the curve DE is again a parabola of free fall. In 
this case the force F becomes a maximum at some point C, when the curvature 
is sharpest, and must be determined from the curve by 

dt^' 
since there is no means of calculating it mathematically. 



,y Google 



298 STEEL RAILS 

The curve can be conveniently obtained by some mechanical device by 
means of which the falling tup makes a tracing on a uniformly revolving drum. 
When F is thus determined the maximum strength values may be calculated. 

To supplement the information furnished by the drop test engineers are turn- 
ing their attention to other means of testing the physical properties of the rail. 



Fig. 208. — Time-dcfleclion Curve, Beam Stressed be; 
Elastic Limit (Joum&l of the Franklin Inatitute. — Ti 

* The Baltimore and Ohio Railroad 
Company, in connection with its inves- 
tigations on rail, has been making use of 
the scleroscope (Fig. 209), an instrument 
for determining the degree of hardness of 
metals. 

t The principle of the scleroscope 
(Greek sclero = hardness) consists of drop- 
ping a small plunger hammer from a fixed 
height onto the surface of the material 
whose hardness is to be measiu"ed. This 
hammer after striking, by no other force 
than its own weight, rebounds to variable 
heights, depending on the hardness or 

amount of resistance to penetration offered Fig.209,— ScieroBcope. (Am. Ry. Eng. Assn.) 
by the metal tested. The rebound of the hammer is used to measure the 
hardness of the metal, and the scale shown on the glass tube is simply for 

• General Infonnation Concerning the Scleroscope and its Use on the Baltimore and Ohio Rail- 
road. A. W. Thompson. ProeeodinRs, Am. Ry. Eng. & M. of W. Assn.. 1910, Vol. 11, Part L 

f See also The Scleroscope, Albert F. Shore, p. 490, Proceedings American Society tor Testinjr 
Materials, Vol. X, 1910. 



,y Google 



STRENGTH OF THE RAIL 



299 



comparative purposes and has no direct numerical value. This scale has 140 
graduations, and a test of very hard steel has resulted in a rehound to the point 
marked 110, while soft brass results in a rebound to the point marked 12, and 
lead is about 2 per cent of hard steel. 

Figs. 210, 211, and 212 present examples of tests. The numbers in these 
figures indicate the degrees of hardness. 

Fig. 210 is an A. R. A. section of open-hearth rail, as rolled by the Bethle- 
hem Steel Company, and is a new section which has not been in the track. It 
will be noted that the hardness on the top of the head of the rail is practically 



O 




Fig. 210. — Scleroscope Tests on Open Hearth 
Rail (New.) (Am. Ry. Eng. Aasn.) 



- Sclero9coi>e Testa on Bessemer Rail. 

(Am. Ry. Eng. Assn.) 



the same as the steel in the section of the rail just below the surface. The center 
of the head appears to be the hardest, as well as a line through the center of 
the web and base. The upper comers of the head are comparatively soft, the 
ends of the base, however, being very much softer than any of the rest of the rail. 
Fig. 211 is a section of a Bessemer rml rolled at Buffalo in 1908. It is a 
crop from the top end of a top ral. Although the specimen was from the top of 
the ingot, there is a difference of but three points in the readings throughout the 
head. The section, where polished and etched, showed rather dimly marked 
segregation. The head and base when planed into and etched showed some 
dark streaks in the head and light streaks and fissures in the base. The top 
of the head for i-inch depth was sound. The experimenter says: " The section 
as a whole is more uniform than is usually to be found in top, middle, or bottom 
rail of a Bessemer ' ingot.' " 



,y Google 



300 



STEEL RAILS 



°i 1 1 




Pig. 212 shows the comparative hardness on different lines on experimental 
titanium rails. This test indicates a skin of soft metal across the top of the 
head, but as soon as this is penetrated the 
hardnres is reached, which compares favor- 
ably with any part of the whole section. 
* Experiments were made at the 
laboratories of McGill University on the 
value of the indentation test for steel rails 
in regard to essential qualities desired in 
service. The study of this method of 
testing was suggested by tests made on a 
large number of rail sections by the Chief 
En^neer of the Canadian Pacific Rmlway, 
a spherical punch .75 inch in diameter 
Fig, 212.— ScieroHcopeTestsonNewTitaoium being used, with a load of 100,000 pounds 
Rail. (Am. Ry. Eng. A«sn.) applied by an Emery testing machine for 

10 seconds after commencing the load, and the indentation was measiu-ed by 
an instrument reading to t^Vu inch. 

The tests, conducted by Mr. Dutcher, were on a set of bars of 2.5 by .75- 
inch section containing known percentages of carbon, which were verified by 
tests, and varied between .11 per cent and .96 per cent. The punches used 
(in addition to the foregoing) were a 60° cone, a 90° cone, and a paraboloid. 

The term " hardness factor " applied to the results was obtained by divid- 
ing the projected area of the indentation on the surface of the specimen into the 
load applied. It was found that the yield point (as determined by tensile 
tests) varies directly as the hardness factor. The percentage elongation curve 
is also fairly straight between 200,000 pounds hardness factor (.10 per cent 
carbon steel) and 450,000 pounds hardness factor (about .70 per cent 
carbon steel) ; and the percentage of carbon varies directly with the hardness 
factor up to about .90 per cent. 

t There have been several methods proposed to test the hardness of the 
metal by ball-pressure tests. In the Brinell testj a hard ball of steel is forced 

• Transactions o( the Canadian Society of Civil Enpneers, Dutcher, Vol. XXI, pp. 47-88. 

t Hardness Testa. Official report by Dr. techn. P. Ludivik, of Vienna. International AsRoda- 
tion tor Testing Materials, 5th Congress, Copenhagen, 1909. McGraw-Hill Book Company, New York, 
al-iO various technical papers on Hardness Testa, Proceeding,?, American Society tor Teatinn MotcrinH 
Vol, XI, Iflll, pp. 707-743, 

I Compare P. Ludwik, " Uber Hartebeetimmung mittelst der Brinnellschen Kugeldnickprobe 
und verwandter Eindruckverfahren," " Zeitachrift des osterr. Ingenieur und Architekt«n-Verein€«," 
19DT, Nr. II und 12 (Mr, 12, p, 205, extensive literature references). 



>, Google 



STRENGTH OF THE BAIL 301 

by quiet pressure into the material to be examined; the diameter of the spherical 
impression is determined, as a rule, with the aid of a special microscope, and the 
area of the cavity is calculated. The quotient of pressure (in kilograms) by the 
area (in millimeters ^) is Brinell's hardness number H. 

The cone-pressure test marks a transition from the ball-pressure methods 
to scratch methods. It is the outcome of efforts to simplify the Brinell test, with 
the further object of making the hardness number Independent of the load and 
of the dimensions of the impression. 

In the Amsler-Laffon instrument (Fig. 213) a cylindrical steel center punch, 
plane above, ground to a right-angled 
cone below, is vertically mounted in a 
casing of bronze, in which it is free to 
turn; it is balanced by a lateral spring. 
The displacement of the cone (with 
regard to the casing) is transferred to 
a pointer by an elastic threaded bush- 
ing and a toothed wheel; the pointer 
allows of easily reading depths up to 5 
millimeters within .01 millimeter. The 
pointer is accurately adjusted by turn- 
ing it with the aid of the milled edge 
of the bushing, in case the top of the 
specimen should not be perfectly plane. 
The cone can easily be exchanged and 
beregromd. The whole instrament f„. 2,3, _ A™ie,-L.l.o„ In..™.™t for 

weighs .7 kilogram (1^ pounds), and Measuring Hardneas. 

its height, from the upper pressure plate to the surface of the specimen, is 
about 10 centimeters (4 inches). 

The question, whether the hardness numbers of a material, obtained by 
these methods, admit of any general conclusions respecting the strength of 
the material, and in particular the yield point and the tensile strength, is of 
high practical interest. 

A direct constant relation between yield point and tensile strength on the 
one side, and hardness on the other, can not exist, since that relation would, 
among other factors, depend upon the shape of the impression and of the stress- 
strain diagram. 

This admission does not, however, at all exclude the possibility of deducing 
from the hardness number, with the ^d of a coefficient which will only hold 



,y Google 



302 STEEL RAII5 

good between certmn known limits, the yield point and the tenale strength with 
an approximation which will frequently be sufficient for practical purposes.* 

t Permanent-way materials have been tested by Ludwik's cone-pressure 
method with two objects: to inquire into the suitability of the method for practi- 
cal purposes, and to ascertain the relation between the cone-pressure hardness 
and the tensile strength. 

The experiments have been conducted in connection with the acceptance 
tests of the materials supplied to the J. R. Austrian State Railways during the 
year 1908, in the Trzynietz Iron Works of the Osterr. Berg- und Huttenwerks- 
Gesellschaft with the aid of an Amsler-Laffon cone-pressure hardness testerj 
on a Mohr and Federhaff testing machine. 

The material experimented upon comprised rails, railway ties, splices, and 
steel crossings. 

The specimens were not prepared in any way apart from being cleaned ; an 
exception was made in the case of the steel points, in which the outer skin con- 
taining coarse impurities had to be removed completely. 

The following are the chief results: 

The ratio of tensile strength to cone-pressure hardness had a mean value 
of about .335, the range of deviation being ±6 per cent. 

The lowest tensile strength of 65 kg./mm.* (42 tons per square inch), ad- 
missible for rails, corresponded to a cone-pressure hardness of about 190. 

The tensile strengths of ties and of smaller parts for the permanent way 
varied between 39 and 47 kg./mm.= (24.75 and 29.8 tons per square inch) and 
the corresponding hardness numbers between 117 and 144. The range of 
variation is hence approximately the same for the tensile strengths as for the 
hardness numbers. 

Othei" methods of testing hardness have been used. The sclerometer of 
Turner makes use of a diamond point which is drawn across the surface to be 
measured. The weight required upon this point to make a barely visible scratch 
determines the degree of hardness. This machine is sometimes used with a 
series of standard weights and the width of a scratch made by one of these 

• For instance, the Pnissian Railway Department stipulatt^ for rails of a minimum tensile strength 
ot60kg,/mm."(38tonsperaqiiareineh), with halb of 19 mm. (Jinch) diameter and .50 tone loads, impres- 
sion depths of from 3,5 to ."(.S mm,; for rails of a minimum tenaile atrenRlliof 70kg./mm.' (44.5tonBper 
equare inch), impression depths ranKinR from 3 to 5 mm. Breaking tentaand ball tests have to be 
made in equal numbers. (Zentralhlatt der Buuver-waltunfr, 1908, No. 77, p. 520.) 

t The Cone-pressure Test for Determining the Hardness of Permanent-way Materials, l)y Dr. 
techn. August Ge6ner, Vienna. International Association for Tesling Materials, 5th Congrees, Copen- 
h^cn, 1909. Mpr.ran--Hill Book Company, New York. 

t Compare P. Ludwik, " Die Kegelprobe, an neues Verfahren lur Hortebestimmung von Mat«i- 
alien." Berlin, 1908, Julius Springer. 



,y Google 



STRENGTH OF THE RAIL 3(» 

measured under a microscope. The Keep test employs an instrument which 
drills into the specimen and pves a measure of the work required to cut out 
the metal, thus testing not only the surface, but also the interior. The Jaggar 
instrument is similar, but uses a small diamond drill in connection with a micro- 
scope. 

Resistance to penetration was long tested by the United States Ordnance 
Department by means of a weighted punch, and a somewhat similar result was 
obtained by means of a series of needles of graded hardness, which were tried 
in succession until one was found that would scratch the material under test. 

While many inconsistencies are found in hardness tests, it is generally con- 
ceded that the test affords an excellent comparison of metals of the same general 
composition and treatment, and the results thus far seem to justify the expecta- 
tion that it will in many cases be possible and advantageous to employ this 
method in the testing of rails in place of the more elaborate and expensive tensile 
strength tests which some foreign engineers require in addition to the drop test. 

The mimetic laboratory of the Bureau of Standards is carrying on an investi- 
gation <m the relation between the magnetic and mechanical properties of steels. 
The reluctance, or the ratio of the magnetizing force to the nu^etic uiduction, of a 
rail is greatly affected by changes in homogeneity, such as may be caused by segre- 
gation, blowholes, or strains due to any cause whatever. By means of the mag- 
netic data taken along tile length of a r^ it is possible to detect the presence of 
these defects. 

Special machines have been devised from time to time for testing different 
properties of the rail, as the machine for testing rail wear illustrated in Fig. 151. 

* The Pennsylvania Steel Company has a machine (Fig. 214) for testing 
rail wear which enables specimens of rails to undergo wear similar to that im- 
posed upon them by every description of traffic, but in a much shorter time than 
if tried in the ordinary road. The rails are fixed to a 20-foot diameter circular 
frame, three specimens being included in the circle. Two standard 33-inch 
wheels having independent axles fixed at each end of a heavy casting, which is 
pivoted at the center of the circle, run upon the rails at speeds up to 85 revolu- 
tions per minute, equivalent to a train speed of about 60 miles per hour. There 
are devices by which, through means of springs, pressure is exerted vertically and 
also centrifugally on the rail, so that the action of the load can be imitated, as 
well as that of its lateral pressure on the rail, and the effect produced by continu- 
ous wear on the rails of different manufacture and composition can be estimated 
in a comparatively short time. 

* R^lway and Engineering Review, Chicago, 1908, Vol. XLVIII, p. 868. 



,y Google 



STEEL RAILS 



Fia. 214. — Machine for Testii^E Rail Wear at Peniuylvania Steei Company. 

Extensive tests have been 
made of the tensile strength of 
the steel in the rail by Mr. M. H . 
Wickhorst, Engineer of Tests 
of the Rail Committee of the 
American Railway Engineering 
Association, covering rails rolled 
at Gary and at the Lackawanna 
Steel Company. 

The rails from the Gary 
works were open-hearth steel 
and 100-pound, A. R. A., Type 
B section. The ingots furnished 
six rails, which were lettered 
A, B, C, D, E, F, the A rails 
coming from the top of the 
ingot, etc. 

Tensile tests were made of 
pieces J-inch diameter by 2-inch 
gauge length, cut from near the 
top end of each rail, as shown 

-DiagramofBoundTe8tPiecea;TenflUeT«t«on '" ^S" ^l^- Five locations in 

Rail Steel. (Wickhorst.) the Sections were selected as 




,y Google 



STRENGTH OF THE RAIL 306 

shown, and t«sts were made in duplicate, the bar being cut sufficiently long to 
make two test pieces. The yield point was determined by means of a Capp's 
multiplying divider, which method ^ves a result somewhat above the elastic 
limit, but which, however, is probably sufficiently definite to make it desirable 
to determine it, and is not subject to the irregularities of the yield jwint as 
determined by the drop of the beam of the test machine, or even by ordinary 
dividers. 

The test pieces were very close to J-inch diameter at the center, but toward 
the ends of the gauge length most of them were from .002 to .004 inch larger in 
diameter. This would tend to make the elongation less than if the diameter 
were perfectly uniform. 

The results of the tensile tests are shown in Table LXXII. 

The duplicates agree well with each other except in a few cases where the 
test pieces broke " short " as follows; One sample from base of the A rail, one 
sample from the interior of the head of the B rail, and one sample from the 
comer of the head of the C rail. One sample from the web of the A rail should 
probably also be classed here. The duplicates from the lower r^ls of the ingot 
agree particularly well, indicating a freedom from local irregularities. The 
ratio of the yield point to the tensile strength averages about 51 per cent, and 
most specimens differ but little from this figure. A comparison of the tensile 
strength is interesting. Table LXXIII shows the tensile strength of the 
sample in each pair that gave the greatest ductility. 

The b samples from the interior of the head and the c sample from the 
web represent what was ori^nally the interior of the ingot, and in the A rail 
these samples show strengths much higher than the samples from the other 
parts of the section representing what was the outer part of the ingot. This is 
also true of the B rail, but to a less extent, and also of the C rail to a slight extent. 
As we continue down the ingot, however, conditions are reversed, and we find 
in the D rail a little lower strength in the samples from the interior than in the 
samples representing what was the outer portion of the ingot. This difference 
is greater in the E r^I and greatest in the F rail. 

The a sample from the comer of the head and the d sample from the 
flange would be metal of similar chemistry, but the flange has a considerably 
lower finishing temperature and is also reduced differently. Table LXXIV 
compares results from these two places. 

The d sample from the flange of the A rail is abnormal, with a low strength 
and high ductility, being evidently taken at a point of negative segregation of 
carbon; but, except for this sample, the d samples from the flange show a little 



,y Google 



STEEL RAII^ 



TABLE LXXII.- 

Tssdlt TwU I 



d-head, comer. . . 
6-heftd, interior. . 

d-flange 

e-base 

o-head, comer. . . 
b-head, iDterior.. 

d-ftange 

a-he«d, comer.. . 

&-head, i 



d-fltu^ 

&-head, interior. . 
(j-flange 

6-head, interior. . 
d-Rangfi 

6-head, interior. . 
I d-flange 



TESTS ON STRENGTH OF RAIL STEEL 

» Oi» Hearth Rul StMl Inm Guy (Widihont) 



61,640 
69,510 
70,020 



66,970 
68,220 
70,310 
70,010 
66,970 
65,690 
64,940 
63,930 
67,710 
65,690 
66,970 
69,230 
66,970 
66,230 
66,490 



68,720 
64,940 
65,930 



64,180 
64,940 

66,460 
67,440 

68,220 
68,220 
69,510 
65,210 



64,180 
70,020 
67,710 
63,670 



Tsuila Stmclh. 


lUtio. 


PoMd-pmSquu, 




124,800 


5.1 


125,800 


48 


134,400 




133,900 


52 






141,000 


52 


121,700 


54 


123.300 


54 


109,600 




120,200 


47 


128.800 




128.300 


51 


135,900 


50 


137,000 


61 


137.000 


51 


132,900 


50 


129,800 


51 


129.400 


50 


126,800 


50 


130,900 


52 


132,900 


50 


133,400 


52 


133,900 


50 




50 


132,900 


50 






128.300 


50 


128,800 


50 


132,400 


51 


130,900 


52 


129,300 


50 


128,300 


52 


130,800 


51 


131.900 


51 


133.900 


52 


134,400 


52 


130.900 


50 


132.400 


52 


127.900 


52 






124,800 


51 


124,300 


52 


128.900 


50 


128,900 


52 


131.800 


51 


132,400 


52 


129.800 


53 


130,400 


53 


128,400 


51 


128,300 


50 


120.800 




123,700 


54 


126,300 








132.400 


53 


131,900 


51 


127,300 


50 


128,300 


51 



11.5 
12.5 
12,5 

10-5 



,y Google 



STRENGTH OF THE RAIL 



Rail. 


s-Hwl. Comer. 


b-Hivl, InUrior. 


fW*. 


i-Fhu^. 


.B„. 


A 

B 

c 

D 
E 
F 


124,800 

128,800 
130,900 

132,400 
127.900 
128,300 


133,900 

136,900 
133,400 
128,300 

124,300 
120,800 


141.000 
137,000 

133,900 
131,900 
128,900 
125,300 


121,700 

129,800 
133,400 
133,900 

131,800 
131,900 


120,200 
129,400 

128,800 
132,400 
129,800 
128,300 


Average 


128,800 


129,400 


133,000 


130,400 


128,200 



Co 


TABLE LXXIV.- 

mparison oF Scrwulh ud Ductility d 3lei 


TESTS ON STRENGTH OF RAIL STEEL 

t Uk« rroDi tbe Corner o[ the Head ud Fbnca of Opw Btmnh Raib (WiokbontJ 




,.,. 


Tenaila Streoglh poundi per 




RedtHtioiiolAitt. 




- 


^ 


• 


- 


• 


- 




124,800 
128,800 

130,900 
132,400 

127,900 
128,300 


121,700 
129,800 
133,400 
133,900 
131,800 
131,900 


Ptrctot. 
10 

11.5 

11.5 

12 

11 

12 


Percent. 

14 

12 

12.5 

13 

12 

12.5 


Percent. 

15 
20 

22 
19 
19 

20 


Per ctot. 






C 

D 


27 
28 








Average BtoF.. 






129,700 


132,200 


11,6 


12.4 


20 


25 



higher strength and also a Httle greater ductility. As the difference in the 
work of rolling is perhaps sufficient to account for this, the conclusion seems to 
be that the difference in roUing temperature has not resulted in any important 
difference in the tensile properties. 

The 6 sample from the Interior of the head and the c sample from the web 
would be of similar chemistry, as representing the interior of the ingot, but the 
web is thinner and gets more work in rolhng and probably is finished at a lower 
temperature. A comparison of these two locations is shown in Table UCXV. 



TABLE LXXV.— TESTS ON 

CorapBitaon ol Btrmsth and Ductility o( Steal talcBB from th. 


STRENGTH ON RAIL 

InieHor of the H«ul and Web ol 


STEEL 

Opea Heartb Rails (Wiclibonl) 




■""*"■ ii[ssi'i«r'*"*' 




Reduction of Area, 




b 


' 


" 




b 


' 




133,900 
135,900 
133,400 
128,300 
124,300 
120.800 


141.000 

137,q00 
133,900 
131,900 
128,900 
125.300 


Percent. 
9 
9 

10 

11 
11 

12 


Per ant. 
9 
11 

11,5 
11.5 
12,6 
13 


Percent, 
16 
16 
17 
21 
21 
22 


Per cat. 
10 




19 








23 






F 


25 






Average BtoF.. 


128,500 


131,400 


10,6 


11,9 


19.4 


22,6 



>, Google 



308 STEEL RAIia 

In the A rail it is probable that the carbon is higher in the web sample than 
in the head sample, but in the other rails there probably are no great differences, 
and the averages shown above are of the B to f r^s inclusive. The tensile 
strength decreases as we go down the ingot and the ductility increases. The 
web samples, as compared with the head samples, show a Mttle greater strength, 
an average of 131,400 pounds, as against 128,500 pounds, and also a little greater 
ductility, an average elongation of 11.9 per cent, as against 10.6 per cent, and 
a reduction of area of 22.6 per cent, as against 19.4 per cent. This difference, 

it would seem, is probably due to 
the increased work in rolling that 
the web gets. It is also interest- 
ing to note that the top rails show 
as good ductility in the head 
samples as the lower rails, allow- 
ing, of course, for the difference in 
tensile strength, which would 
make about 3 per cent decrease 
in elongation for an increase in 
tensile strength of 10,000 pounds. 
The tensile tests on the rail 
steel from the Lackawanna mill 
were from titanium Bessemer 
rails, 90-pound, A. R. A., Type B. 
It was planned to make ten- 
sile tests of pieces from near the 
top ends of each ndl by cutting 

Fia. 216. — Test Pieces 16 inchcH long. Diagram ot Flat flveflat pieceS | by IJ by 16incheS 
Test Pieces. Tensile Testa on Rail Steel. (Wickhorst.) f^om each rail aS shown in Fig. 

216. The test pieces were cut in this manner from the A and D rails, but 
it soon appeared that the time required to prepare the pieces in this manner 
would cause consido'able delays, and then, too, the surfaces are apt to be finished 
in a condition unsatisfactory for test. The plan was then changed so as to 
obtain pieces | inch in diameter and 2-inch gauge length, as shown by Pig. 215. 
The pieces from the B and C rails were prepared in this manner. The results 
of the tensile tests are shown in Table LXXVI. Care must be taken not to 
compare the results of the flat test pieces with those from the round test pieces, 
except, perhaps, as regards the tensile strength, although even here the shape 
of the test piece would make some difference. 




,y Google 



STRENGTH OF THE RAIL 



Rail. 


LoaliMi. 


KindofTen 
Pieoe. 


Taisile Stnmth. 


Eloncstini. 








Flat 
Round 

Flat 


Far Square Inoh. 
103,400 
111,100 
109,200 
103,300 
105.000 
111,700 
122,900 
116,200 
110,700 
111,600 
109,100 
112,100 
112,300 
109,600 
112,400 
109,800 
106,800 
104,700 
108,600 
110,000 


15.5 in 8 inches 

20,0 "8 '■ 
14.5 "8 " 
11.7 "8 " 
13.7 "8 " 
19.0 "2 " 
14.0 "2 " 
17.5 "2 " 
19.0 "2 " 
18.0 "2 " 
20,0 "2 " 
16.0 "2 " 
21.0 "2 " 
20.0 "2 " 
19,5 "2 
12.5 '8 •' 
10.0 "8 " 
13.7 "8 
10,0 "8 " 
9.5 "8 " 


23 flat 




b-head 


A \ 


























6-head 
























a-head 


36 •■■ 


c 




37 '■ 
















a-head 










D 


e-web 


23 " 















The following is a re- 
cord of tensile teste made 
by Waterhouse* on 100- 
pound A. S. C. E. r^l taken 
from stock. The steel was 
made by the acid Bessemer 
process and the ingots rolled 
without reheating. From 
this rail two adjacent pieces 
8 J inches long were cut with 
a slow-speed cold saw, and 
from these pieces tensile- 
test bars were machined in 
the positions shown in Fig. 
217. The results of the 
tensile tests are given in the 
following table, the figures 
being the average of those 
obtained from the duplicate 
pieces: 




Fio. 217. — Location and Numbera o( Test Pieces used in 
Waterhouse'B Tests. (Railroad Age Gawtte.) 



p. 478. 



• Examination of 100-pound RaUs. G. B. Waterhouee. RMlroad Age Gawtte, July 10, 1908, 



,y Google 



STEEL RAllS 



TABLE LXXVII.— TE8TS QN STRENGTH OF RAIL 

Twuile T«u od B«a«nsr Rail Steal (WaUrhouH) 





Elaatic limit. 


Ultimsls Sums. 


•^z- 


^^o, 




P«ilid& 

52,200 
52,200 
54,460 
65,000 
63,100 
63,820 
61,740 
53,340 


Ponndt 

108,400 
109,850 
110,750 
110,160 
110,300 
110,400 
109,850 
111,300 


FaroBC. 

16,75 
16.25 
18.50 
18.50 
18.25 
18.00 
18.26 
17.00 


Pa ant. 

29.9 
28.4 
33.2 
28.6 
29.4 
31.0 
35.4 
36.4 










I :::::.:::::: 









Table LXXVIII shows the chemical composition of the rails tested in the 
three tests just mentioned. 

TABLE LXXVIIL — CHEMICAL COMPOSITION OF RAIL STEEL 
IN TENSILE TESTS OF WICKHORST AND WATERHOUSE 





Wickhont. 


W«Wrt«™e. 


Cry. 


Lack.™.™. 




0.72' 

0.20 
0.72 
0.036 
0.036 


Per cent, 

0.48 

o.u 

0.90 
0,040 
0.091 


Parewrt. 

0,51 
0.147 
0,77 
0,078 
0089 
0,186 






Sulphur 






. 







28. The Strength of the Rail and Proper Weight for Various 
Conditions of Loading 

In determining the proper stress to use for the rail, careful consideration 
must be given to the exact meaning of the terms by which the strength of the 
steel is shown. 

* The dastic limit or yield point may be properly called the limit of pro- 
portionality of stress to deformation, or more briefly, the limit of proportion- 
ality. The limit of proportionality is sometimes called the " true " elastic 
limit, and is frequently regarded as a measure of the load-carrying capacity of 
a member or structure. 

The absolute value of this limit cannot, in general, be determined even by 
the most careful measurements of deformation and load. It has been the 
experience of experimenters that any additional refinement of measurements 
in stress-deformation tests results in the detection of the limit of proportionality 

* Proceedings Americao Society for Testing Materiitk, 1910, Vol, X, pp. 243, 244. Moore. 



>, Google 



STRENGTH OF THE RAIL 311 

at a stress lower than that determined by the less refined methods of measure- 
ment. Perhaps the thermo-electric apparatus used by Turner * and Rasch f for 
measuring deformation is the most delicate yet employed, and both of these 
experimenters showed the existence of a limit of proportionality at stresses 
far lower than those determined by extensometer measurements as ordinarily 
made. 

Members and structures become more and more nearly perfectly elastic 
if subjected to repeated stresses in the same direction, even if these stresses 
are so far beyond the limit of proportionality 
that there is a small but well-defined per- 
manent set upon release of the load. Fig. 218 
shows the result of loading a beam several 
times to a stress well beyond the elastic limit. 
The first load applied to the beam produced 
considerable permanent set, and during the 
application and release of load considerable 
energy was lost, presumably in heat. This 
energy is shown by the shaded area to the 
left of the figure. During the second cycle 
of loading and release much less energy was 
lost, as shown by the central shaded area; and 
during the third cycle the beam behaved as if 

almost perfectly elastic. It should be noted ona Divi»ion.o.iin,Defiectioo 

that the above results would not be obtained *""' 218.- Effect of Repeated ixjads 

on Beams. (Am. Boo. for Testing 

if the direction of the load were reversed. Matenaia — Moore.) 

The ultimate strength, or maximum capacity for resisting stress, has a 
ratio to the maximum stress due to the working load, which, although less in 
metal than in wooden or stone structures, is, nevertheless, made of considerable 
magnitude in many cases. It is much greater under moving than under steady 
" dead " loads, and varies with the character of the material used. For 
machinery it is usually 6 or 8; for structures erected by the civil engineer, 
from 5 to 6.J 

In general, parts of structures are so proportioned as to carry their loads 
without risk of exceeding their elastic limits; and in such cases the factor of 
safety should probably always be referred to the elastic limit. 

• Thermo Electric Measurements of Stress, Transactions American Society of Civil En^neere, 
Vol. XXVIII, January, 1902, p. 27. 

t Proceedings International Aseociation for Testing Materials, No. 11, August 4, 1909, Article VII. 
t Iron and Steel, Materials of Engineering, Thurston, 1909, p. 340. 



>, Google 



312 STEEL RAILS 

The elastic limit is made the basis of estimates by nearly all French engineers, 
while the ultimate strength is taken by German engineers, using a factor of 
safety of larger magnitude. British and American engineers usually base all 
calculations on the ultimate strength, although the former system is extending 
in general practice, and the limit of working load is made to fall well within the 
limit of elasticity. 

The graieral practice at the present time, for railway and highway bridges, is 
to use a unit strain of about one-half the elastic limit of the material. This factor 
is considered correct in places where the load assumed is an absolute maximum, 
as, for instance, where it consists of a definitely determined dead load only. 

In the rail the maximum stress acts during a very short space of time and 
its effect is not the same as the same load applied for a longer period. It is 
possible to apply a much greater stress than the elastic limit of the metal, pro- 
vided the stress be applied very quickly and then released. 

The excellent service ^ven by some of the rails of lighter section, exposed 
to heavy wheel loads, gives evidence that a limited number of excessive stresses 
in the rail will not cause injury when applied for the small fraction of a second, 
as is the case of the stress caused by the wheel load. 

An extreme fiber stress of 20,000 pounds per square inch as applied to 
the base of the rail probably represents a satisfactory mean between the danger, 
on the one hand, of not providing a sufficient margin of safety for the unknown 
quantities of the problem and the liability, on the other hand, of taking too 
great a factor of safety, and thus designing an uneconomical structure. 

The following remarks of Professor Unwin are pertinent to this question: 

" If an engineer builds a structure which breaks, that is a mischief, but one of a limited and 
isolated kind, and the accident itself forces him to avoid a repetition of the blunder. But an engineer 
who from deficiency of scientific knowledge builds structures which don't break down, but which 
stand, and in which the material ia clumsily wasted, commits blunders of a most insidious kind." 

Any consideration of the strength of the rail should take account not only 
of the stresses in the rail itself and the ability of the material of which it is com- 
posed to resist them, but a proper proportion must be made of the rail in order 
that it may distribute the wheel load to the ties in such a manner as not to over- 
strain any part of the track structure. The fact that a rail will not break should 
not be the determining factor in its selection, if, on account of lack of stiffness, 
it will allow too great a proportion of the load to come on a tie. 

The damaging effects of overloading the track, while much less apparent 
than the results attending the use of too great loads in other engineering struc- 
tures, are, nevertheless, of very real importance, and the lack of proper appre- 
ciation of the fundamental principles underlying its design has brought about 



,y Google 



STRENGTH OF THE RAIL 313 

conditions requiring excessive maintenance charges to keep the track in proper 
condition. 

As the real function of the heavier rails is to distribute the wheel load and 
prevent too great a concentration of pressure on the track substructure, we 
have two limiting conditions to consider: First, the rail should be stiff enough 
to enable it to transfer the load in such a manner as not to exceed the maximum 
bearing power of the track substructure, and second, the safe working stress 
of the metal in the rail must not be exceeded. 

Before investigating the proper weight of rail to use with any track struc- 
ture, the weight and types of the locomotives and cars to be run over it should 
be considered, and the maximum wheel pressures ascertained for each type. 

The bearing power of the roadway or subgrade should next be examined. 
The influence of the character of the roadway is well shown by the follow- 
ing case reported by Mr. A. G. Wells, General Manager of the Atchison, Topeka 
and Santa Fe:* 

" From Seligman to Barstow our track is laid with eighty-five-pound rwla; the density of 
the traffic is practically the same over every foot of it. Between Yucca and Barstow, a. distance 
of 227 miles, the subgrade is sandy, porous, and well drained; between Yucca and Seligman, a 
distance of 91 miles, the subgrade is largely clay, of a kind that holds water. From November, 
1907, to October, 1908, we had eighty-three rail breakages on the territory first named, or a per- 
centage of .001; on the other stretch we had in the same period seventy-two breakages, the per- 
centage being .0025, or, in other words, where the subgrade was dense and more or less clay, the 
breakages per mile were two and one-half times greater than where the subgrade was sandy." 

The bearing power of the subgrade is such an important factor in pro- 
portioning the track that it will prove profitable to examine what takes place 
when the soil is subjected to pressure.f 

As in any structure, good judgment must enter into the design; the formulas 
which will be demonstrated mxist be used as guides only. These formulae will 
depend upon the angle of repose of a homogeneous granular mass. For 
ordinary earths for which the angles of repose are known, the results obtained 
by the use of the formulae will compare very favorably with those obtdned 
from examples of the best practice. 

When the angle of repose is not known it should be determined by test-J 

* Railroad Age Gazette, April 9, 1909. 

t The following discussion is based upon Retaining Walls for Earth, Howe, New York, 1896. 
X This can conveniently be done by measuring the force required to cause slipping of two por- 
tions of the earth past each other when subjected to a known pressure, and 

where ^ — angle of repose. 

3 = force required to cause slipping. 
p = pressure on earth. 



>, Google 



314 



STEEL RAILS 



Earth which has an angle of repose of at least 27 degrees may be considered 
as firm. From Table LXXIX it is seen that sand, gravel, and damp clay are 
classed as firm soils; however, these may become so saturated with water that 
their angles of repose will become considerably less than 27 degrees, hence 
precaution must be taken against too much water by draining the ground in the 
immediate vicinity of the roadbed. Particular care must be taken in the case 
of clay, or sand which will become a kind of quicksand when saturated with 
water. 

The water which destroys the bearing power of such soils may come from 
below by capillary attraction,* and the drainage should be carried to a depth 
sufficient to prevent this. Semi-fluid soils, such as quicksand, alluvium, etc., 
should be removed where practicable or the foundation carried to a lower 
stratum. 

TABLE LXXIX. - 



Earth o 
Earth o 
Earth o 
Earth o 
Earth o 



earth, dry sand, clay, and mixed earth. . 
I earth, damp clay 

earth, wet clay 

. earth, shingle, and gravel 




Let Fig. 219 represent a section of the track, and 



"T 



'777777777777777777777m 

SUB GRADE 

Fic. 219. — Resistance of Sub-grade to Pressure of the Track. 

I = the depth of ballast; 

p = the maximum supporting power per square foot of thesubgrade; 

pi the pressure exerted on subgrade midway between ties; 

1 = the weight of one cubic foot of ballast; 
^ ^ the angle of repose of subgrade ; 
xy equals the vertical intensity of the pressure caused by the waght of the 

• Movements of Ground Water, by F- H. King and C. S. Slichter, Government Printing OERce, 
Washington, D. C, 1899, p. 65. 



,y Google 



STRENGTH OF THE RAIL 



315 



ballast on the subgrade midway between the ties. This pressure is augmented 
by the pressure transmitted from the tie, and, while this is much less between 
the ties than immediately underneath a tie, it is, nevertheless, an important 
factor in strengthening the surface of the roadbed. 

If we assume this extra pressure on the roadbed midway between the ties 
to equal in amount tx we will probably be on the safe side and can then write 
Pi = 2 yx.* 

Now when the ballast is about to sink 

p 1 + sin ^ 1 — sin 

*^ = , . -■ or g = p \ . ■ ■ 
q 1 - sm 1 + sin 

But when the roadbed under the tie is on the point of sinking, the part of 

the roadbed between the ties must be on the point of rising, or 

^ _ 1 + sin i» 

Pi 1 - sin * ' 

and the supporting power of the subgrade, or 

j 1 + sin ij, ]* 

n- 

\\ + sin 0P 



p=Pi 



; =2yx 



For convenience the values 
i in Table LXXXI. 



1 - an *1 



' sm 4>) 
are g^ven in Table LXXX and for 



TABLE LXXX 

Values of jli^* I 





ll+_-in*j'. 










' 















1.00 


23 


5.21 


5 


1.42 


24 


5.62 


e 


1.52 


25 


6.07 


7 




26 


6 56 


8 


1>6 


27 


7.09 


9 


1.88 


28 


7.67 


10 


2.02 


20 


8,30 


11 


2.16 


30 


9.00 


12 


2.32 


31 


9.76 


13 


2.50 


32 


10. S9 


U 


2.68 


33 


11,50 


15 


2.88 


34 


12 61 


16 


3,10 


35 


13.62 


17 


3.33 


36 


14.84 


18 


3 69 


37 


16.18 


19 


3.86 




17.67 


20 


4,22 




19.64 


21 


4.48 


40 


21.16 


22 


4.83 







• This apparently would be a safe aMumption for a depth of gravel ballast under the tie of 18 
inchee and 12 inches of stone. For a lees depth of ballast the pressure would be lees and for a greater 
depth the pressure would increafle, the increaae being more rapid in the case of tha stone than of th« 
gravel ballast. 



>, Google 



STEEL RAtLS 
TABLE LXXXr. —VALUES OF ■ 



N.maDlB.llut. 


&£i5i£'Hr 




90 to 106 
90to)0e 
U8 to 129 
90 to 108 











Considering first the weight of r^l which will give a proper distribution of 
pressure to the track, we may adopt a tentative system of classification for the 
track structure based upon the kind of tie, tie spacing, depth and kind of 
ballast, and character of subgrade. As previously noted, a tie spacing of 20 
inches with 18 inches of gravel or 12 inches of stone under the tie, resting on a 
roadbed capable of bearing 1.5 tons per square foot, will sustain safely a load of 
700 pounds per linear inch under each rail. This grade of track we will deag- 
nate as class A. 

Class B and C will represent weaker structures, which may be brought 
about by a departure in any or all of the elements from those found in class A 
track. 

A track would be graded as class B if it was capable of carrying only 600 
pounds per linear inch under each rail. This mi^t occur in several ways. 
A tie spacing of 22 inches, but with all the other elements of class A track, would 
diminish the strength of the structure on account of greater concentration of 
the load on each tie and on the subgrade; similarly, a lesser depth of ballast 
would affect the load on the subgrade. Evidently, a track with all the other 
elements of class A, but resting on a soil having a lower bearing power, would 
offer less resistance to the action of the wheel load. 

Table LXXXII presents descriptions of different kinds of track in each of 
these classes. It will be observed that the limiting factors may be considered 
as being the supporting power of the roadbed and the bearing of the rail on the 
tie. The bending stress in the tie and the bearing of the tie on the ballast are of 
secondary imjwrtance. 

Considering first the bearing under the rail. This is evidently a function 
of the tie spacing and kind of wood of which the tie is made. For class A track, 
capable of supporting 700 pounds per linear inch under each rail, a tie spacing 
of 20 inches would ^ve a load on each tie under the r^l bearing of 14,000 
pounds. This is about the limit of the strength of the woods shown for class A 
track, and therefore excludes the use of weaker woods or greater tie spacing for 
this class of track. 



,y Google 



STRENGTH OF THE RAIL 



317 



In class B track, having a supporting power of 600 poimds per linear inch 
under each rail, the tie spacing may be increased to 22 inches for the woods 
allowed in class A track. It is doubtful whether the group of woods shown in 
the table under class C track should be used for class B track even with a tie 
spacing of 20 inches. 

TABLE LXXXIL — CLASSES OP TRACK 



CblHof 


Ti.. 


^L^-v^'" 


Subcrads 

lOtol.S 
1.0 to 1.5 

l.Otol.5 
1-0 to 1.5 
0,8 to 1-2 
1.0 Ui l.S 

O.Stol.O 
0.5 to 1,0 


BtuHn; Powtc 




Sua. 


CUwoI Wood* Rtpraaeatsd by 


cIS'j* 


Stone. 


Gnvel. 

IDCW. 

18 
18 

18 
14 
18 
10 

18 
12 


-X^" 


At 


iHbO. 

7X9 
7X9 

7X8 
7X8 
7X8 
7X8 

7X8 
7X8 


Oak, locust, hard maple, hickory, 

cherry; not tie plated. 
LongleaX pine, black walnut, beech. 

birch, elm, Kum, hemlock, Douglas 

fir; tie plated. 


IneheB. 

20 

20 

22 
20 
20 
22 

22 
20 


iMhw. 
12 
12 

12 

9 
12 

6 

12 

8 


700 
700 








Bt 




600 


c, 


I/)blolly pine, shortleaf pine (at timee 
this wood is nearly equal to longleaf ) 
sort maple, catalpa, chestnut, white 
pine, Norway pine; tie plated. 


460 















The low-supporting power of class C track permits the use of weaker 
woods. A 22-inch tie spacing for this grade of track ^ves a load at the rail 
bearing of about 10,000 pounds. 

The pressure transmitted to the subgrade is determined by the spacing of 
the ties (or more properly by the distance between adjacent ties) and the depth 
and kind of ballast used. In the table three characters of subgrade are con- 
sidered, the firmest having a bearing power of from 1.0 to 1.5 tons per square 
foot and the least firm a bearing power of from 0.5 to 1.0 tons per square 
foot. 

By applying Uie formula 

(1 -sm<j}) 
we see that the firmer grade corresponds to a soil with an angle of repose of 
from 23 degrees to 31 degrees for 12 inches of stone ballast or 18 inches of gravel 
ballast under the tie, and from 28 degrees to 36 degrees with 6 inches of stone 
or 10 inches of gravel ballast under the tie. Table LXXIX shows that these 
angles (rf repcse fall writbin the limits given for dry sand, clay, and mixed 
eartii. 



,y Google 



318 STEEL RAILS 

The subgrade, having a supporting power of 0.5 ton per square foot, cor- 
responds to a soil having angles of repose varying from 13 degrees to 20 degrees 
under the conditions stated in the table. This agrees with the angle of repose 
for wet clay shown in Table LXXIX. 

Table XLVII (article 19) gives, for a subgrade capable of bearing 1.5 tons 
per square foot, an allowable load applied to the tie at the rail bearing of from 
10,100 to 11,800 pounds, for the size of tie, depth and kind of ballast used in 
class A track. Owing to the rapid application of the load it has been assumed 
in article 19 that these values could be safely increased to 14,000 pounds on 
account of the inertia of the track and roadbed. For 20-inch tie spacing this 
gives a supporting power of 700 pounds per linear inch of Trail. 

For the track designated by Bi in Table LXXXII, owing to the increase in 
distance between adjacent ties this value falls to 600 pounds per linear inch of 
rail. In class B, track we find from equations Nos. 1 and 2, article 18, that the 
allowable load at the rail bearing, as determined by a supporting power of 1.5 
tons per square foot on the roadbed, is 9500 pounds in the case of 9 inches of 
stone under the tie and 8500 pounds for 14 inches of gravel. Making the same 
allowance for the inertia of the roadbed, as in the previous case, it is seen that a 
supporting power of about 600 pounds per linear inch of rail is realized. 

In class Bj track, if we take the supporting power of the roadbed as 1.2 
tons per square foot, we find values from 7700 to 9000 pounds for the load that 
can be applied to the tie at the r^l bearing, which agree fairly well with those 
previously determined for class Bj track. 

The bearing power under each rail, as determined by the permissible load 
on the roadbed for class C track, has been calculated in a similar manner. 

It will be noticed that in each case the upper Mmits of the bearing power of 
the roadbed have been worked to in determining the values given in the table. 
This is a featiue of the analysis which requires careful consideration of the kind 
and volume of traffic over the track. 

The inertia of the roadbed plays an important part in strengthening the 
track when the maximum loads imposed upon it do not occur too frequently, as 
in the case of high-speed passenger trains where the most destructive forces to be 
provided for are those produced by the drivers of the locomotive. In the case 
of dense freight traffic where the heavy loads imposed by the engine drivers are 
followed by the passage of a long train, thus subjecting the track to a contin- 
uing load lasting over a considerable interval of time, the inertia of the road- 
bed is, in a great measure, overcome and a correspondingly lower value for the 
allowable pressure on the roadbed must be used. 



,y Google 



STRENGTH OF THE RAIL 



319 



The all steel 70-ton coal cars, which are coming into use on some of the 
large coal-carrying roads in the Blast, weigh over 50,000 pounds, and have a 
capacity of 140,000 pounds. This weight is carried on four axles and a tr^ 
composed of these cars would prove very destructive to the roadbed unless an 
ample provision was made for the effect of the repeated application of the heavy 
wheel loads. 

Examining now the weights of rail required to distribute the wheel loads 
so as not to exceed the bearing power of the track. Table LXXXIII shows the 
moments of inertia of standard rails*; evidently the rails with the highest 
moment of inertia will give the most favorable loading of the track structure. 



Wtiflit, PoBiid* Ifomtat ot In- Saetioo Modiilua of 



A.S.C.E 

A.S.C.E 

A.S.C.E 

A.R.A. type A 

A.R.A. type A 

A.R.A. type B 

A.R.A. type B 

P.8., Pennsylvania R. R. System. . 
P.S„ Pennsylvania R. R. System. . 
P.R.R., Pennsylvf " ~ 



43.8 


16.11 


30.0 


12.00 


22,8 


9.62 


48 9 


17.78 


28-S 


12 46 


41,3 


15.74 


25-1 


11.08 


41.9 


15.91 


29.1 




38.0 


14 29 


27.4 


11.25 


49.0 


17.00 


30.4 


11.76 


28,5 


n 53 



In Plate XXVI are given the dynamic wheel loads with different axle 
spacing for rails having moments of inertia of 20, 30, 40, and 50. The curves 
shown on the diagrams are calculated by the method explained in the first part 
of article 23 and show the allowable dynamic axle loads, as determined by the 
safe bearing power of the different classes of track given in Table LXXXII. 

An examination of these diagrams shows that for each class of track the 
allowable wheel load increases with the axle spacing up to a certain point when 
a maximum is reached and, as the spacing of the wheels is still further increased, 
the allowable wheel load decreases. The most favorable axle spacing, as might 
be expected, is greater for the heavier rails than for those of lighter section. 

• If the moment of inertia of the section is not known it may be calculated by one of the following 
methods: 

Culmann, C Centralellipse und kenieines schienenprofils, 5 p. 111., 1875. (In hia Die graphische 
statik, ed. 2, p. 475.) 

Morely, Arthur. Graphical detcrmi nation of momenta, centroids, and moments of inertia of 
areas, S p. lU., 1908. (In his Strength of Materials, p. 117.) 

Sankey, C. E. P. Note on the griiphic^ detennination of moments of inertia, 1000 w. dr., 
1810. CIn Engineer, London, v. 110, p. 57.) 



,y Google 



320 STEEL BAII^ 

Plate XXVII presents diagrams showing the maximum bending moments 
in the rails under the conditions given in Plate XXVI. In the table shown on 
Plate XXVII are given bending moments corresponding to an extreme fiber 
stress of 20,000 pounds per square inch in the base of the rail. In using this 
table it should be borne in mind that the relation between the moment of 
inertia and section modulus of different sections vary. The values given in 
the table represent average conditions. 

A comparison of Plates XXVI and XXVII shows that the wheel load is 
determined by the bearing power of the track in the case of classes B and C 
track, but with class A track the working stress of the steel may be exceeded 
without overloading the track. The dotted lines in the diagram for class A ' 
track on Plate XXVI show the correction necessary to apply to the curves of 
this diagram in order to keep within the working stress of the metal of the rail. 

Plate XXVI will now serve as a basis for determining the dynamic wheel 
load corresponding to any section of r^I and axle spacing. For a wheel arrange- 
ment consisting of a series of wheels having the same spacing and each sup- 
porting approximately the same load, the values of the load may be read directly 
from the diagrams. This loading satisfies the condition of the calculations 
which assumes that the tangents to the elastic curve of the rail under the wheels 
and midway between them are horizontal. 

When the wheel spacing is not the same for adjacent wheels, the average 
of the loads ^ven for each axle spacing should be taken. 

In the case of the front and back drivers the conditions are more complex. 
Here we may have a trailing truck and either a two-wheel or four-wheel leading 
truck carrying loads much less than those on the drivers. The preparation of 
charts for all of these conditions and for the case where the wheel spacing of 
adjacent drivers is not uniform would appear to be a refinement which would 
not be warranted by the data upon which the calculations must necessarily be 
based. 

It will be observed that little variation in the wheel load occurs after a 
distance between adjacent wheels of 180 inches is reached, the curve becoming, 
in most cases, nearly horizontal at this point. In other words the wheels are 
so far apart as to have little or no effect upon each other. 

The following formula is proposed to be used in determining the wheel 
loads on the front and rear drivers, it is also applicable to the outside wheels of 
the trucks under the cars. 

W W" ~ W" L_,wz 
2 ^ 2 W"^ 2, 



,y Google 



STRENGTH OF THE RAIL 321 

W = dynamic load of outside driver. 

W = dynamic load corresponding to the wheel spacing between the 

outside and adjacent driver. 
W = dynamic load corresponding to the distance between the outside 

driver and the center of truck. 
W" = the value given below for different moments of inertia and classes 
of track. 
L = dynamic load on truck wheels (one side). 



Uoinat of te- 


'-c£s?sr 


For Bull Drivwi. 

OMofTrKk. 


A. 


B. 


c. 


A. 


B. 


c. 


fiO 
40 
30 


36.000 
30,000 
26,000 


28,000 
26.000 
24.000 
22,000 


18,000 
16,000 
14,000 
12,000 


40,000 
34,000 
30,000 


32,000 
30,000 

28,000 
26,000 


30,000 
18,000 
16,000 
14,000 









W' 



rail outside of the driving-wheel base if there is no leading truck, this is made 
smaller than indicated by the diagrams of Plate XXVI in the case of the leading 
driver on account of the probable extra stress set up in the rail when it is first 
depressed by the weight of the locomotive. 

W" - W" L_ 
'W" 



The term - 



-, is introduced to provide for the extra support 



afforded by the truck wheels. In the ejctreme case where W" = L, the wheels 
ahead of the driver exert the pressure corresponding to their distance from the 



no leading truck the term 



W" - W" 



} becomes equal to zero. 



2 W" 

The dynanuc wheel loads, having been determined, the corresponding 
static loading can be readily computed by the methods given in article 10 for 
steam locomotives, article 11 for electric locomotives, and article 12 for cars. 

Plate XXVIII presents diagrams showing approximately the static loading 
that can be placed on the r^l with different wheel arrangements. From these 
diagrams can be obtained the weight of rail and design of track necessary to use 
in connection with engines where the maximum axle load is fixed, or the diagrams 



,y Google 



322 STEEL RAII5 

may be used in determining whether or not it is safe to run a certfun wdght of 
equipment over an existing line. 

On account of tlie variation in design of engines a separate examination 
should be made in most cases, as the diagrams, of necessity, represent general 
conditions which may be varied from in a considerable degree. 

Fig. A, Plate XXVIII, gives the track diagrams for passenger engines of the 
Atlantic and Pacific types. The main driver in the wheel arrangement of the 
Pacific engine can carry more weight than when it is one of the outside wheels, 
as in the case of the Atlantic engine, and for this reason the former engine is 
generally the most favorable on the track. 

The ten-wheel engine is used extensively for passenger and freight service 
on branch lines. This engine does not have the trailing truck of the two former 
tyi»es, and the rear driver has, therefore, less carrying power than in the Pacific 
engine, although about the same as the Atlantic where the effect of the tr^ling 
truck is offset by the fact that the rear driver is the main wheel. The diagrams 
given on Fig. B, Plate XXVIII, show the axle loads of ten-wheel engines for 
classes B and C track. 

Fig. C shows diagrams for Mogul and Consolidation freight engines. The 
wheel arrangement of these engines is quite similar to that of the ten-wheel 
en^ne, and it will be found that the curves agree very closely with that of the 
ten-wheel engine used in freight service. 

On Figs. D and E are diagrams for cars. It should again be observed 
that these smaller diameter wheels should not be loaded as heavily as the 
drivers, and the diagrams for the loads on car axles are not extended beyond 
axle loads of 45,000 pounds. 

The diagrams of the figures on Plate XXVIII illustrate very clearly the 
effect of the different wheels on the track, and emphasize the fact that the 
entire wheel arrangement must be considered in determining the maximum 
load on any one wheel. 

The assumption made by most foreign writers on this subject, that the 
strains produced by the loads are independent of the position of the wheel is 
obviously incorrect, as has been shown experimentally by Dr. Dudley's strem- 
matograph tests. 

The diagrams of Plate XXVIII are not extended beyond 60,000 pound 
axle loads for drivers or 45,000 pounds for car wheels. With carbon steel 
rails the use of very heavy loads should be approached cautiously until 
further evidence is obtained in regard to the effect of such wheel loads on the 
intensity of the stress at the contact of the wheel and the rail. While axle 



,y Google 



STRENGTH OF THE BAIL 323 

loads of nearly 70,000 pounds are in service on experimental locomotives, they 
have not been used in sufficient numbers to demonstrate fully their effect on 
the rail. 

The lack of proper experimental data presents many difficulties in accu- 
rately determining the functions performed by the rail. Within certain limits, 
however, the duty of the rail can be calculated with a considerable degree of 
accuracy and more attention should undoubtedly be given to the effect of dif- 
ferent loadings on the rail in the design of the engines and cars which run 
over it. 

It must be constantly borne in mind in dealing with the design of the track 
that in many cases the strength of the rail is not the first consideration in the 
selection of the section to be used, and that the question of obtaining a rail of 
the requisite stiffness is of the greatest importance. The sudden failing of 
any part of the track is not to be anticipated within the limits of customary 
practice, but rapid deterioration of the structure may take place which will 
eventually result in failure. 

Economy of train service has become so important that it is safe to say 
that there will be no return to lighter loads the tendency is, and will be constantly, 
in the opposite du-ection. The unportance therefore of pving to the deagn of the 
track the same careful investigation as is considered essential in the design of a 
bridge cannot be over-estimated. The track is, in fact, a continuous girder con- 
necting termini over which pass the same loads as over the bridges. 

The discussion of stresses in the ties, ballast and subgrade which has beai 
made in the preceding pages while suffident to enable the allowable bearing power 
of the supports of the r^l to be determined witiiin reasonable limits, is far from ex- 
haustive enough in its character to serve as a ba^ for the general design of t^e 
track and the proper proportioning of all the elements entering into its construc- 
tion. Such an analysis would be clearly outade the limits of the present work and 
while the various tables and fonnulse that have been developed appear sufficient 
for the purpose intended, any general conclusions based upon their evidence alone 
should be avoided. 



,y Google 



CHAPTER VI 
INFLUENCE OF DETAIL OF MANUFACTURE 

The evidence of the Mlure of rails of heavy section rolled within the 
last few years, equaling and at times exceeding that of lighter rails of earlier 
manufacture exposed to similar conditions of traffic and roadbed, points un- 
mistakably to defective material in some of the later rails. These rails apparently 
do not f^ in the majority of cases due to too great a stress of the metal, and it is 
this irregular failure of mdividual r^ls due to defects in manufacture which has 
giv^ rise to such just feelings of dissatMaction and alarm. 

Inferior quality of the metal in the rail may be attributed to two causes: 
first, the use of imperfect methods of manufacture; second, the influence of the 
form of the section upon the detail of manufacture. 

First, let us examine the methods employed in the manufacture of the earlier 
rails, which gave such satisfactory results, and which have been so constantly 
presented to rail makers as representing that which they ou^t to do. 

* The first steel rails were rolled in mills which had been designed for 
iron rails. Other rolls were used and the number of passes was increased, 
making the reduction very gradual. All blooms were allowed to cool before 
being charged mto the reheating furnace. After the drawing of one heat and 
before the charging of another, the fiUTiace was cooled down, then the heat 
was brought up very gradually and plraity of time was taken to allow the steel 
to " soak." In the converting house, all the possible practices of crudble steel 
teeming were introduced. The ingot molds were carefully brushed out, heated 
and smoked before being used. When the steel was teemed all doors and 
windows of the casting house were closed and time was not spared on any of 
the details. It was expected to produce but 50 per cent as much steel as iron 
rml, and all employees working by the ton were paid twice as much for sted 
as iron. The constant demand for cheaper prices (Fig. 220) and increased 
tonnage rapidly changed these conditions. 

Many of these practices, it is felt to-day, were entirely without reason, and it 
is difficult to say as a general proposition that the steel produced was better than is 

* See paper on Steel Rails, and Specifications for Their Manufaoture, R. W. Hunt. Trana. 
American Institute irf Mining Engineera, Vol. XVII (: " " " 



,y Google 



xnixuenct; of detail of manufacture 



325 



obtainable at the present time. Mr. Buffington stated podtiveiy to &e Indiana 
Kailway Commisdon, at its hearing, that the quaUty of the metal is to-day much 
better than it ever was before, owing to the increased knowledge and bett^ machines 
and mechanical appliances than formerly existed. 

Mr. James E. Howard, in his report of the accident on the Great Northern 
Eiulway, near Shanm, N. D., on Decembra- 30, 1911, states; "It is important to 
consider whether an unprovement in the structural condition of rail steel is att^- 
able. Such seems to be the case, dnce experimental rollings have furnished rails 
*I50, , , , , , ,*I50 




1860 I870 I880 I890 IQi 

YEAR 

FiQ. 220. — Prices of Iron Euid Bessemer Steel Rails, 1855 to 1910. 



which, so far as could be ascotained, were free frcan streaks ... It is believed to 
be metallurpcally feaable to produce better steel than has at times been offered 



No doubt the failures which have their origin in defective metal are conader- 
ably augmented by the character of the stress in the rail. On account of the con- 
centration of tile load on a small area, the stress is not distributed, and consequentiy 
a metal of a great degree of uniformity is required. 

With the large wheel loads now in use the injurious dfect of inferior material 
in the r^ is much more apparent than in other structures not subjected to such 
highly localized stresses. The situation calls for a refinement of manufacture not 
generally realized in practice, and is further comphcated by the dedre for high 
carbon to resist the head stresses, with tiie need for physical properties in the other 
parts of the rail which could best be obtained by the use of a much milder steel. 



,y Google 



326 ste£l rails 

29. Chemical Composition 
It was supposed that the chemical character of the steel in the earlier 
rails accounted for their excellent wear. Among the makers of these rails. 
Sir John Brown & Co., of Sheffield, England, sent to this country those which, 
from their excellent service, were considered by railroad engineers as the type 
of what rails should be. Accepting the chemical theory, rail makers expected 
that the analyses of these celebrated rails would present steel of exceptional 
uniformity and purity. The contrary was proved to be the case. Carbon 
varied from .24 to .70; silicon, .032 to .306; phosphorus, .077 to .156; sulphur, 
.050 to .181; manganese, .312 to 1.046. The following is the variation found in 
thirteen rails made by John Brown & Co., England, all of which had given 
good service; 

Carbon 0.24 to 0.70 

Manganese 312 to 1 .046 

PhoaphoruB 077 to .166 

Sulphur 050 to .155 

Silicon 032 to ,308 

Below is given the analysis of some of the earliest English rails, imported 
between 1860 and 1870. These rails, low in carbon and all other hardening 
constituents, have pven from thirty to thirty-five years' service before wearing 
out, not breaking. 





EUila Uwd on 
Southern Railwiy. 


Raila Used on P.. C. C. A St. L. 
Railoiy. 




PercBit. 
0.158 

.77 
.114 

.067 
,490 


P«CMt. 

0,273 
.28 
.05 
.04 
,025 


Per Dent. 
0.22 
.21 
.05 
.031 
.035 




Phosl^on^' 









Fig. 221 shows the performance of two rails of very similar chemical com- 
position, which, however, possessed quite different wearing qualities. 

* In 1881, Dr. C. B. Dudley, the chemist of the Pennsylvania Raihoad, 
made an investigation to determine the relative relation between the chemical 
and physical characteristics of steel rails and their power to resist wear. Dr. 
Dudley found for the average of 32 slow- wearing rails the following composition: 

Poceat. 

Carbon 0, 334 

PhOBphorufl 077 

Silicon 0,060 

Manganese 0,491 

• The Wearing Oaparaty of Steel Rails in Relation to Their Chemical Compoeition and Phyrical 
Properties, C. B. Dudley, Trans. American InsUtute of Mining Engineera, Vol. IX (1880-81), p. 321. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 327 

and proposed a formula for the correct composition of steel rails, as follows: 

Carbon, between .25 and .35; aim at 0.30 

Phosphorus, not above 0.10 

Silicon, not above. 0.04 

Manganese, between .30 and .40; aim at 0.35 





Fia. 221.— Comparative Wear of Rails of Similar Chemical Composition. (Trans. A. S. C. E. II 

Chemical 0.322 per cent Carbon 0.355 per cent. 

Composition 0.026 per cent Silicon 0.039 per c«nt. 

0.077 percent Phosphorus 0.108 per cent. 

0.492 per cent Manganese 0.490 per cent. 

^earRoUed 1871 1876 

Time in Service August, 1871 to July, 1879 August, 1876 to July, 1879 

7 years, 11 months. 2 years, 11 months. 

Location On hinh side of 5° Curve On high side ot 5" Curve 

Grade 39.6 ft. per mile. Grade 39.6 ft. per mile. 



Tonnage over Rail 40,061,230 t< 



21,504,824 t< 



Table LXXXIV ^ves the result of his experiments. 



Kind ol Sarvice, 



""Kii: 



Ton toLoMS 



Level tangents 

Grade tangents 

Low side level curves. . 
High side level curves. . 
Low aide grade curves. 
High side grade curves, 

Tangents 

Curves 

Low side curves , 

High aide curves 

All conditions 

3 slower wearing 

2 faster wearing 



.0701 

.0706 
.1277 
.0500 
.0911 



.0650 
.1332 

.0767 
,0506 



208.4 
114.1 
113 3 
62.7 



45.7 
147.6 

80.6 
146,8 

80,9 
123.1 

60.1 
104,3 
158,1 

77,8 



,y Google 



328 STEEL RAII5 

Wellington states that the result of this investigaMon, which showed, or 
seemed to show, that very hard rails did not wear so well as softer and tougher 
r^ls, was taken to indicate that softness in itself was a desirable quality in a 
rail; and the painstaking character of the mvestigation and high reputation of 
the road having given these conclusions wide dissemination, manufacturers for 
many years took them as a guide, and produced rails that were too soft. 

Table LXXXV shows the gradual increase of the hardening constituents in 
the steel for rails sance Dr. Dudley's investigation. 

TABLE LXXXV. — COMPARISON OF THE CHEMISTRY OP EARLY AND RECENT RAILS 





DkW 












Ch 


rm.try. 


Nmm. 


CuboD. 




Silio«i. 


Sdphur. 


Phoepb<™. 






Day. 


HiB. U 


^ 


Hill. 


M„. 


Uin. 


Uu. 


uio. 


M». 


MLt 


^ 




ISgl 
1S» 

IBM 
190> 






P« F 

.M 

.30 

1 

41 


1 

K 




FW 


Per 


Per 


Per 


P« 


Pw 


Pw 














«0- TO 
TO- 80 

K 


":m 








.20 




.OS* 








1 

;?» 

.90 

BO 
.SO 


1 IE 

l.O) 

roc 

l.X 
1.01 


.10 
.0* 








The CbTDicie 9Mel Compiiny 























It has often been stated that the reason why the earlier rails seem to last 
80 well was due to the elimination of the poorer quality of rails by the service 
in the track. This statement is not the complete explanation. The older 
rails were cold rolled by the light wheel loads until the surface was sufficiently 
hardened to bear the recent heavier loads without much increased abraaon. 
New rails of the same section and practically the same physical properties 
would, when subjected to heavier wheel loads, lose more of the metal by wear, 
before the surface was rolled as hard as the former sections, and their rate 
would be much faster.* 

The light earlier sections could carry from 60 million to 75 million tons 
before they were rough and imsuitable for passenger traffic, and, when in a 
location where the tonnage was only 250,000 to 500,000 tons per month, 
would last in the track many years. 

The six-inch 100-pound Dudley sections of 0.06 phosphorus and 0.65 car- 
bon laid in 1895 on the New York Central and Hudson River Railroad, when 
taken up in 1907 had carried 375 million tons with a loss of about one-eighth of 
an inch in depth on the head of the rail. 

' Peiper by Di. P. H. Dudley before the Railroad CommioBion of Indiana, February 20, 1912. 



,y Google 



INFLTJENCE OF DETAIL OF MANTJFACTDRE 329 

Carbon is the most important element, except iron, in steel. The me- 
chanical properties of iron-carbon alloys are closely connected with the relative 
amounts of the two elements. The relation between the percentage of carbon 
in an alloy and the tenacity in tons per square inch is indicated* in the following 
table: 



Per cent of carbon - 0.05 0.1 0.2 

Tenacity.tonspersq.m. -26.00 26.0 31.0 



0.4 
36.0 



0.6 
48.0 



0.8 1.0 1.8 
58.0 60.0 44.0 





/ 


A 


\ 


i 

8" 


/ 




\\ 


s. . 
: / 

1= / 


/ 















The results are shown graphically in Fig. 222. 

^con in small proportions hardens the steel and stands intermediate 
between carbon and phosphorus in this respect. It is used to prevent unsound 
or honey-combed ii^ots, but when so used so 
tends to render the steel imduly hard. Silicon 
as high as .2 per cent, in high-carbon steel of 
.5 and .6 per cent carbon, probably has no 
injurious eflEect. 

Ph(Mphorus hardens steel more rapidly 4o 
than either carbon or mlicon. It increases 
its rigidity but impairs its power to resist 
impact. Small proportions render the metal 
harder without materially affecting its 
tenacity, but makes the metal at the same ^^, 
time decidedly cold-short. An excess of phos- Fro- 222. 
phorus also rend^^ t^e steel saudtive to high 
heat. Mr. Robert W. Hunt, in his experim^its m trying to make high phosphorus 
steel in the Clapp-Griffith convwter, found that it was necessary to be very careful 
not to over-heat the steeJ. 

t Owing to the exhaustion of the avmlable low-phosphorus ores, Bessemer 
rail steel is now of necessity a high-phtsphorus and low-carbon alloy, the mean 
carbon being about 0.50t per cent, while the impurity of phosphorus is limited 
to 0.10 per cent. 

Plain baac open-hearth r^l steel is usually a low-phosphorus and medium 
unsaturated carbon alloy, as most of the phosphorus has been reduced by this 

* H. M. Howe, EngJDeering and Mining Journal, p. 241, 1S87. See also Steel by Harbord and 
Hall, LondtHi, 1911, pp. 347, 34S, and a Study of the Elastic Properties of a. Serkfl of Iron-Carbon 
Alloys, Jones and Waggoner. Proceeding American Society tor Testing Materials, Vol. XI, 1911, 
pp. 492-499. 

t See Proceeding American Society for Testing Materials, Vol. XI, 1911, Dudley, Ductility in 
KaU Steel. 

t The chemical compoeitioii nitse, in tins and the following example, to 100-pound rails. 



0-4 0-8 1*2%Carbon 

Tenacity of Iron-Carfoon Alloys. 
(H. M. Howe.) 



,y Google 



330 STEEL RAILS 

process from its content in the ores and iron to 0.04 per cent or under. This 
permits, in this class of steel rails, carbon of 0.63 to 0.75 per cent. 

Sulphm* has little influence on the tensile strength or ductility. The 
real effects of sulphur, however, are seen during the rolling, a very small per- 
centage causing a great red-shortness. Its presence in excess of .06 or a maxi- 
mum of .08 per cent tends to cause cracks to develop during the rolling, which, 
while they close up and are almost imperceptible in the finished rail, neverthe- 
less rem^n as flaws and may form starting points for rupture when the rail 
is subjected to any sudden stress. With sulphur, it is necessary to work the metal 
at a high heat to avoid its cracking during manipulation. The " red-short" term 
means that as the heat approaches the red color ^e tendency to crack becomes 
intenafied, while the effect of phosphorus on heated metal is to make it hot 
short or short undw high heat; in other words, it will work at a low tempera- 
ture, but is sendtive to a high one. 

Apparently the greater part of the sulphur unites with the manganese 
forming manganese sulphide, which is occluded by the metal as a foreign sub- 
stance, preventing its welding and breaking up the continuity of its structure. 
The impurity of sulphur was limited formerly to 0.075 or 0.08 per cent. The 
manufacturers now chaise for this limitation of sulphur five cents extra per 
hundred pounds, and it is, therefore, being omitted from some specifications, 
although in most cases it is required that its content be reported. 

Manganese has a general tendency to increase the tensile strength and 
reduce the ductility; this influence varying with the amount of carbon pre- 
sent in the steel and becoming more marked in the case of high-carbon than 
low-carbon steels. It is possible to keep the manganese down, by the use of a low 
manganese spiegle, and with low-sulphur steel its presence in excess of .8 per cent, 
or its use to bring up the tensile strraigth in place of carbon, is dangerous on account 
of its very distinct hardening effect when above .6 per cent. In the commercial 
run of iron, where the sulphur varies, the practice is to allow the manganese to go 
as high as l.i per (%nt, and some authorities do not consider it dangm)us unless 
above 1.0 per cent even with low sulphur. Manganese tends to neutralize the effect 
of sulphur and prevent the metal becoming red-short, and, to a limited degree, the 
cold-shortness produced by phosphorus. 

The above elements are those generally considered in rail steel, and speci- 
fications rarely refer to the other elements which may be contained in the ore, 
and which either from design or accident are present in the finished product. 
The most important of these are arsenic and copper. 

The effect of arsenic upon steel was quite fully investigated several years 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



331 



ago by Harbord and Tucket.* The conclusions given by them may be sum- 
marized as follows: 

Arsenic, in percentages not exceeding .17, does not appear to affect the 
bending properties at ordinary temperatures, but above this percentage cold- 
shortness begins to appear and rapidly increases. In amounts not exceeding .66 
per cent the tensile strength is raised very considerably. It lowers the elastic 
limit and decreases the elongation and reduction of area in a marked degree. 

Messrs. Ball and Wingham f have investigated the influence of copper on 
the tensile strength of iron and steel. An alloy containing: 

Copper 7.550 

Carbon 2.720 

Manganese 0,290 



Phoaphorus . 130 

Sulphur 0.190 

was bright, white in color, crystalline, and very hard, but did not offer any 
great resistance to impact. Varying quantities of the alloy were melted down 
with Bessemer steel, and teat pieces 1 inch by J inch by fg inch were annealed 
before being tested. The following table shows the results: 



Number. 


Copp«. 


CurboB. 


.KS. 




Pwoenc. 


Percent. 


Tom per «i. in. 


1 


0.847 


0.102 


18.3 


2 


2.124 


0,217 


36.6 


3 


3.630 


0.380 


47.6 


4 


7.171 


0,712 


56.0 



From these experiments it is clear that copper increases the tensile strength 
of iron. J The simultaneous presence of carbon tends to prevent the more 
intimate association of copper with iron. In test piece No. 1, the fractured 
surface was somewhat fibrous, while No. 2 and the others were highly crystallme. 
Even in the absence of carbon, copi>er makes iron extremely hard. Mr. F. 
Stubbs states that the presence of | per cent of copper in steel gives to it the 
property of preventing the oxidation of the steel on being subjected to a burning 
heat. 

Copper in steel rails in small quantity does not materially affect the me- 
chanical properties, but in steels, in which high ductility is required, especially 
in those with high carbon, copper is objectionable. Steel with copper, say up 
to I per cent, appears to resist corrosion better than the same steel without 

* On the Effect of Arsenic on Mild Steel, Journal Iron and Steel Inat., Vol. 1, 1888, p. 183. 

t Iron and Steel Inst., No. 1, 1889. 

t Steel and Iron for Advanced Students, Hioma, London, 1903, p. 322. 



>, Google 



332 ST£EL RAII^ 

copper. * Campbell states that 1 per cent may be present without injuring 
the steel, provided there be but little sulphur; but that if the sulphur be up to 
.08 or .10, the metal will be red-short, and that copper also reduces the welding 
power of the metal, especially if sulphur be present; but he adds: " In all cases 
the cold properties seem to be entirely unaffected." f Richards states that 
"copper causes red-shortness, but much less than sulphur. Five-tenths per 
cent may be allowed in rails and its effect is overcome by manganese." 

t The influence of copper on steel was formerly greatly exaggerated. 
Whereas it was considered to be very harmful, it is now known, when present 
in small quantities, to have no serious influence on the physical properties of steel. 

Mr. H. J. Force reports a case of an 80-pound rail made by the Lackawanna 
Steel Company in 1895, which had given very good service. An analysis showed 
about .40 per cent carbon and about .60 per cent copper.§ 

According to a statement in Professor Howe's " Metallurgy of Steel " || an 
American firm of steel-rail makers habitually made Bessemer tee rails with .51 
to .66 per cent copper and they were so slightly red-short that in spite of the 
thin flanges and low finishing temperature only from 1.25 to 2.5 per cent of 
tbem were so defective as to be classed as second quality. 

Mr. R. W. Hunt states that in the early days of the steel industry excellent 
rails were produced from Cornwall irons. A large number of these r^ls con- 
tained .5 per cent of copper. The Pennsylvania Steel Company, as well as 
the Bethlehem and the Troy Works, used Cornwall iron containing low phos- 
phorus and high copper as their basis for a long time. 

Clamer H has found that the addition of copper and nickel in combination 
seems to have the same effect upon the steel as if they were individually added, 
the copper in its effect really being about the same as so much added nickel. 
It is possible, therefore, to replace part of the nickel in nickel steel by copper, 
without materially altering its physical properties. Recently Messrs. Burgess 
and Aston, working quite independently of Clamer, have confirmed these results. 

The attention of railroad engineers is being directed toward the develop- 
ment of alloy steel, or steel containing a percentage of various materials intro- 
duced to give it special mechanical qualities. In general, however, on account 
of the higher cost of production, these steels are confined to use in special locali- 

* Metallui^ of Iron and Steel, A. Humboldt Sexton, Manchester, 1902. p. 247. 
t Notes on Iron, Robert H. Richards, 1895. 
t Metallurgy of Steel. Harbord, London, 1911. p. 375. 
5 Proceedings American Society for Testing Materials, Vol. X, 1910, p. 279. 
II Metallurgy of Steel, New York, 1891. p. 83. 

^ Proceedinga American Sodety for Testing MateriaJa, Vol. X, 1910, Clamer on Cupro- 
nickel Steel. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 333 

ties where the conditions are especially severe, as on sharp curves under heavy 
traffic or in tunnels where it is a troublesome matter to inspect or renew the 
rails. 

The requirements of steel alloy may be summarized as follows: 

(1) High resistance to shock; 

(2) High elastic limit; 

(3) Resistance to abrasion. 

Some of the alloys best known are manganese, nickel, chromium, and titanium. 
* The record of the chrome nickel on the Central Railroad of New Jersey, 
and of the plain nickel on the. Pennsylvania Lines, Northwest System, is not 
very good. In a period of six months there were 112 failures per 10,000 tons 
laid of 85-pound A. S. C. E. section of nickel steel from the Carnegie Steel Com- 
pany on Pennsylvania Lines, Northwest System, the chemical composition 
being: 

Pnont 

Carbon 44 

Phosphorus 09 

MaDganeee 80 

Silicon 10 

Sulphur 03 

Nickel 3.42 

The 90-pound A. S. C. E. section chrome nickel steel from the Bethlehem 
Steel Company on the Central Railroad of New Jersey for the same period 
showed failures of 41 per 10,000 tons laid, nearly all of which were broken rails. 

The same committee reported for the year, ending October 31, 1910, that 
the record for 90-pound A. S. C. E. open-hearth rail with chromium and nickel 
on the CentrjJ Railroad of New Jersey has been very bad so far as f^lures are 
concerned, there having been 1,129 per 10,000 tons of rail Idd, mostly break- 
ages. This rail is 1909 manufacture. Small lots have also been tested on the 
Baltimore and Ohio and the Erie with a large number of failures. The amount 
of nickel is 2 per cent to 2^ per cent, and the chromium 0.5 per cent to 0.9 per 
cent. In most cases these rails showed a very marked resistance to flange 
wear as compared with ordinary carbon steel rails. 

It has been found dedrable to lower the carbon when the other hardening 
elements are added. A rail with carbon 0.40, chrome 0.50, and nickel 1.25 is about 
equal to a 0.60 carbon ordinary rail. 

Manganese steel with C 0.77, P 0.06, Mn 9.93, Si 0.25, and S 0.038 showed 
about one-third as much abrasion of the head as ordinary Carnegie Bessemer in 
a test, on the Norfolk and Western, lasting nineteen months. 

* Proceeding Am. Ry. Eng. & M. of W. Aasn., Vol. 11, Part 1, 1910, p. 316. 



,y Google 



334 STEEL RAILS 

Chrome ^teel, which usually contains about 2 per cent of chromium and 
.80 to 2 per cent of carbon, owes its value to combining, when in the " hardened" 
or suddenly cooled state, intense hardness with a high elastic limit, so that it 
is neither deformed permanently nor cracked by extremely violent shocks. 

* The tensile strength rises with increase in the percentage of chromium 
till with about 5 per cent it is about 74 tons unannealed, or 55 tons annealed, 
the elongation being 13 per cent in the latter and 8 per cent in the former case. 
The limit of elasticity was 40 tons in the first and 20 tons in the second case. 
As the quantity of chromiiun is increased the metal becomes harder, and with 
about 9 per cent can hardly be touched with the file. In the absence of carbon 
its hardening influence is not so marked. Forging makes the metal hard and 
brittle, but the latter property is removed by annealing, and it is rendered 
excessively hard by quenching. It has a high resistance to shock, and is there- 
fore suitable for the manufacture of rails. 

t On low-carbon steels not annealed, the addition of each 1 per cent of 
nickel up to 5 per cent causes, approximately, an increase of 5000 pounds per 
square inch in the elastic limit and 4000 pounds in the ultimate tensile strength. 
The influence of nickel on the elastic Umit and ultimate strength increases with 
the percentage of carbon present, high-carbon nickel steels showing a greater 
gain than low-carbon nickel steels-t 

The addition of nickel to steel raises the proportion of elastic limit to ulti- 
mate strength and adds to the ductility of the steel. This effect of nickel in 
increasing the ratio of the elastic limit to tensile strength, without sacrifice to 
ductility, accounts for the increase in the working efficiency of nickel steel over 
carbon steel; in other words, its increased resistance to molecular fatigue. 

The exhaustive series of experiments made by Wedding and Rudeloff 
show that the resistance to compression of nickel-iron alloys increases steadily 
with the per cent of nickel present, until 16 per cent of nickel is reached. Had- 
field has also made a very complete series of experiments on the resistance of 
nickel steel to compression. He has found that a steel containing .27 per 
cent nickel shortened, under a compression of 100 tons (224,000 pounds) per 

• Metdlurgy of Iron and Steel, A. H. S«tton, Manchester, 1902, p. 617. 

t Nickel Steel: Its Properties and Applications. Colby. Proceedings American Society for 
Testing Materials, Vol. Ill, 1903. 

t The subject of nickel steel has received considerable attention, notably by D. H. Browne, 
Trans. American Institute of Mining Engineers, Vol. 29, 1899, p. 569, and A. L. Colby, "A Comparison 
of Certain Physical Properties of Nickel and Carbon St^el," Bethlehem Steel Company, 1903. See also 
Guillet, Journal, Iron and Steel Institute, Vol. 2, 1908, p. 177; Wal«rhouse, Proceeding Am. Soc. 
for Testing Materials, Vol. VI, 1906, pp. 249-258; Campbell and Alien, ibid, Vol. XI, 1911, pp. 428- 
438. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 335 

square inch, 49.90 per cent in a length of 1 inch ; a steel with 3.82 per cent nickel 
shortened 41.38 per cent; with 5.81 per cent nickel, 37.76 per cent; and with 
11.30 per cent nickel, only 1.05 per cent. He states that an ordinary mild carbon 
steel without nickel, under similar conditions, would be shortened 60 per cent 
to 65 per cent. He argues that the toughening action of nickel when added to 
steel is caused in a very intimate combination of the molecular structure, and 
that this advantage is further enehanced by the fact that the nickel does not 
show a disposition to segregate in steel like other elements; in other words, it 
appears to be more intimately combined. 

Mr. Campbell, of the Pennsylvania Steel Company, made a series of tests 
to prove what he states to be the current impression among manufacturers 
of nickel steel, — that the presence of this element prevents segregation. His 
conclusion is, that there seems to be good ground for the assumption that nickel 
prevents tiie separation of the metalloids, but that it does not prevent it alto- 
gether, and he states that it is not probable that any other agent will ever be 
found competent for this task. 

* Howe states that nickel steel, which usually contains from 3 to 3.50 
per cent of nickel and about .26 per cent of carbon, combines very great 
tensile strength and hardness, and a very high limit of elasticity, with great 
ductility. 

The combination of ductility, which lessens the tendency to break when 
overstrained or distorted, with a very high limit of elasticity, gives it great value 
for shafting, the merit of which is measured by its endurance of the repeated 
stresses to which its rotation exposes it whenever its alignment is not mathe- 
matically straight. The alignment of marine shafting, changing with every 
I>assing wave, is an extreme example. In a direct comparative test the pres- 
ence of 3.25 per cent of nickel increased nearly sixfold the number of rotations 
which a steel shaft would endure before breaking. 

As has been seen, nickel steel has been used tentatively for railroad rails; 
but while it has the stiffness and resistance to wear which they require, too many 
rails have broken in use. We may hope that this treacheroxisness will be pre- 
vented. It is quite posdble that a change in the percent^e of nickel may give an 
entirely different record. The Mayan ore used by the Maryland Steel Company 
contains a natural percentile of chromium and nickel, and the results with rail 
made from this ore seem, so far, to be pretty good. 

Figs. 223 and 224 j^ve the tensile strength and the ductility of many speci- 
mens of nickel steel from various sources, chiefly, however, from M. Dumas' 

* Iron, Steel, and Other AHoye, Howe, 1903, pp. 316-824. Contains report of M. Dumas' work. 



,y Google 



336 STEEL RAILS 

important monograph.* The curves here given are taken from his work (pages 
18 and 19). A rough resemblance to the manganese steel curves (Figs. 225 and 
226) may be noticed. The great increase of ductility in case of manganese sted 
in the 13 per csat manganese region is reflected in case of nickel steel by a like 
and very abrupt rise at about 25 per cent of nickel. 



fim 














































' 




, 










































' 




. 


" 








































M ' 












\ 




* 










, 




















|200 


























f 




















S"°'°°° 








a 




, 




J 


~; 


— 




% 


[fe 




















« 




i 


" 


, 




r 


































jlWOOO 




1 


«•> 






/ 




. 








. 




\ 


J 


' 














i' 




i 




J 


> 














J 


111 


Ai 


















1 "" 


W 


il 




-< 


i 
















liV 


, \ 








-^ 


— 


"" 




, 




>t 




1 






















" 




r 














i "°° 


^ 
























































































































* 

















Z 4 6 8 10 t2 14 



• ' Tht SiHl ha net ncilwl \ 



ii_ifid*>IkI It h1(hl>i 



20 2Z 24 26 2B 30 32 34 3S 
PERCEKTMEOFWCKa 

H _ Tht StHi h 



" - codIkI IwHtnoitf In 



Fia. 223. — Influence of the Proportion of Nickel and Varying Heat-Treatment upon the Tennle Strength 
of Nickel Steel. (Dumas.) 

As actually made, manganese steel contdns about 12 per cent of manganese 
and 1.50 per cent of carbon. Although the presence of 1.50 per cent of manga- 
nese makes steel brittle, and although a further addition at first increases this 
brittleness, so that steel containing between 4 and 5.5 per cent can be pulverized 
under the hammer, yet a still further increase gives very great ductility, accom- 
panied by great hardness, — a combination of properties which, so far as known, 
was not possessed by any other known substance when this remarkable alloy, 
known as Hadfield's manganese steel, was discovered. 

Its ductility, to which it owes much of its value, is profoundly affected by 
the rate of cooling. Sudden cooHng m^es the metal extremely ductile, and 
slow cooling makes it brittle; its behavior in this respect is thus the opposite oi 

* " B«cherche9 but lea Aciers au Niekel h. Hautea Teneura," M. L. Dumas, Paris, 1902. 



>, Google 



INFLUEKCE OF DETAIL OF MANUFACTURE 

















































" 


























j' 




















" 
























1 






















" 


























!t 














































• 




















" ID 




























; 


1 


















\ 






















i 


I 






1 














§■" 


\* 
























1 


t 


^•! 
















|2. 


i 


V^ 




« 




















j 


• 


»- 




_, 










J** 


3 


X 




■ 














1 1 












■" 




-^ 




i'' 


f 


?f 




V 


I 






' 1 








1 






















1 


!l 




1 


■' 


\' 


' 








' 


^ 


'U 






* 






















V 








. 


'- 




.' 




















'' 




'J 


* 


h 




K 


i 


•1 


4 


/ 


t 


I 
I 


<i 




^ 






















J 








\ 


> 










f 


• 


















' 
















i 












t 
























■ 




1 




2 1 


4 




B : 


2 


2 2 


i 2 


6 28 30 32 M ae 3S M 42 44 



PEaCEHTAGE OF NICKEL 

2-* Elotnillon In 1 Inchn. 8 -* Elaniatkin in 9 or t» taeh«u 

1 >4 II '1 V bflvp tuhjictM to antrtmalv ^^ tampsnturft. 

Av fiMt'tnatm«nt. H inv, not gWmn. 

fi- TKo SIhI hu no< nniivtd frtit-liiHtmant. 

-InBuence of the Proportion of Nickel and Varying Heat-Treatment upon the Ductility of 
Nickel Steel. (Dumas.) 



Fio. 225. — Influence of the Proportion of Manganese on the Tenaile Strength of Manganese Steel. 



• -■ Slowly Cooled Manganese Steel. 
+ ■• Water-tougbened or Suddenly Coaled Manganese Steel. 



>, Google 



338 STEEL RAI13 

that of carbon steel. Its great hardness, however, is not matmally affected 

by the rate of cooling. 

The fact that when cold it is unalterably hard has, howevo", limited its use, 

because of the great difficulty of cutting it to shape, which has in general to 

be done with emery wheels instead of the usual iron-cutting tools. Another 

defect is its relatively low elastic limit. 

Fig. 225 shows the remarkable increase of tensile strength which occurs 

when the manganese rises from 7 to 13 per cent, and the decline of ten^le strength 

as the manganese increases still 
further. By the contrast be- 
tween the position of the crosses 
and the black dots it shows also 
the remarkable effect of sudden 
cooling. 

Fig. 226 shows the corre- 
sponding changes in ductility. 
To show that the maxima for 
tenale strength and ductility 
coincide, the tensile-strength 
curve sketched by eye in Fig. 
225 is reproduced in Fig. 226. 

In Fig. 227 is shown the 
degree to which manganese steel 
combines tensile strength with 
ductility, and in Fig. 228 the 
degree to which it combines 
ductility with elasticity. These 
combinations are often taken 
as a rough measure of the 
general degree of excellence of 
a metal for en^nemng pur- 
poses. For comparison the 











































































































































































































































































































































































































































































































t 






































5 








































M 


























































































































































































































^ 
































1 


















































































. 


+ 


■ 


?] 






^. 






























J 














































c^' 










■ 




. 


























A 








































/. 


























































\ 
































































/;- 






























































" 









0l234367S9t0tt t2l3Uiaie 17 I8192DZI 22 
PERCENTAGE OF NANQAHESE 
Fia. 226, — Influence of the Proportion of Manganese on 

the Ductility of Manganeee Steel. (Howe.) 
Legend: 

• — Slowly Cooled Manganeae Steel. 
+ = Water-toughened or Suddenly Cooled Manganese Steel. 



corresponding properties of carbon steel are shown by small black dots, which 
fall in a pretty well-defined band, much below the manganese-steel crosses. 

These comparisons may, however, give a false idea of the ductility of man- 
ganese steel. If two metals elongate in a like manner, the extent of their elon- 
gation may be a fair comparative measure of their ductility; not necessarily so, 
however, when their mode of elongating is unlike in kind. A bar of carbon sted 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



339 



habitually yields by " necking " when pulled in two, contracting greatly just 
about the place where rupture occurs, while a bar of manganese steel or of brass 
eloi^ates far more uniformly over its whole length. 

The use of manganese frogs in severe service on steam roads and for rails 
on curves of 75 feet radius or less for permanent street railway work has 
been found preferable to ordinary carbon open-hearth or Bessemer material. 

Iltanium steel, while not strictly an alloy steel, may be conveniently treated 
under this head. This 
metal, like vanadium, alu- 
minum or ffllicon, produces 
a soimder ingot, and under 
the usual practice the 
titanium goes into the slag 
and ordinarily there^ is no 
intuition of producing ti- 
tanium alloy steel. 

A progress report of 
the Baltimore and Ohio 
shows that the titanium 
r^I with .70 carbon on 
Kessler's curve is only 
wearing one-third as fast 
as the Bessemer steel 
with .50 carbon, with 
which it is compared. 
The results of dx months' 
service on a New York 
Central crossover carry- 
ing a heavy tonnage show 
that the flange wear of 
titaniimi rails was very 
much reduced as a)m- 
pared with that of ordmary Bessemer rals.* 

t The effect of titanium on steel as understood to-day is to give the 
metal greater dendty and strength. Recent tests on titanium rail steel made by 

* Iron Age, March 25, April 29, and August 5, 1909. 

t The Use of Titaoium Rail on the Baltimore & Ohio Railroad. A. W. Hiompson. Proceedings 
Am. Ry. Eng. & M. of W. Assn., Vol. 11, Part 1, 1910, and Railroad Age Gazette, November 12, 1909. 





























^S^ -~^ 














v t' 




7+7 ♦ - . , 
















10,000 
20,000 
30,000 
40,000 
50,000 
00,000 
70,000 
80,000 
90,000 
00,000 
110,000 
20,000 
30,000 
40,000 
SOflOO 
160,000 
70,000 
180,000 
90,000 
200,000 
210,000 
220,000 
230,000 
240,000 
250,000 



TEHSfLE STRENGTH- POUNDS PER SQUARE INCH 
Fio. 227. — Tensile Strength and DuctiLty of Carbon Steel and of 

Manganeee Steel. (Howe.) 
Legend; 

• = Carbon Steel. 
+ =■ Water-toughened or Suddenly Cooled Mangaocae Steel. 



>, Google 



340 



STEEL RAII£ 



Dr. Waterhouse show the elastic limit to be raised about 6000 pounds above the same 
steel to which the titaiuum alloy had not been added. In the 150 rails examined 
the titanium steel from different parts of the ingot showed a remarkable degree of 
uniformity.* 

The first experiments with titanium alloy in r^ manufacture were made 
by the Maryland Steel Company 
in November, 1907, and this was 
followed in 1908 by the Duquesne 
Works of the Carnegie Steel Com- 
pany, the Cambria Steel Com- 
pany, and the Lackawanna Steel 
Company. During the year 1909 
the process passed the experi- 
mental stage and has since been 
used in a large numbo- of rsdls and 
may be regarded as firmly fixed on 
a commercial basis. 

During the month of June, 
1908, 19 rails were rolled by the 
Maryland Steel Company, of the 
usual composition, to which was 
added 1.5 per cent titanium alloy. 
This alloy was claimed to increase 
the elastic limit, ultimate stroigth, 
and remove a large percentage of 
ELASTIC uHiT-pouHDs PER SQUARE INCH the slag; also to make ^e rail less 

Fio. 228. — EUflticity aad Ductility of Carbon Steel brittle and aVOid extreme SegfCga- 
and o[ Manganese Steel. (Howe.) , ■ ■ • i i i ■ ..i 

j^ ^j. tion and blowholes, teavmg the 

• = Carbon Steel. metal homogeneous, tough, and 

+ = W«ughened or Suddenly Cooled Mangane« fiji^g,^ed. The USe of the alloy 

resulted in a rail with a composi- 
tion high in carbon and phosphorus, which even then successfully passed the 
physical test 

The analysis made by the Maryland Steel Company of this rail shows 
the following: 

C. Mn. P. . S. Si. N. 0. 

0.701 0.92 0.086 0.048 0.079 0.004 Nil 

• See also C. V. Slocum, Mechanical Engineer, Vol. XXUl, 1909, pp. 336-337. 































1 




























































































































































































































































































































1 














































































































































































































































































A 




































































• 




■f 










































































































































































































































































































































H 




;: 




^*l 


: 


1 


■; 


1 


H 


;; 




1 


II 


1 


J 


11 



>, Google 



INTTUENCE OF DETAIL OF MANUFACTURE 341 

The addition of ferrotitanium to the ladle has an important influence on tho 
mechanical structure of the steel by acting as a flux or scavenger, and the cleansing 
effect results in increased solidity and purity of the metal 

Fenotitanium contains 10 to 15 per cent of titanium, 5 to 10 p«- cent of car- 
Ixm, and of other impurities less than 5 per cent; the balance is pure iron which 
has berai electricaUy refined. In Bessemer steel rail manufacture it is the practice 
to add from ^ to 1 per cent crushed ferrotitanium in the ladle as the steel is poured 
from the converter, and then hold the heat in the ladle about three minutes before 
pouring the ingot. Sulphur and phosphorus do not appe^ to be reduced; but in 
combining with oxygen and nitrogen, forming oxides and nitrides, titanium has an 
impcHtant action in ronovii^ these impurities, forming a stable combmation of 
them, which passes into the slag. 

The New York Central lines, prior to 1911, obtaned rmls from the Lacka- 
wanna Sted Company with -^ of 1.0 p^ cent of ferrotitanium alloy added to the 
ladle. The addition to the cost for phin Bessemer was 25 cents per ton for hold- 
ing the metal in the ladle three minutes after the ferrotitanium was added and 
$1.05 per ton for the alloy, or a total of $1.30 per ton. In 1911 ^ of 1.0 per cent 
of metaUic titanium was added to the metal and the price subsequently reduced, 
owing to a reduction in the cost of ferrotitanium. It is claimed that this small 
proportion of ferrotitanium is sufficient to remove the bulk of blowholes and segrega- 
tion usually found in Bessemer ingots and produce a clean, solid, good-wearing itul. 

The following table * gives the production of alloy rails in the years 1909 
and 1910. It appears that greater effort has been made to improve the Besse- 
mer rail by the use of alloys than the open-hearth riul. 





.909 


1«0 




36.945 
1,028 

12,287 
1,245 


195,940 
390 

4*2i6 

81 












Total 


50,505 
37,809 
12,696 


200,621 
174,822 
25,799 







Table LXXXVI gives the specification of chemical composition adopted 
as recommended practice by the American Railway Enpneering Association 
March, 1912, for carbon steel r^ls. Table LXXXVII presents the chemical 
specification adopted January 1, 1909, by the Association of American Steel 

* Railway Age Gaiett«, March 16, 1910 (daily edition), and the Iroa Age, February 23, 1911, 
p. 461. 



,y Google 



342 STEEL RAILS 

Manufacturere for standard Bessemer and open-hearth steel rails for A. S. C. E. 
sections. 

TABLE LXXXVI 

CHEMICAL COMPOSITION OF RAILS — AmMicaa Railway Eii8iiie«rii«A»«oiji»tioii 

The chemical compodtion of the st«el shall be within the following limits: 

( PROCESS 





but under Si lbs! 






0.40tro.50 
0,80 to 1.10 

0,20 

0.10 


0.45 too. 55 
0.80 to 1.10 

0,20 

0.10 




Phosphorus, not to exceed 



(Wb«o lower phoipborui ca 



larbon should be in 



OPEN-HEARTH 



bS.'SSii^i''ibV; 


S5-KI0 Iba. iDcCuiive 


Carbon 


Pet cant 
0,53 to CMS 
0.60 to 0-90 

0.20 

0.04 


PSTMlt 

0.63 to 0.76 
0-60 to 0,90 

0.20 

0.04 


Manganese. 

Silicon, not t 
Phosphorus, 




exceed 

not to exceed 



(When btgher phoqihoriiB a lued, a proper proportionate reductioa 
TABLE LXXXVII. 
CHEMTCAL COMPOSITION OF RAILS. — AMociatboo ot Ai 
BESSEMER STEEL HAILS 



si Manuliwtiirers. 





SI U 90 Pound.. 


.KoiOOP^nda. 


Carbon 


Percent, 

0.43 toO.53 

10 


0-45 to 0-55 

0.10 

0,20 
0,84 to 1.14 




Man^t.P.P 


0.80 to 1.10 





OPEN-HEARTH STEEL RAILS 





«,«■« Pound,. 


BIto too Founds. 


Crbon 


0-59 to 0.72 
0-04 
B.20 


Per cent. 

0.62 to 0-75 

0.04 

0.20 
0.fl0to0,90 




Silicon, not over.... 







Table LXXXVIII ^ves the Pennsylvania specifications revised January 
10, 1912, for 85-pound and 100-pound carbon steel r^ls. 

TABLE LXXXVIII 

CHEMICAL COMPOSITION OF RAILS,— P«inayl»»ni« Railroad Syatem. 
BESSEMER STEEL RAILS 





Lower Umlt- 


DseiredCom- 
poaition. 


Upper Limit. 


Carbon 

Manganese 


0.45 
O.RO 
0,05 


0.50 

1.00 
0,12 


0.55 

1.20 
0.20 
0,10 











>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 

OPEN-HEAHTH STEEL RAILS 







C^fi^tiooB. 




Lower Limit. 


Deaiivd Com- 

pMiUOIl. 


IIPFH Limit. 


lAwer Limit. 


pomli™. 


Upper Limit. 


Cubon 


0.70 


0.75 


0.83 

0.80 
0.20 
0,03 


0,62 


0.70 ' 








SilicoD 


0.05 


0.12 


0.05 


0.12 





















BIBLIOGRAPHY 
General 

LiTTT, B. E. V. — Mathematical relations between increases in coet and increases in durability 
in Bteel nub. 1500 w. 1911. (In RaUway Age Gazette, Vol. 51, p. 1223.) 

DcoLBT, P. H. — Ductility in rail steel. 1800 w. 1911. (In Railway Age Gaiette, Vol, 51, 
p. 289.) Paper before the American Society (or Tenting Materials. 

Considers varying composition of rail steel and its influence on the wear, 

Sandberq, Chmsteb P. — Chemical composition of steel rails and latest developments. 
4000 w. 1908. (In Bulletin of the International Railway Congress, Vol. 22, Part 1, p. 13.) 

The same. (In Engineering, Vol. 83, p. 827.) 

Discusses effect al different elements on quality and wear of metal. 

Manganese Steel 

Rolled manganese steel rail. 1200 w. 111. 190S. (In Railroad Age GaEette, Vol. 45, p. 1536.) 

Gives results of tests and showf comparative life of rails. 

Rolled manganese steel rails. 1600 w. 111. 1909. (Id Iron Age, Vol. 83, Part 2, p. 1261.) 

Discusses wear of Manard rail of the Pennsylvania Steel Company. 

Steward, H. M. — Life of manganese steel rail on curves from service tests made on the 
elevated division of the Boston Elevated Railway Company. 1500 w. 1908. (In Proceedings of 
the American Street and Interurban Railway Engineering Association, Vol. 6, p. 333.) 

The Mme. (In Electric Railway Journal, Vol. 32, p. 1196.) 

Titanium Steel 

DuDLET, p. H. — Use of ferro-titanium in Bessemer rails. 300( 
Industrial and Engineering Chemistry, Vol. 2, p. 299.) 

Gives ductility teats of ferro-titanium rails showing them to average several per cent higher 
than ordinary Bessemer rails in ductility. Believes that range of ductility can be prescribed by 
proper study of chemical composition. 

Maltttz, Ed. von. — Dcr einfluss des titans auf stahl, besonders auf schienenstahl. 6000 w. 
lU. 1909. (In Stahl und Eiscn, Vol. 29, Part 2, p. 1593.) 

The same, condensed. 1200 w. (In Iron Age, Vol. 84, Part 2, p. 1790.) 

Gives results of experiments on effect of additions of titanium lo Bessemer rail steel. 

Slocttm, Charles V. — Titanium alloy in rails and car wheels. 9000 w. LI. 1909. (In 
Proceedinga of the Railway Club ot Pittsburg, Vol. 8, p. 176.) 

Emphasizes the increased wear and soundness of titanium rails. 

Slocum, Charles V. — Use of titanium in steel for rails, car wheels, etc. 2000 w. Dl. 
1909. (In Electrochemical and Metallurgical Industry, Vol. 7, p. 128.) 

Slows the increased durability and strength of titanium stee! and its product*. 

Sprinoer, J. F. — Titanium steel. 2000 w. 111. 1911. (In Cassier's magazine, Vol. 40, p. 483). 

Considers especially its properties and importance as rail steel. 

Thompson, A. W. — Use of titanium rail on the Baltimore and Ohio Raikoad. 2500 w. 
1900. (In Canadian Engineer, Vol. 17, p. 238.) 



1910. (In Journal of 



,y Google 



344 STEEL RAIIS 

Gives properties and testa. 

WAT£RHOtiaE, G, B. — Influence of titanium on segregation in Beesemer rail steel. 3000 v. 
111. 1910, (In Proceedings of the American Society for Testing Materials, Vol. X, p. 201.) 

Results indicate that presence of titanium in rail steel lessens segregation and promotes 

unifonnity. 

30. Extraction op the Iron from Its Ore 
Before the process of reduction or " smelting " is attempted at the blast 
furnace the ore is usually subjected to some preliminary treatment.* 
The preparatory processes are: 

(a) " Grading " the ore; 

(b) Calcination or roasting; 

(c) Mixing to make up the desired proportions of ore chai^. 

The grading of the ore is not necessary at the furnace when it has already 
been properly done at the mine. When the sorting at the mines has not been 
carefully done, or when a greater number of grades than usual are required, 
sorting is also practiced at the furnace, and the ore is then distributed to the 
several bins of the stock house, which building is erected as near the furnace 
stack as possible. 

t The purpose of roasting is to remove sulphur, carbonic add, and water 
and to increase the porosity of the ore. It is accomplished in two ways, — 
by roasting the ore in a heap, or in a kiln using wood, coal, or gas for fuel. 
Fig. 229 shows the ore roasters used at the Norway furnace, Bechtelsville, Pa., 
inl883.t 

Lake Superior ores require no roasting, and for this reason very little roasting 
of the ore is necessary at the present time. The iron ores in the vicinity of 
Johnstown, which were formerly used by the Cambria Steel Works, cont^ 
high sulphur and phosphorus content. The iron content in the ore was but 
30 per cent, necessitating roasting before charging into the furnace. These 
works now use Lake Superior ores having an iron content of from 50 to 65 per 
cent, and the process of roasting is not necessary. 

Making up the furnace charge is an operation which demands both a knowl- 
edge of the chemistry of the blast furnace and of ores. The proportions of 
the charge are determined by the character of the ore, the fuel, and the flux, by 
the size and method of working the furnace, and by the character of product 
required. 

• Iron and Steel, Materials of Engineering, Thuraton, Part 2, 1909, p. 91. 
t Not«e on Iron, Rjcharda. 

t Roasting Iron-Ores, by John Birkjnbine. Trans. American Institute of Mining Engineers, 
Vol. Xil (1883-4), pp. 361-379. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 345 

The location of the plant is usually chosen according to the cost of assembling 
these matetials and getting the product to the market. Other things being 
equal, that furnace will be most economically located which is placed near the 
mines. Where the ores and fuel are widely separated, location is often deter- 
mined by the facilities for marketing the iron, and the furnace is so placed that 
the total of all the costs of transportation and of working shall be a minimum. 



VERTICAL SECTION 



FiG. 229. — Ore Roasters, Norway Furnace, 1883. {Am. Inst, of Mining Engineers.) 

If the quantities transported are 0', 0", 0'" respectively, and the cost 
of carriage is e dollars per ton, the distance for each being S', S", S'", the total 
cost (Thurston), 

K = cO'S' + cO"S" + cO"'S"', 

should, other things being equal, be made a minimum. 

The notable present tendency in the iron industry is the lower average iron 
content in the ores used. * This tendency will undoubtedly continue in the 
future as the more easily accessible portions of the richer deposits are worked out. 
As a corollary to this is the observed tendency toward a decentralization of the 

* Iron Ores of the United States. Report of the National Conservation Conuniaaion, Vol. Ill, 
p. 483, February, 1909. Government Printing Office, Waehinglon. 



,y Google 



i 

s 

s 

"■S 
3 1 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



If 

I 



>, Google 



STEEL RAII5 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 349 

iron industry, and with a decrease in the hon content of the ore used, involving 
a corresponding increase in cost of transportation per unit of iron, there will he 
an increase in the proportion of fuel which goes to the region producing the ore. 
Sir I. Lowthian Bell in 1884 stated * that while "Wages (in America) are 
high ... the geographical position of the ore and coal and of the markets 
themselves constitute obstacles of a far more insurmountable description. The 
distances over which ore is conveyed are sometimes very great; as an example, 



Fio. 233, — Steamer " Augustus B. Wolvin," 560 ft. in length, capacity about 12,000 tons. 

the produce of the Lake Superior mines is carried to Pittsburg, involving car- 
riage of 790 miles. The cost of transport on the minerals consumed for 
each ton of pig iron I have calculated t to average 10s., 9d., at the eight chief 
seats of the iron trade in Great Britain; whereas, in the United States the mean 
charge at fourteen of the large caiters is 25 s., 8 d." The introduction of improved 
methods for handling the ore in transport and the deepening of the waterways of 
the Great I^es X has in a measure overcome the adverse conditions mentioned 
above. 

* Manufacture of Iron and Steel, Bell, London, 18S4, p. 473. 

t Report to Her Majesty's Government on Iron Manufacture of the United States compared 
with that of Great Britain. 

t William Chandler, History of St. Mary's Falls Ship Canal, 1877. The Great Lakes and Our 
Commercial Supremacy, John Foord, North Am. Review, Vol. 167, p. 155. Saint Mary's Falls Canal 
Semicentennial, History of the Canal, John H. GoR, 1907. 



>, Google 



STEEL RAILS 



General View of the Dock. 



Side View of the Dock with Ore Care oa the Structure. 

— Great Northern Railway Ore Dock at AUouei Bay, Superior, Wis. 
(From ScientM Conapectua.) 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 351 

Figs. 230 to 234 illustrate some of the features of the Lake Superior ore 
industry. Figs. 230 and 231 show the 
method of mining the ore by steam 
shovels employed in northern Minne- 
sota. The shovels are large, with 
about 5-ton dippers. The amount of 
stripping required at these mines is 
often heavy, amounting in some cases 
to as much as 100 feet and costing 
from $0.25 to $0.40 per cubic yard of 
material removed from on top of the 
bed of ore. It is generally considered 
profitable to strip up to a maximum 
depth which does not exceed the thick- 
ness of the layer of ore uncovered. 

Figs. 232, 233, and 234 show the 
ore docks and the type of vessels used 
in transporting the ore. * Down to 
late in the fifties the ore product of 
Lake Superior was handled over a 
mule-tram road to Marquette, and as 
late as 1870 a 700-ton ship was an 
enormous craft, the loading of which re- 
quired two days and the unloading be- 
ing sddom accomplished in that time. 

In 1871 the largest ore barge 
carried 1050 tons, now the cargoes 
reach 14,000 tons. 165,000 tons of 
ore has been loaded into sixteen steam- 
ships in one day at the docks of the 
Duluth, Missabe and Northern Rail- 
way. The loading of the steamer " H. 
E. Corey " of 10,000 tons capacity at 
the Duluth and Iron Range Steel 

Ore Dock, at Two Harbors, Minn., _ 

was accomplished in 39 minutes.t 

• The Development of Lake Superior Iron Ores. Bacon. Trans. American Inatitute of Mining 
Engioeera, Vol. XXVII (1897), p. 341. f Scienti6c American, December II, 1909. 



,y Google 



352 STEEL RAILS 

The construction of a special type of ship of large tonnage for ore trade, 
coupled with the invention of unloading machinery of great capacity at the 

terminal ports, has brought the 
cost of transportation down to 
a very low figure. Thus, a ton 
cd ore is now hauled one hun- 
dred miles by rail from the 
most distant mines in the Lake 
Superior range to a Lake Su- 
perior port, is loaded into cars 
or into the stock pile at a Lake 
Erie port at a cost of less than 
$1.80 per ton. 

Fig. 235 presents an in- 
board profile and cross section 
of the " Wolvin," a representa- 
tive of the type of present ore 
steamers. This vessel is 560 
feet in length, 56 feet beam and 32 feet deep. The largest single cargo of ore 
carried by the " Wolvin " was 11,536 tons, a feat which she performed in 1904. 



Fig. 236. — Ten-ton Bucket of Uuloader in Hold of the " Wol- 
vin." {Scientific American.) 




Fio. 237. — General View of Ore Unloader with Steamer at the Dock. (Railroad Age Gazette.) 

The " E. H. Gary " in 1905 carried a single cargo of 12,368 tons. In 1906 the 
"J. P. Morgan" carried a single cargo of 13,272 tons of ore and in 1907 she 
carried 13,800 tons. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



il 
'I 

Mi 

1 
I 



>, Google 



354 STEEL RAILS 

Fig. 236 shows the bucket of the Hulett ore unloader. Four of these 
machines located at the docks at Conneaut, Ohio, are credited with having 
taken out of the "Wolvin" 7257 grass tons of ore in four hours and six minutes.* 
The Hulett unloaders at Gary are showing an average rate of 300 tons per hour 



Fio. 239. — Blast Furnace with Stovea and Buildings. (Thurston.) 

for each machine. On July 10, 1912, the "Morgan" discharged 10,091 tons of ore 
in three hours and ten minutes at Conneaut. This was apparently the fastest time 
ever made in unloading, but on July 24, 1912, it was surpassed when the "Wm. 
P. Palmer" was relieved of 11,044 tons in three hours and seventeen minutes, or at 
the rate of 56 tons per minute. 

* Saint Mary's Falls Canal Semicentennial, Commerce of the Great Lakes, Ralph D. Williams, 
1907, p. 201. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 865 

1 
I 



i 

I, 



1^ 



>, Google 



STEEL RA.11^ 




,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 357 

• The coal and ore docks of the Baltimore and Ohio Raih-oad at liOrain, 
Ohio, are among the largest on the Great Lakes. The machinery for unload- 
ing the steamers is the latest design of Brown hoist unloader, driven by elec- 
tricity and equipped with three grab buckets having a total capacity of 1000 
tons of ore an hour. Figs. 237 and 238 illustrate a steamer being unloaded at 
the dock. 

The large grab buckets employed scoop up from seven to ten tons of ore 
each time they are lowered into the hold of the vessel, after which they are 
hoisted and carried in over the dock on a movable girder, or ram, carried in a 
heavily braced portal frame, which is itself movable lengthwise of the dock. 

The buckets may either be dumped into a 75-ton weighing hopper, from 
which the ore is discharged directly into cars on any one of the four tracks 
spanned by the unloader, or dropped 
into the trough space, which has a 
capacity of 100,000 tons, and is sepa- 
rated by a concrete wall from the 
tracks. Once deposited in the trough, 
the ore may either remain in tempo- 
rary storage, or be conveyed to the 
larger storage space covered by the 
ore bridge. 

The combination of fast unload- 
ing plants on the dock front with 
buckets moving at high speed over a 
short travel, with a storage bridge of 
long span, carrying a larger bucket 
over the storage space, is found on 
all modem lake docks. 

The blast fimiace is shown by 
Figs. 239, 240, and 241. It is a brick ^ „,„ -,,„.. ,„, ^ 

* Fia. 242. — Top R\ggins of Blast Furnace. 

structure, usually cu-cular in section 

and built in two parts; the upper part resting on columns, while the lower 
portion rests directly on the foundation. The upper portion is sheathed with 
boiler plates. Fig. 242 shows the top rigging of a modem blast furnace. The 
charge is automatically elevated and dumped into the hopper. 

In the United States furnaces are worked up to 100 feet high. The best 
modem practice is, however, about 90 feet high, with a product of 400 to 500 

* Riulway Age Gazette, July 28, 1911, p. 178. 



,y Google 



3 

a. 

|1" 






>,Goog[e 



INFLUENCE OF DETAIL OF MANUFACTURE 359 

tons per day. The following dimensions of the Gary furnaces are typical erf 
the best practice. The blast furnaces (Fig. 243} are 88 feet in height from the 
tap hole to the top of the furnace lining, and the capacity of each is 450 tons- per 
day. Each furnace has four blast stoves. The interior diameter of the blast 
furnace is 15 feet at the hearth, 21| feet at a height 13 to 21 feet above the hearth, 
and 16 feet at the top. 



Fig. 244. — The Whitwell Hot-blast Stove. (Thuraton.) 

The earlier blast furnaces were blown with cold air, but later a hot blast 
was used with an aim to saving fuel, and the air from the blowing engines 
passed through stoves which were heated by the waste gases from the furnace. 
Fig. 244 shows the Whitwell stove and Fig. 245 a more modern stove. 

The fh^ stoves in use were of cast-iron. The gases were burned around 
and circulated among U-shaped cast-iron pipes enclosed In a fire-brick struc- 
ture. This process was continuous — a recuperative process. However, it was 



,y Google 



aw STEEL RAII5 

subject to a number of defects, among which was the bxuning out of tiie tubes, 
making it impossible to obtain more than 900° F. in the blast. This type of 
stove was followed by the fire-brick stove operated on the regenerative princi- 
ple, and by its use a hot-blast tem- 
perature of approximately 1500°F. can 
be obt^ed. 

The atmosphere is the most vari- 
able element involved in the blast- 
furnace process, which consumes air 
in large quantities. In furnaces 
using ore from the Lake Superior 
district the raw material, amounting 
to about 7200 pounds per ton of iron, 
varies in composition within 10 per 
cent, but the atmosphere, of which 
11,700 pounds are consumed per ton 
of iron, varies in its content of 
moisture from 20 to 100 per cent 
from day to day and often in the 
same day. 

Many experiments have been 
made to determine the most feasible 
method for extracting the moisture 
from the air. Various schemes for 
its direct absorption were worked out 
and in turn abandoned, and finally 
Mr. Gayley* designed and put in 
successful operation the dry-blast 
Fia.245.-JuUan Kennedy Stove. proccss wWch bears his name. This 

(Harbison- Walker Refractories Co.) • . ■ r ■ ,i 

consists m freezmg the moisture out 
of the air. The Gayley process not only reduces the cost of producing the 
pig ircm, but, which is very much more important, gives a more effective 
control of the operation and product of the furnace. 

The product was first put in operation on the Isabella furnaces of the 

* The Application of Dry-tur Blast to the Manufacture of Iron. James Gayley. Trans. American 
Institute of Mining Engineere, Vol. XXXV (1905), p. 748. 

The Application of I>ry-air Blast to the Msnufacture of Iron — Supplementary Data. Jamee 
Oayley. Ibid, Vol. XXXVI (1906), p. 315. 

Oayley'fl Invention of the Dry Blast. R. W. Raymond. Ibid, Vol. XXXIX (1908), p. 695. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



361 



Carnegie Steel Company, situated at Etna, Pa., a suburb of Pittsburg, on August 
11, 1904. The lines and dimensions of this furnace, shown in Fig. 246, repre- 
sent tiie usual construction of furnaces in the Pittsburg 
district. 

Fig. 247 shows graphically the opwation of each 
day, averaged with all the preceding days from August 1 
to S^tember 9, 1904, nclusive; the increase in output 
and reduction in coke consumption corresponding to the 
increase in burden; the varying conditions of humidity 
from day to day, which represent the average humidity 
for each twelve-hour period; and the change in humidity 
after treatment in the dry-blast apparatus. 

The materials for smelting are iron ores, limestone 
(flux), and fuel. Charcoal was first used and the iron 
from this fuel was of excellent quality on account of the 
low ash and sulphur of the charcoal and its great 
porosity. It has so little strength, however, that its use 
in the modem high furnaces is prohibited. 

Coke is now generally used. Anthracite as a blast- 
furnace fuel is inclined to decrepitate and give trouble 
from its fineness. Bituminous coal is not used, as it 
cakefi and absorbs heat for distillation of volatile consti- 
tuents. 

* At Gary, Plate XXIX, between the stock pile and 
the furnace is a hne of elevated storage bins arranged in 
two parallel row^. One row is tor coke and the other *^"^" ^^^ 
for ore and limestone. Above the bins are four tracks 
on which travel two 60-ton electric transfer cars. The 
ore is loaded into the transfer cars by the buckets of the overhead ore bridges. 
The coke and limestone are brought up ova* the bins by rail and deliver thrar 
load directiy by gravity. 

At the bottom of the bms are spouts controlled by electrically operated 
gates, and below these are tracks which run the full length of the bins. Traveling 
on these tracks are electrically operated lorries into which the ore, coke, and 
limestone are delivered fro-n the bin spouts. The lorries carry the materials 
to what are known as the " furnace skips," of which there is a p^ to each 
furnace. The skips run upon an inclined rsulway which runs db-ect from a pit 

* Scientific Americ&n, December 11, 1909. 




— Isabella Fiuv 

Carnegie Steel 
Company. (Am. Inat. 
of Mining Engra.) 



>, Google 



STEEL RAIIS 







4-- 


4 t 4 


t 


t 


A- 


- t - 


■_ 


■ t 


t 


a 




* 










i- 4 


\ ^ ■ ^i 


-' 4 




8 


: 1- : '- s 


a 


^ s 


s 


T J s 


s 


- n - ~ 


s 


- ^^' ' 


s 


J 8 


" " T 


X V ' 


' t- 


- ^^ - A ' 


s 


i' i 5 


5 


" 5 \- " ~ = 


8 .. ^ . 


- t I - s 


= -■ J - 


- 1 A = 


5 


^ n ^ •- 


:: . i*Y>oii»^<i 


3 J V = 


5 


-41 Z : 




1 I = 


=- J - 


-it A^ - 


' 7 


- i U V ' 


= C"=t^;^ 


" ti "^ J 


' 3 4 


" ^" 5 = 


° ^ 1 


4^ - 




-l\ 


i i] : 


I r 




! » 


* — r " -4 




" Jl 2 ' 


'14 -1" 


' J li 


J I' 


■ i »J 


« t. 


■■ 4 it 


Ac 1- 


i' :j1 r. 


' M 


s i ni 1 


1 -8° 1 II 

5 8 S , • 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 363 

below the transfer cars to the charging platform at the top of the blast 
furnaces. 

The operation is entirely automatic. Each trip of the skip is made in about 
axty seconds, and its average load consists of about 7000 pounds of ore, or 
6000 pounds of limestone, or 3600 pounds of coke. 

At Cambria* from the port of entry the ore is hauled to the works and there 
unloaded by a car-dumping machine, and again handled and stacked by a travel- 
ing gantry. From the stock pile it is brought back to special bin cars that are 
run into the charging house, where they take the place of the usual bins and 
discharge into hoppers that empty into the loading skips, which are hoisted to 
the charging door at the top of the furnace. 

Slag is run off either continuously or at short intervals. The iron is tapped 
regularly into a runner, through which it flows either into the molds of the pig- 
bed or else into the direct metal ladles. 

To control the kinds of pig iron produced by a furnace, we can vary the 
compOMtion of the slag, and change the burden. The burden is made heavy by 
increasing the amount of ore and flux to a charge. 

Mr. Wickhorstt gives the following as the day's burden of "A" blast fur- 
nace at the Maryland Steel Company: 

Tons. 

El Cuero ore 346-299 

Nicolaieff ore 60.676 

Sierra Morena ore 17.759 

Coke 346.875 

Limestone 82.366 

Dolomite 82.266 

The records of the ore analyses were as follows: 





nCatro. 


Ni«,b««. 


Iron, natural 

Iron, dried at 212"' F 


PBtHOt. 

58.31 
69.62 
2.03 
9.52 

-32 
.017 


Per ™i. 

65.75 
66.86 
1.66 
2.16 
.13 
.017 
trace 




Manganese 

Phosphorus 

Sulphur 



No nickel, cobalt, copper or chromium in either one. 

The metal from the blast furnace waa poured into an 85-ton receiver, from 
which it was weighed and poured into an 18-ton converter. 

The same authori^ gives the following for the blast-furnace practices at 

' Rwlway Age Gazette, August 19, 1910. 

t Report to Rail Committee, Proceedings Am. Ry. Eng. A M. of W. Assn., Vol- 12, Part 2, 1911. 



,y Google 



364 STEEL RAILS 

the Gary works. Lake Superior ore was used, reduced in blast furnaces using 
ordinary air and having the following average charge: 

Pouids, 

Coke 12,000 

Stone 5,100 

Ore 26,500 

.C 




PLAN. SECTION 

Fio. 248. —300-ton Mixer. E, FiUer; H. Pouring Spout. (Haibord and Hall.) 

A mixture of ores was used, but consisted largely of Chapin ore, showing 
analysis as follows: 

Iron - 54.63 

Aluminum (about) 2.00 

Manganese 21 

Phosphorus 059 

Sulphur trace 

Silicft(SiO,) fi.51 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



The limestone used analyzed about 
as follows: p^^i. 

54 87 



Magnesium oxide.. 
Itoa and alui ' 



.53 



The iron from the blast furnaces 
was poured into a mixer, from which 
it was weighed. 

The direct process is one by which 
the steel is made from the ore in one 
operation. In ordinary coke blast- 
furnace practice successive casts vary 
too much in Si and S to allow of taking 
the metal as it flows from the furnace 
into ladles and from them to the con- 
verter or open-hearth furnace. It is, 
therefore, poured first into large reser- 
voirs or mixers, the casts from diffa-ent 
furnaces being mixed together. 

Capt. Wm. R. Jones was the first 
to use the mixer in anything like the 
form which has now become universal 
practice. He bmlt and successfully 
operated his mixer at the Edgar Thom- 
son Works, and although it may have 
previously been used in some modified 
form the introduction and practical use 
of the mixer in connection with the Bes- 
semer process is apparentiy due to his 
^orts. 

The larger the mixer the better 
are the results obtiuned, both with re- 
spect to the purification and also in 
retaining the available heat of the 
metal. Mixers capable of holding 
600 tons of metal are now in use. 
fig. 248 shows the general arrange- 
ment of a 300-ton metal mixo- in use 
at the Cambria Iron Company's Works. 




-Ten-foot Iron Cupola, MoiylaiKl 
Steel Company. 



Joogle 



366 STEEL RAILS 

The molten metal for supplying the converter when not taken direct from 
the blast furnace or the mixer is melted in cupolas. The modem cupola is 
really a small blast furnace, as shown in Fig. 249. 

31. Conversion of the Steel 
The principal points in connection with the conversion are given below: 

1. The temperature of the conversion must be controlled; 

2. The recarbonizer must be thoroughly mixed; 

3. Time and opportunity must be allowed for the escape of gases im- 

prisoned in the molten steel. 

There are three methods of converting the metal from the blast furnace 
into steel: the Bessemer converter, the Open-hearth furnace and the Electric 
furnace. The first or Bessemer process has had an important influence upon 
railroad history. The out-growth of an attempt to make wrought iron cheaply, 
it came in just at the time when the wrought-iron rml was be^nning to demon- 
strate its unfitness to stand the increased wheel loads then coming into use. It 
perhaps is not too much to say that the Bessemer steel rail has made the modem 
railroad possible, and that without it, or its equivalent, the world's development 
would be half a century behind its present advanced position. 

The pneumatic method of steel making, generally known as the Bessemer 
process, was until the last few years the most extensively practiced and the most 
productive, by far, of all known methods of making ingot metal. 

* It has been known since the time of Cort that the aptation of molten 
cast iron in presence of oxygen will produce combustion and removal of carbon, 
and the reduction of the cast iron to the state of malleable iron or of steel. 

The pijeumatic process secures such an agitation and a very thorough inter- 
mixture of the fluid iron with the oxidizing atmosphere, by causing the latter to 
stream up through the molten mass in innumerable minute bubbles; the rapid 
combustion thus secured is sufficient to supply all heat needed, not only to retain 
the metal in a fused condition, but, also, so rapidly and so greatly to devate its 
temperature during the operation that the product, even when entirely deprived 
of carbon, remains a perfectly fluid wrought iron in the converting vessel. 
Fig. 250 shows earlier experiments of blowing air through the bath.f 
The process was invented independently by Henry Bessemer, in Great 
Britain, and by William Kelly, in the United States. 

• Iron and Sted, Materials of Engineering, Thuraton, New York, Part 2, 1909, p. 241. 

t Sketch of the Oripn of the Bessemer Process, by Sir Henry Bessemer. Trans. American 
Society o[ Mechanical Engineers, Vol. XVIII, 1897, p. 455. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE. 367 

* It is even clamed for America that it was the birthplace of the pneu- 
matic process of steel making, Kelly having begun a series of experiments based 
upon this theory as early as 1851. As Kelly, soon after Beessmer's patent 
was taken out, succeeded in showing that he had previously had similar views, 
Bessemer's patent rights in the United States became limited to certain me- 
chanical arrangements, and a lawsuit arose between the company which bought 



Fia. 250. — Early Experimento of Blowing Air through Bath. (Am. Soc. M. E.) 

Bessemer's patent rights for that country and that which took over both Kelly's 
right and Munshet's patent for taking away the red-shortness of the final 
product by the addition of spiegeleisen. This lawsuit, together with the Civil 
War, prevented the development of the Bessemer process in the United States 
up to the year 1866, when an agreement was at last entered into between the 
two companies. Both these companies had, indeed, before this time con- 
structed their experimental works; but it was only after the compromise was 
concluded between the two companies that there could be any steps taken for 
erecting Bessemer works on a larger scale. 

• Stoel; Ita History, Muiuf&cture, Properties, and Uses, J. S, Jceidb, London, 1880, p. 144. 



>, Google 



STEEL RAIl^ 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



Fia. 252. — Plan. 

American 5-toii Bessemer Ftaot. 



Pio. 253. — Sectitm on Line HR. 

(Thurston.) 



The plant shown by Fig. 251 may be taken as illustrative of an efficient 
arrangement. The general arrangement of the Bessemer plant is shown in the 
accompanying drawings. Fig. 252 represents the ground plan as designed by 



Fio. 254. — Arrangement of Converters at Maryland Steel Company. (Am. Soc. of Mech. Ei^rs.) 



>, Google 



370 STEEL RAII5 

Holly, and Fig. 253 is a section laterally on the center line of the pit surround- 
ing the converter. 

The pig iron is melted in cupolas, A, A, A, A, Fig. 252 in plan, and seen 
in elevation in the section resting upon the second floor of the converting house, 
at the right of the converters, C, C. Materials are hoisted from the lower levels 
by hydraulic elevators placed at each end of the charging floor, the one for fuel, 
the other for metal. 



Fig. 255. — IS-ton Converter, Maryland Stee! Company. (Am. Soc. of Mech. Engra.) 

Rgs. 254 and 255 show 18-ton converters at the Maryland Steel Works, 
at Sparrows Point, Md., in 1897. Fig. 256 illustrates a typical English Bessemer 
converter, while Figs. 257 and 258 show converters in operation. 

The blow generally requires about ten minutes. The molten iron is 
poured into the converter when it is lying on its side, as shown by Fig. 257, the 
converter is then placed in a vertical position and the air, compressed to about 
20 pounds per square inch, is turned on, the pressure of the blast being suf- 
fident to prevent the molten metal entering the tuyeres in the bottom of the 
converter. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



,2 -^ 

■III 

lil 



S'iS 

111 
1 a 8 

llj 




i i 5 i - 5 



S = s S 8. 1 



>, Google 



>, Google 



INifXUENCE OF DETAIL OP MANDTACTORE 



Fia. 258. — Bestwmer Converter in FuU Blast. (Am. Tech. Soc.) 



>, Google 



374 STEEL RAII5 

The molten pig iron contains a large proportion of carbon which is almost 
burned out during the blow. The combustion of this carbon increases the heat 
of the metal and the flame, shown in Fig. 258, is at first red, but rapidly be- 
comes brighter until it can hardly be looked upon by the naked eye. The sudden 
dropping of the flame after nine or ten minutes gives evidence that the carbon 
is almost burned out, and the operator turns the converter down and shuts 
off the blast. Spiegeleisen or ferromanganese is then added to recarbonize the 
metal. 

Mr. "Wickhorst* gives the following description of the process of making 
Bessemer steel at the Maryland Steel Company: The metal from the blast 
furnace was poured into an 85-ton receiver, from which it was weighed and 
poured into an 18-ton converter. In addition to the hot metal from the blast 
furnace, cupola metal was used, which ordinarily is the same metal that has 
been run into pigs and then remelted in a cupola, this being necessary when 
the Bessemer plant cannot take care of all the metal from the blast furnaces. 
In the case of this heat, two-thirds of the cupola metal was Lebanon iron. The 
converter charge was as follows: 

Pounda. 

Metal from No. 2 receiver 22,500 

Cupola metal 18,000 

Scr^ steel 1,000 

After blowing, 4300 pounds of spiegel was added to the converter and 260 pounds 
ferromanganese and 30 pounds ferrosilicon added to the ladle during the pour- 
ing. The analyses of the metal in the converter before starting to blow, and 
before the addition of scrap and of spiegel, were as follows, special samples 
bdng taken for these analyses: 





CoDTtrMc 
UmsL 


S|ri<«el 


HntAulyns. 


Carbon 


3.68 
.040 
.060 
,24 

1.33 
.35 


3.92 

.064 
trace 
3.»6 

.65 


.51 
.046 
.060 
.89 
.101 















The bade open-hearth is r^idly supplanting the Bessemer process. This is 
probably due to the supply of low phoeqijhorus ores being exhausted and the re- 
duced price of scr^, as on account of the great capacity o( the Bessemer prorass 
the open-hearth would oth«-wise have httle chance. 



' Report to Rail Coram 



i, Proceedinge Am. Ry. Eog. i 



[. of W. Assn., Vol. 12, Part 2, 1911. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 375 

The first experiments which eventually led to the development and per- 
fection of the open-heart;h process were carried on by Josiah Heath about 1845. 
Siemens began his experiments about 1861, while at the same time Martin was 
independently working on the same problem in France. The first open-hearth 
furnace introduced into the United States for the production of steel was built 
by Frederick J. Siade for Cooper, Hewitt and Company, then proprietors of the 
New Jersey Steel and Iron Company, at Trenton, N. J. 

This method is sometimes regarded as one of decarburization of cast iron 
by the addition of uncarborized metal; it must not be forgotten, however, that the 
carbon and alicon of the molten pig metal are not entirely taken out or neutralized 
by the addition of the uncarborized metal, but that the oxidizing flame from the 
gas which is burned in the furnace plays an important part. One of the principal 
objects in adding a large amount of scrap is to save time and cost of fuel in the de- 
carborizing and dealiconizmg process as well as to also save Uie lining of the furnace. 

* The furnace (B^g. 259) consists of a rectMigular bath, hearth, or basin, 
open at each end for the admission of gas and air at the ports. This hearth is 
arched by a roof from 9 inches to 12 inches in thickness. At each end of the 
furnace are two checker chambers, one for the preheating or regeneration of 
the Mr, the other of the gas. Before starting the furnace a wood fire is built 
in one set of chambers (or in the furnace) and after these have attained a dull- 
red heat the gas and air are passed through them, entering at one end of the 
furnace, are deflected downward by the direction of the ports, unite in com- 
bustion over the hearth, and the gases, the products of combustion, leave the 
furnace through the ports at the opposite end, pas^ng downward through the 
checkers or regenerative chambers, there giving up their heat to the checkers, 
thence through the flues to the stacks. 

At frequent intervals, say from 15 to 20 minutes, dependent on the quality 
and amount of fuel, charge, working of furnace, etc., the currents of gas and 
air are reversed, now entering the furnace at the opposite ends and having 
passed through the checker chambers, heated up during the previous period, 
take this stored-up heat to create a more intense flame over the bath. These 
waste gases in turn pass out through the chambers, ^ving up their heat. This 
reversal is maintained with regularity until the charge is ready to tap. 

Fig. 260 illustrates the general arrangement of an open hearth plant. 

The Talbot continuous open-hearth process employs a tilting furnace 
which may be operated at a capacity of from 20 tons upward; 100 to 150 or 
evai 200 tons are entirely practicable. The charge is run in from the cupola, 

• A Study ot the Open Hearth, HarbisoD-Walker Refractories Companj, Rttaburg, 1909. 



>, Google 



STEEL RAI15 



blast furnace, or mixer, desiliconized wholly and decarbonized lately by a 
blanket of slag rich in oxides, and reduced ultimately in the usual way. The 



[SECTIONAL PLAN OF BEQENEEtATIVE CHAMBERS 



TBANSVEHSE SECTION OF LONQITUDINAL SEOTIOM AT TRANSVERSE SECTION AT 

REGENERATIVE CHAMBERS CENTER LINE OF FURNACE CENTER UNE OF FURNACE 

Fia. 259. — Modem OpeD-heartb Furnace. (HatbiBon-Walker Refractories Co.) 

charge is run off and its place supplied by a new charge, the bottom bdng at 
no time allowed to become exposed. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 377 

Figs. 261 and 262 show tilting open-hearth furnaces. Fig. 261 shows a 
Wellman tilting open-hearth furnace and Fig. 262 was taken at the Jones and 



laughlin Steel Company's plant, where the Talbot open-hearth process is 
employed. More Talbot tilting open-hearth furnaces have been installed, both 
in this country and in England, than the Wellman furnace. 



,y Google 



378 STEEL RAILS 

The tilting furnaces do away with a great portion of the tap-hole troubles, 
the taphole being above the metal and slag lines with the furnace in the normal 
position, and it is consequently only necessary to fill the tap hole with a very 
light tamping. They also enable the melter to thoroughly drain the furnace 
bottom of any slag or metal, it being in the stationary furnace often a difficult 
matter to rabble or splash out all depressions, and any portion of the heat left 



Fia. 261, — Wellman Tilting Open-hearth Furnace. (Am. Tech. Soc.) 

in such a hole very soon tends to permeate and disintegrate the surrounding 
bottom 

The process of open-hearth steel production* at the Gary works is illus- 
trated by the following description of an open-hearth heat. This consisted of 
charging limestone and ore into a basic open-hearth furnace heated with pro- 
ducer gas and piling on scrap. The charging was started at 7.29 A.M. After 
2J hours liquid mixer metal was added, and the whole was melted down until 

• Report of Teats of Open-hearth Rails — ^Gary Works. Wickhorat. Proceedings Am. Ry. Eng- 
& M. of. W. Assn., Vol. 12, Part 2, 1911. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 379 

the tapping test showed carbon .26 per cent by fracture. During the melting 
small quantities of fluor spar were added at intervals to make the slag more 
fluid and assist in the melting. The additions of fluor spar started at 2.50 P.M. 
and amounted to about 1300 pounds total. At 3.00 P.M. a fimiace sample was 



FiQ. 262. — Pouring Steel into Ladle of Open-hearth Furnace, {Copyright, Keystone View Co.) 

taken and the phosphorus found to be .012 per cent. At 4.25 p.m. 150 pounds 
of ore was added. Mixer metal to recarbonize was added to the furnace at 
5.02 P.M. The furnace was tapped at 5.12 p.m. and ferromanganese (80 per 
cent) and ferro-silicon {50 per cent) were added to the ladle, shortly after 
tapping. 



,y Google 



STEEL RAILS 
The amounts of the various materials used were as follows: 

Limestone 25,000 

Ore (Chapin) 15,000 

Scrap steel 60,400 

Mixer metal, first charge 110,900 

Mixer metal to recarbonize 24,000 

Ferromangsnese 1,000 

Ferrosilicon 500 

A sample of the mixer metal used to charge the furnace gave the followiztg 

Carbon 3.85 

Silicon 1.73 

Manganese 1.50 

Phospborus 215 

Sulphur 030 



(The silicon in the mixer metal ordinarily averaged about 1.25 per cent 
instead of 1.73, as shown above.) 

The ferromanganese and ferrosilicon had compositions about as follows: 





Par cml ~ 
80.0 
1.0 

6.5 
.3 
.04 


Pcrrceilicon. 
.3 

54.0 
.3 
.02 
-02 















* The hot metal is tapped from the blast furnaces into 40-ton ladles, in 
which it is hauled to two 300-ton mixers. The metal is poured from the 
mixers into 60-ton charging ladles, in which it is conveyed to the open-hearth 
furnaces on electric transfer cars. From these cars the ladles are picked up 
by a 75-ton traveling crane and the metal is poured into the opai-heartJi fur- 
naces through a runner (Fig. 263). 

The fact that the Bessemer process has ah-eady passed the zenith of its 
growth is one which has now become well recognized by metallurgists generally. 

Mr. Talbot has given a very clear presentation of this subject and the 
author is indebted to him for the abstract of his paper which follows.! 

The three mam causes bringing about the supersession of the Bessemer 
process are: 1. The ever-growing scarcity of iron ores suitable either for the 
acid or baac Bessemer process; 2. The superiority of the product obtmned 

* SdentiGc American, December 11, 1009. 

t Benjamin Talbot, in the London Times Engineeiiog Supplement, Febmaiy 13, 1907. 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 381 

by the open-hearth processes of manufacture; 3. The cheapening of the pro- 
duction of the steel ingot by modaii open-hearth methods of manufacture.* 

The increasing scarcity of iron ores suitable for use in the acid Bessemer 
process is, perhaps, the most cogent of the three causes named. In the United 



Fig. 263. — Cborging FUtfonn of the Open-hearth Furnaces at Gary. (Scientific American.) 

States, apart from the Southern States and the northern portion of New York 
State, there are practically no ores at present available tor the manufacture of 
pig iron suitable for the basic Bessemer process. The rivahy is, therefore, 
between the acid Bessemer and the basic open-hearth. The former had. a 
long lead, but the growth of the open-hearth r^I manufacture has been rapid 
and in the immediate future the development of the open hearth will be out of 
all proportion to the further development of the acid Bessemer process. 

t The total annual capacity for the production of open-hearth rails in the 

• The cheapening of the cost of scrap, while not mentioned by Mr. Talbot, haa been an impor- 
tant factor. It should also be observed that open-hearth raila are subject to the same mechanical de- 
fects as Bessemer rails, and it has not yet been proved that their superiority over Bessemer rails is very 
marked; in fact, some extremely poor open-hearth rails have been made. 

t Railway Age Qaxette, March IS, 1910, daUy edition. 



,y Google 



382 STE^L RAILS 

United States is now about 1,500,000 tons, the principal nsill being that at 
Gary, which is turning out 500,000 tons; then Ensley, with 400,000 tons; Bethle- 
hem, 200,000 tons; Colorado, 200,000 tons; Lackawanna and others of smallo- 
capacity, 200,000 tons. The total capacity of all mills making open-hearth r^ls 
up to 1907 was less than 200,000 tons, and in that year production reached 
253,629 tons. In 1908 it reached 571,841 tons, and in 1909, 1,256,674 tons. 

The production * of open-hearth steel rails in 1910 was 1,715,899 tons, 
against 1,256,674 tons in 1909. The increase in 1910 over 1909 was 459,225 
tons or more than 36.5 per cent, while the increase in 1909 over 1908 was 
684,833 tons or over 119 per cent. 

The production of Bessemer steel rails in 1910 amounted to 1,917,900 tons, 
agEdnst 1,767,171 tons in 1909, an increase of 150,729 tons or over 8.5 per cent. 
Included in the total for 1910 is 68,497 tons of re-rolled rails. In 1911 the produc- 
tion of open-hearth steel rails was less than in the previous year, but on account 
of the smaller tonnage of Bessemer steel rails rolled than in 1910 more r^ls were 
made from open-hearth steel than Bessonenf 

Table LXXXIX gives the production from 1907 to 1911. 

TABLE LXXXIX. - 





1M7, 


im 


igoe. 


i.io. 


laii. 




Ton, 

3,380,025 

253,629 

3,633,654 


1,349,153 

671,811 

1,920,994 


Tons. 

1,767.171 
1,256,674 
3.023,845 


T«B. 

1,917,900 

1,715,899 
3,633,799 




Open-hearth rails 


1,676,923 
2,815,556 





X In any consideration as to the future of the acid Bessemer process in 
the United States a thorough understanding of the ore situation is essential. 
As is well known, the Lake Superior, particularly the Mesaba, ores are the 
mainstay of pig-iron production in the north. Each year this ore becomes 
leaner, and there is a difficulty in keeping the phosphorus content of the pig 
iron manufactured from it below the .1 per cent of phosphorus which is the 
standard for Bessemer steel in the United States. Sted made from such pig 
is dangerously near the limit of safety for some purposes, when it is manu- 
factured by the acid Bessemer process, but when treated in any form of the 
basic open-hearth process such pig produces a metal of most excellent quahty, 
with phosphorus, when desired, down to .02 per cent, or even less. The carbon 
content of the steel can also, in the lattw class of process, be varied within v&y 

* The Iron Age, Febni&ry 23, 1911, p. 461. 

t Railway Age Gazette, July 19, 1912, p. 125. 

t Benjamin Talbot, London Times Engjueeriog Supplement, Februaiy 13, 1907, cUuIy edition. 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 383 

wide limits, while it is not so easy to produce .6 to .7 per cent carbon steel 
in the acid Bessemer process, and even if made, steel with such high carbon 
and with .1 per cent pho^honis, or thereabouts, is certmly not a material 
that should be looked upon with favor for rail purposes. 

All the facta point in one direction. The Bessema- process, while the 
actual cost of conver^on, apart from the question of waste, is perhaps the 
cheapest, is yet one which requires, ather for acid or basic working, a spedal 
quality of pig iron, — a quality which is ever tending to become dearer. The 
waste of metal in the Bessemer must of neces^ty be higher than in any form 
of the open-hearth process, and this fact accentuates the importance of the ques- 
tion of the cost of the pig iron; the higher the price, the greater the cost due 
to waste. Roughly sp^ildng the loss in a Bessemer is from 8 to 10 per cent and 
in the open-hearth from 3 to 6 per cent 

The margin for economies in the Bessemer process is less than any which 
can be made in the bade open-hearth process. Unless a radical change is 
effected in the operation of the Bessemer furnace, only small further savings 
appear possible. It is true that in some Bessemer the blowing power is still 
raised by steam obtsdned from coal burnt imder boil«^, but even in cases in 
which the blowing power is obtained from sxuplus blast-fumace gas, products 
are absorbed which could otherwise be economically and usefully employed in 
creating power for other purposes, if the open-hearth process were employed. 

* The electric furnace is rapidly coming into use as an important factor 
in steel manufacture, and where water power is abundant and fuel is scarce 
it is extending the boundaries which have for a long time confined the iron 
and steel districts. Experience with the electric furnace in foreign countries 
has shown that it will purify the metal to a larger ejrtent than the gas furnace 
or the Bessemer converter, and it is proposed to use it as an adjunct to the 
ordinary processes of steel manufacture for the purpose of reducing the amount 
of phosphorus and sulphur and to deoxidize the bath. 

After a careful investigation by its metallurgists, the United States Steel 
Corporation has decided to use a 15-ton Hdroult electric furnace at the South 
Chicago works. Three-phase alternating current will be used, and it is pro- 
posed to refine the blown metal from the Bessemer converter in the H^oult 
furnace, reducing the percentage of phosphorus and sulphur, and to use the 
product for high-grade steel rails. The capacity of one furnace is sufficient for 
the production of 500 tons of steel in 24 hours. 

* Railroad Age GaKtte, March 12, 1909, end Compoeition and Heat Treatment of Steel, E. F. 
Lake, 1910, pp. 42-63. 



,y Google 



381 STEEL RAIIS 

The Hfiroult steel-refining furnace is of the crucible type with a tilting rack. 
The heating is initially effected by means of the electric arc which forms between 
the surface of the slagging materials which float on the metal bath and the two 



k 
I 

i 
i 

3 ^ 
1 I 



massive carbon electrodes which are suspended above it. The impurities of 
the steel are removed by renewing the slag. The refining operation thus be- 
comes a " washing out " one. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 385 

The lining is the same as the basic open-hearth and the phosphorus is first 
reduced and then the sulphur. Recarbonizing is done in the bath by addii^ 
crushed electrodes which are 98 per cent pure 
carbon. 

Fig. 264 shows a transverse section 
through the pouring spout of the Heroult 
furnace at La Praz, and also longitudinal sec- 
ti(»is of the furnace through the roof. The 
eJectrodes, of which there are two passing 
through the roof, are shown at E on the figure. 
An. alternating current at 110 volts is used. 

In the Stassano arc fiunace the neces- 
sary heat is obt^ned by direct radiation 
from the arc. It is shown in Fig. 265. 
Vaiious experiments have been made in this 
furnace to produce steel direct from the ore, 
but, owing to the difficulty of controlling the 
composition of the slags with average ores, 
the production of steel of any required grade 
is far from easy.* 

t More than 5000 tons of rails have been 
made from steel from the electric furnace 
at the Roechling Iron and Steel Company, 
Voelklingen, Germany. The furnace is of a 
spedal combination electrode and induction 
type, known as the Roechling Rodenhauser, 
and takes three-phase current at 25 periods. 
The pig iron is blown in a basdc lined Bes- 
semer converter, then transferred to the 
electric furnace for refining at an expenditure ^"'- 2«6.-sta««n^^ Electric Fumaoe. 

of power of 125 kilowatt hours per ton. Re- The fumace is inclined to the veridical and 

cently some tests have been published, made ""'"^ ^^ '^^ '"^''^"i^ «»«>"" '>^'°" 
November 27, 1908. The analysis of the three pieces then tested was as follows : 

„ , PbT (Ml. 

Carbon . , , , 75 

Silicon 0. 10 

Manoaneae 57 

Sulphur 0.044 

Phosphorus 0.023 

The rails were of flange section, 82.65 pounds to the yard. 

■ Steel, Harbord and Hall, London, 1911, pp. 261-283. t Railroad Age Gazette, July 2, 1909. 



,y Google 



STEEL RAILS 



Physical tests were made on these rails; the pieces have a length between 
punch marks of 7.94 inches and a diameter of almost 1.0 inch, being .975, 
.966, .984 respectively. These results are given below; 



.Vumbsr. 


UlUnute 


Elontatun. 
Per'wnt. 

12.25 
12.25 
13.80 


ReductioBof 

21.00 
12.60 
20.40 


1 

2 


123,341 
126,172 
122,765 





They show excellent ductility, in conjunction with tenacity. 



Fia. 260. — Hoechling-Rodenhauser Furnace. (Lake.) 

The latest development * in connection with the furnace is its operation 
by a three-phase current, with a frequency of 50 periods for a 15-ton furnace. 
Fig. 266 shows this furnace in sectional elevation and plan. It is claimed that 
a special feature of the furnace is the rotation of the charge due to the presence 
of a rotatory field, as in an induction motor, which insures an automatic circu- 
lation in the bath. The furnace is essentially a transformer with a primary 
winding A round both iron cores H of the transformer. The secondaries are 
two in number; one is the molten bath in the form of an 8, the channel D 

• Steel by Harbord and Hall, London, 1911, pp 261-283, and The Report of the Canadian Com- 
mission Hppoinleil to investigate the Different Eleclro-Thermic ProceBses for the Smelling of Iron 
Ores and the Manufacture of Steel in Operation in Europe. 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



387 



between the two cores being very broad. The other secondary is the copper 
winding B, which is connected with the metal plate E. 

The dectric furnaces, just described, show the three distinct types which 
are daiming the serious attention of metallurgists. In the present state of 




Fia. 266. — Roeehliog-Rodenhauser Furnace. (Lake). (Continued). 

development of the electric furnace it cannot compete as regards cost of pro- 
duction with a modem large open-hearth furnace for the manufacture of rails, 
and it is only when a superior quality is reqmred that it can be employed. 

The steel from different parts of the ingot shows a very regular compo- 
sition and is remarkably free from segregation of impurities. The mechanical 
properties are extremely good. 



,y Google 



388 STEEL RAII5 

Some recent experiments in the United States show that steel made in the 
electric furnace has a greater density, and within the range of .08 to .75 carbon, 
shows 10 per cent greater strength than open-hearth steel of the same chemical 
composition. 

* The duplex process is a combination of the Bessemer and the open- 
hearth, and is particularly applicable to pig iron containing too high alicon 
for advantageous working in either basic Bessemer or basic open-hearth. 

In the acid Bessemer conva-ter the preliminary blast removes the silicon, 
together with a considerable portion of the manganese and a certsun amount 
of the carbon. The desiliconized metal is then transferred to the basic open- 
hearth, where the phosphorus and the reminder of the carbon is eliminated 
in accordance with the usual practice. 

t The Jones and Laughlin Steel Company has been experimenting for some 
time with the duplex process in its present Pittsburg plant, with the idea of 
using a portion of its Bessemer capacity for preparing metal for its open-hearth 
furnaces, thus decreasing its output of Bessemer steel and correspondingly 
increasing the open-hearth output. The Maryland Steel Company has com- 
pleted five open-hearth furnaces and is using a considerable portion of its 
Bessemer capacity to duplex with the new open-hearth furnaces. 

These moves in the direction of duplexing represent a distinct dedre to 
find a new use for Bessemer capacity because there Ls not sufficient employment 
for it in its old function. 

The casting ladle, or the ladle which receives the finished steel for casting 
into molds, is shown in Fig. 267. If slag is allowed to pass into the ingot molds 
with the steel the latter is liable to be spoiled, and in consequence the steel can- 
not be poured from a lip into the molds, but has to be tapped or teemed from 
a hole in the bottom of the ladle. 

The time allowed after the conversdon of the steel and when it is held 
in the converter or casting ladle exercises considerable influence upon the 
finished product. The thorough mixing of the recarbonizer, and the oppor- 
tunity for the impurities to separate from the metal and the gas to escape from 
the molten steel are of importance. Dr. P. H. Dudley requires a definite in- 
terval of time between the additions of the spiegel and the teeming of the steel. 

He says:t " Restricting the ingots to three-rail lengths and holding the steel 
three minutes afta* recarbonizing, in connection with the dry blast at South 

• A Study of the Open-hearth, by Harbison-Walker Refractories Company. 

t Railway Age Gazette, March 18, 1910, daily edition. 

t ProceedingB Americaa Society for Testing Materials, Vol. VllI, 1908, p. 112. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



389 



Chicago, shows a marked reduction in seams and cracks in tiie bases of rails. 
In a lot of 2500 tons of these rails hardly a trace of seam has been found." 

The dry blast referred to is the Gayley process of furnishing sar, practically 
free from aqueous vapor, to the converters while blowing the charge. This 




Fra. 367. — Details of Casting Ladle. (Hufoord and Hall.) 
A, Goose-neck; B, atoppra rod; L, sliding bar carrying gooHe-neck; M, M', M', bolts for attaching 
lever, P; P, lever bar; Q', Q*, brackets on ladle to guide sliding bar, L; K, screw bolt for holding 
sliding bar rigidly in position previous to teeming; F, fire-clay sleeves threaded on steel stopper rod, B; 
Z, teeming nozzle; E, fire-clay stopper head; G, nozzle box; H, trunnion; C, C',C, C*, cotter pins; S, 
forged bead on sliding bar through which end of goose-neck is passed, and is fixed by cotter pin, C'. 

decreases the amount of iron oxide in the bath which von Maltitz claims to be 
a principal agent in producing blowholes. 

That time should be ^ven for the necessary chemical reaction after the 
addiUon of the recarbonizer and before casting the metal into ingots, has been 
known for at least thirty years. It was known that such time was also impor- 
tant for the escape of the occluded gases, and the value of this latter knowledge 
was manifested by the several devices for accelerating their escape which, years 
ago, were either propcsed or actually used. These ranged from the thrusting 
into the ladle full of molten steel a wooden pole, or placing in the ladle before 
it received the steel from the converter pieces of wood saturated with kerosene, 
to more elaborate devices, for agitating the steel while in the casting ladle by 



,y Google 



390 STEEL RAIIB 

power-driven, refractorily protected screws. It was the practice at some of 
the Bessemer works to j^n put on the blast after the introduction of the 
recarbonizer, and then to partially turn the vessel up, thus a^tating the charge. 

Mr. Robert Forsyth sought to accomplish the desired results through his 
transfaring ladle arrangement, by which liie ladle after receiving the steel from 
the converter was transferred by a hydraulic ram from the receiving crane to 
the ladle or casting crane. This he did when remodeling the Union Steel Plant, 
Chicago, about 1886. Later he put the same arrangement in the South Works 
of the Illinois Steel Company. Later Mr. \Tilliam R. Widker carried this further 
by pouring the steel over the top of the recdving ladle into the casting ladle 
through a nozzle in the bottom of which it was cast in the usual way. 

Prof. Henry Fay * has observed the moon-shaped fractures in the base and 
the thermal cracks in the head of the rail which he beheves to be gener^ly 
found along a streak of manganese sulphide, extending in the direction of the 
rolling. The cause of the manganese sulphide being in the steel he infers is 
probably the lack of time given for the steel to purify itself after the addition 
of the recarbonizer. When the spiegel or ferromanganese is added to the 
bath, the manganese combines with the sulphur, and, given time enough, man- 
ganese sulphide, having a spedfic gravity less than that of steel, will rise to the 
surface. 

Manganese sulphide melts at 1162° C. Its specific gravity is 3.96, or 
about half that of steel. It is a glassy, hard, and extremely britUe material. 
The steel from which the r^I is made sohdifies at about 1450° C, and the 
manganese sulphide will not solidify imtil it reaches 1162° C. Therefore, 
the manganese sulphide is in a fluid state some time after the steel sohdifies. 
If the rolling of the rail starts, we will say, at a temperature above 1162° C, 
this material will be rolled out in thin strips in the direction of rolling. It is 
plastic below the melting point and it is capable of being rolled out while in the 
plastic condition into long, thin strips. 

He states: f " That manganese sulphide when existing in certain forms is a 
harmful constituent of steels can no longer be doubted. The remedy seems to 
be a very simple one. Specifications should be so drawn as to Umit the amount 
of sulphur in the steel. At the present time most of the specifications do not 
even mention sulphur. Having done this, the next step is to allow the metal 
to stand a longo- time after the addition of the ferromanganese. With the 

* JounuJ AsaociatiOQ of En^neeriag Societies, July, 1908, p. 2S. 

t A Microscopic Investig&tion of Broken Steel Kaila; Manganese Sulphide s8 a. Source of Danger. 
Fay, Vol. VIH (1908), ProcepdinRs American Roniety for Testing Materials. Further Investigationa 
of Broken Steel RaUs, Fay and Winl, ibid., Vol. IX. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 391 

specific gravity of manganese sulphide 3.966 and steel 6.82, it should rise to the 
surface and be skimmed off with the slag if given sufficient time. Usu^y this 
time interval between charging of the ferromanganese and the pouring of the 
ingot is very short. The desire of the manufacturer to increase his output 
has led him to cut down this interval to the shortest possible Kmit, with the 
natural consequence of a large number of defective riuls. A longer time interval 
will allow the metal to purify itself." 

Mr. E. von Maltitz* found that where recarbomzing is done in the ladle 
and insufficient time allowed for the complete reduction of the iron oxide in 
the bath, an excessive number of gas holes may be formed. The presence 
of gas seams tends to cause unsound metal. Mr. Robert Job has pointed out 
that in nearly every case of f^lure due to crushed heads the section shows 
marked unsoundness, and the vertical flaws of the gas seams weaken the head 
greatly. 

The fact should be empha^zed that it is not alone sulphide of manganese 
which is a source of danger, but other forms of slag are also to be looked upon 
with suspicion.t These may be summed up as follows: 

1. Excessive slag (manganese sulphide, silicate, etc.). 

2. Segregation of slag concentric with the section. 

3. Remnants of slag in the large split portion of the head. 

4. Slag in those areas where flow of metal has occurred, or where mi^o- 

scopic cracks have developed. 

The rails illustrated in Figs. 268 and 269 show the effect of unsound metal in 
the head of the rdl. Referring to Fig. 268, view 1 shows a section of the rail 
which has been polished and etched with add. It shows some segregation and 
flow of metal, as expected from the rolling. The rail was slivered, breaking off 
the right of the head as indicated. This view shou% also the portions into which 
this section was cut and the marks by which they are identified ; these were in 
part polished and examined microscopically for defects as shown in the other 
views of this figure. 

View 2 shows the grmn size at the center of the head and view 3 shows the 
fin«- grain near the surface and the distortion of the grains by whed action. 
The photograph is taken at the end of a crack, and shows, besides this, some 

* Blowholes in Steel Ingots. E. von M&ItitE. Tr&na. American Institute of Mining Engi- 
neen. Vol. XXXVin (1907), p, 412-447. 

t IroD and Steet MsgaEine, August, 190S (Job); and Journal of the Iron and Steel Institute, 
p. 301, 1905 (CfVtuD Howorth). 



,y Google 



392 STEEL RAILS 

small cavities, which were more noticeable before etching, and which might have 
been minute oxide pits or pockets. 

View 4 shows the metal, which is the white ground mass, to be badly con- 
taminated by slag, which is extended longitudinally by the process of rolling. 
These slag lines were found to some extent over the whole surface of this piece, 
but were worst in the neighborhood of the point indicated by the dot By, view 1. 
This is shown on the picture, which was taken at the edge of the localized portion. 



"WW 1. View 2. CroBB Section at point C„ Mag. 100, 

68,000 graiiiB per eq. in. 
FiQ. 268, — Crushed Head. 

Fig. 269 illustrates another example of a crushed head due to unsound 
metal. View 1 shows a section of the head taken at the point of greatest dis- 
tortion. The cavity in the top of this view is a drilled hole. On one side of 
the head a cavity which did not show on the surface, but indicated marked 
breaking down of the metal, was revealed. The metal, however, is more uiu- 
form throughout this rdl than was the case in the ral of the preceding figure. 

^lew 2 shows the gran at the center of the head ; view 3, Kke view 3 of 
Rg. 268, is taken at the end of the crack. It shows the finer grain and distor- 
tion of the same, and shows as well the further distortion of the metal at the 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 393 

end of the crack as a sort of tearing action. The end of the crack is at the 
comer of the picture; the furUier direction of progress of the failure is shown 
by the black defects extending across the photograph. 

A longitudinal section of this rail made on portion G showed slag lines, as 
in the rail of Kg. 268, somewhat most abundant at the point q, though the num- 
ber was not so great as in the rail of the preceding figure. 

I ' >; ' ' it' 



i 






View 3. Cross Section of point H„, Mag. 100. View 4. Lon^tudinal Section on Top of Portion B, 

Mag. 50, Unetohed. 
Fio. 268. — Crushed Head. (Continued.) 

• The deleterious influence of slag inclosures in steel has perhaps escaped 
attention to some extent owing to the fact that in ordinary tensile tests, taken 
in a direction parallel to that of rolling, these inclosures only occupy a very 
small proportion of cross-sectional area and possess a tapered shape which allows 
of gradual distribution of the stresses imposed on the material. If, however, we 
consider the case of transverse stresses, or of shock or vibration, it will be seen 
that these inclosures will be fractured as soon as the metal undergoes any material 
ddormation, and then each such inclosure practically represents an internal 

* " Slag Inclosures " in Steel, by Walter Rosenhain. International Aasociatioa for Testing 
MalerialB, Sth Congress, Copenhagen, 1909. McGraw-Hill Book Company, New York. 



>, Google 



STEEL RAILS 



8 

i 
^. .1 

d.s K 



IS £ 

o 

I 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 395 

fissure which is ready to extend — and actually does extend — in any direction 
compatible with the applied stresses. 



Fia. 270. — Teeming Ingots at Beeaemer Converter. (CopjTight, Keystone View Co,) 

32. The Ingot 
The principal points in connection with this jjart of the process are as follows: 

1. Care must be exercised in casting the ingot. 

2. The ingot must be in a perpendicular position until the interior has 

had time to solidify. 

3. The steel must not be overheated in the heating furnace or soaking pits. 

4. Defective material from the top of the ingot must be excluded from the 

finished product. 



,y Google 



396 STEEL RAII^ 

From the casting ladles the steel is fun into cast-iron ingot molds located 
on cars, as illustrated in Figs. 270 and 271. Table XC presents data lowing 
the time required to pour the steel at different mills. After solidifying, the 



Fig. 271. — Teeming Ingots at Open-hearth Furnace. (Copyright, Keystone View Co.) 

ingot mold cars are run under the stripper, shown in Fig. 272, from which hooks 
are lowered and engage the lugs on either side of the mOld and lift it off the 
ingot. 

The ingot is then taken up by a traveling crane and conveyed to the re- 
heating furnaces or soaking pits, shown by Figs. 273 and 274, to allow the tem- 
perature in all parts of the ingot to become equalized before rolling. 



,y Google 



FiQ. 272. — Stripping the Mold from Ingots. (Lake.) 



Fia. 273, — Soakbg Pita — Gary. (Scieotific American.) 397 

DignzcdDv Google 



STEEL RAII^ 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 
TABLE XC. — TEEMING PRACTICE AT AMERICAN RAIL MILLS 

(CompiM by CommittB on Rul. Am. Ry. Emt. Asm.. ieO», ud revised by tbe author IS12) 



Rail UiU. 


LDCslion. 


^Om Meft. 


TimeinPourin* 
into Ingot Mold». 


Tons Poured 
per MiDule. 




CanadianSoo,Can. | 

Bethlehem, Pa 

Johnstown, Pa 

Braddock, Pa 

South Chicago, 111. 
Gary Ind. 


Tons. 
Bess. 5 
O.H. 40 
50 
12 
15 
13 
80 
12 
15 
100 


4 min, 
20 min. 
20 min. 

6 min. 30 sec. 

7 min. 15 sec. 
9 to 11 min. 

20 min. 
3 min. 15 sec. 
6 min. 
t20 to 25 min. 


1.25 








2,15 
















Buffalo. N.Y 

Sparrows Point, Md 
Birmingham, Ala... 








Tenn. Coal, Iron & R.R. Co 


t4to5 



(*). whieb waa obtainad direct trtna TDuufactnnn by tbs 



t Compiled by author. 



The unsoundness of the ingot results from several causes: 

1. A funnel-shaped cavity or pipe at the top of the ingot. 

2. Dispersed cavities or blowholes throughout the ingot. 

3. Segregation of the impurities of the steel, as silicon, phosphorus, man- 

ganese, etc., from the mass of the metal and their concentration 
in different parts of the ingot. 

The pipe is due to the contraction of the interior of the mass after the out- 
side has set. After molten steel has been cast into an iron mold, the metal in 
contact with the bottom and the £ddes bepns first to solidify. 
After a relatively short while the top of the ingot, which is 
exposed to the cooling action of the air, also becomes soUd 
and the ingot now consists of a rigid metallic shell holding 
a mass of molten steel, as shown in Fig. 275. As the cooUng 
proceeds this solid shell increases in thickness; but since steel, 
like most substances, undergoes a considerable contraction in 
passing from the liquid to the sohd state, the mass of metal 
which when hquid was sufficient to fill the space within the 
solid shell will, after it has in turn solidified, be unable to fill Fm. 275. — Forma- 
it and a cavity must necessarily be formed in the upper part *"*" "^ ^'p* ^ 
of the ingot. 

The piping of ingots has been known for a number of years.* Robert 
Forsyth at the Union Steel Works in 1888 demonstrated the relation of the 
length of the pipe to the position of the ingot while its interior metal was solidi- 



* The Manufacture of Beesemer Steels by R. W. Hunt, Lecttu* delivered at The PrankLn In- 
stitute, January 21, 1889. See Journal of the Franklin Inst., May. 1888. 



,y Google 



STEEL RAILS 



fying, by breaking a number of ingots which had been differently handled — 
some placed in a horizontal position as soon as possible after being cast, and 



Fia.276. — Section of Ingot, 17 ins. Squareat Top, 19 Fio. 277. ^ Bloom from an Ingot of some 

ins. Squ&re at Base, and 50.5 ina. long. Containing Heat and of same Size aa Fig. 276, ^ow- 

Cavity of 128 cubic inches. (Am. Inet. of Mining ing Reduction of Cavity. (Am. Inst, of 

Engrs.) Mining Engra.) 

others so placed at varying intervals up to having been kept vertically until 
all of the steel was thoroughly set. 

The best modem practice is to charge the hot ingots into the reheating 
furnaces to equalize their heat for blooming as soon as possible after they are 
teemed, stripped, and weighed. 

An interesting experiment was tried by Dr. P. H. Dudl^ to determine 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 401 

the relation between the pipe in an ingot which had been allowed to get cold 
and one which had been promptly charged into the reheating furnace.* 

Fig. 276 is a photograph of a three-rail ingot, for lOO-pound rails, teemed in 
a mold 19 inches square on the base, 17 inches square on the top, and 66 inches 
long. The ingot, poured 50.5 inches long, was well deoxidized, and therefore 
had a large cavity. The ingot had a volume of 7.4 cubic feet, inclosing a shrink- 
age cavity of about 128 cubic inches, practically 1 per cent of its volume. This 
is a larger percentage than would be found in nul steel not so well deoxidized, 
or which cont^ned numerous blowholes. 

Fig. 277 is a photograph of the bloom of an ingot of the same heat and 
length, cut for a 9 per cent mill discard. The ingot, after stripping and a subse- 
quent ride of 500 feet, was charged directly into the reheating furnace without 
allowing the temperature to fall below the recalescence point, while the bulk 
of the steel was severe hundred degrees above, and in about 2 hours the ingot 
was drawn and bloomed. The cavity was small and less than one-tenth of that 
of the cold ingot of the same heat. 

Blowholes generally form in the upper half of the ingot, which is permeated 
by honeycombs or dispersed cavities, due to the liberation of imprisoned gases, 
principally hydrogen, as well as nitrogen and carbon monoxide. These gases 
are absorbed, dissolved, or occluded in the molten steel, but are wholly or partially 
evolved and collect into bubbles when the metal bepns to solidify. These 
bubbles are generally more numerous towards the side of the ingot. 

The evolution of the gases in the mold seems probably to be due to two 
causes; First, by the reduction of the temperature, the solvent power of the 
steel for the gases is lowered and, consequently, cert^n proportions of the gas 
are libOTated; and, second, an evolution of carbon monoxide (CO) or carbon 
dioxide (COi), due to chemical action. 

t According to Howe, blowholes may be lessened or even wholly prevented 
by adding to the molten metal shortly before it solidifies either sihcon or 
aluminium, or both. An addition of manganese has a like effect.^ These ad- 
ditions seem to act in part by deoxidizing the minute quantity of iron oxide and 
carbonic oxide present, in part by increaang the solvent power of the metal for 
gas, so that ev^i after freezing it can ret^n in solution the gas which it had 
dissolved when molten. But, since preventing blowholes increases the volume 
of the pipe, it is often better to allow them to form, but to control their posi- 

* DbcusnoQ of Henry M. Howe's paper on Pipii^ and Segregatioa in Steel logota, Trans. Ameri- 
can Institute of Mining Engineers, Vol. XL (1909), pp. 821-830. 

t Iron,Steel,andOtber Alloys, Howe, 1903, pp. 369-372. Contains record of Brinell's experiments, 
t Titanium deozidiies the steel in a very marked manner, as shown in Fig. 285. 



>, Google 



402 STEEL RAIIS 

tion, so that they shall be deep-seated. In case of steel which is to be foiled 
or rolled, this is done chiefly by casting the steel at a relatively low tempera- 
ture, and by limiting the quantity of manganese and Edlicon which it contains. 

Brinell finds that, for the conditions which are normal at his worics at 
Fagertsa, Sweden, if the sum of the percentage of manganese plus 5.2 times 
that of the sihcon is as great as 2.05, the steel will be so completely free from 
blowholes as to have an undedrably large pipe. If this sum is 1.66, there will 
be just that small quantity of minute, hardly visible blowholes which, while 
sufficient to prevent any serious pipe, is yet harmless. If this sum is less than 
1.66, blowholes will occur and will be injuriously near the surface unless this 
sum is reduced to .28. He thus finds that this sum should be either about 
1.66, so that the quantity of blowholes shall be harmlessly small, or as low as 
.28, so that they shall be harmlessly deep-seated. 

These numbers must be varied with the variations in other conditions. 
In general, either a higher casting temperature, or a smaller cross section of 
the ingots, or the use of hot or that of thin-walled molds, calls for a smallCT 
quantity of silicon and manganese. 

Brinell also finds that an addition of .0184 per cent of duminum is ap- 
proximately equivalent to the presence of manganese and silicon in the pro- 
portions Mn + 5.21 Si = 1.66 per cent; i.e., it imaided gives rise to structure 
B (Table XCI). Naturally, little or none of this aluminum remains in the 
steel. It oxidizes to alumina, which rises to the surface of the molten metal, 
or is found lining the walls of the pipe. 

Table XCI and Eigs. 278 to 284 give some of Mr. Brinell's results. 

TABLE XCI. — INFLUENCE OF MANGANESE AND SILICON UPON BLOWHOLES 
AND PIPES 



Quality of tbe 3t«l u Resnrdt Blowholat 



Cast too hot 
Cast too cold < 



: No blowholes, but asmallpipe. 

, No visible blowholes, no pipe. 

j External blowholes, no pipe. 

. Fewer blowholes and somewhat 
deeper seated. 

: The blowholes are very deep-seated. 
Many external blowholes and a pipe 
Many blowholes, both external atid 

I internal. 



the 



Injured by the pipe. 
Juxt compact enough; excellent. 
Injured by the external blowholi 
Blowholes still harmfully 

surface. 
Excellent. 

Injured by the external blowholes. 
[injured by the external blowholee. 



The structures and H are those induced by too high and too low a casting 
temperature respectively. The steel which here has structure would, if cast 
at a normal temperature, have had structure A. It was thought that the reason 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 403 



Fia. 278. — Structure A. — Brinell'e Teats. Fia. 279. — Structure B, — Brinell'a Testa, 



Fia. 280. — Structure C. — Brinell's Tests. Fia. 281 . — Structure D, — Brinell's Testa. 



FiQ. 282. — Structure E. — Brinell's Teste. 



Fio. 283. — Structure 0. — BriDell'a Teeta. Fia. 284. — Structure H. — Brinell's Tests. 



,y Google 



404 STEEL RAII^ 

why the excessivdy high temperature caused these external blowholes was that it 
caused the carbon of the molten steel to react on the iron oxide on the surface 
of the mold, with the formation of carbonic oxide gas, which itself forms these 
blowholes. 

Von Maltitz* gives the following as the means for the prevention of blow- 
holes in sted ingots: 

1. Medium temperature of the heat during the last period of the process 
in the converter or open-hearth. 

2. Careful avoidance of overblowing or overordng of the heat; careful 
boiling out of the last jwrtion of ore added to the bath. 

3. A finishing slag not too rich in oxygen and having the proper degree 
of fluidity. 

4. The destruction, by stirring the heat before tapping, of the ferrous 
oxide formed. 

5. Addition of sufficient deoxidizing material to the heat, and the allowance 
of sufficient time for the complete separation of the manganese protoxide, 
silicate of manganese or alumina thus formed, into the slag. 

t Howe nmntmns that the gas contained in the blowholes is partially 
absorbed and the blowhole walls to some extent weldable during the process 
of rolling. This action probably is less favorable in direct rolling of rails (i.e., 
rolling direct from ingot to rml at a single heat) than in reheating practice, in 
which the bloom into which the ingot is rolled is held in a special bloom-heating 
furnace before rolling into a rail. 

Segregation is one of the important questions before the steel maker. It 
is, ther^ore, natural that for many years it should have engaged the attention 
of iron and steel metallurgists in different countries, and have given rise to an 
important literature. 

Steel contains different impurities, as silicides, phosphides, carbides, sul- 
phides, etc., whose freezing or soUdifying points vary, and all have a lower 
melting point than the metallic iron, consequentiy those having the lowest 
melting point will tend to gradually segregate from the iron and concentrate 
in the hottest part of the ingot. The top and center of the ingot always con- 
tains the larger proportions of impurities. 

All steeU do not necessarily exhibit excessive concentration of impurities. 

* Blowholes in Steel Ingota, E. voa MaJtiti, Trana. American loatitute ot Miiung EngineecB, 
Yd. XXXVni (1907), p. 445. 

t The Welding of Blowholes in Sted, Heniy M. Hone, Proceedings AmericsD Society for Tesl^ 
ing Materials, Vol. X, 1910, p. 168. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 405 

The highly segregated portions of an ingot are often very small isolated areas 
in the interior of the mass. 

It is highly probable that a large part of the segregates in steel ingots is 
directiy traceable to the formation of blowholes. The pressure of the cooling 
gas on the mixture of pure solids and impure liquid in which it forms must 
squeeze out some of the impure Uquid, which passes outward, ascends to the 
top of the ingot, or finds its way into previously formed blowholes. 

Below are presented the results recently obtained by Waterhouse* on 
segregation in add Bessemer rail steel. They point in the same direction as 
those published in 1905 by Talbot, which showed the important influence of 
alununium in greatly retarding segregation in cert^n cases. 

In the present instance, titanium, when rightly applied in the proper amount, 
was also found to retard segregation of sulphur, phosphorus, and carbon, in what 
is normally quiet, quick-setting steel. 

The ingots used were from an ordinary rail-steel heat, and from a heat to 
which had been added 64 pounds of ferrotitanium, which amounted to only 
.25 per cent of the weight of the heat. The ordinary steel was made in the usual 
way. After the heat was turned down, the proper amount of molten spiegel 
was poured into the vessel. The heat was held there a short time, poured 
into the ladle, and tiien through a IJ-inch nozzle into the ingot molds. 

There were six molds. After three ingots had been poured a ladle test 
was taken. The third ingot of the six was allowed to cool while still standing 
on its stool, and was then cut through lon^tudinally. 

The heat in which ferrotitanium was used was made in exactly the same 
way. As the steel was poured into the ladle, the alloy was shoveled in; the heat 
was then held in the ladle for three minutes before pouring the ingots. In this 
case also the third ingot of the heat was cooled in an upright position and cut 
through longitudinally. An analysis of the ferrotitanium gave: 

Carbon 10.50 

Titanium il , 60 

Iron 74, 12 

Silicon 1 .60 

Manganese 0.30 

Calcium trace 

Photographs of the ordinary and titanium ingots are shown in Fig. 285. 
The noticeable feature is the increased soundness of the titanium steel, due to 
the concentration of the blowholes in the pipe cavity. 

* The iBflueace of Titanium an Segregation in Bessemer Rail St«d, G. B. Waterhouse, Proceed- 
ings American Society for Testing Materiala, Vol. X, 1910, p. 201. 



>, Google 



STEEL RAll£ 



The analyses of the ladle tests from the two heats were as follows: 




CsrboD. Sulpbur. 


Phoephom.. 


SiLiufi. 






Poreent. 

0,44 
0.47 


Per cent. 

0.098 
0.068 


Percent. 

0.088 
0.093 


P«™t. 

0.117 
0.118 


Permt. 


Titanium ateel 


0.95 



Ordiasry Steel. Tituium Steet. 

Fig. 285. — Ordinary Steel Ingot and Titanium Steel Ingot. (Am. Soc. for Testing Materiafa.) 

No trace <rf titanium could be found in the latter steel, so that it is not, 
strictly speaking, a titanium steel, but will be called so for the purpose of dis- 
tinction. 

In the accompanying diagrams, Figs. 286 to 291 inclusive, are shown 
the results obtained from determinations for sulphur, phosphorus, and carbon. 
A If-inch drill was used, and drillii^ were taken to a depth of | inch. The 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



407 



diagrams are drawn to scale, so that the location of the diillings can be readily 
seen. The methods used were the same for all the samples and all the deter- 
minations were made by one man. The results are briefly considered in order. 
Sulphur segregates the most. In the ordinary steel it has risen from .098 
per cent in the ladle test to a maximum of .223, and there is a considerable 




Fio. 286. — Nonnal Ingot, Half Section. Fia. 287. — Titanium Ingot, Half Section, 

Sulphur in Ordinary Steel. Sulphur in Titanium Steel, 

(Am, Soc. for Teeting Materials — Waterhouse.) 

area with more than .147, which is 50 per cent more than the ladle test. In 
the case of the titanium ingot, the contrast is very remarkable. The greatest 
result is .101, which is not quite 50 per cent more than the ladle test, .068, 
and the segregated area is very much smaller than in the ease of the ordinary 
ingot. It is true that there are two factors which may partly account for this 
difference in results: the titanium ingot is somewhat smaller, and there is less 
sulphur in the steel as a whole. The fact remans, howevw, that there is much 
less segregation of sulphur in the titanium than in the ordinary steel. 



.yGoogle 



408 



STEEL RAII5 



In the ordinary steel the phosphorus has risen from .088 per cent in the 
ladle test to a maximum of .167, and here also the segregated area is seen to 
be conaderable. In the titanium steel the maximum is .124, starting with a 
ladle test of .093, and the segregated area is more restricted than in the case 
of the ordinary steel. The ordinary steel is in a better condition to start with 



'^n 



'7^'- 



^ — 3 — 3 — i 

. -V V T . 

ve , 'If f . 






^4 



"5%0~ 



. -y *> «■ 
If'/ . r »• 

t T V "i' f 
• '* ^ 1^ ■.T' 



:j^ 



IT-. . 



-/*/'- 



W^^*' 



Fig. 288. — Normal Ingot, Halt Section. Fig. 289. — Titanium Ingot, Half Section. 

Phoaphorus in Ordinary St«el. Fhosphonia in Titanium Steel. 

(Am. See. for Testing Materiab -~ Waterhouee.) 

than the titanium steel, as it has slightly less in amount (.088 as compared 
with .093), so that the behavior of the phosphorus is a good test of the pre- 
ventive power of the titanium. The results given in the diagrams Rgs. 288 
and 289 show it to have been effective. 

The titanium steel shows less segregation of carbon than the ordinary steel, 
and it must be remembered that it starts with .47 as compared with .44 per 
coit. The highest results found are .67 in the one case and .69 in the other. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 409 

and the diagrams show that the ordinary steel ag^n has the larger segregated 
area. 

The results of the sihcon and manganese determinations are not given; 
they are somewhat erraUc in each case, but do not exhibit segregation. 

Mr. Henry M. Howe presented a very complete discusaon on Piping and 



'4^^--^. 



•r r . •» -f 

y f. ^ f ^ 

^ .^ .St ^ 



-"■/!- 



.' ' • * 

«• f * ? g 

H 9 t «• 

. -? -sr r 
. . r. 



*'' i«i 



-/*«•- 



B 



-/*/•- 



-^^ 



Fia. 290. — Normal Ingot, Halt Section. Fin. 291. — Titanium Ingot, Half Section. 

Carbon in Ordinary Steel. Carbon in Titanium Steel. 

(Am. Soc. for Teeting Materials — Waterhouse.) 

Segregation in Steel Ingots, at the London meeting (July, 1906) of the American 
Institute of Mining Engineers,* of which the following is an abstract. The 
first part of this paper treats of the causes and the restraining of piping in steel 
ingots, the second considers the causes and the restraining of segregation, and 
the third proposes certain precautions in engineraing specifications concerning 
these two defects. An article coming from such a source is necessarily of value 

* Imping and Segregation in Steel Ingots. Howe, Trans. American Institute cf Mining En- 
gineoB, Vol. XXXVIII (1907), p. 3-108. 



,y Google 



410 STEEL RAILS 

in any conaderation of these subjects, and it will be pertinent to briefly review 
it. 

Professor Howe infers that the pipe is chiefly due to what may be called 
the virtual expansion of the outer walls of the ingot in the early part of the 
freezing, and finds that the upper and smooth-faced part of the pipe probably 
forms while the interior is still molten, but that the lower, steep, and crystal- 
faced part probably forms in metal which is already firm. Of the causes 
which may cooperate to limit the depth of the pipe, it is suggested that three, 
namely, blow holes, s^ging, and the progress of freeing from below upwards are 
usually effective. 

The pipe may be lessened by easting (1) in wide ingots; (2) in sand molds; 
(3) at the top instead of at the bottom;* (4) slowly; (5) and with the large end up; 
(6) by the use of a sinking-head or other means of retarding the cooling of the 
top; (7) by permitting blowholes to form; (8) and by liquid compression. 

It is believed that although the reasons why (1) casting in wide ingots 
and (2) in sand or clay-lined molds shortens the pipe do not apply to show that 
they should raise the segregate, yet the position of the segregate should be 
raised by the six other means by which the pipe is shortened (see 3, 4, 5, 6, 7, 
and 8 in preceding paragraph). 

The means proposed for lessening the degree of segregation, as distinguished 
from raising the position of the segregate, are next considered. 

These are: 
9. Quieting the steel by adding aluminium. 

10. Casting in small instead of in large ingots, and hastening the solidifica- 

tion, not only by casting in small ingots hut also 

11. By casting at a low temperature. 

12. By casting in thick-walled iron molds (i.e., those of high thermal con- 

ductivity); and 

13. By casting slowly. 

It is pointed out that quieting the steel has materially lessened segregation 
in certain cases, and that segregation is probably much less in small than in 
large ingots. 

The effectiveness of the different methods of fluid compression (see Rgs. 
294-297) is considered, and it is concluded that the benefidal lifting effect on 
the segregate should be the greatest in Williams' system, which compresses the 

* logoU are now practically all cast from the top, absolutely so la regard to rail st«el. About 
twenty years ago there was a great deal of bottom casting practiced, in casting the ingots at the Peon- 
Bylvania Steel Company's Worlcs, which were rolled into rails at the Cambria Company's Worits, Mr, 
A, L. Holley first used bottom casting. 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 411 

ingot chiefly in the middle of its length; it should be the least in Whitworth's 
system, which compresses the ingot more at its top than elsewhere; and it should 
be intermediate in the systems of lUingworth and Harmet, which compress the 
ingot equally in all parts of its length. 

Finally, stress is laid on inspection at the rolls and shears, and especially 
on axial drilling of the billets or other products. 

Figs. 292 and 293* present experimental verification of the following pre- 
dictions made in this paper: 

A. That the pipe is shortened and the segregate raised: 

1. By slow casting. 

2. By casting with the large end up instead of down. 

3. By retarding the cooling of the top, e.g., by means of a anMng 

head. 

B. That the pipe is shortened by slow coohng. 

C. That the pipe and segregate lie in the last freezing part. 

The procedure was to cast ingots of wax containing a little bright-green 
copper oleate (usually 1.5 per cent) under varying conditions; to saw each ingot 
open along a longitudinal plane pasdng through its axis; and to examine the 
longitudinal section thus laid bare. The segregated or enriched parts are shown 
by the darker areas in the photographs, indicating the green of the segregated 
copper oleate. 

Taking up the evidence in detail, the influaice of the rate of casting is 
shown in ingots Nos. 1, 2, and 3, Figs. 1, 2, and 3. The casting of No. 1 was 
finished in 30 seconds and that of No. 2 was so slow that, though it was continuous 
except for momentary interruptions for heating the wax, it lasted 1 hour and 
13 minutes. 

The pipe in the fast-poured No. 1 stretches down 90 per cent of the ingot's 
length, and, except for some very thin bridges, is practically continuous for 
49 per cent; whereas in the slowly cast ingot the pipe stretches down only 14 per 
cent of the length of the ingot. In this particular ingot (No. 2) there is a second 
riidimentary pipe near the bottom, caused by the accidental pouring at first 
faster than was intended. In ingot No. 3, which was poured slowly from the 
start, this second pipe is absent. 

The segregate in the fast-poured ingot No. 1 can be traced at A near the 
bottom of the ingot. The slow-cast ingot No. 2 has a succession of local axial 

• The Influence of the Conditiona of Cuting on Piping and Segregation, aa shown by Meana of 
Wax Ingots. Howe and Stoughton, Trans. American Institute of Mining Engineers, Vol. XXXVIII 
(1W7), p. 109. 



,y Google 



412 STEEL RAILS 

horizontal segregates, and in the still more slowly cast ingot No. 3 these local 
segregates are so small as almost to escape notice 

The effect of casting with the large end up instead of down is shown in 



if 



ll a I 



•B 

si 8 



a 



Figs. 4 and 5, which represent two ingots cast in immediate succession and under 
otherwise like conditions. The pipe stretches down only 30 per cent of the in- 
got's length when the large end is up, but 82 per cent when the large end is down. 
The segregate Ues well above the center in the ingot with the large end up, but 
very near the bottom in that with the large end down. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 413 

The effect of retarding and of hastening the cooling of the top of the ingot 
s shown in Figs. 6 and 7 and in Figs. 8 and 9. The depth to which the pipe reaches 
IS a nearly continuous cavity is only 26 per cent of the mgot's length in the hot 



4 



i f 



^1 



el .as 
it! I" 

1 1 

a * 

* •B 
III 

3| I 



3j 



topped ingot No. 6, but 85 per cent in the cold-topped No. 7. The pair of ingots 
Nos. 8 and 9 which viere slow cooling do not show so well the influence of the 
distribution of temperature on the position of tiie pipe. 

The effect of the rate of cooling is shown in Figs. 8, 9, and 10. Figs. 11 and 



,y Google 



414 



STEEL RAII£ 



13 illustrate what Professor Howe calls "surface tensdon bridges." In the middle 
and lower part of No. U there are five bridges from F to J and at K and L, as 
well as at Af of Fig. 13, there are the remains of bridges. These are in each case 
greener than between the bridges and appear to have been formed oa account of 




FiQ. 294. — illingworth's Fresa for Compressing Steel In^te Homontally while Solidifjing, Sectional 

Plana. 
In I, the mold is shown ready for receiving the molten steel. Two distance-barB, DD, are set between 
the halves o( the split mold B and C. After the steel has been poured into the mold, these distance-bars 
are pulled out lengthwise, and the two halves of the mold are then forced towards each other by meaoa 
of the ram F, shown in II. The convex edges of the distance-bars are for the purpose of making an 
initial depression in the side of the ingot lest part of its side should be forced out as a fin or welt into the 
crevice between the two halves of the mold. 

(Trans. Am. Inst, of Mining Engrs., Vol. XXXVIII.) 

the local enrichment of the oleate, making the wax so fusible and plastic that it 
stretches instead of cracking open when the pipe is being formed. The same phe- 
nomma was noticed at E, F, G, of Fig. 12. The mgot of this figure had its cooling 



I 



II 



ill 



Fto. 295. — Williams' Abdominal Liquid Compression of Solidifying Steel Ingots. 

The ingot is cast in Its mold as shown at I. Aftt^r its outer crust has solidified the mold is opened, as 
shown at II, and a hner B is slipped between mold and ingot. A strong cap A is then fastened down, 
and by means of pressure applied through the ram C the abdominal protuberance on the ingot is forced 
in, so as to close the pipe and Uft the segregate into it, as shown at III. 

(Trans. Am. Inst, of Mining Engrs., Vol. XXXVIII.) 

hastened on the right-hand side and retarded on the left-hand ade, which resulted 
in shifting the pipe distinctiy to the left of the axis. 

Several methods have been used to produce sound ingots, as stirring the 
steel in the casting ladle to allow the gases to escape; casting on a turntable 
that is made to revolve and the metal run into a mold at its center; bottom 
casting, with a closed top instead of the ordinary open-topped molds; and casting 
under pressure. 



,y Google 



\ 




1 




a 





INFLUENCE OF DETAIL OF MANUFACTURE 415 

There are different methods of exerting the pressure on the fluid metal 
in the mold. At the Krupp mills, fluid compression was tried by applying 
the pressure exerted by carbonic anhydride 
in its fluid state, the ingot mold being 
capped after the ingot was cast and con- 
nected to a reservoir contmning the car- 
bonic anhydride. 

At the Edgar Thompson works the same 
principle was appUed, u^g steam under a 
pressure of about 200 pounds per square 
inch. Under this steam pressure the ingot 
of 5 or 6 inches diameter shortened in length 
from 1^ to 2 inches. 

Illingworth's process (Rg. 294) conasts 
of casting in vertical molds, split lengthwise. 
The two halves are separated during the 
casting, but when the crust is formed they 
are brought together by a ram. 

Williams' system (Fig. 295) employs the 
split mold, and the two sides are pressed to- 
gether with a liner between. 

The Whitworth process (Fig. 296) con- 
sists of u^ng a steel mold which is placed 
in a hydraulic press and the fluid steel sub- 
jected to a pressure of about 6 or 7 tons 
per square inch of horizontal section. Un- 
der this pressure the ingot shortens about 
li inches per foot of its length. This pro- 
cess produces an ingot of uniform quality 
throughout and in a great measure over- 
comes the difficulty experienced from the <Traiw. Am. inat. of Miniog Eogrs., Voi. 
formation of blowholes and piping. 

The Harmet process (Fig. 297) consists in using a tapering mold and com- 
presang the fluid by means of a hydraulic ram acting on the open end of the 
mold. The effect of the tapering mold is to exert a lateral pressure which tends 
to close up any axial pipes. The French Government require 28 per cent crop 
in uncompressed ingots and only 5 per cent in compressed ingots made at the 
St. Etienne works by the Harmet process. 




Fia. 296. — Whitworth'fl Hydraulic 1 
for the Compression of Steel Ingota while 
Solidifying. 

A, main compreaeion-cylinder. B, ita 
plunger, C, the carriage on which the 
mold or flask sits. G, boes against which 
the ateel in the mold is forced. KK, steel 
jaclfets fof the mold. LL, the mold 
proper. MM, perforated cast-iron lag* 
ging. NN, inner sand lining. 



>, Google 



416 STEEL RAILS 

The test pieces cut from the compressed ingots diow, without forging 
or roUing, as good results under tensile and impact tests as test pieces cut from 
ingots of the same composition which had been forged with a reduction of two 
times in the cross-sectional area. There is a very marked diminution of segre- 
gation, the chemical analyses at the top and bottom of the ingot being sub- 
stantially the same. The compressed ingot has a grain of a visibly finer structure 
and the large cleavages often found in sections cut 
from uncompressed ingots are not found. The 
metal is sound and thoroughly homogeneous. 

Fluid compression of the steel in the ingot, when 
in proper hands, according to our present evidence, 
prevents pipes, blowholes, and cracks almost com- 
pletely, and, to a limited extent, segregation. 
While its introduction would certainly lead to 
complication at the mills, the benefits to be derived 
warrant a trial of this method. 

As has been seen the greatest defects are found 

Fio.297.-H«met'« Liquid Com- »» ^^^ ^PV^T part of the ingot, and to obtain a 

pression by Wire Drawing. souud ingot it was generally specified that a certain 

The ingot. 4^, iaoaat in a strong discard or cfop should be made from the top of 

conical mold, 27, reinforced 

with hoopa, 28. Strong pres- the iugot, which it was supposcd would contain all, 
Bure at the base of the ingot, 26, qj. ^g^riy all, of the Imperfect metal. 

forces it leogthwise of ^he mold, 

thus compressing it radially. To get definite knowledge on this pomt Dr. P. 

(Trans. Am. Inst, of Mining H. Dudley experimented by lettering the rails fwrned 

Engra., Vol. XXXVIIL) , j-^r i. _.. / .^u • i. an. -i i * 

from different parts of the mgot. The rails were let- 
tered "A" for the top rail, "B" for the second rail, and "C" for the third, or rail from 
the bottom part of the ingot. These letters could be found in the tracls, and, as 
was to be expected, the "A" rails have a larger percentage of impurities than the 
" B " or "C " rails. They wear faster, developmg more surface defects, and at seveal 
points upon the road (the New York Central), under heavy traffic, after 10 or 12 
years' service, have become practically wean out for main-line traffic, while the 
"B" or "C" rails are still good. 

The Committee on Standard R^ls of the American Railway Assodation 
reported at the meetmg of the Association, April 22, 1908, on this subject, 
as follows: 

All rails are to be branded with the name of the maker, the weight of the rail, and the month 
and year; and the number of the heat, and a letter indicating the portion of the ingot from which 
the rail was made, shall be plainly stamped on the web of each rail, where it will not be covered 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 417 

by the splice bare. Rails to be lettered consecutively " A," " B," " C," etc, the rail from the 
top of the ingot being " A." In case of a top discard of 20 or more per cent, letter " A " will be 
omitted. All rails marked " A " shall be kept separate and be shipped in separate cars. 

While railway engineers formerly specified a definite jjercentage of discard 
from the ingot, they now unanimously agree that the specifications should not 
state definitely how much should be sheared from the bloom, but that sufficient 
material should be discarded from the top of the ingots to insure sound rmls. 
Records of r^l failures which have been kept for a number of years disclose 
the incontrovatible fact that, where a 15 per cent discard might do for one 
ingot, 50 per cent would not be adequate for another.* 

In its report the Committee of the American Rmlway Association says: 
^ith regard to the discard question, the Committee has always held that it 
would be preferable to test the finished product rather than specify as to the 
details of mill manufacture, and the Committee arranged for a trial lot of ruls 
to be rolled from the ingot without any discard whatever, except such as was 
necessary to enable the bloom to enter the rolls, and after these rails had been 
cut into small pieces they were broken under the hammer and the fracture ex- 
amined. This test proved to the satisfaction of the Committee that if "pipes" 
or other physical defects were present they could be detected by this means. 
The test also proved quite conclusively that it is posdble so to conduct the 
process of manufacture- that the "pipes" or other physical defects will be re- 
duced to a minimum, and that these defects may not occur at all, even in rails 
rolled from the top portion of the ingot. 

In order to avoid an imnecessary waste of good material, the Committee 
set about to devise means by which the rejection of defective material could 
be insured without requiring an arbitrary and definite percentage of discard 
in every case, and a committee of the Pennsylvania Railroad, pursuing the 
same line of investigation, adopted a tentative specification which provided 
for a phy^cal test of this nature, and which further provided that when physical 
defects were discovered all top rails of the heat should be rejected. A trial 
lot of rails, of a section corresponding to " type B " (Plate VIII),was rolled un- 
der this specification as to discard, and the results convinced the conunittee 
that a development of this idea would prove the best solution of the discard 
problem. 

Some of the advocates of a fixed and arbitrary discard have argued that 
the mere provision of a discard to insure the elimination of " piped " rails, or 
rails containing physical defects, was not sufficient, and urged the rejection of 

* See Papa by W. C. CuBhiDg before the Indiana Railroad Conausaion, Fdiniary 20, 1912. 



,y Google 



418 STEEL RAIIS 

a fixed percentage from the top of Hie ingot, because of the well-known fact that 
segr^ation occurs in the upp&r portion. This question of segregation was given 
careful conaderation by the Committee, and while it is a fact that, due to the 
rearrangement of the constituent parts of the metal during the process of cooling 
and solidification in the ingot mold, any analysis of the metal in the finished 
rail will often show a wide departing from the analysLs required by the speci- 
fications, it is also true that an analysis of the metal taken from the different 
parts of the finished rail will frequently show similar wide variation. This 
discrepancy is due to the fact that the test ingot referred to in the specifications, 
and upon which the chemistry specification is based, is taken from the ladle before 
the metal is poured into the ingot mold, and, consequently, before the segrega- 
tion takes place. 

It has been assumed that, because of this variation from the standard 
composition of the metal in the finished rail, the rejection of all segregated 
metal would be warranted. But, on this assumption, it would be nerassary 
to discard more than a third of the upper part of the ingot to be on the safe 
side, as the segregation frequently extends that far; and, while our knowledge 
of the subject is not so complete as we could wish it to be, we have a great deal 
of evidence that rails of good physical condition can be made from the upper 
portion of the ingot. Furthermore, the analyses of a large number of rails, 
taken after years of service, indicate that these wide variations in chemical 
compoation may occur without apparently affecting the safety or wearing 
quality of the rail; and, since it is impossible to check the analyses of the finished 
rail with that of the test ingot, the question arises as to what limits should be 
placed on the variation which will be permissible. None of the experts con- 
sulted are ready to say what this limit should be, and all admit that no facts 
are available as the results of actual experience which would warrant the adoption 
of any fixed limit to govern the rejection of material. 

The provision in the new specifications for stamping the rails to show 
their position in the ingot will enable us to obtain more definite information 
on this point in the future. 

BIBLIOGRAPHY 

Beikirch, F, 0. — Verfahren mu verhUtung der lunkerbildung in Bchweren robatahlbtdckeo, 
1800 w. lU. 1905. (In Stahl und Eisen, Vol. 25, Part 2, p. 865.) 

Describes uae of sinking-head, afterwards cut oS, Piped portion is thus removed. 

Daglen, K, M. — Die verfoiireD Eur verhUtung der lunkerbildung in stahUilocken. 1800 w. 111. 
1905. (In Stahl und Eisen, Vol. 25, Part. 2, p. 923.) 

The same. {In Zeitflchrift des Vereines Deutscher Ingenieure, Vol. 49, Part 2, p. 1398.) 

Describes fluid compression methods, methods of heating moulds and continuous procesi, proposed 
by writer. 

DoRHUs, A. V. — Die blasen- und lungerbildung des flusseisens. 1000 w. Q]. 1902. (In Zeit- 
schrift dea Oeterreiduschen Ingenieure- und Archit«kten- Vereines, Vol. 54, Part 1, p. 279.) 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 419 

The tame, translated. (In Siblev Journal of Engineeri 
Describes forniation of b'ow-hofeB and pipes in steel a 
prevention. 

Dudley, P. H. — Dark carbon streaks in segregat«d metal in split heads of rails. 2000 W. 111. 

1909. (In Proceedings of the American Society for Testing Materials, Vol. 0, p. 98.) 
The tame. (In Railway and Engineering Review, Vol, 49, p. 626.) 

Hbboult, p. L. T. — Presence and influences of gases in steel. 800 w. 1910. (In Transactions 
of the American Electrochemical Society, Vo . 17, p. 135.) 

Concludes that blow holes are the result of disengagement of carbon monoxide in sl«el. 

Howe, Henry M. — Does the removal of sulphur Mid phosphorus lessen the segregation of carbon? 
2500 w. 1907. (In Proceedings of the American Society for Testing Materials, VoL 7, p. 75.) 

Results from examination of many cases indicate that removal of sulphur and phosphorus does 
not lessen segregation of carbon. 

Howe, Hbnrv M. — Further study of segregation in ingots. 5000w. 111. 1907. (In Engineer- 
ing and Mining Journal, Vol. 84, p. 1011.) 

Suggests that quiet, resulting from increased ingotr^Ize and slow cooling, may explain why these 
CMiditions do not alvrays favor segregation. 

Howe, Henry M. — InQuence of ingot«ize on the degree of segregation in steel ingots. 800 w. 
III. 1909. (In Transactions of the American Institute of Mming Engineers, Vol. 40, p. 614.) 

Howe, Henry M. — Influence of top-lag on the depth of t£e pipe in steel ingots, 1000 w. 1909. 
(In Transactions of the American Institute of Mining Engineers, Vol, 40, p. S04,) 

Howe, Henry M. — Piping and s^regation in steel ingots. 106 p. Ill, 1906. (In Transactions 
of the American Institute oF Minmg Engineers, Vol, 38, p. 3.) 

Discussion, Vol. 39, p, 818. 33 p. 

Considers causes and prevention of piping and of segregation and precautions to be taken regarding 
these in specifications. 

Howe, Henry M. — Segregation in steel ingola. 1000 w. 1908, (In School of Mines Quarterly, 
Vol. 29, p, 238,) 

Summarizes results of author's investigations. 

Howe, Henry M., and Stovqhton, Bradley, — Influence of the conditions of casting on piping 
and segregation, as shown by means of wax ingots. 3000 w. 111. 1907. (In Transactions of the 
American Institute of Mining Engineers, Vol. 38, p. 109.) 

VeriGes by experiments views expressed by Howe in earlier paper. 

Huston, CH*RLBa L. ^Experiments on the segregation of steel ingots in tie relation to plate 
specifications, 3000w. 111. 1906. (In Proceedings of American Society for Testing Materials, Vol 6, 
p. 182.) 

Illustrates and eimlains cases of segregation. 

Job, Robert. — Investigation of defective open-hearth steel rails. 1500 w. 111. 1909. (InPro- 
ceediniraof the American Society for Testing Mataials, VoL 9, p. 90.) 

Discussion, p. 106. 

The game. (In Railway Age Gazette, Vol. 48, p. 523.) 

Shows failures to be due to unsoundness of meUl. 

KNiaar, S. S. — Observations on segregation phenomena as applied to cast steels. 3000 w. 111. 

1910. (In Iron Trade Review, Vol. 46, p. 476.) 

Paper before Philadelphia Foundrymen's Association. 

Illustrates examples of segregation and emphasizes need of further investigations. 

LiLiENHERG, N ^Piping in steel ingots. 2000 w 111, 1906. (In Transactions of the American 
Institute of Mining Engineers, Vol. 37, p, 238.) 

Considers methods for prevention of piping, with special attention to Illingworth's side-compression 
method. 

Mathbsius, W. — Herstellung von poren- und lunkerfrciem grauguss, stablguss und schmiede- 
stUckendurchanwendungvonthermit 2500w. 111. 1903. (In Stahlund Eiaen, VoT.23, Part 2,p.925,) 

RiEUER, JuLios. — Ein neues verfahren zura verdichten von stahlblOcken in flUssigem zustande. 
2000 w, lU. 1903. (InStahlundEisen, Vol. 23, Part2, p, 1196.) 

The same, translated. (In Iron and Coal Trades Review, Vol. 67, p. 1776.) 

New method ci — '"'' "' ■-■.-•. ... 

Saovedr. Alb: 
method. 1500 w, lU, 1903. 
p. 129.) . , 

Segregation in soft steel ingots; its effect on rolled material as shown in tensile and shock tests. 
2500 w._ 111. 1910. (Inlron Age, Vol, 86.P. 730.) 

Gives results and conclusions reached by WQst and Felser in paper in " Metallurgie," Shows 
s^regation u> have matest infiuence in shock tests, as segregated material is very brittle 

SPRiNaER, J. F, — Piping in irteel ingots; methods for its reduction and elimination. 4000 w. 
III. 1909. (In Caasier'sMagazine. Vol. 36, p. 426.) 

Springer, J. F, — Thermal treatment ofsteelingots. 1200w. HI. 1910. (In Scientific American, 
Vd. 116, pp. 282. 269.) 

Describes piping and s^regation phenomena and briefly reviews methods for prevention. 

Stead, J. E. — Crystallization and segregation of steel ingots. 53 p- III. 1906. (In Proceedings 
of the Cleveland Institution of Engineers, 1905-06, p. 163.) 

Review of work done on the subjeot, with conclusions. 



>, Google 



420 STEEL RAILS 

Talbot, Benjamin. — Segregation in steel ingoU. 44 p. III. 1905. (Id Journal of the Iron aod 

Steel Institute, Vol. 68, p. 204.) 

Includes extensive tabulated data on the eGFect of additions of small amounts of aluminuni lo 
the ingot. 

Waulbbuq, Axel. — InSuence oF chemical composition on soundness of steel ingots 38 p 111 
1902. (In Journal of the Iron and Steel Institute, Vol. 61, p. 333.) 

Wedding, H. — Unterauchung Uber den uisprung eines blasenraumee in einen flusseisenblocke 
2000 w. 111. 1905. (In Stahl und Eisen, Vol. 25, Part 2, p. 832-) 

Considers probable CMigin of a blow-hole in a 2-ton ingot. Faulty shape of mould cooaideied to 
be explanation. 

Weitere entwicklung des Rieraerschen verfabreos zur herst«llung dichter stahlbldcke. lOOO w. 
111. 1904. (In Stahl und Eisen, Vol. 24, Part 1, p, 392.) 

Illustrates great reduction in piping from uae of new method of Riemer. 

WiCKHOBST, M. H. — Low-carbon streaks in open-hearth rails. 1200 w. III. 1910. (In Procent 
ings of the American Society for Testing Materials, Vol. 10, p. 212 ) 

Studies rails that developed a peculiar kind of failure, shown to be due to streaks of metal low in 
cartion. 

WiCKHORST, M. H. — Sep^ation and other rail properties as influenced bv siae of ingot 97 p. 



33. Influence of Mechanical Work 

The principal points in connection with the rolling are given below: 

1. Resistance to wear is a function of fineness of grain. 

2. Fln^ess of grain is principally a result of mechanical treatment at 
proper temperature.* 

3. Work done on steel above SSO^-IOSO" C. (1742° F.-1922° F.) has less 
effect on changing the size of gr^n from the normal crystallization of the ingot 
than when the roUing is done at a lower temperature. 

Fig. 298 shows the steel entering the rolls. Figs. 299, 300, 301, and 302 
illustrate views taken by Mr. Howard and show the gradual reduction of the 
bloom to the finished rail as it passes through the successive rolls. 

In 1909 a further investigation was made of the steel at different stages of 
the rolling by James E. Howard at the Watertown Arsenal.t In these tests, 
beginning with the ingot, the structural state of the metal was examined by 
taking cross sections and longitudinal sections. This method was carried 
through the various successive derivative shapes, and the results obtained are 
shown in the large number of illustrations which form the body of the report. 

The greater part of the work was devoted to Bessemer rail steel, five add 
Bessemer heats being made for this series of tests, each heat furnishing sai 
ingots about 19^ by 20^ inches at the bottom and about 5 feet high. 

One of the most important results of the tests was to throw light on the 
question of the amount of work or reduction necessary in rolling to develop 
the full physical qualities of the steel. Mr. Wickhorst draws the following 

■ This should not be interpreted as meaning that resistance to wear is not a function of the chan- 
ical composition. 

t Tests of Metals, etc., 1900, Vol. 1 and Vol. 2, Government Printing Office, Waehingtoo. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



,y Google 



422 STEEL RAILS 

conclusion from the tenale tests made of specimens taken at various stages 
from the ingot to the finished rail. 



Fm. 2d9. — Cross Sectioa of 8 by 8 in. Rail Bloom Rolled from an Ingot 20 ins. Square. 
(Am, Ry. Eng. Aaao.— Howard.) 

The results indicate that the metal in the walls of the ingot takes com- 
paratively little work or reduction to impart to it what may be called its full 



Fio. 300. — Rail from an £arly Pass in Roughing RoUb, Rolled from Bloom Shown in Fig. 299. 
(Am. Ry. Eog. Assd. — Howard.) 

physical properties of tensile strength and ductility. These are reached in the 
bloom, except at the top end. The axial metal at the bottom of the ingot also 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



Fio. 301. — S&tne Rail aa Shown in Fig. 300 after Further Reduction. 
(Am. Ry. E^g. Asan. — Howard.) 

soon reaches its full physical properties, but in the upper half of the Ingot it 
must be carried well toward the finished rail before these properties are fully 
developed. 



Fia. 302. — Finiahed Rail froni Same Ingot as Bloom and Pieces from Roughing Rolls. 
(Am. Ry. Eng. Assn. — Howard.) 



>, Google 



424 



STEEL RAIM 



Where the metal ia of fairly even compoMtion and free from sponginess, it 
reaches its full physical qualities of tensile strength and ductility at about ten 
reductions or a reduction to one-tenth of the original cross section of the ingot, 
but the interior portion of the upper part of the ingot requires twenty-five or 
more reductions to have its full physical qualities developed, that is, the cross- 
sectional area must be reduced below one twenty-fifth of its original amount. 

Table XCII pves the result of tests by Sauveur on the relation between 
the size of the grain and the physical properties of the same piece of steel.* 

TABLE XCIL — RELATION BETWEEN SIZE OF GRAIN AND 
PHYSICAL PROPERTIES OF STEEL 



Sin of CraiE. 


Teoiile Stnnstta. 


Elouolioii 


ReduMkoaot 


MiilimBWr. 


Number p«r 
Squan Inch. 
AppmiiMtBly. 


'^;;p:- 


Pound* per 
8qi»r« l»b. 


148 
118 
62 


44,000 
54,000 
101,000 


69.6 
70.3 

77.7 


99,000 

100,000 
110,000 


15 
19.0 
22.5 


20 
22 
36 



Mr. Robest Job, chemist of the Philadelphia & Reading Railroad, states 
that " in a lot of over 75,000 tons of rail observed during a period of five years, 
we have 15 times as many fractures in service from rails of coarse grain, or 
19,600 cells per square inch, as from rails of medium fine structure, 48,400 cells 
per square inch, and there is also a marked difference in capacity for wear in 
favor of the finer structure rml. Out of sevaal thousand tons of rail now 
(1905) in our tracks, made with a clause in the specifications requiring more 
than 40,000 cells per square inch, only one rail has fractured in service, and that 
owing to pipes in the steel in process of manufacture." 

Dr. P. H. Dudley states that rails of 100-pound section with 48,000 to 70,000 
cells per square inch after having sustained 250,000,000 tons in the track were 
still in good condition. 

The roiling tends to break down the grain and give a finer structure, but im- 
mediately after the work stops the grain commences to grow ^ain, consequentiy 
the lower the finishing temperature the smaller the grain size. If the steel is 
worked below the critical point, strains are developed which injure the metal 
and may rupture it. Work at too low a temperature distorts the gr^n or 
flattens and elongates the crystals in the direction of the rolling. 

* N. LjomiD, Chem. Zeit, 1899, Baumateri alien, 1S99, finds the tenacity in different Bt«dB Tarka 
directly as the sise of the pearlite grains, at the same finiBhing temperaUiie. 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 426 

To thoroughly understand the effect of the rolling, it is necessary to study 
the structural changes that take place in the cooling metal.* 

Let us first consider the cooling curve of a copper bar, shown in Fig. 303. 
We find here no evidaice of any sudden change in the nature of the cooling copper. 



ISO' 

so' 


\ 








\ 










\ 










^ 


-^ 



Tim* 

Fio. 303. — Cooling Curve ot Solid Copper. 
(J. W. Mellor.) 



V 








\ 


■nezinK 








\ 


\ 








X 


■-^ 



TbM 

Fia. 304. — Cooling Curve of Water. 
(J. W. Mellor.) 



If, however, a curve is drawn for water cooling down from 20° to -20° C, 
we get a terrace in the coohng curve, as shown in Fig. 304. This tells us that 
some change has taken place in the nature of the substance at 0°. We see at 



























\ 




























■-^ 






l\ 
























l, 


\ 


























^ 


L^ 


^ 




























■ — 



20 40 60 80 too Time 

FiQ. 305. — RecaleBcence. (J. W. MeUor.) 

once that this change corresponds with the passage of water from the liquid 
to the solid state. 

When a steel bar is coohng, an evolution of heat occurs at about 690° C. 
The amount of heat evolved is so great that the metal visibly brightens in color. 
The phenomenon is called " recalescence." The cooling curve is shown in 
Fig. 305. 

* See The Ciystallization of Iron and Steel, J. W. Mellor, 1006. 



,y Google 



426 



STEEL RAIIfl 



The cooling curve of iron from the molten condition is shown in Fig. 306. 
The iron was practically pure, containing only .01 per cent of carbon. 

Osmond maintains that the existence of the tranation points, Ar, and Arj, 
in the cooling curve of the solidified metal points to the existence of three 
allotropic modifications of solid iron. Each critical point is found to be associ- 
ated with a change in the mechanical properties, the microscopic appearance, 
and the specific gravity of the metal. 





V. 


Fr..,lnj 










\, 






600' 




86C 


S^ 










V 



V 






/ 


\ 




Ae,,.^ 


i 


«s 


f^ 


/ 




^ 


./ 


V 


^ 



Fia. 306. — Cooling Curve of Iron. 

(J. W MeUor.) 



20 40 60 Time 

Fio. 307. — Cooling and Heating Curves of 

Steel. (J. W. Mellor.) 



The changes which occur during the cooling of a substance are reversed 
when the substance is heated. The cooling curve of steel, with 1.2 per cent 
of carbon, shown in Fig. 307, is reversed on heating, as shown by the heating 
curve in the same diagram. 

The critical points on the heating curve of mild steel are generally a few 
degrees higher than the corresponding points on the cooling curve. 

Let us now consider what takes place when steel containing 0.6 per cent 
carbon cools from 900° C. The cooling curve shows nothing very remarkable 
until a temperature of about 720° C. is attuned. Here the critical points 
Ar% and Ar^ of pure iron coalesce into one. At this point pure iron, or 
ferrite, separates from the solid solution. The sejKiration of ferrite goes on 
along the curve AF (Fig. 308) until the temperature reaches about 690" C, 
when another recalescence point occurs (Ari). No other essential change, as 
far as we are concerned, occurs as the system cools down to the normal tem- 
perature of the atmosphere. 

n,g. 308 is derived from Roozeboom's diagram,* the carbon-iron diagram, 

• H. W. B. Rooseboom, Zeit. Phys. Chem., 34.437, 1900; improved in Zeit. Elektrochem., 
10,489, 1904; MetallographiBt, 3.293, 1900: H. le ChateUer, ibid., 3.290, 1900; 4.161, 1901; F. Osmood, 
4.150, 1901; H. Jiiptner von JonBtorft, ibid., 5,210, 1902. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



427 




400, 



given by Howe,* based upon later researches, shows the temperatures some' 
what higher than those of the figure. 

When the temperature is above the hne APB the iron is in a form known 
as "austenite." Whatever carbon is present is dissolved in this austenite, 
which is what is called a "solid solution" as distinguished from a mechanical 
mixture or conglomerate, just as salt and water, when brought in contact, merge 
in each other and pass from the condition nooV 
of a mixture or conglomerate to that of a 
^ngle substance. 

As the iron with 0.6 per cent carbon 
cools and the line AP is reached the 
austenite begins expeUing from itself part 
of its iron in the form of ferrite. As the 
ferrite thus expelled is nearly or quite free 
from carbon, the remaining austenite be- 
comes relatively richer in carbon, until, 
when the temperature reaches Ati, it 
contains 0.9 per cent carbon which is the 
carbon content of pearlite. On cooling 
past this point all the austenite changes into pearlite, with no change in the 
ferrite which it has generated in the passage, along the line AP, so that the 
steel now con^sts of a conglomerate of ferrite and pearlite. This conversion of 
the austenite into pearhte is accompanied by a considerable evolution of heat, 
and is shown by the recalescence curve of Fig. 305. 

Steels containing just 0.90 per cent carbon, and hence consisting of pearlite 
alone, are called "eutectoid " steels. Those containing less than this are called 
"hypo^eutectoid," and those more than this, "hyper-eutectoid" steels. 

As previously stated, work below the Ar^ point distorts t^e grain or flattens 
and elongates the crystals in the direction of the rolUng. 

The result of work above the An point is to retard the growth of the 
grains. Howef explains the relation between the temperature of the hot work 
and the size of the grain as follows: 

The mechanical distortion in rolling elongates these gnuns in the direction 
of rolling and shortens them in the plane of the pressure; this appears to throw 
the metal crystallographically into unstable equilibrium, with the result that 



Fia. 308. 



1-S?i Carbon 



- Cooling of "Solid Steel." 
(J. W. Mellor.) 



''life History of Network and Ferrite Grains lit CarboD St«el, H. I 
Atnericfm Society for Testing Materials, Vol. XI, 1911, p. 266. 
t H. M. Howe, Iron, Steel, and Other AlloyH, 1903, p. 262. 



, Howe. Proceedings, 



>, Google 



428 STEEL RAIia 

the old grains thus distorted break up, and that the metal rearranges itself into 
new and equiaxed grains. 

But these new grains assume a size normal, not to the temperature at which 
the old ones had formed, but rather to the temperature now existing; during the 
rolling the temperature is constantly falling; each pass through the rolls tends 
more or less fully to break up the preexistii^ grain, and to substitute for it a 
new grain of a aze more nearly normal to the 
now lower temperature. To speak more accu- 
rately, the new grain size approaches that 
normal for the existing temperature; but the 
result is much the same. For if each of a 
succes^on of passes through the rolls breaks 
up the existing grain, and substitutes for it a 
new one, then ^ch new grain will be smalla- 
than the preceding, because the normal towards 
which it tends is smaller than the normal 
towards which its' predecessor tended at t^e 
higher temperature then existing. 

Fig. 309 attempts to express this condi- 
tion of affairs graphically. Here ordinates 
represent temperature and abscissae coarseness 
of grain. The line AciA may be taken as 
representii^ roughly the normal size of grain, 
D", which steel of g^ven composition tends to 
assume with varying temperature, or the line 
of normal coarseness of gnun. If the grain is 
smaller than the normal for existing tempera- 
ture, it always tends to grow and to approach 
that normal. If it is coarser than that normal, it does not tend to shrink 
back towards the normal, except when the temperature is riang past Act. 

Let us suppose that we cease rolKng a piece of steel while its temperature 
is at B, the mechanical work of the rolls having broken the grain down. During 
subsequent cooling the grain will grow, somewhat as sketched in the Hne BCE. 
If, however, we resume rolling when the grain has reached C, we will break down 
' the grain, and drive it back, say to D. And so, keeping on, between passes the 
grain grows and the temperature simultaneously falls, while at each pass the 
squeeze which we give the metal breaks up the grain, and the curve of grain 
and tempraature follows the zigzag line BCDG. 





B 4 . 


^^' 


G 




^ 


1 I 

1 i 






A -*> ,: 





Ora//7 ' ^/>e 

Fio. 309.— The Influence of the Finisb- 
ing Temperature on theSizeof Grain. 

(Howe.) 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 429 

If we cease rolling when the temperature has fallen to G, then the grain 
will grow as the metal cools, till the line of the actual size of grain intersects 
that of the normal size, the line AciA ; with further cooling no further growth 
ensues, and the final size of grain is OP. If we had quenched the metal while 
at G, the final size of grain would have been OH. If we had ceased rolling 
when the temperature was at B, the final size of grain in the cooled steel would 
have been OE. 

Generally speaking, the grain aze will be the coarser the higher the finish- 
ing temperature. Fig. 310 illustrates this principle. This shows the micro- 
structure of two like bars of the same steel, of which each had first been heated 



Cooled to 993° C. Cooled lo S17" C. 

Fio. 310. — Influence of Fioishing Temperature on the Size of the Grain of Steel 
of 0.50 per cent CarboQ (Howe,) 

to 1394° C, then cooled slowly to the temperature indicated in the figure, then 
rolled, and then cooled slowly, so that these temperatures are the " finishing 
temperatures." Note how much coarser the meshes are in A, finished at 
963° C, than in B, finished at 837°. 

Professor Sauveur's micrographs of r^l structure show that in the section 
of a given rail, the network, or size of the walled cells, is the coarser the higher 
the temperature at which the rolling is finished.* 

According to Howe, when a piece like a rail, which is highly heated, is 
rolled with such heavy reduction as to distort the austenite grains greatly, the 
distorted and hence unstable grains immediately shatter, and their remans im- 
mediately begin growing again by coalescence. This is repeated as often as the 
piece is greatly reduced by the rolling. Each of the grains of austenite, formed 
by coalescence after the last of these reductions, in cooUng down to the Ari 
point, gives birth to a walled cell by ejecting to its outside the ferrite which it 
generates. Hence it is the size which these austenite grains reach after this last 

• Trans. Americtm Institute of Mining Engineers, Vol. XXII, 1893, pp. 546-557, and especially 
Plat« IV and V. 



,y Google 



430 STEEL RAIIS 

effective reduction that determines the network size of the cold steel, and as 
regards the opportunity for network growth, the finishing temperature is the 
equivalait of the highest temperature reached by objects not rolled. 

This well-marked network common to rail steels is probably due to their 
large manganese content, as otherwise, on account of their slow cooling, the 
ferrite would coalesce and break up the network.* 



















\ 




















r ._v-v>^ 


IH^r 




'K^^^r^ 






^ 


M 








c< 


■S 








i 


^! 










1 

1 


PMCTICW. 



Fig. 311. — Diagram of Results of EicperimeDts on Rolling at Dififerent Temperatures. 

This principle of governing the grain size by means of the finishing tempera- 
ture is of very great importance. In general, we should be inclined by considera- 
tions of economy of power to roil steel as hot as we dare, because the hotter it is 
the softer it is, and the less power is consumed in rolling. But this would natu- 
rally lead to a high finishing temperature, and thus to coarseness of gnun and 
brittleness. Hence a high temperature is desirable as regards power consump- 
tion, but unde^rable as regards the quality of the steel. 

t Fig. 311 shows graphically the results of experiments made at the Spar- 

• For iliufltration of thia cellular structure in rails aee Job, The Metallographist, Vol. 5, 1902, 
pp. 177-191; P. H. Dudley, ibid., Vol. 8, 1903, p. HI. 

t Tbe Manufacture and Properties of Iron and Steel, H. H. Campbell, 1904, p. 410. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 431 

rows Point plant of the Maryland Steel Company. As in Fig. 309 ordinates 
represent temperature and abscissae coarseness of grain, the grain growing 
coarser from left to right. An ingot was rolled into blooms and two adjacent 
blooms, "A" and " B," were rolled into rails without further heating, the first, 
"A," bang held before rolUng in order to allow it to cool so that all work 
should be done at as low a temperature as possible, without, of course, reaching 
the lower critical point, while the second, "B," was rolled as quickly as possi- 
ble through all the passes, except the last, but was then held at the finishing 
pass If minutes; the result being that both pieces went through the finishing 
pass at the same temperature, which was about 750° C. (1382° F.). 



Fia. 312. — Rail " B" Near Surface, Fia. 313. — Rail "A" Near Surface, 

46 Dia. — CampbeU. 46 Dia. — Campbell. 



Fig. 312 shows " B " rail near the surface. 
Fig. 313 shows " A " rail near the surface. 



Fio. 314. — Rail "B" Center of Head, Fra. 315. — Rail "A" Center of Head, 

46 Dia. — CampbeU. 46 Dia. — CampbeU. 

Fig. 314 is from the center of the head of " B " rail. 
Fig. 315 is from the center of the head of " A " rail. 
While Figs. 312 and 313 appear similar. Figs. 314 and 315 show the real 
difference between the two rollings. The last pass does very little work; there- 



,y Google 



STEEL RAILS 



fore, holding the rail hefore the last pass does little good, except on the outer 
surface of the rail, and a low shrinkage or finishing temperature does not neces- 
sarily mean that the r^ will have a good grain throughout. 



Fia. 316. — Top View at Top of Head, TO lb. Rail, Fic;. 317. — Top View at Ceoter of Head, 

50 Dia. (Am. Ry. Eng. Aasn.) 70-lb. Rail, 50 Dia, (Am, Ry. Eng. Amd.) 



Fia. 318. —Side View at Top ot Head, 70-lb. Rail, Fig. 319. —Side View at Center of Head, 

50 DU. (Am. Ry. Eng. .\s.sn.) 70-lb. Rail, 50 Dia. (Am. Ry. Eng. Assn.) 

Figs. 316 to 321 presented by Wickhorst* illustrate the different grain found 
in the top and center of a new 70-pound Bessemer ral. A section about { inch 

• Flow of Rail Head under Wheel Loada, M. W. Wickhoral, Am. Ry. Eng. & M. ot W. Aflsn., 
Vol. 12, Part 2, 1911, p. 535. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 433 

tJuek was taken from, the rail and two pieces cut from it for microscopic test, 
as shown in Fig. 322. These pieces were polished on three sides, etched with 



Fia.320.— Transverse View at Top of Head, 70-lb. Flo. 321, —Transverse View at Center of 

Rail, 50Dia. (Am. Ry. Eng. Assn.) Head, 70-lb. Rail, 50 Dia. (Am. Ry. 

Eng. Assn.) 

10 per cent solution of nitric acid in alcohol, and microphotographs made, mag- 
nified 50 diameters. Thus, horizontal, vertical transverse, and vertical longi- 
tudinal sections were obtained at 
the top and at the center of the 
head. 

There will always be some 
difference between the structure of 
the center of the head and the por- 
tion near the surface, but when the 
rail is rolled at a proper temperature 
during the passes, when considerable 
work is put upon the piece, this 
difference will not be serious. 

The effect of finishing temper- 
ature is not fully agreed upon, and 
many rolling-mill men feel that 
the properties of the steel depend 
quite as much on the amount of 
reduction in the rolls as upon Qie finishing temperature. 




Fia. 322. — Pieces tor Microscopic Viewa shown in 
Figs. 316 to 321. 



,y Google 



434 STEEL RAILS 

* To try to arrive at some concluaon in this matter, a number of tempera- 
ture readings were taken with both the Fery and Wanner pyiximeters and 
checked against a thermo-couple at one or two large plants. For steel rails the 
finishing temperatures as indicated by the optical pyrometers averaged 1050 
to 1100° C.;t for structural steel, 950 to 1000° C. And yet such material is not 
coarse-grained. (Reheating to such a temperature would give rail steel a very 
coarse grain.) A difference of over 100° C. in finishing temperature could not 
be detected in the size of grain, but a difference in section could very soon be 
noticed. Similar results can be reached experimentally by rolling out small 
sections at different temperatures. A section heated to 1300° C. and rolled 
out with 30 per cent reduction showed about the same sized grain as one heated 
to 1300° C, cooled to 900°, and rolled out, the finishing temperature being about 
700° C. 

The question of rolling at a low temperature is one that has occupied the 
minds of engineers for a long time, and the fact that at present no solution has 
been made which is satisfactory to both the manufacturer and the consumer 
is evidence that it is not easily put aade. 

In rolUng early rails it was recognized that mechanical defects would be de- 
veloped to a greater or less ejrtent by the rolling process, and therefore the bloom, 
about 7 inches square, was conveyed from the blooming rolls to a steam hammer 
by which all visible cracks or defects were chipped out, care being taken to cut 
to the bottom of the imperfection, and not leave any pronounced shoulders at 
the edges of the resulting depressions; and until the adoption of automatically 
operated tables attached to the rail rolls, if the partially formed r^ls still showed 
defects, the operation of rolling was halted, while such places were chipped out 
by hand. These were usual practices, and were not abandoned because of their 
results being unsatisfactory, but on account of the time consumed and the 
expense incurred. 

In the formation of the grooves in the rolls much damage can be and often 
is done to the steel. With the object of increasing the product of a given mill, 
the ingot is rolled off at one heat, with heavy reductions in each pass so as to 
reduce the number of passes and consequentiy the time taken in rolling. 

It is interesting to turn to the following review of the English practice by 
Mr. Talbot: t " Our practice '& to take large ingots and have a furnace between 

* Some Practical ApplicatiooB of Metallography, Campbell, Proceedinp American Society for 
Testing Materials, Vol. VIII, 1908, p. 353. 

t These temperatures are from 100 to 200 degrees hinher than the usual practice. 
t On Rail Steel as Manufactured by the ContiouoUB Open-hearth Procees, Talbot. Proceedings 
American Society for Testing Materiab, Vol. VII, 1007. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 436 

the cogging and finishing mills,* which has the effect of acting as an equalizer 
so that the blooms are delivered to the finishing mill at ^i even temperature, 
making the bar more easily shaped, and the flange of the rail is sent out of the 
finishing groove at a temperature nearer to that of the head than has hitherto 
been possible. This, no doubt lessens the strains set up in cooling on the hot 
banks. Our practice is to increase the number of passes, decrease the amount 
of reduction per passj and get the product by increased speed of the rolls and 
not by digging into and tearing the metal, as is done in the case where too few 
passes and heavy drafts are adopted. 

" With regard to rolling temperature, we may say we roll a 100-poimd rml 
in lengths which give, after crops are cut off, three lengths of 10 meters. The 
first length is cut within 15 to 20 seconds after leaving the finishing groove of 
the mill, and on this we allow 71 inches shrinkage. The next length is cut within 
35 to 50 seconds from leaving groove, and here 7} inches is allowed. The third 
is within 60 to 80 seconds, and 7 inches is allowed for shrinkage. Of course these 
allowance only apply to 100-pound rails; less allowance is made in lighter rmls." 

t At one of the large rail mills they formerly had a table on which the rails 
were held before they went through the finishing pass. The scheme of holding 
the rail before the finishing pass, where only a small amount of work is put upon 
the metal, as illustrated in P^gs. 312, 313, 314, and 315, while giving a low 
finishing temperature, does not necessarily decrease the size of the grain. 

Starting with such a scheme as a basis and designing the rolling mill to 
hold the pieces to allow them to cool before the passes where the most work 
is done, and also arranging for the sorting out of the blooms to equalize the 
finishing temperature, would give a better arrangement and would have merit 
if applied to mills where the rails are finished too hot as a direct result of pro- 
ducing large tonnage. However, even with the most careful manufacture, a 
certain amount of heat on the bloom is necessary in order to take the A. S. C. E. 
sections through with the flanges properly filled out, and it will not be possible 
to reduce this temperature without the danger of the breakage of rolls, improper 
filling out of the flange, and additional strains in the steel, which are necessarily 
detrimental, unless the design is changed to make all parts of the section more 
nearly in balance as to the temperature of finish. 

An interesting experiment may be tried on a tee rail, which has been finished 
and str^ghtened. Take 6 or 8 feet of rail and place it on a planing machine 
and cut the head off the web at the point where the web joins the head, and both 

* Thia is customary at a. number of rolling mills in America. 

t Eenoedy-MorrisoD Frocwa. The Iron Age, December 20, 1900, pp. 16-18. 



>, Google 



436 



BTEEL RAIIfi 



the head and bottom portion will spring out of a str^ght line, sometimes to a 
very marked extent, thus showing that great internal strains are there. This 
is a condition that cannot be avoided by the manufacturer without some help 
from the rail deagner. 

The better distribution of metal in the new American Railway Association 
sections gives a rail that can not only be rolled at a lower temperature, but 
which is much less liable to injury in the straightening press. 

The rolling of these rails has developed some surprises. It was expected 
that the rails could be rolled at a lower temperature than the old sections, and 
that the shrinkage allowance could be reduced; but it was found that under 
the same conditions the new section would require a greater shrinkage allowance 
than the old. The rails were unquestionably rolled colder than the old section, 
with the exception of the thin flange, but it was this thin flange that determined 
the shrinkage of the old rails. In going through the cambering wheels the head 
was stretched, giving the hot head a greater length for shrinkage than the base. 
In the new section the temperature is nearly uniform and much colder than the 
head of the old rail was, but no part of the new rail is as cold as the thin base 
of the old rail; consequently a greats shrinkage allowance is required. 

At Gary* the ingots are bloomed to 8 by 8 Inches in 9 passes and finished 
in 9 passes, making a total of 18 passes from ingot to rail. The reduction of 
the rail is from 400 square inches at the bottom of the ingot to 10.1 square inches, 
or a reduction of 39 times. The areas of the various passes, as furnished by the 
Steel Company at Gary, of their section 10030, which is the A. R. A. type " B " 
100-pound ral, are as follows: 



P^K^b.. 


Am 


I Pw Number. 


a™.. 




SqWre InehesT 




Square Inches. 


Ingot 


400 


10 


43,2 




376.6 


1 11 


32,9 


2 


2S2.4 


1 12 


25.2 


3 


214. S 


13 


21.5 


4 


164 8 


14 


17.8 


5 


130 3 


15 


16.4 


6 


107 9 


16 


13.2 


7 


88.9 


17 


10.7 


8 


70.8 


IS 


10.1 


' 


589 







At the Maryland Steel Company f the blooms remain in the soaking pit 
about 1 hour and 25 nunutes, and are rolled to 7f by 7f-inch blooms in 13 passes, 
the top end of the ingot forward, and turned after each two passes. The blooms 

■ Report to Rail Committee, ProceedinRB Am. Ry. Eng. & M. of W. Aasn., Vol. 12, Part 2, 
1911, p, 430. 

t Report to RaU Committee, ProceedinKs Am. Ry. Eng. Assn., Vol. 12, Part 2, I9!l, p. 388. 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



437 



are rolled directly into rails without reheating, in 11 passes, making a total of 
24 passes from ingot to rail. The area of cross section in square inches in the 
various passes for the A. R. A. type " B " 90-pound rail, Maryland Steel 
Company, section No. 162, is about as follows: 

TABLE XCIII.— REDUCTION OF AREA IN 90-LB. A. R. A. TYPE "B" RAILS ROLLED 
AT MARYLAND STEEL COMPANY 



Ingot 

Blooming.. 
Blooming. . 
Blooming.. 
Blooming. . 
Blooming. . 
Blooming. . 
Blooming, . 
Blooming. . 



Blooming. 



Blo< 



ling 



I' Blooming. 
' Blooming. 
I Blooming. 
. Roughing. 
! Roughing 
[ Roughing 
Roughing. 



10 


1.7- ' 


.• n 




■1 12 


74 


1 13 


58 


'1 1 




,1 2 


34.4 


3 


31.3 


.1 i 


23.1 



Roughing. . . . 
Roughing. . . . 
Intermediate. 
Intermediate. 
Intermediate, 
Intermediate, 
Finishing. . . . 



21.5 
J7,8 
16-0 
15.2 
12.3 



Table XCIV presents data on American rolling-mill practice. Fig. 323 
shows the arrangement of the rail mill of the Algoma Steel Company's plant. 
This consists of one 32-inch rev. bloom mill, three Seimens heating furnaces 
for reheating the blooms, and one 23-inch three-high rail mill. 

Plate XXX illustrates a reversing cogging mill and Plate XXXI and 
f^g. 324 three-high mills. The left hand illustration of Fig. 325 shows a set 
of rolls for use in a three-high mill, according to common English practice, in 
which the bottom and middle rolls are grooved to receive the rail, while the 
"closers" are on the middle and top rolls, and the guards to peel the bars out 
of the grooves rest by gravity on the bottom and middle rolls. 

In the American three-high mill, shown on the same figure, the top and 
bottom rolls are grooved and the middle roll serves as the "closer" for each of 
the others, itself carrying no grooves at all; to enable the bare to be got out of 
the top groove, the upper guard must be placed on the upper dde of the bar as 
it issues from the roll, and as it will not lie there by gravity, the guard has ^o 
be kept up to the roll by a counterweight or spring, and is known as a balanced 
guard.* 

f^g. 326 shows the three-high rolls in the r^l mill at Gary. fThis mill is 
equipped with 12 sets or stands of roll trains, all operated at varying speeds by 
General Electric alternating-current motors, some of which are of the largest 

* steel, by Harbord and HaU, London, 1911, p. 625. 

t Raiboad Age Gazette, November 13, 1908, and The Iron Age, April 1, 1900. 



,y Google 



STEEL RAILS 
TABLE XCI v. — DATA ON AMERICAN ROLUNG-MILL PRACTICE 

(CompilalbyCamiiutUionRul, Am. Ry. E«. Amb., im, ukl revioKiby themotbor IllS) 









N 


mbarof 


! 


1 


pl 


■ 






Bloami(« Hill. 


PuMin 


p- . 


1 










8ailT««. 


IJ 


2 


taS). 






r ,■ 






P 


m 




RullfUI. 






1 














¥ 


II 

r 


1 


i 


i 


! 

So. In 


I' 
i 












m. ». 










Al«onuiSt«JC(.. 


Ou>adiaoeoo, 
Cauda. 


181 -xim 


tl 30 


ts'xs- 


It 


B 


* 




M 




Bknnia rehaitMl. 


tBtthlabem Slsd 


B«Lhl«baiD. 


1B-X23' 


1 20 


S-X8- 








M 8 


2« 


1 




Co. 


Pa. 




















hit taken Inm molik 
■ltd are railed direcl irith- 
oul lurlhcr reheati^. 


'Cambria SMel Co 


)obanawi.P.. 


M'xa- 


1 ii 


si'xio- 






t 


iS 00 


a 


S M 


liwoli! an nbHted, rolled 
into blooins. mtxia w«- 
heated and nlM iaio 

prior' to GabgbiH pui. 


Cani«ii*3l»ICo 


Breddock. Pa. 


"'"' 




SJ'XW 




m 






(M 




liKOU are rebHted. roUsd 
into bk»nik mil- re- 
bMtad and railed inia 
laUs. Uaib an «iren a 

EniahiBi PH. oThom 














































1 min, lo t mia. 41 sec. 


CanacieSlaelCo 


Braddock, P.. 


Hi'xisr 


















Old mill. Lii^ltHla. 
























Raila sol bM belir^ 


ICiniMig Steri Co. 


Y«™iowa. 
Ohio, 


l»'X!l' 


» 


8'X8- 


» 


10 


■ 


3im 


JO 


4 » 


pa». 


Colnrado Fud A 


Pueblo, Colo. 

SydMy. 
Canada. 


wxzr 






17 


10 


; 




38 






Iron Co. 
t Dominion Iron A 








StaalCo. 








•IlllBoaaUBiCo, 


a Chiowo.III 


tl8'XI9i" 


tso 


8'XS' 








31 W 






Raib bald 30 to 80 ncoida 


Indiwia StaaL Co. 


Gary, Ind. 








9 


8 


, 




18 




before Rniahint Pua. 
No rett prior to fiaiabinc 


•L.^™™ St-l 


Bdffnio, N. Y. 


IB'XIB* 


i « 


8'XS' 


e 


4 


« 


4» i 


IS 


S i 


'MaryUiud Stari 


Sparrows PI., 


M'XSI* 


! 10 


^'XTI' 


13 


, 


, 


m 


M 


5 00 


l^u an rebaated aflar 




Md. 




















lad aie rolled direcl wiib- 
OBt further batii«. No 
r«t i. livs. ocapt ibai 




SleelloD. Pa. 


!8('X181' 




rrx7i- 





6 






30 






StnlCo. 






















bSrslSTcS^" " 


tTnuiMH Coal. 


















M 


■to«" 




lmi.*R.R.Co. 


Ala. 























in Irom R. W. HuoiA Cto., t 



which waa obtained di 



t Compiled by autbor. 



sizes ever constructed for industria service. These are housed in a separate 
bay, or lean-to, running parallel with the rolls. The rotors are 20 feet in 
diameter and have a speed of 83 revolutions per minute. All of the motors are 
connected directly to the roll trdns by regular mill couplings. Although the 
motors are provided with flywheels and run in one direction, provision is made 
for reverang in case of necessity. The control system has been worked out 
with the greatest nicety, all operations being under the instant control of the 
operator by means of a master controller. 

The first group of rolls consists of four stands of continuous 40-inch mills, 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 




>, Google 



STEEL RAILS 




Fio. 324. — Housing for 28-inch Three-high MiU. (HEirbord and Hall.) 





ish Three-high Rail Mill, American Three-high Rail Mill. 

Fro, 325, — Rolb used in Three-high Rait Mills. (Harbord and Hall.) 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 441 

each two of which are driven by a 2000-h.p. motor. They are arranged in 
tandem, requiring no manipulation from stand to stand. Here, as elsewhere 
through the plant, sufficient distance is left between successdve sets of rolls to 
enable a quarter turn of the ingot or bloom to be made, so that it is worked 
equally on all sides. The first two mills are equipped with 42-inch rolls, 
enabling 20-inch by 24-inch ingots to be used. After passing these four mills 
the ingot is sent to a 40-inch three-high blooming mill equipped with lifting 



Fia. 326. —Three-high Rolle in the lUil MiU at Garr. (Scientific American.) 

tables and arranged with a combined hydraulic and pneumatic balancing device. 
This mill, which is driven by a 6000-h.p. motor, gives the ingot five passes. 

After being bloomed the ingot is sheared in a 10-inch by 10-inch horizontal 
blooming shear, and the crop ends or butts are taken outside of the mill by a 
butt conveyor of unusual construction, which was designed and built by the 
engineers of the Indiana Steel Company. Each bloom then goes through a 
28-inch roughing null, which is three-hi^ and equipped with tilting tables. 
This null has three stands or rolls. The roughing stand, however, is the only 
one that is three-high, the other two stands being two-high. The mill is driven 



.y'Google 



4^ STEEL RAILS 

by a 6000-h.p. motor and ^ves the bloom three passes. After leaving the rough- 
ing mill the bloom goes through a two-hi^ 28-inc:h forming mill driven by a 
2000-h.p. motor, receiving but one pass. Then it is sent to finishing mills, which 
consist of five stands of 28-inch mills driven by two 6000-h.p. motors. 

After the dummy pass, the bloom is transferred to the first edging, which 
is in this same mill but the second stand, and turns back on an elevated table 
to the second edging, which is in line with the 28-inch rou^ng mill. It then 
travds by chain transfer to the lower tables and on the leading pass goes through 



DUMMY 2e''x40" 




l»T ROUGHING F»SS 
28"X60" 



Pass Diagram, Rail Mill, Illinois Steel Company, South Works. 



2>f> R0U6HING FV^SS 
ZB'x-56" 

Pia. 327. - 

a stand, which is also in hne with the roughing mill and driven by the same 
motor, and continues on to the third stand of the 28-inch finishing mill, this 
being the eighteenth and last pass. After the finishing pass the rail travels 
through to the saws, of which there are five provided, thus cutting four rails to 
length. These four rail lengths consist of half the ingot. As the capacity of 
this mill is 4000 gross tons per 24 hours, it will be seen that there must be a 
four-rail length sawed about every half-minute. The saws have 42-inch 
blades, arranged to be raised and lowered in unison by one controller from 
the hot-saw operator. 

The pass diagram of the rail mill at the South Works plant of the Illinois 
Steel Company is illustrated in F^g. 327. The Bessemer ingot is 18 inches by 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 




*%! 



>, Google 



444 STEEL RAILS 

19i inches, the heating capacity is 192 ingots (24 single-hole pits). The ingot 
is worked direct to rail without reheating. The blooming mill is 40-inch pitch 
diameter three-high, and the ingot is ^ven 9 passes and reduced to an 8-inch by 
8-inch or 8-inch by 8J-inch bloom. The number of rail lengths rolled are three 
and four. 

The finishing mill eonasts of one stand 28-inch P. D. three-high first rough- 
ing rolls, one stand 28-inch P. D. two-high second roughing rolls, one stand 28- 
inch P. D. two-high dummy rolls,one stand 28-inch P.D. three-high finishing rolls. 

The number of passes from ingot to rail is as follows: 

Blooming 9 

First roughing 3 

Second roughing 1 

Dummy 1 

Finiahing. 4 

Total 18 

Fig. 328 presents the general arrangement of the rail mill. 
Table XCV shows the shrinkage allowed by American rail mills and 
Fig. 329 the lengths of saw runs. 





TABLE XCV. - SHRINKAGE ALLOWED BY 

(Compitod byCommitlaeon Ruil. Am. Ry. En*. Amu., 190 


AMERICAN RAIL 

V. und reviBKl by the author 


MILLS 

912) 


No. of 


Rail Mill. 


LooatioD. 


Tim. 

of 

Saw 

Ruu.t 


SbriDkac* Alloirad by Milla on 33-IaoI Raili. 
iMhs. 




I. 


T* 


BO 


ss 


M 


M 1 IH 




Algoma Steel Co... 
■Bethlehem Steel Co. 
•Cambria Steel Co. . . 
Carnegie Steel Co. . 
Carnegie Steel Co. . . 
Carnegie Steel Co. . , 
Colo. Fuel & Iron Co 
EJominion Iron & 

Steel Co 

•Illinois Steel Co.... 
Indiana Steel Co.. , 
•Lackawanna Steel 

Co 

•Maryland Steel Co... 
•PennBylvania Stee! 

Co 

Tenn. Coal, Iron & 
R.R.CO 


Canadian Soo, Canada. 

Bethlehem, Pa 

Johnstown. Pa 


35 
20 

17-19 
12-14 

'12-17 

*i7" 
10-20 

16 
26 

5 

27 






6 

6 

il 


6i 






2 

3 


6 
51 


6A 


1 


.,'*.l ? 




Braddock! Pa 


I' 

6i 

6 " 

ti 


61 






•1 

6« 


11 


it 

is 

V 

61 

ei 








8 
9 


Sydney, Canada 

South 6hicago, 111. 


::::;. 1 


11 

12 
13 

14 


Buffalo, N. Y 

Sparrows Point, Md. 

StMlton, Pa 


6 

1 a 

1 



Nm. ~ labmutiao Imm R. W. Hmt A Co.. ucspt that marksd (■}. wbicb vas ot 
t Time It wcoDili anwuiMd ham tbe tima rail luvas Hniihinc rolts tilt sair dropa. 



After leaving the saw the rals pass through the cambering machine and 
are ^ven a head sweep (Fig. 330) of from 3 to 8 inches, the A. S. C. E. sectjon 



,y Google 



INFLUENCE OF DETAIL OF MAKUFACTHEE 



SAW RUNS 



SAW NO I 2 



- 2O7-0" - 



' I ,., °1 



SAW NO , I I 1 1 RAIL STOP 



— 105'- O" 

SANV NO. 



ii—^ 



I ll RAIL i 



"J 'r 



-154-2" ► 



^fcfc 



— 2J8'-0" - 



JL ' I ° l 



- 69 -O' — » 

SAW NQI I a I 5 1 



Fio. 329. — Saw Runs of Amerioan Rail MMIa. (See Table XCV for number of millfl.) 

requiring a greater ordinate than the A. R. A. rails. The rails then pass to the 
hot beds and after bang allowed to cool are transferred to the gagging or cold- 
straightening presses (Rg. 331). 



Fia. 330. — Head Sweep. 

In the heavier A. S. C. E. sections the gagging or straightening of the rail 
after it leaves the rolls and is cooled in the hot beds tends to develop injurious 
strains in the base and web. In some of the most modem English plants the 



,y Google 



STEEL RAILS 



rails are run while still hot through straightening rolls, thus much reducing 
the labor of final straightening.* 



Fiu, 331 . — Cold StraightenJDg Press, MarylaDd Steel Company. 

After straightening, the rails are inspected, drilled, reinspected, and loaded 
on cars for shipment. Many rails were formerly damaged in loading, but at 
several mills the use of a magnetic crane is now employed for loading the rails. 

The brand on rails gives the name of the manufacturer, a number or abbrevia- 
tion by which the rail section is designated, the month and year of manufacture, 

* 3. von Schukowski advocates Btrai^htening rails while still hot. Das richlen voa eisenbafan- 
Bchienen im katten und warraen zuatande. Stahl und Eiaen, Vol. 27 (1907), Ft. 1, p. 797. 



,y Google 

^ i 



INFLUENCE OF DETAIL OF MANUFACTURE 



447 



and, if the metal is open-hearth steel, the letters "O. H." are also added. Some- 
times the letters "F. T." are added to signify ferro-titanium steel. Square 
block letters and figures about one inch high are commonly used, and, as these 
are cut into one of the rolls of the last pass, the brand will always appear slightly 
r^sed at regular intervals on the web of the rail. The month is generally 
shown by Roman numerals, as VII for July, and sometimes by a series of I's, 
as mil for May. 





- 




*^ 










































1 

r 


17 






J 








\ 


^ 








■b 






•s 


^ Jk 






rr 








o 


















-~ ~~ 


te 














EMPERAT 


RE IN • 








66 1020 sao 



Fia. 332. — Value of V/E for Tables XCVI, XCVII and XCVIIL 

The number representing the heat, blow or melt of steel, and the letter 
to indicate the position of the rail in the .ingot is stamped on the web of the 
rail with dies while it is still red hot, but after it has been completely rolled and 
sawed to length. As the brand always appears in raised letters, and the heat 
number and letter is stamped on no confusion of the two should exist.* 

A series of experiments were made in Germany by Dr. Puppef to determine 
the power required to roll different sections. As this investigation presents 
many features connected with the design of the section and rolls and the effect 
of the temperature on rolling, it will be of interest to briefly review Dr. Puppe's 
work as it relates to the heavier weights of rails examined. 

The reverang mill on which these tests were made consisted of one cogging 
mill houang and three finishing mill housings. A flywheel converter set of 
the Ilgner system served to equalize the fluctuations of the power taken. The 
mill was driven by three motors rigidly coupled together, and connected in ; 

* R. W. Hunt and Company, 1121, The Rookery, Chicago, 111., have published in convenient 
Torm full information in r^^axd to the practice of branding and stamping at the different mills. 

t Experimental Investigationa on the Power Required to Drive Rolling Mills, J. Puppe, London, 
19L0. See also iot a fuU treatise of this subject Steel, Hatbord and Hall, London, 1011, pp. 666-741. 



,y Google 



44S 



STEEL RAI1£ 



series electrically; they had an aggregate output of 3600 H.P. normal and 
10,350 H.P. maximum at a speed of 110 r.p.m. 

Tables XCVI, XCVII and XCVIII contain the results of the tests with 
the heavier rails. The second line of the tables gives the time taken by the 
pass (in seconds) determined from the rise and fall of the current and jrowa- 



Line 3 gives the time intervals (in seconds) between successive passes which 
were determined in the same way as the times of the passes. The time interval 
between the first and second passes is given in the colunm headed " pass 1," 
that between second and third passes in the column headed " pass 2," and so on. 
The last column contains the sum of all the figures in the preceding columns. 




Fia. 333. — Diagnun of Coding Rolls, Tables XCVI, XCVU and XCVIII, Dimensiona in Millimetera. 

The energy given up, or stored by, the flywheel and other rotating masses 
was calculated from the speed ou^es and the moments of inertia. 

To arrive at the light-load loss, the average speed was determined, and the 
light-load loss was obtained by multiplying the power taken to drive the mill 
hght at this speed by the time. 

The total work done in rolhng the bar up to any pass is the sum of the work 
done in each individual pass up to that point, taking due account of the motor 
eflSciency and leaving out the work expended m accelerating the moving parts 
or in light-load losses, copper losses, etc. Frequently, the section of the ingot 
during the first few passes could be determined only roughly, and then the laigth 
of the bar after the second or third pass was taken as unity, and the summation 
of the work done in rolling commenced correspondingly later. 

The areas of the cross sections of the bar or ingot were obtained wherever 
posable by cutting trial pieces off the bar, and measuring the area directly by 
a planimeter. It was not always possible to cut samples from the bars during 
the roughing passes, as the bar was too large for the bloom shears or hot saws, 
and in such cases the cross-sectional area was obtained from the scale showing 
the position of the rolls, while sections were cut from the bar in the last two 
passes, and the area measured as a check. It is often found that billets are 



,y Google 




>, Google 



p 



'Sy^rmnmW^lP'lh 



"' ^i 1P»^gPi§"l i^''l ^ ! 



5^; iHj ;||i8sg§i§|^s;| i=i"i = g 



-:-^ap^^P#-iiii"i 



= 5 



«;:»i;j|;ssp8j|5"|j§'^iij 



i;^S6:jiis3psgri|i=--iSj 



2'S jSss 



p^ 



is:,5s:#.p^r=iir-|ig 



==*-^ 



S:: S;, 



:6.|Si§|!Bgf-S. 



1% 



"3 ;;=8 ;pss»;8a8.f "I 



..B„ , 



:=::S.B:||ss.p8Bf"||is--|i| 







Mi 



13 



II 



I 



an 

m 

m 

in 

! ill; i 



lili 



i 



iili! 
ipil 

■-"II « 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



I m ill 1 1 



P 







"'' i^"! :S}°^-iP W"| 



I ^ 



114 

III! 



fijllll 






>, Google 



-=5 ;3SJ ;i|SJS||gJiS3-| P'^"| 



'5 



^Hw^^ww^i 



"-Mmum^-s^lij 



"'N:jpss|i5gf"|P--|ig I 



^mmnimw^s^lij ] \ 



I 

11 



^m^mmmw^l^ih 



21=;: 



' :!F8^g88ap=-| 



ir-jij 



IFili 



^ «i ;||Bss|u||S::| 



2'! ;3Si 



:S.E.iS!SISaiS"i 



S3s;aigp,,„„,;:g,|ffigij 




>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 




>, Google 





STEEL RAILS 


= 1 








'== ;=5i ||isajis8.|'="5 |5s""i i § 




-" H ;gps=ps|f ==| pi-"! s 1 




"5= Issi ;|iB8ss.sasr^3| |5s--i s = 






d 

^1 • 




Ii ■ 


-"^ ;S=i ;tS!5sp5||5:-| |jj-=i i 5 


a ■ 


-sis pSj ;|p3S|5|!|»==| ||i"-i i 3. 




"-= ;'=i ^S|B|5|.SSi§»=^| A^"l 3 § 




-"= :s5i ;g|ij!.iS5p5S=j |p--| s 8^ 








nmimyiiHiNi 


i: 

a : 

ft 

1 

1 



ii 

i 

it 






! 



>, Google 



a 


INFLUENCE OF DETAIL OF MANUFACTURE 
asia sslg § « 

!al8 s5l3 S 1 


s ; ; . : . — : 


1. 


a 


= := ;=«S. ||.»3S||s|,|S=5 ; 


.9 


r 


a 


5-!iS3=isssisiS»8IS=" : 


If 

u 


S^S ig =: g 


» 


== - — : 


1 




s 




5| 


'■" i- ' 1 


a 


S2o « „„„ 


P 


— 5 ■ * 


s 


'aSEesiiijJjjajpjs:! : 


J 


F 


sM ;| 5 1 


5 


"■ ■ ;S=j ||=«5S||8f s= 


|j 


■8 


= 


""^:^^iiil=-pilf= 


P 




= 


— ;2S| sssssiisgic" 


P 


8S= :| n 1 


3 


•2= ;=s| ii|8ss|58.||»='-| 1 ie-"| = s. 


a 


•S= ;»!ii ;?s5gsiissss::j | liiii i j 


J 

■a 

1 


li Bi. 

iii IL 



i 

1 

11 

SI 

ll 

p 



>, Google 



456 STEEL RAILS 

rolled to different cross sections in the same grooves in the cogging mill, but this 
is because the position of the screwing-down gear is altered. 

The cross sections of the ingot in each pass with the reversing cogging mills 
were determined from the position of the top roll. A special pointer was fixed 
to the top roll, which indicated the position of this roll on a vertical scale pro- 
vided for the purpose. The length of ingot or bar was calculated from its cross 
section and weight. The crop ends were frequently cut off by bloom shears 
after the roughing passes, in which case a corresponding allowance was made 
in the calculations for the finishing passes. 



nzn 



Fig. 335. — Sectiona in which only "Direct Presaiire" occuta in the Proeeaa of Rolling. (Puppc.) 

The elongation was calculated from the cross sections and the lengths, such 
calculations often being based on the cross section after the second or third pass, 
in order to obtain greater accuracy, as already mentioned. The " volume dis- 
placed " is obtained by multiplying the reduction in cross section by the length 
of the billet before the pass in question, i.e., {Qi - Qa) X L„, where d = the 
cross section of the ingot or billet before the pass, ft = the cross section after 
the pass, and L,, - length of billet before the pass. 
In rolling there are two means by which the 
pressure is applied to the bar, viz., " direct " and 
" indirect " pressure. By direct pressure " is 
meant pressure which exists between the surfaces 
of two rolls, as, for example, with rolls grooved 
as shown in Fig 335. 

The term ' indirect pressure " will be used 
to denote the pressure usually existing between 
the groove sides of one and the same roll, which 
produces principally a reduction in a horizontal 
direction (width) and not in a vertical direction 
(height), as is the case with the sections shown 
Fia. 336. — lUustratioo of "Indirect In F^g. 335. The following example will make 

Pressure." (Puppe.) this dear: 

If the flange of a rail (Fig. 336) be pressed in such a manner that its thick- 
ness 0-6 and a'-b' is reduced, then this must be due to indirect pressure along 
the whole line from c to d and c' to d'. It will be noticed that indirect pressure 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 457 

takes place here both between the sides of the groove of the same roll as also 
between the two rolls. 

Where the j)articles move solely in an axial direction along the bar, the 
energy required to accomplish this displacement is directly proportional to the 
product (Qi — Qi) X L,„ in which Qi and Qi represent the cross sections of two 
consecutive passes, and L„ the length of bar corresponding to the cross section 
Qi. For shortness, we will denote this product, which represents the volume 

displaced, by V. The fraction (Q^ - Q^) ^X L^A^'^ f- ^^-^ depends on the 
*^ •' work done m rolling {m m. kg.) 

plasticity of the material, and, therefore, to a large extent on the temperature. 

By plotting a curve with the above fraction calculated for each pass as ordi- 

nates, and the corresponding temperatures as abscissae, we can obtain from this 

the number of cubic millimeters of material which will be displaced per m. kg. 

at various temperatures. 

Let us now investigate the case in which the particles of the metal do not 
move mainly in an axial direction along the bar, but also at right angles thereto. 
A simple example is that of a bar which is flattened out by rolls having no grooves 
to restrict the movement sideways. 

The angle of inci^on into the bar should be such as not to cause excessive 
spreading. Incisions at acute angles naturally cause greater spreading than 
inciaons at obtuse angles, but very often the proper form at the first passes can 
only be obtained by incision of the bar at an angle from about 40° to 60°,* in 
which case a considerable amount of lateral 
spreading cannot be avoided. It is then 
advisable to facilitate lateral spreading and 
not to hinder the movement too much by 
Uie sides of the grooves. 

When designing rolls, successive profiles 
should be arranged in such a way as to 
obtain a smooth carve for V/E free from 

such sudden jumps as occur, for instance, f,«. 337. _ Effect of Inclin^Jion of inner 
in th6 curves in Fig. 332. It need hardly Surface ot the Rail Flange on Energy re- 
be said that the roUing of simple sections, "^"^ " ^^'"'- f^""^-* 
such as flats, squares, rounds, etc., requires less power than the more compli- 
cated rail sections in which the flow of the metal is brought about by indirect 
pressure, and where the loss due to friction gainst the ades of the grooves is 
great. This loss should, of course, be kept as low as posable, and equally dia- 

* Bartholme, Stahl usd Eiaea, 1907, p. 68. 




,y Google 



458 STEEL RAII^ 

tributed amongst the various passes. It can be reduced to a nunimum l^ 
making the angle oo (Fig. 337) as small as possible, so as to obtain a large com- 
ponent a. 

From the point of view of economy in energy consumption when rolling 
rmls, it is very desirable that the inner surfaces of their flanges be inclined as 
much as possible. If, however, the inclination of the inner surfaces, or the 
thickness of the finished flange, has been fixed, the amount of indirect pressure 
required for the formation of the flange can be reduced to a minimum by 
working it as thin as possible at the first forming pass. 

To sum up, the power required for complicated sections is greater than for 
simpler sections of the same final cross-sectional area, but this extra power 
depends to a very large extent on the skill with which the rolls are designed. 
If the rolls have the most favorable shape, the values for Y JE for the various 
passes will be consistent with one another. 

Plate XXXII contains curves which are calculated from the tests. The 
shaded areas represent the energy supplied by the motors to the mill, and cover, 
therrfore, not only the work required for actual rolling, but also the power 
required for running tiie mill light and for accelerating. The power taken by 
the cogging mill generally increases towards the end of the pass, owing to the 
acceleration of the rotating masses and the increase in the no-lead losses. 

The same holds good for a few of the roughing and finishing passes. At liie 
last passes, however, the power taken decreases when the maximum speed has 
been attained, and the rotating masses are no longer accelerated. The peaks 
at the beginning of the ciUTes for the last passes are due to this cause. The 
relatively high peak in the lower curve for pass No. 23 may be explained by the 
fact that maximum speed was reached in a very short time, as indicated by 
the sharp rise in the speed curve. 

The speed curve (Plate XXXII) usually drops rapidly when the ingot enters 
the rolls, and sometimes it even falls to zero, and then rises again. Frequently 
the speed increases suddenly at the end of the pass, especially in cogging mills. 
These irregularities in speed are due to careless manipulation of the driver's 
lever. In reversing mills the kinetic energy of the rotating masses is seldom 
used to assist the motor, but it does happen occasionally. 

From the acceleration curve (Fig. 338). it will be seen that a very large 
amount of energy is requu^ to accelerate the rotating masses of a reverang 
mill; for instance, about 5300 h.p.-minutes are reqmred to bring the rotating 
masses of this reverang mill up to a speed of 120 r.p.m. Assuming that tiie 
average speed is only 80 r.p.m., and that there are 20 passes, this means that 



,y Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



459 



48,000 h.p.-minutes are expended in accelerating the rotating masses per ingot. 
This is a large percentage of the total energy required for rolling. It must be 
noted, howevo", that in the case of electrically driven reversing mills the energy 
expended in accderating the rotating masses is largely returned again in the 
form of electric energy when the speed decreases, so that the net amoimt of 
energy required for acceleration purposes is not considerable. Where reverang 





- 






~ 






~ 






'~ 






'~ 


~ 






































J 


































t 




































































^~ 


JIW 


































































""IZI- 




































































_ -iZ: - 




































































I 




































































\-l-^. _ 






































































aeoo 
































:z : : 




































































_,. _ . . 
































/ 




^ 
































































































} 
































/ 












































































Jllk 




















































/ 
































































' 


























































» 




-A 








f 


L 


1 


'i« 


«: 


9 


.! 


ulr. 


ti ao m m M 



Fig. 338. — Work done in Accelerating the Rotating Masses in Reversing Mill. (Puppe.) 



steam en^nes are employed, however, the work done in accelerating the rotating 
masses is lost, and this should he borne in mind when deciding whether for a 
given mill a reverable engine or a flywheel engine is best. 

Rails after becoming so worn in the track as to be unfit for further service may 
be rerolled into new sections. A consdderable tonnage of such r^ has been re- 
rolled by the McKenna process at a cost of about $7.00 pa- ton. The advantages 
claimed for this process are: 

BTret. That the worn rail is selected material, as the imperfect rails have been 
eliminated to a large extent during the time the rails have been in service. 



,y Google 



460 



STEEL RAILS 



Second. The rerolling puts additional mechanical work upon the material, 
which should improve its quaKty. 

The practical difficulties in connection with the process may be summarized 
as follows: 

Ilrst. The rail must not be worn beyond a co-tain degree, as tiiere must re- 
main in the head of the rml a suffident amount of metal to form the new head. 
The metal from the base and flange cannot be made to flow into the head to make 
up for its worn conditiffli. In general it is not desirable to remove rails from the 
main track, which show only the amount of wear best adapted to rerolling; how- 
ever, wheal it is necessary to remove rails from some cause otho" than reduction of 
section on account of wear it may prove ewnomical to reroll them into sections of 
lesser wdght. 

Second. It is desirable that the rerolled rails may conform to standards already 
in use. This seems to be in a practical way difficult to accomplish, owing to the 
varying shape of the head of (iie rerolled rail, as the rail is much or little worn. 

The following figures are taken from a record of a mUe oi 80-pound r^ 
rerolled by the American McKenna Process Company. 

10,560 ft. of rail taken out of track Tooi. 

Weight when new 125.71 

Loss in track after 10 years' use 5,33 

Weight sent to mill 120.38 

10,151 ft. of rail received from mill 103.20 

Scrap received from mill 17 . 18 



The sections pven in Fig. 

— -Ji- *i 



* and Plates VII and VIII of the rails used 





Pennayivania, 100-lb. 

Fio. 339. — Recent Rail Sectiooa. (R^lroad Age Gacette.) 



U. <j- 

Penntylvania, 8B-lb. 



>, Google 



INFLUENCE OF DETAIL OF MANUFACTURE 



j. 2^. 




Canadian Pacific, 8G>lb. 




Baltimore A Ohio, go.lb. 



r* — -« — •-! 



--*--1 



r — "' — I/- f> TT 




Santa Fe, es-lb. 



Burlington, 90'lb. 





Hsad. 


WA. 


But. 


ToW. 




42.2 
33.8 
37.0 
36.2 


17.8 

22,2 
22.8 
24.0 


"lo™'' 

40.0 
41.0 
40.2 
39.8 

























Fia. 339 (continued). — Recent Rail Sections. (Railroad Age Gazette.) 



>, Google 



462 STEEL RAIIB 

at the present time show clearly the willingness on the part of the railroads to 
meet the criticism of the manufacturers in regard to the faults in tbe deagn of 
the heavier A. S. C. E. sections. 

These new thick-base sections, adopted after the studies of 1907, cool with 
less curvature than the former thin-base types and require less cold-straighten- 
ing. There appears to have been a material reduction in the number of base 
fsdlures in the new sections as compared with the A. S. C. E. design. 

A number of students of this subject think that there is room for still 
further improvements along this line to eliminate to an even greater extent the 
failures in the base of the rail. This is evidenced by the new Chicago and 
North-Westem section for 100-pound rail in which the base is |^ inch thick 
at the outer edge. Dr. P. H. Dudley has designed a new section ^ving to the 
fillet between the base and the web a radius of 1 inch. The Illinois Steel 
Company at its South Works plant is also expmmenting with a new section 
of 110 pounds in wei^t which is similar to that used on the foreign railways 
in that the upper surface of the base is broken at two angles. 



,y Google 



CHAPTER VII 
RAIL SPECIFICATIONS 

34. COMPABISON OF AMERICAN SPECIFICATIONS 

The r^ committee of the American Railway En^neering and Maintenance 
of Way Association revised their specifications in the latter part of the year 
1909; these were subsequently withdrawn and in March, 1912, the committee 
presented to the American Railway Engineering Association, which was now 
the name of the association, specifications for carbon steel r^ls. 

Some par^raphs, such as those relating to carbon under remarks in Table 
XCIX and Nos. 14 and 15, relatmg to physical requirements, were not conddered as 
final, it being thought that the committee did not have sufficient information in 
its possession to make these sections in the specifications mandatory. The re- 
quirements in section 14 for ductility were somewhat lower than some of the 
members thought desirable. Paragraph No. 15, referring to deflections as a 
method of clas^ying rails, is also tentative, and it is the intention, when suffi- 
dent data is at hand, to prescribe maximum and minimum limits for deflections 
under the drop test. The committee, will continue its investigations and the 
gpedfications in these respects will be subject to change. 

On January 10, 1912, the Pennsylvania revised its specifications for 
Bessemer and open-hearth rails. 

On January 1, 1909, the Association of American Steel Manufacturers 
issued standard specifications for Bessemer and open-hearth rails. These rail 
spedfications, adopted by the steel manufacturers of the United States and 
Canada, which are practically the same for the different companies, indicate 
the views of the rail makers as to the proper tests and chemical composition for 
securing good rails. The A. S. C. E. sections are still officially regarded as 
standard practice. 

Standard specifications for Bessemer and open-hearth steel rails were 
adopted by letter ballot on August 16 1909, by the American Society for Testing 
Materials. 

In October, 1909, ihe Harriman lines adopted standard spedficatimis for 
Bessemer and open-hearth steel rails; the open-hearth specifications were subse- 
quently revised in February, 1910. 



,y Google 



^4 STEEL BAII£ 

The above spedfications ^w the development during recent years of rail 
siiecifications in this country and an examination of th^ requirements will prove 
of inters. The American Railwi^f Engineering Association specifications of 1912 
reflect the latest thought and are noticeable for the increase in the number of 
physical tests ovk- those required in earlier specifications. A great many defects, 
such as piping of the ingot, can be adequatdy guarded against by proper physacal 
tests, and in gaia^ it would appear dedrable to leave the producer free in 
such cases to adopt his own methods of manufacture. Within certain limits, 
however, the specifications may well be drawn to racclude practice which is known 
to result in defective material. The desirability of doing this is emphasized by the 
great difference in quality found in rolhngs from different mills and in some cases 
for rails from the same mill, but rolled in different years. The specifications giv«i 
m Article 35 are a good example of spedfications drawn with a view to eliminating 
defective practice at the mill. 

The trend of recent specifications is to increase tiie amount of inspection which 
is brai^ ^ven the rail at the mills. The plan of the R. W. Hunt and Company of 
placmg inspectora throughout the mill to watdi the entire process of manufacture 
is evidence of this. 



Inspection 
Am. Ry. Eng. Assn.: 

1. Inspectors representing the purehasra" have free entry to the works of the 
manufacturer at all times while the contract is being executed, and shall have all 
reasonable facilities afforded them by the manufacturer to satisfy them that the 
rails have been made in accordance with the terms of the spedfications. 

2. All tests and inspections shall be made at the place of manufacture, prior to 
shipment, and shall be so conducted as not to mterfere unnecessarily with the oper- 
ation of the mill. 

All of the specifications are substantially the same as the above. 

Maierial 
Am. Ry. Eng. Assn.: 

3. The material shall be steel made by the Bessemer or open-hearth proce^ 
provided by the contract. 

The clause ui reference to material in the Pennsylvania Specification is the 
same, but in the other spedfications it is omitted and a separate specification 
written for each class of material, i.e., Bessemra- or opwi-hearth sted. 



,y Google 



RAIL SPECIFICATIONS 465 

Chemical R&piirements 

Am. Ry. Eng. Assn.: ' 

4. The chemical composition of the steel from which the rails are rolled, de- 
termined as prescribed in Section 7, shall be within the following limits: (See Table 
XCIX.) 

Table XCIX presents a comparison of the chemical requirements of the 
different specifications. 

The Committee on Standard Rail and Wheel Sections of the American 
Railway Association, in its report of March 23, 1908, to the association, stated: 
" In the matter of chemistry specifications for Bessemer steel rail, statistics 
were obtained from the officers of the Ore Producers' Association which con- 
vinced the committee that it would be impossible for the mills to furnish more 
than a small percentage of the total rail requirements of the railways with a 
phaiphorus specification less than .10. 

" The optional specification for .085 phosphorus prepared by the joint 
conmiittee of manufacturers and railway men is now in the hands of all members, 
and is, therefore, available for use by those who are able to obtain low-phos- 
phorus Bessemer rails. It is not considered proper, however, to require less than 
.10 phosphorus in a specification intended for general use. Members desiring 
to obtain a low-phosphorus rail will have the further option of using open- 
hearth steel. 

The committee conferred with a nimiber of disinterested experts on the 
phosphorus question, and among the principal authorities consulted were 
William Metcalf, of Rttsburg, Robert Forsyth, of Chicago, and Henry M. 
Howe, of Columbia University. These gentlemen all agreed that it would be 
unreasonable to require less than .10 phosphorus in a specification for Bessemer 
rails intended to cover purchases for all American railways." 

The Pennsylvania specification for open-hearth rails makes the upper limits 
for classification A, phosphorus .03 and carbon .83; for elasafication B, .04 
phosphorus and .75 carbon. The desired carbon for the two grades is .75 for 
the lower phosphorus, and .70 for the higher. The reason for making two 
classifications for open-hearth rails relates principally to the cost of manufacture. 
It was thought desirable to specify phosphorus as low as .03 so that high carbon 
could be used and the wearing quality of the rails, particularly on curves, would 
be materially improved. But the extra time required in the open-hearth furnace 
to reduce phosphorus from .04 to .03 results in some increase in the cost of 
manufacture, and a slight addition to the normal price per ton is added for the 
class A rails. 



,y Google 



STEEL RAILS 











HI 




niJi- li 


p 


1! 


1^1 


i 
1 




i 

1 




1 f''M I l^'k 
' j-iiii '-Hi 


II 








ii 




m 


Hi 








■"Hi 






i 


^■a 


■™ 1 


J_ 


H 


■"H 


s 


3 




ss 


s 


■"!W 1 


-U 


i 


-Hi 


g 








8 


•"!H 1 : 


s 


f.- 


:"« I _ _ 


^ 


I 




a 


S 










e 








° 




__ 








£j:3 






















SB 
























SS3 










































£ 


■o-W 














« 






















n 






2 


S 




S 


i 






















■"!H 




















s 




i 


1 
















■"!K 




s 












s 


s 






t« 






































,,^ 


* ""^ 




* 












9 S«=i 






sss 






1 


















■p."a 














S S9S 




s 


S33 








"" 




.^i. 


_o_S = = 








: 


<j 


s 


im 


liliS 




1 

i 


3SS 


1 




2 

S 


! ! 


= 


11 
















1 






1 

1 


^ 1 


u 

1 


1 

! 


1 
IE 



>, Google 



RAIL SPECIFICATIONS 467 

Sulphur is not generally mentioned in the chemical requirements, but the 
trend of modem specifications is to require that this element be reported. This 
is illustrated by the American R^way Engineerii^ Assodation specification 
for analyses pven below. 

Amdyses 

Am. Ry. Eng. Assn. : 

7. In order to ascertain whether the chemical compo^tion is in accord- 
ance with the requirements, analyses shall be furnished as follows: 

(o) For Bessemer process the manufacturer shall furnish to the inspector, 
daily, carbon detaminations for each heat before the rails are shipped, and two 
chemical analyses every twenty-four hours representing the average of the 
elements, carbon, manganese, silicon, phosphorus, and sulphur contained in the 
steel, one for each day and night turn respectively. These analyses shall be 
made on drillings taken from the ladle test ingot not less than one-eighth inch 
beneath the surface. 

(6) For Open-hearth process, the makers shall furnish the inspectors with 
a chemical analysis of the elements, carbon, manganese, silicon, phosphorus, and 
sulphur, for each heat. 

(c) On request of the inspector, the manufacturer shall furnish drillings 
from the test ingot for check analyses. 

The Pennsylvania spediications are die same as t^e above. The other speci- 
fications only require one complete chemical analysis every twenty-four hours ui 
the Bessemo* process. The manufacturers and American Society for Testing 
Materials require that the analyses shall be made an drillings taken not less than 
(me-fourth inch beneath the surface of the test ingot. The Harriman Lines speci- 
fication does not give the depth at \duch the drillings should be taken. 

PkysiccU Requirements 
Table C compares the physical requironents of the different spedfications. 

Am. Ry. Eng. Assn.: 

Pkyaedl QnaWiet. 

8. Testa dudl be made to determine: 

{a) Ductility or toughness as opposed to brittleness. 
(ft) Soundness. 

MeOiod of reeling. 

9. The phymcal qualitiee shall be determined by the Drop Test. 

Drop Tegtirtg Maehine. 

10. The drop-teeting machine used shall be the standard of the American Railway Engineering 
Assodation. 

. (a) The tup shall weigh 2000 pounds, and have a striking face with a radius of five inches. 

(b) The anvil block shall weigh 20,000 pounds, and be supported on springs. 

(c) The supports for the tcatpieces shall be spaced 3 feet between centers and shall be a part of, 
and finnly secured to the anvil. Tne bearing surfaces of the supports shall have a radius of S mches. 

Dionizodb, Google 



468 STEEL RAILS 

Pieeei for Drop Te»t. 

11. Drop testa sh^ be made on pieces of rail not less than 4 feet and not more than 6 feet kmg. 
These test pieces shall be cut from the top end of the top rail of the ingot, and marked od the base or 
head with gaee marks 1 inch apart for 3 inches each mde of the center of the test piece, for measimng the 
ductility of the metal. 

Temperattare of Tesf Pieces. 

12. The temperature of the test pieces shall be between 60 and 100 d^reea Fahrenheit. 

Height of Drop. 

13. The test piece shall, at the option of the inspector, be placed head or base upwards on the 
auppcarts, and be subjected to impact of the tup falling free from uxe following heights: 

For 70-pound rail 16 feet 

For 80-, 85-, and 90-pouiid raU 17 feet 

Fot 100-pound rail 18 feet 

Elongation or Ductility. 

14. Under these impacts the rail under one or more blows shall show at least 6 per cent elongation 
for 1 inch or 5 per cent each for two consecutive inches of the 6-inch scale, marited as described rn 
Section 11. 

Permanent Set. 

15. It is desired that the permanent set after one blow under the drop test shall not exceed that 
in the following table, and a record shall be made of this information. 



Rail. 






Ordinate in Inches in a Length ol 3 [t. 


SwtiOB. 


"7^- 


HomentDTIurtia. 


Beesamei Pnxwa. 


O. H. Proom. 


A.R.A.-A 


100 
100 
90 
90 
80 
80 
70 
70 


48.94 
41.30 
38.70 
32.30 
28.80 
25,00 
21.05 
18,60 


1.65 

2,05 
1.90 
2.20 
2,85 
3.15 
3.50 
3.85 


1 45 


A.R.A.-A 

A.R.A.-B 

A.R.A.-A 

A.R.A.-B 

A.R.A.-A 


1.65 
2.00 
2,45 
2.85 
3,10 







Teel to Destruction. 



r subsequent blows shall be nicked and 



B««emer Process Drop Tetta. 
17. One piece shall be tested from each heat of Bessemer steel. 

(a) If the test piece does not bre^ at the first blow and shows the required elongation (Section 14), 
all of the rails of the beat shall be accepted, provided that the test piece when nicked and broken does 
not diow interior defect. 

" ' IF does not show the requu^ elongation (Section 14), 

u, but when nicked and broken shows interior defect, 



piece selected by the inspector from the top end of 
ingot. If the test piece does not break a( *""" 



(d) If the piece breaks at the first blow, a 
or if the test piece shows the required elongatio: 
ail of the top rails from that heat shall be rejectt 

(c) A second test shall then be made of a tes 

any second rail of the same heat, preferably of the „--. ^ -. .. 

first blow, and shows the reauired elongation (Section 14), all of the remainder of the rails of the heat 
shall be accepted, provided that the test piece when nicked and broken does not show interior defect, 

{d) If the piece breaks at the first blow, or does not show the required elongation (Section 14), ot 
it the piece shows the required elongation, but when nicked and broken shows interior defect, all of the 
second rails from that heat ^all be rejected. , 

(e) A third test shall then be made of a test piece selected by the inspector from the top end of 
any third rail of the same heat, preferably of the same ingot. If the test piece does not break at the 
firet blow and shows the required ek>ngation (Section 14), all of the remainder of the rails of the heat 
shall be accepted, provided that the teat piece when nicked and broken does not show interior defect. 

if) If the piece breaks at the first, blow, or does not show the required elongation (Section 14), or 
if the piece shows the required elongation, but when nicked and broken shows interior drfect, all of we 
remainder of the rails from that h^it shall be rejected. 



>, Google 



RAIL SPECIFICATIONS 469 

Open-hearth Process Drop Teslt. 
18. Test pieces ghall be selected from the second, middle, and last full ingot of each open-bearth 
heat. 

(a) If two of these test pieces do not break at the lirat blow and show the required elongation 
(Section 14), all of the rails of the heat shall be accepted, provided that these test pieces when nidied 
and broken do not show interior defect. 

(b) If two of the test pieces break at the first blow, or do not show the required elongatitm, or if 
any of the pieces that have been tested under the drop when nicked and broken show interior defect, 
all of the top rails from that heat shall be rejected. 

(c) Second tests shall then be made from three test pieces selected by the inspector from the top 
end of any second rails of the same heat, preferably of the same in^ta. If two of these test pieces do 
not break at the liiat blow and show the required elongation (Section 14), all of the remainder of the 
rails (rf the heat shall be accepted, provided that the pieces that have been tested under the drop when 
nicked and broken do not show interior defect. 

(d) If two of these test pieces break at the first blow or do not show the required elongation 
(Section 14), or if any of the pieces that have been tested under the drop when nicked and broken show 
interior detect, all of the second rails of the heat shall be rejected. 

(e) Third tests shall then be made from three test pieces selected by the inspector from the top 
end of any third rails of the same heat, preferably of the same ingots. If two of these test pieces do not 
bieak at the first blow, and show the required elongation (Section 14), all of the remainder of the rails 
of the heat shall be accepted, provided that the pieces that have been tested under the drop when 
nicked and broken do not show interior defect. 

(f) If two of these test pieces break at the finrt blow or do not show the required elongation 
{Section 14), or if any of the pieces, that have been tested under the drop when nicked and broken 
show interior defect, all of the remainder of the rails from that heat shall be rejected. 

The drop-testing machine has been standardized, and it is claimed that the 
lower drop called for under the new conditions is equivalent to the higher drop 
of 22 feet previously specified. 

The provision in the American Railway Engineering Association and Penn- 
sylvania specifications that drop testing shall be continued to the destruction of 
the specimen is a precaution which should result in a material benefit to the railway 
by reducing the number of piped rails, which, in spite of the usual inspection and 
tests, get into the main track. It seems far more desirable not to specify any defi- 
nite discard, but to test to destruction a number of rail butts representing a certain 
proportlcHi of the total output, and to base rejections on the results of these 
tests. 

The Cambria Steel Company rolled a considerable tonnage of rmls under 
these specifications, and the testing to destruction unquestionably detected the 
pipes. To find to what depth the pipes extended, they poUshed the ends of the drop- 
test pieces and cut the top rail adjoining the test pieces into small lengths, examin- 
ing carefully each cut for pipes. It was found that of the heats showing pipes in 
the drop-test piece when tested to destruction, sixty per cent contained pipes so 
diort that they were confined raitirely to the crop end. Of the remaining forty 
per cent which extended into the top rml, nearly half showed a pipe extendmg less 
than four feet. 

Most of the spedfications require that two test pieces be tested in the open- 
hearth process but only one in the Be^emer process. This is because of the greater 
tonnj^ of metal in an open-hearth heat as compared to a Brasemer heat, there 
being about five times as much metal. All the basic open-hearth rml steel now 



,y Google 




Uig:izcci ^y vjiv*. 'Wi\. 





1 

1 

1 


n 

m 


J 


1 


il 


II 


m. 




Is 


'hi 




ti 






■111 


Jli: 




°JI 


H^ 


^ 






■ 


1 

•s 








1 

i 






is 


1 












*0 

1 






1 






1 






|1 






fi 






i 






1=^ 
1^ 









>, Google 



472 STEEL RAILS 

being made in the United States is melted in furnaces of a minimum capacity (rf 
40 tons, and the majority of it is made in 50-ton or 80-ton furnaces. 

The feeling is growing among railway eng^eers that present specifications do 
not go far enough in specifying tests of a cotain number of ingots from each heat, 
but that specimens should be tested from each individual ingot. On account of the 
uncertainty attending the formation of the pipe in the ii^t an increase in the 
number of tests per heat would appear desirable. 

The Pennsylvania is the only spedfication that provides for a Umiting deflec- 
tion, althou^ the American Railway Ei^:ineering Association give limits that it is 
dedred not to exceed and state that it is proposed to presmbe the requirements in 
r^ard to deflection as soon as proper limits have been decided on. In the absence 
of a tendon test it would se^n desirable to make provisions for fixing maximum 
and minimum limits for the deflection. 

The American Railway Engineering Association speeifieatdons are the only 
specifications which call for a ductility test. This test has beai used for some 
time by Dr. P. H. Dudley, as shown in the New Yoric Central lines qiecifications 
^ven in Article 35. 

No. 1 and No. 2 RaOs 
Am. Ry. Eng. Assn.: 

19. No. 1 classification rails shall be free from injurious defects and flaws 
of all kinds. 

20. (a) R^ls which, by reason of surface imperfections, or for causes men- 
tioned in Section 30 hereof, are not classed as No. 1 rails will be accepted as No. 2 
rails, but No. 2 rails which contain imperfections in such number or of such char- 
acter as will, in the judgment of the inspector, render them unfit for recognized 
No. 2 uses will not be accepted for shipment. 

(6) No. 2 rMls to the extent of 5 per cent of the whole order will be received. 
All rails accepted as No. 2 rails shall have the ends painted white and shall have 
two prick punch marks on the dde of the web near the heat number near the end 
of the rail, so placed as not to be covered by the splice bars. 

All of the specifications state that "No. 1 rails shall be free from injurious 
defects and flavra of all kinds." 

The Pennsylvania spedfications for No. 2 r^ls are the same as the above with 
the clause added that rails which exceed the prescribed limits of deflection in tiie 
drop test may be accepted as No. 2 rails. The manufacturers' specifications are 
somewhat different and are given below. 

Rails which, by reason of surface imperfections, are not classed as No. 1 r^ 



,y Google 



RAIL SPECIFICATIONS 473 

shall be conEodered No. 2 rails, but No. 2 rails shall not be acc^ted for shipment 
which have flaws in the head of more than ^ inch, or in the flange of more than 
i inch in depth, and these shall not, in the judgment of the inspector, be, in any in- 
dividual rail, so numerous or of such a character as to render it unfit for recognized 
No. 2 rail uses. Both ends of No. 2 rails shall be p^ted white. 

The Harriman lines spedfications are the same as the manufacturers' with 
the additional clauses tiiat No. 2 rails will be accepted up to 5 per cent of the 
whole order, and m conmion with the American Sodety for Testing Materials they 
will not accept rails as No. 2 from heats which failed under the drop test. 

Quality of Manufaetwe 
Am. Ry. Eng. Assn. : 

21. The entire process of manufacture shall be m accordance with the best 
current state of the art. 

This section is common to all specifications except the manufactures'. 

Bledlng<as 
Am. Ry. Eng. Assn.: 

22. Bled ingots shall not be used. 

This is specified by aH except the manufacturers', with an addifional clause 
providing that the ingots be kept in a vertical position until ready to be rolled, 
or until the metal in the interior has had time to solidify. 

Discard 
Am. Ry. Eng. Assn.: 

23. There shall be sheared from end of the bloom formed from the top of 
the ingot, sufficient metal to secure sound rails. 

The Pennsylvania requirements are the same as the American Railway 
En^eering Association, while the Manufacturers' specifications do not refer 
to this part of the process. 

The American Society for Testing Materials and the Harriman linK sped- 
fications call for a definite discard as follows: 

A.S.forT.M.: 

There shall be sheared from the end of the blooms formed from the top of 
the ingots not less than z per cent,* and if, from any cause, the steel does not 
then appear to be sofid, the shearing shall continue until it dora. 



* The percentage of luinimum discard in any case to be subject to agreement, and it should be 
recogniied that the hi^er this percentage the greater will be the coat. 

DionizodbyGOOgle 



474 



STEEL RAILS 



HanimaQ lines: 

1. (d) There shall be sheared from the end of the bloom and rail formed 
from the top of the ingot a total discard of not less than nine (9) per cent of the 
weight of the ii^ot, and if, from any cause, the steel does not then appear to 
be solid, the shearing shall continue until it does. If, by the use of any improve- 
ments in the process of making ingots, the defects known as " piping " shall be 
prevented, the above shearing requirements may be modified. 

On this system, by excluding the top, or A, rdls from main-line use, the 
effect of an additional 30 per cent discard is obtained, and at the same time 
the discard is saved for use on sidings and other locations where second-hand 
rails usually are used; the road thus doing its own discarding beyond the 
manufacturers' allowance, and saving the product without risk to the quality 
of the main-track rails. 

In view of the uncertjdnties, as to the length of the pipe it appears that 
the position taken by the American Railway Engineering Association and the 
Pennsylvania System in their spedfications is the most reasonable one, viz., to 
leave the discard to the manufacturers, and to safeguard the product by proper 
tests, especially by chooang the test piece from such a location, and making the 
rejections such, that it will be to the interest of the manufactiu'er to volun- 
tarily discard the metal which will not stand test. 

LengA 

Am. Ry. Eng. Assn.; 

24. The standard l^igth of rails shall be 33 feet, at a temperature of 60 de- 
grees Fahrenheit. Ten per cent of the entire ordra- will be accepted in shorto" 
lengths varying by 1 foot from 32 feet to 25 feet. A variation of i inch from the 
specified lengths will be allowed. No. 1 rails less than 33 feet long shall be painted 
green on both ends. 

TABLE CL — LENGTH OF RAILS IN STANDARD AMERICAN SPECIFICATIONS 



SpeciEiaitioiUL 


'^^ 


^-,XbilX^" 




33* 

30 or 33 

30 or 33 

33 

33* 


32, 31, 30, 29 
28, 27, 26. 25 
J Varying by e^'en 
J feet to twenty- 
) four (24) feet. 
30, 27i, 25 













All of the specifications agree m allowing ten per cent (rf the entire orda- to 
be shorter lengths than the standard and permit of a variation of i uich in length 
from that specified. All call for the short-length rails to be painted green aa both 



,y Google 



__^ I 



RAIL SPECIFICATIONS 



aids. The standard length and short lengths that will be accepted, however, 
vary, and tiie requirements given in the various specifications are shovm m Table CI. 

Shrinkage or Control of Finishing Temperature 
Am. Ry. Eng. Assn.: 

25. The number of passes and speed of tram shall be so regulated that, on 
leaving the rolls at the final pass, the temperature of the rml will not exceed 
that which requires a shrinkage allowance at the hot saws, for a rail 33 feet in 
length and of 100 pounds' section, of 6f inches, and ^ inch less for each 10 pounds' 
a in section. 





























MANU 


=ACTUR 
AS.FO 


ERS 1 
RT.M/ 


V^ 


-^ 






^^ 


^ 








■^ 


AAE.t 


M.wX/ 


x^ 


^ 






^^ 


^ 


AS.C.E. 








'^ 





















o*^ 



«i 

lu 

1 

z 



70 SO 90 lOO 

WEIGHT OF RAIL. LBS. PER- YD. 

Fia. 340. — Shrinkage Allowed in American Specificatione in 1909. 

26. The bars ^lall not be held for the purpose of reducing their tempera- 
ture, nor shall any artificial means of cooling them be used after they leave 
the finishing pass. Rails, while on the cooling beds, shall be protected from 
snow and water. 

The other specifications are the same, except that in the Manufacturers' 
American Society for Testing Materials and Harriman Lines, the statemwit that 
rails, while on the cooling beds, shall be protected from aiow and water, is omitted. 

A greater shrinkage Is allowed in present specifications than was formerly the 
case. fig. 340 shows the allowance of different specifications in 1909; it will be 



,y Google 



476 STEEL RAILS 

noted that the shrinkage has been increased to agree with the higher figure ^ven by 
the mamifacturers. In the Pennsylvania requirements of that date, it was pro- 
vided tiiat the shrink^:e allowance be decreased at the rate of t Ju inch for each 
second of time elapsed between the rail leaving the finishuig rolls and being sawed. 
The A. S. C. E. specifications called for sV i"*^ in place of j}^ inch. 

The control of the finishing temperature by the amount of contraction which 
the rail undergoes in cooling from the finishing to the atmospheric temperature 
appears to be the only method practical to use. Other efforts have been made 
to determine the finishing temperature by the use of pyrometers and by the ex- 
amination of the microstructure of the rails. The use of pyrometers naturally 
suggested itself at iirst as the most promising means of accomplishing that pur- 
pose, but it was soon found that no pjTometric device existed which could be 
applied in a practical way to the detection of the temperature of quickly moving 
r^ls. The micro-test, although attractive and useful, can only be applied to a 
very small percentage of the rails manufactured, and this is its greatest 



Section 
Am. Ry. Eng. Assn.: 

27. The section of rails shall conform as accurately as possible to the tem- 
plet furnished by the railroad company. A variation in height of ^ inch less 
or sV inch greater than the specified height Mid jV inch in width of flange will 
be permitted; but no variation shall be allowed in the dimensions affecting 
the fit of the splice bars. 

The other specifications are substantially the same as those given above. The 
Maniifacturers' specify A. S. C. E. sections; the Harriman Lines, A. R. A. section 
"A" 90-pound, Railroad Company's "Common Standard" 75-pound, A. S. C. E. 
section, 65-pound; and the American Society for Testing Matraials say, unless 
otherwise specified, A. S. C. E. sections. 

Weighi 
Am. Ry. Eng. Assn.: 

28. The weight of the rails specified in the order shall be mmnt^ned as 
nearly as possible, after complying with the preceding section. A variation of 
one-half of one per cent from the calculated weight of section, as applied to an 
entire order, will be allowed. 

29. R^ls accepted will be paid for according to actual weights. 
The other specifications are substantially the same as the above. 



,y Google 



RAIL SPECIFICATIONS 477 

Straightening 
Am. Ry. Eng. Assn.: 

30. The hot straightening shall be carefully done, so that gagging under 
the cold presses will be reduced to a minimum. Any rail coming to the straight- 
ening presses showing sharp kinks or greater camber than that indicated by a 
middle ordinate of 4 inches in 33 feet, for A. R. A. type of sections, or 5 inches 
for A. S. C. E. type of sections, will be at once classed as a No. 2 rail. The 
distance between the supports of rails in the straightening presses shall not be 
less than 42 inches. The supports shall have flat surfaces and be out of wind. 

All of the specifications are substantially the same. 

DriUing 
Am. Ry. Eng. Assn. : 

31. Circular holes for joint bolts shall be drilled to conform accurately in 
every respect to the drawing and dimensions furnished by the Railroad Company. 

Substantially the same for all specifications. 

Finishing 

32. (a) All rails shall be smooth on the heads, str^ght in line and surface, 
and without any twists, waves or kinks. They shall be sawed square at the 
ends, a variation of not more than one-thirty-second inch being allowed; and 
burrs shall be carefully removed. 

(b) Rails improperly drilled or straightened, or from which the burrs have 
not been removed, shall be rejected, but may be accepted after being properly 
finished. 

Substantially the same for all specifications. 
Branding 

Am. Ry. Eng. Assn.: 

33. The name of the manufacturer, the weight and type of rail, and the 
month and year of manufacture shall be rolled in raised letters and figures on 
the side of the web. The number of the heat and a letter indicating the portion 
of the ingot from which the rail was made shall be plmnly stamped on the web 
of each rml, where it will not be covered by the spKce bars. The top rails shall 
be lettered "A," and the succeeding ones "B," "C," "D," etc., consecutively; 
but in case of a top discard of twenty or more per cent, the letter "A" will be 
omitted. Open-hearth rails shall be branded or stamped "0. H." All mark- 
ings of rails shall be done so effectively that the marks may be read as long as 
the rmls are in service. 

The Pemisylvania spedficalions are the same as the above. 



,y Google 



478 



STEEL RAILS 



All of the other spedfications omit the clause in the American Railway Engi- 
neering Association specifications in refCTence to omitting the letter "A" in case of 
a top discard of twenty per cent or more, but are substantially the same in otha* 
respects. The Harriman lines specify that all "A" rails i^iall have the top of the 
flange at each end painted yellow, and the American Society for Testing Materials 
only require rails weighing 70 pounds per yard or over to be stamped with a letter 
to indicate the portion of the ingot from which the ral was rolled. 

Separate Classes 
Am. Ry. Eng. Assn.: 

34. All classes of rails shall be kept separate from each other. 

Loading 
Am. Ry. Eng. Assn.: 

35. All rails shall be loaded in the presence of the inspector. 

The Pennsylvania specifications are the same for both of the above clauses. 
The Harriman lines specifications contain the following clause: 

The following classes of rail shall be loaded separately as far as practicable, 
excepting at the finishing of an order or the end of a rolling. In this case the differ- 
ent classes shall be kept separate by placmg strips of wood between each class, and 
each shipping notice shall contain full mfcHmation as to the contents of each car: 

No. 1 rails, B, C, D, etc, full lengths. 

No. 1 rails, B, C, D, etc., short lei^ths. 

No. 1 "A " rails; that is, r^ from the top of the ingot, full length. 

No. 1 "A" r^ls, short length. 

No. 2 r^, all lengths. 

35. SPEcinCATioNS (New York Central Lines) for Basic Open-hearth 
Rails 
i6t — Chemical Composition: 





» Lb. 


» Lb. 


lOD Lb. 




.65 to .68 

.70 to 1.00 

.10 to .20 

.04 


.60 to .73 

,70 to 1.00 

.10 to .20 

.04 


.62 to .75 

.70 to 1.00 

.10 to .20 

.04 











To adjust the chemical composition to the special conditions of manu- 
fecture at each mill, the engineer representing the railroad company, from the 
inspection of the ingots, their heating, blooming, and rolling into r^ls, shall 
have the right to select the lower or average limit of either the sificon or man- 



,y Google 



RAIL SPECIFICATIONS 479 

ganese, or both, with the average carbon content as the working haras for mak- 
ing the steel, as he may find requisite for good setting ingots with freedom from 
pipes and rolling into tough steel by the plant of the manufacturer. 

2nd — Process of Manufacture : The entire process of manufacture and 
testing shall be in accordance with the best current state of the art, and special 
care shall be taken to conform to the following instructions: 

(a) Excessive use of material thrown into the teeming ladle to set the 
stopper must be avoided. 

(&) The steel must be well deondized and the waste products eliminated 
before the ingots are teemed. 

(c) The steel must be made to set quiet by the chemical composition in the 

molds without the addition of aluminum, either in the ladle or molds. 

(d) Spattering the ulterior sides of the molds in pricking the melts and 

teeming the ingots to be avoided as much as possible. 

(c) "Hme must be allowed for the tops of the ingots to set without spray- 
ing with water. 

(J) The ingots should be stripped as soon as the metal caps over on top, 
then sent to the scales to be weighed, th^i sent to the reheating 
furnaces to be charged promptly, to avoid the cooKng of the interior 
metal and thus check the large shrinkage which occms in it from 
unnecessary loss of temperature due to delays. It has been found, 
in good practice, possible in this way to confine the interior shrinkage 
to 0.05 to 0.1 of one per cent per cubic foot of the metal of rail 
ingots. The total shrinkage of an ingot depends upon its volume, 
chemical composition and loss of temperature at the time it is charged, 
yet in fair practice it may be confined to such small limits that it is 
removed in the usual discard of the bloom. Piped rails come from 
cold ingots or those which have been unduly delayed before charging 
into tiie reheating furnace. 

(g) Cast Iron Cut Out qf the Ingot Stools: Care to be taken m teeming the 
ingots to prevent cutting out of the cast iron of the stools or ingot 
molds by the falling stream of hot metal from the ladle, avoiding a 
frequent cause of carbon streaks found in the segregated steel of 
"split heads." The most disturbing factor of the small amount of 
ordinary segregation in rail steel is the diffused cast iron in some 
ingots cut out from the stools. 

(A) Ingots shall be kept in a vertical position on the ingot cars and in the 
reheating furnaces until their heat is equalized ready to be rolled. 



,y Google 



480 STEEL RAILS 

(t) Bled ingots shall not be used. (" Bled ingot " — one from the centa- 

of which the liquid steei has been permitted to escape.) 
(j") There shall be sheared from the end of the bloom formed from the top 
of the ingot sufficient discard to secure sound rails. (All metal from 
the top of the ingot, whether cut from the bloom or the r^l, is the 
"top discard.") 
(k) One-hundred-pound (100 lb.) rails not to be rolled from blooms exceed- 
ing three (3) thirty-three-foot (33') lengths in a continuous bar; 
eighty-pound (80 lb.), or lighter, rails in not over four (4) lengths 
of thirty-three feet (33') in a continuous bar, when inserted in the 
contract. 
3rd — Shrinkage : The number of passes and speed of train diall be so reg- 
ulated that, on leaving the rolls on the final pass, the temperature of the rails 
will not exceed that which requires a shrinkage allowance at the hot saws for 
a 33-foot rail of 100 pounds section of 6f inches, and ^^ inch less for each five 
pounds decrease of section. No artificial means of cooling the steel shall be 
used between the leading and finishing passes, nor after the rails leave the 
finishing rolls; neither shall rails be held before sawing for the purpose of 
reducing their temperature. 

4th — Drop and Ductility Tests : A drop test to be made of a crop from 
the top bar of the second, the middle and the last full ingot of the melt. The 
crop 4 to 6 feet long to be stamped with a spacing bar of dx one-inch spaces on 
the base, head or side as desired. 

Each butt must show under a angle blow of the drop, of 18-foot, for the 
80-pound or 90-pound section, and 20-foot for the 100-pound section, at least 
MX per cent elongation for one inch or five per cent each for two consecutive 
inches before fracture for acceptance of the melt. 

The crop or butt is Uable to be chilled accidentally in entering the rolls 
several times, or it may be caused by other delays, and should it break under 
a single blow without showing the percentage of elongation specified, it shall 
be considered as indicating deficient ductility or chilled metal, and the results 
must be rejected. 

The Inspecting Engineer representing the Rmlroad Company must then 
take a duplicate test from the same ingot at the top end of the " A " or " B " 
rail, according to the nine or greater percentage of discard, and the results 
taken in lieu of those from the first crop or test to determine whether or not 
the piece had the requisite ductility in accordance with the specifications. 
The distinction between a chilled test crop and those of inadequate ductility 



,y Google 

^ 1 



RAIL SPECinCATIONS 481 

must be ascertained according to above preBcribed tests before rejections are 
made or rails accepted. 

Should any test piece under the first blow of the drop not break, but fail 
to show the percentage of elongation specified, the test piece shall be subjected 
in the same position to a second blow and the results so obtained govern in 
passing the test. 

The ductility of at least one specimen of each melt to be exhausted by one 
or more blows of the drop, and a record made of the respective elongations of 
each test. 

The drop-testing machine shall have a tup of 2000 pounds weight, the 
striking face of which shall have a radius of not more than five inches (5"), 
and solid supports, centers three feet (3') apart, for the test butts. The anvil 
block shall weigh at least 20,000 pounds and the supports shall be part of or 
firmly secured to the anvil. The report of drop test shall state the atmospheric 
temperature at the time the test was made. The testing shall proceed con- 
currently with the operation of the mill. The temperature of the test butts 
to be between 40 degrees and 100 degrees Fahr. 

5^1 — Section : The section of rail shall conform to the dunensions fur- 
nished by the purchaser as accurately as possible consistent with the paragraph 
relative to specified weight. 

A variation in height of rails of sV of an inch over or g^ of an inch under, 
also tV of an inch in width of flange will be permitted, but no variation will 
be allowed in dimensions affecting the fit of the splice bars. 

6th — Weight: The weight of the rail shall be maintained as nearly as 
possible, after complying with the preceding paragraph, to that specified in the 
contract. 

A variation of one-half of one per cent, from the calculated weight of sec- 
tion, on the entire order, will be allowed. 

Rails will be accepted and paid for according to actual weight. 

7th — Length: The standard length of r^ls shall be thirty-three feet (330. 
Ten per cent of the entire order will be accepted in shorter lengths varying as 
follows: Thirty feet (30'), twenty-eight feet (28'), twenty-six feet (26') and 
twenty-four feet (24'). A variation of i of an inch from the specified length 
will be allowed. 

Three rmls in every 100 tons to be thirty-two feet and ax inches (32' 6") 
long, the ends painted red, when inserted in the New York Central Contract. 

All other No. 1 rails less than thirty-three feet (33') long shall be painted 
green on both ends. 



,y Google 



482 STEEL RAILS 

8th — Branding: The name of the maker, the weight of the raO and the 
month and year of manufacture, together with "0-H," shall be rolled in raised 
letters on the side of the weh, and the number of the melt and letter to deag- 
nate the position of the rail in the ingot shall be so stamped on each r^ as not 
to be covered by the splice bars. 

When the rails are to be rolled with twenty per cent (20%) discard the 
first rail in the ingot shall commence with the letter "B," the second "C," 
the third "D" and the fourth "E." 

When the "A" rmls are to be taken they are to be loaded separately upon 
cars for shipment and the flanges at the ends painted yellow, when inserted 
in the contract. 

pth — Drilling: Circular holes for splice bars shall be drilled in accordance 
with specifications of purchaser. They shall in every respect accurately con- 
form to drawing and dimensions furnished and shall be free from burrs. 

loth — Straightening: Care must be taken in cambering the rails and 
with the hot-bed work, which must result in the rails being left in such con- 
dition that they shall not vary throughout their entire length more than four 
inches (4") for the "A. R. A." thick bases and not more than five inches (5") 
for the "DUDLEY" section or "A. S. C. E." sections from a straight line in 
any direction when delivered to the cold-straightening presses. Those which 
vary beyond that amount, or have short kinks, shall be classed as second qual- 
ity rails and be so marked. Rails while on the "hot-beds" shall be protected 
from coming in contact with water or snow. The distance between supports 
of rails in the gagging press shall not be less than forty-two inches (42"); 
supports to have flat surfaces. 

Rails shall be straight in line and surface and smooth on head when fin- 
ished — final straightening being done while cold. They shall be sawed square 
at ends, variations to be not more than jV of an inch, and prior to shipment 
shall have the burr caused by the saw cutting removed and the ends made 
clean. 

1 1 til — Inspection: The inspector representing the purchaser shall have 
free entry to the works of the manufacturer at all times while his contract is 
being executed, and shall have all reasonable facilities afforded him by the 
manufacturer to satisfy him that the rails are being made in accordance with 
the terms of the contract. All tests and inspection shall be made at the place 
of manufacture priw to shipment, and shall be so conducted as not to imneces- 
sarily interfere with the operation of the mill. 

The manufacturer shall furnish the inspector with a chemical analysis of 



,y Google 



RAIL SPECIFICATIONS 483 

each melt of steel covering the elements spedfied m the section No. 1 hereof, 
and also report sulphur and copper. 

Analysis shall be made on drillings taken from small test ingots, the drill- 
ing being taken at a distance of not less than J of an inch beneath the surface 
of said test ingots. On request of the inspector the manufacturer shall furnish 
drillings for check analysis. 

rath — No. 2 Rails : Rails which by reason of surface imperfections are 
not classed as No. 1 rails shall be considered No. 2 rails, but No. 2 rails shall 
not be accepted for shipment which have flaws in the head of more than J of 
an inch; or in the flange of more than j of an inch in depth; and these shall 
not, in the judgment of the inspector, be, in any individual rail, so numerous 
or of such a character as to render it unfit for recognized No. 2 rail uses. 

13th — Designation of No. 2 Rails and Short Lengths of No. i Rails: 
Both ends of all No. 2 rails shall be painted white. 

Both ends of all short lengths No. 1 r^s shall be painted green, except 
Uie 32-foot and 6-inch rails, which are to be painted red. 

(Sgd.) P. H. DUDLEY, 
New York Central Lines. 
(Specifications of Oct. 1st, 1909. 
Revised Jan. 11th, 1911, to con- 
form to Manufacturers' sale per 
100 pounds.) 
Note 1. "Process of Manufacture" (b): The elimination of the deoxidation 
products and impiuities from the bath of metal is more important than has 
yet been appreciated. This prevents minute portions of the deoxidation prod- 
ucts from becoming entrained in the setting metal and therefore will avoid 
their being rolled in the steel, where in the rail head or base they would be 
subjected to alternate unit fiber strmns under moving trains and contribute 
the needed factor to develop the interior transverse checks recently observed 
in a few rail heads. 

Time is required for the deoxidation products and impurities to rise after 
the steel is tapped into the ladle. 

These heterogeneous portions of the deoxidation products or impurities 
in the steel, as well as small flaws and interior cavities, are theoretically and 
practically known to be zones of weakness, and interrupt the normal unit strains 
and increase them in the surrounding metal, which often result in detailed 
fractures. 

Note 2. "Process of Manufacture" (/): The percentage of interior shrink- 



.yGoogle 



4S4 



STEEL RAILS 



age per cubic foot of the metal of the ingots there mentioned was reduced the 
past year by good mill practice and well organized train service. The lattff 
was to transport promptly the ingots after they were teemed and stripped 
so that they could be charged with the least possible delay into the reheating 
pits and then as soon as the heat of the metal was properly equalized, they 
were bloomed, which restricted the reduced cavity to the discard. 
Note 1 and 2 added for information. 
P.H.D. 
1/3/12 

86. Brttkh Standasd Specifications of Bull Head Railway Rails 

(Iteport No. 9, Revised July, 1909.) 
laaued by The Engineering Standards Committee 

Supported by: The Institution of Civil Engineers; The Institution of Mecbanical EoEiiieera; Tb« 
Institution of Naval Architects; The Iron and Steel Institute; The Institution of ElectrictuEngineen. 
(Reprinted by permisBion of the Committee). 

1. The steel for the Rails shall be of the best quality made by the Bessemer, Siemens-Martin, 
or other process, aa may be approved by the Engineer (or by the Purchaser). 

The Rails shall show on analysis that in chemical composition they conform to the following 
limits: 

Carbon from 0.35 to 0.5 per cent. 

Manganese " .7 to I.O " " 

Silicon not to exceed 0.1 " " 

Phosphorus " " 0.075 " " 

Sulphur " " 0.08 " " 

2. The Manufacturer Ehall make and furnish to the representative of the Engineer (or of 
the Purchaser) carbon determinations of each cast. 

A complete chemical analysis, representing the average of the other elements contained 
in the steel, shall be similarly given for each rolling. Such complete analysis shall be made fronj 
drillings taken from the rail or tensile test piece or pieces. When the roiling exceeds 200 tons, 
an additional complete analysis shal be made for each 200 tons or part thereof. 

Should the Engineer (or the Purchaser) desire to make independent chemical determinations, 
the necessary specimens and samples shall bo furnished by the Manufacturer. For this purpoee 
not more than two rails in every hundred tons manufactured shall be selected by representative 
of the Engineer (or of the Purchaser) and drillings taken with a drill of 2 inches diameter from 
the top of the head of the rail, unless otherwise specified by him, and if, upon being subjected to 
the specified tests, cither fail to comply therewith, then all the Rails in the cast of which the let 
pieces form a part may be rejected. 

The representative of the Engineer (or of the Purchaser) may then take similar samples 
from a further two rails out of the same 100 tons, and should either fail to conaply with Ibe 
specified analy^a the whole 100 tons may be rejected. 

In case of difference between the Engineer (or between the Purchaser) and the Manufacturer 
as to the accuracy of any analysis, cither party shall have the right to have samples of the Bt«l 
analyzed by an independent metallurpst, to be mutually agreed upon. The expenses attendant 
upon such independent analysis shall be borne by the party adjudged to be in the wroi^. 

3. Each Rail shall be made from an ingot not less than 12 inches square at the smaller and 
14 inches square at the larger end, and must be cogged down Into blooms, and sufficient crop 
then sheared from each end to ensure soundness. 



., Google 



RAIL SPECIFICATIONS 



All stn^htening shall be done by pressure tmd not by hammering. Pwmissible 

4, A rolling margin of J per cent under to J per cent above the calculated wdght will be Variation 
pennitted, but the calculated weight only will be paid for. "* weight 



TABLE OF GENEBAL DIMENSIONS AND WEIGHTS OF "B. S." RAILS 

(See PIftte XIV) 



Numb«of-B.8." 








S«tk>n and Nomina 


HeiihtotlUil. 


Width otHsHi. 


C.I™l;.gdW«<b. 


Pounda. 


lHh« 


laebm. 


Pounds iwr Yard. 


60 


4! 


U" 


59.79 


65 


4{ 


64.58 


70 


5 


2A 


70.13 


75 


5 




74,56 


80 


5 




21*3 


79.49 


85 


5 




2| 


S4.8S 


90 


5 




2 


89-77 


95 


5 




2 


94.59 


100 


5 




2 


99.84 



General 
DinensiooB 
of Kails. 



6. Before the general manufacture of the Rails is commenced the Manufacturer shall, if 
required by the Engineer (or by the Purchaser), supply two sets of templates, internal and external, 
of approved material, for each " B. S." Section of Rail. 

Each template shall be suitably engraved with the Purchaser's name, the number of the 
" B. S." section (being the nominal weight of the Rail in pounds per lineal yard), the Manufac- 
turers' name and address, and the date of the Contract. 

These templates shall be submitted to the Engineer (or to the Purchaser) for his approval, 
and at the commencement of rolling the Engineer will have a competent person present to approve 
of the section. 

7. Each Section of Rail under this Contract shall be accurately rolled to its respective 
template. 

8. The whole of the Rails shall be of uniform section throughout, true to templates, per- 
fectly sound and straight, and free from splits, cracks, burrs and defects of every kind. 

9. A quantity of abort lengths will be taken in such lengths and quantities as may be 
ordered by the Engineer (or by the Purchaser), provided that these short lengths are cut down 
from longer lengths found to be defective at the ends only, and that the total quantity taken 
does not exceed 7J per cent of the contract. 

N. B. — The Committee recommend the adoption of the following, as the normal lengths 
of Rails, viz.: — 30 feet, 36 feet, 45 feet, or 60 feet. 

10. The Rails shall be the specified length at the temperature of 60° Fahr. No Rail 
will be accepted which is more than three-sixteenths of an inch (t^ inch) above or below the 
length specified, whether for curved or straight line. 

11. When required by the Engineer (or by the Purchaser) rails are to be supplied from 1 to 
6 inches shorter or longer than the normal specified lengths, and these special lengths are to have 
about one foot at each end painted with such colors as may be ordered. 

12. Rails shall be supplied for sn-itches and crossings when so ordered, and such Rails shall 
be of the required lengths and shall be cut from sound long Rails. 

13. The Brand shall be rolledonthewebof each Rail to show that the Rail is of British Standard 
Section and made under the conditions of this Specification,' the number of the " B. S." Section 
(being the nominal weight of the Rail in pounds per yard), the process by which the Ralls have 
been manufactured, the Manufacturer's name, initials, or other recognized mark, and the month 
and year of manufacture shall be rolled, in letters three-quarters of an inch (I inch) in size, on one 



Length of Rails 
for Straight 



Pennissible 
Variation 
in Length. 

Rails of Special 
Length for 
Matching in 
Curved Line. 
Rails for 
Switches and 
Crossings. 
Branding. 



>, Google 



486 



STEEL RAILS 



aide of the web of each Rail, e.g., w B.S. 95, B.A.* 4.04; and tbe aumber of the cast or 

blow from which it has been rolled efaall be stamped on tbe end of each Rail in half-inch (1 inch) 
block figures. * 

14. From each caat one rail shall be selected by the representative of the Engineer (or of 
the Purchaser). Prom this a piece 5 feet long shall be cut which shall be placed in a horizontal 
position with the bullhead uppermost upon two iron or steel supports resting on a solid founda- 
tion and placed so that their centers are 3 feet 6 inches apart, the upper surfaces of the supports 
being curved to a radius of 3 inches. The test shall comprise two blows delivered midway 
between the bearings from a falling iron weight of 2240 pounds, the striking face of which shall be 
roimded to a radius of not more than 5 inches. The heights of tbe drop for tbe various sections 
of Rails shall be as tabulated below. The blows must be sustained without fracture, and the 
Rail must show a deflection between the limits given below. 

FALLING WEIGHT TEST 







Pint Bkiv 




SecoiHlBlow 


NnmbM ol ■■ B. g.'' 
Sectioo and Nom- 

inal Weight ol Rails 
per <% iD lb.. 


Drop. 


Deaeetio.. 


Drop. 


DeflocdoD. 




From. To. 




From. To. 


Pooad.. 


F»m. 


lB0b«a. 


F«M. 


Incbm. 


60 


5 


I lA 


10 


3 3 


65 


5 


1 H 


12 


3 3 


TO 


6 


1 lA 


12 


3 3 


75 


6 


1 lA 


12 


3 3 


80 


6 


15 




85 


6 




15 




90 


7 




20 




95 


7 


lA 


20 




100 


7 


lA 


20 


3 il 



Should the length cut from the selected Rail fail to comply with the test specified for its 
weight, two other Rails from the same cast will be selected and similar lengths cut and tested, and 
the acceptance or rejection of the cast will be decided by the result of the three tests, so that if 
two of the Rails selected fail to comply with the test, the entire caat will be rejected. 

15. From each 100 tons of Rails the Manufacturer shall (if required by the representative 
of the Engineer or of the Purchaser) cut a test piece from any Rail selected as a sample Rail ; such 
test piece to be stamped to correspond with the sample Rail. It shall then be placed in a testing 
machine of approved pattern, and shall have an ultimate tensile strength equivalent to not less 
than 40 tons per square inch, nor more than 48 tons per square inch, with an elongation of not 
less than 15 per cent upon the Standard Test Reces C or D (see Fig. 341). Should the test piece 
fail to fulfil these conditions, the representative of the Engineer (or of the Purchaser) may require 
the Manufacturer to test two other Roils from the same cast in the same manner, and the blC- 
ceptance or rejection of the east shall be decided by the results of the three tests so that if two of 
the three Rails selected fail to comply with the test the entire cast will be rejected. 

The representative of the En^oeer (or of the Purchaser) may then take similar test pieces 
fn»n a further two rails out of the same 100 tons, and should either f«l to comply with the test 
the whole 100 tons may be rejected. 

Should the Engineer (or the Purchaser) desire to have independent tests made, the Manu- 
facturer shall provide the necessary test pieces, viz., two for every 200 tons, properly shaped and 
prepared as described in Fig. 341. 



• The following abbreviations are recommended : 
S.A. SieiDens-Martin Add. 
S.B. Kemeos-Martin Ba«c. 



B.A. Bessemer Add. 
B.B. Bessemer Basic. 



>, Google 



RAIL SPECIFICATIONS 



487 



16, TTie boles for fishbolts must be drilled through the web from the solid at each end of Hdes in Rails. 
the Rails, of the sizes and in the position shown in the British Standard specification for Fish 
plates for Bull Head Rails (Report Xo. 47) or on a drawing to be supplied by the Engineer (or 
the Purchaser). These holes must be clean and square with the web, without burrs on either 
side, and will be checked with the gai^es to be furnished to the Manufacturer by the Engineer 
(or by the Purchaser). Should any of the holes vary from the correct size or position more than 
one thirty-second of an inch (A inch) the Rails in question will be liable to rejection. 



— a'awOE LENOTH— 



AMA-MS^m. 



TEST PIECE C. 




DIA^.TMIN. 
MCA-MaO-IH. 



TEST PIECE D. 

Fio. 341 . — Test Pieces C and D, British Standard Specifications of Rails. 

The gauge length and the parallel portion are to be as shown, the form of the ends to be as 
required in order to suit the various methods employed (or gripping the test piece. 



17. The Manufacturer shall give to the Engmeer (or to the Purchaser), or his representative, 
at least seven clear days' previous notice, in writing, b^ore the rollii^ of the first lot of Rails, 
and at least three clear days' previous notice, in writing, Ijefore the rolling of any subsequent lot 
of Rails, is commenced, in order that arrangements may be made for the presence of the repre- 
sentative of the Engineer (or of the Purchaser) at the rolling. 

18. The Engineer (or the Purchaser) or his representative shall have access to the works 
of the Manufacturer at all reasonable times. He shall be at liberty to examine the Rails during 
any stage of their manufacture, and to reject any material or finished Rail which does not conform 
to the terms of this specification. 

Before the Rails are put before the representative of the Engineer (or of the Purchaser) 
for inspection the Manufacturer shall have them examined, and all Rail which he admits to be 
defective are to be sorted out and placed in a separate stack; the representative of the Engineer 
(or of the Purchaser) being empowered to refuse to inspect any lot of Rails not put in uniform 
lengths and sorted. 

19. The Manufacturer shall supply the material required for testing free of chai^ and shall, 
at his own cost, furnish and prepare the necessary test pieces, and supply labor and appliances 
for such testing as may be carried out at his premises in accordance with this specification. FmI- 
ing facilities at his own works for making the prescribed tests the Manufacturer shall bear the 
cost of carrying out the tests elsewhere. 

20. All Rails accepted by the representative of the Engineer (or of the Purchaser) shall be 
" Q his presence. 



Notice of 
Rolling to be 
Given. 



>, Google 



STEEL RAILS 



Pennissible 
b^ariatioii in 
Weight 



37. British Standard Specifications of Flat Bottom Railway Raii^ 

(Report No. 11, Revised July, 1909) 

Issued by The Engineering Standanla Committee 

Supported by: The Institution of Civil Engineers; The Institution of Mech&nical Engjaeers; The 
Institution of Naval Architects; The Iron and Steel Institute; The Institution of Electrical Engineers. 
(Reprinted by permissioa of the Committee.) 

1. The steel for the Rails shall be of the best quality made by the Bessemer, Siemens- 
Martin, or other process, as may be approved by the Engineer (or hy the Purchaaer). 

The Rails shall show on analysis that in chemical composition they cooform to the following 
limits: 

Carbon from 0.35 to 0.50 per cent. 

Manganese " 0.70 to 1.00 " " 

Silicon not to exceed 0.10 " " 

Phosphorus... .. " " " 0.07 " " 

Sulphur " " " 0.07 " " 

2. The Manufacturer shall make and furnish to the representative of the Engineer (or of 
the Purchaser) carbon and phosphorus determinations of each east. 

A complete chemical analj^is, representing the average of the other elements contained in 
the steel, shall be similarly given for each rolling. Such complete analysis shall be made from 
drillings taken from the Rail or from the tensile test piece or pieces. When the rolling exceeds 
200 tons, an additional complete analysis shall be made for each 200 tons or part thereof. 

Should the Engineer (or the Purchaser) desire to make independent chemical determinations, 
the necessary specimens and samples shall be furnished by the Manufacturer. For this purpose 
not more than two Rails in every 100 tons manufactured shall be selected by the representative 
of the Engineer (or of the Purchaser) and drillings taken with a drill of 2 inches diameter from the 
top of the head of the Rail unless otherwise specified by him, and if, upon being subjected to the 
specified tests, either fail to comply therewith, then all the Rails in the cast of which the test 
pieces form a part may be rejected. 

The representative of the Engineer (or of the Purchaser) may then take similar samples 
from a further two rails out of the same 100 tons, and should either fail to comply with the specified 
analysis the whole 100 tons may be rejected. 

In ease of difference between the Enpneer (or between the Purchaser) and the Manufac- 
turer, as to the accuracy of an analysis, either party shall have the right to have samples of the 
steel analyzed by an independent metallurgist, to be mutually agreed upon The expenses at- 
tendant upon such independent analysis shall be borne by the party adjudged to be in the wrong. 

3. Each Rail shall be made from an ingot not less than 12 inches square at the smaller 
end and 14 inches square at the larger end, which must be cogged down into blooms, and have 
sufficient crop then sheared from each end to ensure soundness. 

All straightening shall be done by pressure and not by hammering. 

4. A rolling margin of J per cent under to J per cent above the calculated weight will be 
permitted, but the calculated weight only will be paid for. 

6. Before the general manufacture of the Rails is conmienced the Manufacturer shall, if 
required hy the Engineer (or by the Purchaser), supply two sets of templates, internal and external, 
of approved material, for each " B. S." Section of Rail, 

Each template shall he suitably engraved with the Purchaser's name, the number of the 
" B, S." section (being the nominal weight of the Rail in pounds per yard), the Manufacturer's 
name and address, and the date of the Contract. 

These templates shall be submitted to the Engineer (or to the Purchaser) for his approval, 
and at the commencement of rolling the Engineer will have a competent person present to aj^rove 
of the section. 



>, Google 



RAIL SPECinCATlONS 



. TABLE OP GENERAL DIMENSIONS AND WEIGHTS OF "B. 8." 

(3« PlaU XVI 



Niimbar al "B. S." 








Vaid iBPounds. 


"*"■»■"■ 


WidlhofH«d. 


oIRul. 




jMh«. 


iDChM. 


Pound, per YKd. 


20 


2) 






19.96 


25 


21 






24.95 


30 








29.98 


35 








35.03 


40 








39,98 


45 






1 


45,10 


50 


fj* 


2Vi 


49.94 


65 


22 


54,78 


60 


4^ 


2 




60.11 


65 


2 


fk 


64.86 


70 






6S.77 


75 




It 


74,79 


80 


6 


79.94 


85 


It 


2A 


84.87 


90 


89,92 


95 


5A 


2H 


94,76 


100 


5i 


99.95 



Genend 
Dimensiona 
and Wei^ts 
of Kails. 



7. Each Section of Rail shall be accurately rolled to ite respective template. 

8. The whole of the Rails shall be of uniform section throughout, true to templates, per- 
fectly sound and straight, and free from splits, cracks, burra, and defects of every kind. 

9. A quantity of short lengths will be taken in such lengths and quantities as may be ordered 
by the Engineer {or by the Purchaser), provided that these short lengths are cut down from longer 
lengths found to be defective at the ends only, and that the total quantity taken does not exceed 
71 per cent of the Contract. 

10. The Rails shall be the specified length at a temperature of 60° Fahr, No Rail will be 
accepted which is more than three-sixteenths of an inch (.-A inch) above or below the length 
specified, whether for straight or curved Unes. 

11. When required by the Enpneer (or by the Purchaser) Rails are to be supplied from 1 
to 6 inches shorter or longer than the normal specified lengths, and these special lengths are to 
have about one foot at each end painted with such colors as may- be ordered. 

12. Rails shall be suppUed for switches and crossings when so ordered, and such Rails shall 
he of the required lengths and shall be cut from sound R^ls. 

13. The Brand (see sketch) shall be rolled on the web of each Rail to show that the Rail 
is of British Standard Section and made under the conditions of this Specification; the number 
of the " B. S." Section (being the nominal weight of the Rail in pounds per yard), the process* by 
which the Rails have been manufactured, the Manufacturer's name, initials, or other recognized 
mark, and the month and year of manufacture shall also be rolled, in letters three-quarters of 

an inch (} inch) in laie, on one side of the web of each R^I, e.g., W B,8. 95-B.A,* 4.04; 

and the number of the cast from which it has been rolled shall be stamped on the end of each Rail 
in half-inch () inch) block figures, 

14. From each cast a piece of Rail (which may be a crop end) shall be selected by the repre- 
sentative of the Engineer (or of the Purchaser) and stamped with his mark and the number of 
the cast. From this a piece 5 feet long shall be cut which shaU be placed in a horizontal position, 
with the bead uppermost, upon two iron or steel supports resting on a solid foundation, the upper 

* The ftdlowing abbreviations are recommended: — 
S.A. Siemens-Martin Acid. 
S.B. Siemens-Martin Bauc. 



Rails to Conform 
to Templates. 
Rails to be Free 
from Defects. 
Length of Rails 
for Straight 



Penniaaible 
Variation in 
Length. 



Curved Line. 
Rails for 
Switcbes and 
Crossings. 
Branding. 



>, Google 



490 



STEEL RAILS 



surfaces of the supports being curved to a, radius of 3 inches. The test shall comprise one blow, 
delivered midway between the bearings, from a falling iron weight or tup, the striking face of 
which shall be rounded to a radius of not more than 5 inches. The weight of the tup, the span 
of the test piece between the centers of the bearings, and the height of the drop for the various 
sections of Rails shall be as tabulated below. The blow must be sustained without fracture. In 
addition to the above test the representative of the En^eer (or of the Purchaser) shall select one 
finished Rjul from every 200 offered, and a piece S feet in length cut from this R^I shall be si " " 
tested as specified above. 



Nuiiib«'ot"B,S." 


FdliuWtwhtTaK. 




*'°4'5#''"' 








WBghtoITnp. 


C«MnotBi>rii«L 


Drop. 




CwU. 


F»t. 


Few. 


20 


5 


3 


8 


25 


5 


3 


9 


30 


10 


3 


10 


35 


10 


3 


12i 1 


40 


10 


3 


15 


45 


15 


3 


15 


60 


15 


3 


15 


&S 


15 


3 


171 


80 


20 


3 


20 


65 


20 


3 


20 


70 


20 


3 


20 


75 


20 


3 


20 


80 


20 


3 


22 


85 


20 


3 


24 


90 


20 


3 


26 


95 


20 


3 


28 


JOO 


20 


3 


30 



Should the length cut from the selected Rail fail to comply with the test specified for its 
weight, two other Rails from the same cast will be selected and similar lei^hs cut and tested, 
and the acceptance or rejection of the cast wili be decided by the result of the three tests, so that 
if two of the Rails selected fail to comply with the test the entire cast will be rejected. 

15. From each 100 tons of R^ls the Manufacturer shall (if required by the representative 
of the Engineer or of the Purchaser) cut a teat piece from any IRaii selected as a sample Rail; such 
test piece to be stamped to correspond with the sample Rail. . It shall then be placed in a testing 
machine of approved pattern, and shall have an ultimate tensile strength of not leas than 40 tons 
per square inch, nor more than 48 tons per square inch, with an elongation of not less than 15 per 
cent upon the Standard Test Keces C or D (see Fig. 341). Should the teat piece fail to fulfil these 
conditions, the representative of the Engineer (or of the Purchaser) may require the Manufacturer 
to test two other Rails from the same cast in the aame manner, and the acceptance or rejection 
of the cast shall be decided by the result of the three testa, so that if two of the three Rails selected 
fail to comply with the test the entire cast will be rejected. 

The representative of the En^neer (or of the Purchaser) may then take similar test pieces 
from a further two Rails out of the same 100 tons, and should either fail to comply with the test 
the whole lOO tons may be rejected. 

Should the Engineer (or the Purchaser) desire to have independent tests made, the Manu- 
facturer shall provide the necessary test pieces, viz., two for every 200 tons, properly shaped and 
prepared as described in Fig. 341. 

16. The holes for fishbolts shall be drilled through the web from the solid at each end of 
the R^ls, of the azes and in the position shown in the British Standard Specification for Fish 
Plates for Flat Bottom Rails (Report No. 47), or on a drawing to be supplied by the Ei^neer 
(or by the Purchaaer), These holes must be clean and square with the web, without burrs on 



>, Google 



RAIL SPECinCATlONS 491 

either side, and will be checked with the gauges to be furnished to the Manufacturer by the 
Engineer (or by the Purchaser). Should any of the holes vary from the correct size or position 
more than one thirty-second of an inch (Vi inch) the Rails in question wiU be liable to rejection. 

17. The Manufacturer shall give to the Engineer (or to the Purchaser), or his representative, Notice of 

at least seven clear days' previous notice, in writing, before the rolling of the first lot of Rails, R?'!''** *» Be 
and at least three clear days' previous notice, in writing, before the rolling of any subsequent lot 
of Rails, is commenced, in order that arrangements may be made for the presence of the repre- 
sentative of the En^eer (or of the Purchaser) at the rolling. 

18. The Engineer (or the Purchaser), or his representative, shall have free access to the Inspection and 
works of the Manufacturer at all reasonable times: he shall be at Uberty to examine the Rails Testiiij. 
during any stage of their manufacture, and to reject any material or finished Rail which does not 

conform to the terms of this Specification. 

Before the Rails are put before the representative of the Engineer (or of the Purchaser) 
for inspection, the Manufacturer shall have them examined, and all R^ls which he admits to be 
defective shall be sorted out and placed in a separate stack; the representative of the En^neer 
(or of the Purchaser) being empowered to refuse to inspect any lot of Rails not put in uniform 
lengths and sorted. 

19. The Manufacturer shall supply the material required for testing free of charge and shall. Testing 
at his own cost, furnish and prepare the necessary test pieces, and supply labor and appUances PaeUitieB. 
for such testing as may be carried out on his premises in accordance with this Specification. Failing 
facilities at his own works for making the prescribed tests, the Manufacturer shall bear the cost 

of carrying out the tests elsewhere. 

20. All Rails accepted by the representative of the En^eer (or of the Purchaser) shall be Haiking nt 
stamped in hie presence. Accepted Kails 

38. Specifications for Street Railway Raiis 

American Society for Testing Materials, AfUialed mth the International Asaociation for Testing 
Materials. — Standard Specifications for Open-hearth Sted Girder and High Tee Rails. Adopted 
June 1, 191S. 

I. Mahofacttirz 

1. The steel shall be made by the open-hearth process. The entire process of manufac- Process, 
ture and testing shaU accord with the best current practice. 

2. Bled ingots, and ingots or blooms which show the effects of injurious treatment, shall Bled Ingots, 
not be used. 

3. A BuflScient discard from the top of each ingot shall be made at any stage of the manu- Discard. 
facture to obtain sound rails. When finished rails show pipii^, they may be cut to shorter lengths 

until all evidence of this is removed. 

II. Chemical PnopEETiEe and Taars 

4. The steel shall conform to either of the following requirements as to chemical composi- Chemical 
tion, as specified in the order: Composition. 

Clah a. Clam B. 

Carbon, per cent 0.60-0,75 0,70-0.85 

Manganese, per cent 0.60-0.90 0,60-0-90 

Silicon, percent notoverO,20 not over 0.20 

Phosphorus, per cent not over 0.04 not over 0,04 

5. To determine whether the material conforms to the requirements specified in Section 4, J'^je 
an analysis shall be made by the manufacturer from a test ingot taken during the pouring of ^^^' 
each melt. Drilling for analysis shall be taken not less than i inch beneath the surface of the 

teat ingot. A copy of this analysb shall be given to the purchaser or his representative. 

6. A Check analy^ may be made from time to time by the purchaser from a test ingot or Check 
driUii^ therefrom furnished by the manufacturer. Analyses. 



>, Google 



STEEL RAII^ 



Ntunber 
of Tests. 
Retests, 



III. Physical Properties and Tests 
7. (o) The teat specimen shall be test«d on a drop-teat machine of the type recommended 
by the American Railway Engineering Association. The specimen shall be placed head upwards 
on the supports of the machine, and shall not break when tested with one blow in accordance with 
the following conditions : 



Waithl^dHaitbtof RaU. 



Rails weighing over 100 lb. per yd. 

and over 7 in. in depth ... 

ig 100 lb. or less per 
. or less in depth. , . . 



Rails y 

yd..- 



60-120 
60-120 



Hai«ht at Etrop. 



2000 
2000 



(b) The atmospheric temperature at the time of testing shall be recorded in the test 

report. 

(c) The testing shall proceed concurrently with the operation of the works. 

8. (a) Three rails, each from the top of one of three ingots from each melt, shall be selected 

by the inspector, and a test specimen shall be taken from each of two of these. 
(b) Drop test specimens shall not be less than 4, nor more than 6 feet in length. 

9. Two drop tests shall be made from each melt. 

10. If the result of the drop test on only one of the two specimens representing the rails in 
a melt does not conform to the requirements specified in Section 7, a retcst on a specimen from 
the third rail selected shall be made and this shall govern the acceptance or rejection of the raib 
from that melt. 



IV. Standard Secttons, Lengths, and Weights 
11. (a) The cold templet of the manufacturer shall conform to the specified section as shown 
in detail on the drawing of the purchaser, and shall at ail times be maintained 
perfect. 
(6) The section of the rtul shall conform as accurately as possible to the templet, and 
within the following tolerances: 

(1) The height shall not vary more than A inch under nor more than ^ inch 

over that specified. 

(2) The over-all width of head and tram shall not vary more than j inch fron; 

that specified. Any variation which would affect the gage line mort 
than I*! inch will not be allowed. 

(3) The width of base shall not vary more than j inch under that specified for 

widths less than Gi inches; -^n inch under for a width of 6} inches; and 
i inch under for a width of 7 inches. 

(4) Any variation which would affect the fit of the splice bars will not be allowed. 

(5) The base of the rail shall be at right angles to the web; and the convexity 

shall not exceed V: inch, 
(c) When necessary on account of the type of track construction, and notice to that 
effect has been given to the manufacturer, special care shall be taken to mainlaio 
the proper position of the gage line with respect to the outer edge of the base. 
12. (a) Unless otherwise specified, the lengths of rails at a temperature of 60° F. shall be 
60 and 62 feet for those sections in which the weight per yard will peiroit. 
(6) The lengths shall not vary more than 1 inch from those specified. 
(c) Shorter lengths, varying by even feet down to 40 feet, will be accepted to the extent 
of 10 per cent by weight of the entire order. 



>, Google 



RAIL SPECIFICATIONS 493 

13. (a) The weight of the rails per yard as epecified in the order shall be maintained as Weight. 

nearly aa possible after confonmng to the requirements specified in Section 11. 

(b) The total weight of an order shall not vary more than 0.5 per cent from tliat specified. 

(c) Payments shall be based on actual weights. 

V. Workmanship and Finish 

14. (o) Rails on the hot beds shall be protected from water or snow, and shall be carefully Straightening. 

manipulated to minimize cold straightenii^. 
(b) The distance between the rail supports in the cold-strmghtening presses shall not 
be less than 42 inches, except as may be necessary near the ends of the rails. The 
gag shall have rounded comers to avoid injury to the rails. 

15. (a) Circular holes for jomt bolts, bonds, and tie rods shall be drilled to conform to the Drilling aiul 

drawings and dimensions furmahed by the purchaser. Punching. 

(b) In Class A rails the tie-rod holes may be punched. 

16. The ends shall be milled square laterally and vertically, but the base may be undercut Hilling. 

1*1 inch. 

17. (a) Rails shall be smooth on the head, straight in line and surface without any twists, Finiah. 

waves, or kinks, particular attention being given to having the ends without kinks 

(6) All burrs or flow caused by drilling or sawing shall be carefully removed, 

(c) Raih shall be free from gag marks and other injurious defects of cold-straightening. 

VI. CnASaiFiCATioN of Raiu 

18. Rails which are free from injurious defects and flaws of all kinds shall be classed as No. 1 No. i Rails, 

Rails. 

19. (o) Rails which are rough on the head or which by reason of surface or other imper- No. a Sails. 

fections are not classed as No. 1 rails, shall be classed as No. 2 rails; providing 
they do not, in the judgment of the inspector, contain imperfections in such num- 
ber and of such character as to render them unfit for No. 2 rail uses, and pro- 
viding they conform to the requirements specified in Section II. 

(b) Rails which have flaws in the head exceeding t inch in depth, or in the base exceed- 

ing i inch in depth, shall not be classed as No. 2 nuls. 

(c) No. 2 riuls will be accepted to the extent of 10 per cent by weight of the entire order. 

VII. Marong and Loaping 

20. (a) The name or brand of the manufacturer, the year and month of manufacture, the Marking. 

letters "O. H.," the weight of the rail, and the section number shall be iegibly 
rolled in raised letters and figures on the web. The melt number shall be legibly 
stamped on each rail where it will not be covered subsequently by the joint plates. 

(it) Both ends of ail shortr-length No. 1 rails shall be pmnted green. 

Both ends of all No. 2 rails shall be painted white and shall have two heavy center- 
punch marks on the web at each end at such a distance from the end that they 
will not be covered subsequently by the joint plates. 

21. (a) Rdls shall be loaded in the presence of the inspector, and shall be handled in such Loading. 

a manner as not to bruise the flanges or cause other injuries. 

(b) Rails of each class shall be placed together in loading. 

(c) Rails shaU be paired as to length before shipment. 

VIII. Inspection 

22. The inspector representing the purchaser shall have free entry, at all times while work Inqwctioii. 
1 the contract of the purchaser is being performed, to all parts of the manufacturer's works 



>, Google 



494 STEEL RAILS 

which concern the nmnufacture of the material ordered. The manufacturer shall afford the 
inspector, free of cost, all reasonable facilities to satisfy him that the material is being furnished 
in accordance with these specifications. All tests and inspection shall be made at the place of 
manufacture prior to shipment, and shall be so conducted as not to interfere unnecessarily with 
the operation of the works. 

39. BiBUOGEAPHY OF RAIL SPECIFICATIONS 

Search furnished by the Secretary of the American Society of Civil Engineers, and made 
in its library, January 24, 1910, and January 19, 1912, supplemented by the Teclmology 
department of the Carnegie Library of Pitteburgh. 

1877 
" Permanent-way Rolling Stock and Technical Working of R^lways," Vol. 1, p. 512, by 
Ch. Couche, tr. by James N. Shoolbred. Paris, 1877. Dunod, 49 Quai des Augustins. (Contains 
specifications for cliemical composition of steel rails.) 

1879 

"The Chemical Composition and Physical Properties of Steel Rails," by C. B. Dudley. 
Trans. Am. lust. Min. Engra., Vol. 7, p. 172 (1879). (Recommends a formula for the chemical 
composition of rails for the use of the Pennsylvania Railroad.) 

" Does the Wearing Power of Steel Rails Increase with the Hardness of Steel? " by Charles 
B. Dudley. Trans. Am. Inst. Min. Engrs , Vol. 7, p. 202 (1879) (four pages). 

" Discussion of Dr. Charles B. Dudley's Papers on Steel Riuls." Trans. Am. Inat. Min. 
Engrs., Vol. 7, p. 357 (1879) 

1380 

" The Wearing Capacity of Steel Rails in Relation to their Chemical Composition and 
Physical Properties," by Charies B. Dudley. Trans. Am. Inst. Min. Engrs., Vol. 8, p. 321 (1880). 
(Contains references to specifications for chemical composition.) 



" Specifications for Steel RiuU and Track Fastenings." Eng. News, Vol. 20, p. 172 (Sept. 
1, 1888). (Specifications drawn up and used by Frank Ward & Bro., of Kttaburg, Pa.) 

Same. R. R. Gai., Vol. 20, p. 587 (Sept. 7, 1888). 

" Steel Rails and Specifications for their Manufacture," by Robert W. Hunt. Trans. Am. 
Inst. Min. Engrs., Vol. 17, p. 226 (1888). (Contains specifications for Bessemer steel rails.) 

Abstracts of same. R. R. Gaz., Vol. 20, p. 697 (Oct. 26, 1888); Eng. and Min. Jour., 
Vol. 46, p. 370 (Nov. 3, 1888). 

1893 

" Steel Rdls: their Manufacture and Service." Eng. News, Vol. 30, p. 172 (Aug. 31, 1893). 
(Proposed spet^fications for steel rails.) 

1895 

" Specifications for Steel Rails of Heavy Sections Manufactured West of the Alleghenies," 
by Robert W. Hunt. Trans. Am. Inst. Min. Engrs., Vol. 25, p. 653 (1895). (The author states 
that " the only important features in which the present specifications differ from those of 1888 is 
in providing for a chemical composition and for drop tests.") 



" Brief Note on Rail Specifications," by Robert W. Hunt. Trans. Am. Inst, Mjn. Engrs,, 
Vol. 27, p. 139 (1897). (One page; report of progress.) 



,y Google 



RAIL SPECIFICATIONS 



" Specifications on Stnictural Steel and Rails," by W. R. Webater. Journal of the EVankiiii 
Institute, Vol. 147, p. 1 (Jan., 1899). (General diacusdon of the subject.) 

1900 

Proceedings American Railway En^neering and Maintenance of Way Association, Vol. 1, 
p. 116 (1000). (Refers to specifications for steel rails.) 

" Recent Practice in Rails: an Informal Discussion." Trans. Am. Socy. of Civ. Engrs., 
Vol. 44, p. 489 (Paper 887, Dec, 1900). (Gives tiie standard rail specifications of the Louisville 
& Nashville R, R, Co., Robert W, Hunt's specifications, and specifications of the Western rail 
mills, and a review of foreign rail specifications.) 

"American Standard Specifications and Methods of Testing Iron and Steel," by Albert 
Ladd Colby. Journal of the Iron and Steel Institute, Vol. 158, p. 215 (1900). (Gives specifi- 
cations for steel rails.) 

1901 

" Proposed Standard Specifications for Steel Ruls." In Proc. Ant. Ry. Ei^. and M. of 
W. Assn., Vol. 2, p. 192 (1901). (Specifications recommended by Committee No. 1 of the Ameri- 
can Section of the International Association for Testing Materials.) 

"Nature of Metal for Rwb, Report (United States)," by P. H. Dudley. International 
Rwlway Congress. Proceedinp, Sixth Session, 1900, Vol. 1, Question 1, pp. 205, 248, Brussels, 
1901. P. Weissenbruch, 49 Rue du Poingon. (Gives specifications for steel rwls.) 

" Examen des Specifications Normales Americaines Proposes Eprouvettes et M4tbodes 
d'Easai du Fer et de I'Acier," by Albert Ladd Colby. In " ConununieatJons Pr^nt^ devant 
le Congres International des M^hodes d'f^ssai des Mat4riaux de Construction tenu a Paris du 
9 au 16 Juillet, 1900," Vol. 2, Pt. 1, pp. 147, 162. Paris, 1901. Vve. Ch. Dunod, 49 Qum des 
Grands-Augustins. (Contuns a review of foreign specifications for steel rails aiid proposed 
American standard specificatJona.) 

" Specifications for Steel Ralls," by P. H. Dudley. R. R. Gaz., Vol. 33, p. 158 (March 8, 
1901} (one page). 

" Some Suggestions as to SpecificatJons for Steel Rails," by E. P. Kenney. Eng. News, 
Vol. 46, p. 226 (Oct. 3, 1901). 

1902 

American Society for Testing Materials, Proceedings, Vol. 1, pp. 101, 264 (1899-1902). 
(Proposed standard specifications for steel rails recommended by American Branch of Coomiittee 
No. 1, American Section of the International Association for Testing Materials. 

" Review and Text of the American Standard Specifications for Steel, adopted in August, 
1901," p. 41, by Albert Ladd Colby. Ed. 2. Easton, Pa., 1902. The Chemical Publishing Co. 
(Contdns specifications for steel rails.) 

" Proposed Modifications of the Standard Specifications for Steel Rails, Topical Discussion." 
Proc. Am. Socy. for Testing Materials, Vol. 2, p. 9, 23 (1902). 

Proceedings of the Am. Ry. Eng. and M. of W. Assn., Vol. 3, p. 201 (1902). (Specifications 
recommended in 1901 with some amendments.) 

" Speofications for Steel Rails/' by W. R. Webster, Trans. Am. Inst. Min. Engrs., Vol. 31, 
pp. 449, 967 (1902), (Contains proposed standard specifications recommended May, 1900, by 
the American Branch of Committee No. 1 of the International Association for Testing Materials.) 

" Steel Rails: Specifications," by Robert Job. Am. Eng. and R. R. Jour., Vol. 76, p. 310 
(Oct., 1902). (Givesspecificationsof thePhiladelphia&ReadingRailwayCompanyforsteelrailB.) 



" The Present Situation as to Specifications for Steel Rwls," by William R. Webrter. Trans. 
Am. Inst. Wn. Engrs., Vol. 33, p. 164 (1903) (five pages). 



,y Google 



496 STEEL RAILS 

"Proposed ModilicationB in the Specifications for Steel Rails adopted by the Anierican 
Railway En^neering and Maintenance of Way Association in March, 1903." Proc. Am. Socy. 
for Testing Materials, Vol. 3, p. 74 (1903). 

1904 

" British Standard Specification and Sections of Bullheaded Railway Rails." Engineering 
Standards Committee. Report No. 9, Lond., 1904. (lOs. 6d. net.) 

" Specifications for Steel Raiia of the American Railway Ei^neering and Maintenance of 
Way Association, aa Amended and Adopted in March, 1904," with Introduction by William R. 
Webster. Proc. Am. Socy. for Testing Materials, Vol. 4, p. 195 (1904). (Showii^ main difference 
in specifications adopted by the two societies.) 

" Standard Specifications for Bessemer Steel R^ls." In Proc. Am. Ry. Eng. and M. of 
W. Assn., Vol. 5, p. 465 (1904). 

" Specifications for Bessemer Steel R^Is." Eng. News, Vol. 50, p. 275 (March 24, 1904). 
(Gives specifications adopted by the American Railway En^neering and Maintenance of Way 
Association.) 

1905 

" Railroad Construction," p. 243; by Walter Loring Webb, M. Am. Socy. C. E. Ed. 3. 
N. Y., 1905, John Wiley & Sons, 43 E. 19th St. (Contains proposed standard specifications for st«el 
rails of the American Railway Enpneering and Maintenance of Way Association, March, 1902.) 

British Standard Specification and Sections of Flat-bottomed Railway Rjuls. Eng^eering 
Standards Committee. Report No. 11. Lond., 1905. Leslie S, Robertson, Secy., 28 Victoria 
Street, Westminster, S. W. (IDs. 6d. net.) 

" Steel Rails," by William R. Webster. R. R. Gaz., Vol. 38, p. 440 (May 5, 1905). (Gives 
specifications of the Ataerican Railway Ei^neering and Maintenance of Way Association.) 

1906 

" On Specifications for Steel Rails." Proc. Am. Socy. for Testii^ Materials, Vol, 6, p. 35 
(1906). (Gives proposed standard specifications for steel rails.) 

Proc. Am. Ry. Eng. and M. of W. Assn. Vol. 7, p. 553 (1906). (Gives comparison of 
spedfications of American Railway Engineering and Maintenance of Way Association and the 
American Society of Civil Engineers.) 

" Rails for Lines with Fast Trains." Reports; by P. H, Dudley and Van Bogaert. Inter- 
national Railway Congress, Seventh Session, 1905, Vol. 1, Question 2, pp. 141, 194. Brussels, 
1906. P. Weissenbruch. {Contains very brief data on rails specifications.) 

"Specifications for Steel Rails." R. R. Gaz., Vol. 40, p. 280 (March 16, 1906). (Specifi- 
cations recommended by the American Railway Engineering and Maintenance of Way Association, 
American Society for Testing Materials, and American Society of Civil Engineers.) 

1907 

" Report of the Special Committee on Rail Sections to the American Society of Civil En- 
gineers." Proc. Am. Socy. of Civ. Engrs., Vol. 32, p. 52; Vol. 33, p. 290 (1906, 1907). {Contains 
recommended specifications for Bessemer steel rails.) 

" Manual of Recommended Practice for Railway Engineering and Mtuntenance of Way," 
p. 55. Am. Ry. Eng. and M. of W. Assn. Edition of 1907. (Contains specifications for rails.) 

" Proposed Standard Specifications for Steel Rails." Proc. Am. Socy. for Testing Materials, 
Vol. 7, p. 40 (1907). (Specifications adopted Sept. 1, 1907.) 

" The Steel-rail Discussion, American Society for Testing Materials." Ry. and Eng. 
Review, Vol. 47, p. 570 (June 29, 1907). 

" Proceedings of the Session of the American Rwlway Association, October 30, 1907," p. 176, 
N, Y., 1907. (Report of the Committee on Rwl Sections givii^ specifications for Bessemer 
steel rails with explanatory notes.) 



>, Google 



RAIL OTECIFICATIONS 497 

" Standard Specifications for Steel Rails." Proc. Am. Soc. for Testing Materials, Vol. 7, 
p. 44 (1907). 

Same. Eng. Record, Vol. 55, p. 774 (June 29, 1907). 

Comparison of American and Foreign Rail Specifications, with a Proposed Standard Speci- 
fication to Cover American Rails for Export," by Albert Ladd Colby. Trans. Am. Inat. Min, 
Ei^rs., Vol. 37, p. 576 (1907). (Contains bibliography.) 

Same. Iron and Coal Trades Review, Vol. 73, p. 357 (July 27, 1906). 

" Rail Sections and Specifications." R. R. Gaz., Vol. 43, p. 250 (Sept. 6, 1907). (Com- 
pares rail specifications of the American Society of Civil Engineers, American Railway Engineering 
and Maintenance of Way Association, American Society for Testing Materials.) 

" Some Progress Toward Getting Better Rails." (Editorial.) R. R. Gaz., Vol. 43, p. 577 
(Nov. 15, 1907). (Brief discussion of rail specifications.) 

" Proposed Standard Rail Sections of the American Railway Association." R. R. Gaz., 
Vol. 43, p. 627 (Nov. 22, 1907) (illustrated). 

" Rail Specifications." R. R. Gaz., Vol. 43, p. 736 (Dec. 20, 1907). (Contains specifiea- 
tjons of the American Railway Association.) 

1908 

" Railway Track and Track Work," p. 78; by E. E. Russell Tratman, Assoc. M. Am. 
Sooy. C. E. Ed. 3. N. Y., 1908. Enpneerii^ News Publishii^ Co., 220 Broadway. S3.50 
net. (Contains a comparative table of specifications for chemical composition of rails.) 

Proceedings of the American Railway AssocialJon, Special Session, February 7, 1908; 
Regular Session, April 22, 1908, p. 359, N. Y., 1908. (Specifications for Bessemer and open- 
hearth steel rails, accompanying the Report of the Committee on Standard Rail and Wheel Sec- 
tions, dated March 23, 1908.) 

" Standard Specifications for Steel Rails." Proc. Am. Soc. for Testing Materials, Vol. 8, 
p. 44 (1908). (Specifications adopted Aug. 15, 1908.) 

" The Present Status of Rail Specifications." R. R. Age, Vol. 45, p. 76 (Jan. 17, 1908). 
(A review of the action taken hy the American Railway Association.) 

American Society of Civil En^neers, Report of Special Committee on HaH Sections. Eng. 
News, Vol. 59, p. 105 (Jan. 23, 1908). (Recommended specifications for Bessemer steel rails.) 

" New Steel-rail Specifications of the Pennsylvania Railroad." Eng. News, Vol. 59, p. 426 
(April 16, 1908). (Gives specifications for chemical composition, process of manufacture, mechani- 
cal requirements, tests, and inspection.) 

" The Pennsylvania New Rail Sections and Specifications." R. R. Gaz., Vol. 44, p. 539 
(April 17, 1908). 

" New Rail Sections and Rail Specifications of the American Railway Association." Eng. 
News, Vol. 59, p. 530 (May 14, 1908). (Specifications for Bessemer and open-hearth steel rails.) 

"American Railway Association's Rail Committee." (Editorial.) Eng. News, Vol. 59, 
p. 533 (May 14, 1908). (Comments on the rail specifications; one and a half colunms.) 

" Steel-rail Breakages; Questions of Design and Specifications," by Harold V. Coes. En- 
gineering Magazine, Vol. 35, p. 417 (June, 1908). (Gives specifications for the Union and Southern 
Pacific railways and British standard chemical specifications for steel rails.) 

" Some Features of the Present Steel Rail Question," by Charles B. Dudley, Proc. Am. Soc. 
for Testing Materials, Vol. 8, p. 19 (1908). (Discusses changed demands on steel rails and pro- 
posed specifications.) Same. Engineering News, Vol, 60, p. 9. 

1909 
" Proceedings of the Session of the American Railway Association held in Chic^o, Novem- 
ber 17, 1909," p. 995. N. Y., 1909. W. F. Allen, Secy., 24 Park Place. (American Railway 
AssociatJoD specifications for Bessemer and for open-hearth steel nuls, adopted as lecoomiended 
practice April 22, 1908.) 



,y Google 



498 STEEL RAII5 

Proceedings Am. Ry. Eng. and M. of W. Assn., Vol. 10, Pt. I, pp. 369, 374 (1909). (Recom- 
mended changes in epecificaUoDS as previously adopted by the Association.) 

" New Rail Section and Specifications, Canadian Pacific Ry." Ry. and Eng. Review, VoL 
49, p. 27 (Jan. 9, 1909). (Gives specifications for open-hearth and Bessemer rails.) 

" New Ruls for the Canadian Pacific Ry." (Editorial.) Ry, and Eng. Review, Vol. 49, 
p. 34 (Jan. 9, 1909). (Discusses specifications and rail gections.) 

"New R^ Specifications of the Pennsylvania R. R. System." Eng News, Vol. 61, p. 50 
(Jan. 14, 1009). (Revision of specifications of Feb. 4, 1908.) 

" Pennsylvania RmI Specifications." R. R. Age Gaz., VoL 46, p. 101 (Jan. 16, 1909). 
(Specifications of the Pennsylvania Riulroad evised under date of Dec. 10 1008.) 

"The New 85-pound Hai! Section of the Canadian Pacific Ry." Eng. News, Vol. fil, 
p. 272 (March 11, 1909) (illustrated). 

" Recent Developments in Rail Dengn and a Comparison of Ridl Sections." (Editorial.) 
Ei«. News, Vol. 61, p. 276 (March 11, 1909). (Compares rail specifications.) 

" Recent Rail Sections." R. R. A e Gaz., Vol. 46, p. 537 (March 19, 1909) (one page, 
illustrated). 

" New Raii Orders and Specifications." (Editorial.) R. R. Age Gaz., Vol. 46, p. 535 (March 
19, 1909). (Discusses nul specifications of various nuiroads.) 

" Riul Specifications." (Editorial,) R. R.A^Gaz., Vol. 46,p.925 (April 30,1909). (Very 
brief.) 

" R^l Spetafications " (letter), by R. Trimble. R. R. Age Gaz., Vol. 46, p. 1018 (May 14, 
1909), (Brief letter correcting error in the above editorial.) 

" Comparative RaU Specifications." R. R. Age Gaz., VoL 46, p. 1066 (May 21, 1909). 
(Compares specifications of the American Railway Association, Steel Manufacturers of America, 
American Society of Civil Engineers, American Railway En^neering and Mfuntenance of Way 
Association, and American Society for Testing Materials, with comments.) 

" Rul Sections and Specifications." (Editorial.) R. R. Age Gaz., Vol. 46, p. 1060 (May 21, 
1909). 

" Specifications for 90-pound Bessemer and Open-hearth Steel Rails for the Harriman 
Lines." R. R. Age Gaz., Vol. 47, p. 185 (July 30, 1909). (Specifications to which the Harriman 
Lines are ordering thdr 1909 rails. 

" On the Question of Strengthening the Traclt and the Bridges with a View to Increasing 
the Speed of Trains Subject II, for Discussion at the Eighth Session of the Railway Congress," 
by M. L. Byers. Bulletin of the International Railway Congress Association, VoL 23, p. 908 
(Sept., 1909). (Gives rail specifications proposed by he American Railway Association and by 
the Pennsylvania Railroad Committee.) 

" Report of Committee on R^, American R^way Engineering and Miuntenance c^ Way 
Association. Bulletin 113 (Dec., 1909). (Specifications for steel rtuls and review of previous 
r^wrts.) 

Abstract of same. "Specifications for Steel Rails." R^way and Engineering Review, 
Vol. 50, p, 118 (Feb. 5, 1910). 

" Standard Specifications for Bessemer Steel Riuls." Proceedings American Society for 
Testing Materials, Vol. 9, p. 62 (1909). (Adopted Aug. 16, 1909.) 

" Standard Specifications for Open-hearth Steel Rails." Proceedings American Society tor 
Testing Materials, VoL 9, p. 66 (1909). (Adopted Aug. 16, 1909.) 

" La Voie Courante des Chenains de Fer de I'Etat Beige," by Pierre Decamps. Revue Gen- 
eral des Chemins de Fer et des Tramways, VoL 32, Pt. 2, p. 267 (Oct., 1909). (Appendix pv«B 
r^ specifications of the state railroad of Belgium.) 

" Revised Rail Specifications, Pennsylvania Railroad System." En^eering, VcA. 87, 
p. 218. 

British Standard Sections, No. 47. En^neering Standards Committee (1909). (British 
standard specifications for bull headed and flat bottom railway rfdls.) 



>, Google 



RAIL SPECIFICATIONS 499 

" Specifications for Standard Open Hearth Steel Rails for A. S. C. E. Sections," Cam^^ 
Steel Co., Jan. I, 1909. (Two leaflets.) 

" Specifications for Standard Bessemer Steel R^ls for A. S. C. E. Sections," Carnegie Steel 
Co., Jan. 1, 1909. (Two leaflets.) 

" Specifications for Steel Rails." Baltimore and Ohio R. R. Co., No. 163C., Jan. 25, 1909. 
(Two leaflets.) 

" Specifications for Open Hearth Steel Rails." Proceedings American Street and Inter- 
urban Railway Engineering Association, Vol. 7, p. 59 (1909). (Includes specifications adopted 
by the Traout Supply Co., Lorain Steel Co., Pennsylvania Steel Co., and the Manganese Steel 
RaUCo.) 

1910 

" Permanent Way." R. R. Engr., Vol. 31, p. !8 (Jan., 1910). (Gives Pennsylvania Railroad 
System specifications for steel rails.) 

" Final Report of Special Committee on Riul Sections." Transactions American Society of 
Civil En^neers, Vol. 70, p. 456 (Paper 1177, Dec, 1910). (Contains reprint of rwl specifics^ 
tions of the American Railway Engineering Association.) 

" The American Railway Association, The American Railway En^neering and Maintenance 
of Way Association, Specifications for Steel Raits." Proceedings American Railway Engineering 
and Maintenance of Way Association, Vol. 11, Pt. 1, p. 254 (1910). 

Abstracts of same. "Specifications for Steel Rails." Railway and En^neering Review, 
Vol. 50, p. 118 (Feb. 5, 1910). "Rail Specifications and Sections," Engineering News, Vol. 63, 
p. 384 (Mar. 31, 1910). 

" Track Standards and General Rules." Depari^ment of Munt. of Way, Metropolitan 
Street Railway Co. Elec. Ry. Journal, Vol. 35, p. 863. 

" Recent Work of the German Street and Interurban R^lway Association." Elec. Ry. 
Journal, Vol. 35, p. 38. (Considers specifications and standards agreed upon.) 

Hunt (Robert W.) & Co., Engineers. Bureau of Inspection, Testa and Consultation. 
(Includes "Specifications for Standard Open-hearth Steel Girder and High Tee-Rails," 1910, 
American Street and Interurban Railway En^eering Association, p. 5; and "Specifications for 
Standard Open-hearth Steel Girder and High Tee-Rails," Lorain Steel Co., Jan. 1, 1910, p. 14.) 



1911 

" Standard Specifications for Bessemer and Open-hearth Steel Rwls." March 21, 1910, 
United States Steel Products Export Co." (Year-book, American Society tor Testing Materials, 
1911, p. 202.) 

" Rail Sections and Specifications." Elec. Ry. Journal, Vol. 37, p. 8. (Editorial, dis- 
cusui^ progress toward uniform specifications in 1910.) 

" Interborough Rails for Tangents and Curves." Elec. Ry. Journal, Vol. 37, p. 82. (Gives 
recent modifications of specifications of open-hearth steel rails.) 

Same, abstract. Journal of the Iron and Steel Inst., Vol, 34, p. 619. 

" Manufacturers' Standard Specifications for Bessemer Steel Rails," Association of American 
Steel Manufacturers. Year-book, American Society of Testing Materials, 1911, p. 199. 

" Specifications for Basic Open-hearth Rails," New York Central Lines. (Specifications of 
Oct. 1, 1909, revised Jan. 11, 1911, to conform to manufacturers' sale per 100 pounds.) 

" Specifications." Report of Committee on Rail. Proceedings American Railway En- 
gineering and Maintenance of Way Association, Vol. 12 (1911), Pt. 2, p. 12. (Gives short report 
of pn^press.) 

Report of Conmiittee A-1. Proceedinp American Society for Testing Materials (1911), 
Vol. XI, p. 48. (Contains reference to international specifications for kuIs.) 



,y Google 



STEEL RAII£ 



1912 



" Specifications for Carbon Steel Rails." Proceedings American Railway Engineering As- 
sodation (1912), Vol. 13, p. 665. 

" New Specifications for Steel Rails." Iron Age, Vol. 89, p. 816. (Gives report of rail com- 
mittee at 1912 meeting of the American Railway Ei^neering Association and specifications 
adopted.) 

" Specifications for 85-pound and 100-pound Carbon Steel Rails," 1912, Pennsylvania Rail- 
load Company. (Two leaflets.) 

" Specifications for Standard Bessemer Steel Tee Rails," 1912 Catalogue, Maryland St«et 
Company, p. 10. 

" Specifications for Standard Open-hearth Steel Tee Ruls," 1912 Catalogue, Maryland 
Steel Company, p. 12, 

" Specifications for Standard Open-hearth Steel Girders and Kgh Tee R^ls," 1912 Cata- 
logue, Pennsylvania Steel Company, p. 14. 



,y Google 



APPENDIX 

REPORTS AND RECORDS 

The forms recommended by the Rail Committee of the American Railway 
Engineering Association, and contained in the 1911 Manual of the Association, 
are typical of the best practice, and are shown on ¥ig^. 342 to 359 incluave 
and Plate XXXIII. The explanation of the forms as ^ven by the committee 
is as follows: 

Group I. Repor'k of Rail Inspection and Shipment at the Mill 
This set of forms. Figs. 342-344 and Plate XXXIII, is for the use of the 
railroad company's Inspector at the mills where the rwl is rolled, and gives 
all the information necessary to inform the purchaser that his order has been 
manufactured in accordance with the specifications and shipped. 



,y Google 



>, Google 





ji. B. ^ C R. R. Co. 

Rwportof Chamloal and Rhyaloal .g^, 


-.■S.-;- 




Co,M 






Oidv 
No.of 




1 






P««ii.RoIl. _B«hdd«..v«aeeo( «™d4« .P«. 


Wc«ht ol Tup, 2000 Hm. 


«• Stnight«>Dg Pp 
Height of Drop .„ 


tt DBianct between Supportt. 3 f( 






HdtKi. 


,™l.^ 


1 


1 


1 


1 


1 






f|l 


^ 


i4°u 


^r 


£JS 




















































2 


























3 


























4 


























S 


























« 


























; 


























B 


























9 


























10 


Pen 


Uvln 


ntbtT 




otgmn 


ihouM 




dau 


lb. si 


lejKh 




■, th.1 


iu«l 11 






J 




s 






s 


3 


3 




t 


" 12 










luto 






<niu|i 


Kba. 








13 


























14 


























IS 


























18 


























IT 


























IS 


























IB 


























20 


























2t 


























23 


























23 


























24 


























2S 




















































27 


























28 


























sg 


























30 


























31 


























32 


Nol«— R«|uir«nenu of Standanl SpeciScaimnt an lo be su 


ed t>n line 1 


ihtC 


-fEn, 


meer 


« o( « 








App 


™.. 


i^.S^x-"~ 



Fio. 342. 

M. W. 401. — Report of Chemical and Physical Examination: 
This blank is filled out from the mill records under the supervision of the 
Inspector, and ^ves the chemical contents taken from the ladle analysis and 
the result of the drop test. (503) 



-lOOgle 



/i. B. & a R. R. Co. 
Certificate of inspection 



(OpnHoiUi.Bawi 
Uanufactured by 



_ Praces Riik lt«. pwyii._. 



Mt ..Chief Engineer M. o( W. Date 

The following Steel Rails have been in^iecieil and aciepted according to contract. 
Rail! are certified to be within the limits of the Specilkationi of the 



AU Ruils have been inspected and approved (or Chcmiqal Analysis, Physical Test*, Section. 
Weight. Straightening. Drilling, Saving. Length, Stamping. Finish, Quality. 

end poution occupied in Oie ingot. Dateol Rolling j_j£_ 

No.ot Rails Rolled _ No. o( Rails Accepted ^ 

No. of Rails temporarily rejected and cause 






;#'" 



This Cotilicate coven the run fimi ^ 

Heat No to Meat No both incl 



Calculated Weight. 



«pted under this Ctnitv 



Hours Wis. Houn Wit. 



Approved : , 



Fia. 343. 

M. W. 402. — Certificate of Inspection: 

This is the Inspector's written statement that the material which he has 
witnessed rolled has been turned out strictly in accordance with the specifications 
and tlie order of the railroad company. (Sm) 



.yGoogle 



13 

e IS 
le 

17 

23 
24 
2S 
56 

11 
28 
29 
30 

32 


RCRC 


5RT 


£^ C *. ^. Ci. 

OF SHIRMENT. 




No 

_ Secuon 












Coiuigns 
Orf«N 


1 










LoxdedODCan. 


Nambei of R«Is of each Length. 


""t:!"^ 


[nilui 


No 




33 


30 




2TJ 




:u 




?js: 


























































































































































































































































































































F™sih.t«.-iih 


-p^nl 


colon 


U.bd.: 


Jh>p 


«lm 


nllu^ 


10 UK 


ueivc 


hy .h 


Bs^rei, 


c 


g 


1 


i 


Jl 


Jl 


J 


J 


J 


s 


e 


8 








S.I 


gl it. 


Ilnqu 


'da>i 


































































































































































































































































































































































































































Toial Wciuhl Eipressed iii Grcsi Tons and D«imiiJi. 








.«.T«ueT,o«. ,;^,^ 






Onto 
Chirf E 
the Gen 


■T»l Supe 


s 


tnd 
ndcnt 


^:T? 


Th'^ 


Ap 


^ 


— ?^i^¥i:" 



Fia.344. 

M. W. 403. — Report of Shipment: 

This blank is used for reporting the number and length of rail shipped in 
each car from the works, and, when properly checked by the Receiving Officer, 
it furnishes the basis for payment of the bill. 

M. W. 418. — Results of Drop Tests and Surface Inspection of Rails Rolled 
(Plate XXXIII). (605) 



This form is intended for tabulating the results of drop tests and surface 
inspection of r^ls rolled. 

Group II. Reports from Division Officers 
This group, Figs. 345-347, contains all the regular reports which come from 
the division officers concerning the rsuls which have been put in service in track. 



A. B. & a J?. J?. Co. 



Report of RAIL FAILURES in Main Tracks 

Date of Report 191... 



Brand an Rail? ("D" on back) 

Kind of Steel? ("E" on back) 

He»l No. on Rail? ("F"on back),.,. 
Rail No. or Letter f ("F" on back).. 
* if Hail? 



Oririnal Length of R 
Montb and V^r Rail 



Wbich fraekV.! ".';:; '!.'";; 

Which Rail? 

On Curve or Straight Line?.. 

No. of Curve? 

Degree of Curve? 

High or Lo* Rail, if on Corr 



) Rail "Broken"? , 

r "Detccti™"? 

r Damaged? 

e "Description of Failuret" oi 



Wai Rail much or titti 
By whom diEcovered.. 
Dale and Time found? 
Was Rail removed?., 



Exact gage of Track »t "Break"? 

Was "Break" over or between Tie*?.. 

Was "Break" square or angntar? 

DisUnce between edges of Tia at 



ci each tide of "Brei 











Kind of materialin roadbed unde 


bii: 



Was Track well A 



er? (Wet, dtj. warm or cold, fre. 



of flaws or defects, and If pofsible. 



)raw on Diagram lines of "Break," or partial fracture, aoch ai long mccca from 
nde of head and hatt-moon piecei from ba», ihowing dimensons. Hollows in 
head should be ihovm on "End Section." Defect! miT also be Indicated en 



ing^nd.' 



lark distance from end ti 



ng End," draw pen through words "Leaving End;" if nearest "Leaving End." 
Iraw pen throuah words ''Receiving End." ('Refers to track upon which the 
urrent of ttaffic is En one direction.) Indicale "Gage Side" on "Diagram" 






I "Damaged," describe ni 



ing pen through words "Gage Side" on 




lesTdc 




|. g « 1 


* m m m m m 


m 


OB 


IBB 



and cause if known. (See "Description of Failures" 



Approved: 
. . . Foreman. 

I, and Description of Failures, on Back. 



Fia. 345, q^ace of Fonn.) 

- Report of Rail Failures in Main Tracks: 



This is the basic report of all rail failures and is sent by the Track Foreman 
to his Supervisor and by him transmitted to the Division Engineer. It contains 
a clasfflfication of r^ failures which is used in the tabulations employed in t&e 
following blanks. 



INSTRUCTIONS 



-ward tbis Report direct to the Divirio 
r will have copies af Ihii Report mu 
:pj 10 the CbLef Engineer M. o£ W. 
raised characters oa the web of tl 



e day the break li 
the daj it ia taken 

ERgineer. 
: iiDiaediitelr upon 



D. The answer I 

E. The answer to 4 is "Bessemer" <I»; "Open-Heartb" (O.H.); "Nidwl" (N.); 

"Ferto-tilaniura (F.T.)i "Chrome Nicker CC.N.); or other method of iranu- 
factare or allof. 

F. The answers lo S and 6 are stamped into the metal on side of web— figures 

for 5 and a teller for 6. 

G. Mile Fast No. from 1 1^^^"* end of DiTJslon to be nsed. 

DESCRIPTION OP RAIL PAILURES 

When descrlMng Failures of Rails, Che following terms shoDld be used 




h *- Sn.TT Hud. This tern includes rails split throagh or near the center line 
P of the held, or rails with jHeces split off the side of the bead. When this 

bt term is used it sfaould be further defined by stating whether it i> or is not 



1 ± 



Sn.IT Wn. This term la a longitudinal split along the arts of the web. 



generally sUrtiog f, 



lil through the bolt holes. 



I ~ia 



Fio. 345 (continued). (Back of Form.) 



>, Google 



1 

1 

■s 

i 

^ J 
^ i 

i 


•"H'wi 




- 


« 


« 


- 


»■ 


«■ 


"-" 


■; 


7 


s 


3 


z 


a 


s 


s 


s 


s 


s 


2 a 


1 


1 J 

1 » 


"11^8 |o ""O 


iS 










































"!I«S"I 


s 










































ana 


i: 










































g 1'«'!14S 


S 








































E 


3 V^HWS 


K! 








































■:, 




Z 












































c 








































ii 


i i«vua 


a 








































1 


Ittlnna 


^ 










































jf ' 


£ 












































& 1 


a 








































1 i 


.-"w 


s 








































.J"'» 


s 












































.3|j 


a 










































1 I^ 


, H 


a 










































si-SS 


= 












































g I"™! 


3 












































U 1 


K 










































J ■ 


1 1 


S 










































? 1 


.(i>.»i 


s 










































f = 


".7 


- 








































? 


^^f,' 


« 


, 








































3 


^l^i 


^?l^ 


" 








































? 


^•U 


. .1 


- 








































1 




=is j 


„ 




















































































1 




1' 


" 








































j 


=sli 


'S 


■■ 










































iH 


li 


- 










































'l"l 



M. W. 405. — Superintendent's Monthly Report of Rail Failures in Main 

Tracks: 
On this blank the Division Engineer informs his Superintendent of the 
total number of rail fmlures for the month, tabulated from the Track "t'oreman's 



,y Google 



INSTRUCTIONS 

A. The DiTidoD EnclDMr will make out two coi^ei of tbi* report at the end of 

the month from the Section Foremen's Reports, and lend one copj to the 
Chief Eniineer M. of W. and one to the General SuperiiitendenL 

B. Mile Port No. frotn [ |^("' end of Diyinon to be niad. 

DESCRIPTION OF RAIL FAILURES 

When descrihing Failure! oE RaUa, the followinf lerms should be nied: 

1. BiOKEN R^iL. Thig tern Is to be coaGned to a rail which ii broken through, 
Bcparaling it into two or more parts. A crack which misht result in ■ complete 
break will come under this head. 



' 2. Flow or Metal. Thin term means a "RollinR Out" of the metal on 
of the head towards its sides without there heing any indication of ■ b 
ing down of the head ttruciure, thai is, the under side of Ihe bead it 



i. CIUSHED Head. This term is used 
and is usually accompanied by a 



"Flattening" of t. 



fi A. Sfltt Head. This term includes rails split through or near the center line 
id of the head, or rails with pieces split off the side of (he head. When this 

lb term is used it should be further defined by stating whether it is ot is not 



Fia. 346 (continued). (Back of Form.) 

report, and other officers who are interested, such as the Chief Engineer, Chief 
Engineer of Mmntenance of Way, or General Superintendent, are furnished with 
copies. In eases where a copy of the Track Foreman's report is sent to the Chief 
Engineer or Chief Engineer of Maintenance of Way, the monthly report serves 
as a check on the receipt of all individual rail reports. 



,y Google 







J. B. ^ C. R. R. Co. 




Stitemenl of Suel Rails nisting in Main Tnc 
M.in Track. 






December SI 19 


Location. 


i 
1 


Brand. 


« 


1 

■s 


I.«nelhofF«tinTn:k 


From Tg 


Uid Previous 

Kll9 


K.-S«,L*,d 


Steel Laid 
19 


Raiurla. 


U.P. 


.ft. 


M.P 


*f>. 












































































































































































































































































































































































































































































































































































1 


IcSlli 
























































.1 


*«. 
























































































































































































































































































■■ ■■ 100 
Column a lobtiind ler any (pedal nn.n 
and Rc-drOkd 

To be made out aod fonnided by the Bnii 
Engineer M, ol W., as loon »ft« the cloK o( It 


KhatR^-nlkdorSawed 

Comet: 
iw*rM,ofWtolheCiael 
»yeara>poDib)e. ■aita.a..<> 



FiQ. 347. 

M. W. 406. — Annual Statement of Steel Rails Existing in Main Tracks: 
This is an annual report sent by the Division Engineer to the Chief Engineer 
or Chief Engineer of Maintenance of Way, for the permanent record of the com- 
pany, to show the different kinds of steel in the main tracks at the end of tiie 
year. This may be used in conjunction with the rail chart, or take its place 
altogether, because the rail chart may not be in convenient form for a per- 
manent record, which may be referred to, after many years, for information 
concerning the kind of rail in use at a stated period. (sio) 



,y Google 





li 

1 i 
el 

oj . 

0! ■ 

y 1 
S 1 

|l! 

3 1 j 

u 


d 


ll 

i 


i 


1 M ' 


m • 


i^i\ 1- 


1 fi r 




5 


f.5 




1, 
1 I 


1 ^ 
1 


„ I 


I. t- 


1^ '' 




. 


II = 



>, Google 



Group III. Laboratory Examination of Special Raii^ 

This group is, at present, represented by the angle form shown in Rg. 348. 
It is used for making chedc analyses against Uie mill analyses and for reporting the 
result of chemical analyds and physical test of special rail or other test pieces which 
may be sent to the laboratory, from time to time, for examination. Fig. 349 shows 
standard locations of borings for chemical analyses and also the standard tensile 
test pieces of the association. 



FOR CHEMICAL ANALYSES 



FOR TENSILE TEST RECE8. 




IF RAIL IS FLiNBE WtfH, THE 

eoniMBS WO test piece from the upper 

PART Cr HEAD BHALL BE TAKEN FROM JUt 
OPPOerTE GWNER 



|^i*-~«f— ifrrj 



tfF 



Group IV. Compilation of Results for Study 

This group. Figs. 350-354, exhibits the different ways for compiling quan- 
titative statistics of rail failures. 

M. W. 408 (Fig. 350) is intended for compiling the information relative to 
rail failures for a period of one year. 

The columns for " specified chemical analysis " are intended for recording 
the analysis of the particular lot of rml as given in the specification, and is 
inserted in this blank in order to give an idea as to whether the rail is high or 
low in carbon, or high or low in phosphorus, etc. 

M. W. 409 (Jig. 351) has been provided on which the results from M. W. 
408 will be recorded at the end of the year, thus making a continuous 
record. 

(512) 



,y Google 



I 
I 

lisi 

<t;i 
m 





li 


j' li 


s : ! : ! t e s 8 = 




31 




1 




II 




f 




--.- — T- — J — J — ^ 




i 

i 

J 


1 1 






J ^ 




J T 




1- 

i 


5 T 






P 1 






1 I "^^T 




s 


n ii , 






i 






f 








5 T T 






; ^ 








J 




1 


r 




.lj.|j.|j.|j.|/.|i.|jWi|.|(.|i.tf 


-!«l-!3I-i«i-?.(-i-i 




■•»?• 1 1 






1 

! 
i 


1 






fi 






1 








• 






. 






j 






. 




. 


"1 






ih' 






J! 






-«n ---..--••s: 


H5-- = !r5SSS 



St iMfntiqi kfnD (pn ■• 



f»axfoi'^'"'vm'> 



m 



_^^ -Biflui Jnt""- )<>•»< «>^''*°>™p>™' JOB"!!' 



-worn suu to □httiu loi sKouoma 






jOOglc 



„_, -«A,«er.a<> 


a = a22sssasj 


?3 


___J 


t ,ll 


9 


U ,1 :.. 


1 


! h 


9 


i 3^ 




U n ; 




1 i] : 


:::::::::::::: ::i 


^ \ 11""" : " 




1 llii 




]il*l 




\ru 




1 ir;::::::::::::::: 


I 1 


M |J ...:.. 




i 3i 


j 


|j :. 


1 




I I I I I I I I I I I r ! 7[i , I I t ) 


.i.3-!;i.|.i<i;iss 


"I "I -l-i »^ *l*l ^*S -1 






1 - 




i ^ 




i 1- 




1 „ 




:l f -t 




1? T ,- 




J i"i 




?i 1 -I 








i] I "" 




1 ■= ^, 1 




s ^ ^ " 








! 




i p 




13 




,K,.l -««-«».-iBO 


s =22 -; s ^ A s ' 



-^ 



1 


e J 


II 


u 

i 

c 
i 1 , 

r IE 


II . - J, 


1 


1 


J T" a 


s t- 1 


J 

1 


1 f • t 


' %' 1 


1 
1 


i 1' ' 1 


1 

2 

J 


lii 

■si 
1' 


i %' '■ n 


1 


j ] ' ] 1 


i 1 " 1 1 


1 

1 


1 1 1' 1 1 


i 1" 1 


1 

■i 


- - - --i- -- ti --| 

3 I----J4-- I 


i : f 


1 


1 t: ! 


Z 

• 
> 

5 
5 

E 







111 • 1 


™;- 




p 





Fia.352. 

M. W. 410. — Comparative Number of Failures of Steel R^ls of Different 
Section or Pattern, Rolled by Different Steel Companies: 

In order to compare the product of different mills, and also to compare 
different weights per yard and different sections together, this blank has been 
provided. It cont^ns the totals taken from M. W. 408 or M. W. 409 as deared. 

(515) 





if 




■ 
S 


I 


1 . 


il 


il 


- 


i 

1, 




h 1 


1 


M 






a 


, 


"^ 


_ - <-- _ 


1 


""""!"]- 












^ 


1 s 


---|-1T---- !; 




jj ^ 


■j •" -^ 




""f]-~~" A 


w s 


___.'j-..__t 


Q \ 


___.._-. __JI 




- ' 1-i -«^ 


ai 


.i'^ 


^ 1 


____|_i>i___? 


Is 


"3 


1 


! >i il 


'- 
s 

i 

1 

1 


1 1 


1 "i II 


^ 




^ : 


M " |« 


1 


;:::r -ii::^] 




■* ^-i 




a ' 


" il- - i« 


" "d 


"---n - — i'A 


■ ::::::r:: 


___jj-____li 


L 






.1^ - -ill 


[ 


J s 


1 J" " 


r 


5 ; 


-,- - — -ly 


S 


= ' iSt 


■ 


sis 


a 1 


____,_. ^_„_iii 


o ;3 


- " '13 


" :::"""i:"" 


__.,.. _ 1^ 






< 


-"■■"■ °l5 











Fig. 353. 

M. W. 411. — Position in Ingot of Steel Rails which Failed: 
This is intended to furnish data on the number and character of raol ffulures 
according to the original position in the ingot held by the rail in question. 

(ei6) 



,y Google 



(COTW Ptgt tor Formi M. W. 406. 409. 410. 411.) 



A. B. & C. H. /i. Co. 



Offl« of CHIBP EK6INBER U. of W. 



M. W. 412: 

The information in this group should be bound together in one book; this 
cover has been provided for convenience and neatness. 

Group V. Progressive Wear of Special Rail under Observation 

In order to keep track of special rail, from time to time, and determine 
the value of the results being given, it is necessary to have a systematic plan 
of procedure for examinations and records. This group. Figs. 355-359, is fur- 
nished for that purpose, and is provided with a cover, as in the case of the 
previous group. (517) 



,y Google 









nil 

Pill! 

3* 1 












d 

^ 






* 






^ 












5 










t 

S 

. 1 




















.' 










s 


. 








f 










iiiili 






< 





M. W. 413. — Location Diagram: 

This blank is on a scale one inch equals one mile, and is intended fw ^' 
grams showing the location in different places of the same kind of rail under trial 

(SIS) 



,y Google 









m 



Fio. 336. 

M. W. 414. — Location Diagram: 

This is similar to M. W. 413 except that it is on a scale of two inches equal 
one mile, and is intended to show the location of a particular portion of the 
raU given in M. W. 413. It is made on a larger scale, so as to locate the points 
of measurement. A place U provided on each blanli for the summary of the 
wear or area abraded m percentage of total area of head. (sis) 



Joogle 




M. W. 415. — Diagram Showing Lines of Wear: 

The measurement of rail section at a spedfied point is shown on this blanlt 
and its position on M. W. 414 is given by the number in the circle of the blank 
at the top. All statistical information of interest and importance is given on 
the blank. (520) 



-vGoOgi- 







1 

1 


if 

il 


1 

1 


ii 


- 




1 

! 

I 

1 
i 
s 

1 

1 
1 

■8 

1 
1 


s 
« 

f 1) = 


!^ 


• 




1 


1 

^•5 


• 




;l 


• 




UodSM ;o M^P 


-• 






1. 




RT"C 


?. 




I 

1 

1 


% 




::::i;:" 






a. 




n 


t: 




(J 


^3 

e 

1 

u 


S 


- - 2- 6 -1 




1 


.._jA 




1 


B. 


.-.-l-i 




Cj 


5.1 




1 


i 




:::: if::::: 






|.? 




» 


""^ l-l 




« 






a 






a. 


5* 




■s 


u 


,* 




ami»]nu.|, 


f= 




.^»s^-^^ 


i- 




lu.Xj«[„»«„ 


£. 




p.rIIuOJ,Joo^ 


• 




1 


= 




il 


= 







M. W. 416. — Record of Comparative Wear of Special R^l: 
This blank is intended for compiling the information given in the previous 
ones, so as to ^ve a general summary of the results, (sai) 



.yGoogle 



(CoMT Pag* tor Fonni H.W. 413. (14, 4U. 4U.) 



A. B. & C. R. R. Co, 



RAIL SECTIONS 

Showing ProgTcssive Wear 



OBm of CHIBP RNGINEBR H. of W. 
(or other afflcer] 

Fio.359. 

M. W. 417: 

The information in this group should be bound together in one book; this 
cover has been provided for convenience and neatness. 

In some cases it may prove desirable to use charts or diagrams which illus- 
trate graphically the information found in the records. The diagrams of rail 
f^lures foimd in the Proceedings of the American Railway Engineering Asso- 
ciation are examples of this. Fig. 360 shows a method of recording the failures 
of different groups of rails during a period of years, and shows as well the dis- 
tribution of the failures during each year. It will be noted from the figure that 
the rails f^Ied, with a few exceptions, during the winter and spring months. 
A modification of this diagram can be obtained by making the record cumu- 
lative, which affords a ready comparison of the behavior of different rollings 
after having been in service any considerable length of time. 



,y Google 




; s I 






i \ 



oSoOO 



am 



,y Google 



Mg. 361 presents a further example and shows a diagram of rail failures < 
the Harriman Lines. 























































\ 


1 






















" 


•i 


1 






















' 


% 


«/ 


I 




















' 


> , 


/ 


1 




















' 


* 


/ 


1 






















« 


/ 


I 






















1 " 


1 


1 




















I " 




1 






















■J 


























1 
























' 






































































































































































































































































? ' 














^v- 










5 = 


, 


. 















Lg 




. J,. 



rAILl/^^ 0*f ALL I 



i ou/fiA/s yTA/r aoT. 



*mn90/6.r»,iiof/l> 



pMct during t-ht tt/^1- ma 



« >arv*r mtimimr «/ iramUt ttrli 



- Diagr&m of Rail FaUures, Harriman Li&ea. (Am. 1 



>, Google 




id to Area of Head 1^ 
J " " " Web 3.73 
2 " « « Ban 352 



Aim of Head 4.18 aq. in. 42% Batio Feripfaery of Head to Area of Head 1.66 

" " Web 2.06 " " 21% Web " " " Web 3.43 

" " Bam 3.83 " " 87% " " " Base " " " Base 3.06 

Total 9.82 sq. in. 100% RAtio Total Perii^ier; to Total Area 2M 

Mranent of Inertia, 43.8 
Section Modulus, Head, 14.44 
Section Modulus, Base, 16.11 



lociety of Civil Engineen (adopted 18SS). 



^, 



., Google. 



>, Google 




lEW YORK CENTRAL ft.R. 
Pura n.— Rail 



6I>'*' SANDBER& 187a 



>, Google 



>, Google 





1885 Recommended bt Committee to * 
m.c.b.as5.ak0 received majority of 
Votes but mot enough to make IT the 

MECOMMGNDED STANDARD 



"N 



1886 Presented TO M.M, ASS. 

BUT NOT ADOPTED 



,y Google 



>, Google 




■ Uig:izcd by 



Google 



>, Google 



Ratio POTphwy 



B sq. in. 100.0% lUUo Total Pen 

Moment <rf Inertia, 28^ 
Sectum Modulus, Head, 10.2' 
Sectkn Modulus, Base, 12.40 

1 Sail Section of the American Raihr^ 




.Google 



>, Google 



E 






(rfHd 
"Ba 
otal 



,y Google 



>, Google 




■ Web 3.58 
' Base 2.43 



Total 9.97 aq. in. 100% B»Uo Total Perq)ha7 to Total Area 

MMONit of Inertu, 41.9 
Section Moduhu, Head, 13.71 
Section Modulus, Base, 15.SI 

FtDiUiTlratu lUilroad Byetem (adopted 1907). 



,y Google 



>, Google 




FUtb X.—RailSeoUoos 



,y Google 



>, Google 




Geiman Raihntyi. 



-■■\ 



.Google' 



>, Google 




.Google 



>, Google 




^^^ 



□IL— L. A 




1 lai Inj r«MOT«l> 




vngji 54 Ibe. each. The ordinary or mtcnnedi&te chaiis are 7| itia. wide, and weigh 45 lbs. each. 



>, Google 



>, Google 




GALVMIItO SCDIW 



^ 



^ wrigji 64 &». eadi. The ordinary or intermedtftte ch^rs are 7| ins. wid^ and wei^ 45 lbs. each. 



>, Google 



>, Google 



.^2S£._. 




lie Ibfl. per yd. 

"B£." Section No. So. 

For use w cunree. 



.Google 



>, Google 



1 




^ 


■ 


^ — _^ 






15 18 

TIME IN SE 



' I>riTinK Wheel Spring, Consolidation Engi 



>, Google 



>, Google 



.Google 



>, Google 





»wra XXL— Freif^tLocomol 



.Google 



>, Google 



^ .., 



□ 
□ 



D 



taOnxd Tie Plate. 



>v Google 



>, Google 




4l 



\i\KK\\\KWmii 



:^^^m 



\\\\\\\\\\\w, yro^ 



-II 



li 

1% 



II 






III 
ill 

in 

S -si 

I H 

H 

|8t 

I 
li 



>, Google 



«y 



>, Google 




Google 



■yGOOglJ 




>, Google 



>, Google 



Dynnn^ Wbeel Loads for Various BaSa sad Axle ^Mtdng. 




:/ny Inci 



75 100 /25 

MKle Spacing Inches 



CUueCTnck, 



>, Google 



>, Google 



/ 


"Sr^tS,- 


^^p-sr*- 


' .^ 50 

46 
40 


18.0 
18.5 
IS.O 
13.5 
12.0 
10.5 
9.0 


360,000 
330,000 
300.000 
270,000 
240,000 
210,000 
180,000 


30 
25 



ding moments in different rails which 
"Till cause an extreme fiber stress of 
),000 lbs. per sq. in. in the base of the 
;ail. 



.ing In. 




I' SO 



.75 /OO /ZS 

Me spacing inches 



>, Google 



>, Google 



Platb XXVnL— Weight of Riul for Varioua Conditioiu of Loading and Ckaeea d Tfadc 
Class A Track. 



; 



1 

6 



Sn 



n 



L 



f 



I 



in 



-^^ 



Xx 



£0,000 3O.000 ^000 

static Axle load lbs. 

. K — Poor Wheel I^sBBUger Can on Steam Roodfl and Interaib&n Cam on Electric Roads, 






Dwizodb, Google J 



>, Google 




PL4T1 XXrX. — Plan oi 



,y Google 



>, Google 




3 



>, Google 



>, Google 



>, Google 



>, Google 













i 




1 




\^ 


h 


^ 








' 






































so- 
li™ 3a 


6* 
JtU. 


K)' 

Pappe.) 






>, Google 



>, Google I 






iffphu >te Atinf. Phot. Sit. W 



•E&TS AND INSPECTION. 



'shelff 
aiibeHsied 

:ef^asNo.l 






Avarfff* Anmlyti9 
of Httft Pmj*et»d 



Di»e»rJ«d 



Ifftcfef 



effh€ soma heat Ifhto eufeffhne ^fheaeaecond Msf pieeet 
breeK fhar«m9in^r^H»r»iboffhahesfmllglsob4reiechd 
tfh¥0outotfheeeffheseaeeoniittstpie^sdonotbnak,fh9 
remainder of Hie rails afffte/teaf mU be aceepfad, previ^thaf 
conform to fhe efhar raguiremertta of these apeeif/tattona,aal4ett 
or/itpL z cb$$ificgf/on. aeeord/nq as the deflection ia lass or/nor^ 
raspeeflvefy. tttm the pr^ scribed limit. ' 

ii) Ifsny Asff piaea, tasf'^dbaa not break; but mhen nicked and 
tested h destruction shoMs mteriordetact ttTst^ rails ftem 
esett ingot ^fftaftiOBtsttai/ be r^fseted ' 



ptacaa 

' second rm'ls 



(r\ 



>, Google 



>, Google 



Ar,, Ar,, ^r. points 426 

Aberdarc rails, effect of cold on 385 

Adriatic Railway tie 102, 104 

Affleck tie 104 

Algoma Steel Compaoy: 

raU mill 437. 438, 444 

shrinkage allowed at 444 

teeming practice 399 

Alloy steel (see Special steels). 

Alsace-Lorraine Railways, data on track 218 

Aluminum: 
effect of, in casting steel 401, 40S 

America {see alto American) : 

early steel rails used in 2 

length of rails used In 267 

production of rails in 382 

steel manufacturers of (see Association of 
American Steel Manufacturers). 

use of English rails in 2, 4 

American (see also America and United 
States) ; 

engines, examples of 32, 33, 34, 72 

joints, emmples of 264 

rail mills, practice at 399, 43S, 444 

specifications for rails 463 

speed of railway trains 21, 23,27,28 

Steel Manufacturers, Association of {tee 
Association of American Steel Manu- 
facturers). 
Bteel rails: 

examples of girder rails 19 

T-raila, early 6 

T-roils, present 10, 460 

prices of 326 

tie platen, examples of 122 

type engine (see Eight-wheel type engine). 
American cypress, resistance to pulling 

of spike 140 

American Electric Railway Engineering 

Association, rail sections 19 

American Forest Congress, ties 106 

American Institute of Mining Engineers: 

blowholes 404 

effect of recarbonizing 391 

Gayley dry blast 360 



American Institute of Mining Engineers: 

Lake Superior ores 351 

piping in ingots 401, 409 

relation between chemical composition and 

strength of rails, Dudley on 326 

roasting iron ores 344 

Sauveur on rail structure 429 

American Locomotive Company : 
examples of modem locomotives. 32, 33, 34, 72 

excess pressure of driving wheels 35 

American McKenna Process Company, 

rerolling rails 459 

American Railway Association: 

committee on standard rail and wheel sec- 
tions, report of 14 

discard, investigation of 417 

lettering rails from different parts of the 

ingot 416 

rail sections, principles governing design 

of 16 

proposed by ... : 14 

seriea A, description of 15 

failures of 10 

B, description of 15 

test to determine piping 417 

American Railway Engineering and Main- 
tenance of Way Association («ee 
American Railway Engineering As- 
sociation). 
American Railway Engineering Association: 

classification of defective rails 10 

comparison of rail failures of different 

sections 10 

distribution of pressure to subgrade 185 

drop testing machine, proposed by 290 

tests 293 

flow of rail h&d under wheel loads 205 

grain size in head of rail 432 

impact teats 64 

report on flat spots on car wheels 56, 61 

manufacture of rails, 

363, 374, 378, 439 

metal and composite ties 90 

scleroaeope 298 

reports and records proposed by 10, 501 

screw spikes 142 



,y Google 



American Railway EiigineerinB Association: 

size and spacing of ties 121 

special steels 333, 339 

specifications for rails 463 

Standard locations of borings for chemical 

analysts and tensile test pieces 512 

statistics of defective rails 10 

strei^h of head and web of rail 252 

tail steel 304 

stresses in the rail 210, 218 

t«Bts on flat spots in car wheels 61 

joint* 264 

strength of roil steel 304 

strength of wood 169 

tie plates 133, 170 

tree plantations for ties Ill 

unit stresses of different woods 16S 

American Railway Masto' Mechanics' 
Association: 

excess pressure of driving wheels 35 

tire wear 197 

American Society (or Testing Materials: 

bibliography of impact tests 288 

cupro-nickel steel 332 

drop tests 288 

ductility m rail steel 286, 287 

experiments on repeated Htress 278, 280 

finishing temperature 434 

hardness tests 298 

Howe, on welding blowholes 404 

ferrite grains 427 

impact tests 288 

influence of titanium on segregation 405 

manganese sulphide in steel 390 

manufacture of car wheels 57 

specifications for girder rails 4fll 

T-rails 463, 491 

stremmati^raph experiments 236 

Talbot, roUing practice in England 434 

tests on nickel steel 332, 334 

American Society of Civil Engineers: 

committee on rail sectioiw 6 

finishing temperature m rails 7 

rail section 6 

difficulty in rolling heavier rails. . . 7, 14, 435 

failures of 10 

necessity for rolling at Bigh tem- 
peratures 435 

strains produced in, by straightening. . . 445 

report on early steel rails 4 

specifications for rails, bibliography of.. . . 494 
American Society of Mechanical Engineers: 

Bessemer process 366 

electric locomotives 74 

■e of locomotive drivers 32, 62 

e tests of Professor Goes 62 



American Wood Preservers' Association, 

report on timber supply 113 

Amsler-LafFon machine for testing hardness.. 301 
Analysis: 

converter metal 374 

copper alloys used by Ball and Wingham. 331 

early Enghsh rails 326 

ferromanganese 374, 3S0 

ferrosiLcon 3S0 

ferrotitanium 405 

iron oree 363, 361 

location of standard borings for (American 

Railway Engineering Association) ... 512 

mixer metal 380 

specification for [see Specifications). 

steel, Bessemer II, 253, 310 

electric 385 

manganese 333 

nickel 333 

open hearth 310 

titanium 340, 406 

Angle bare (see Joints). 
Angle: 
fishing (»ee Joints). 

of friction of soils 311 

of incision in roUing 457 

of repose of various soils 314 

Ann Arbor Railroad, concrete ties on 105 

Annealed rails 20S 

Annealing, effect of, on special steels. . . 336, 337 

Anthracite coal in blast furnace 361 

Archduke Albrecht Steel Works at Carl- 

shQtte 3 

Arnold, J. O.: 

causes of rupture in steel 271 

infiuence of bismuth on copper 2TI 

Arsenic, effect of, in steel 330 

Articulated compound engine (aee Articu- 
lated engine). 
Articulated engine: 

classification of 29 

dimensions of 34, 72 

typical dynamic wheel loads 73 

weights of 31. 72 

Aah, physical properties of IW 

Association of American Sleel Manufacturers: 

chemical specifications for rails 34- 

drop testing machine 230 

specificatiora for rails 342, 463 

Association of Engineering Societies, Howard, 

on rail failures 203 

Ast^n, on cupro-nickel steel 332 

Atchison, Topeka and Santa Fe Eailway: 

effect of aubgrade on track 3'3 

rail section *fi' 

screw spikes on "' 



,y Google 



Atchison, Topeka and Santa Fe Railway; 

speed o( trains 25, 26, 27, 28 

Atlanta, Birmingham and Atlantic Railroad, 

ate and spacing of ties 121 

Atlantic City Railway, speed of trains, 

23, 25, 26, 28 

Atluttic Coast line Railroad: 

size and spacing of ties 121 

speed of trains 24 

Atlantic type engine: 

allowable asle loads 322 

classification of 29 

description of 21 

dimensions of 32, 72 

effect of excess balance and angularity of 

main rod 40,70 

pressure rounding curve 259 

rail stresses caused by 212 

speeds of 21, 24, 25, 26, 70 

strength of track required for 322 

typical dynamic wheel loads 72 

weight of rail required for 3^ 

weights of 32, 72 

Atlas, Barrow and Dowl&is, steel rails made 

by 2 

Austenite 427 

Australia, length of rails used in 267 

Austria, early steel rails 3 

Austrian State Railways, hardness tests 

on 302 

Axle loads: 

caia 85, 89 

dynamic (tee Dynamic). 

effect of siie of wheel on 202 

given in modem bridge specifications 211 

increase in 15, 30 

locomotives, electric 74, 79, 80 

steam 31, 32, 33, 34, 72 

maximum albwable, on rail 319 

used on Paris, Lyons and Mediterranean 

Railway 19 

Axle spacing: 

effect of, on load 247 

load diagrams of rail for different 319 

Back driver (»ee Driving wheels). 

Baggage car 84 

Bairstow and Stanton, experiments on re- 
peated atresB 281 

Baker, Benjamin, experiments on repeated 

stress 282 

Bald cypress, physical properties of 164, 168 

Baldwin Locomotive Works, ptf^rees in loco- 
motive building 30 

BaO, influence of copper on steel 331 

BaQ pressure tests 300 



bearing power of 179 

effect of dynamic load on 189 

CuSnot's experiments on 172 

depression of tie in, U. S. Government 

experiments 219, 222, 233 

depression of track m 172, 176, 189 

depth required for different classes of 

track 317 

distribution of load to subgrade 180, 185 

experiments on, in Germany 180 

frozen, effect of, on bending moment of 

raU 229,288 

frozen, tests on, by U. S. Government 225 

influence of kind of, on stresses in rail .... 229 

Pennsylvania Railroad testa 184 

weights of 316 

Baltic Sr (lee al»o Fir), force required to puU 

spike from 140 

Baltimore and Ohio Railroad: 

articulated compound engine 34 

chrome-nickel rails on 333 

electric locomotives 79 

impact tests on bridges 64 

ore docks 357 

rail section 461 

size and spacing of ties 121 

speed of trains 24, 28 

titanium rails on 339 

use of scleroscope on 298 

Bangor and Aroostook Railroad, size and 

spacing of ties 121 

Barnes, rail pressures of locomotive driving 

wheels 32 

Base rail: 

broken (see Broken base). 

indirect pressure used in rolling. . . , 457 

principles governing design of 16 

strength of st«el from 306 

of cars 85,89 

of electric locomotives 79, 80 

of steam locomotives 32, 33, 34, 72 

Basquin, O. H,, endurance tests 280 

Bavarian State Railways: 

arrangement of joints on 144 

daU on track 218 

spikes used on 144, 145 

tie plate 128, 132 

weight of rail on 144 

Bearing power (see Supporting poww). 

Beck, effect of cold on rails 285 

Beech: 

force required to pull spike from 140, 152 

ties, amount purchased in the United 

States 158 



,y Google 



Beech ties, coet of 156 

gn French Eaet«ni Railway .... 117 

Belgian State Railways, rail and tie plate, 125, 126 

Bell, Sir I. Lowthian: 

coet of transporting ore. 349 

depression of rails at different speeds 190 

early iron rails 1 



□ for different loading of rail 3^ 

rail, calculation of 2 

determined by Love 2 

effect of joint on 2< 

unequal tie pressure on . . 239, 21 

proposed solutions of 2 

stremmati^aph tests of 21 

U. S. government experiments 2 

tie 171, 177, r 

tie plat« 122, 1' 

Bending strength of wood: 

effect of moisture on li 

steaming on 1 

treatment on 1. 

in natural state 158, 164, 166, 1 

Bending tests on worn rails 3 

Benjamin, Professor C. H., flat spots on 

Berlin-iSoHaen line, speed of electric loco- 

motivee 

Bessemer, Sir Henry, origin of the Bessemer 



comparison of, with open hearth process. . 

description of plant 

early rails, in America 

invention of process 

process, description of 370, 

economies possible in 

steel, analysis of 11, 253, 

segregation in 

strength of 309, 

rails, amount in tracks of American 

railroads 

branding 

ductility of 

failures of 

hardness determined by sclero- 

prices of 

production of 

rolling tests on, at Watertown 

Arsenal 

specifications for (see Specifi- 
cations). 

Steel Works at Troy, early steel rails 

1 Lake Erie Railroad, steel 



Bethlehem Steel Company: 

chrome nickel steel rails 333 

raU mill 438, 4M 

shrinkage allowed at 444 

teeming practice 399 

use of iron with high copper 333 

Bibhography: 

of chemical composition 343 

of impact teste 28S 

of jointa 268 

of literature on steel manufacture (see 
Preface). 

of piping and segr^^tion 41S 

of rail specifications 494 

Birch ties: 

amount purchased in the United States. . . 156 
cost of 156 

Birkinbine, John, roasting iron ores 344 

Bittemut hickory, physical properties of.. . . 164 

Bituminous coal in blast furnace 361 

Black locust, phmting, for tie timber 109 

Bland, J. C, stresses in the rail 211 

Blast furnace: 

description of 357 

dry blast 360 

fuel used in 361 

location of plant 345 

operation of, at Gary 361 

Maryland Stoel Company 363 

performance of, with dry blast 361 

stoves 359 

typical examples of 357 

Bled ingote, specifications for 473 

Blessing, concrete tie 104 

Bloom; 
elimination of defecte in, early mill 

practice 434 

entering rolls 421 

size of 438 

Bloombg: 

passes, reduction m 437, 43S 

practice at American mills 43S 

Blowholes: 

Brinell's oxperimente on 402 

caused by iron oxide 3S9 

description of ■*(" 

effect of fluid compression on 4IG 

m titanium steel 406 

means for preventing 401 

Blum: 

report on speed of trains as affecting track . 2S 

stresses in the rail 21' 

Bolt, track (tee aUo Joints): 
holes, arrangement of (see Joints). 

specifications for (tee SpecificatioDS, 
drilling). 



>, Google 



Bolt, track, pull required to tighten 260 

strength of 260 

used on English buU head rail 19 

Bonzono joint 264 

Borings, location of standard, for chemical 
analysis, American Railway £ngineer- 

ing Association 512 

Boehia and Albany Railroad: 

angle bar t«ats 260 

early steel rails on 2 

track experimenta on, by U. S. Govern- 
ment 218, 225 

Boston and Lowell Railroad, granite ties 

on 90 

Boston and Maine Railroad: 

rail section 10 

faQurea 10 

size and spacing of ties 121 

teeta on locomotive springs 52 

Boston and Providence Railway, early steel 

rails on 2 

Box car 81 

Brakes, effect of application of, on locomo- 
tive springs ^ 

Branding: 

practice at different mills 447 

specifications for (see aUo Specification in 

question) 477 

Breuil, P., impact teats 292 

Bridge: 

effect of velocity of load on 69 

unpact on, caused by moving train 64 

specifications, impact allowed in 210 

limiting weights on axles.. . 211 
Brinell: 

ball pressure testa 300 

experiments on blowholes 402 

hardness test 300 

British Standards Committee; 

bull bead rails 19 

flat bottom rails 19 

rail sections 19 

specifications for bull head railway rails . . 4S4 

flat bottom railway rails. 488 

tramway rails 19 

Broken base: 

ctassiffcation of, American Railway Engi- 
neering Association 10 

due to defects in casting 390 

photographs of typical failures 11 

rail failures, six months ending April 30, 

1909 10 

Broken rails: 

caused by defective equipment 57 

classification of, by American Railway 

Engineering Association 10 



Broken rails: 

failures, six months ending April 30, 
1909 

photographs of typical failures 

Broken stone: 

ballast (see Ballast). 

weight of 3 

Brown hoist machine for unloading ore 3 

Brown, John, and Company: 

chemical composition of rails rolled by.. . . 3 

early steel rails 

rails rolled tor Ashbel Welsh 

Brown, J. P., on use of catalpa for ties 1 

Brunson concrete tie 1 

"B. S." rail sections (_ue British Standards 

Committee). 
Buffalo, Rochester and Pittsburgh Railway, 

sise and spacing of ties 1 

Buhrer: 

composition tie 

concrote tie 97, 1 

8t«el tie 

Bull head rail (see also British Standards 
Committee) : 

speciheations for 4 

types of 

Burbank tie 1 

Burden of blast furnace 363, 3 

Bureau of Forestry (see U. S. Forest Service). 
Bureau of Standards, magnetic testing of 

Burgess, on cupro-nickel steel 3 

Burlington (see Chicago, Burlington and 
Quincy Railroad), 

Calcination of iron ore 3 

Caledonian, Lancashire and Yorkshire Rail- 
way, early steel rails on 

Caledonian Railway: 

raU fastenings on 1 

speed of trains 

Camber, amount of, in rails 444, 477, 4 

Cambering machine 4 

Cambering, specifications for (see Specifi- 
cations). 
Cambria Steel Company: 

early steel rails 

handling ore at 3 

mixer used at 3 

oil tempered joints ! 

ores formerly used at 3 

railmill 438,4 

shrinkage allowed at 4 

teeming practice 3 

tests to detect pipes 4 

titanium steel rails S 



>, Google 



Camden and Amboy Railroad, first use of 

T-raU 6 

Camden and Atlantic Railroad, speed of 

trains 25 

Campbell, H.H.r 

experiments on nickel steel 335 

rolling 430 

Campbell tie 105 

Canada, length of raila used in 2€7 

Canadian Pacific Railway: 

indentation test 300 

rail section 461 

speed of trainB 24 

ten-wheel type engine 33 

Canadian Society of Civil Engineers: 

Dutcher on indentation test 300 

long ties used on Musk^ swamp 188 

Capillary attraction, movement of water by, 

insubgrade 314 

Carbon: 

content in rails: 
increase in following Dudley's experi- 
ments 328 

of Bessemer steel 11, 263, 310 

of open hearth steel 310 

of special steel (see Special steels) , 
specifications for (see ^»o Specifications), 

328, 342, 466 

effect of, in steel 329 

on ductility of rail steel 287 

iron diagram 427 

modification of, for low phosphorus 466 

segregation of 409 

Carbon dioxide, evolution of, in cooling 

steel 401 

Carbon-iron diagram 427 

Carijon monoxide, evolution of, in cool- 
ing steel 401 

Carbonic anhydride, compression of the ingot 

by means of 415 

Carnegie Library of Pittsburgh, bibliography 

of rail specifications 494 

Carnegie Steel Company: 

blast furnaces 361 

chemical composition of rails 32S 

dry blast at 361 

nickel steel rails 333 

rail miU 438, 444 

shrinkage allowed at 444 

steel ties in tracks at Duquesne Plant 94 

teeming practice 399 

titanium steel rails 340 

Carnegie steel tie 90 

Cars: 

allowable wheel loads of 322 

dynamic augment of wheel pressure — 85, 89 



Cars: 

strength of track required for 322 

weight of rail required for 322 

weights of 85, 89 

Cartesian coordinates, stress-repetition 

curves drawn to 281 

Cast iron: 

manufacture of 357 

wheel, effect of, on rail 195 

Casting: 

effect on steel 390, 395 

fluid compression during 410, 415 

the ingot 393 

Casting ladle: 

effect of agitation of steel in 390 

example of 3S9 

Catalpa: 

force required ta puQ spike from 139 

plantii^ for tie timber 109 

Cedar: 

physical properties of 164, 168 

ties, amount purchased in the United 

SUtcB 154,156 

cost of 154,156 

Cedar elm, physical properties of 164 

Cell {see Grain). 

Cementite 437 

Census, United States: 
distribution of lumber products in the 

United States 106 

ties purchased in the United States Ia6 

Center of gravity in locomotives 74, 259 

Central of Georgia Railway, size and spacing 

of ties 121 

Central Railroad of New Jersey: 

site and spacing of ties 121 

special steel rails on 333 

speed of trains 24, 28 

Centrifugal force: 

of counterbalance 3S 

of locomotive when rounding curve 2^ 

of wheel on irregular track 45 

Chair used with bull head rail 19 

Chanute, Octave; 

investigation of proper design of rail ^ 

report on steel rails 5 

Chapinore 394 

Charcoal in blast furnace 361 

Charge of blast furnace 363, 361 

Charpy, G., impact tests 292 

Chemical composition: 

bibliography of recent literature on ^ 

comparison of early and rec^it, in rails. . . 3% 

Dudley's formula 321 

effect on physical properties of steel (tee 
Element in question). 



>, Google 



Chemical composition: 
effect OB rail breakages in cold weather ... 2 
form for reporting, American Railway En- 
gineering Association 5 

increase in hardening constituents in rails . 3 
location of etandard borings for analysis, 
American Railway Engineering 

Association 5 

of Bessemer Bteel 11.253,3 

of different steels (see Analysis). 

of early rails 3 

of joints 264, 2 

of open hearth steel 3 

of special steels (aee Special steels). 
epeciScations for, in rails (see also Specifi- 
cation in question) 328, 342, 4 

Chemins de Fer d'Orl^ans (see Orleans Rail- 

Ch'emms de Fer de I'Est («« French Eastern 

Railway). 
Chemins de Fer de I'Etat {see State Railways 

of France). 
Chemins de Fer de I'Ouest («ee Western Rail- 
way of France). 
Chemins de Fer de Paris k Lyon et k la Medi- 
terran6e {see Paris, Lyons and Medi- 
terranean Railway). 
Chemins de Fer du Midi (see Southern Rail- 
way of France). 
Chemins de Fer du Nord (see Northern Rail- 
way of France). 

Chenoweth concrete tie I' 

Chesapeake and Ohio Railway articulated 

compound ei^ine 

Chestnut: 

force required to pull spike from i: 

ties, amount pm^jhased in the United 

States 154,1. 

cost of 154, 1 

Chicago and Alton Railroad: 

concrete ties on 90, li 

early steel rails on 

Chic^o and Eastern Illinois Railroad, size 

and spacing of ties l; 

Chicago and North Western Railway: 

concrete ties on II 

early steel rails on 2, 

new rail section 4i 

siie and placing of tjea t! 

speed of trtuns 24, 26, : 

ten-wheel type en^e I 

Chicago, Burlington and Quincy Railroad : 

concrete ties on II 

plantations tor tie timber i: 

prairie type engine ; 

rail section 10, 41 



Chicago, Burlington and Quincy Railroad : 

rail section failures 10 

size and spacing of ties 131 

speed of trains 24, 25, 26, 27 

track experiments 210 

Chicago Great Western Railroad, size and 

spacing of ties 121 

Chicago, Indiana and Southern Railroad: 

consolidation type engine 34 

size and spacing of ties 121 

Chicago, Indianapolis and Louisville Rail- 
way, size and spacing of ties 121 

Chicago Junction Rail\ray, concrete ties on.. 104 

Chicago, Milwaukee and Puget Sound Rail- 
way, size and spacing of ttea 121 

Chicago, Milwaukee and St. Paul Railway: 

excess pressure of driving wheels 36 

flat spot in engine wheel 56 

Pacific typo engine 32 

size and spacing of ties 121 

speed of trains 25 

Chicago, Rock Island and PaciSc Railway: 

Atlantic type engine 32 

early steel rails on 2 

size and spacing of ties 121 

Chicago, St. Paul, Minneapolis and Omaha 

Railway, size and spacing of ties. . . . 121 

Chrome-nickel steel raib: 

production of 341 

use of 333 

Chromium: 

effect of, in steel 334 

in combination with 

nickel 335 

steel rails, production of 341 

Cincinnati, Hamilton and Dayton Railway, 

size and spacing of tics 121 

Cinder ballast (see Ballast). 

City core 88 

Clamer, copper and nickel in steel 332 

Clary tie plate 122 

Classification: 

of locomotives 29 

of track 317 

Whyte's system 29 

Clausen, L. R., example of flat wheels 66 

Clay: 

ai^le of friction of 314 

bearit^ power of 187 

Cleveland, Cincinnati, Chicago and St. 
Louis Railway: 

plantations of tie timber 112 

size and spacing of ties 121 

Cleveland, Painesville, Ashtabula Railway, 

car used on 88 



>, Google 



Coal: 
can {tee idao Care], effect of heavy, on 

track 82, S 

used in blast furnace S 

Coefficient: 

of ballast I 

of friction of soils ! 

of slip in driving wbeela ISS, 1 

of yielding in ballaat on German railways, ^ 
Coes and Howard, experiments on rocking 

of engine 

Cogging mill 437, 4 

Cogging rolls used in Puppe's tests 448, 4 

Coke car 

Coke used in bkst ruraace 361, 363, i 

Colby, A. L., experiments on nickel steel , . , , S 
Cold: 

defective tires due to 

effect of, on strength of rail i 

roUing of head of rail caused by wheel 

rolling of rails 4 

shortness in steel S 

straightening press 4 

straightening rails 4 

work, effect of, on structure of steel . 424, 4 

CoUet trenail 148, ] 

Colorado and Southern Railway, siie and 

spacing of ties 1 

Colorado Fuel and Iron Company: 

rail mill 438,4 

shrinkage allowed at 4 

Columbia type engine, ctasBification of 

Composite ties used in CuSnot's experi- 
ments ] 

Compression: 

fluid, of ingot 410, 4 

modulus of, at point of contact of wheel 

and rail 194, 1 

cfTect of bearing surface on. , . 1 
strpuRth of special steels (see Steel in 

strength of steel 1 

effect of chemical composition on (see 
Element in question). 

strength of wood 158, 165, 167, 168, 1 

effect of treatment on ] 

Concrete ties: 

(see alto Tie in question) 97, 1 

service tesU on 1 

Cone pressure tests J 

Conical tires 

Connecticut River Railroad, early stw^l rails 



Consolidation type engine: 

classification of 29 

dimensions of 34, 72 

effect of excess pressure and angularity of 

main rod 44, 70 

pressure rounding curve 259 

rail stresses caused by 228 

speeds of 70 

strength of track requu^ for 322 

tests on driving wheel springs 52 

typical dynamic wheel loads 73 

U. S. Government experiments with 228 

weight of rail required for 322 

weights of 34, 72 

Continuous: 

girder, rail as 189, 240, 247 

joint 264 

process of malting steel 375 

rails 267 

record of rail failures, form (or reporting 

graphically 523 

Conversion of iron into steel 366 

Converter (see Bessemer). 

Converter metal, analysis of 374 

Cooling curve of different substances. . . 425, 426 

Coombs, R. D., on concrete ties 104 

Cooper, Hewitt and Company, first open 

hearth furnace in America 375 

cooling curve of 425 

effect of, in steel 331 

granular structure 270 

Cornwall irons, high copper in mils made 

from 332 

Corrugations of rails 209 

Cort, early methods of making steel . 366 

Coat: 
of Bessemer compared to open hearth 

proces 383 

of forest land 114 

of plantations tor tie timber Ill 

of rails, 1855 to 1910 325 

ferrotitanium 341 

rerolling 459 

of ties, annual charge of 115 

treated 115 

in the United SUtes 154,156 

of track of German railways 218 

with screw spikes U2 

of transporting ore 349, 352 

Codard, depreasion of ties in ballast.. . . 172, 190 
Counterbalance pressure: 

absence of, in electric locomotives 7S 

amount of, for different types of 

engines 35, 70 

calculation of M 



>, Google 



Counterbalance preaaure: 
effect of inertia of track on stresses pro- 
duced by 69 

on bridges 68 

tire wear 202 

Profess