LIBRARY UNIVERSITY OF CALIFORNIA. Class WORKS OF PROFESSOR GOSS PUBLISHED BY JOHN WILEY & SONS. Locomotive Performance. The Results of a Series of Researches Conducted by the Engineering Laboratory of Purdue Uni- versity. 8vo, xvi-J- 439 pages, 229 figures. Cloth, $5.00. Locomotive Sparks. 8vo, vii-f- 172 pages, 69 figures. Cloth, $2.00. LOCOMOTIVE PERFORMANCE THE RESULT OF A SERIES OF RESEARCHES CON- DUCTED BY THE ENGINEERING LABORATORY OF PURDUE UNIVERSITY BY WILLIAM F. M. GOSS, M.S., D.E. Director of the Engineering Laboratory arid Dean of the Schools of Engineering Purdue University FIRST EDITION FIRST THOUSAND or THF UNIVERSITY J ^.-.-' NEW YORK JOHN WILEY & SONS LONDON: CHAPMAN & HALL, LIMITED 1907 Copyright, 1906, BY WILLIAM F. M. GOSS ROBERT DRTTMMOND, PRINTER, NKW YORK PREFACE. FOR a number of years the Engineering Laboratory of Purdue University has concerned itself with problems relating to the performance of locomotives, and the results of its researches have from time to time appeared in the proceedings of various scientific and technical societies. This process of publication has extended over a period of fourteen years, and has run through many different channels, with the result that the record now exists in widely scat- tered parts, which are often difficult of access and, therefore, of lim- ited usefulness. The purpose of this volume is to combine the most important of these results with other material not before published, and thus make a permanent and accessible record of the work of the labora- tory. If it should appear that the pages have been burdened with too great an array of detail, it should be rememhered that many prefer such a minute presentation of facts as will enable them to work out conclusions for themselves. Primarily, the volume is de- signed as a record rather than a text, though it is hoped that it will prove of interest and value to any who wish to increase their acquaintance with the action of steam locomotives. As a whole, the researches are the outcome of many influences. The trustees and president of the University have supplied means, professors and instructors have assisted in working out the problems- of the laboratory, students have given their aid as observers, and in addition to these, skilled assistants have had a part in checking and arranging data. Much work and many plans necessarily preceded the actual work of the laboratory, and whatever may have been accomplished in the development 'of a system for testing locomotives is due, in the first instance, to the part which was taken by the late President Smart. It required some courage for the president of a university not rich in funds to respond to the suggestion of a department, and, iii 156495 IV PREFACE. in the absence of favorable advice, and even in the face of some adverse criticism, to give his approval and support to an untried process involving the expenditure of a considerable amount of money; but President Smart carefully considered every favorable plea and weighed every objection, and when he finally decided to enter upon a line of work so novel as that of testing loco- motives in a laboratory, he took a bold but nevertheless a well-con- sidered and intelligent step. He afterward imparted so much of his spirit of enthusiasm to the Board of Trustees as led them out of the limited resources at their disposal to supply means for executing his plan; and later still, when difficulties appeared in the development of mechanical matters, he was ever patient with delays, helpful in advice, and unfailing in his support. Another whose help counted for much was the late A. J. Pitkin while General Superintendent of the Schenectady Locomotive Works. In order to make the undertaking a success, it was necessary that funds at the disposal of the University should be supplemented by outside aid, and it was through the friendly influence of Mr. Pitkin that arrangements were finally made by which the Schenectady Locomotive Works agreed to supply a suitable locomotive in return for the small sum which was available. Without this cooperation the establishment of the testing-plant would have been greatly delayed, or it would have proceeded under conditions far less favor- able than those which afterward existed. Among others who early lent encouragement to the work of the testing-plant should be mentioned Professor James E. Denton, Mr. William Forsyth, and the late Mr. David L. Barnes. These mechanical engineers were among the first visitors who came to inspect the initial testing- plant, and their subsequent interest and frequent commendations went far to gain for the laboratory a degree of recognition which, in many incidental ways, has ever since been helpful. It is impos- sible to acknowledge properly the assistance of every one concerned in the actual operation of the laboratory, but mention should be made of Mr. Richard A. Smart, who for a considerable time was in immediate charge of the testing-plant, and of Messrs. Daniel Royse and Robert S. Miller, who each for a period of one year checked the numerical work of students. W. F. M. G. PURDUE UNIVERSITY, LAFAYETTE, INDIANA, March, 1906. CONTENTS. I. LOCOMOTIVE TESTING. CHAPTER I. THE DEVELOPMENT OF THE PURDUE TESTING PLANT. PAGE 1. The Growth of Engineering Laboratories at Purdue; 2. Considerations Lead- ing to a Locomotive Testing Plant; 3. Arrival of the Locomotive; 4. The First Testing Plant; 5, The Supporting Axles; 6. The Alden Friction Brakes; 7. Traction Dynamometer; 8. Behavior of the Mounting Mechanism of the First Plant; 9. The Work of the First Plant; 10. The Second Testing Plant; 11. The New Wheel Foundations; 12. The Emery Dynamometer; 13. The Superstructure; 14. The Building; 15. Work with the New Plant; 16. Sale of Locomotive Schenectady; 17. Schenectady No. 2 1 CHAPTER II. GROWTH OF INTEREST IN LABORATORY TESTS OF LOCOMOTIVES. 18. Locomotive Operation under Conditions Other than Those of the Track; 19. Growth of Interest in Locomotive Testing; 20. Interest in Purdue's Work; 2L New Plants 42 CHAPTER III. LOCOMOTIVE SCHENECTADY NO. 1. 22. Controlling Conditions Affecting the Choice of a Locomotive; 23. Specifi- cations; 24. Drawings; 25. Constants; 26. Steam Passages 46 CHAPTER IV. METHOD OF TESTING AND DATA. 27. Method of Testing; 28. Data.. . , 69 v vi CONTENTS. II. LOCOMOTIVE PERFORMANCE: A TYPICAL EXHIBIT. CHAPTER V. LOCOMOTIVE PERFORMANCE AS AFFECTED BY CHANGES IN SPEED AND CUT-OFF. PAGE 29. Purpose; 30. The Tests; 31. The Valves and Their Setting; 32. Indicator Cards; 33. Events of the Stroke; 34. Wire-drawing; 35. Mean Effective Pressure; 36. The Indicated Horse-power; 37. The Steam Consumption; 38. Critical Speed; 39. Cylinder Condensation; 40. Boiler Performance; 41. Performance of the Locomotive as a Whole; 42. Maximum Power De- pendent upon Efficiency 102 III. THE BOILER. CHAPTER VI. BOILER PERFORMANCE. 43. Selection of Data; 44. The Boiler; 45. General Conditions; 46. Actual Evaporation; 47. Quality of Steam and Equivalent Evaporation; 48. Power of Boiler; 49. Coal and Combustible; 50. Thermal Units; 51. Draft, Rate of Combustion, and Smoke-box Temperature; 52. Evaporative Per- formance; 53. Power and Efficiency; 54. Efficiency as Affected by the Quality of Fuel; 55. Derived Relations; 56. Conclusions 124 CHAPTER VII. HIGH RATES OF COMBUSTION AND BOILER EFFICIENCY. 57. General Statement; 58. The Tests and the Results; 59. Grate Losses; 60. Spark Losses as a Factor Affecting Grate Losses; 61. Losses Due to Incom- plete Combustion and Excess Air; 62. Losses along the Heating Surface; 63. Conclusions. . .156 CHAPTER VIII. THE EFFECT OF THICK FIRING ON BOILER PERFORMANCE. 64. The Conception Underlying these Tests; 65. The Tests and Their Results; 66. Interpretation of the Results; 67. The Influence of the Fireman 167 CHAPTER IX. SPARK LOSSES. 68. Sparks; 69. The Spark Trap; 70. Conditions affecting Tests; 71. Observed Weight of Sparks; 72. The Heating Value of the Sparks; 73. Volume of CONTENTS. vii PAGE Sparks as Dependent upon Quality of Fuel; 74. Refuse Caught in the Ash-pan; 75. Distribution of Sparks throughout the Stack; 76. The Size of Sparks; 77. Conclusion 173 CHAPTER X. RADIATION LOSSES. 78. The Amount of Heat Radiated; 79. Loss of Heat from a Locomotive Stand- ing in a Building; 80. Radiation Losses upon the Road; 81. Plan of the Tests; 82. The Experimental Boiler and its Equipment; 83. Observers; 84. The Track; 85. Movement during the Tests; 86, The Coverings Tested; 87. The Tests; 88. Standing Tests and Results; 89. Running Testa and Results; 90. Conditions Affecting Results for Which No Corrections Have Been Applie' 1 ; 91. A Summary of Results; 92. Efficiency of Coverings; 93. Radiation and Its Power and Coal Equivalent; 94. The Effect of Condi- tions Other than Those Which Prevailed during the Tests; 95. Conclusions. 185 CHAPTER XL THE FRONT END. 96. Definitions; 97. Draft and Its Distribution; 98. The Action of the Exhaust- jet; 99. Form and Character of the Jet; 100. The Jet as Affected by Changes in Speed of the Locomotive; 101. The Effect upon the Jet of Changes in the Height of the Bridge; 102. Jets Formed by a Steady Blast of Steam; 103. The Form of the Jet as Influenced by Different Tips; 104. The Form and Efficiency of the Jet as Affected by Bars over the Tip; 105. The Form of the Jet as Af- fected by Stack Proportions; 106. The Jet as Affected by Cut-off ; 107. The Stack Problem; 108. The Plan of the Tests; 109. Conditions at the Grate; 110. The Experimental Stacks and Nozzles; 111. The Tests; 112. Results; 113. A Basis of Comparison; 114. The Effect upon Stack Proportions of Changes in Speed and Cut-off; 115. A Review of Best Results; 116. Relation of Height to Diameter of Stack; 117. The Effect of Changes in the Height of the Exhaust Nozzle upon the Diameter of the Stack; 118. Equations Giving Stack Diameters for Any Height of Stack between the Limits of 26 Inches and 56 Inches, and Any Height of Nozzle between the Limits of 10 Inches below the Center of the Boiler and 20 Inches above the Center of the Boiler, and for Any Diameter of Front End; 119. Unavoidable Loss in Draft with Reduction in Height of Stack; 120. Relative Advantage of Straight and Tapered Stacks; 121. A Summary of Results; 122. Later Experiments ... 209 CHAPTER XII. SUPERHEATING IN THE SMOKE-BOX. 123. An Experimental Determination ,....,. . 262 viii CONTENTS. IV. THE ENGINES. CHAPTER XIII. INDICATOR WORK. PAGE 124. Concerning Indicator Work; 125. The Effect upon the Diagram of Long Pipe Connections for Steam-engine Indicators; 126. Experiments upon a Stationary Engine; 127. Different Lengths of Pipe; 128. The Form of the Cylinder Diagrams; 129. The Effect of the Pipe at Different Speeds; 130. The Effect of the Pipe at Different Cut-offs; 131. Conclusions 267 CHAPTER XIV. THE EFFECT OF LEAD UPON LOCOMOTIVE PERFORMANCE. 132. Lead; 133. Tests Involving Different Amounts of Lead; 134. Effect of Lead upon the Events of the Stroke; 135. Effect of Lead upon Valve Travel and Port Opening; 136. Effect of Lead upon the Form of Indicator Cards; 137. Steam Consumption; 138. Lead and Machine Friction; 139. Conclusions 282 CHAPTER XV. THE EFFECT UPON LOCOMOTIVE PERFORMANCE OF OUTSIDE LAP. 140. Outside Lap; 141. Events of Stroke; 142. Changes of Form in the Indi- cator Cards Resulting from Changes in Outside Lap; 143. Power Variation; 144. Steam Consumption 291 CHAPTER XVI. THE EFFECT UPON LOCOMOTIVE PERFORMANCE OF INSIDE CLEARANCE. 145. Inside Clearance; 146. Maximum Opening of Steam-port into Exhaust; 147. Changes in Events of Stroke Resulting from Inside Clearance; 148. Changes in the Form of the Indicator Cards Resulting from Inside Clear- ance; 149. The Blowing-through Effect; 150. Mean Effective Pressure; 151. Steam Consumption as Affected by Increased Clearance. -. 298 CHAPTER XVII. LOCOMOTIVE VALVE-GEARS. 152. The Function of a Valve-gear; 153. A Stephenson Valve-gear; 154. What the Stephenson Gear Does; 155. Devices for Increasing the Acceleration of the Valve; 156. Wire-drawing as a Factor Controlling Valve-gear Design; CONTENTS. IX PAGE 157. Improved Valve-gears; 158. Foreign Valve-gears; 159. Adaptability of Valve-gears; 160. The Conclusion 310 CHAPTER XVIII. ACTION OF THE COUNTERBALANCE. 161. The Problem of Balancing; 162. Experimental Method; 163. The Balance of the Locomotive; 164. Results; 165. Conclusions 321 CHAPTER XIX. MACHINE FRICTION. 166. A Statement of the Problem; 167. Methods; 168. Difficulties Encoun- tered in Measuring Draw-bar Stresses; 169. Friction Tests and Their Results; 170. A Comparison of Results; 171. Conclusions 333 V. LOCOMOTIVE PERFORMANCE. CHAPTER XX. THE EFFECT OF THROTTLING. 172. Throttling; 173. The Tests; 174. Indicator Cards; 175. Numerical Re- sults; 176. Steam Consumption; 177. Machine Friction 352 CHAPTER XXI. EFFECT OF HIGH STEAM PRESSURES ON LOCOMOTIVE PERFORMANCE. 178. Power and Efficiency; 179. Thermal Advantage of High Steam Pressures; 180. The Arguments for and against the Use of Higher Pressures; 181. Tests at Different Pressures; 182. Pressure versus Capacity; 183. Summary. 363 CHAPTER XXII. CONCERNING DIAMETER OF DRIVING-WHEELS. 184- Practice with Reference to Wheel Diameters; 185. A Study Based upon Observed Facts; 186. A Recapitulation 373 CHAPTER XXIII. ATMOSPHERIC RESISTANCE TO THE MOTION OF RAILWAY TRAINS. 187. Atmospheric Resistance; 188. The Plan of the Experiments; 189. Con- duit; 190. Air Supply; 191. The Determination of the Velocity of the Air CONTENTS. PAGE Currents; 192. The Model Cars; 193. Observations; 194. Observed and Calculated Results; 195. One Model; 196. Two Models; 197. Trains of Three, Five, Ten, and Twenty-five Models; 198. The First Model of a Train; 199. The Last Model of a Train; 200. The Second Model of a Train; 201. Models between the Second and Last of a Train; 202. Distribution of Forces Acting throughout the Length of the Model Train; 203. Relation of Force and Velocity; 204. A Summary of Conclusions; 205. Atmospheric Resistance to Actual Trains; 206. Application of Results Obtained from Models; 207. Resistance Offered to Locomotive and Tender; 208. Resist- ance Offered to Trains of Freight Cars; 209. Resistance Offered to Trains of Passenger Cars; 210. Resistance Offered to Any Train in Terms of Its Length; 211. Conclusions 377 CHAPTER XXIV. A GENERALIZATION CONCERNING LOCOMOTIVE PERFORMANCE. 212. Application of Data; 213. Boiler Performance; 214. Cylinder Perform- ance; 215. Draw- bar Pull; 216. Losses between Cylinder and Draw- bar; 217. The Application of Results to Several Typical Locomotives; 218. A Summary of Results 411 ILLUSTRATIONS. no. PAQB 1. The Purdue Locomotive, Schenectady No. 1 3 2. The Course Followed from Track to Laboratory 4 3. The Engineering Laboratory, 1891 5 4. Plan of the Engineering Laboratory 5 5. The Locomotive in the Laboratory, 1891-94 6 6. Elevation of Locomotive Mounting 8 7. Plan of Locomotive Mounting 10 8. The Exhaust-fan above the Engine 12 9. Supporting Axles 13 10. Alden Friction-brake 14 11. Arrangement for Oil Circulation 15 12. Friction-brake with One-half of Case Removed 16 13. Locomotive Dynamometer 19 14. The Dynamometer and ihe Mechanism for Controlling Pressure on Brakes . , 22 15. The Completed Engineering Laboratory prior to January 23, 1904 24 16. The Locomotive Testing Plant Immediately after the Fire 25 17. Elevation of the Second Testing Plant 26 18. Plan of the Second Testing Plant 28 19. The Dynamometer and the Mechanism for Controlling Pressure on Brakes, Second Plant. 29 20. Section of Building, Second Plant . . ; 32 21. Floor Plan, Second Plant 33 22. An Interior View, Second Plant. 35 23. 24. Locomotive Laboratory, 1894 37, 38 25. Plan of the Reconstructed Engineering Laboratory 38 26. The Departure of Schenectady No. 1 39 27. Schenectady No. 1 in a New Role 39 28. The Second Experimental Locomotive, Schenectady No. 2 40 29. Elevation of Schenectady No. 1 46 30. Stack 53 31. Exhaust-pipe and -nozzle 53 32. Dry Pipe 54 xi xii ILLUSTRATIONS. FIG. PAGE 33. Branch Pipe ; , 54 34. Netting and Deflector-plate 55 35. Boiler 56 36. Cylinder and Saddle 57 37. Valve-box and Cover 58 38. Cylinder-heads 59 39. Piston and Piston-rod 60 40. Cross-head 60 41. Guides and Guide-yoke 61 42. Valve .' 62 43. Valve-yoke 62 44. Steam-chest Valve-rod 62 45. Rocker 63 46. Link and Link Block 63 47. Eccentric Rod 64 48. Eccentric Strap 64 49. Eccentric 65 50. Reverse-shaft 65 51. Reverse-lever and Quadrant 66 52. Throttle -lever 66 53. Throttle and Throttle-pipe 67 54. Steam-passage Areas 68 55. Running Log 70 56. Feed-water Log 71 57. Fuel Log 72 58. Calorimeter Log 73 59. Indicator Record 74 60. Summary Sheet 75 61. Indicator- cards, showing Effect of Speed and Cut-off 104 62. Indicator-cards, showing Effect of Speed upon Size of Card 105 63. Diagram showing Mean Effective Pressure 108 64. Diagram showing Indicated Horse-power 109 65. 66. Diagram showing Steam per I.H.P. per Hour 110, 111 67. Diagram showing Percentage of Total Steam Accounted for by Indicator. . . 113 68. Diagram showing Percentage of Total Steam Accounted for by Indicator at Cut-off 114 69. Diagram showing Equivalent Evaporation per Hour 116 70. Diagram showing Evaporative Efficiency 117 71. Diagram showing Coal per I.H.P. per Hour 119 72. Diagram showing Coal per D.H.P. per Hour 120 73. Diagram showing Draw-bar Pull 120 74. Diagram showing Dynamometer Horse-power 121 75. Diagram showing Friction Horse-power 121 76. Diagram showing Coal per Mile-run 122 77. Chart from Recording-gauge showing Boiler Pressure 127 78. Chart from Recording-gauge showing Draft 139 79. Chart from Recording-gauge showing Back Pressure 140 ILLUSTRA TIONS. xiii FIG. 80. Diagram showing Coal per Foot of Grate Surface, as Related to Draft and Smoke-box Temperature .......................................... 141 1. Diagram showing Water per Foot of Heating-surface, as Related to Draft and Smoke-box Temperature ...................................... 142 82. Diagram showing Rate of Evaporation and Evaporative Efficiency ....... 144 83. Diagram showing Rate of Evaporation and Evaporative Efficiency as Obtained from Five Samples of Bituminous Coal .................... 150 84. Diagram showing Relation of Coal Burned and Water Evaporated ........ 152 85. Diagram showing Rates of Combustion and Evaporative Efficiency ....... 153 86. Grate used in Test 2 ................................................ 158 87. Grate used in Test 3 ................................................ 158 88. Grate used in Test 4 ................................................ 158 89. Diagram showing Evaporative Efficiency and Rate of Combustion. * ...... 164 90. Diagram showing Losses in Evaporative Efficiency ...................... 165 91. Diagram showing Evaporative Efficiency Obtained as a Result of Thick Firing as Compared with a Curve of Normal Efficiency ............... 170 92. Spark- trap ........................................................ 174 93. Plan of Stack, with Assumed Divisions Employed in Determining Spark-losses 175 94. Diagram showing Spark-losses .......................... . ............ 177 95. Diagram showing Heating Value of Sparks ............................. 179 96. Pounds of Sparks Passing out of Stack per Hour .............. . ......... 183 97. Pounds of Sparks Passing out through Areas Indicated .................. 183 98. Cross-section of Stack showing Density of Spark Discharge ............... 183 99. Sample Sparks ..................................................... 184 100. Head of Experimental Train ......................................... 187 101. Experimental Boiler ................................................ 189 102. Diagram showing Effect of Speed on Radiation Loss ................... 206 103. Diagram showing Distribution of Draft ................................ 210 104. Apparatus Used in Exploring the Exhaust- jet .......................... 213 105. The Form of the Jet ................................................ 216 106. A Diagrammatic Illustration of the Process of Envelopment by the Jet. . . . 217 107-109. Form and Character of the Jet as Affected by Changes in Speed of the Locomotive, Low-bridge Pipe. ... .............................. 219 110-112. Form and Character of the Jet as Affected by Changes in Speed of the Locomotive, High-bridge Pipe. . . ............................... 221 113-1 15. Form and Character of the Jet as Affected by a Steady Flow of Steam . . 222 116-118. Form and Character of the Jet as Affected by External Conditions. . . . 223 119. Three Forms of Exhaust-tips ......................................... 224 120. Bars or Bridges for Exhaust-tips ....................................... 225 121. The Fire-box as Prepared for Tests of Stacks ........................... 228 122. Dimensions of Stacks Employed in Experiments ........................ 229 123. Dimensions of Exhaust-nozzles Employed in Experiments ............... 229 124. Design of Exhaust-pipe and -nozzle ................................... 230 125. Diagram showing Draft Obtained from a Straight Stack 26" High and of Various Diameters ... ........................................... 237 126. Diagram showing Draft Obtained from a Straight Stack 36'' High and of Various Diameters .............. ......... ... 237 Xiv ILLUSTRATIONS. FIG - PAGE 127. Diagram showing Draft Obtained from a Straight Stack 46|" High and of Various Diameters 238 128. Diagram showing Draft Obtained from a Straight Stack 56|" High and of Various Diameters 238 129. Diagram showing Draft Obtained from a Tapered Stack 26" High and of Various Diameters 238 130. Diagram showing Draft Obtained from a Tapered Stack 36" High and of Various Diameters 238 131. Diagram showing Draft Obtained from a Tapered Stack 46" High and of Various Diameters 239 132. Diagram showing Draft Obtained from a Tapered Stack 56|" High and of Various Diameters 239 133. Diagram showing Combination of Stack Diameter and Nozzle for Best Results Obtainable from a Straight Stack 26" High 242 134. Diagram showing Combination of Stack Diameter and Nozzle for Best Re- sults Obtainable from a Straight Stack 36" High 242 135. Diagram showing Combination of Stack Diameter and Nozzle for Best Re- sults Obtainable from a Straight Stack 46|" High 242 136. Diagram showing Combination of Stack Diameter and Nozzle for Best Re- sults Obtainable from a Straight Stack 56" H'gh 242 137 Diagram showing Combination of Stack Diameter and Nozzle for Best Re- sults Obtainable from a Tapered Stack 26 \" High 243 138. Diagram showing Combination of Stack Diameter and Nozzle for Best Re- sults Obtainable from a Tapered Stack 36" High 243 139. Diagram showing Combination of Stack Diameter and Nozzle for Best Re- sults Obtainable from a Tapered Stack 46" High 243 140. Diagram showing Combination of Stack Diameter and Nozzle for Best Re- sults Obtainable from a Tapered Stack 56|" High 243 141. Diagram showing Relation of Height to Diameter of Stack for Best Results . 246 142. Proportions for Tapered Stack and Nozzle 250 143. Proportions for Straight Stack and Nozzle 251 144_147. Diagrams showing Relation of Draft to Height of Stack 252 148. The Best Arrangement of Front End assuming an Outside Tapered Stack 29" High 259 149. The Best Arrangement of Front End assuming an Inside Stack of Usual Form, Outside Projection 29" 259 150. The Best Arrangement of Front End assuming a False Top with Stack having an Outside Projection of 29" 259 151. The Best Arrangement of Front End assuming a Modified Form of Inside Stack 260 152. The Best Arrangement of Front End assuming an Inside Stack with Bell. . 260 153. The Best Arrangement of Front End assuming a Single Draft-pipe 260 154. The Best Arrangement of Front End assuming a Double Draft-pipe 260 155. A Proposed Standard Front End 261 156. Points Selected for Observing Degree of Superheating in Smoke-box 263 157. Thermometer Cup 264 158. Indicator Rigging 268 ILLUSTRATIONS. xv FIG. PAGE 159. Long and Short Pipe Connections for Locomotive Indicators 269 160. Typical Cards from Indicators Differently Connected with a Locomotive Cylinder 270 161. Arrangement of Pipes and Indicators on Buckeye Engine 272 162. Pipe and Cylinder Indicator- cards, Five-foot Pipe, 200 R.P.M 274 163. Pipe and Cylinder Indicator-cards, Ten-foot Pipe, 200 R.P.M 274 164. Pipe and Cylinder Indicator-cards, Fifteen-foot Pipe, 200 R.P.M. 274 165. The General Effect upon the Form of Indicator-cards of Very Long Pipes . . 275 166. Pipe and Cylinder Indicator-cards, Ten-foot Pipe, 100 R.P.M 278 167. Pipe and Cylinder Indicator-cards, Ten-foot Pipe, 200 R.P.M 278 168. Pipe and Cylinder Indicator-cards, Ten-foot Pipe, 300 R.P.M 278 169. Pipe and Cylinder Indicator-cards, Ten-foot Pipe, 200 R.P.M., -| Cut-off. . 280 1-70. Pipe and Cylinder Indicator-cards, Ten-foot Pipe, 200 R.P.M., J Cut-off. . 180 171. Pipe and Cylinder Indicator- cards, Ten-foot Pipe, 200 R.P.M., Cut-off. . ^80 172. Events of Stroke as Affected by Changes in Lead 285 173. Device for Measuring Valve Travel 286 174. Indicator-cards from Tests for which there were Changes in Lead 288 175. Indicator-cards showing the General Effect of Changes in Lead 289 176. Steam Consumption as Affected by Lead 289 177. Valves having Different Amounts of Outside Lap 292 178. 179. Valve Diagrams 294 180-182. Indicator-cards from Tests for which there were Changes in Outside Lap 295 183. Indicator-cards showing the General Effect of Outside Lap 296 184. Valve having Different Amounts of Inside Clearance 299 185. Diagram of Tests to Determine the Effect of Inside Clearance 300 186. 187. Indicator-cards from Tests for which there were Changes in Clearance 302, 303 188. Diagram showing Exhaust Interference with Excessive Clearance 304 189, 190. Diagram showing Inside Clearance and Steam Consumption 307, COS 191. Valve-motion Diagram 312 192. 'Indicator-cards representing Different Speeds and Cut-offs 314 193. Stationary Link 317 194. Allen Link 317 195. Joy Gear 318 196. Walschaert Gear 318 197. Arrangement of Guide-pipe and -wire for Counterbalance Experiments .... 32 j 198. Initial End of Test Wire 324 199. Diagram of Revolving and Reciprocating Parts 324 200-202. Plotted Measurements of Wires 325, 326 203, 204. Diagram having Reference to Position of Drivers on Supporting Wheels 336, 337 205. Tell-tale showing Position of Locomotive on the Mounting Mechanism 339 206. Method of Supporting Locomotive during Friction Tests 342 207. Diagram showing Draw-bar Pull Due to Engine Position 343 208. Diagram showing Tests Run under Throttle 353 209-212. Indicator-cards from Tests Run under Throttle 354-357 213. Diagram showing Steam Consumed when Running under Throttle 359 214. Diagram showing Steam Consumption under Different Pressures 366 xvi ILLUSTRATIONS. FIG. PAGE 215. Diagram showing Economy Resulting from Increase in the Capacity of Boiler. 369 216. Section of Conduit within which Model Trains were Exposed to the Action of Air-currents 379 217. Pitot's Tube 379 218. The Pitot Tube as Used in Experiments 380 219. A Portion of the Conduit with Attached Tubes 381 220. Diagram showing Velocities at Several Points in the Cross-section of the Conduit 382 221. The Model Dynamometer Car 383 222. Elevation of Model Train 385 223. A View \vithin the Conduit 386 224. The Influence of the Air-currents upon Different Portions of the Model Train v 396 225. Diagram showing Resistance of Different Portions of a Train at Different Speeds 397 226. Diagram showing Relation of Draw-bar Stress to Speed 416 227. Diagram showing Losses between Cylinders and Draw-bar 419 228. Diagram showing Absorption of Power between Cylinders and Draw-bar, and Traction Power 4 ' 21 229. Diagram showing the Performance at the Draw-bar of Several Typical Locomotives 422 A UNIVERSITY T f LOCOMOTIVE PERFORMANCE. I. LOCOMOTIVE TESTING. CHAPTER I. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 1. The Growth of Engineering Laboratories at Purdue.' Purdue University was opened as a school of Science, Agriculture, and Mechanic Arts in 1874. Five years later instruction was given students in shop-practice, and in 1882 a regular four-years course in Mechan- ical Engineering was established. From this time there was a gradual increase in the amount of apparatus available for the use of engineering students, but it was not until ten years after the establishment of the courses in shop-work that the development of an engineering laboratory was entered upon. In the spring of 1890 the policy with reference to the equipment of such a laboratory found expression in an order given for a cross- compound Corliss engine which during the summer of that year was established as a complete testing-plant with condenser, air- pump, weighing -tanks, and all other accessory apparatus needful for experimental work. A year later, in 1891, a considerable sum of money became available for use in the extension of laboratory facilities, and, with a view to future development, plans for an exten- sive building were adopted. Two important results followed: one consisting in the erection of a portion of the building which had been planned, and the other in the establishment of a locomotive testing-plant as a part of its equipment. 2. Considerations Leading to a Locomotive Testing-plant. The locomotive testing-plant was the outgrowth of natural condi- 2 LOCOMOTIVE PERFORMANCE, tions. The Trustees and the President of the University had already shown their interest in the development of laboratories for engi- neering students, and it was easy for them to foresee the great advan- tage to be derived from a study of locomotive performance under conditions as favorable to the work of testing as those which had been so often declared necessary in connection with stationary engines. It was known that under conditions of service locomotives were tested with difficulty, results being far less satisfactory than similar results obtained from stationary engines under service conditions. If, in stationary practice, progress had been made by the establish- ment of experimental plants, how much more might be achieved by the installation of experimental locomotives! These and similar arguments made it clear that the process of building up the equipment of an extensive engineering laboratory might easily and logically be made to involve the erection of a loco- motive testing-plant, the value of which both as a means to the instruction of students and for the purposes of research could not be doubted. The general outlines of the plant having been defined, its design and establishment became largely a matter of routine. The locomotive, stripped of the enchantment which may perhaps attend it on the road, became simply a steam-engine and a steam-boiler. Its action was not different in principle from that of other engines and boilers already included in the plans of the laboratory, and hence great experience in the management of locomotives, as such, was not necessary. In fact, the Purdue staff, which was instrumental in establishing the plant, had no member who had been trained in the motive-power department of a railroad. 3. Arrival of the Locomotive. In September, 1891, the loco- motive which had been named for its builders, "Schenectady," arrived upon a switch of the Lake Erie & Western Railway about one mile distant from the laboratory, and at a point having approxi- mately the same elevation with the University grounds. There was no track to the laboratory. The surface of the ground between it and the laboratory which was to be traversed was slightly rolling, and from a considerable portion of the intervening territory a wheat- crop had been taken eight or ten weeks previous to the arrival of the engine. The delivery of the engine upon the switch aroused great enthu- siasm on the part of the students of Purdue, and out of respect for THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 3 the interest shown a holiday was declared, and a call was made for volunteers to assist in receiving the engine and in starting it on its overland journey. It was rather late in the morning before the work began. The privilege of breaking a joint in the track, to assist in rolling the engine out upon the temporary skids which had been prepared, having been denied, it was found necessary to block and carry the flange of the wheels over the top of the switch-rails, a task which involves some difficulty even when managed by experienced men with plenty of tools, whereas in the case described both experi- ence and tools were lacking. But many willing hands, aided by a single team of horses, made light work of what might otherwise have been a laborious undertaking, and when night arrived " Schenectady " PIG. 1. -The Purdua Locomotive, Schenectady No. 1. had been pulled across the right-of-way ditch and well out toward the middle of the first field. By the close of the first day the measure of the task had been taken and time was therefore given to a reconstruction of equipment. Three sections of track were made, each of a rail's length, built in the form of skids and capped by 56-pound rails. The foundation of each consisted of two 5X12 yellow-pine pieces, laid flatwise, across which 2X12 pieces were spaced as ties, the rail-spikes passing through the ties and into the foundation beneath. The reorganized force included three pairs of horses with drivers and two or three men to handle blocking. One pair of horses was employed to give forward movement to the engine ; a second to draw the skids one after another from rear to front ; and the third to pull the heel of the advancing skid LOCOMOTIVE PERFORMANCE into line with the one previously placed. The men soon became so skilled in their respective parts that where the ground was smooth the engine could be kept in constant motion for considerable dis- tances, one skid being drawn from rear to front and placed in posi- tion while the engine was passing over the other two. Even with the new apparatus it was found impossible to make the locomotive follow the skids if laid on a curve, and w r herever a change of direction was necessary it was made by laying cross- blocking under the skids, upon which one end of the skid bearing the engine was slipped bodily. The course involved four turns, each somewhat less than a right angle, and the whole distance =^^B=MiK=3BBI^=^ p^= Sz^SS^^^^EjBc^ij^^^^S^^^^*"' Y" ~~ir ~ir Wheat Field* - 1 i , J t ' i WEST LAFAYETTE j i UN /VERSITY FARM i Clover Field I [Gravel Road 1 1 / \ Gravel Road N x 1 1 PURDO CAMPUS \ N ] First Laboratory 22 -^ I Present Laboratory!^-- -^- - -^ [ [ '" Rout i traveled first trip ion of present railway Loca FIG. 2. The Course followed from Track to Laboratory. traversed by the engine was in the neighborhood of \\ miles. On the eighth working day after the start, the engine arrived at the laboratory without accident and without having once touched the ground. * 4. The First Testing-plant. It has been already noted that before the locomotive was ordered, lines had been laid down for an extensive engineering laboratory, and the construction of a portion of this had been entered upon early in the spring of 1891. The building thus begun constituted a definite portion of the larger struc- * The success attending the transportation was chiefly due to the skill and energy of Mr. Robert Lackey, who had immediate charge of the operations after the first day's work. Mr. Lackey at the time was a student-assistant in the labora- tory. By his good judgment an undertaking which, if less perfectly managed, might have been an expensive one, cost but a trifling sum. O THF UNIVERSITY ] THE DEVELOPMENT OF THE PURDUE TESTING-PLANT 5 ture of which it was eventually to form a part. A view of this por- tion of the building is shown by Fig. 4 3, and the location of the testing-plant within it by the floor-plan, Fig. 4. FIG. 3. The Engineering Laboratory, 1891. Between the time of ordering the locomotive and its delivery, the details of the mounting mechanism were designed and put in place, so that when, in September, 1891, the locomotive arrived, the FIG. 4. Plan of the Engineering Laboratory, 1891. plant was practically ready for its reception. The following descrip- tion of this first locomotive testing-plant is based upon a paper presented before the American Society of Mechanical Engineers.* * "An Experimental Locomotive." Proceedings of the American Society of Mechanical Engineers, 1892. 6 LOCOMOTIVE PERFORMANCE The plan of mounting, in its inception, involved (1) support- ing wheels carried by axles running in fixed bearings, to receive the locomotive drivers and to turn with them; (2) brakes which would have sufficient capacity to absorb continuously the maximum power of the locomotive, and which should be mounted on the axles of the supporting wheels; and (3) a traction dynamometer of such form as would serve to indicate the horizontal moving force and at the same time allow but a slight horizontal motion of the engine on the supporting wheels. It was believed that a locomotive thus mounted could be run either ahead or aback under any desired load and at any speed; that while thus run, its performance could be deter- FIG. 5. The Locomotive in the Laboratory, 1891-94. mined with a degree of accuracy and completeness far exceeding that which it is possible to secure under ordinary conditions of the road; and that the whole apparatus would be extremely valuable to students in steam-engineering. It was not assumed that every condition of the track would be perfectly met, but it was expected that the results obtained would prove valuable in extending a knowledge of loco- motive performance. Fig. 5 from a photograph is a view of the locomotive in place upon the plant; Fig. 6 shows a complete general view, in elevation, of the locomotive and its mounting machinery; and Fig. 7 shows in plan the mounting machinery only. Reference to these figures will show that there was a heavy THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 7 rubble foundation, capped at convenient points with cut stones rising 10 inches above grade-line. Upon these stones were placed well- seasoned oak timbers arranged in two lines, each composed of three lengths, 4X14 inches. The timbers of each line were well bolted to each other and were securely anchored to the foundation. Upon the timbers rested the bearings of the supporting axles, which were thus given 14 inches of oak to constitute an element of elasticity between them and the foundation. The supporting wheels were of the same diameter with the loco- motive drivers, and similar in other respects save that the cranks and counterweights were omitted. Their faces were turned flat, with the inside edge rounded as in a rail. The four friction-brakes which provided the load for the sup- porting shafts, and which are shown in position in Figs. 6 and 7, were designed on the principle developed by Professor George I. Alden and already described by him.* The details of the brake design under consideration will be given farther on, but it is im- portant to state here that the principle as developed by Professor Alden provided extensive rubbing surfaces of cast iron and copper. Excessive wear was prevented by thorough lubrication. The inten- sity of the brake action was controlled by water pressure, by which means the rubbing surfaces were brought into contact more or less intimate, and the heat evolved was carried off by water circulation. By reference to the plan and elevation, Figs. 6 and 7, it will be seen that there was no provision for measuring the load at the brakes, where, instead of a weighted lever, anchor-rods were used to secure the case of the brakes to the foundation. The value of the load ap- peared at the dynamometer connected with the draw.-bar of the loco- motive. The water supply for the brakes was furnished by a 3-inch pipe. It passed first a balanced valve A, around which there was a by-pass controlled by valve B. From the tee C the pipe was branched for the several brakes, 2^-inch piping serving for two brakes, and 2-inch for each individual brake. Valves, D, were provided in the supply-pipe for each brake, so that any one might be entirely cut out or have its action modified to any desired extent. The water from each brake was returned by a separate pipe to a point E, where valves were provided by which the amount of water allowed to pass each brake was regulated. From these valves the water * Transactions of the American Society of Mechanical Engineers, Vol. XI, p. 959 et seq. . LOCOMOTIVE PERFORMANCE. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 9 flowed in an open stream and was finally discharged into a sewer. The water pressure within each brake was indicated by one of the fcur gauges at F, Fig. 6. The balanced valve A had its spindle connected with one of the levers of the dynamometer in such a way that its position was con- trolled by the pull, or if backing, by the push exerted by the loco- motive. Thus, suppose the locomotive to be in motion and the outlet-valves at E adjusted to allow the passage of enough water to keep down the temperature of the brakes; suppose also that the pull of the locomotive were such as to bring the weighted lever of the dynamometer to its mid-position, then there would be a defi- nite opening of the balanced valve, and a definite water pressure within the brakes would result. If now, for any reason, the weighted lever should fall, there would be a corresponding increase in the opening of the balanced valve and, hence, an increase of water pressure within the brakes. The greater pressure would result in greater resistance to the movement of the supporting wheels and, hence, in a stronger pull of the locomotive on the dynamometer, and this increased pull would tend to lift the weighted lever again. Similarly, if for any reason the pull of the locomotive were sufficient to raise the lever beyond its central position, the balanced valve would respond by reducing the water pressure within the brakes; the tractive force of the engine would then decrease and the dyna- mometer lever would fall. When, therefore it was desired to increase the load on the locomotive it was necessary only to place the addi- tional weight upon the lever of the dynamometer, and the corre- sponding increase in the load was furnished automatically by the brakes. By a proper adjustment of the by-pass valve, B, the lever could be made to stand exactly in its central position. 'The traction dynamometer was made up of a system of levers. The first lever in the system, shown by dotted outline at G in Figs. 6 and 7, had direct connection with the locomotive draw-bar. The last lever, shown at H, carried an ordinary weight-holder. The whole arrangement was such that, whether the engine moved ahead or aback, the stress was transmitted by the draw-bar, and its value shown by the weight necessary to balance the lever H. The dynamometer levers were carried by a heavy framework, which was well secured to the locomotive foundation and to sur- rounding parts of the building. The character of the framing is but imperfectly shown by the drawings. 10 LOCOMOTIVE PERFORMANCE. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 11 While the draw-bar was the only active agent by which the horizontal movement of the locomotive was controlled, there was ample provision of chains and buffers to check any excessive move- ment which might occur. Above the levers of the dynamometer a floor was laid which chiefly served the purposes of a tender. It gave room for the stor- age of a limited quantity of coal, and for a tank from which the locomotive injectors drew their supply. Connected with the water- tank was a glass gauge, 7, and above the tank a weighing-barrel through which the tank received its supply. Scales for weighing fuel were also given a place on the " tender floor." The three counters at J were connected, respectively, with the rear driving-axle and with each* of the two supporting shafts. These gave a ready means for determining the speed of the engine, and the per cent of slip between the drivers and their supporting wheels. The telltale at K showed the position of the locomotive relative to the supporting wheels. The board a was fastened to the loco- motive and consequently moved with it; the rod b was connected at one end, c, to an iron column as a fixed point, and at the other end to the pointer d. This pointer was pivoted to the board a at e, so that any backward or forward movement of the locomotive was greatly multiplied in the similar movement of the lower end of the pointer d. A tangent-wheel and screw were provided at L for the purpose of turning by hand the forward supporting axle and hence the engine, whenever it might be desired to do so, as, for example, for convenience in valve-setting. When not in use, the screw could be disengaged. The truck-wheels of the engine rested upon light rails which were fixed at the level of the laboratory floor and extended in front of the engine a distance sufficient to allow the whole machine to be moved forward off the supporting wheels, whenever the latter needed . to be take out for repairs. A Sturtevant 4JX6J steam-blower, located above the engine (Figs. 6 and 8) but not in pipe connection with it, removed from the room everything given out by the locomotive stack, without chang- ing, materially, the draft conditions under which the locomotive worked. The cylinder-cocks and the overflow-pipes from the injectors 12 LOCOMOTIVE PERFORMANCE. were all in loose connection with the sewer. The discharge from the overflow-pipes could be directed into weighing-barrels. The boiler was in pipe connection with the fixed boiler which supplied steam for general use in the laboratory, so that the loco- motive might be used to supply steam to other apparatus, or the fixed boiler to supply the locomotive. In the latter case the loco- motive boiler was drained by the steam-trap M; by this means the boiler was freed from water, and as a consequence all its parts were kept at the same temperature. The greater convenience attending Stack. FIG. 8. The Blower above the Engine. the operation of the fixed boiler made its use desirable when problems were studied which affected only the mechanism of the engine. The following is a description of some of the more important parts making up the plant, the details of which are not clearly shown by the drawings thus far referred to. 5. The Supporting Axles were of hammered iron and of the form and dimensions shown by Fig. 9. The bearings for these shafts were 8" in diameter and 16" long. Each bearing was fitted to a cast-iron plate 14"X36"X2", which plate, in turn, was carefully bedded upon the oak timbers (Figs. 6 and 7). The whole was made secure by four bolts which passed through the bearing, plate, and timber. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 13 6. The Alden Friction-brakes, which supplied the load to the sup- porting axles and, hence, to the locomotive itself, are shown in position by Figs. 6 and 7, and in detail by Fig. 10. They were designed and constructed with the consent of Professor Alden, to whom the author is indebted for many courtesies. The cast-iron moving disk K, Fig. 10, Sec. AB, is keyed to the supporting axle (not shown), and hence turns with it. The power of the locomotive is transmitted by the supporting wheels and axle to the disk, on either side of which are light copper plates, L, which form a part of the enclosing case. The sides M of this case have Section of rim at Supporting vheel. Of wheel 03 inches d li < hub of imeter _ j c ! > mm radius of round corner %* ItoM of Wheel BOM Of ^1 eel 7* > 1 I 1 2 Brake fit journal Wlee - fit smooth hammered cot turned sn-allest Clj. ti be EJt less than 7^* WheeJ Journal * Brake Ht | | ^ 5fiIJ_'^_ -,JLg For wormwhcel on une shaft > not occupied on other FIG. 9. Supporting Axle;. bearing on the hub of the moving disk, and are connected around their outer edges by the distance-ring N, and by through-bolts (Sec. CD). The copper plates already referred to are clamped at their outer edges between the sides of the case and the distance-ring N. At their inner edges the joint is made by means of wrought-iron rings, secured by closely spaced machine-screws tapped into the case. The copper plates are thus held very near to the moving disk, but they are not necessarily in absolute contact with it. Between the copper plates and the sides of the case are annular spaces within which water may be circulated. It will be seen that the case with its copper plates is quite free to revolve upon the hub of the cast-iron disk ; or if, as in practice, the case is at rest, the disk may revolve freely within it. It will be seen, also, that water under pressure in the annular spaces will force the light copper plates against the moving cast-iron disk; that, as a result of this contact, there will be a tendency on the part of the case to turn with the cast-iron disk, and that if this tendency is overcome by an outside resistance, power will be absorbed. This, in brief, is the action of the brakes; the case, in the present instance, is prevented from turn- ing by anchor-rods connecting it with the foundation (Fig. 10). LOCOMOTIVE PERFORMANCE. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 15 The circulating water enters at the opening Q (Fig. 10), passes from the annular space on this side of the case to that on the other side by eighteen f-inch holes through the copper plates and the distance-ring (Sec. EF), and is finally discharged at the opening R } sufficient water being allowed to pass to carry away the heat result- ing from the friction. To prevent the water pressure within the brake from spreading the sides of the case, the rings S are fitted over a feather to the hub of the moving disk, and are held to their place by nuts which screw up to a shoulder. When the brake is in use, all clearance space between the cop- per plates and about the moving cast-iron disk is filled with oil. The distribution of the oil is secured by thirty-two radial grooves on each face of the cast-iron disk (Fig. 10), and by a spiral groove ARRANGEMENT FOR OIL. CIRCULATION FIG. 11. extending from the inner edge to the cirqumference of the rubbing surfaces of the disk, with a pitch of about 4 inches. The entire rubbing area of the cast-iron disk is thus split up into surfaces the length or breadth of which in no case greatly exceeds 4 inches. The spiral oilway gives these surfaces such positions that in passing a given point on the copper plate they continually change their align- ment, which fact, it is believed, greatly assists in the distribution of the oil. The radial grooves give rise to a slight pumping action, by means of which the oil may be kept in circulation between the center and circumference of the brake. Provision for this circula- tion is made as shown by Fig. 1 1 ; the oil passes the valve and piping a from the highest point in the brake, is received by the filter-can. 16 LOCOMOTIVE PERFORMANCE and is thence delivered to the center of the brake. The filter-can also serves the purpose of a supply reservoir, and always contains surplus oil when the brake is in action. This circulation helps to maintain the oil at a uniform temperature, subjects all parts to the same service, and gives a ready means for detecting any defect which may arise in the lubrication of the brake. One of the brakes with one-half of the case removed, exposing one of the rubbing surfaces of the moving disk, is shown by Fig. 2. As preliminary to the design of these brakes, the load which they would need to carry was determined as follows: The locomo- tive drivers were 63 inches in diameter, and the sum of the moments about the two driving-axles when the engine was exerting a tractive force of 16,000 pounds (assumed 63 to be maximum) was therefore 16,000 = 42,000 foot-pounds, which, since the support- ing wheels were of the same diameter with the drivers, was the moment under which the brakes on the supporting axles must work. Other considerations led to an early decision FlG '- A. 2 '7T A ? n 1 c , tioi }~ b ^ ake to use two brakes on each axle, making it neces- with One-half of Case Removed. sary, therefore, for each brake to act under a maximum moment of 10,500 foot-pounds. Before fixing upon the dimensions of the proposed brakes, a rather extensive series of experiments was made upon a small Alden brake then in use in the laboratory, for the purpose of determining the probable value of the coefficient of friction which would attend the action of such a brake. These preliminary experiments were made on a 21-inch disk-brake while driven at speeds varying from 300 to 450 revolutions per minute. From these experiments it appeared that lubrication could be maintained with certainty under a water pressure of 40 pounds per square inch, and this pressure was adopted as the maximum to be used in the design of the larger brakes herein described. It also appeared that the apparent coefficient of friction * varied from 2.7% to over 4%, depending largely on the * By apparent coefficient of friction is meant that factor which is obtained by assuming that the entire moment of the brake is due to the friction between the rubbing surfaces of the moving disk and the copper plates. Since there are other rubbing surfaces, it is clear that the apparent coefficient of friction is larger than the actual coefficient. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 17 viscosity of the oil used and upon the temperature of the brake. It was thought that 3.5% would be a safe coefficient for moderate speeds, and this factor was accordingly used. The moment in foot-pounds required to revolve one disk upon another against the action of friction, when the two are bearing face to face, is the product of the area in contact, the pressure of contact per unit area, the coefficient of friction, and two-thirds the radius of the disk in feet. Representing this statement by an equa- tion, we write ., where M = moment of force in foot-pounds; A = area in contact in square inches -*; p = pressure in pounds per unit area; / = the coefficient of friction ; r = the radius of the circular plate in inches. But each brake of the design in question (Fig. 10) consisted of a revolving disk with a fixed copper plate on either side, thus giving two rubbing surfaces, and for such a brake the moment will be (1) Again, the copper plates forming the rubbing surfaces were not com- plete disks, but rings. The moment necessary to revolve such a disk against such a ring will be the same as the moment necessary to revolve a disk of a radius equal to the outer radius of the ring, less the moment necessary to revolve a disk of a radius equal to the inner radius of the ring. Now if ri = the outer radius of the ring in inches, 7*2 = the inner radius of the ring in inches, we may write, by the aid of equation (1), for the moment of the ring in foot-pounds, (2) 18 LOCOMOTIVE PERFORMANCE. I "o 3 I THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 19 As already stated, previous experiments with a trial brake had shown that a coefficient of friction of 0.035 could be depended upon. The maximum water pressure assumed for the design was 40 pounds per square inch, and trial solutions which were carried out as the design was in process of development made r\ equal to 28 inches and T2 equal to 10 inches. Substituting these values in equation (2) gives M = -X40X0.035X (21952- 1000) = 10234, which is to be compared with the required moment of 10,500 foot- pounds. It is evident that the best available information indicated that the dimensions chosen were such as to make the brakes well suited to the requirements of their anticipated service. 7. Traction Dynamometer is shown by Fig. 13. The outline plan and elevation in this figure show the relative position of the different levers when in place. The details of each lever are shown by the remainder of the plate. The main lever A being first referred to, the following description will be seen to apply. This lever was supported by a round steel pin which connected it with the cast plate E, and which in turn was securely bolted to the heavy timbers which formed a part of the framework behind the locomotive. The round rod F was an extension of the locomotive draw-bar, and the pull, or push, of the locomotive was exerted along the line of its axis. Stress was transmitted from this rod to the blocks / by nuts as shown, the fit between the nuts and the blocks being spherical (detailed view, Fig. 13) to allow slight changes in the direction of F. The blocks / were connected with the short arm of the lever A by the round steel pins g, and the long arm of this lever was engaged by the hook I, which was free to move within the link k. To illustrate the action of this part of the dynamometer, let it now be assumed that the pull of the locomotive on the draw-bar F is ahead, that is, in the direction of the upper arrow; then the stress on the draw-bar will result in a tendency to raise the hook I, and the lever B with its weight will serve only as a counterbalance to the lever A. But if it be assumed that the locomotive is working aback, that is, that the stress in the draw-bar is in the direction of the lower arrow, the long arm of the lever A with the hook / will tend to fall. Motion of I in this direc- tion, however, is soon arrested by the link k, which by virtue of its 20 LOCOMOTIVE PERFORMANCE. connection with the lever B rises as I falls, until it engages the check- nuts i] these are thus made a means of transmitting the stress to the hook I, which as before will move upward, while the shackle at the upper end of the hook I is entirely free from stress. It will thus be seen that, whether the locomotive is assumed to be working in forward or in backward gear, its tractive force is made manifest by an upward movement of the hook Z, from which point two simple levers complete the dynamometer. The direction of the lever C, which connected with hook Z, was at right angles with that of the lever A, its purpose being to bring the last lever D of the system out from behind the locomotive. The lever D carried a weight-holder arranged to receive twenty 10-lb. weights. The ratio of the whole system was as 1 to 100; each weight, therefore, that was balanced at the weight-holder represented a tractive force exerted by the locomotive of 1,000 Ibs. A small dash-pot (not shown) was attached to the last lever, D. The lever C had depending from it a light rod, which controlled a balanced valve B (Figs. 6 and 7) in the pipes upplying the brakes with water. By means of this valve, as previously described, the load on the brakes was made to vary automatically with the position of the last lever of the dynamometer. The locomotive while in motion could be given a final adjust- ment to its place on the supporting wheels by means of the nuts on the draw-bar. Having now examined the more important principles underlying the design of the first plant, we may make inquiry concerning its behavior in service. 8. Behavior of the Mounting Mechanism of the First Plant. Any considerable arrangement of machinery which is new in the details of its design, or which is required to serve purposes which are new, must for a time be considered experimental. The mounting mechanism of the locomotive testing-plant was novel in many of its details, and the conditions surrounding its use were new. Each element entering into the make-up of the plant, therefore, demanded in its turn its full share of patiently-bestowed and long-continued attention. The most important element was the friction-brakes by which the power of the locomotive upon the mount was absorbed. These, as already stated, had been designed upon a principle which had been previously developed by Professor Alden, but no brakes had before been made approaching in capacity those of the Purdue plant. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 21 Their ability to perform the work for which they were intended was early settled beyond question. When they were first used, the oil became very hot about the outside of the case, while that in other parts of the apparatus remained cool. To avoid this accumulation of heat around the outside of the case, closed circulating pipes were added to convey the hot oil from the outer portions of the case back to the center, the radial curves in the disks being sufficient to maintain a pumping action by which there was a continuous flow of oil from the outside to the center of the brake. These pipes took the place of the open circulation provided for in the original design. They served, also, to maintain all portions of the brake at an approxi- mately uniform temperature, the excess of heat being carried away by the circulating water as already described. The brakes were lubricated by a cheap grade of cylinder-oil, which was supplied by the cans shown by Fig. 11. There was always some leakage at the center, but no more than was desirable to insure lubrication of that point. The drip was caught and returned to the can. In the original design of the brakes it was assumed that the work absorbed would be proportional to the water pressure exerted upon the copper plates, an assumption which would be true provided the lubricant between the copper plates and the cast-iron moving disk was always in the same condition. In practice it appears that the work is all done on the oil and any change of condition resulting in a change in the amount of work absorbed produces a correspond- ing effect upon the temperature and, hence, upon the viscosity of the oil. Other things being equal, therefore, increasing or diminish- ing the pressure does not result in a proportional increase or dimi- nution of the amount of work absorbed. The effect of this obser- vation is expressed by saying that the amount of work absorbed by the brakes depends quite as much upon the temperature of the oil in the brakes as upon the pressure exerted between the copper plate and the moving disk; and in practice, excepting when the speed of revolution is very small, it is found best to use a moderate water pressure, and to vary the load by varying the brake tempera- ture, this being easily accomplished by controlling the amount of circulating water. The water pressure under ordinary conditions of service rarely exceeds 10 pounds. It has been stated that the work absorbed by the brakes is prac- tically limited to that which is represented by the force necessary to overcome the viscosity of the oil. Evidence of the truth of this 22 LOCOMOTIVE PERFORMANCE. statement is to be found in the fact that the bearing pressure between the copper plate and the moving disk, being seldom over 10 pounds per square inch, causes no perceptible wear of the rubbing surfaces so long as lubrication is maintained. The copper plates and the cast-iron disk which constitute the rubbing surfaces of the four brakes, after a total of six million revolutions show wear in only in spots on the copper plates where, owing to imperfections in the surface, small areas have received a concentration of pressure. Since the work absorbed by the brakes depends largely upon the temperature of the oil which lubricates their rubbing surfaces, it will appear that the FIG. 14. The Dynamometer, Mechanism for Controlling Pressure on Brakes, and the Revolution Counters. First Plant 1891. maximum load will be greatest when the heat developed between the rubbing surfaces is most quickly conducted away. The copper plates in the brakes under consideration were made -& of an inch in thickness. It is probable that less cooling water would be required and the action of the brakes would be improved if the thickness of these plates were considerably reduced. They can perhaps be reduced until only a sufficient amount of metal remains to withstand the shearing forces to which the action of the brake subjects them. The dynamometer to which the draw-bar of the engine was con- nected (Fig. 13), while reasonably satisfactory, was not well cal- culated to absorb the heavy vibrations which at high speed the loco- motive brought upon it. The structure behind the locomotive was cr THF ^\ UN1VERSITY THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 23 several times strengthened. First, the woodwork upon which the dynamometer was mounted was reenforced, the anchorage behind it increased, and a small dash-pot added to its last lever. Later a larger dash-pot having a diameter of 12 inches was interposed between the fixed tender-frame behind the locomotive and a bracket attached to the locomotive foot-plate. This, being entirely indepen- dent of the dynamometer, served to absorb the larger part of the vibrating forces incident to high speed, and after its addition no difficulty was experienced either in holding the engine or in measuring approximately its pull. The dynamometer was not, however, a piece of apparatus whose accuracy could be relied upon under all condi- tions of service. In view of the enormous force which a locomotive is capable of exerting it would appear, at first sight, that an error of 50 or even 100 pounds in the determination of draw-bar stresses would be of slight consequence, and that great accuracy in this matter would not be required. Under some conditions this conclusion is correct, but under others it is far from true. The work done at the draw-bar is the product of the force exerted multiplied by the space passed over. If the force exerted is great and the speed low, a small error in the draw-bar stress is not a matter of great importance, but if the reverse conditions exist if the force is small and the speed is high then it is absolutely necessary that the draw-bar stresses be determined with great accuracy. Moreover, high speeds necessarily involve low draw-bar stresses. A locomotive which at ten miles an hour may pull 12,000 pounds will have difficulty, when running at sixty miles an hour, in maintaining a pull of 2,500 pounds. When, therefore, in the general progress of events, the opportunity came to improve the character of the dynamometer, considerations such as have been stated were deemed of sufficient importance to justify the purchase of the most accurate apparatus which could be found. The exhaust-fan over the engine (Fig. 8) was a detail made necessary by the character of the room in which the locomotive was located and the variety of equipment by which it was surrounded. It never failed to serve its purpose, and when steam was being raised it was invaluable. A minor difficulty encountered in the course of the early work upon the plant was that due to the tendency of oil used upon the bearings of the supporting axles to follow the arms of the supporting wheels out to the rim and thence to reach the locomotive drivers. 24 LOCOMOTIVE PERFORMANCE. This oil by spreading over the faces of the supporting wheels and drivers reduced adhesion and greatly interfered with the action of the plant. A remedy was found in the attachment of a ring to the face of the supporting-wheel hub of such form as would receive and retain by centrifugal force all drip occurring during a run. Still other minor changes will be found worked out as accomplished facts in Purdue's second plant, a description of which is to follow. FIG. 15. The Engineering Laboratory, Purdue University, prior to Jan. 23, 1 94. 9. The Work of the First Plant. During the school year of 1891-2, following the installation of the plant, work upon it was directed more to the perfection of mechanical features than to the acquistion of scientific data. Nevertheless, during this year twenty efficiency tests were run, many of them at light power and almost all with the throttle only partially open, and later an experimental investigation concerning the action of the counterbalance was under- taken. It was not until the fall of 1893 that the plant was in such com- plete working order that tests at high power could be run with cer- tainty. During the latter part of this year, however, a consider- able number of such tests were carried out and much was expected from an early study of the results, but no essential fact was ever established from these data. On January 23, 1894, the Engineering Laboratory, which by a succession of additions had increased greatly in size, was burned. All experimental data which had not been published were lost, and the locomotive went down in the wreck. The fire entailed a heavy burden of labor and expense. Tut with THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 25 the new responsibilities which it brought there came also new oppor- tunities. All details of the mounting mechanism were most care- fully reviewed, and every fragment of experience was made to serve a useful purpose in the design of a new plant. A permanent track 8,000 feet in length was laid to connect the laboratory with the rail- ways of the country. The damaged locomotive was extricated from the ruin, sent out over the track and thence to Indianapolis for repairs. Upon its return a few weeks later, it was put under its own steam and backed in over the Purdue track directly to its place upon the supporting wheels of a new testing-plant. The ease and rapidity FIG. 16. The LocDmative-testing Plant Immediately after the Fire. with which this trip was made were in striking contrast with the laborious methods which attended its first trip across the same terri- tory. Four months after the fire the new work had been completed and the reconstructed engine was in position. 10. The Second Testing-plant. The new plant, as shown by the accompanying plates, occupies a building especially planned to receive it, and is arranged for the accommodation of any locomotive. The dotted outline in Fig. 17 is to the scale of the University's loco- motive Schenectady. 11. The New Wheel Foundation. By reference to Figs. 17 and 18 it will be seen that there is provided a wheel foundation of nearly twenty-five feet in length. This is more than sufficient to include the driving-wheel base of any standard eight-, ten-, or twelve-wheeled 26 LOCOMOTIVE PERFORMANCE. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 27 engine. For engines having six wheels coupled, a third supporting axle will be added to those shown, and for engines having eight wheels coupled four new axles, having wheels of smaller diameter than those shown, will be used. The wheel foundation carries cast-iron bed-plates, to which are secured pedestals for the support of the axle-boxes. The lower flanges of the pedestals are slotted and the bed-plates have threaded holes spaced along their length. By these means the pedestals may be adjusted to any position along the length of the foundation. The boxes in use at present are plain, babbitted shaft-bearings, and between each bearing and its pedestal a wooden cushion is inserted. A bearing has been designed, for use in some special experiments which provides for the suspension of the axle from the springs, but has not yet been used. The outer edges of the wheel foundations are topped by tim- bers to which the brake-cases are anchored. The brakes which absorb the power of the engine are those which were used in the original plant, and which have been described. The fire caused no serious damage to this portion of the apparatus. 12. The Emery Dynamometer. The vibrating character of the stresses to be measured makes the design of the traction dynamom- eter a matter of some difficulty. The dynamometer of the original plant consisted of an inexpensive system of levers attached to a heavy framework of wood, the vibrations being controlled by dash- pots. In the present construction wood as a support is entirely abandoned and a massive brick pier, well stayed with iron rods, has been substituted. The dynamometer connects with the draw- bar at the rear of the locomotive. It consists of the weighing-head of an Emery testing-machine, the hydraulic support of which is capa- ble not only of transmitting the stress it receives, but also of with- standing the rapid vibrations which the draw-bar transmits to it. The apparatus is of 30,000 pounds capacity, and at the same time is so sensitive that one standing in front of the locomotive may press with his fingers upon its pilot and cause a deflection of the needle of the dynamometer. As is well known, the arrangement of the hydraulic support of the Emery testing-machine permits the weighing-scale to be at any convenient distance from the point where the stresses are received. Figs. 17 and 18 show only the receiving end of the apparatus. The draw-bar connects with this apparatus by a ball-joint, which leaves 28 LOCOMO Tl VE PERFORMANCE. I I 1 1 THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 31 its outer end free to respond to the movement of the locomotive on its springs. A threaded sleeve allows the draw-bar to be lengthened or shortened for a final adjustment of the locomotive to its position upon the supporting wheels; and, finally, to meet the proportions of different locomotives, provision is made for a vertical adjust- ment of the entire head of the machine upon its frame. 13. The Superstructure. Figs. 20 and 21 show the arrangement of floors. The "visitors' floor" (Fig. 21) and the fixed floors adjoin ing are at the level of the rail. The open space over the wheel foun- dation is of such dimensions as will easily accommodate an engine having a long driving-wheel base, movable or temporary floors being used to fill in about each different engine, as may be found convenient. The temporary flooring shown is that employed for the Purdue loco- motive. The level of the " tender floor" is at a sufficient height above the rail to serve as a platform from which to fire. At the rear is a runway leading to the coal-room, the floor of which is somewhat lower than the tender floor. A platform scale is set flush with the floor at the head of the runway. During tests the scale is used for weigh- ing the coal which is delivered to the fireman. The feed-water tank, from which the injectors draw their supply, is shown in the lower right-hand corner of Fig. 21. Above this supply-tank are two small calibrated tanks so arranged that one may be filled while the other is discharging. The steam-pump shown on the visitors' floor is for the purpose of supplying water under pressure to the friction-brakes which load the engine. The conditions under which the engine is operated are at all times within the control of a single person, whose place is just at the right of the steps leading to the tender floor. From this posi- tion he can see the throttle and reverse-lever and observe all that goes on in the cab. At his right is the dynamometer scale-case, wherein is shown the load at the draw-bar; in front are the gauges giving the water pressure on the brakes; and under his hand are the valves controlling the circulation of water through the brakes. No attempt has been made in these drawings to show small acces- sory apparatus, neither does it seem necessary to give an enumera- tion. 14, The Building. Fig. 23 presents several views of the loco- motive building. The entrance-door, which opens upon the visi- LOCOMOTIVE PERFORMANCE. 34 LOCOMOTIVE PERFORMANCE. tors' floor, is shown in the south elevation. It is approached from the general laboratory, 150 feet away. The north and west elevations show the roof construction by which the upper end of the locomotive stack is made to stand out- side of the building. The roof sections shown may be entirely removed, and a door in the cross-wall, which extends between the removable roof and the main roof, provides ample height for the admission of the locomotive to the building. A window in this door (Fig. 20) serves to give the fireman a clear view of the top of the locomotive stack from his place in the cab, a condition which is essential to good work in firing. Above the stack is a pipe to convey the smoke clear of the building. To meet a change in the location of the stack this pipe may be removed to any position along the length of the removable roof. An illustration from a photograph of the interior, taken when the plant was in operation, is given as Fig. 22. In connection with the plan of the building (Fig. 23) there is shown the arrangement of tracks for the locomotive and of those used for supplying coal. Fig. 24 is an exterior view of the locomotive labora- tory when the plant is working, and Fig. 25 gives the position of the locomotive or annex laboratory, relative to the whole group of buildings of which it forms a part. 15. Work with the New Plant began in the fall of 1894 and con- tinued without interruption for more than two years. In the course of this time most of the results were secured which are to be dis- cussed in the succeeding chapters. Fifty or more efficiency tests of boiler and engine, under various conditions of speed, load, cut-off, steam pressure, and valve proportion, were run. A large amount of time was spent in determining the amount of work lost in machinery friction under different conditions of running. A study was made of the effect of high rates of combustion upon boiler efficiency, of the conditions affecting the draft action produced by the exhaust, of steam distribution by plain and Allan-ported valves, and of the power necessary to move balanced and unbalanced valves. In the course of this work, and of that done upon the first plant, not less than 10,000 indicator-cards were taken, 20,000 miles were run, and 25,000 different observations made. From these data about 12,000 derived results had been obtained, making a total of nearly 50,000 facts which were then available to disclose the performance of this experimental locomotive. It is perhaps safe to say that there was nowhere such an accumulation of locomotive data obtained THE DEVELOPMENT OF THE PURDUE TESTING-PLANT 37 LOCOMOTIVE PERFORMANCE. FIG. 24. The Locomotive Laboratory, 1894. LOCOMOTIVE I COAL LABORATORY | WOOD ROOM ENGINEERING LABORATORYU MACHINE ROOM LOCKERS ! II LOCKERS FIG. 25. A Plan of the Reconstructed Engineering Laboratory, Purdue University, 1898. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 39 under conditions so favorable to experimentation, and arranged so systematically, as the material then existing in the Purdue laboratory. FIG. 26. The Departure of Schenectady No. 1, FIG. 27. Schenectady No. 1 in a New Role. 16. Sale of Locomotive " Schenectady." The six years which fol- lowed the introduction of Schenectady into the Purdue laboratory 40 LOCOMOTIVE PERFORMANCE. had been marked by unusual progress in locomotive design, and by the end of that period the experimental engine had ceased to Le representative of the most approved practice. For example, steam pressures of 180 or 200 pounds per square inch were then common, while the boiler of Schenectady had been designed for a maximum pressure of 140 pounds. Early in 1897, therefore, it was decided that locomotive Schenectady should be disposed of and another engine which would better serve the purposes of the FIG. 28. The Second Experimental Locomotive Schenectady No. 2, 1897. laboratory secured to take its place. It thus happened that on a bright morning in May, 1897, Schenectady, hereafter to be known as Schenectady No. 1, steamed out of its " rectangular round-house " and slowly made its way across the campus under the escort of a throng of students. At the Art Hall a stop was made and the sig- nificance of the event was emphasized by brief addresses.* Passing from Purdue's possession, Schenectady became the' property of the Schenectady Locomotive Works, and later entered regular service on the Michigan Central Railway. Our last glimpse * The speakers were President Smart, Colonel H. G. Prout, and two members of the faculty. THE DEVELOPMENT OF THE PURDUE TESTING-PLANT. 41 of the machine shows it at the head of a train on this road, where its identity is established by its number, 422.* 17. Schenectady No. 2 arrived at the laboratory, fresh from its builder's hands, in October, 1897, and entered at once upon the work for which it was designed. Since, however, the greater part of -the data discussed in succeeding chapters is the result of the work done on Schenectady No. 1, a description of Schenectady No. 2 wiU for the present be omitted. * The photograph from which Fig. 27 has been made was obtained through the courtesy of the late Mr. Robert Miller, Superintendent of Motive Power and Equipment of the Michigan Central Railway. In transmitting the photograph Mr. Miller wrote, under date of Dec. 11, 1897: *' I take pleasure in forwarding you to-day ... a photograph of our engine, No. 422, known to us as the Purdue Locomotive. Our men call her 'the schoolmarm. ' " CHAPTER II. * GROWTH OF INTEREST IN LABORATORY TESTS OF LOCOMOTIVES. 18. Locomotive Operation under Conditions other than those of the Track. A testing-plant was not required to demonstrate the practicability of operating a locomotive with the machine as a whole at rest, for, previous to the existence of such a plant, locomo- tives had been made to serve the purposes of stationary engines by being blocked up until their drivers cleared the rails enough to carry belts; * and before the days of the injector it sometimes hap- pened that an engine held for an unusual period on a siding was in an emergency pumped without changing its position. This was accomplished by setting a jack under the rear of the engine in such a manner as to take a considerable portion of the weight of the engine from the drivers, which, thus relieved of pressure and aided by the application of oil to the rail, could be slipped at moderate speed through the action of its cylinder while the pumps delivered feed- water to the boiler. Nor has the student of locomotive perform- ance failed to take advantage of such possibilities as these. Thus, the late Alexander Borodin, when Engineer-in-Chief of the Russian Southwestern Railway, presented the results of an elaborate series of tests made upon a small locomotive blocked clear of the track, with belts applied to the drivers in such a way that the power developed could be absorbed in driving shop machinery, f But in neither of these cases was it possible to operate the locomotive under con- ditions normal to the track. The Purdue plant was the first to receive a normal engine in the condition in which it might be as it * A locomotive thus arranged was seen driving the machinery of a railway repair- shop in Montgomery, Ala., in the winter of 1866. t "Experiments on the Steam- jacketing and Compounding of Locomotives in Russia,'' by Alexander Borodin, Mem. Inst. M. E. Proceedings of the Meetings of the Institution of Mechanical Engineers, London, August, 1886. 42 GROWTH OF INTEREST IN LABORATORY TESTS. 43 came from the road, and to allow such an engine to be loaded and run in a perfectly normal way.* 19. Growth of Interest in Locomotive-testing. The year 1891, which marked the installation of the Purdue locomotive testing- plant, and those immediately succeeding, represent a period of unu- sual interest in locomotive design and performance. The preceding twenty-five years had witnessed the development of the vast rail- way systems of this country, new engines had been constantly in demand for the operation of new track, and shop facilities for repair- ing them were required to be constantly increased. When there was an end to the period of enormous track extension, the condi- tions affecting locomotive development became more settled; rail- way managers began to enforce economy in operation, and attention was thus turned from matters of construction to questions of per- formance. As a consequence, much activity was developed in test- ing locomotives while in operation on the road, but there was no standard method of making such tests; and as each investigator fol- lowed his own method, a comparison of results proved to be of little value. These conditions led naturally to a desire for greater uni- formity in practice, which, in the spring of 1890, found expression at a meeting of the American Society of Mechanical Engineers, which resolved "That a Committee of Seven be appointed by the Chair to report on Standard Methods of Conducting Tests of Efficiency of Locomotives, including the engine, the boiler, the quality of the steam, and the comparative efficiencies of simple and compound locomotives." f The debate upon the resolution there adopted revealed the limited and indefinite character of the information available concerning the performance of locomotives, and gave expression to a desire for fuller knowledge upon the subject. In the early summer of the following year it happened that the American Society of Mechanical Engineers and the American Rail- way Master Mechanics' Association met during the same week (June, * In this connection, it will be well to note that in the course of a discussion before the American Society of Mechanical Engineers in May, 1890 (Proceedings of the Society, Vol. XI, p. 886), Mr. Louis S. Wright proposed a scheme for testing locomotives embracing many of the important features which appear in the Purdue plant, as designed a few months later, and it is probable that the importance of the problem had led others to consider it as a subject for speculation. It is but proper to add that the Purdue designer did not draw his inspiration from any such suggestion. f Proceedings of the American Society of Mechanical Engineers, 1890, p. 591. 44 LOCOMOTIVE PERFORMANCE. 1891), one at Providence, R. I., and the other at Cape May, and it happened also that the subject of locomotive testing was introduced in both conventions. The Mechanical Engineers received a report of progress * from its committee appointed a year previous, and the Master Mechanics appointed a committee "to investigate the prac- ticability of establishing a standard system of tests to demonstrate the fuel and water consumption of locomotives^ also to ascertain the value of the steam-engine indicator for locomotive use."f I n the succeeding year (June, 1891, to June, 1892) the Mechanical Engineers' Committee made no report, but in December, 1892, it .again reported progress. J During this interval the Master Mechanics 7 -Committee had asked and received authority to confer with the 'Committee of the Mechanical Engineers, but the official records show :no further reference to the work of either committee or to any joint "work done by them until the summer of 1893, when the Committee of trie Master Mechanics' Association presented their report after conference with a similar committee appointed by the American Society of Mechanical Engineers. The report gave in elaborate form the specifications to be observed in conducting tests of loco- motives upon the road. It is an interesting fact that up to this time the record contains no reference to laboratory tests or to loco- motive-testing plants. 20. Interest in Purdue's Work. A paper describing Purdue's locomotive-testing plant was presented at the San Francisco meet- ing of the American Society of Mechanical Engineers in the summer of 1892,|| where it failed to draw out any discussion. But a year later, when the Master Mechanics' Committee already referred to rendered its report, the Convention after due discussion resolved, first, "That the Committee on Locomotive Tests be instructed to cooperate with the Committee from the Mechanical Engineers' Society in a series of comparative shop tests of compound and simple loco- motives to be made on the locomotive- testing apparatus of Purdue University," and, second, "That the Executive Committee be instructed * Proceedings of the American Society of Mechanical Engineers, 1891, p. 613. f The committee appointed in response to this resolution consisted of J. N. lander, J. Davis Barnett, Albert Griggs, John D. Campbell and F. W. Dean. (Pro- ceedings of the American Railway Master Mechanics' Association, 1891, p. 206.) J Proceedings of the Society, 1893, p. 21. Proceedings of the American Railway Master Mechanics' Association, 1893, p. 22. || "An Experimental Locomotive." Transactions of the American Society of Mechanical Engineers, Vol. XIII, p. 427. GROWTH OF INTEREST IN LABORATORY TESTS. 45 to provide funds to pay the necessary expenses connected with such shop tests, and solicit subscriptions on this account from railroad companies and locomotive-builders." A few weeks later, at the summer meeting of the American Society of Mechanical Engineers, a paper giving results of twenty efficiency tests upon the Purdue locomotive was presented,* and at this meet- ing, also, the committee appointed to report upon standard methods of conducting locomotive tests presented a document f dealing in a comprehensive manner with the question at issue, giving liberal atten- tion to shop tests and recommending such tests as the only satisfactory method of settling the important questions of performance. As a re- sult of this report a resolution was proposed calling for a committee to cooperate with the committee of the American Railway Master Mechanics' Association in carrying out investigations at Purdue. At the next regular meeting of the Society, however, the council to whom the above-mentioned resolution was referred reported that such a resolution was deemed inexpedient, and further consideration of the matter by the Mechanical Engineers was, therefore, dropped. Meanwhile it was being urged by the Master Mechanics' Associa- tion that funds and locomotives should be made available for work upon the Purdue plant, and while the movement here was also des- tined to failure, so far as the immediate purpose in view was con- cerned, the discussion and reports during the years 1893 to!895 leave no doubt as to the attitude of the Association with reference to the value of laboratory tests of locomotives. 21. New Plants. t In view of the value of the locomotive-testing plant as a means of determining the performance of locomotives, it was reasonable to expect that the number of such plants would increase. The first to be completed, following lines developed at Purdue, was that of the Chicago & Northwestern Railroad Com- pany, which went into commission in the year 1895; the next that of Columbia University, built in 1899. Following these came various plants for railroad companies and technical institutions both in this country and abroad, among which the most deserving of mention at this time (1906) is probably that of the Pennsylvania Railroad Company installed at St. Louis during the Exposition of 1904 and afterwards removed to the Company's shops at Altoona. * " Tests of the Locomotive at the Laboratory of Purdue University." Trans- actions of the American Society of Mechanical Engineers, Vol. XIV, pp. 826-854. f Transactions of the American Society of Mechanical Engineers, 1893, pp. 1312-1339. J Transactions cf the American Society of Mechanical Engineers, Vol. XXV, CHAPTER III. i LOCOMOTIVE SCHENECTADY NO. 1. 22. The Controlling Conditions Affecting the Choice of a Loco- motive which should serve the purposes of the laboratory were the outcome of a desire to have the machine, in size and in the char- acter of its details, fairly representative of good American practice at the time the order was given; that is, in 1891. The type of loco- motive and the size of its cylinders were the only points prescribed by the University authorities. The details of the design were left to FIG. 29. Elevation of Schenectady No. 1. the discretion of the builders, who, being thus untrammeled, pro- duced an engine which it is presumed represented their standard practice. There were heavier and more powerful engines in use, but there were also very many which were lighter and of less power. The locomotive was in every respect normal. No fixture or fitting was omitted which was then in common use on the road. It had air-brakes and a headlight, and had it been coupled to a tender and placed at the head of a train, it would have lacked nothing for the service which might then have been demanded of it. Neither 46 LOCOMOTIVE SCHENECTADY No. 1. 47 was there any detail of consequence added or modified by the builders to especially adapt the machine to the purposes of the laboratory. The locomotive, therefore, being representative in size and nor- mal in character, it was but reasonable to expect that results obtained from it would, in their general aspect at least, apply to a very large class of American locomotives. 23. Specifications as prepared by the builders are as follows: SCHENECTADY LOCOMOTIVE WORKS. SCHENECTADY, N. Y, April, 1891. SPECIFICATION No. 1491. Eight-wheel Locomotive Engine for Purdue University: Engine Schenectady, No. 1. GENERAL DESCRIPTION. Gauge 4 feet 8 inches. Fuel, bituminous coal. Cylinders 17" diam., 24" stroke. Drivers 63" diam. Driving-wheel base 8 ft. 6 in. Total wheel-base 22 ft. 11 in. Weight in working order, on drivers, about 56,000 Ibs. Total, about 85,000 Ibs. CONSTRUCTION. BOILER. To be of the best workmanship and material, to be capable of carrying with safety a working pressure of 140 Ibs. per square inch, and of sufficient capacity to supply steam economically. All horizontal seams quadruple-riveted with welt-strip inside. A double-riveted seam uniting waist with fire-box. All plates planed at edges and calked with round-pointed tools. Boiler to have extended front end. Waist, dome, and outside of fire-box of steel %%' thick. Diam. of waist at front end 52", made wagon-top, with one dome 30" diam. placed on wagon- top. FIRE-BOX. Of best quality fire-box steel, 72" long, 34f" wide, 73" deep. Crown-sheet three-eighths, tube-sheet one-half, side and back sheets five-sixteenths thick. Water-space 4" front, 3" sides, 3" back. All sheets thoroughly annealed after flanging. Stay-bolts \ and 1 inch diam., screwed and riveted to sheets, and placed not over 4 inches from center to center. Crown-sheet supported by crown-bars made of two pieces of wrought iron 5" wide, f " thick, placed not over 4 inches apart, reaching across crown and resting on edge of side-sheets. Crown-bars riveted to crown-sheets, with seven -eighth -inch rivets placed not over 4| inches from center to center, each bar having four stay-braces to top of boiler or dome. 48 LOCOMOTIVE PERFORMANCE. TUBES. Of charcoal-iron, 200 in number, 2" diam., 11 ft. 6 in. in length. Set with copper ferules at both ends. Cleaning-holes at the corners of fire-box, and blow-off cock on back. GRATE. Grates, rocking. Ash-pan with dampers front and back. STACK. Smoke-stack straight. Deflecting-plate and netting in smoke- box. FRAMES. Of best hammered iron, main frame in one section with braces welded in. Forward section securely bolted and keyed to main frame. Pedestals protected from wear by cast-iron shoes and wedges, and locked together at bottom by bolt through cast- iron thimble. Width of main frame 4". CYLINDERS. Of close-grained hard charcoal-iron. Cast with half-saddle attached, the right and left cylinders from the same pattern and interchangeable. Fitted together in a substantial manner and securely bolted and keyed to frame. Valve-face and steam- chest seat raised above face of cylinder to allow for wear. Cylinders oiled from N. & Co. No. 8 sight-feed lubricator placed in cab, with copper pipe under boiler-lagging to steam-chest. THROTTLE. Balanced valve placed in dome, with wrought- iron dry- pipe and cast-iron steam-pipe connecting to cylinders. PISTONS. Made with removable followers, or with solid heads, and fitted with approved steam-packing. Piston-rods of hammered iron. VALVE-MOTION. Approved shifting link-motion graduated to cut off equally at all points of the stroke. Links, sliding-blocks, plates, lifting- links, pins, and eccentric-rod jaws of the best hammered iron, thoroughly case-hardened. Steam-chest valves, Richardson balanced. GUIDES. Of hammered iron, case-hardened. CROSS-HEADS. Of cast steel with brass gibs. Style for four bar-guides. DRIVING-WHEELS. Four in number, 63" diam. Centers cast of the best charcoal- iron and turned to 57" diam. to receive tire. TIRES. Of American steel, 3" thick, both pairs flanged 5f" wide. AXLES. Of hammered iron, with journals 7" diam. 8" long. Driving- boxes of cast iron, with heavy Damascus bronze bearings and large oil- cellars. SPRINGS. Made of the best cast steel, tempered in oil. Secured to a system of equalizing- beams to insure the engine riding in the best possible manner. RODS. Connecting and parallel rods of hammered iron or steel, each forged solid, fitted with all necessary straps, keys, bolts, and brass bearings. LOCOMOTIVE SCHENECTADY No. 1. 49 CRANK-PINS. Of steel. WATER-SUPPLY. To have two injectors (Monitor No. 8 R. H., No. 7 L. H. ) with well- arranged cocks and valves for convenience in working. ENGINE-TRUCK. With square wrought-iron frame, cast-iron pedestals, and center bearing suitable for rigid center, with approved arrangement of equalizing beams and springs. WHEELS. Four double-plate chilled wheels of first-class manufacture, 28" diameter. AXLES. Of hammered iron, with inside journals 5" diam. , 9" long. Springs of best cast steel, tempered in oil. CAB. Constructed of seasoned ash, substantially built, and secured by joint- bolts and corner- plates. Furnished with seats and tool- boxes for engineer and fireman. PILOT. Of wood, strongly braced, and provided with substantial draw- bar. FINISH. Boiler lagged with wood and jacketed with planished iron, secured by planished iron bands. Dome lagged with wood, with sheet-iron casing and cast-iron rings painted. Cylinders lagged with wood, with sheet- iron casing and cast-iron head- covers finished. Steam-chests cased with sheet iron, with cast, iron covers finished. Hand-rail of iron finished, with columns secured to boiler. Boiler front and door of cast iron. FIXTURES AND Engine provided with sand-box, support for headlight, bell, FURNITURE. whistle, steam-gauge, gauge-cocks, cab-lamp, blower, oil-cans, torch; also all necessary wrenches, fire-tools, hammers, chisels, packing- tools, etc. Two jack-screws and a pinch- bar. Principal parts of engine fitted to gauges and templates, and interchangeable. All finished removable nuts case-hardened. I. A. Williams & Co. headlight. All threads United States standard. PAINTING. Engine to be well painted and varnished, with the road, number mark, and name put on in handsome style. PATENTS. All patent fees not covered by this specification excepted. BRAKE. Westinghouse automatic air-brake on drivers. 24. Drawings of the locomotive as covered by the preceding specifications follow the text of this chapter. 25. Constants. In calculating the performance of the locomotive from observed data, use is made of various dimensions, relationships, 50 LOCOMOTIVE PERFORMANCE. and calculated values, which for convenience are herein grouped under a single head as " constants." The dimensions upon which some of these values are based, as, for example, cylinder diameters, are subject to slight change resulting from use or repairs; and as the tests in question extended .over a period of several years, and as they were interrupted by the necessity of giving the locomotive general repairs, the values used as constants were carefully watched and frequently checked. All changes observed were, however, too small to merit consideration excepting those which occurred in the diameter of the cylinder at the time the engine was put through general repairs in 1894. After this date certain values which had previously prevailed could no longer be employed, and it will be noted that the tabulated statement exhibits two sets of constants for such factors. At the time these tests were made, authorities disagreed as to the method of measuring the extent of heating surface in boilers, some advocating the use of the inner surface, while others favored the use of the outer surface of the tubes. The general practice in locomotive work had been to employ the outside of the tube. In the Purdue work, for reasons which appeared logical the heating sur- face was defined as the inside surface of the fire-box plus the inside surface of the tubes plus the effective surface of the front head. A summary of the principal dimensions of the experimental locomotive and of constants used in calculating results derived from tests is as follows: Total weight (makers' figures) 85,000 pounds. Weight on four drivers (makers' figures) 56,000 " Total wheel-base 22 ft. 11 in. Driving-wheel base 6 " 6 " Nominal cylinder diameter 17 in. Nominal stroke of piston 24 " Diameter of piston-rods 3 " Area of piston-rod section 7.07 sq. in. Prior to Subsequent to Jan. 24, 1894. Jan. 24, 1894. Actual cylinder diameter (inches): Right side 17 17.047 Left side 17 17.031 Actual stroke of pistons (inches): Right side 24 23.936 Leftside.. 24 23.889 LOCOMOTIVE SCHENECTADY No. 1. 51 Prior to Jan. 24, 1894. Effective area of pistons (square inches): /Headend 226.98 Rl g htSlde \Crankend 219.91 /Headend 226.98 Leftside \Crankend 219.91 Piston displacement, cubic feet: /Headend 3.1525 Right side \ ~ . , AK/m I Crank end o . Uo4U /Headend. ... 3.1525 LeftSlde \Crankend 3.0540 Clearance volume, per cent of piston displacement: / Head end 9.390 Right side \ 1A Oft _ I Crank end 10 . 265 /Head end. ... 9.980 Left side s _, , , .,. _._ I Crank end 10.019 INDICATED H.P. CONSTANT, or the horse-power for one pound mean effective pressure and a speed of one revolution per minute. /Headend 01375 Right side ( Crankend 01332 f Head end 01375 I Crank end. .. .01332 Subsequent to Jan. 24, 1894. 228.23 221.16 227.82 220.75 3.1617 3.0637 3.1497 3.1520 9.103 10.046 10.086 9.880 .013796 .013368 .013744 .013318 ractive H.P. Constant, or the horse-power devel- oped at the draw-bar when there is a pull of one po.ind and a speed of one revolution per minute. . .0004981 .0004956 Drivers: Nominal diameter, inches 63 Actual circumference, inches 197.25 Actual diameter, inches 62. 7 63 196.25 62.47 Ports: Length, inches 16 Width of steam-ports, inches 1 Width of exhaust-port, inches 2 J Valves: Type "D" slide. Design Richardson balanced. 52 LOCOMOTIVE PERFORMANCE. The proportions and the setting of the valves were changed for different series of tests as follows: Reference Symbols. No. appearing in Col. 5, Table I. I II III IV V VI VII VIII IX XI Series. Outside Lap. Inches. Inside Lap. Inches. Inside Clearance. Inches. A Valve-setting. Builders' Lead Reduced Eccentrics Advanced Boiler: Diameter waist at front end, inches 52 Diameter tubes, outside, inches 2 Diameter tubes, inside, inches 1 . 78 Number of tubes 200 Length of tubes, feet. 11.5 Thickness of tubes, inches .11 Area of flameway through tubes, feet 3.46 Width of fire-box, inches 34.5 Length of fire-box, inches 72 Height of fire-box, inches 73 Heating surface, square feet: In fire-box 132. 1 In front head 10.5 In tubes calculated from outside diameter 1204.3 In tubes calculated from inside diameter 1071 .8 Total heating surface assuming the tube surface to be calculated from the outside diameter 1346.9 Total heating surface assuming the tube surface to be calculated from the inside diameter.* 1214.4 Grate area, square feet 17.25 Ratio of heating surface to grate area:. Assuming heating surface to be 1214.4 sq. ft 7.04 Pounds of water in boiler when filled to middle gauge 8450 Steam-space in boiler when filled to middle gauge, cubic feet 52 . 8 Ratio of steam-space to entire cubical capacity of boiler .29 Exhaust nozzle, double. * This factor has been used in all calculations involving heating surface. LOCOMOTIVE SCHENECTADY No. 1. 53 Weight of parts connecting with crank-pins: Piston and piston-rod, pounds 297 .0 Cross-head with part of indicator rigging attached, pounds 170.5 Main rod, pounds 344. 5 Side rod, pounds 278 .0 26. Steam-passages. The areas of the several steam-passages between the throttle and the exhaust-tips through which the steam passes are shown graphically with great clearness in the diagram Fig. 54. By use of this diagram in connection with the recorded data of tests it should be possible to make an approximate deter- mination of the steam velocity at various points in its course. HALF PLAN OF PIPE HALF PLAN OF NOZZLE FIG. 30. Stack. FIG. 31. Exhaust-pipe and -nozzle. 54 LOCOMOTIVE PERFORMANCE. FIG. 32. Dry-pipe. FIG. 33. Branch-pipe. LOCOMOTIVE SCHENECTADY No 1. 55 p E. I 56 LOCOMOTIVE PERFORMANCE. LOCOMOTIVE SCHENECTADY No. 1. 57 58 LOCOMOTIVE PERFORMANCE. FIG. 37. Valve-box and Cover. LOCOMOTIVE SCHENECTADY No. 1. 59 60 LOCOMOTIVE PERFORMANCE. FIG. 39. Piston and Piston-rod. I o jit \ . *- i w^-* co co a 1 L!r ^ _ .___. ~" O" 5 \ f < \ ! 1 ^r t (I a FIG. 47. Eccentric Rod. Drip Ploy FIG. 48. Eccentric Strap. LOCOMOTIVE SCHENECTADY No. L 65 FIG. 49. Eccentric. FIG. 50. Reverse-shaft. 66 LOCOMOTIVE PERFORMANCE. c FIG. 51. Reverse-lever and Quadrant. FIG. 52. Throttle-lever. LOCOMOTIVE SCHENECTADY No. 1. 67 FIG. 53. Throttle and Throttle.pipe. 68 LOCOMOTIVE PERFORMANCE < Throttle Port Throttle Pipe 90 ~rOD" : ~G. i do Elbow between Throttle and Dry Pipe Dry Pipe Body of T- Fitting in Smoke Box 33.6 C " 32.131?; 3i.81 D " ^Branches of T-Fitting in Smoke Box Branch Pipes in Smoke Box Openings into Saddle ( Steam Ports in Valve Seat G0 n Exhaust Pass;io-es in Exhaust Tips FIG. 54. Steam-passage Areas. CHAPTER IV. METHOD OF TESTING AND DATA. 27. Method of Testing. In preparation for a test the locomo- tive was run for an hour or two in the morning at low speed and light load. After the noon stop it was again started with the reverse lever in the position which it was to occupy during the test, while the load was gradually thrown on and the speed increased until the predetermined conditions for the test had been attained. At the stroke of a gong, usually between ten and thirty minutes after start- ing, all readings were taken and the test was commenced. The throttle was seldom touched after starting a test. Variations in boiler pressure during a test were so slight that changes in speed caused thereby were easily controlled by increasing or decreasing the brake load. At the end of the test care was taken to have the water in the boiler at the same level and the fire in the same con- ditions as at the start. Then upon the stroke of the gong the throttle was closed and the fire immediately dumped and quenched. Four regular attendants of the laboratory, including a fireman, prepared the locomotive for test, and during the tests fired the boiler and gave all necessary attention to the mounting and accessory machinery, A corps of student observers working under the direction of an in- structor acting as supervisor of tests was responsible for all the observations. Students thus employed were generally seniors in the Department of Mechanical Engineering. Twelve observers were re- quired, stationed as follows: Number of Observers. To keep running log, time, and gong ' 1 To weigh fuel; also to note time of starting and stopping injector, and to keep log of same 2 To weigh feed- water, to observe temperature, and to keep log of same 2 To operate indicators; also to take calorimeter readings, dry-pipe pressure, draft-gauge, and pyrometer readings ". 4 To read engine and supporting- wheel counters and dynamometer; also to read boiler pressure, and regulate the water pressure upon the several brakes 2 To occupy the engineer's box 1 12 69 70 LOCOMOTIVE PERFORMANCE. ENGINEERING LABORATORY. Of Tat Jfa 2l/__Sri,_ RUNNING LOG PURDUE UNIVERSITY. Diameter af CylinAm l7.O$f nin m *,t~. n f nriim ins),,, (j4(>T Reversing Lever /sr notch front A/OV. to Mfl.4- Time ind Lof Ketper A/ll'-'-* T)a.tf. (For other data see Fuel Log. Injector lag. Feed Water Lojrand Calorimeter Log.} R.a^.n.,, regular gong.. uch a tl m. n ..w"r),ake n th. run" ng'tlmeta'tweenTt''.^^'^. Dr"e*v fou. io'ng'.o^f ]' to' the t^Tf Si r.Vular"mrv7l'b7t "e'eVgong. 1 ) 3 ' f 1 I 1 ! l ! i lli 1 1 F I P 1 J = 1 $ \ I |J { | ll I I j 3 * /^f 6-?23&s. 659 2.6 76 9.Z /,r f 2 '20 .r !&?. fZ3. 2.6 q 2;25 ,-r 7023/6 2-99O. f5 /2f /zs 2.^ 76 Z9.Z 10 2.30 T /2f /zz 2.f II 2:35 5- 2-77% O o 130 /26. 2.J" 76. &* 12. 2>70 ,T /27 tzs. 26 13 2:-7j f 25 29.% 29 WO v f /JO I&S. g.f .................. . ______ Jg.7z.ze~ , __________ ZSLffL xxTi. XXIII. MOTES; No water JL^^^ ,_. FIG. 55. METHOD OF TESTING AND DATA. 71 As the readings were taken they were entered on the appropriate logs, facsimiles of which are shown in Figs. 55-60. The card log and the summary sheet (Figs. 59 and 60) were filled out later as the test was worked up, the card log containing all the indicator-card measure- ments, while the summary sheet, as its name implies, comprises all FEED-WATER LOG. Test No... 45-J.. ..Series./? ...... ..... Locomotive.. .". Observer. ____Ha8LJin. ______________ Date. . .Nov.. ,. 1695 LEFT BARREL. RI&HT B/JRREL. Check. " Time. Check. Time,. Check. Time Temjo. Chech. Time,. Temp. /:47 ' 3'/6 ' 1.50 5f ' 3.14 Y l'55 r 3 2/ ^ Z'OO ^ 3.23 f z:oi * 3:25 K Z.QT ^ 3:27 X 210 i/ 3.30 f Z'lZ 3.32 56. X z.w ' 335 Z.I6 57 ^ 3-37 y Z.IT S 3.40 V Z..ZQ , 3.^3 r 2:22 3:^6 s Z'.Z* * 3.50 r 2'.27 ^ 3:53 2 30 57 * 3-55 r 2--3E. ^ 357 If 2:3^ y 3.57 56. r 2.36 / 3.0Z v 2.3f r 4.05 r 2!^/ X q.oq s 2 W .56 ' q:il r 2!^7 r 4 13 r Z-50 q:is 56. v 2.52 , Z 55 y 2.5<7 s 301 56. / 3-0-7 , 306 I/ 3.01 ' 3.13 56. Left Barrel - - -Tiroes mpt/7 . .2f Caltirote^ ^..3/^. . M. of Water JO fZ?, Right Barrel - - - Times Emptied. . 2.7# . Calibrated WL.32J. _M. of WaUr.lQ?9ti Rveraje Temperature.. .364 : Total Wt. ofWater.l6f_?& /Vote. ..OflLld. OF LOST J3/IRRL . UfiEO, , _ i FIG. 56. the final calculated and observed results from the test, together with the constants and dimensions of the engine. 28. Data. The tables at the end of this chapter present the results of forty-four efficiency tests made upon locomotive Schenec- tady No. 1. These tests are not all that were run during the period 72 LOCOMOTIVE PERFORMANCE. covered, but they comprise all the later tests and have been chosen as those concerning the reliability of which no doubt can be enter- tained. Only final results are given, which, in the case of observed data, are generally the average of observations made at five-minute intervals and checked either by two different observers or by one observer and some form of automatic recording-instrument. In ENGINEERING LABORATORY. FUBL LOG 1 2 3 4 5 1 2 3 4 5 Tim of Cong. tVeisMof r u n BO*. Weight 'of Empty Bo,. Weight of Tim* of Gang. Mi? of FuZI Bar. Weight of Em^ty flo. WeioAt of Coal. I :is 2.55 332, 238. 323. 93. 2=50 ^IS 45. 222 3.'00 .527 !! taemJ fte NOTES '"K* / IMS /2a 5. 28a t 'j t:55 10 75-9. 5. 80. 5 2 '051 JO /27 5. 273. 1 7 ?-tf 10 /?fi 5 279.5 q 2:25 10 /2a 5. 279. J . II ?'tt If) 130. 5. 279.5 13 ?'15 in 730. 279.5 IS 2:55 10 /2fi. 5. . 273.5 n ,TfW m ' 130. ,5 279.5 11 M 10 123, 5. 279.5 2.1 3,'25 10 /23, 5. 2795 23 3\35 m IZ7. 279.5 25 3W 10 /28. 5. Z79.S 27 S.55 10 130. 5. 260. 2 QUALITY OF STEAM I Per cent of Priming /.S? 19. 226.07 FIG. 58. distinguished the various series are listed under Constants. Through- out this set of tables the tests are arranged in series, with low speeds and early points of cut-off first in each series, and for con- venience in comparison and reference, the Consecutive Numbers and Laboratory Symbols are reprinted with each table. In succeed- ing chapters it has been found convenient in some instances to 74 LOCOMOTIVE PERFORMANCE. select certain data and present them rearranged to suit the matter in hand. In all such cases the Laboratory Symbol affords a ready means of identification. The following concerns the several items appearing in the tables and the methods of computation: aronraanrauBORATOBT. PURDUE UNIVERSITY. PRESSURES AS OBTAINED FROM INDICATOR CARDS. fell Jfa^ ir /jS~ I . fieri ** Q_ O f 7./*vmt,/rfiW $ C. ^4 gftf ^ T/f Q Y- _, L - $faf7jt-NQVM$ER_0 ___t89__ Sprint of Card* 60 R> (*M OS . [ StU 1(0 63 14 11 1.7 20.7 101 70 IT 22 3. 2/.7 101 7/ If 25 3.3 22.-? 114 61 M ZO 26 2/2 No Ca r t 2/.3 106 74 17 24 4. 22.7 No Ca . /560. 405.$ 20?. 3/7 52. 213.1 1555. WO. 243. 32 4/.I J/6./ 1653. 105. 205. VT. Jflj 3/27 1615. 1005. 245 3 3* 333.7 ^H ;//.* Ml 11.1 116 2.2 20.1 ///./ 70.7 /7.3 21 2.7 22.6 IIT. 6U I1.b 1.1 LI 22.3 117.5 71.6 17.5 J/ C ZiT JL 3.3 'Of 30. 3T3 37. 70? jfhvlute Preuurt at Cut-off, TLZOf Absolute Mean Back-Fn 16.46% FIG. 59. Item 1. Consecutive Number. Item 2. Laboratory Symbol. Item 3. Date of Running Test. Item 4. Duration of Test in Minutes. METHOD OF TESTING AND DATA. 75 CNeiNEumOLAiORATOiir. SUMMARY OF RESULTS. m '......mm.: _ Afav,..Q+.., ..iJ'..o, LOCOMOTIVE.... .."5cHt-A r C7/.o.y TffAQQT fiNO ..CJK.Wot.ee Checked by R^.M'UjZfi .., CONCERNING THE LOCOMOTIVE. Tolal weight, Ibs TSfOOQ...., Cireumlereneo of driven, It 16.3.53.... Total are. M healing sarhee. &9..fttt. Worghto. ./?... .eViven ....56,000. Ota. ol boiler, inches ...*$. Area d grate surface . -J.9- . Fut. . . . . . rtfe el piston 23,.^/^..... Length d tube, Art ..//,f. Dt'i ot drivers CONSTANTS FOR TEST. | l.r in ./J.t.....wlck Irom oenlw Tractive lorte assumed to be necessary to draw one ten on level track at speed o( tesl . ., ...,.. Clearance, per cent, ol piston displacement . 3,!6l>tti>. .. A?f ...,9/XMfT ...OJ.37WI. &M... ... ./,.?* ..:. Beginning ol compression, per cent ol itroko. .LOJ3.6&79 JJffllSVI. ... 1,000 Jtf.6 :.**7!ffKf RESULTS FROM THE INDICATOR. Initial pressure ///.#.. ......l/Lt. I.I?,Q. ....U.7..?. //,t cut** .... , , 63*7... .......%..7 67,6 \ 7/:6 67,.?.. acupressure... A .. t ; ^.2 |?,.ff. .'.,&7. &-T. '.f*.' Prewife at the beginnlnj ol compression t '.. .- '.-.i.. .?.:(! .?.?.'. f&*3. ....?.?;/.. n efl.we press^ : -..., ...Jfat.. ...,.&ft.. ZZ3 .^f.f.. RESULTS FROM. LOOS OF TEST. lie Temperature el smoke bm......'........6$$'3S... Hi Kind ol fuel Ki Temperature ol laboralorj ?$(>. ..... $ Pound, ol dry coal Lv Temperature ol teed water .' $6:3.... Tr Pounds ol dry ash . Ell Seller pressure -.-. .....: tf*T&... IJTOJ Pounds o( waler delivered to belkjr m* U, Cio aromelric pressure ..../&30T.. Ntn Pounds water lost Iron boiler f/.:.7 Mil Height el water I* gauge- Pounds (lean uppll tUM . -..-.. t . puly ol let In I CALCULATED RESULTS. Mt RtoHT " 6t - ci - . ""'""..';. - "- t Weight ol steam at euH( ,,.., ..,...., Weight ol steam at releite .... ,^.., 4 Weight el steam during compression .-..., *e^aporlionperH.P..eflK)iir....:.....4t^f66 .;. . ejiin ^ .,..., ./ !. :...._...... f,?f< M Peds ol eej * tadlctW H. P. .1 Weight ol sleam per rev. by W........:^?T?:?*.,/ 18 'Ormometer H. P. ,,...,...., ,.,.... iw Me .....'. 4<3tf.f.. 5 Weight ,1 mixture In cylinder, 19 Friction ol e-vjin, I. H. P. ... > 29 Pounds ol teal per mile run ... """ * *M*K~. 20 FrhDM^lM.rMri.rf 30 P^, ^ c. M, .<<. 9 Per cent ol mixture accounted lor M , I. H.>. .'.. ; 31 p,,,^, ,, ^, ewewttef eer steam at cut-ol! /7i..T/ 2l OyMmonrtef worli In lol lon, ieime 1 ol etl '. 10 Per cent, ol mklure accounted (or a, per pound ol stean by tank 32 steam, I release... ...,...f /..< o^m.mrt w werl, h, tart toM I2F.. 11 Pounds ol steam per indicated H. P. _ per pound ol dry coal 3) g, T. U. take p.. K^t,, Indicator. f.KfXftf. 23 CquW^ wei,bl K trrt,-*. 34 B. T. U. tak, p p ^nd M ,BHM... QSH/*. M Pounds ol steam per indicated H. P. -. lens ,.ttt>.i. , 35 e. T. U. taken up by boiler per H. P. I* ..-.. , ., cars ol 33 Ions each -,.. r, ,. 35 Pounds coal per loot ol grate / ** ' !zni. . 28 ,.-.; 37 indkated H. P. per fool ol grate. ' - ~ 2. . ... FIG. 60. 76 LOCOMOTIVE PERFORMANCE. Item 5. Valve-setting and Proportions. The Roman numerals here refer to the table of valve proportions under Con- stants (Paragraph 25). Item 6. Approximate Cut-off is a predetermined condition for the test. Item 7. Reverse- Lever, notch Forward of Center. Item 8. Throttle Position. Item 9. Draw-bar Pull is derived from the known value of the work done in cylinders and of that lost by friction = Item 134 -Item 136. Item 10. Total Revolutions is the record from a counter attached to the rear driver, which was read every five minutes as a check on the regularity of running. In addition there was a Boyer railway speed-indicator reading in miles per hour. Item 11. Revolutions per Minute=Item 10-7-Item 4. Item 12. Miles Run Equivalent to Total Re volutions = Item 10 X Circumference of Driver in feet -f- 5280. Item 13. Miles per Hour=Item 12 --(Item 4-T-60). Item 14. Feed-water Temperature by Thermometer. Item 15. Laboratory Temperature by Thermometer. Item 16. Smoke-box Temperature is the record obtained from a copper-ball pyrometer used as follows: A copper ball of known weight was held in the smoke-box till it had presumably attained the temperature of its surround- ings, usually about forty minutes, and was then allowed to roll quickly down a tube into a water calorimeter. The original temperature of the ball was computed from data thus obtained. Item 17. Boiler Pressure by Gauge is the record of readings from an ordinary dial-gauge at intervals of five minutes as checked by a recording-gauge. Item 18. Atmospheric Pressure in Pounds. Item 19. Absolute Boiler Pressure =Item 17+Item 18. Item 20. Dry-pipe Pressure by Gauge. Item 21. Times Injectors were Started includes the record for both right-hand and left-hand injectors. Item 22. Minutes one or both injectors were in action. Shows the regularity of feed-water supply. Item 23. Draft is assumed to be the difference in pressure between METHOD OF TESTING AND DATA. 7T the atmosphere and the interior of the smoke-box, as measured by a U tube in inches of water. This tube- was fastened securely to the laboratory wall, having one end in pipe connection with the interior of the smoke-box and the other open to the atmosphere. The pipe entered the smoke-box at a point marked C, Fig. 35, Chapter III. Subsequent experiment has shown that this is not the point of least pressure, and that to obtain the actual greatest draft these values should be mul- tiplied by a factor of 1.3. (See Chapter XI.) Item 24. Water Delivered to Boiler and Presumably Evaporated is the actual weight of water delivered to the injectors minus the weight caught from the injector overflow. The water was measured by means of two carefully cali- brated barrels, one of which was filled while the other was being emptied into the receiving-tank. A water- meter in the supply-line served as a check to prevent any gross error in measuring. Item 25. Water Lost from Boiler is the weight of steam used by the calorimeter as determined by previous experiment. In no case were boiler leaks permitted. Item 26. Steam Supplied Engine = Item 24 -I tern 25. Item 27. Water Evaporated by Boiler per Hour=Item 24 (Item 4-60). Item 28. Steam Used by Engine per Hour=Item 26-=- (I tern 4-60).. Item 29. Quality of Steam as indicated by the throttling calorimeter. Item 30. Water Evaporated per Square Foot Heating Surf ace = Item 24-1214.4. Item 31. Water Evaporated per Square Foot Grate Surface=Item 24-17.25. Item 32. Dry Coal Fired is the total weight of coal less the amount of accidental moisture, as found by sample taken during each test, a correction which seldom exceeded one per cent. Item 33. Dry Ash Caught in the Ash-pan during the Test. Inas- much as under locomotive conditions a large per cent of the non-combustible material goes out of the stack y the weight of ash was not used to determine the amount of combustible fired. Item 34. Dry Cinders Caught in the Smoke-box. 78 LOCOMOTIVE PERFORMANCE. Item 35.--Dry Coal fired per hour = Item 32 -f- (Item 4-60). Item 36. Combustible Fired per Hour = Item 35X0.9. Through- out these tests the fuel used was Brazil block coal, mined at Brazil, Indiana. Frequent analysis of the chemical composition showed a very uniform content of combustible, the average being 90 per cent. Item 37. Dry Coal per Square Foot Heating Surface per Hour = Item 35-- 1214.4. Item 38. Dry Coal per Square Foot Grate Surface per Hour = Item 35 -f- 17.25. Item 39. Same as Item 29. Item 40. B. T. U. taken up by each pound of Steam is the result obtained by the formula B. T. V.=q^-q +xr, where q = ihe heat in one pound of water at boiler tem- perature, go = heat in one pound of water at tempera- ture of the feed-water, x = quality of steam as given in column 39, and r = the latent heat of evaporation at boiler pressure. (Peabody's Steam-tables used for all computations involving thermal quantities.) Item 41. B. T. U. taken up by Boiler per Minute =Item 40Xltem 27-60. Item 42. B. T. U. per Pound of Coal = Item 40Xltem 27 -f- 1 tern 35. Item 43. B. T. U. per Pound of Combustible = Item 40x1 tern 27 -f- Item 36. Item 44. Water Evaporated per Pound of Coal = Item 27 Item 35. Item 45. Water Evaporated per Pound of Combustible = Item 27 -r- Item 36. Item 46. Same as Item 27. Item 47. Equivalent Evaporation per Hour = Item 27 X Item 4 latent heat of evaporation at atmospheric pressure ( = 965.8 B. T. U.). Item 48. Equivalent Evaporation per Square Foot Heating Surface per Hour=Item 47-1214.4. Item 49. Equivalent Evaporation per Square Foot Grate Surface per Hour = I tern 47-17 25 Item 50. Equivalent Evaporation per Pound of Dry Coal = I tern 47-Item35. METHOD OF TESTING AND DATA. 7 Item 51. Equivalent Evaporation per Pound of Combustible = Item 47 + Item 36. Item 52. Boiler Horse-Power=Item 47-J-34.5. Items 53 to 102. Events of Stroke and Pressures above Atmosphere from Indicator-cards need no explanation. The same point on the card was used to determine the event of stroke and the corresponding pres- sure. Items 103 to 117. Weight of Steam present at the Several Events of the Stroke is the result obtained by the formula where w=the weight in pounds, v = the per cent of stroke plus clearance per cent multi- plied into the cylinder volume, and s = the volume of one pound of steam at the pressure above atmosphere as measured on the card. Items 118 to 122. Indicated Horse-power from the Card Measure- ments. Item 123. Steam per I. H. P. per Hour = Item 28 -I tern 122. Item 124. Steam per I. H. P. per Hour by Indicator = (Item 112 Item 117)Xltem llX60-r-Item 122. Item 125. Dry Coal per I. H. P. per Hour=Item 35^Item 122. Item 126. Weight of Steam per Revolution by Tank = Item 26-f- Item 10. Item 127. Weight of Mixture in Cylinder per Re volution = Item 126 +Item 117. Item 128. Per Cent of Mixture present as Steam at Cut-off = Item 107 -i- Item 127. Item 129. Per Cent of Mixture present as Steam at Release = Item 112 -Item 127. Item 130. Reevaporation per Re volution = Item 112 Item 107^ when Item 1 12 > Item 107. Item 131. Condensation per Revolution=Item 107 Item 112, when Item 107 > Item 112. Item 132. Reevaporation per I. H. P. per Hour = Item 130 X Item 11 X 60 H- I tern 122. Item 133. Condensation per I. H. P. per Hour = Item 131 X Item. 11X60^ Item 122. 80 LOCOMOTIVE PERFORMANCE. Item 134. Draw-bar Pull Equivalent to Average M. E. P. The D. H. P. =draw-bar pull X. 0004956 XR. P. M. and I. H. P. = average M. E. P.X.0542XR. P. M. Assumed that all the I. H. P. was transmitted to the draw-bar without friction losses. Then, these expres- sions are equal and their second terms may be equated. Performing the operation indicated and solving will give Draw-bar pull=M. E. P.X 109.419. Item 134 therefore = Item 102X109.419. It was found imprac- ticable to get the correct observed value of draw- bar stress, and therefore it was assumed that at equal cut-offs, friction was constant at all speeds. (See Chap- ter XIX.) Items 135 to 139 are computed on this basis. Item 135. Friction Expressed as M. E. P. is assumed constant for all tests at same cut-off. Item 136. Friction Expressed as Draw-bar Pull = Item 135X109.419. Item 137. Assumed Friction Horse-power = Item 135 X. 0542 X Item 11. Item 138. Draw-bar Horse-power=Item 122 -Item 137. Item 139. Mechanical Efficiency = I tern 138-=- 1 tern 122. Item 140. I. H. P. per Square Foot Heating Surf ace = Item 122 H- 1214.4. Item 141. I. H. P. per Square Foot of Grate Surface = I tern 122 -f- 17.25. Item 142. Steam Used per Dynamometer Horse-power per Hour = Item 28-7- I tern 138. Item 143. Coal Used per D. H. P. per Hour = Item 35 -v- Item 138. Item 144. B. T. U. Used by Engine per Minute = I tern 28 X Item 40-60. Item 145. B. T. U. Used by Engine per I. H. P. per Minute = Item 144 -Item 122. Item 146. B. T. U. Used by Engine per D. H. P. per Minute = Item 144 -I tern 138. Item 147. Coal Used per Mile Run = Item 32-Item 12. Item 148. D. H. P. per Square Foot Heating Surface = Item 138-7- 1214.4. METHOD OF TESTING AND DATA. 81 Item 149. D. H. P. per Square Foot Grate Surf ace = I tern 138-=- 17.25. Item 150. Water Evaporated from and at 212 F. per Hour = Item 47. Item 151. Coal fired per Hour = Item 150 10.08-. 000244 X Item 150' Item 152. Coal burned per Square Foot of Grate per Hour = Item 151^ 17.25. Item 153. Water per Pound of Coal=Item 27-r-Item 151. Item 154. Equivalent Evaporation per Pound of Coal = Item 150 -r- Item 151. Item 155. Coal per I. H. P. per Hour = Item 151 -Item 122. Item 156. Coal per D. H. P. per Hour = Item 151 -Item 138. Item 157. B. T. U. Taken up by Boiler per Pound of Coal= Item 40Xltem 153. Table XIX., of Locomotive Performance based on uniform firing conditions, is composed of several items so corrected as to eliminate the effect of irregularities in firing. In this one table the total weight of coal fired per hour, which is a factor in all these items, is not that which was found by experiment, but was calculated from the known weight of water evaporated per hour by use of the formula c w 10.08 -. 000244 JF' where C is the pounds of coal fired per hour, and W is the total pounds of water evaporated from and at 212 per hour. The significance of this formula has been developed in the following chap- ter, pages 150 and 151. For the purposes of this table, W, Item 153, was taken from the experimental results, and C, Item 151, computed. Items 151 to 155 are calculated, using the value C instead of the ex- perimental results. The coal used in these tests was the same throughout, namely, Indiana block from the Brazil field, a light fuel which burns freely to a fine ash. The following is the average result of four analyses: 82 LOCOMOTIVE PERFORMANCE. Per cent fixed carbon 51.06 Per cent volatile matter 39.26 Per cent combined moisture 3.14 Per cent ash. . 6.54 Total 100.00 B. T. U. per pound of dry coal 13,000 It has been shown by tests made subsequent to those herein discussed that 0.8 of a pound of first-class Pittsburg or West Virginia coal will give in locomotive service approximately the same results as one pound of Brazil block. The record of the observed and calculated results is as follows: METHOD OF TESTING AND DATA. 83 TABLE I. GENERAL CONDITIONS. Valve T)n-rn Setting Approx- Reverse- Labora- tory Symbol. Date. .L'lira- tion of Test, Min- and Propor- tions (SPP imate Cut-off, Per Cent of lever Notch Forward of Throttle Position.' Draw- bar Pull. 1 utes. V.occ Con- Stroke. Center. stants). fc 1 2 3 4 5 6 7 8 9 1 15-1&-V Nov. 23, '94 190 I 25 1st Fully open 4227 2 25-1-V Nov. 26, '94 240 I 25 1st Fully open ! 3657 3 35-1-V Dec. 4, '94 140 I 25 1st Fully open j 2796 4 55-1-V Dec. 18, '95 120 I 25 1st Fully open j 1544 5 15-1-A Dec. 12, '94 240 II 25 1st Fully open 4180 6 25-1-A Dec. 14, '94 255 II 25 1st Fully open 2893 7 35-1-A Dec. 17, '94 180 II 25 1st Fully open 2600 8 45-1-A Nov. 20, '95 150 II 25 1st Fully open 1870 9 55-1-A Nov. 25, '95 120 II 25 1st Fully open 1342 10 15-2- A Nov. 13, '95 180 II 35 2d Fully open 6409 11 25-2-A Oct. 25, '95 180 II 35 2d Fully open 5259 12 35-2-A Dec. 19, '94 180 II 35 2d Fully open 4157 13 45-2-A Nov. 18, '95 140 n 35 2d Fully open 3132 14 55-2-A Nov. 22, '95 68 ii 35 2d Fully open 2431 15 25-3-A Nov. 1, '95 122.5 n 45 3d Fully open 6666 16 35-3-A Nov. 15, '95 120 ii 45 3d Fully open 5050 17 15-9 -A Nov. 6, '96 150 ii 80 9th Partly open 7224 18 35-2-B Jan. 14, '95 170 in 35 2d Fully open 2756 19 35-2-C Jan. 23, '95 120 IV 35 2d Fully open 5100 20 35-2-E Jan. 16, '95 180 V 35 2d Partly open 2722 21 35-2-F Jan. 21, '95 180 VI 35 2d Partly open < 2851 22 15-1-G Nov. 9, '96 180 VII 25 1st Fully open : 4704 23 35-1-G Nov. 20, '96 170 VII 25 1st Fully open 2860 24 35-l 6 -G Dec. 2, '96 170 VII 25 1st Fully open 2903 25 55-1-G Nov. 23, '96 120 VII 25 1st Fully open 1796 26 35-2-G Nov. 13, '96 160 VII 35 2d Fully open 4467 27 35-3-G Dec. 4, '96 140 VII 45 3d Partly open 3423 28 15-9-G Nov. 12, '96 160 VII 80 9th Partly open 7473 29 15-1-H Dec. 9, '96 180 VIII 25 1st Fully op zn 4418 30 35-1-H Dec. 18, '96 160 VIII 25 1st Fully open 2857 31 55-1-H Feb. 11, '97 120 VIII 25 1st Fully open 1783 32 35-2-H Dec. 16, '96 120 VIII 35 2d Fully open 3833 33 35-2 6 -H Feb. 10, '97 160 VIII 35 2d Partly open 2857 34 35-3-H Dec. 14, '96 120 VIII 45 3d Partly open 3316 35 15-9-H Dec. 11, '96 180 VIII 80 9th Partly open 7114 36 15-1-1 March 27, '97 270 IX 25 1st Fully open 37 35-1-1 March 29, '97 170 IX 25 1st Fully open 38 55-1-1 April 1, '97 120 IX 25 1st Fully open 39 15-9-1 March 26, '97 190 IX 80 9th Partly open 40 15-4^T April 14, '97 300 X 25 4th Fully open 41 35-4-J April 13, '97 180 X 25 4th Fully open 42 15-15-J April 9, '97 160 X 80 15th Partly open 43 15-2-K April 27, '97 300 XI 35 2d Fully open 44 35-2-K April 24, '97 180 XI 35 2d Fully open i LOCOMOTIVE PERFORMANCE. TABLE II. GENERAL CONDITIONS Continued. Speed. Temperature, Degrees F. Labora- Miles 1 tory Symbol. Total Revolu- tions. Revolu- tions per Minute. Run Equiva- lent to Total Revolu- Miles Hour. Feed- water. Labora- tory. Smoke- box. tions. 1 2 10 11 12 13 14 15 16 1 15-1 6 -V 14,891 78.37 46.12 14.56 53.76 73.93 549.50 2 25-1-V 29,174 121.56 90.36 22.59 53.85 69.62 605.60 3 35-1-V 26,111 186.51 80.88 34.66 51.90 69.00 633.00 4 55-1-V 36,080 300.67 111.75 55.88 58.10 75.70 666.60 6 15-1-A 19,367 80.69 59.99 15.00 53.20 65.46 553.20 6 25-1-A 33,024 129.50 102.29 24.07 53.75 67.34 567.0 7 35-1-A 34,373 190.96 106.47 35.49 53.17 74.91 628.10 8 45-1-A 37,359 249.06 115.71 46.29 56.40 75.60 644.30 9 55-1-A 36,693 305.75 113.65 56.83 56.00 75.90 675.40 10 15-2- A 13,878 77.10 42.98 14.33 55.19 79.30 620.60 11 25-2-A 22,973 127.63 71.15 23.72 56.00 84.80 696.30 12 35-2-A 34,204 190.02 105.94 35.31 52.54 72.20 720.30 13 45-2-A 34,454 246.10 106.72 45.74 55.00 83.70 740.50 14 55-2-A 20,808 306.00 64.45 56.87 58.40 76.40 754.60 15 25-3-A 15,819 129.10 49.00 24.00 53.30 76.40 762.10 16 35-3-A 22,054 183.78 68.31 34.15 55.50 76.70 798.10 17 15-9- A 11,886 79.24 36.82 14.73 55.62 78.50 724.20 18 35-2-B 32,029 188.41 99.20 35.01 50.00 65.90 664.30 19 35-2-C 22,746 189.55 70.45 35.23 51.68 70.80 737.70 20 35-2-E 34,947 194.15 108.24 36.08 51.88 70.60 652.00 21 35-2-F 33,673 187.07 104.30 34.76 52.72 71.50 653.10 22 15-1-G 14,589 81.05 45.18 15.06 53.33 68.70 582.50 23 35-1-G 32,361 190.36 100.23 35.38 55.00 74.40 684.70 24 35-l 6 -G 32,341 190.24 100.17 35.35 52.47 77.60 618.30 25 55-1-G 36,300 302.50 112.43 56.22 54.17 79.40 26 35-2-G 29,952 187.20 92.77 34.80 53.55 77.90 654.50 27 35-3-G 26,612 190.09 82.43 35.33 52.62 69.60 719.00 28 15-9-G 12,619 78.87 39.08 14.66 53.13 77.40 724.00 29 15-1-H 14,660 81.44 45.41 15.14 54.63 71.50 570.00 30 35-1-H 30,397 189.98 94.15 35.31 52.59 67.10 655.00 31 55-1-H 35,308 294.23 109.36 54.68 52.08 70.60 695.00 32 35-2-H 23,183 193.19 71.81 35.90 53.01 71.80 33 35-2 6 -H 30,653 191.58 94.94 35.60 50.56 74.30 689.20 34 35-3-H 23,289 194.07 72.13 36.07 53.00 73.50 35 15-9-H 14,421 80.12 44.67 14.89 53.22 81.10 695.70 36 15-1-1 21,853 80.93 67.69 15.04 52.80 66.40 37 35-1-1 32,444 190.84 100.49 35.47 56.30 74.50 38 55-1-1 34,868 290.57 108.00 54.00 54.50 79.50 39 15-9-1 15,067 79.30 46.67 14.73 51.90 68.00 40 15-4-J 24,934 83.11 77.23 15.45 53.77 41 35-4^1 34,308 190.60 106.26 35.42 54.00 42 15-15-J 12,987 81.17 40.23 14.62 51.10 43 15-2-K 24,671 82.24 76.41 15.28 56.00 44 35-2-K 33,722 180.73 104.45 34.82 56.00 METHOD OF TESTING AND DATA. 85 TABLE III. GENERAL CONDITIONS Continued. Pressures, Pounds per Square Inch. Injectors. Labora- Draft, Inches Number. tory Symbol. Boiler by Gauge. Atmos- phere by Barom- eter. Absolute Boiler. Dry Pipe by Gauge. Times Started. Minutes One or Both in Action. of Water.* 1 1 2 17 18 19 20 21 22 23 15-1 6 -V 127.27 14.54 141.8 130.00 14 129 2.04 2 25-1-V 127.12 14.34 141.4 130.40 7 213 2.60 3 35-1-V 128.87 14.32 143.2 127.20 1 136 3.43 4 55-1-V 128.40 14.38 142.8 123.04 1 120 3.20 5 15-1-A 125.94 14.46 140.4 124.58 22 130 1.72 6 25-1-A 120.04 14.52 134.6 113.37 10 236 1.93 7 35-1-A 129.72 14.62 144.3 127.23 2 180 3.00 8 45-1-A 128.84 14.31 143.1 124.90 2 146 2.68 9 55-1-A 124.91 14.18 139.1 121.29 1 120 2.58 10 15-2- A 129.48 14.39 143.9 125.10 9 149 2.42 11 25-2-A 129.32 14.35 143.7 124.90 180 3.37 12 35-2-A 131.65 14.49 146.1 120.67 180 4.42 13 45-2-A 126.67 14.29 141.0 122.10 140 4.93 14 55-2-A 124.00 14.43 138.4 118.80 68.5 4.58 15 25-3-A 127.19 14.47 141.7 123.20 122.5 5.45 16 35-3-A 125.28 14.34 139.6 122.08 120 7.49 17 15-9- A 122.47 14.34 136.8 74.17 150 4.56 18 35-2-B 98.36 14.39 112.8 95.57 8 158 3.28 19 35-2-C 143.28 14.42 157.7 143.08 3 125 5.13 20 35-2-E 128.05 14.49 142.6 95.46 3 175 2.89 21 35-2-F 155.35 14.68 170.1 93.11 9 149 2.57 22 15-1-G 123.58 14.39 138.0 120.54 12 127 1.87 23 35-1-G 125.00 14.55 139.6 123.91 1 170 3.02 24 35-1&-G 128.23 14.56 142.8 126.44 1 170 2.98 25 55-1-G 126.90 14.46 141.4 123.18 1 120 3.57 26 35-2-G 121.04 14.54 135.5 118.34 1 160 4.65 27 35-3-G 125.93 14.34 140.2 86.62 140 4.88 28 15_9_G 124.06 14.49 138.6 76.33 154 4.76 29 15-1-H 123.55 14.32 137.9 119.97 126 1.93 30 35-1-H 121.16 14.57 135.8 117.09 160 3.00 31 55-1-H 127.48 14.39 141.9 121.52 120 3.44 32 35-2-H 112.04 14.58 126.6 108.36 115 4.33 33 35-2 6 -H 122.64 14.54 137.1 92.97 160 3.18 34 35-3-H 116.48 14.29 130.8 74.52 120 4.52 35 15-9-H 122.72 14.37 137.1 73.94 180 4.99 36 15-1-1 127.95 14.53 142.5 124.43 37 35-1-1 128.05 14.45 142.5 124.97 38 55-1-1 125.43 14.47 139.9 122.90 39 15-9-1 131.61 14.35 145.9 72.91 40 15_4_J 127.51 14.50 142.0 126.49 41 35-4-J 131.95 14.42 146.4 124.05 42 15-15^1 128.94 14.41 143.3 89.61 43 15-2-K 125.50 14.21 139.7 123.31 44 35-2-K 126.02 14.15 140.2 124.57 * Multiply these values by 1.3 to obtain draft at point of maximum draft (in front of dia- phragm). LOCOMOTIVE PERFORMANCE. TABLE IV. BOILER PERFORMANCE. WATER AND STEAM. 1 Labora- tory . Symbol. Water Delivered to Boiler and Pre- sumably Evap- orated. Water Lost from Boiler. Steam Supplied to Engine. Water Evapor- ated by Boiler Hour. Steam Used by Engine Hour. Quality of Steam in Dome. Mois- ture. Water Evap- orated per Hour per Square Foot of Heating Surface. Grate Sur- face. g Per fc Lbs. Lbs. Lbs. Lbs. Lbs. Cent. Lbs. Lbs. 1 2 24 25 36 27 28 29 30 31 i 15-VV 17,789.6 67.6 17,721.0 5,617.8 5,596.4 .63 4.626 325 2 25-1-V 29,475.7 82.3 29,393.4 7,368.9 7,348.4 .68 6.068 427 3 35-1-V 19,899.2 48.6 19,850.6 8,528.2 8,507.4 .63 7.022 494 4 55-1-V 19,418.4 41.6 19,376.8 9,709.2 9,688.4 .88 7.995 563 5 15-1-A 22,100.2 81.6 22,018.6 5,525.1 5,504.7 .49 4.549 320 6 25-1-A 25,894.6 84.5 25,810.1 6,092.8 6,072.9 .78 5.017 353 7 35-1-A 24,367.5 62.8 24,304.7 8,122.5 8,101.6 .83 6.688 471 8 45-1-A 21,672.6 51.7 21,620.9 8,669.0 8,648.3 .87 7.138 502 9 55-1-A 17,882.6 40.6 17,842.0 8,941.3 8,921.0 1.22 7.363 518 10 15-2-A 21,840.6 62.7 21,777.9 7,280.2 7,259.3 .76 5.995 422 11 25-2-A 28,906.0 62.7 28,843.3 9,635.3 9,614.4 .90 7.934 558 12 35-2-A 34,324.5 64.0 34,260.5 11,441.5 11,420.2 .06 9.422 663 13 45-2-A 29,100.4 47.7 29,052.7 12,471.0 12,451.2 .24 10.269 722 14 55-2-A 15,911.4 22.9 15,888.5 14,039.5 14,019.2 .13 11.561 813 15 25-3-A 26,401.0 41.8 26,359.2 12,931.1 12,910.6 .11 10.648 749 16 35-3-A 29,874.8 40.6 29,834.2 14,937.4 14,917.1 .11 12.300 865 17 15-9-A 28,522.0 50.0 28,472.0 11,408.8 11,388.8 .94 9.395 661 18 35-2-B 24,554.7 46.5 24,508.2 8,666.3 8,649.9 .69 7.136 502 19 35-2-C 26,008.5 45.6 25,962.9 13,004.2 12,981.3 .70 10.708 754 20 35-2 -E 25,696.7 62.1 25,634.6 8,565.6 8,544.9 .68 7.053 496 21 35-2-F 24,333.3 74.1 24,259.2 8,111.1 8,086.4 1.00 6.679 470 22 15-1-G 18,695.7 60.3 18,635.4 6,231.9 6,211.8 1.06 5.132 361 23 35-1-G 26,562.7 57.5 26,505.2 9,375.1 9,354.8 1.14 7.720 543 24 35-l 6 -G 26,743.3 58.9 26,684.4 9,438.8 9,418.0 .40 7.772 547 25 55-1-G 20,760.9 41.0 20,719.9 10,380.4 10,359.9 .31 8.548 602 26 35-2-G 32,846.0 52.8 32,793.2 12,317.3 12,297.4 .44 10.143 714 27 35-3-G 27,060.2 47.6 27,012.6 11,597.2 11,576.8 .62 9.550 675 28 15-9-G 30,872.4 55.7 30,816.7 11,577.1 11,556.2 .29 9.533 671 29 15-1-H 18,223.4 60.3 18,163.1 6,074.5 6,054.4 .62 5.002 351 30 35-1-H 24,723.3 52.8 24,670.8 9,271.2 9,251.4 .49 7.634 537 31 55-1-H 20,513.6 41.0 20,472.6 10,256.6 10,236.3 .29 8.446 595 32 35-2-H 24,733.3 36.8 24,696.5 12,366.6 12,348.2 .34 10.183 717 33 35-2&-H 25,829.0 50.7 25,778.3 9,685.9 9,666.9 .20 7.976 561 34 35-3-H 23,385.7 38.2 23,347.5 11,692.9 11,673.8 .27 9.628 677 35 15-9-H 34,373.4 60.3 34,313.1 11,457.8 11,437.7 .29 9.435 664 36 15-1-1 27,608.3 93.6 27,514.7 6,135.2 6,114.4 5.052 356 37 35-1-1 26,781.0 58.8 26,722.2 9,453.2 9,432.5 7.780 548 38 55-1-1 22,509.0 40.6 22,468.4 11,254.5 11,234.2 9.268 652 39 15-9-1 34,374.3 67.1 34,307.2 10,857.3 10,836.1 8.940 630 40 15-4-J 32,783.5 102.8 32,680.8 6,556.7 6,536.2 5.390 380 41 35-4-J 30,646.7 63.6 30,583.1 10,215.6 10,194.4 8.413 592 42 15-15-J 30,879.1 57.5 30,821.6 11,608.7 '11,587.0 9.558 673 43 15-2-K 33,443.0 101.0 33,451.0 6,710.4 6,690.2 5.525 388 44 35-2-K 30,934.0 61.2 30,872.8 10,311.3 10,290.9 8.490 597 METHOD OF TESTING AND DATA. 87 TABLE V. BOILER PERFORMANCE Continued. COAL CONSUMPTION BRAZIL BLOCK COAL. Total Dl 7 Dry Cinders Dry i"* rt _l Combus- tible Dry Coat Fired per Hour per Square Foot of Labora- Dry Coal Asn Caught Caught in Coal Fired Fired per J5 tory Symbol. Fired.* in Ashpan. Smoke- box. per Hour.* Hour by Analysis. Heating Surface.* Grate Surface.* a 3 * Lbs. Lbs. Lbs. Lbs. Lbs. Lbs. Lbs. 1 2 32 33 34 35 36 37 38 1 15-VV 2,577.2 261.5 31.0 813.9 732 .670 47.2 2 25-1-V 4,460.0 320.0 79.5 1,115.0 1,003 .918 64.6 3 35-1-V 3,180.0 272.0 124.5 1,362.9 1,227 1.122 78.9 4 55-1-V 3,330.0 238.5 255.0 1,665.0 1,498 1.371 96.5 5 15-1-A 3,399.0 292.0 57.0 849.7 765 .700 49.3 6 25-1-A 3,864.3 93.5 60.0 909.2 818 .748 52.8 7 35-1-A 3,784.9 299.0 135.0 1,261.6 1,135 1.038 73.1 8 45-1-A 3,272.2 231.5 135.7 1,308.9 1,178 1.078 75.9 9 55-1-A 2,980.3 219.5 162.0 1,490.2 1,341 1.227 86.4 10 15-2- A 3,296.7 257.5 71.0 1,098.9 989 .905 63.7 11 25-2-A 4,826.0 25.0 227.0 1,608.7 1,448 1.324 93.3 12 35-2-A 5,932.9 257.5 239.0 1,977.6 1,780 1.628 114.6 13 45-2-A 5,720.2 242.5 380.5 2,451.5 2,206 2.019 142.1 14 55-2-A 2,994.8 205.0 330.0 2,642.5 2,378 2.176 153.2 15 25-3-A 4,684.1 91.0 231.0 2,294.3 2,065 1.889 133.0 16 35-3-A 6,266.3 298.0 118.2 3,133.2 2,819 2.581 181.6 17 15-9- A 5,014.0 265.0 174.0 2,005.6 ,805 1.651 116.3 18 35-2-B 4,369.7 460.5 177.5 1,542.2 ,388 1.270 89.4 19 35-2-C 5,356.1 384.0 348.0 2,678.1 2,410 2.205 155.3 20 35-2-E 4,106.5 353.0 113.5 ,368.8 ,232 1.127 79.3 21 35-2-F 3,859.1 294.5 109.5 ,286.3 ,158 1.059 74.6 22 15-1-G 2,678.0 262.0 43.0 892.7 803 .735 51.8 23 35-1-G 4,288.4 341.5 141.5 ,518.1 ,362 .246 87.7 24 35-1&-G 4,150.3 380.0 179.0 ,464.8 ,318 .206 84.9 25 55-1-G 3,768.9 241.0 314.5 ,884.5 ,696 .552 109.2 26 35-2-G 6,182.7 363.0 168.5 2,318.5 2,087 .909 134.4 27 35-3-G 4,707.9 365.0 157.0 2,017.7 ,816 .662 117.0 28 15-9-G 5,486.7 354.0 168.0 2,057.5 ,852 .694 119.3 29 15-1-H 2,374.5 240.0 37.0 791.5 713 .652 45.9 30 35-1-H 4,152.3 320.0 169.0 1,557.1 1,401 .283 90.3 31 55-1-H 3,203.1 213.5 233.5 1,601.6 1,441 .319 92.8 32 35-2-H 4,250.3 297.0 204.0 2,125.1 1,912 .750 123.2 33 35-2&-H 4,321.2 331.5 100.5 1,620.4 1,458 .334 93.9 34 35-3-H 4,248.3 346.0 300.0 2,124.1 1,912 .749 123.2 35 15-9-H 6,363.1 344.0 234.0 2,121.0 1,909 1.746 123.0 36 15-1-1 37 35-1-1 38 55-1-1 39 15-9-1 40 15-4^1 41 35-4-J 42 15-15-J 43 15-2-K 44 35-2-K * To obtain an approximate measure of boiler performance in terms of West Virginia or Pittsburgh coal, multiply the value of the columns 32, 35, 37. and 38 by 0.8. 88 LOCOMOTIVE PERFORMANCE. TABLE VI. BOILER PERFORMANCE Continued. B. T. U. Taken Up by Actual Water Evaporated Quality Labora- of Steam tory in Dome, Each Boiler Boiler Pei^ Per tl Symbol. Moisture. Pound Boiler per per Pound Pound 1 of per Pound Pound of Coal.* Combusti- Water. Minute. of Coal. Combus- ble. 3 tible. fe Per Cent. Lbs. Lbs. 1 2 39 40 41 43 43 44 45 1 15-1 6 -V 0.63 1,162.6 108,858 8,026 8,918 6.903. 7.670 2 25-1-V 0.68 1,161.8 142,688 7,678 8,531 6.609 7.343 3 35-1-V 0.63 1,164.7 165,543 7,288 8,098 6.258 6.953 4 55-1-V 0.88 1,156.3 187,109 6,743 7,492 5.831 6.479 5 15-1-A 0.49 1,164.1 107,194 7,569 8,410 6.502 7.224 6 25-1-A 0.78 ,160.0 117,798 7,771 8,638 6.701 7.434 7 35-1-A 0.83 ,161.7 157,268 7,479 8,310 6.438 7.153 8 45-1-A 0.87 ,158.1 167,336 7,671 8,523 6.623 7.379 9 55-1-A 1.22 ,154.7 172,079 6,929 7,699 6.000 6.666 10 15-2-A 0.76 ,160.2 140,778 7,686 8,541 6.625 7.361 11 25-2-A 0.90 ,158.2 185,994 6,937 7,709 5.990 6.633 12 35-2-A .06 ,159.3 221,072 6,707 7,452 5.786 6.429 13 45-2-A .24 ,155.9 240,262 5,881 6,534 5.087 5.652 14 55-2-A .13 ,153.9 269,996 6,130 6,812 5.313 5.903 15 25-3-A .11 ,158.8 249,743 6,531 7,256 5.636 6.262 16 35-3-A .11 ,156.3 287,871 5,513 6,126 4.768 5.296 17 15-9-A 0.94 ,157.2 220,038 6,583 7,314 5.688 6.320 18 35-2-B 0.69 ,160.5 167,624 6,521 7,245 5.619 6.243 19 35-2-C 0.70 ,166.7 252,873 5,666 6,296 4.856 5.395 20 35-2-E 0.68 ,164.1 166,188 7,283 8,093 6.258 6.953 21 35-2-F .00 ,165.1 157,502 7,348 8,163 6.305 7.005 22 15-1-G .06 ,158.6 120,340 8,088 8,988 6.981 7.756 23 35-1-G .14 ,156.5 180,706 7,164 7,961 6.194 6.882 24 35-l 6 -G .40 ,157.4 182,072 7,458 8,287 6.444 7.160 25 55-1-G .31 ,156.1 200,015 6,369 7,076 5.508 6.120 26 35-2-G .44 ,154.6 237,025 6,134 6,815 5.313 5.903 27 35-3-G .62 ,154.8 223,206 6,637 7,375 5.748 6.386 28 15-9-G .29 ,157.0 223,243 6,510 7,234 5.627 6.252 29 15-1-H .62 ,152.4 116,668 8,842 9,823 7.675 8.528 30 35-1-H .49 ,155.3 178,516 6,879 7,643 5.954 6.615 31 55-1-H .29 ,158.5 198,046 7,419 8,243 6.404 7.115 32 35-2-H .34 ,153.5 237,746 6,713 7,459 5.819 6.465 33 35-2&-H .20 ,160.0 187,260 6,934 7,704 5.977 6.641 34 35-3-H .27 ,155.9 225,266 6,366 7,071 5.505 6.116 35 15-9-H .29 ,156.6 220,866 6,248 6,942 5.402 6.002 36 15-1-1 37 35-1-1 38 55-1-1 39 15-9-1 40 15-4-J 41 35-4-^J 42 15-15-J 43 15-2-K 44 35-2-K J To obtain an approximate measure of the evaporative performance of the boiler, when using West Virginia or Pittsburgh coal, multiply column 44 by 1 . 25. METHOD OF TESTING AND DATA. 89 TABLE VII. BOILER PERFORMANCE Continued. Equivalent Evaporation from and at 212 F. 1 Labora- tory Symbol. Water Evaporater Hour. Per Hour. Per Square Foot of Heating Surface Per Square Foot of Grate Surface Per Pound of Dry Coal.* Per Pound of Com- bustible Boiler Horse- power (34.5 Evapo- rative JP s 3 per Hour per Hour units = 1B.H.P.) S? Lbs. Lbs. Lbs. Lbs. 1 2 46 47 48 49 50 51 52 1 15-1 6 -V 5,617.8 6,762 5.57 392 8.31 9.24 196 2 25-1 -V 7,368.9 8,864 7.30 514 7.95 8.84 257 3 35-1-V 8,528.2 10,286 8.47 596 7.54 8.38 298 4 55-1-V 9,709.2 11,650 9.57 673 6.98 7.76 337 5 15-1-A 5,525.1 6,660 5.48 386 7.83 8.70 193 6 25-1-A 6,092.8 7,318 6.03 424 8.05 8.95 212 7 35-1-A 8,122.5 9,751 8.03 566 7.74 8.61 283 8 45-1-A 8,669.0 10,395 8.56 603 7.96 8.83 301 9 55-1-A 8,941.3 10,690 8.80 619 7.17 7.95 310 10 15-2-A 7,280.2 8,745 7.20 507 7.95 8.84 253 11 25-2-A 9,635.3 11,555 9.51 669 7.18 7.98 335 12 35-2-A 11,441.5 13,733 11.31 796 6.94 7.72 398 13 45-2-A 12,471.6 14,926 12.29 865 6.09 6.77 433 14 55-2-A 14,039.5 16,773 13.81 973 6.34 7.04 486 15 25-3-A 12,931.1 15,515 12.77 899 6.76 7.51 450 16 35-3-A 14,937.4 17,878 14.73 1,037 5.71 6.36 518 17 15-9- A 11,408.8 13,670 11.25 792 6.81 7.57 396 18 35-2-B 8,666.3 10,404 8.57 604 6.75 7.50 302 19 35-2-C 13,004.2 15,709 12.93 911 5.86 6.52 456 20 35-2-E 8,565.6 10,324 8.51 598 7.54 8.38 299 21 35-2-F 8,111.1 9,785 8.06 567 7.61 8.45 284 22 15-1-G 6,231.9 7,476 6.16 433 8.37 9.31 217 23 35-1-G 9,375.1 11,226 9.24 650 7.41 8.24 325 24 35-1&-G 9,438.8 11,312 9.31 655 7.72 8.58 328 25 55-1-G 10,380.4 12,425 10.23 720 6.59 7.33 360 26 35-2-G 12,317.3 14,725 12.12 853 6.35 7.06 427 27 35-3-G 11,597.2 13,866 11.42 804 6.87 7.63 402 28 15-9-G 11,577.1 13,869 11.42 804 6.74 7.49 402 29 15-1-H 6,074.5 7,248 5.97 420 9.15 10.17 210 30 35-1-H 9,271.2 11,090 9.13 643 7.12 7.92 321 31 55-1-H 10,256.6 12,302 10.13 713 7.68 8.54 356 32 35-2-H 12,366.6 14,769 12.16 856 6.95 7.73 428 33 35-26-H 9,685.9 11,633 9.58 674 7.18 7.97 337 34 35-3-H 11,692.9 13,993 11.52 811 6.59 7.32 406 35 15-9-H 11,457.8 13,752 11.32 796 6.47 7.18 398 36 15-1-1 6,135.2 37 35-1-1 9,453.2 38 55-1-1 11,254.5 39 15-9-1 10,857.3 40 15-4-J 6,556.7 - 41 35-4-J 10,215.6 42 15-15^1 11,608.7 43 15-2-K 6,710.4 44 35-2-K 10,311.3 * To obtain an approximate measure of the evaporative performance of the boiler, when using West Virginia or Pittsburgh coal, multiply column 50 by 1.25. 90 LOCOMOTIVE PERFORMANCE. TABLE VIII. ENGINE PERFORMANCE. EVENTS OF STROKE BY INDICATOR. Admission Per Cent of Stroke. Cut-off Per Cent of Stroke. Labora- tory Right Side Left Side. Right Side. Left Side. Symbol. Aver- Aver- 1 age. age. 1 H.E C.E H.E C.E. H.E. C.E. H.E. C.E. 1 2 53 54 55 56 57 58 59 60 61 62 1 15-U-V 3.05 6.10 4.60 6.60 5.09 24.5 24.8 25. OC 26. OC 25.08 2 25-1-V 3.05 6.10 4.60 6.60 5.09 24.5 24.8 25. OC 26.00 25.08 3 35-1-V 3.05 6.10 4.60 6.60 5.09 24.5 24.8 25. OC 26.00 25.08 4 55-1-V Car ds un satis! actory 26.04 25.1 26.09 23.95 25.31 5 15-1-A 2.25 4.25 3.50 3.00 3.25 23.5 25.2 25.75 24.25 24.68 6 25-1-A 2.25 4.25 3.50 3.00 3.25 23.50 25.2 25.75 24.25 24.68 7 35-1-A 2.25 4. '25 3.50 3.00 3.25 23.50 25.25 25.75 24.25 24.68 8 45-1-A 1.79 2.64 1.79 2.39 2.15 21.36 23.04 22.97 22.36 22.43 9 55-1-A Car Is un atisf actory 23.16 23.54 24.98 22.53 22.61 10 15-2-A 1.10 1.70 1.40 1.80 1.50 39.10 36.30 38.80 37.10 37.80 11 25-2-A 1.09 1.86 1.28 1.91 1.53 33.26 34.6 34.92 34.69 34.37 12 35-2-A 1.00 2.50 2.00 1.30 1.70 32.75 33.00 35.25 33.75 33.69 13 45-2-A 1.39 2.40 1.11 2.10 1.75 32.40 33.00 33.80 33.10 33.07 14 55-2-A Car s un atisf qtory 35.30 34.70 36.60 36.00 35.40 15 25-3-A 1.20 1.90 1.00 1.20 1.32 44.40 43.70 44.80 44.00 44.22 16 35-3-A 1.81 1.69 0.46 1.08 1.51 43.71 44.97 43.98 42.63 43.82 17 15-9-A 0.00 0.00 0.00 0.00 0.00 80.16 79.88 81.30 79.93 80.32 18 35-2-B 1.58 1.77 1.55 1.66 1.64 29.30 31.19 35.13 30.75 31.59 19 35-2-C 1.28 .06 1.21 1.75 1.57 29.06 31.23 35.55 31.70 31.88 20 35-2-E 1.61 .08 1.67 1.79 1.79 29.01 30.73 29.03 27.13 28.97 21 35-2-F 1.83 2.21 1.83 1.83 1.92 28.20 30.23 25.33 28.14 27.97 22 15-1-G Car s un atisf ctory 25.69 26.06 26.15 25.32 25.80 23 35-1-G Car s un atisf ctory 26.89 25.31 27.46 25.72 26.34 24 35-1 6-G 4.42 6.47 3.72 4.95 4.89 25.94 24.79 26.84 24.53 25.52 25 55-1-G 6.99 6.51 4.95 5.09 5.89 25.58 24.62 25.36 23.16 24.68 26 35-2-G 3.18 4.12 3.20 3.83 3.58 35.29 35.36 33.77 34.30 34.68 27 35-3-G 2.25 2.56 2.24 2.18 2.31 45.39 44.18 45.89 45.04 45.12 28 15-9-G .95 1.04 1.10 1.12 1.05 81.07 79.83 80.91 79.69 80.38 29 15-1-H 2.35 3.77 2.64 3.19 2.99 26.93 25.90 26.64 24.23 25.92 30 35-1-H 3.24 5.17 3.31 3.74 3.86 24.52 24.80 27.60 24.21 25.28 31 55-1-H 3.21 3.29 2.84 3.02 3.09 25.33 24.58 25.82 24.10 24.96 32 35-2-H Car s un atisf ctory 35.00 36.42 37.35 36.23 36.25 33 35-26-H 2.77 2.78 2.47 2.22 2.56 35.94 34.45 37.09 35.81 35.83 34 35-3-H 0.32 2.00 1.07 1.77 1.54 42.51 42.53 45.50 47.25 44.45 35 15-9-H Car s un atisf ctory 83.92 82.87 83.90 87.67 83 35 36 15-1-1 2.20 3.61 1.35 3.35 2.63 23.72 26.34 24.05 25.85 24.99 37 35-1-1 2.50 4.58 1.44 3.29 2.95 20.85 22.61 19.38 21.11 20.99 38 55-1-1 3.16 3.65 2.00 2.93 2.94 17.33 22.10 18.37 19.76 19.39 39 15-9-1 0.00 0.00 0.00 0.00 0.00 82.34 81.77 83.03 82.97 82.53 40 15-4-J 1.27 1.71 1.41 1.65 1.51 26.70 24.03 27.63 23.49 25.46 41 35-4-J 1.39 1.58 0.63 1.51 1.53 23.90 21.92 25.06 21.96 23.21 42 15-15-J 0.00 0.00 0.00 0.00 0.00 72.26 68.31 70.65 67.85 69.76 43 15-2-K 2.10 2.68 4.70 4.70 3.55 29.30 27.51 29.61 27.11 28.38 44 35-2-K 1.99 2.75 4.14 4.33 3.30 28.30 28.22 30.86 26.75 28.53 METHOD OF TESTING AND DATA. 91 TABLE IX. ENGINE PERFORMANCE Continued. EVENTS OF STROKE BY INDICATOR. Release Per Cent of Stroke. Compression Per Cent of Stroke. Labora- tory Right Side. Left Side. Right Side. Left Side. Symbol. Aver- Aver- 1 i H.E. C.E. H.E. C.E. age. H.E. C.E. H.E. C.E. age. 1 2 63 64 65 66 67 68 69 70 71 72 l 15-1 &-V 65.50 70.00 66.00 69.10 67.65 38.00 43.50 42.00 45.00 42.125 2 25-1-V 85.50 70.00 66.00 69.10 67.65 38.00 43.50 42.00 45.00 42.125 3 35-1-V 65.50 70.00 66.00 69.10 67.65 38.00 43.50 42.00 45.00 42.125 4 55-1-V 72.90 68.27 73.31 69.35 70.96 21.96 23.53 22.37 23.75 22.90 5 15-1-A 37.80 72.00 73.00 71.50 71.08 31.00 33.00 35.50 34.00 33.37 6 25-1-A 67.80 72.00 73.00 71.50 71.08 31.00 33.00 35.50 34.00 33.37 7 35-1-A 67.80 72.00 73.00 71.50 71.08 31.00 33.00 35.50 34.00 33.37 8 45-1-A 73.07 73.04 75.07 72.18 73.34 21.89 24.11 20.82 23.96 22.69 9 55-1-A 76.16 74.13 73.35 73.12 74.20 22.56 23.79 21.15 22.96 22.62 10 15-2- A 75.60 75.10 76.00 76.60 75.80 23.50 23.70 20.30 19.40 21.72 11 25-2-A 74.94 72.61 76.00 74.50 74.51 28.12 27.11 26.31 28.44 27.49 12 35-2-A 77.75 78.50 78.00 77.50 77.94 27.75 28.50 28.75 29.50 28.62 13 45-2-A 77.70 78.10 79.20 78.00 78.25 22.60 23.90 25.60 25.90 24.50 14 55-2-A 76.90 77.50 77.60 75.60 76.90 27.40 27.20 27.40 27.30 27.32 15 25-3-A 78.40 79.30 80.90 79.80 79.60 15.80 17.50 16.00 17.20 16.62 16 35-3-A 81.00 78.75 82.06 80.37 80.54 23.29 25.85 23.62 25.21 24.49 17 15-9- A 93.54 93.75 94.10 93.04 93.61 7.82 6.57 8.90 7.54 7.71 18 35-2-B 77.75 80.13 77.30 81.65 79.21 29.38 30.66 29.72 30.22 29.99 19 35-2-C 77.58 78.33 78.06 78.95 78.23 30.78 30.78 25.03 28.20 28.69 20 35-2-E 78.70 79.98 77.35 79.30 78.83 30.10 31.40 28.45 28.95 29.72 21 35-2-F 77.21 78.00 77.69 78.05 77.74 29.93 29.45 30.10 28.45 29.48 22 15-1-G 83.10 61.14 64.32 63.00 62.89 31.67 28.18 31.24 30.29 30.34 23 35-1-G 63.88 62.12 65.90 63.38 63.82 35.15 33.81 33.31 32.03 33.57 24 35-1 6-G 64.12 63.09 85.11 64.08 64.10 33.63 35.71 33.38 34.14 34.21 25 55-1-G 65.53 65.73 35.77 66.61 65.91 34.73 37.18 33.91 35.94 35.44 26 35-2-G 69.89 70.28 72.37 68.63 70.29 26.25 27.37 26.90 25.40 26.'48 27 35-3-G 76.00 75.61 76.89 74.93 75.86 26.04 23.57 24.57 22.93 24.28 28 15-9-G 92.39 92.77 92.30 91.21 92.17 6.84 5.45 7.49 6.39 6.54 29 15-1-H 59.16 59.00 60.04 57.74 58.98 23.67 24.66 24.04 23.25 23.90 30 35-1-H 60.33 57.64 61.84 58.18 59.50 27,99 27.86 28.70 26.88 27.86 31 55-1-H 62.08 59.67 58.95 59.35 60.01 23.44 21.37 20.43 21.65 21.72 32 35-2-H 66.29 67.78 69.10 68.00 67.79 25.29 25.80 24.35 24.73 25.04 33 35-2&-H 65.42 65.27 68.16 65.53 66.09 24.87 23.70 23.78 22.97 23.83 34 35-3-H 71.41 71.65 73.42 71.90 72.10 21.34 21.20 16.44 19.40 19.60 35 15-9-H 92.46 91.47 92.69 91.54 92.04 6.86 6.28 6.65 6.40 6.55 36 15-1-1 47.02 50.82 48.61 51.22 49.42 15.74 19.82 14.60 19.14 17.325 37 35-1-1 49.95 53.05 49.79 52.28 51.27 16.64 19.94 15.00 18.86 17.61 38 55-1-1 53.80 55.33 53.50 50.50 53.28 19.46 19.40 17.87 19.55 19.07 39 15-9-1 89.84 89.08 90.50 89.92 89.84 9.66 9.03 8.76 8.95 9.10 40 15-4-J 71.53 68.74 70.78 67.48 69.63 33.35 31.46 34.48 31.95 32.81 41 35-4^1 71.35 68.21 70.50 69.56 69.90 34.03 31.10 32.68 30.99 32.20 42 15-15-J 91.62 89.25 90.65 89.53 90.26 9.41 8.64 11.01 9.80 9.71 43 15-2-K 70.76 68.70 67.15 64.86 67.87 31.86 28.85 35.58 32.66 32.24 44 35-2-K 72.47 71.30 70.08 68.41 70.57 32.22 32.03 34.47 31.64 32.59 92 LOCOMOTIVE PERFORMANCE. TABLE X. ENGINE PERFORMANCE Continued. PRESSURES ABOVE ATMOSPHERE BY INDICATOR. Initial Pressure, Pressure at Cut-off, Pounds per Square Inch. Pounds per Square Inch. Labora- tory Right Side. Left Side. Right Side. Left Side. Symbol. Aver- Aver- j age. age. 1 H.E. C.E. H.E. C.E. H.E. C.E. H.E. C.E. fc 1 2 73 74 75 76 77 78 79 80 81 82 1 15-1&-V 123.10 126.00 126.00 128.65 125.94 92.11 93.68 103.10 93.41 95.57 2 25-1-V 127.90 132.80 127.20 135.00 130.72 83.10 85.40 88.60 86.90 86.00 3 35-1-V 111.78 115.00 116.70 117.30 115.20 69.80 78.14 75.64 78.80 75.59 4 55-1-V 118.17 123.45 119.12 118.09 119.71 57.12 67.64 67.25 65.95 61.99 6 15-1-A 118.20 120.54 122.88 122.60 121.06 89.11 91.75 91.29 91.81 90.99 6 25-1-A 101.48 100.58 104.42 105.58 103.02 67.04 68.88 70.25 72.46 69.66 7 35-1-A 121.80 120.60 128.10 130.00 125.13 67.33 71.20 70.80 73.10 70.61 8 45-1-A 111.40 111.10 118.00 117.50 114.50 64.70 70.70 64.60 71.60 67.90 9 55-1-A 116.93 120.82 119.20 119.64 119.15 57.43 63.75 54.30 63.86 59.83 10 15-2 A 115.30 122.30 121.60 123.60 120.70 85.20 90.90 84.90 90.80 87.90 11 25-2-A 118.40 117.40 121.50 122.80 120.02 79.60 81.50 83.20 84.20 82.12 12 35-2-A 109.38 119.90 114.94 118.17 115.60 68.65 78.00 72.66 74.19 73.37 13 45-2-A 111.10 116.90 118.60 117.30 115.97 60.10 68.20 60.10 67.20 63.90 14 55-2-A 118.00 118.30 118.10 117.40 117.95 51.60 59.30 52.70 56.40 55.07 15 25-3-A 113.30 119.70 120.80 120.90 118.67 80.00 85.40 84.10 85.60 83.70 16 35-3-A 110.88 122.29 117.04 115.83 116.51 71.08 75.58 73.68 78.66 74.75 17 15-9-A 74.80 75.00 75.13 75.76 75.17 68.03 68.16 70.73 70.33 69.30 18 35-2-B 91.25 93.53 88.19 93.22 91.55 55.75 58.39 53.00 58.47 56.40 19 35-2-C 134.00 140.75 134.46 140.00 137.30 89.67 92.67 86.83 90.75 89.98 20 35-2-E 87.59 89.94 87.44 88.72 88.42 58.82 59.18 62.38 61.72 60.52 21 35-2-F 86.72 87.95 85.67 88.20 87.16 58.78 59.56 65.50 61.00 61.21 22 15-1-G 119.86 119.72 123.30 124.03 121.98 92.94 96.64 98.17 96.75 96.12 23 35-1-G 106.64 107.87 112.56 112.57 109.91 68.44 73.28 71.38 70.50 70.90 24 35-1 6-G 110.67 111.06 116.44 118.37 114.14 71.52 74.44 71.53 75.81 74.08 25 55-1-G 99.50 102.75 113.00 114.30 107.38 60.50 65.00 65.54 68.50 64.88 26 35-2-G 116.35 113.09 119.53 123.40 118.09 69.96 73.64 77.06 77.23 74.47 27 35-3-G 81.21 83.00 83.75 82.43 82.60 48.93 52.18 52.00 51.25 51.09 28 15-9-G 80.00 78.37 78.75 80.00 79.28 70.03 69.96 72.66 72.66 71.33 29 15-1-H 118.36 119.78 123.44 123.53 121.28 89.14 93.54 90.62 93.78 91.77 30 35-1-H 107.81 106.53 114.90 115.21 111.11 66.63 72.35 65.33 69.50 88.45 31 55-1-H Cards unsati sfactor V 56.42 60.58 57.27 59.80 58.52 32 35-2-H 106.21 103.87 108.41 110.08 107.14 60.96 62.29 60.75 60.58 61.14 33 35-26-H 84.41 85.47 88.45 87.91 86.56 47.16 50.69 49.31 47.97 48.78 34 35-3-H 71.63 75.09 75.09 73.25 73.76 45.45 49.91 46.04 45.50 46.72 35 15-9-H 76.21 74.42 75.00 76.44 75.52 64.92 65.36 68.50 68.22 66.75 36 15-1-1 114.51 120.04 118.52 123.22 119.07 88.19 94.27 92.96 93.29 92.18 37 35-1-1 110.65 109.76 115.47 119.35 113.81 74.83 83.35 79.12 82.76 79.87 38 55-1-1 98.91 114.91 111.25 115.58 110.16 69.50 69.41 71.91 73.41 71.06 39 15-9-1 71.72 69.72 69.42 72.16 70.76 63.52 65.61 65.31 64.58 64.76 40 15-4-J 119.77 126.15 123.91 124.50 123.58 93.92 96.97 97.28 99.33 96.82 41 35-4^1 119.16 122.08 127.91 128.91 124.52 80.86 85.58 85.86 87.00 84.83 42 15-15^1 88.59 88.18 88.28 90.59 88.91 78.46 78.75 79.09 82.00 79.58 43 15-2-K 119.42 122.97 122.92 123.13 122.11 91.70 92.47 97.83 96.87 94.72 44 35-2-K 115.50 117.19 113.25 120.36 116.57 71.75 72.03 74.78 78.86 74.36 METHOD OF TESTING AND DATA 93 TABLE XI. ENGINE PERFORMANCE Continued. PRESSURES ABOVE ATMOSPHERE BY INDICATOR. Pressure at Release, Pressure at Beginning of Com- Pounds per Square Inch. pression, Pounds per Square Inch. Labora- tory Right Side. Left Side. Right Side. Left Side. Symbol. Aver- Aver I age. age. I H.E. C.E. H.E. C.E. H.E. C.E. H.E. C.E. fc 1 2 83 84 85 86 87 88 89 90 91 92 1 15-lft-V 32.50 29.53 35.42 30.76 32.05 3.47 2.95 2.66 2.15 2.81 2 25-1-V 28.90 28.06 31.50 27.90 29.09 6.33 3.98 3.94 3.46 4.43 3 35-1-V 23.14 24.50 26.14 23.70 24.37 7.78 8.57 7.50 7.57 7.86 4 55-1-V 15.96 20.64 14.62 18.54 17.44 31.21 30.23 27.12 27.04 28.90 5 15-1-A 29.16 28.79 28.02 27.63 28.40 8.05 8.71 4.58 4.38 6.43 6 25-1-A 20.40 19.33 19.67 20.33 19.93 10.40 9.15 6.00 6.96 8.13 7 35-1-A 20.47 20.20 20.30 20.80 20.44 12.25 11.60 9.30 10.50 10.91 8 45-1-A 14.90 17.30 14.60 17.50 16.10 22.60 23.00 24.90 23.10 23.40 9 55-1 A 13.52 15.45 13.18 15.32 14.37 24.69 27.73 27.30 28.18 26.98 10 15-2- A 38.30 42.80 38.60 39.30 39.70 4.60 7.60 8.60 9.30 7.50 11 25-2-A 32.70 36.70 33.20 34.03 34.16 6.00 7.90 6.80 6.50 6.80 12 35-2-A 25.05 29.00 29.02 27.19 27.57 9.50 12.03 10.90 11.50 10.98 13 45-2-A 21.00 23.70 21.30 24.00 22.50 17.25 21.10 16.10 18.80 18.31 14 55-2-A 19.60 23.00 19.40 22.30 21.07 17.30 21.20 17.70 20.40 19.15 15 25-3-A 40.50 44.40 41.30 42.40 42.15 13.40 13.40 12.30 11.80 12.72 16 35-3-A 34.25 39.95 35.36 38.41 36.99 13.75 16.58 15.12 15.66 15.28 17 15-9-A 56.10 55.73 58.43 57.73 57.00 1.80 2.00 1.77 1.73 1.82 18 35-2-B 18.25 18.53 19.56 17.72 18.51 9.06 9.30 7.81 9.34 8.88 19 35-2-C 31 . 17 34.21 34.88 32.89 33.28 11.79 12.83 10.08 12.42 11.78 20 35-2-E 17.68 18.12 20.00 17.28 18.27 9.30 9.09 9.56 9.67 9.40 21 35-2-F 18.06 18.89 18.67 18.47 18.52 8.28 9.17 7.67 9.53 8.66 22 15-1-G 36.44 42.83 37.44 35.69 38.10 4.22 6.94 5.25 4.88 5.32 23 35-1-G 24.85 27.72 26.76 24.86 25.05 7.68 9.25 9.47 9.13 8.88 24 35-1 6-G 25.59 28.33 26.59 24.97 26.37 9.53 8.73 9.03 7.87 8.79 25 55-1-G 19.91 20.33 20.64 19.20 20.02 14.33 12.42 14.10 12.20 13.26 26 35-2-G 32.25 33.90 33.20 34.13 33.37 11.14 11.25 9.90 10.86 10.78 27 35-3-G 25.93 26.93 27.61 26.78 26.81 11.82 11.36 10.43 10.18 10.95 28 15-9-G 59.16 57.70 61.00 60.28 59.53 2.50 2.67 2.22 2.00 2.35 29 15-1-H 39.50 43.28 38.31 36.91 39.50 12.78 11.41 10.15 10.20 11.13 30 35-1-H 24.42 29.54 26.37 25.75 26.52 10.69 12.04 9.40 9.75 10.47 31 55-1-H 19.33 22.33 22.09 20.20 20.99 20.92 27.25 23.45 21.10 23.18 32 35-2-H 29.25 31.04 30.00 28.71 29.75 9.21 9.66 9.00 8.54 9.10 33 35-2 6 -H 22.25 24.03 23.53 22.75 23.14 9.34 9.97 9.37 8.34 9.25 34 35-3-H 23.86 26.64 24.86 26.33 25.42 7.90 8.68 10.00 8.42 8.75 35 15-9-H 57.03 57.25 59.41 59.19 58.22 2.03 2.31 2.47 1.50 2. OS 36 15-1-1 44.70 51.54 45.93 44.41 46.64 19.70 14.08 15.48 13.00 15.57 37 35-1-1 30.65 34.71 31.18 31.59 32.03 21.76 18.88 19.47 16.35 19.12 38 55-1-1 21.75 26.16 22.66 26.83 24.35 20.41 22.75 19.50 19.00 20.42 39 15-9-1 56.47 58.66 58.00 57.32 57.61 2.18 00.43 1.50 1.26 1.34 40 15-4^1 32.55 36.10 33.31 31.32 33.32 2.10 2.08 2.73 2.51 2.35 41 35-4^1 24.06 25.69 26.85 23.38 24.99 8.03 9.31 8.05 7.35 8.18 42 15-15^7 56.84 57.28 57.68 56.46 57.07 2.71 1.62 2.25 1.81 2.09 43 15-2-K 36.32 37.82 39.60 37.71 37.86 3.96 4.65 3.80 3.83 4.06 44 35-2-K 24.37 25.53 29.03 25.80 26.18 10.19 10.97 9.22 9.78 10.04 LOCOMOTIVE PERFORMANCE. TABLE XII. ENGINE PERFORMANCE Continued. PRESSURES ABOVE ATMOSPHERE BY INDICATOR. Least Back Pressure, Mean Effective Pressure, Pounds per Square Inch. Pounds per Square Inch. i i Labora- tory Right Side. Left Side. Right Side. Left Side. Symbol. Aver- Aver- 1 age. age. 1 H.E. C.E. H.E. C.E. H.E. C.E. H.E. C.E. 1 2 93 94 95 96 97 98 99 100 101 102 1 15-U-V 2.18 1.53 1.22 0.92 1.46 41.04 41.97 47.60 45.12 43.93 2 25-1-V 2.65 2.23 1.88 1.19 1.94 35.35 36.98 43.28 39.31 38.73 3 35-1-V 3.53 3.03 2.90 2.18 2.91 27.27 28.68 35.12 32.29 30.84 4 55-1-V 3.21 4.68 2.12 3.28 3.32 17.29 20.67 19.54 20.19 19.42 5 15-1-A 1.62 1.60 1.06 0.94 1.31 39.95 42.83 46.44 44.76 43.50 6 25-1-A 1.80 2.22 1.60 1.43 1.76 27.05 30.08 33. SC- 32.32 30.83 7 35-1-A 3.70 3.05 2.50 2.77 3.01 26.01 28.40 SI. 08 30.73 29.06 8 45-1-A 2.20 2.90 2.70 2.80 2.70 20.90 22.60 22.30 23.80 22.40 9 55-1-A 2.70 3.39 2.25 2.90 2.81 16.63 17. Gl 17.69 18.33 17.56 10 15-2-A 0.90 0.70 1.20 0.40 0.80 60.93 62.93 62.96 64.32 62.78 11 25-2-A 3.00 2.90 2.47 2.90 2.82 49.12 52.09 53.98 53.85 52.26 12 35-2-A 5.05 6.40 5.16 5.36 5.49 39.38 42.65 44.44 42.21 42.17 13 45-2-A 5.07 5.90 5.10 5.60 5.47 30.69 33.52 32.94 34.14 32.82 14 55-2-A 4.90 6.75 5.10 6.60 5.84 25.33 26.67 26.55 27.10 26.41 15 25-3-A 3.50 3.90 3.40 3.30 3.52 61.18 64.40 66.34 66.14 64.51 16 35-3-A 8.33 10.08 7.91 10.09 9.10 47.86 49.19 50.73 51.14 49.73 17 15-9- A 1.C3 1.75 1.30 0.64 1.18 66.68 66.39 68.68 68.98 67.68 18 35-2-B 3.08 3.41 2.93 2.88 3.07 27.59 28.89 31.13 29.88 29.37 19 35-2-C 6.79 7.41 7.79 6.75 7.19 47.92 49.85 55.13 50.27 50.79 20 35-2-E 3.24 3.35 3.47 2.89 3.24 27.82 28.64 31. 1C 28.63 29.08 21 35-2-F 2.89 2.94 2.46 3.13 2.86 28.33 29.54 30.59 28.83 29.32 22 15-1-G 0.36 1.80 0.80 0.14 0.78 47.48 48.07 49.93 47.69 48.39 23 35-1-G 2.91 2.06 2.15 1.43 2.14 29.71 29.92 33.88 32.23 31.43 24 35-U-G 3.00 2.53 1.56 1.87 2.24 29.75 31.32 33.52 32.69 31.82 25 55-1-G 2.92 3.29 2.95 2.85 3.00 19.63 21.21 23.29 22.72 21.71 26 35-2-G 3.96 5.47 4.00 4.56 4.50 42.89 43.07 47.08 46! 66 44.92 27 35-3-G 6.61 5.11 4.50 4.46 5.17 31.47 34.54 37.15 36.32 34.87 28 15-9-G 0.69 1.43 0.50 0.12 0.69 69.00 69.00 70.41 71.07 70.02 29 15-1-H 0.70 2.23 0.69 0.48 1.01 45.08 46.11 46.55 44.83 45.64 30 35-1-H 1.23 3.75 1.80 1.37 2.04 29.22 30.93 33.57 31.88 31.40 31 55-1-H 3.04 3.29 2.18 2.10 2.65 20.36 19.92 23.55 22.49 21.58 32 35-2-H 3.21 5.33 3.33 3.33 3.80 37.57 38.17 40.82 40.31 39.22 :33 35-2&-H 2.59 2.42 2.34 2.10 2.36 28.17 29.92 31.90 31.20 30.30 :34 35-3-H 3.31 4.31 2.81 3.17 3.40 31.53 33.07 35.27 35.72 33.90 35 15-9-H 0.79 1.54 1.06 0.46 0.96 65.52 65.32 67.44 68.50 66.70 36 15-1-1 1.28 0.17 0.76 0.61 0.71 37.30 44.92 42.17 44.94 42.33 37 35-1-1 2.82 1.97 1.29 1.14 1.81 25.98 33.33 30.71 34.30 31.08 38 55-1-1 4.00 3.66 2.50 3.16 3.33 20.49 24.15 24.37 25.72 '23.68 39 15-9-1 2.10 0.03 1.42 0.92 1.09 59.80 62.24 61.76 63.04 61.71 40 15-4-J 0.63 0.68 1.19 0.36 0.72 50.84 50.13 52.78 49.58 50.83 41 35-4-J 3.00 3.47 3.08 2.08 2.91 34.55 34.94 40.11 37.75 36.84 42 15-15-J 1.71 0.75 0.94 0.70 1.03 73.24 74.10 74.95 76.86 74.79 43 15-2-K 0.92 0.57 1.08 1.00 0.89 52.63 51.80 53.09 50.05 51.89 -44 35-2-K 3.14 3.88 3.25 2.28 3.14 34.17 33.78 40. oe 38.03 36.51 METHOD OF TESTING AND DATA. 95 TABLE XIII. ENGINE PERFORMANCE Continued. WEIGHT OF STEAM SHOWN BY INDICATOR-CARDS. Pounds Steam at Cut-off. Pounds Steam at Release. Labora- tory Right Side. Left Side. Right Side. Left Side. Symbol. 1 Total. Total.. 1 H.E. C.E. H.E. C.E. H.E. C.E. H.E. C.E. 1 2 103 104 105 106 107 108 109 110 Ill 113 1 15-1&-V .26352 .25402 .29059 .26872 1.07685 .26819 .25583 .28372 .26187 1.06961 2 25-1-V .24217 .23532 .25616 .25309 0.98673 .24776 .24565 .26182 .24518 1.00041 3 35-1-V .21107 .21924 .22629 .23399 0.89059 .21660 .22724 .23372 .22223 0.89978 4 55-1-V .18443 .20335 .18946 .19103 0.76827 .11283 .20507 .18726 .19421 0.77937 5 15-1-A .24892 .25295 .26866 .25199 1.02252 .25725 .25780 .26627 .25177 1.03309' 6 25-1-A . 19828 .20173 .21841 .20891 0.82733 .20895 .20489 .21728 .21099 0.84211 7 35-1-A .19993 .20723 .21998 .21062 0.83776 .20969 .21039 .22162 .21423 0.85593 8 45-1-A .17548 .19777 .18948 .19392 0.75665 .18243 .19673 .19062 .19471 0.76850 9 55-1-A .16958 .18517 .17613 .17824 0. 709 12 .18418 .18764 .17742 .18355 0.73279 10 15-2-A .34479 .33852 .34738 .34153 1.37223 .S3426 .35169 .34023 .33524 1.36142 LI 25-2-A .28702 .29871 .31464 .30476 1.20513 .29843 .30696 .3075C .29677 1.20966 12 35-2-A .25858 .27234 .28397 .27162 1.08651 26542 .28001 .28875 .26782 1.10200 13 45-2-A .22603 .25015 .23811 .24589 0.96018 23531 .24814 .24305 .24827 0.97477 L4 55-2-A .21642 .23419 .22518 .23040 0.90619 22533 .24305 .22760 .23225 0.92823 i5 25-3-A .36433 .37349 .38739 .37369 1.49890 35918 .37890 .37711 .36703 1.48222 16 35-3-A .32731 .34660 .34338 .34007 1.35736 32928 .34914 .34288 .34416 1.36546 L7 15-9- A .53454 .52264 .56200 .53284 2.15202 53132 .51815 .55392 .52534 2.12873 18 35-2-B .20289 .20840 .22277 .21062 0.84468 22118 .21961 .22706 .21967 0.88752 19 35-2-C .29154 .29929 .32882 .30377 1.22342 30198 .31095 .32490 ,30679 1.24462 20 35-2-E .20986 .20832 .21765 .20008 0.83591 22054 .21729 .23056 .21191 0.88030 21 35-2-F .20589 .20719 .20484 .20419 0.82211 22042 .21833 .22427 .21748 0.88050 22 15-1-G .26699 .27713 .28967 .26941 1.10320 27554 .29416 .28809 .26464 1.12243- 23 35-1-G .21709 .21799 .23313 .21207 0.88028 .21940 .22444 .23783 .21264 0.89431 24 35-U-G .21879 .21743 .22966 .21698 0.88286 .22404 .23067 .23452 .21528 0.90451 25 55-1-G .19028 .19452 .20578 .19230 0.78288 .19738 .19641 .20383 .19147 0.78909 26 35-2-G .27234 .28099 .28904 .28264 1.12501 .27903 .28393 .29562 .27764 1.13622 27 35-3-G .25516 .25782 .27295 .25677 1.04270 .26112 .26055 .27624 .25613 1.05404 28 15-9-G .55337 .53405 .57241 .54601 2.20584 .54767 .52771 .56316 .53403 2.17257 29 15-1-H .26712 .26851 .27492 .25433 1.06488 .27474 .28710 .27541 .25097 1.08822 JO 35-1-H .19869 .21268 .21850 .20094 0.83081 .20666 .21909 .22327 .20182 0.85084 11 55-1-H .17910 .18410 .18812 .17819 0.72951 .18491 19009 .19235 .17779 0.74514 }2 35-2-H .24364 .25282 .26043 .24477 10016C .25035 .26006 .26621 .24681 1.02343 J3 35-2 6 -H .20597 .20779 .22197 .20418 0.83991 .21008 .21504 .22689 .20791 0.85992 U 35-3-H .22899 .24185 .24796 .24487 0.96367 .23484 .24664 .24857 .24418 0.97423 J6 15-9-H .53796 .52304 .56394 .53606 2.16100 53232 .51711 .55336 .52769 2.13048 16 15-1-1 .24169 .27401 .26128 .26575 1.04272 .24700 .28755 .26234 .25827 1.05515 17 35-1-1 .19346 .22260 .19810 .20822 0.82238 .20145 .22598 .20555 .20868 0.841C7 8 55-1-1 .16121 .18978 .17755 .18210 0.71064 .17473 .19599 .17111 .18308 0.72492 39 15-9-1 .51982 .51815 .53847 .51591 2.09236 .51459 .51416 .53161 .50688 2.06720 10 15-4^1 .27682 .26256 .29949 .26118 1.10005 .28630 .29009 .29032 .25849 1.1252ft 11 35_4_j .22652 .22238 .22927 .22366 0.90183 .23650 .23179 .25199 .22170 0.94199- 12 15-15^1 .54516 .51025 .53423 .52072 2.11037 .52687 .50627 .53081 .49947 2.06341 13 15-2-K .31595 .28291 .29104 .27764 1.16754 .30944 .28079 .30300 .29747 1.19070 44 35-2-K .26119 .23660 .23306 .23175 0.96261 .26092 .22987 .24006 .23853 0.96938 96 LOCOMO TIVE PERFORMANCE. TABLE XIV. ENGINE PERFORMANCE Continued. WEIGHT or STEAM AS SHOWN BY INDICATOR- CARDS. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Pounds of Steam at Compression. Labora- tory Right Side. Left Side. Symbol. Total. H.E. C.E. H.E. C.E. 2 113 114 115 116 117 15-1&-V .00911 .07131 .07116 .07116 .28274 25-1-V .07870 . 07452 .07523 .07504 .30409 35-1-V .08383 .09194 .08910 .09195 .35682 55-1-V .10704 .10986 .10201 .10224 .42115 15-1-A .07287 . 07444 . 06854 . 06388 .27973 25-1-A .08022 .07587 .07351 .07228 .30182 35-1-A .08611 .08363 .08501 .08378 .33853 45-1-A .08761 .09453 .09211 .09352 .36778 55-1-A , . 09399 .10440 .09815 .10197 .39851 15-2-A . 04931 .05679 .05483 .05266 .21359 25-2-A .06009 . 06323 .06070 .06109 .24511 35-2-A .07127 . 07550 .07653 .07768 . 30098 45-2-A .07733 . 08941 . 08373 .08814 .33861 55-2-A .08954 . 09873 .09269 .09610 .37706 25-3-A . 05408 .05797 . 05433 .05368 .22006 35-3-A .07087 .08327 .07682 .07882 .30978 15-9-A .02194 .02112 .02448 .02171 .08925 35-2-B .07276 .07179 .06925 . 07261 . 20641 35-2-C .08366 .08216 .06689 .07746 .31016 35-2-E .07510 .07280 .07217 .07152 .29158 35-2-F .07232 .07006 .07036 .07086 .28360 15-1-G .06049 .06250 .06426 . 05946 .24671 35-1-G .07763 . 07953 .08161 . 07535 .31412 35-U-G . 08090 .08128 . 08036 .07518 .31772 55-1-G .09815 .09605 .09739 .09210 .38369 35-2-G .07106 .07318 .07069 .06775 .28268 35-3-G .07187 . 06553 .06708 .06095 .26543 15-9-G .02169 .02063 .02346 . 02078 . 08656 15-1-H .06931 .06772 .06530 .06155 .26388 35-1-H .07340 .07636 .07280 .06776 . 29032 55-l^H .08827 .09637 .08798 .08295 .35557 35-2-H .06434 .06613 .06364 .06083 .25494 35-2&-H .06379 .06292 .06341 .05717 .24729 35-3-H .05333 . 05479 .05042 . 05060 .20914 15-9-H .02101 .02116 .02350 .02006 . 08573 15-1-1 .06643 .06442 .05723 .06014 .24721 35-1-1 .07150 .07411 .06528 .06614 .27703 55-1-1 .07658 .07232 . 07285 .07327 .29501 15-9-1 .02489 .02207 .02393 .02284 .09373 15-4^1 . 05652 . 05349 .06121 .05500 .22621 35-4-J .07638 .07439 .07551 .06786 .29413 15-15-J .02537 .02332 .02808 .02474 .10151 15-2-K .06534 . 05905 . 05894 . 05441 .23774 35-2-K .08145 .07520 .07900 .08028 .31594 METHOD OF TESTING A ND DATA 97 TABLE XV. ENGINE PERFORMANCE Continued. Indicated Horse -power. Steam Used per I.H.P, per Hour. Dry Coal Used Labora- tory Right Side, Left Side. Measured Measured iSfp. Symbol by by per 1 Total. Tank. Indicator. Hour.* 1 H.E. C.E. H.E. C.E. fc Lbs. Lbs. Lbs. 1 2 118 119 120 121 122 123 124 125 I 15-H-V 44.37 43.98 51.27 47.09 186.71 29.973 19.999 6.903 2 25-1-V 59 . 28 60.10 72.31 63.64 255.33 28.780 19.891 4.355 3 35-1-V 70.17 71.51 90.03 80.20 311.91 27.275 19.480 4.359 4 55-1-V 71.72 83.08 80.75 80.84 316.39 30.621 20.423 5.262 5 15-1-A 44.47 46.20 51.51 48.10 190.28 28.929 19.171 4.449 6 25-1-A 48.33 52.08 60.27 55.74 216.42 28.062 29.288 4.188 7 35-1 -A 68.52 72.52 81.57 78.15 300.76 26.937 19.710 4.184 8 45-1 -A 71.81 75.24 76.34 78.94 302.33 28.604 19.476 4.329 9 55-1-A 70.15 71.98 74.34 74.64 291.11 30.645 21 . 069 5.119 10 15-2-A 64.81 64 86 66.72 66 04 262.43 27 . 661 20.233 4.187 11 25 -2 -A 86.47 88.86 94.67 91 51 361.51 26.595 20.432 4.449 12 35-2-A 103.24 108.35 116.06 L06.82 434.47 26.286 21.021 4.543 13 45-2-A 104.20 110.28 111.41 111.89 437.78 28.445 21.457 5.600 14 55-2-A 106.93 109.10 111.66 110.44 438.13 31.997 23.094 6.031 15 25-3-A 108.97 111.14 117.71 113.71 451.53 28.593 21 . 657 5.081 16 35-3-A 121.35 120.86 128.15 125.17 495.53 30.103 23.492 6.323 17 15-9- A 72.90 70.33 74.80 72.79 290.82 39.162 33.343 6.896 18 35-2-B 71.71 72.76 80.61 74.97 300.05 28.821 22.646 5.130 19 35-2-C 125.32 126.32 143.62 126.90 522.16 24.861 20.354' 5.120 20 35-2-E 74.52 74.34 83.15 74.02 306.03 27.922 22.410 4.465 21 35-2-F 73.12 73.88 78.65 71.82 297.47 27.184 22.523 4.324 22 15-1-G 53.09 52.09 55.62 51.48 212.28 29.263 20.062 4.205 23 35-1-G 78.03 75.14 88.64 81.71 324.52 28.827 20.420 4.664 24 35-H-G 78.08 79.66 87.64 82.82 328.21 28.696 20.408 4.463 25 55-1-G 81.92 85.78 96.83 91.53 356.06 29.096 20.665 5.293 26 35-2-G 110.77 107.79 121.13 116.33 456.02 26.967 21.023 5.084 27 35-3-G 82.53 87.78 97.06 91.95 359.31 32.220 25.030 5.615 28 15-9-G 75.08 72.75 76.32 75.28 299.43 38.593 32.966 6.871 29 15-1-H 50.65 50.20 52.11 48.62 201.58 30.032 19.980 3.926 30 35-1-H 76.59 78.56 87.65 80.66 323.46 28.602 19.753 4.814 31 55-1-H 82.65 78.36 95.23 88.13 344.37 29.725 19.972 4.651 32 35-2-H 100.14 98.58 108.39 103.71 410.82 30.058 21.683 5.173 33 35-26-H 74.46 76.63 84.00 79.60 314.69 30.719 22.378 5.149 34 35-3-H 84.42 85.81 94.07 92.31 356.61 32.735 24.982 5.956 35 15-9-H 72.42 69.97 74.26 73.09 289.74 39.476 33.924 7.320 36 15-1-1 41.64 48.60 46.90 48.44 185.59 32.943 21.140 37 35-1-1 68.39 85.04 80.55 87.17 321.15 29.366 21.317 38 55-1-1 84.22 96.21 99.81 102.08 382.34 29.383 19.603 39 15-9-1 65.42 65.98 67.32 66.57 265.30 40.865 35.393 40 15-4^1 58.29 55.70 60.29 54.88 229.16 28.520 19.563 41 35-4-J 90.86 89.02 105.06 95.82 380.76 26.772 19.458 42 15-15^1 79.53 77.91 81.05 80.51 319.02 35.130 29.040 43 15-2-K 0.01 54.81 59.71 56.94 231.47 28.810 20.310 44 35-2-K 84.21 81.61 99.50 91.54 356.87 27.820 20.580 * To obtain an aoproximate measure of the performance of the locomotive when using West Virginia or Pittsburgh coal, multiply column 125 by 0.8. LOCOMOTIVE PERFORMANCE. TABLE XVI. ENGINE PERFORMANCE (Continued]. RESULTS FROM INDICATOR- CARDS. Labora- Weight Steam per Revo- Weight Mixture in Cyl- inder Per Cent of Mix- ture Present Per Cent of Mix- ture Present Reevap- oration per Conden- sation per Reevap- oration I H P Conden- sation I if T p tory lution per Revo- Revo- j Symbol. by Tank. Revo- lution. as Steam at as Steam at lution. lution. per Hour. Hour. I Cut-off. Release. 1 Lbs. Lbs. Lbs. Lbs. Lbs. Lbs. 1 2 126 127 128 129 130 131 132 133 1 15-1&-V 1.19011 1.47285 73.11 72.62 .00724 .18267 2 25-1-V 1.00752 1.31161 75.23 76.27 0.01367 0.39060 3 35-1-V 0.76024 1.11706 79.73 80.55 .00919 0.32970 4 55-1-V 0.53705 0.95820 80.18 81.34 .01110 0.63293 5 15-1-A 1.13690 1.41663 72.18 72.22 .01057 0.39944 6 25-1-A 0.78156 1 . 08338 76.37 77.73 .01478 0.53073 7 35-1-A 0.70709 1.04562 80.12 81.86 .01817 0.70708 8 45-1-A 0.57873 0.94651 79.94 81.18 .01185 0.58587 9 55-1-A 0.48625 0.88476 80.15 82.82 .02367 1.49286 10 15-2-A 1.56924 1.78283 76.97 76.36 .01081 . 19032 11 25-2-A 1.25553 1.50064 80.30 80.69 .00453 0.09615 12 35-2-A 1.00165 1.30263 83.41 84.60 .01549 0.40628 13 45-2-A 0.84323 1.18184 81.24 82.48 .01459 0.49207 14 55-2-A 0.76357 1.14063 79.44 81.37 .02204 0.92330 15 25-3-A 1.66629 1 . 88635 79.46 78.58 .01668 .26461 16 35-3-A 1.35278 1.66256 81.64 82.13 .00810 0.18023 17 15-9-A 2.39542 2.48467 86.61 85.67 .02329 .38076 18 35-2-B 0.76518 1.05159 80.32 84.40 . 04284 1.61392 19 35-2-C 1.14142 1.45159 84.28 85.74 .02120 0.46180 20 35-2-E 0.73353 1.02511 81.55 85.87 .04439 1.68960 21 35-2-F 0.72044 1 . 004C4 81.88 87.70 .05839 2.20343 22 15-1-G 1.27736 1.524C7 72.38 73.65 .01923 0.44054 23 35-1-G 0.81905 1.13317 77.68 78.92 .01403 0.49379 24 35-1 6-G 0.82510 1.14282 77.25 79.15 .02165 0.75295 25 55-1-G 0.57080 0.95449 82.02 82.57 .00621 0.31655 26 35-2-G 1.09486 1.37754 81.67 82.48 .01121 0.27611 27 35-3-G 1.01506 1.28049 81.43 82.32 .01134 0.35995 28 15-9-G 2.44208 2.52864 87.23 85.92 .03327 .52578 29 15-1-H 1.23895 .50283 70.86 72.41 .02334 0.56587 30 35-1-H 0.81161 .10193 75.39 77.21 .02003 0.70587 31 55-1-H 0.57981 0.93540 77.99 79.66 .01563 0.80128 32 35-2-H 1.06528 .32022 75 . 87 77.52 .02177 0.61425 33 35-26-H 0.84097 .08826 77.18 79.02 .02001 0.73092 34 35-3-H 1.00251 .21165 79.53 80.41 .01056 0.34481 35 15-9-H 2.37938 2.46511 87.66 86.43 .03052 .50635 36 15-1-1 1.25908 1 . 04869 82.81 83.80 .01243 0.32532 37 35-1-1 0.82364 1.10067 74.71 76.46 .01942 0.69249 38 55-1-1 0.64438 0.93939 75.64 77.16 .01427 0.65088 39 15-9-1 2.27697 2.37070 88.26 87.19 .02516 .45132 40 15-4-J 1.31060 1.53681 71.58 73.22 .02516 0.54740 41 35-4^J 0.89140 1.18553 76.07 79.46 .04015 1.20500 42 15-15^1 2.37320 2.47470 85.28 83.38 .04696 .69510 43 15-2-K 1.35150 1.58920 73.47 70.49 .02316 0.49379 44 35-2-K 0.91551 1.23150 78.17 78.18 . 00677 0.02132 METHOD OF TESTING AND DATA. 99 TABLE XVII. LOCOMOTIVE PERFORMANCE. Machine Friction of Engine I.H.P. per Square in Terms of Foot of Draw- Mechan- bar Pull Draw- ical Labora- Equiva- bar or Effi- 1 tory Symbol. lent to Average ME.P. Average M.E.P. Draw- bar Pull I.H.P. Dynamo- meter H.P. ciency of Engine. Heating Surface. Grate Surface. fc Lbw. Lbs. Per Cent. 1 2 134 135 136 137 138 139 140 141 I 15-U-V 4,806 5.29 579 22.5 164.2 87.9 .154 10.8 2 25-1-V 4,236 5.29 579 34.9 220.4 86.2 .210 14.8 3 35-1-V 3,375 5.29 579 53.5 258.4 82.8 .256 17.1 4 55-1-V 2,123 5.29 579 86.3 230.1 72.7 .260 18.3 5 15-1-A 4,759 5.29 579 23.1 167.2 87.9 .157 11.0 6 25-1-A 3,372 5.29 579 37.1 179.3 82.9 .178 12.5 7 35-1 -A 3,179 5.29 579 54.8 246.0 81.8 .248 17.4 8 45-1-A 2,449 5.29 579 71.4 230.9 76.4 .249 17.5 9 55-1-A 1,921 5.29 579 87.7 203.4 69.9 .239 16.9 10 15-2- A 6,867 4.18 458 17.5 244.9 93.2 .226 15.2 11 25-2-A 5,717 4.18 458 28.9 332.6 92.0 .297 20.9 12 35-2-A 4,615 4.18 458 43.1 391.4 90.1 .357 25.2 13 45-2-A 3,590 4.18 458 55.8 382.0 87.3 .359 25.4 14 55-2-A 2,889 4.18 458 69.5 368.6 84.1 .361 25.3 15 25-3-A 7,057 3.57 391 25.0 426.5 94.5 .370 26.1 16 35-3-A 5,441 3.57 391 35.6 459.9 92.8 .407 28.7 17 15-9-A 7,409 1.69 185 7.3 283.5 97.5 .239 16.8 18 35-2-B 3,214 4.18 458 42.8 257.3 85.7 .247 17.4 19 35-2-C 5,558 4.18 458 43.0 479.2 91.8 .430 30.2 20 35-2-E 3,180 4.18 458 44.1 261.9 85.6 .252 17.7 21 35-2-F 3,209 4.18 458 42.4 255.1 85.7 .244 17.2 22 15-1-G 5,283 5.29 579 23.2 189.1 89.1 .174 12.3 23 35-1-G 3,439 5.29 579 54.6 269.9 83.2 .267 18.8 24 35-1 6-G 3,482 5.29 579 54.6 273.6 83.4 .270 18.9 25 55-1-G 2,375 5.29 579 86.8 269.3 75.6 .293 20.6 26 35-2-G 4,915 4.18 458 42.5 413.5 90.7 .375 26.4 27 35-3-G 3,814 3.57 391 36.8 322.5 89.8 .296 20.8 28 15-9-G 7,658 1.69 185 7.2 292.2 97.6 .246 17.4 29 15-1-H 4,997 5.29 579 23.4 178.2 88.4 .166 11.7 30 35-1-H 3,436 5.29 579 54.5 269.0 83.1 .266 18.6 31 55-1-H 2,362 5.29 579 84.4 260.0 75.5 .284 20.0 32 35-2-H 4,291 4.18 458 43.8 367.0 89.3 .338 23.8 33 35-26-H 3,315 4.18 458 43.5 271.2 86.2 .259 18.2 34 35-3-H 3,707 3.57 391 37.6 319.0 89.5 .294 20.7 35 15-9-H 7,299 1.69 185 7.3 282.4 97.5 .238 16.8 36 15-1-1 4,632 5.29 579 23.2 162.4 87.3 .153 10.8 37 35-1-1 3,401 5.90 646 61.0 260.2 81.1 .264 18.6 38 55-1-1 2,591 5.90 646 92.9 289.4 75.6 .315 22.2 39 15-9-1 6,752 1.69 185 7.3 258.0 97.3 .219 15.4 40 15-4^J 5,562 5.29 579 23.8 205.4 89.7 .189 13. a 41 35-4^1 4,031 5.50 602 56.8 324.0 85.1 .313 22.1 42 15-15^1 8,185 2.00 219 8.8 310.2 97.2 .263 18.5 43 15-2-K 5,678 4.90 536 21.8 209.7 90.6 .191 13.4 44 35-2-K 3,995 4.90 536 48.0 308.9 86.6 .294 20 7 100 LOCOMOTIVE PERFORMANCE. TABLE XVIII. LOCOMOTIVE PERFORMANCE (Continued). Steam Used Dry Coal B.T.U. Used by Engine. Coal Used D.H.P. Developed per Square Foot. Labora- tory D!H!P. T) TT P J Symbol. Hour. j-'.jn. ST. per Hour.* Per Minute. Per I.H.P. per Per D.H.P. per Run. Heating Surface. Grate Surface. s Minute. Minute. 55 Lbs. Lbs. Lbs. 1 2 143 143 144 145 146 147 148 149 1 15-U-V 34.08 4.86 107,890 5,778 6,570 55.88 .135 9.5 2 25-1-V 33.34 5.08 141,610 5,446 6,425 49.36 .181 12.9 3 35-1-V 32.92 5.24 164,275 5,267 6,357 39.32 .212 14.9 4 55-1-V 42.11 6.95 185,707 5,869 8,071 29.80 .189 13.4 5 15-1-A 32.92 4.69 106,250 5,584 6,354 56.66 .138 9.7 6 25-1-A 33.87 4.90 116,805 5,397 6,514 37.78 .148 10.4 7 35-1-A 32.93 5.14 156,083 5,189 6,344 35.55 .203 14.3 8 45-1-A 37.45 5.95 166,053 5,492 7,191 28.28 .190 13.4 9 55-1-A 43.86 7.01 170,938 5,872 8,404 26.22 .167 11.8 10 15-2-A 29.64 4.48 139,656 5,322 5,702 76.69 .202 14.2 11 25-2-A 28.91 4.77 184,636 '5,107 5,551 67.82 .274 19.3 12 35-2-A 29.18 5.19 219,844 5,060 5,616 56.00 .322 22.7 13 45-2-A 32.59 6.06 238,677 5,452 6,248 53.60 .315 22.1 14 55-2-A 38.03 7.59 267,981 6,102 7,270 46.47 .304 21.3 15 25-3-A 30.27 5.77 248,051 5,493 5,815 95.60 .351 24.7 16 35-3-A 32.44 6.79 285,931 5,770 6,217 91.74 .379 26.7 17 15-9- A 40.17 7.13 218,483 7,512 7,706 136.19 .233 16.4 18 35-2-B 33.62 5.36 166,387 5,545 6,466 44.05 .213 14.9 19 35-2-C 27.09 5.23 251,132 4,809 5,228 76.02 .395 27.8 20 35-2-E 32.63 5.22 164,943 5,389 6,298 37.94 .216 15.3 21 35-2-F 31.71 4.97 156,202 5,251 6,125 37.00 .210 14.8 22 15-1-G 32.85 4.77 119,319 5,620 6,310 59.26 .156 11.0 23 35-1-G 34.66 5.65 179,366 5,527 6,645 42.78 .222 15.6 24 35-U-G 34.42 5.63 180,721 5,506 6,605 41.43 .225 15.9 25 55-1-G 38.38 6.53 198,585 5,577 7,374 33.52 .222 15.6 26 35-2-G 29.74 5.48 235,413 5,162 5,680 66.64 .341 24.0 27 35-3-G 35.89 6.40 221,649 6,168 6,873 57.12 .266 18.7 28 15-9-G 39.55 7.07 221,657 7,403 6,169 140.38 .241 16.9 29 15-1-H 33.97 4.87 115,673 5,738 6,491 52.29 .147 10.3 30 35-1-H 34.39 5.57 177,190 5,478 6,587 44.10 .222 15.6 31 55-1-H 39.37 5.89 196,619 5,709 7,562 29.30 .214 15.1 32 35-2-H 33.90 6.20 236,330 5,752 6,439 59.19 .302 21.3 33 35-2&-H 35.64 5.91 185,912 5,907 6,855 45.51 .223 15.7 34 35-3-H 36.60 6.57 223,693 6,272 7,012 58.89 .261 18.5 35 15-9-H 40.50 7.19 219,289 7,568 7,783 142.46 .233 16.4 36 15-1-1 37 35-1-1 38 55-1-1 39 15-9-1 40 15-4-J 41 35-4^1 42 15-15-J 43 15-2-K 44 35-2-K * To obtain an approximate measure of the performance of the locomotive, when using West Virginia or Pittsburgh coal, multiply the values of column 143 by 0.8. METHOD OF TESTING AND DATA. 101 TABLE XIX. LOCOMOTIVE PERFORMANCE (Continued). The values of this table were computed on the assumption of uniform fire conditions. The process involved is developed in Chapter VI. Equiva- lent Coal Burned Water /L __ Equiva- lent B.T.U. Labora- Evapo- ration from Coal Fired per Square Foot T^k) Evapo- Evapo- ration from Coal I. Si. P. Coal D.H!P. Taken up by Boiler j I tory Symbol. and at 212 F. Hour. Hour. Grate Surface Hour. rated per Pound of Coal. and at 212 per Pound of Coal. per Hour. Hour. per Pound of Coal. fc Lbs. Lbs. Lbs. Lbe. Lbs. Lbs. Lbs. 1 2 150 151 153 153 154 155 156 157 1 15-1&-V 6,762 798.0 46.2 7.04 8.44 4.270 4.86 8,151.3 2 25-1-V 8,864 1,113.1 64.5 6.62 7.93 4.359 5.05 7,658.8 3 35-1-V 10,286 1,349.3 78.9 6.32 7.59 4.325 5.22 7,330.4 4 55-1-V 11,650 1,590.3 92.1 6.10 7.27 5.026 6.91 7,021.5 5 15-1-A 6,660 791.6 .45.8 6.98 8.37 4.160 4.73 8,083.7 6 25-1-A 7,318 876.6 50.8 6.95 8.31 4.050 4.89 8,025.8 7 35-1-A 9,751 1,261.2 73.0 6.44 7.71 4.193 5.13 7,446.3 8 45-1-A 10,395 1,367.3 79.2 6.34 7.56 4.522 5.92 7,300.1 9 55-1-A 10,690 1,416.8 82.1 6.31 7.49 4.867 6.97 7,253.8 10 15-2-A 8,745 1,093.1 63.3 6.66 7.96 4.165 4.46 7,687.7 11 25-2-A 11,555 1,590.0 92.1 6.06 7.23 4.398 4.78 6,982.7 12 35-2-A 13,733 2,028.6 117.7 5.64 6.75 4.669 5.18 6,519.1 13 45-2-A 14,926 2,301.0 133.4 5.42 6.46 5.256 6.02 6,239.1 14 55-2-A 16,773 2,802.3 162.2 5.01 6.02 6.396 7.78 5,814.1 15 25-3-A 15,515 2,444.4 141.6 5.29 6.02 5.414 5.73 6,103.8 16 35-3-A 17,878 3,099.0 179.6 4.82 5.75 6.254 6.74 5,553.3 17 15-9- A 13,670 2,040.9 118.2 5.59 6.67 7.018 7.20 6,441.8 18 35-2-B 10,404 1,371.5 79.5 6.32 7.56 4.571 5.33 7,300.4 19 35-2-C 15,709 2,496.0 144.6 5.21 6.27 4.780 5.21 6,055.5 20 35-2-E 10,324 ,355.3 78.5 6.32 7.58 4.428 5.17 7,320.7 21 35-2-F 9,785 ,263.4 73.2 6.42 7.71 4.247 4.95 7,446.3 22 15-1-G 7,476 900.0 52.1 6.93 8.27 4.239 4.76 7,987.1 23 35-1-G 11,226 ,517.0 87.9 6.18 7.36 4.674 5.62 7,108.3 24 35-1 6-G 11,312 ,532.2 88.8 6.16 7.34 4.668 5.60 7,089.0 25 55-1-G 12,425 ,747.5 101.2 5.94 7.07 4.908 6.49 6,828.2 26 35-2-G 14,725 2,256.0 130.7 5.46 6.50 4.947 5.46 6,277.7 27 35-3-G 13,866 2,052.6 113.1 5.65 6.72 5.712 6.36 6,490.2 28 15-9-G 13,869 2,049.0 118.8 5.65 6.72 6.843 7.01 6,490.2 29 15-1-H 7,248 866.5 50.2 7.01 8.32 4.298 4.86 8,035.2 30 35-1-H 11,090 1,490.5 86.3 6.22 7.40 4.608 5.54 7,146.9 31 55-1-H 12,302 1,675.8 97.1 6.12 7.34 4.866 6.45 7,052.2 32 35-2-H 14,769 1,959.8 113.6 6.31 7.50 4.770 5.34 7,243.5 33 35-26-H 11,633 1,593.0 92.3 6.08 7.26 5.062 5.87 7,011.7 34 35-3-H 13,993 2,080. 6 120.7 5.62 6.69 5.834 6.52 6,461.2 35 15-9-H 13,752 2,031.5 117.8 5.64 6.72 7.012 7.19 6,490.2 36 15-1-1 37 35-1-1 38 55-1-1 39 15-9-1 40 15-4-J 41 35-4-J 42 15-15-J 43 15-2-K 44 35-2-K II. LOCOMOTIVE PERFORMANCE, A TYPICAL EXHIBIT. CHAPTER V. LOCOMOTIVE PERFORMANCE AS AFFECTED BY CHANGES IN SPEED AND CUT-OFF. 29. Purpose. It is the purpose of the present chapter to define and summarize certain fundamental facts concerning cylinder per- formance. It shows, by means of data derived from experiments, the effect of changes in speed and cut-off upon the power and efficiency of a locomotive when running under a wide-open throttle. 30. The Tests which form the basis for this discussion, twelve in number, are those which, in the tabulated record of Chapter IV., are designated as Series A. They cover practically all conditions of speed and cut-off at which it is possible to operate continuously the experi- mental locomotive with a wide-open throttle. Five tests were run with the reverse lever in the first notch, giving a cut-off of approxi- mately 25 per cent, the speeds being 15, 25, 35, 45 and 55 miles per hour, respectively; five were run with the reverse lever in the sec- ond notch, giving a cut-off of approximately 35 per cent, the speeds being the same as for the shorter cut-off, and two were run with the reverse-lever in the third notch, giving a cut-off of approximately 45 per cent., the speeds being 25 and 35 miles per hour, respectively. The steam pressure, the setting of the valves, the fuel used, and the fireman were the same for all tests. Excepting as to speed and cut- off, which were as specified above, the conditions affecting all tests were as nearly as possible identical. Table XX. presents the laboratory symbol of the several tests and a summary of the more important controlling conditions. 102 CHANGES IN SPEED AND CUT-OFF. 103 TABLE XX. ESSENTIAL CONDITIONS. Speed. Cut-off. Steam Pressure. 1 H a g ||] JS.S o jD O o 53 stfzu |l 1 PH PH < f Brazil ] 15-1-A 15.00 80.7 1 25 125.9 124.6 Wide open { Indiana 1- Chas. Reyer [ Block J 25-1-A 24.07 129.5 1 25 120.0 113.4 " " " 35-1-A 35.49 191.0 1 25 129.7 127.2 " ' ' 11 45-1-A 46.29 249.1 1 25 128.8 124.9 " " M 55-1-A 56.83 305.7 1 25 124.9 121.3 ' ' ' ' " f Brazil 1 15- 2- A 14.33 77.1 2 35 129.5 125.1 Wide open { Indiana \ Chas. Reyer L Block J 25-2-A 23.72 127.6 2 35 129.3 124.9 " " 35-2-A 35.31 190.0 2 35 131.6 120.7 t c c c " 45-2-A 45.74 246.1 2 35 126.7 121.1 c c " " 55-2-A 56.81 306.0 2 35 124.0 118.8 ( C C I ' ' f Brazil 1 25-3-A 24.00 129.1 3 45 127.2 123.2 Wide open { Indiana } Chas. Reyer I Block J 35-3-A 34.15 183.8 3 45 125.3 122.1 31. The Valves and their Setting. The slide-valves used in this series of tests had f" outside lap, -%" inside lap, and a maximum travel of 5.57". Their setting was that which, after considerable experimenting, had been found to produce the most efficient results- When the cut-off was 25 per cent of the stroke, they gave ^y lead. At full stroke, the lead was nil. (See A setting, Chapter III.) 32. Indicator-cards. In Fig. 61 are presented average cards f :om the left side of the locomotive for each test. These cards, when studied in connection with the results derived from them, will serve to disclose the effect upon the distribution of steam in the cylinder, of changes in speed and cut-off. Comparisons along vertical lines show the effect of changes in speed at constant cut-off, along hori- zontal lines, of changes in cut-off at constant speed. Examining the cards along either of the vertical lines, as, for example, along that which represents a cut-off of 25 per cent, it appears 104 LOCOMOTIVE PERFORMANCE. that as the speed is increased, the size of the card is diminished. At very high speeds the card is comparatively small. This change occurs notwithstanding the fact that there is no change in cut-off or in any other event of the stroke. It is due wholly to wiredrawing past the valves. While the extent of port opening may be unaffected by changes in speed, the period of opening is reduced as the speed is 45 I 25 a 15 188 135 81 25 35 45 Cut-off -Notches and Per Cent of Stroke FIG. 61. increased, from which it necessarily follows that there is admitted less steam per stroke as the speed is increased. A further discussion of this matter will be found in Chapter XVII., which deals especially with valve-gears. It will be sufficient for the present to observe that the size of the card is controlled to a limited extent only by the cut- off. Thus, by reference to Fig. 61, it will be found easy to select cards representing a cut-off of 35 per cent which are actually smaller than others representing a cut-off of 25 per cent. CHANGES IN SPEED AND CUT-OFF. 105 An interesting illustration of the influence of speed upon the size or the card is afforded when an attempt is made to operate a locomotive with the reverse-lever in its extreme forward position at high speed. Fig. 62 represents two cards, both taken with the reverse-lever in its extreme forward position. The dotted line is the normal card taken at low speed, the full line a card taken at 35 miles an hour. It will be seen that the total pressure range for the card at speed 'is divided into three parts, more than a third representing the loss in passing 130 100- 50 , M.E.P. 52.3 FIG. 62. from boiler to cylinder, a quarter from cylinder to exhaust, and approximately a third only the mean effective pressure. It is evident that speed as well as cut-off have an important influence on mean effective pressure. In this connection, also, attention should be called to the changed form of the card which results from change of speed (Fig. 61). At low speeds the events of the stroke are rather clearly marked, but as the speed is increased these become less and less distinct. These statements will be sufficient to show how much more complicated is the study of locomotive performance than is that of stationary engines designed to run at fixed speeds, for in locomotive service the element of speed appears as an ever-present influence affecting other factors; for example, the mean effective pressure, the distri- bution of steam, the thermodynamic efficiency, and even the machine friction. 33. Events of the Stroke. In the early work upon the loco- motive it was assumed that a fixed location of the reverse-lever would serve to maintain all events of the stroke constant. It was soon found, however, that the cut-off, for example, could not be depended upon to remain constant even though the reverse-lever was not moved, but that a change in the speed of the locomotive or in the lubrication of its valves was quite sufficient to materially affect the time of action of the valves. This is to be -accounted for by the fact that the width of port-opening is small, the force required 106 LOCOMOTIVE PERFORMANCE. to move the valve is considerable, and the mechanis mwhich drives it is not sufficiently stiff nor so free from lost motion as to impart to the valve a motion which is absolutely positive. The precise values representing the several events of the stroke for each of the twelve tests under consideration, as obtained from indicator-cards taken under the conditions defined, are given in Table XXI. Cards taken at high speeds do not show clearly the point of admission, the value of which is therefore omitted from the table. TABLE XXI. EVENTS OF STROKE AS DETERMINED FROM INDICATOR-CARDS. The values given are the averages for the four cylinder-ends. Per Cent of Stroke. D ' of Test. Admission. Cut-off. Release. Beginning of Compression. 15-1-A 3.25 24.68 71.08 33.37 25-1-A 3.25 24.68 71.08 33.37 35-1-A 3.25 24.68 71.08 33.37 45-1 -A 2.15 22.43 73.34 22.69 55-1 -A 22.81 74.20 22.62 15 -2- A 1.50 37.80 75.80 21.71 25-2-A 1.53 34.37 74.51 27.49 35-2-A 1.70 33.69 77.94 28.62 45-2-A 1.75 33.07 78.25 24.50 55-2-A 35.40 76.90 27.32 25-3 -A 1.32 44.22 79.60 16.62 35-3-A 1.51 43.82 80.54 24.49 34. Wiredrawing between boiler and cylinder in a locomotive is unavoidable, the extent of its influence in any given locomotive depending upon the, rate at which steam is used. The areas of the wide-open throttle, steam-pipe, branch pipes, etc., for locomotive Schenectady No. 1 are shown graphically by Fig. 54, Chapter III. The facts concerning the drop in the pressure of the steam from the boiler to the exhaust-pipe, with related data of interest for the twelve tests under consideration, are shown by Table XXII. 35. Mean Effective Pressure, as affected by speed and cut-off, is shown diagrammatically by Fig. 63. If the values of this figure are compared along vertical lines, 'the very striking effect resulting from changes in speed may be seen. With a given cut-off each CHANGES IN SPEED AND CUT-OFF. 107 increment in speed reduces the amount of work done per stroke, and while values measuring such changes depend somewhat upon the size of ports and the characteristics of the valve-gear employed, the presence of such a change is unavoidable in the action of a loco- motive. (Chapter XVII.) TABLE XXII. DROP IN STEAM PRESSURE. WIDE-OPEN THROTTLE. EXHAUST TIP DOUBLE 3" DIAM. Desig- nation of Test. 15-1-A 25-1-A 35-1-A 45-1- A 55-1 A Steam Ex- hausted per Hour. Lbs. Maximum Port- opening, Inches. Pressure of Steam in Steam. Ex- haust. Boiler. Branch Pipe. 124.6 113.4 127.2 124.9 121.3 Cylinders At Cut-off. At Release . At Com- pression. Least Back. 5,505 6,073 8,102 8,648 8,921 .20 .20 .20 .20 .20 .92 .92 .92 .92 .92 125.9 120.0 129.7 128.8 124.9 91.0 69.7 70.6 67.9 59.8 28.4 19.9 20.4 16.1 14.4 6.4 8.1 10.9 23.4 26.9 1.3 1.8 3.0 2.7 2.8 15-2-A 25-2-A 35-2-A 45-2-A 55-2-A 7,259 9,614 11,420 12,451 14,019 .25 .25 .25 .25 .25 .97 .97 .97 .97 .97 129.5 129.3 131.6 126.7 124.0 125.1 124.9 120.7 122.1 181.8 87.9 82.1 73.4 63.9 55.0 39.7 34.2 27.6 22.5 21.1 7.5 6.8 11.0 18.3 19.1 0.8 2.8 5.5 5.5 5.8 25-3-A 35-3-A 12,910 14,917 .30 .30 1.02 1.02 127.2 125.3 123.2 122.1 83.8 74.7 42.1 37.0 12.7 15.3 3.5 9.1 Another fact of interest in connection with Fig. 63 is that the values given cover practically the entire range of action under which the experimental locomotive can be operated with a wide-open throttle. An attempt to run a test at 15 miles and 10" cut-off resulted in so high a tractive power that the drivers slipped, while a test at 45 miles and 10" cut-off could not be run because of the failure of the boiler to supply steam. 36. The Indicated Horse-power for different speeds and cut-offs is shown by Fig. 64. Here it will be seen that at constant cut-off, the power increases with increase of speed up to a certain point, after which it remains practically constant. In the case of the locomo- tive experimented upon, the maximum power was practically reached at a speed of 35 miles an hour. The limit, however, as applied to 108 LOCOMOTIVE PERFORMANCE. locomotives in general will depend upon the proportion of the cylinders, the diameter of the drivers, and the capacity of the boiler, and while this point may not always be as clearly defined as it is by the data from Schenectady No. 1, its presence will appear in all locomotives which are designed to run at high speeds of rotation. It will be seen that the highest power reached was 496 horse. Prior to the tests herein described, no one knew what was the maximum power of a locomotive under constant operating con- 3 Cut-Off Reverse Lever Notch FIG. 63. Mean Effective Pressure. ditions, though it was not uncommon for locomotives of the size of Schenectady No. 1 to be credited with as high as 800 horse-power. The facts presented by Fig. 64 and the fuller exhibit of Table XV. suggest that such an estimate is too high, though if a superior grade of coal had been used instead of the light and friable Brazil block, it is probable that the maximum could have been made to approach 600. It may easily be shown, by an analysis based on the dimensions of the locomotive and the experimental facts herein presented, that for all speeds below 18 miles an hour the cut-off, CHANGES IN SPEED AND CUT-OFF. 109 with a fully open throttle, is limited by the adhesion of the drivers : : for speeds above 18 miles the limit is found in the capacity of the boiler to supply steam. The cut-off which at any given speed will serve under a fully open throttle to operate the experimental loco- motive at 500 horse-power is that shown by the curve, Fig. 64. 3 Reverse Lever Notch Cut-Off FIG. 64. Indicated Horse-power. It is well, also, to review the values of Fig. 64 with reference to a. matter which is always of interest to locomotive designers, namely,, that of controlling speed by the reverse-lever vs. the throttle. The reverse-lever quadrant of Schenectady No. 1 was notched at intervals of { of an inch, and the first three notches forward from the center gave 25, 35, and 45 per cent cut-off respectively. Fig. 64 discloses the fact that at all speeds a change of one notch in the reverse-lever position makes a very marked alteration in the power output, the differences ranging from 70 to 146 horse-power. Between the first and second notches there is an increase of about 50 per cent over the power developed with the lever in the first notch. Evidently no very fine gradation of power, and consequently of speed, can be secured by 110 LOCOMOTIVE PERFORMANCE manipulating the reverse-lever of this locomotive. Quadrants of the modern locomotive are, however, much more closely notched, and as, in present-day practice, locomotives are allowed to go as fast as they will whatever load may be attached to them, the question outlined now has much less significance than formerly. 37. The Steam Consumption per horse-power hour for different speeds and cut-offs is shown numerically by Fig. 65 and graphically 55-fMin -[Ho] Cut-Off Reverse Lever Notch FIG. 65. Steam per I.H.P. per Hour. by Fig. 66. Engineers unfamiliar with the performance of the loco- motive have often characterized it as an extremely wasteful engine, whereas from Fig. 65 it appears that its performance compares favor- ably with that of any other class of single-cylinder, non-condensing engine. With open throttle, the consumption of steam per indicated horse-power does not under any conditions of speed or cut-off exceed 32 pounds, and under favorable conditions it falls to about 26 pounds. In this connection it should be noted that Schenectady No. 1 carried but 140 pounds pressure, and that the tests were run at about 130. When favored by a higher pressure, this engine has given one horse- CHANGES IN SPEED AND CUT-OFF. Ill power on a consumption of less than 25 pounds of steam per hour. The results show that the minimum consumption is obtained at the 35 per cent cut-off, or, say, one-third stroke ; this for an engine carrying 140 pounds of steam. The losses resulting from the employment of a shorter cut-off are, however, slight as compared with those which attend a lengthening of the cut-off. An exception to this statement is r however, to be found at high speed, where the 25 per cent cut-off is FIG. 66. Steam per I.H.P. per Hour. better than the 35 per cent. It should be said, also, that experiments involving higher steam pressures go to show that the most economical cut-off for all speeds is somewhere between one-quarter and one-third stroke. It appears, also, whatever may be the cut-off, that, beginning with slow motion, a gradual increase of speed is attended by a de- crease in the amount of steam required up to a certain point, after which the consumption increases. Both an inspection of the values of Fig. 65 and a glance at the curves of Fig. 66 show the most economi- 112 LOCOMOTIVE PERFORMANCE. cal performance to have been obtained at a speed of but 35 miles per hour, or 188 revolutions per minute. In explanation of this fact, it would appear that below this limit any increase of speed is of advan- tage through reduced cylinder condensation, an advantage which is roften assumed to attend all increase of piston speed, but above this limit some other influence enters which is so strong in its effect as to more than neutralize the advantage of the higher piston speed. This neutralizing influence is without doubt the wiredrawing, the presence of which becomes more and more marked as the speed rises. 38. Critical Speed. In the two paragraphs immediately preceding, attention has been called to the fact that the maximum power and the maximum efficiency of locomotive Schenectady No. 1 were found at a speed of approximately 35 miles an hour, or 190 revolutions per minute. To give significance to the conditions involved, the author has called that speed at which the power and efficiency of a locomotive become maximum the " critical speed." It will, of course, not always be found at 35 miles an hour, for much depends upon the diameter of drivers, but it will be found not far from 200 revolutions per minute. With higher steam pressures the limit is probably raised somewhat, .and with a superior valve-gear the point may not be quite so well marked as in the data, but the preceding statement will be found substantially true in its application to all simple locomotives. 39. Cylinder Condensation. The effect of changes in speed and f cut-off upon the percentage of the total steam used, which is shown rby the indicator, is given by Fig. 67. Two facts are clearly presented iby this figure. The first is that the percentage of the steam used -which is accounted for by the indicator, is greatest at the critical speed. This appears to be true for all cut-offs. The data well illustrate a rather commonly accepted theory, that for a given cut-off anything ~which tends to suppress cylinder condensation improves the perform- ance of the engine. Fig. 67 shows that either increasing the speed .above or diminishing it below 35 miles an hour increases the conden- sation, and it has already been shown (Fig. 65) that similar changes operate to increase the water consumption of the engine. A second fact, which is apparent from Fig. 67, is that at constant ^speed any increase of cut-off increases the percentage of steam shown by the indicator, or diminishes the condensation. The data show some exceptions to this rule, but they involve values which are small. The tendency of the exhibit is clear. It will be elsewhere shown (throttling test, Chapter XX.) that by throttling and making the CHANGES IN SPEED AND CUT-OFF. 113 cut-off very late in the stroke more than 90 per cent of all the steam used is shown by the indicator. Dealing with the more detailed action which marks the inter- change of heat between the steam and the walls of the cylinder, atten- tion is first called to the percentage of mixture present as steam at cut-off. (Fig. 68.) The difference between the values given and 100 will represent the percentage of the mixture which is water. Here Reverse Lever Notch FIG. 67. Percentage of Total Steam shown by Indicator. again, it appears that at any given cut-off the percentage of the mixture which is present as steam is nowhere greater than at the critical speed. Other values of interest are shown by Table XXIII. Thus, the last- four columns of this table show the extent of the reevaporation during expansion when the cut-off is constant. This reevaporation in all cases increases with increase of speed, while at constant speed length- ening the cut-off diminishes the reevaporation. For example, referring to the 25-mile series, the test with the reverse-lever in the first notch gave a reevaporation per revolution of .0148; with the reverse-lever 114 LOCOMOTIVE PERFORMANCE. in the second notch the reevaporation decreased to .0045 pound; and with the reverse-lever in the third notch all reevaporation dis- appeared, and the condensation amounted to .0167 pound. The weight of mixture in the cylinder per revolution and the weight exhausted (Table XXIII.) shows, when compared, the char- acter of the exhaust action at different speeds. For example, refer- ring to the series for which the reverse-lever was in the first notch, at 15 miles an hour, of the 1.4 pounds in the cylinder, 1.1 pounds cs per Hour Revolutions per Minute 35 80.1 79.4 o 43 79.9 81.2 42 4 33 80.1 KB 4 81.6 , i S8 1 i 25 76.4 80.3 79.5 i 35 g I 13 72.2 77.0 Si- 25 35 45 p jrqen of Stroke 2 Reverse Lever Notch 3 Cut-Off FIG. 68. Percentage of Total Steam shown by Indicator at Cut-off. were exhausted, leaving 0.30 of a pound during compression. When the speed is increased to 55 miles per hour, of the 0.88 of a pound in the cylinder, but 0.49, or a trifle more than half, is exhausted. It is, however, worthy of note that while the relative amount of mixture retained during compression is much greater at the higher speed, the actual increase is small since the weight of mixture during compression at the higher speed is but 0.39. A more detailed exhibit of data derived from these tests, relating to the interchange of heat, will be found in Chapter IV., Series A. CHANGES IN SPEED AND CUT-OFF. 115 TABLE XXIII. THERMAL ACTION WITHIN THE CYLINDERS. Designa- tion of Test. Weight of Mixture in Cylinder per Revolution. Lbs. Weight of Mixture Exhausted per Revolution. Lbs. Reevaporation. Condensation. Per Revolution. Lbs. Per I.H.P. per Hour. Lbs. Per Revolution. Lbs. Per I.H.P. per Hour. Lbs. 15-1-A 25-1 -A 35-1-A 45-1-A 55-1-A 15-2-A 25-2-A 35-2-A 45-2-A 55-2-A 1.4166 1.0834 1.0456 0.9465 0.8848 1.1369 0.7816 0.7071 0.5787 0.4862 0.0106 0.0148 0.0182 0.0118 0.0237 0.3994 0.5307 0.7071 0.5859 1.4929 .7828 .5006 .3026 .1818 .1406 1.5692 1.2555 1.0016 0.8432 0.7636 0^0045 0.0155 0.0146 0.0220 6!o96i 0.4063 0.4921 0.9233 0.0108 0.1903 25-3-A 35-3-A .8863 .6626 1.6663 1.3528 0^0081 0^1802 0.0167 0.2646 40. Boiler Performance. Since the performance of the boiler is fully treated in another portion of this work, it is unnecessary to present an elaborate discussion in this connection. It will, however, serve in the further discussion of the performance of the locomotive, if some attention be given the limitation upon its capacity. This is well set forth in Fig. 69. The values of this figure show the pounds of water from and at 212 degrees, which were required to be evapo- rated each hour to sustain the operation of the locomotive under the conditions of speed and cut-off indicated. The values on shaded background represent an hourly evaporation of more than 12 ; 500 pounds, and correspond to a rate of power which approaches closely to the maximum of the boiler when fired with Indiana block coal. With a better fuel it is probable that a total of 18,000 can be evapo- rated, an amount which is equal to 15 pounds of water per foot of heating surface per hour. Accepting this latter value as representing the maximum capacity of the boiler, then the points of cut-off and speed, for which the boiler will serve to meet the demands of the cylinders under a fully open throttle, will be represented by the curve in Fig. 69. Thus, at 25 miles steam can be supplied for a cut-off of over 55 per cent; while at a speed of 50 miles the cut-off cannot be much greater than 45 per cent. An attempt to employ longer cut-offs will result in failure through lack of steam. This curve 116 LOCOMOTIVE PERFORMANCE. represents maximum performance, under a wide-open throttle, for the engine in question. It shows, also, the effect of changes in speed upon the total consumption. Thus, comparing tests for which the cut-off is the same, it will be seen that the demand upon the boiler is not greatly increased by increasing the speed. The reason for this has already been explained, but it will be of interest to note that the upper portion of the curve of maximum performance is so nearly vertical that, after the speed has reached 25 miles, but slight reduc- Reverse Lever Notch 'FiG. 69. Equivalent Evaporation per Hour. tions in the cut-off are required to enable the boiler to meet the demands of the cylinders under further increments of speed. Doub- ling the speed from 25 miles to 50 miles only requires the cut-off to be reduced from 55 per cent to 45 per cent; whereas, if it were not for the wire-drawing action, it might be expected that doubling the speed would necessitate the use of half the cut-off. The pounds of water evaporated per pound of coal, as actually obtained for the tests, are given by the points in Fig. 70, and, as rectified by the use of the equation E= 10.08 -.296H. CHANGES IN SPEED AND CUT-OFF. 117 by the line in the same figure. A full statement of observed and cal- culated data, derived from the boiler as a result of the twelve tests under consideration, will be found in the recorded data, Chapter IV. 41. Performance of the Locomotive as a Whole. Dealing first with the locomotive as a power-plant, consisting of boiler and engines, its performance is briefly set forth by Table XXIV. and Fig. 71. From these exhibits it appears that an indicated horse- power is under most conditions of operation obtained on a consump- i 11 10 I 3 I 1 2 3 4 5 6 7 8 9 10 11 12 13 " Equivalent Evaporation per Sq. Ft. Heating Surface per Hour 14 FIG. 70. Evaporative Efficiency. tion of from 4 to 6 pounds of Brazil block coal. With higher grades of coal it should be possible to lower the record by about one-fifth, making the range 3.2 to 4.8 pounds. These latter values may fairly be used in comparing the performance of a simple locomotive opera- ting under ordinary conditions with steam-plants of other types. The rate at which power is developed is well shown by the table Each foot of heating surface in the boiler yields from .15 to .47 of a horse-power, and each foot of grate surface from 11 to 28 horse- power. Results which will approach these in value are not to be found in any other class of service. 118 LOCOMOTIVE PERFORMANCE. TABLE XXIV. BOILER AND ENGINE PERFORMANCE. Coal per I.H.P. Indicated Horse-power Designation of per Hour. Test. Lbs. Per Square Foot of Heating Surface. Per Square Foot of Grate Surface. 15-1-A 4.16 0.157 10.90 25-1-A 4.05 0.178 12.40 35-1-A 4.19 0.248 17.23 45-1-A 4.52 0.249 17.27 55-1-A 4.87 0.239 16.63 15-2-A 4.16 0.226 14.99 25-2-A 4.40 0.297 20.65 35-2-A 4.67 0.357 24.87 45-2-A 5.26 0.359 25.01 55-2-A 6.40 0.361 25.03 25-3-A 5.41 0.370 25.80 35-3-A 6.25 0.407 28.31 TABLE XXV. PERFORMANCE AT THE DRAW-BAR. Designation of Test. Steam Used per D.H.P. per Hour. Lbs. Dry Coal Used per D.H.P. per Hour. Lbs. Dynamometer Horse-power Per Square Foot of Heating Surface. Per Square Foot of Grate Surface. 15-1-A 25-1-A 35-1-A 45-1-A 55-1-A 32.92 33.87 32.93 37.45 43.86 4.69 4.90 5.14 5.95 7.01 0.138 0.148 0.203 0.190 0.167 9.7 10.4 - 14.3 13.4 11.8 15-2-A 25-2-A 35-2-A 45-2-A 55-2-A 29.64 28.91 29.18 32.59 38.03 4.48 . 4.77 5.19 6.06 7.59 0.202 0.274 0.322 0.315 0.304 14.2 19.3 22.7 22.1 21.3 25-3-A 35-3-A 30.27 32.44 5.77 6.79 0.351 0.379 24.7 26.7 Between the cylinders and the draw-bar there is, of course, some loss of power (Chapter XIX.), which, in the case of a locomotive on a testing-plant, amounts only to the friction of the machine. Since, however, it is the whole purpose of a locomotive to exert force at CHANGES IN SPEED AND CUT-OFF. 119 the draw-bar, it is of importance to possess some measure of its per- formance in terms of force exerted at the draw-bar. Table XXV. gives a record of the weight of steam and coal used per hour in developing a horse-power at the draw-bar. The coal used is also shown diagrammatically by Fig. 72. The table gives also the horse- power developed at the draw-bar (dynamometer horse-power) per foot of heating surface and per foot of grate surface respectively. 2 3 Reverse Lever Notch Cut-Off FIG. 71. Coal per I.H.P. per Hour. The effective pull exerted at the draw-bar is shown by Fig. 73 and the equivalent horse-power by Fig. 74. The horse-power equiva- lent to machine friction is shown by Fig. 75. A correct understand- ing of the general effect of increments of speed upon the forces exerted at the draw-bar is a matter of great importance to any one inter- ested in locomotive performance. The effects are so well shown by the figures referred to as to make them well worthy of careful study. The coal used per mile run is shown by Fig. 76. 42. Maximum Power Dependent upon Efficiency. It has already been stated that anything which operates to increase the LOCOMOTIVE PERFORMANCE. 3 Cut-Off Reverse Lever Notch FIG. 72. Coal per D.H.P. per Hour. Miles per Hour Ti 1342 2431 1870 3132 e 40 2600 4157 5050 88 >JU n^ 2793 5259 6&66 i ao ~ 1 I'er ?.er t< f S tr( ke 3 Cut-Off FIG. 73. Draw-bar Pull. Reverse Lever Notch CHANGES IN SPEED AND CUT-OFF. 121 Reyolutions pe^JMinute 203.4 368.6 % ") oc 5-j T IB 230.9 382.0 JO ^ MS 246.0 391.4 459.9 1 10 OB 179.3 332.6 426.5 j JB l ~, 167.2 244.9 g l 2 \ 8 I 4 5 Iferpetit ojf S,trc ke 1 2 - 3 Reverse Lever Notch Cut-Off FIG. 74. Dynamometer Horse-power. 87.7 H~ 69.5 71.4 55.8 35 54.8 43.1 35.6 -188 * 37.1 28.9 25.0 135 23.1 17.5 81 25 35 45 Percent of Stroke 2 3 Cut-Off FIG. 75. Friction Horse-power. Reverse Lever Motch 122 LOCOMOTIVE PERFORMANCE. efficiency of a locomotive may, at the normal limit of power, be utilized in the development of more power. The importance of this statement entitles it to some further consideration. It is evident that the maintenance of pressure in the cylinders demands steam from the boiler, and, hence, that the limit of cylinder work is reached when the boiler can no longer meet the demand which is made upon it. If now some improvement in fire-box or ; 3 . Reverse Lever Notch Cut-Off FIG. 76. Coal per Mile Run. boiler construction is adopted which will increase the evaporative efficiency of the boiler, making it possible to evaporate more water in return for each pound of fuel burned, then, at the sacrifice of the saving, it will become possible to establish a new and a higher limit in the capacity of the boiler. The more efficient boiler can be made to deliver more steam, which, in turn, may be utilized in the development of more power. Again, assuming the boiler performance to remain unchanged, an improvement in the cylinder action will be found to give a simi- lar result. For example, with a definite maximum quantity of CHANGES IN SPEED AND CUT-OFF. 123 steam, the cylinders under normal conditions can deliver a certain maximum power. If now, by the adoption of compound cylinders, or of some other improvement, an arrangement can be had which delivers a horse-power upon the consumption of less steam than before, then the definite maximum quantity of steam which can be delivered by the boiler will not be entirely consumed in developing the normal cylinder power, and the surplus can be employed in developing more power. In a similar manner it can be shown that any improvement which will increase the efficiency of any essential detail of the machine establishes a credit which, when the locomotive is pushed to its high- est performance, may be invested in an increase of power. There is, therefore, a twofold purpose to be served by any device or arrangement which adds to the efficiency of any important ele- ment of the locomotive: first, under normal conditions of opera- tion, that of saving fuel, and, second, when the need is for the largest possible output of work, that of raising the maximum output of power. III. THE BOILER. CHAPTER VT. BOILER PERFORMANCE.* 43. Selection of Data. The tabulated data in Chapter IV. con- stitute the results obtained from forty-four efficiency tests of the experimental locomotive. It will be noted that the first thirty-five tests were all made within a period of two years between November 23, 1894, and February 11, 1897. The discussion in this chapter is based on these thirty-five tests. Other tests were made earlier and later than these, and some of the earlier results were published ; f but inasmuch as the earlier results were for the most part obtained at a time when the mounting mechanism of the testing-plant was being perfected, and when the accessory apparatus for detecting errors was limited, they do not possess the same degree of reliability as those chosen for the present purpose. During .the early existence of the plant the conditions affecting the operation of the boiler could not be maintained with the same degree of constancy as was after- ward possible, and, to add to the irregularities, different men were employed in firing, which as a study of the results has shown, intro- duced a personal variable. The results of a considerable number of later tests were destroyed by fire. The analysis which follows will show that the experimental results are in most respects consistent. Some of the facts which at first appear to be irregular will be accounted for, while others * This discussion was presented as a paper before the American Society of Mechan- ical Engineers in December, 1900, forming part of Vol. XXII of the Transactions. f Tests of the Locomotive at the Laboratory of Purdue University, Trans. M. E., Vol. XIV. 124 BOILER PERFORMANCE. 125 must be passed over as at present impossible of explanation. It is probable, however, that a full knowledge of all the conditions affecting the performance of the locomotive boiler may serve to> make clear those phenomena which are at present inexplicable. 44. The Boiler. A complete description of the boiler under consideration, together with a statement of all dimensions, will be found presented in Chapter III. The boiler, during its stay in the laboratory, suffered little or no change of condition. The water used carries some solid matter in solution, but it does not form hard scales, unless it is allowed to remain in the boiler for a considerable period. The boiler of Schenectady was never neglected, and at the close of its service in the laboratory it was clean and otherwise in as good condition as when work was commenced upon it. 45. General Conditions (Table XXVI). To those accustomed to reviewing data derived from stationary boilers, the duration of the tests (Column 4) will appear insufficient. The arguments sus- taining the practice involving short tests for locomotives on a test- ing-plant were presented to the committee having in charge the revision of the American Society of Mechanical Engineers' code relative to a standard method of conducting boiler-tests, and will be found in the published correspondence of that committee.* The tests herein recorded were run several years in advance of the presen- tation of the committee's final report, and for this reason it will be of interest to call attention to the fact that fifteen out of the thirty- five tests perfectly fulfill the requirements of the present code. In defense of those which are shorter than allowed by the code, it should be said that the writer's opinion on the subject was embodied in a recommendation to the committee to the effect that the limit apply- ing in such cases be the burning of a total for the test of not less than 100 pounds of coal per foot of grate surface. The committee adopted the general form of the recommendation, but fixed the limit at 250 pounds. For two of the tests presented the total fuel per foot of grate is slightly below 150 pounds, and for four others it is below 200. The boiler pressure (Column 5) was practically the same for afT tests save two, and for these it was intentionally higher than the nor- mal in the one case and lower in the other. Pressures were observed from an ordinary dial-gauge at five-minute intervals, and also recorded * Proceedings Am. Soc. Mech. Eng'rs, Vol. XXI, p. 112. 126 LOCOMOTIVE PERFORMANCE. TABLE XXVI. GENERAL CONDITIONS. The several tests represented in this table are arranged in order of the rate of evaporation, No. 1 representing the test for which the rate is least, and No. 35 that for which it is greatest. Identification of Test. Average Pressure, Pounds per Square Inch. Average Temp., Deg. Fahr. Dura- tion of tt Test in Consecuti Numbe Labora- tory Designa- tion. Date of Test. Min- utes. Steam Pressure in Boiler by Gauge. Atmos- pheric Pressure. Absolute Steam Pressure. Of Labora- tory. Of Feed- water. 1 2 3 4 5 6 7 8 9 1 15-1-A Dec. 12, 1894 240 125.9 14.5 140.4 65 53.2 2 15-U-V Nov. 23, 1894 190 127.3 14.5 141.8 74 53.8 3 15-1-H Dec. 9, 1896 180 123.6 14.3 137.9 72 54.6 4 25-1 -A Dec. 14, 1894 255 120.1 14.5 134.6 67 53.8 5 15-1 G Nov. 9, 1896 180 123.6 14.4 138.0 69 53.3 6 15-2- A Nov. 13, 1895 180 129.5 14.4 143.9 79 55.2 7 25-1-V Nov. 26, 1894 240 127.1 14.3 141.4 70 53.9 8 35-1-A Dec. 17/1894 180 129.7 14.6 144.3 75 53.2 9 35-2-F Jan. 21, 1895 180 155.4 14.7 170.1 71 52.7 10 35-1-V Dec. 7, 1894 140 128.9 14.3 143.2 69 51.9 11 35-2-E Jan. 16, 1895 180 128.1 14.5 142.6 71 51.9 12 45-1- A ; Nov. 20, 1895 150 128.8 14.3 143.1 76 56.4 13 35-2-B Jan. 14, 1895 170 98.4 14.4 112.8 66 50.0 14 55-1-A Nov. 25, 1895 120 124.9 14.2 139.1 76 56.0 15 35-1-H Dec. 18, 1896 160 121.2 14.6 135.8 67 52.6 16 35-1-G Nov. 20, 1896 170 125.0 14.6 139.6 74 55.0 17 35-1&-G Dec. 2, 1896 170 128.2 14.6 142.8 78 52.5 18 25-2-A Oct. 25, 1895 180 129.3 14.4 143.7 85 56.0 19 55-1-V Dec. 18, 1895 120 128.4 14.4 142.8 76 58.1 20 35-26-H Feb. 10, 1897 160 122.6 14.5 137.1 74 50.6 21 55-1-H Feb. 11, 1897 120 127.5 14.4 141.9 71 52.1 22 55-1-G Nov. 23, 1896 120 126.9 14.5 141.4 79 54.2 23 15-9- A Nov. 6, 1896 150 122.5 14.3 136.8 79 55.6 24 15-9-H Dec. 11, 1896 180 122.7 14.4 137.1 81 53.2 25 35-2-A Dec. 19, 1894 180 123.0 14.5 135.7 72 52.5 26 35-3-G Dec. 4, 1896 140 125.9 14.3 140.2 70 52.6 27 15-9-G Nov. 12, 1896 160 124.1 14.5 138.6 77 53.1 28 35-3-H Dec. 14, 1896 120 116.5 14.3 130.8 73 53.0 29 35-2-G Nov. 13, 1896 160 121.0 14.5 135.5 78 53.6 30 35-2-H Dec. 16, 1896 120 112.0 14.6 126.6 72 53.0 31 45-2-A Nov. 18, 1895 140 126.7 14.3 141.0 84 55.0 32 25-3-A Nov. 1, 1895 122.5 127.2 14.5 141.7 76 53.3 33 35-2-C Jan. 23, 1895 120 143.3 14.4 157.7 71 51.7 34 55-2-A Nov. 22, 1895 68 124.0 14.4 138.4 76 58.4 35 35-3-A Nov. 15, 1895 120 125.3 14.3 139.6 77 55.5 BOILER PERFORMANCE. 127 by a Bristol gauge. To better show the fluctuations in pressure, the chart of this and other recording-gauges used about the loco- motive had a rapid motion, making a complete revolution in six hours. A sample chart is presented as Fig. 77. 46. Actual Evaporation (Table XXVII). This table shows (Columns 10 and 11) the regularity with which water was delivered FIG. 77. Chart from Recording-gauge showing Boiler Pressure. to the boiler; the total pounds delivered to the injector during the test (Column 12); the pounds caught from injector overflow (Column 13) ; the pounds received by the boiler (Column 14) ; and the rate at which the evaporation proceeded (Column 15). In all tests an effort was made to keep the injector constantly in action, but for those of low power the rate of delivery could not be made sufficiently small to permit this being done. The data for a few of the tests show the injector to have been started two or more times, while its 128 LOCOMOTIVE PERFORMANCE. TABLE XXVII. ACTUAL EVAPORATION. The several tests represented in this table are arranged in order of the rate of evaporation, No. 1 representing the test for which the rate is least, and No. 35 that for which it is greatest. Identification Water and Steam. of Test. Consecutive Number. Labora- tory Designa- tion. Dura- tion of Test in Min- utes. Number of Times Injector was Started. Minutes One or Both In- jectors were in Action. Total Pounds of Water Delivered to In- jectors. Pounds of Water Lost from Injector Overflow. Pounds of Water De- livered to Boiler and Presum- ably Evap- orated. Pounds of Water Evap- orated per Hour. 1 2 4 10 11 12 13 14 15 1 15-1-A 240 22 130 22,415 315 22,100 5,525 2 15-1 &-V 190 14 129 17,925 135 17,790 5,618 3 15-1-H 180 7 126 18,349 126 18,223 6,075 4 25-1-A 255 10 236 25,9.35 100 25,895 6,093 5 15-1-G 180 12 127 18.768 72 18.696 6.232 6 15-2-A 180 9 149 21,885 44 21,841 7,280 7 25-1-V 240 7 213 29,508 32 29,476 7,369 8 35-1-A 180 2 180 24,468 100 24,368 8,123 9 35-2-F 180 9 149 24,793 460 24,333 8,111 10 35-1-V 140 1 136 19,899 19,899 8,528 11 35-2-E 180 3 175 25,742 45 25,697 8,566 12 45-1-A 150 2 146 21,688 15 21,673 8,669 13 35-2-B 170 8 158 24,728 173 24,555 8,666 14 55-1-A 120 1 120 17,970 87 17,883 8,941 15 35-1-H 160 1 160 24,793 70 24,723 9,271 16 35-1-G 170 1 170 26.598 35 26,563 9,375 17 35-1 b -G 170 1 170 26,753 10 26,743 9.439 18 25-2-A 180 1 180 28,911 5 28,906 9,635 19 55-1-V 120 1 120 19,526 108 19,418 9.709 20- 35-26-H 160 1 160 25,834 5 25,829 9,686 21 55-1-H 120 1 120 20,514 20,514 10,257 22 55-1-G 120 1 120 20.791 30 20,761 10,381 23 15-9 A 150 1 150 28,582 60 28,522 11,409 24 15-9-H 180 1 180 34,403 30 34,373 11,458 25 35-2-A 180 1 180 34,351 29 34,325 11,442 26 35-3-G 140 1 140 27,060 27,060 11,597 27 15-9-G 160 1 154 30,877 5 30,872 11,577 28 35-3-H 120 1 120 23,446 60 23,386 11,693 29 35-2-G 160 1 160 32,886 40 32,846 12,317 30 35-2-H 120 3 115 24,793 60 24,733 12,367 31 45-2-A 140 1 140 29,313 213 29,100 12,472 32 25-3-A 122 1 122 26,415 14 26,401 12,931 33 35-2-C 120 3 120 26,034 25 26,009 13,004 34 55-2-A 68 1 68 16,188 277 15,911 14,039 35 35-3-A 120 1 120 29,984 109 29,875 14,937 BOILER PERFORMANCE 129 period of action is recorded as coincident with the length of the test. This apparent inconsistency is explained by the fact that it was sometimes convenient to change from one injector to the other during the progress of the tests, and at other times both injectors were in action at the same time. Column 10 merely shows the number of starts made, and includes the record for both injectors. The last column of this table discloses something of the significance of the conditions attending the action of the locomotive boiler. For example, test 35 shows that nearly 15,000 pounds of water were delivered to the boiler and presumably evaporated each hour, or, approximately, 250 pounds per minute. This rate of very nearly a barrel a minute is sufficient to evaporate an amount of water equal to the full water capacity of the boiler in 34 minutes. At this rate, had the injectors ceased in their action, the water level would have fallen between the upper and the middle gauges at the rate of one inch each minute. 47. Quality of Steam and Equivalent Evaporation (Table XXVIII). The quality of steam, assuming dry saturated steam to be unity, is shown by Column 16, and the percentage of moisture by Column 17. Results were obtained by the use of a throttling calorimeter of an approved form, taking steam from a perforated pipe extending horizontally into the dome of the boiler at a point A, Fig. 35 (Chapter III.) , near the throttle-opening. Observations were made at five-minute intervals. An examination of the table will show , that, in general, the amount of moisture in the steam increases as the rate of evaporation is increased, though variations in individual results are so great that they do not fall in any well-defined line. The reason for apparent inconsistencies is to be looked for either in the methods employed, or in actual variations in the performance of the boiler. The methods employed were the same for all tests, while the condition of the boiler was necessarily subject to change. The boiler had to be frequently washed. It is conceivable that dryer steam may be furnished by the boiler when newly washed than when in a condition requiring washing, though there is nothing in the data either to sustain or to discredit such a conclusion. It has, however, been shown by Professor Jacobus * that almost all moisture which may be intermixed with steam passing a horizontal pipe separates itself * Tests to show the Distribution of Moisture in Steam when flowing in a Hori- zontal Pipe, Transactions of the American Society of Mechanical Engineers, Vol. XVI, p 1017. 130 LOCOMOTIVE PERFORMANCE. TABLE XXVIII. QUALITY OF STEAM AND EQUIVALENT EVAPORATION. The several tests represented in this table are arranged in order of the rate of evaporation, No. 1 representing the test for which the rate is least, and No. 35 that for which it is greatest. Identification of Test. Results of Calorimeter Tests. Equivalent Evaporation from and at 212 Fahr. per Hour. Dura- tion of Test in Quality of Steam in Assuming Assuming all Water Deliv- Consec- utive Labora- tory Min- utes. Dome of Boiler, Dry, Percentage of Moisture Quality of Steam as ered to Boiler to have been Num- Desig- Saturated in Steam. Shown by Completely ber. nation. Steam being Column 16. Evaporated Taken as into Dry . Unity. Steam. 1 2 4 16 17 18 19 1 15-1-A 240 .9951 0.49 6,659 6,683 2 15-U-V 190 .9937 0.63 6,763 6,794 3 15-1-H 180 .9838 1.62 7,249 7,337 4 25-1-A 255 .9922 0.78 7,318 7,360 5 15-1-G 180 .9894 1.06 7,476 7,535 6 15-2-A 180 .9924 0.76 8,745 8,796 7 25-1-V 240 .9932 0.68 8,865 8,910 8 35-1-A 180 .9917 0.83 9,771 9,832 9 35-2-F 180 .9900 1.00 9,785 9,855 10 35-1-V 140 .9937 0.63 10,284 10,332 11 35-2-E 180 .9932 0.68 10,324 10,377 12 45-1-A 150 .9913 0.87 10,395 10,463 13 35-2-B 170 .9931 0.69 10,413 10,467 14 55-1-A 120 .9878 1.22 10,691 10,788 15 35-1-H 160 .9851 1.49 11.090 11,214 16 35-1-G 170 .9886 1.14 11,226 11,322 17 35-1 6-G 170 .9860 1.40 11,312 11,430 18 25-2-A 180 .9910 0.90 11,554 11,633 19 55-1-V 120 .9912 0.88 11,624 11,700 20 35-2&-H 160 .9880 1.20 11,634 11,737 21 55-1-H 120 .9871 1.29 12,303 12,422 22 55-1-G 120 .9869 1.31 12,426 12,548 23 15-9-A 150 .9906 0.94 13,670 13,766 24 15-9-H 180 .9871 .29 13,722 13,853 25 35-2-A 180 .9894 .06 13,735 13,843 26 35-3-G 140 .9838 .62 13,867 14,035 27 15-9-G 160 .9871 .29 13,869 14,002 28 35-3-H 120 .9873 .27 13,995 14,128 29 35-2-G 160 .9856 .44 14,725 14,884 30 35-2-H 120 .9866 .34 14,771 14,932 31 45-2-A 140 .9876 .24 14,926 15,065 32 25-3-A 122.5 .9889 .11 15,515 15,644 33 35-2-C 120 .9930 .70 15,709 15,788 34 55-2-A 68 .9887 .13 16,773 16,903 35 35-3 -A 120 .9889 .11 17,883 18,032 BOILER PERFORMANCE. 131 from the steam by gravitation, and forms a rill in the bottom of the pipe, the steam above being approximately dry. Experiments by the writer, involving a visual examination of the steam-space of a boiler while in action, revealed no haze or mist above the surface of the water, thus sustaining the conclusion that the steam w r ithin the steam-space of the boiler is ordinarily dry and saturated; the water is present as water and the steam as steam. They are not intermixed. If this is true, the steam which passes the throttle cf a locomotive should be expected to be dry, and it would be entirely so if it were not that the violence of the circulation projects small beads of water upward, far beyond the general surface. Some of these enter the throttle with the steam. This action explains why the moisture increases with the power of the boiler, and makes it not unreasonable to assume that the purity of the water in the boiler may actually affect the quality 'of the steam. The comparative dryness of the steam under all conditions is a fact worthy of emphasis, for the locomotive is often credited with carrying over a great deal of water to the cylinders. The tests show that this does not happen under constant conditions of running. When it occurs it is probably the result of too high a water level or of a sudden demand upon the boiler. For example, if the throttle of a locomotive, which has been for some time inactive, is quickly opened, large volumes of steam-bubbles leave the heating surface and crowd to the upper part of the boiler, making spray in the dome, a portion of which may pass out with the steam. A similar action occurs when an engine, which has been working at a light load, is suddenly required to increase its power. But these are exceptional conditions. Under uniform conditions of running, such as prevailed throughout the tests herein presented, the moisture passing the throttle is never great. If, as the writer believes, it is fair to conclude that variations in moisture are largely due to incidental conditions, no serious mistake would be made if the indications of the calorimeter were entirely disregarded, and all calculations based on the assumption that the steam generated is dry and saturated. Results thus obtained should be somewhat more consistent, one with another, than those cor- rected for moisture, and hence for general purposes more satisfactory. In accord with this view, while none of the calorimeter work has been ignored in calculating results, many of the derived results have been carried out in duplicate. Thus, the equivalent evaporation is first 132 LOCOMOTIVE PERFORMANCE. determined on the assumption that the steam has the quality shown by the calorimeter (Column 18), and is also determined on the assump- tion that all water delivered to the boiler is evaporated into dry and saturated steam (Column 19). In accord with the usual practice, however, comparisons which follow are based on the corrected results (Column 18). 48. Power of Boiler (Table XXIX) .The power developed by the boiler is proportional to the rate of evaporation (Column 18). A few comparisons will serve to show something of the peculiar con- ditions under which the boiler of a locomotive performs its service. Thus, the water evaporated per foot of grate surface (Column 20) varies from less than 400 to more than 1,000 pounds per hour. These figures reflect well the intensity of the furnace action, which must provide for the combustion of sufficient fuel to produce such a result. Similarly, the weight of water evaporated per square foot of heating surface (Column 21) varies from 5J to almost 15 pounds per hour, the maximum rate being nearly the equivalent of a boiler horse-power for every 2 square feet of heating surface in the boiler. The total boiler horse-power (Column 22), the horse-power per square foot of heating surface (Column 23), and the horse-power per square foot of grate surface (Column 24) are also of interest, especially for the higher power tests. 49. Coal and Combustible (Table XXX). Attention has been called to the very large amount of water evaporated by the boiler tested. It follows that a correspondingly large amount of fuel is needed to bring about this evaporation. .The record appears in Table XXX. The coal used for all tests was Indiana block, mined in the neigh- borhood of Brazil. It burns to a white ash, without clinkers, and is light and friable, which qualities prevent its giving maximum results in locomotive service. The composition of several representative samples proved 'to be as follows: 1334 Per cent fixed carbon 49.65 51.84 51.09 51.59 Per cent volatile matter 40.29 39.00 38.93 38.87 Per cent combined moisture 3. 15 3.62 2.35 3.44 Per cent ash 6.91 5.54 7.63 6.10 100.00 100.00 100.00 100.00 In each test the coal was weighed a barrowful at a time as it was dumped at the feet of the fireman. A sample of from 50 to 100 BOILER PERFORMANCE. 133 TABLE XXIX. POWER. The several tests represented in this table are arranged in order of the rate of evaporation, No. 1 representing the test for which the rate is least, and No. 35 that for which it is greatest. Identification of Test. Dura- Equivalent Evaporation from and at 212 Fahr. Pounds, assuming Quality of Steam as shown by Column 16. Rated Horse-power 34.5 Evaporative Units, assuming Quality of Steam as shown by Column 16. tion of Test in Water Water >nsecutive Number. Labora- tory Desig- nation. Min- utes. Water Hour. Evaporated per Square Foot of Grate Surface Evaporated per Square Foot of Heating Surface Total. 5 er Square Foot of Heating Surface. Per Square Foot of Grate Surface. a per Hour. per Hour. i 2 4 18 20 21 22 23 34 i 15-1-A 240 6,659 386 5.48 193 .159 11.2 2 15-1 6-V 190 6,763 392 5.57 196 .161 11.4 3 15-1-H 180 7,249 420 5.97 210 .173 12.2 4 25-1-A 255 7,318 424 6.03 212 .175 12.3 5 15-1-G 180 7,476 433 6.16 217 .179 12.6 6 15-2- A 180 8,745 507 7.20 253 .208 14.7 7 25-1-V 240 8,865 514 7.30 257 .212 14.8 8 35-1-A 180 9,771 566 8.05 283 .233 16.4 9 35-2-F 180 9,785 567 8.06 284 .234 16.4 10 35-1-V 140 10,284 596 8.47 298 .245 17.3 11 35-2-E 180 10,324 598 8.51 299 .246 17.3 12 45-1-A 150 10,395 603 8.56 301 .248 17.4 13 35-2-B 170 10,413 604 8.57 302 .249 17.5 14 55-1-A 120 10,691 619 8.80 310 .255 17.9 15 35-1-H 160 11,090 643 9.13 321 .264 18.6 16 35-1-G 170 11,226 650 9.24 325 .268 18.8 17 35-1 &-G 170 11,312 655 9.31 328 .270 19.0 18 25-2-A 180 11,554 669 9.51 335 .276 19.4 19 55-1-V 120 11,624 673 9.57 337 .278 19.5 20 35-2&-H 160 11,634 674 9.58 337 .278 19.5 21 55-1-H 120 12,303 713 10.13 357 .294 20.7 22 55-1-G 120 12,426 720 10.23 360 .296 20.9 23 15-9- A 150 13,670 792 11.25 396 .326 22.9 24 15-9-H 180 13,722 796 11.29 398 .328 23.1 25 35-2-A 180 13,735 796 11.31 398 .328 23.1 26 35-3-G 140 13,867 804 11.42 402 .331 23.3 27 15-9-G 160 13,869 804 11.42 402 .331 23.3 28 35-3-H 120 13,995 811 11.52 406 .334 23.5 29 35-2-G 160 14,725 853 12.12 427 .352 24.8 30 35-2-H 120 14,771 856 12.16 428 .353 24.8 31 45-2-A 140 14,926 865 12.29 433 .357 25.1 32 25-3-A 122.5 15,515 899 12.77 450 .370 26.1 33 35-2-C 120 15,709 911 12.93 455 .375 26.4 34 55-2-A 68 16,773 973 13.81 486 .400 28.2 35 35-3-A 120 17,883 1,037 14.93 518 .427 30.0 134 LOCOMOTIVE PERFORMANCE. TABLE XXX. COAL AND COMBUSTION. The several tests represented in this table are arranged in order of the rate of evaporation, No. 1 representing the test for which the rate is least, and No. 35 that for which it is greatest. Identification Fuel: Indiana Brazil Block Coal, Pounds. Rate of Com- of Test. bustion. Dura- 8,31 1 ls || 1* Labora- tory Desig- nation. tion of Test in Min- utes. Total Dry Coal Fired. Total Refuse Caught in Ash- pan. Total Com- bustible by Analy- sis. Dry Coal Fired Hour. Com- bustible Fired Hour. y Coal Fired 3q. Ft. of Gn Surface per H y Coal Fired !q. Ft.of Heat Surface per H Q _i/j per Hou FIG. 95. Heating Value of Sparks. cent of the weight of the coal fired, and a pound of the sparks was equivalent only to four-sixths of a pound of coal, whereas, in Test No. 7, when the engine was working hard, about 14 per cent by weight of the coal escaped as sparks, and each pound of sparks was equivalent to nearly five-sixths of a pound of coal. Under the con- ditions of this test, therefore, more than 10 per cent of the fuel which enters the furnace may completely pass the heating surface unconsumed. 73. Volume of Sparks as Dependent upon Quality of Fuel. In 1898, in connection with locomotive Schenectady No. 2, a series of tests designed to determine the relative value of five different samples of coal were undertaken by the Engineering Laboratory for the C. C. C. & St. L. Railway Company, in connection with which the extent of spark losses was determined. The several samples of coal were designated as A, B, C, D, and E, and the evaporative efficiency 180 LOCOMOTIVE PERFORMANCE. obtained from them in locomotive service is that set forth in Table XLIL, the values of which are arranged in the order of merit. From the evaporation obtained the value of the several samples may be estimated. Sample E was a West Virginia coal of very high quality, while sample C was a Western coal, probably from an Indiana or Illinois mine. TABLE XLII. EQUIVALENT EVAPORATION. Pounds of Water Evaporated per Pound of Coal. Relative Value of Sample, calling the Value of Sample E 100. of Sample. When the Rate of Evaporation is 5. When the Rate of Evaporation is 10. For Use under Light Power. For Use under Heavy Power. I. II. III. IV. V. E 10.7 8.7 100 100 A 10.0 8.1 93 93 B 9.4 7.7 87 89 D 8.3 7.0 79 82 C 8.5 7.1 77 81 The spark loss is assumed to be the weight caught in the front end plus that which passes out of the stack. Values for spark losses thus obtained for the several samples of coal tested, expressed as a per- centage of the weight of coal fired, are presented in Columns II. and III. of the table below. TABLE XLIII. SPARK LOSSES. Designation of Sample. Percentage of Weight of Coal Fired, Accounted for as Sparks Entrapped in Front End and Passing out of the Stack. Relative Weight of Sparks Produced by each of the Several Samples in Generating the Same Weight of Steam, assuming the Weight of Sparks resulting from Sample E to equal 100. When the Rate of Evaporation is 5. When the Rate of Evaporation is 10. When the Rate of Evaporation is 5. When the Rate of Evaporation is 10. I. II. III. IV. V. E A B D C 6.6 6.2 4.8 5.7 3.3 21.2 15.6 16.4 17.4 13.8 100 101 84 109 65 100 79 87 100 80 SPARK LOSSES. 181 From the preceding table it appears that under conditions of running common in every-day service, from 14 to 21 per cent of the coal fired disappears in the form of sparks (Column III). It is true, however, that the fuel loss is not as great as this, since the sparks represent fuel which has been partially consumed. The sparks result- ing from the samples under consideration have not been analyzed, but the investigations already described indicate that they have from 60 to 85 per cent of the fuel value of a similar weight of coal. A fair estimate of the fuel losses in these tests, resulting from the passage of sparks through the tube, may be taken as 75 per cent of the values given in Columns II. and III. From a review of the tables it appears that those samples giving the highest evaporation also give the largest spark losses. Two con- ditions probably account for this fact. First, the purer coals are lighter, and hence respond to the draft action more easily than those intermixed with non-combustible matter; secondly, in general, the better the coal the lighter the ash, a large percentage of which, instead of falling through the grate passes out with the sparks, and adds its mass to their weight. It may be urged as an objection to those coals giving a high ef- ficiency, that their use is attended by a large spark loss. The argu- ment, while good, is not true to the extent indicated by the values in Columns II. and III. For example, Column III. shows the percentage of the coal fired, accounted for as sparks, but a pound of sample C, producing 0.138 pound of sparks, did not make as much steam as a pound of sample E, producing 0.212 pound of sparks. The relative spark-producing qualities of the several samples, based upon the weight of steam generated, are given in Columns IV. and V. These A r alues, therefore, serve as a logical basis from which to determine the relative spark-producing qualities of the several samples. 74. Refuse Caught in the Ash-pan. Closely allied with the matter of spark production is that of refuse in the ash-pan, since, as already noted, when the ash is light much of it passes up the stack. The facts with reference to ash, as applying to the five samples of coal dealt with in the preceding paragraph, are given in Table XLIV. The table shows that when the engine is running light, the five samples give nearly the same amount of refuse in the ash-pan, whereas, when the power is increased to make the rate of evaporation 10, sample C gives nearly three times as much deposit in the ash-pan as sample E. In general it may be said that the better the fuel, the less deposit 182 LOCOMOTIVE PERFORMANCE. there will be in the ash-pan. Comparing the values of Table XLIV. with those of Table XLIII. it appears that as the character of the coal changes with reference to refuse in the ash-pan, an inverse change results with reference to spark losses. Such a result is logical, and is quite in accord with the explanation given in the discussion of spark losses. TABLE XLIV. REFUSE IN ASH-PAN. Relative Weight of Ash resulting Designation Percentage of Weight of Coal Fired, Accounted for as Refuse in Ash- pan. from Different Samples when the Same Weight of Steam is Gener- ated, calling the Weight resulting from Sample E 100. of Sample. When the Rate When the Rate When the Rate When the Rate of Evaporation of Evaporation of Evaporation of Evaooration is 5. is 10. is 5. is 10. I. II. III. IV. V. E 11.0 4.9 100 100 A 10.7 9.2 105 203 B 13.9 11.6 146 267 D 14.8 10.3 170 257 C 14.4 11.0 171 276 75. Distribution of .Sparks Throughout the Stack. To inter- cept the sparks at some point between the front tube-sheet and the top of the stack is one of the problems which confront designers of draft appliances. The development of a successful device for the pur- pose is likely to be advanced by a knowledge of the course taken by the sparks in leaving the boiler. It is therefore thought proper to present here whatever information may have been procured in the course of this investigation with regard to spark distribution in the stack. For this purpose the data of Test No. 4, which are considered to be rep- resentative, are presented graphically in Figs. 96, 97, and 98. In Fig. 96 the numbers express the weight of sparks passing out of the stack per square inch per hour at the places designated, while in Fig. 97 they indicate he total pounds of sparks per hour passing each of the several areas into which the stack was arbitrarily divided. Comparing these results it appears that the sparks follow the out- side of the exhaust steam rather than the center. The weight of sparks per unit area increases steadily from the center to the circum- ference, with the result that ever 50 per cent of the whole weight is credited to a ring two inches broad, measured from the eutsicle circum- SPARK LOSSES. 183 ference, which ring contains only about 40 per cent of the total area. The velocity of the exhaust is necessarily less on the outside of the stream, and, apparently, the sparks most readily follow the portion of stream issuing from the stack with the least velocity. In this con- FIG. 96. Pounds of Sparks passing out of Stack per Square Inch per Hour. FIG. 97. Pounds of Sparks passing out of Stack per Hour in the Areas Indicated. nection it should be noted that these observations were all made on a cross-section of the stream as it issued from the stack, and therefore do not necessarily represent conditions actually existing in the stack. FIG. 98. Cross-section of Stack showing Density of Spark Discharge. 76. The Size of Sparks varies with the extent of the spark losses. Thus, under low rates of combustion, when the total spark loss is small, it consists of a very fine, almost sooty, deposit (A, Fig. 99). but when the total loss becomes large, the sparks themselves are large (B, Fig. 99). Fig. 99 is a photograph of two lots of sparks, and of a pile of buckshot, with which the sparks may be compared. The sample A was obtained in Test No. 1, when the total loss was only 22 pounds per hour, while the sample B represents Test No. 7, for which the loss was more than ten times as great. 77. Conclusion. The results of the investigation described in the preceding paragraphs seem to justify certain conclusions which, while 184 LOCOMOTIVE PERFORMANCE. susceptible of rather general application, should nevertheless be accepted only for the fuels and boilers involved. Both locomotives involved, Schenectady No. 1 and No. 2, have narrow fire-boxes and a grate area of about 17.5 square feet. The conclusions are as follows : 1. The weight of sparks, which passes the heating surface of a boiler as cinders or sparks, increases steadily with the rate of com- bustion, and may reach a value of from 10 to 15 per cent of the coal fired. 2. Sparks are composed of coals more or less completely burned, or they may be wholly non-combustible, in which case they are entirely of ash. FIG. 99. Sample Sparks. 3. The fuel value per pound of sparks increases as the total amount increases. 4. The weight of sparks produced and the weight of refuse caught in the ash-pan are inversely related. When the spark production is high, much of the refuse which would otherwise drop into the ash-pan passes out from the top of the stack. 5. The distribution of sparks throughout the cross-section of the stream issuing from the stack is such that the greatest weight of sparks follows the slowest currents on the edges of the stream, more than one- half the total weight passing through the annular area which com- prises the two inches nearest the stack. 6. The size of the individual sparks increases with the total amount produced up to the limit allowed by the openings in the netting. CHAPTER X. RADIATION LOSSES. 78. The Amount of Heat Radiated from a locomotive boiler is necessarily large, but few attempts have been made to measure it. It is chiefly for this reason that the experiments herein described are of interest. 79. Loss of Heat from a Locomotive Standing in a Building. By means of a Bristol recording pressure-gauge attached to the boiler of locomotive Schenectady No. 1 many charts were obtained, showing the rate at which the boiler pressure declined after the locomotive had been shut down for the day. These charts indicate that ordinarily the pressure -would fall from 120 to pounds in from 12 to 15 hours, the exact time, of course, depending upon many different conditions. On January 25, 1895, at the close of a test, the water level was brought to the 5" mark on the glass, and the pressure noted. The chart on the Bristol gauge gave the subsequent record of time and pressure. From the known dimensions of the boiler and these data> the following was readily derived : Water in boiler, 140.6 cu. ft ^ 8787.5 Ibs. Steam in boiler at 110 Ibs. absolute pressure, 53.0 cu. ft.= 13.2 Ibs. Total in boiler , 8800.7 Ibs. Total weightXheat of liquid = 2,685,974 B. T. U. Weight of steamXheat of vaporization = 11,565 B. T. U. The heat of the steam is but .43 of one per cent of the heat of the liquid, and is negligible. In determining heat losses it will be sufficient to assume that the thermal units dissipated during any given interval are the product of 8800 and the fall in 'temperature in that interval, Results thus derived are given in Table XLV. 185 186 LOCOMOTIVE PERFORMANCE. TABLE XLV. HEAT LOST FROM LOCOMOTIVE BOILER IN BUILDING. AVERAGE TEMPERATURE OF BUILDING, 71.5 F. Time. Absolute Pressure , Lbs. Temperature, Degrees F. Fall in Temperature, Degrees F. Fall in Temoerature per Hour, Degrees F. B.T.U. Lost per Hour. 5.07 110 334.56 5.20J 100 327.58 6.98 31.02 273,007 5.37 90 320.04 7.45 26.64 234,459 5.59 80 311.80 8.24 22.98 202,247 6.25 70 302.71 9.09 20.58 181,125 6.59 60 292.51 10.20 18.24 160,530 7.40 50 280.85 11.66 17.04 149,969 8.36 40 267.13 13.72 14.70 129,374 9.52 30 250.27 16.86 13.14 115,645 11.45 20 227.95 22.32 11.88 104,556 The losses shown have by implication been referred to as radiation tosses, but they really include heat losses arising from all sources. A consideration of the methods employed will make it evident that they include losses resulting from leakage either of steam or water. The boiler itself is known to have been tight, and the throttle and other valves are thought to have been so, but experience proves that in so large and complicated a structure as a locomotive boiler and its fittings it is impossible to absolutely avoid leakage. The locomotive was in an exceptionally good condition, and the values given, are to be accepted as representing such losses as occur from causes stated under most favorable conditions. The coal equivalent of the heat lost may be found by assuming the boiler to absorb 8000 B.T.U. per pound of coal. This assumption is justified by v the facts presented in Chapter VI. By its use the heat loss per hour when the gauge shows 95 pounds pressure (abso- lute 110) is equivalent to 273007 8000 = 34. 1 pounds of coal. As the pressure and temperature fall, the heat loss diminishes. 80. Radiation Losses upon the Road. Tests to determine the heat losses from the boiler of a locomotive on the road were undertaken in cooperation with the Chicago & Northwestern Railroad Company and several manufacturers of boiler-covering in 1898. The immedi- ate purpose of the test was to disclose the relative value of different RADIATION LOSSES. 187 materials employed in lagging boilers, but the plan was sufficiently broad to make the results of value in a discussion of the more general subject of heat losses on the road.* Si. Plan of the Tests. In carrying out the tests two locomotives were employed; one to be hereafter referred to as the " experimental locomotive" was subject to the varying conditions of the tests; the other was at all times under normal conditions serving to give motion to the experimental locomotive, and as a source of supply from which steam could be drawn for use in maintaining the experimental boiler at the desired temperature. The experimental locomotive was coupled ahead of the normal engine, and, consequently, was first, when running, to enter the undisturbed air. The action of the air-currents upon FIG. 100. Head of Experimental Train. it, therefore, was in every way similar to those affecting an engine doing ordinary work at the head of a train. The boiler of the experimental locomotive was kept under a steam pressure of 150 pounds by a supply of steam drawn from the boiler of the normal engine in the rear. There was no fire in the experimental boiler. It was at all times practically void of water. Precautions were taken which justified the assumption that all water of condensa- tion collecting in the experimental boiler was the result of radiation of heat from its exterior surface. This water of condensation was collected and weighed, thus serving as a means from which to calculate * These tests were outlined and conducted by the author under the direction of the Chicago & Northwestern Railroad Company, as represented by Mr. Robert Quayle, Superintendent of Motive Power, in cooperation with the manufacturers of various materials employed as boiler-coverings. See also "Proceedings of Western Railway Club," January, 1899. 188 LOCOMOTIVE PERFORMANCE. the amount of heat radiated. The head of the experimental train is shown by Fig. 100. 82. The Experimental Boiler and its Equipment. The Chicago & Northwestern locomotive, No. 626, the boiler of which served in the experiments, is of the eight-wheeled type, weighing about 90,000 pounds. An outline drawing, used in ordering covering, is shown by Fig. 101. The principal dimensions of the boiler are as follows : TABLE XLVI. DIMENSIONS OF BOILER. Diameter in inches 52 Heating surface (square feet) 1391 Total area of exterior surface, not including surface of smoke-box 358 Area of surface covered (square feet) 219 Area of steam-heated exposed surface not covered 139 Ratio of surface covered to total surface .61 It should be noted that the values given in Table XLVI. are based upon projected areas of the plain boiler. No account has been made of the edges of plates at joints, nor of surface due to the projection of rivet-heads, nor of the surface of various attached projections, such as running-board brackets and frame fastenings. While all such pro- jections above the general surface of the boiler are active agents- in conducting heat from the interior, the present study does not require them to be taken into account. The extent of area covered for this boiler is entirely normal for the class of locomotives to which Xo. 626 belongs, which gives added interest to the fact that but 61 per cent of the exposed surface of the boiler was covered. The whole grate of the experimental boiler was deadened by brick- work, and as a further precaution against the movement of air-currents through the fire-box, tubes, etc., the top of the stack was securely filled with wood. The furnace and front-end doors were also care- fully closed and fastened. A steam-separator in the supply-pipe within the cab of the experimental boiler was assumed to deliver to it steam of a uniform quality. These and other precautions justify the assump- tion that all condensation occurring in the boiler was due to radiation from its exterior surface. As a safeguard against air-pockets and to further insure a uniform temperature of "all portions of the interior of the boiler, steam was allowed to waste from it through a small orifice at the end of a pipe connecting with the front end, and leading outside to the top of the RADIATION LOSSES. 189 I I i 190 LOCOMOTIVE PERFORMANCE. stack, and some leak was allowed also at the whistle-valve. The loss of steam from the experimental boiler in no way affected the value of the measurements made, since they neither increased nor diminished the amount of condensation taking place within the boiler. A 12-inch water-glass was attached to the water-leg of the boiler close to the mud-ring. A thread around this glass served as a reference- line. The water condensing within the experimental boiler was led through a f-inch pipe from the blow-off cock at the bottom of the boiler to a valve at the rear of the cab, thence to the top of the tender- tank, at which point it connected with a coil submerged in the water of the tank. The discharge from this coil was delivered to a weighing- barrel set up within the coal space of the tender. By these means the water of condensation was made a measure of the amount of heat radiated, its level was maintained constant a few inches above the level of the mud-ring, the excess was drained out, cooled to avoid all chance of loss from vaporization, and weighed. As the scales could not be balanced during a run, and as the weigh- ing-tank was of insufficient capacity to hold all of the water accumu- lating during a test, a calibrated small-necked can was used between stops to reduce the level of the water in the barrel. Each can, as emptied, was charged against the barrel, a full can counting 45.7 pounds. 83. Observers. Three observers were ordinarily employed during each test. One was assigned the duty of observing the force and direc- tion of the wind, the character of the weather, and to so manipulate the valve in the pipe through which the condensed steam was discharged to the weighing-barrel, that the water level within the experimental boiler would at all times remain near the reference-line. Another weighed the condensed steam, and a third recorded five- minute readings of the steam pressure within the experimental boiler, and attended to the discharge drain of the steam separator, in order that the water level within this apparatus might be kept within fixed limits. He also rang the locomotive bell for crossings. The rear engine carried its usual crew. The train conductor also rode on the rear engine. 84. The Track used for the running tests is a single line extending from Clinton to Anamosa, Iowa, a distance of seventy- two miles. This stretch of road was chosen because of the light traffic upon it, and the assurance against interruption which this condition gave. It leads over rolling country, and throughout its length is rather RADIATION LOSSES. 191 sinuous. For the first twenty-five or thirty miles the general direction is northwesterly, and for the remainder of the distance nearly west. For four miles out of Clinton it extends through the yards of that city and the adjoining city of Lyons, and for several miles it follows along the Mississippi River, from which it finally leads out upon a more open country. The wind and temperature conditions were generally differ- ent for that portion of the road along the river than for portions extend- ing across the more open country. Crossing stops were necessary just beyond Lyons and at Delmar, thirty-three miles from Clinton. There are fifteen stations intermediate between terminals, but, with one exception, the plan of the tests did not involve them. 85. Movement During the Tests. The work in connection with each covering occupied ordinarily a single day. A test under condi- tions of rest, hereafter referred to as a "standing test," was first made, followed by a test on the road, hereafter to be referred to as a " running test." Each running test involved a trip from Clinton to Anamosa and return, with an intermediate stop each way at the station of Maquo- keta, thirty-eight miles from Clinton and thirty-four miles from Ana- mosa. By means of these stops it was possible to divide each running test into four parts of nearly equal length, which not only gave oppor- tunity for ascertaining something of the character of the results, which were being obtained, but served as a safeguard against the loss of a whole test in case of an accident on the road. For convenience these parts of the running tests are hereafter referred to as " quarters," but they are not of equal value. The quarters may be defined as follows : 1st quarter, Clinton to Maquoketa, 38 miles, approximate running time, 83 minutes. 2d quarter, Maquoketa to Anamosa, 34 miles, approximate running time, 68 minutes. 3d quarter, Anamosa to Maquoketa, 34 miles, approximate running time, 71 minutes. 4th quarter, Maquoketa to Clinton, 38 miles, approximate running time, 84 minutes. In anticipation of a test the locomotives were coupled, the pipe connections made, and steam turned on the experimental boiler at as early an hour as practicable. In most cases this was between 6 and 7 o'clock in the morning. After the normal pressure had been secured in the experimental boiler, the cooling coil within the tender was put 192 LOCOMOTIVE PERFORMANCE. under pressure and examined for leaks. The engines were then moved to the yard stand-pipe and the tender- tanks of both engines filled. As soon as practicable after this, and to insure the same level of the experimental boiler for all tests, the engines were moved to a point lower down in the yard where the rear driver of the experimental engine rested over a certain marked tie. The water within the experi- mental boiler, resulting from condensation, was then brought to the reference-line, time was taken, and the scales of the weighing-barrel balanced. At fifteen-minute intervals thereafter this process of bringing the water to line and balancing the scales was repeated, usually for a period of from one to two hours, the locomotive remaining in its place upon the marked tie. When the rate of condensation became uniform the standing test was assumed to have commenced. In due time, usually at about 9:35 in the morning, the water was brought to line for the last observation of the standing test, the scales balanced, and as soon as practicable thereafter the engines were started for the running test. The time of this balancing of the weigh- ing-tanks marked the end of the standing test and the beginning of the running test. The first few miles were, necessarily, at varying speed, but after passing Lyons and the crossing just beyond, a speed of thirty miles an hour was soon secured, and was thereafter maintained until Maquo- keta was reached. In all tests the stop at Maquoketa was made with the rear tender under the spout of the water-tank, water was taken by the pushing engine, and the engines were oiled. While this was being done the water of condensation in the experimental boiler was brought to line and the weighing-barrel balanced, thus ending the first quarter and beginning the second quarter of the running test. After a ten minutes' stop start was again made and the run continued to Anamosa, where the stop was made with the rear driver of the experi- mental locomotive on a certain marked tie. As soon as practicable thereafter the water was brought to line and the weighing-barrel bal- anced, thus ending the second quarter of the test. At Anamosa the engines were uncoupled and turned one at a time, coupled again, and the rear driver of the experimental boiler brought over the same marked tie upon which stop had been made. Steam was then turned on the experimental boiler, and the pressure which, during the process of turning, usually dropped to about eighty pounds, was restored to normal conditions. The start from Anamosa was made at about 1 :45, or an hour and a half after the scheduled time RADIATION LOSSES. 193 of arriving. This interval gave time for the work of balancing the weighing-barrel, turning the engines, and for dinner; it also gave a sufficient period, after normal pressure had been restored in the experimental boiler, to allow everything to become thoroughly warm before starting. When all was ready the condensed steam in the boiler was brought to line, the weighing-tank balanced, and the third quarter of the test thus commenced. As soon as practicable after the balancing, the train was started on the run to Maquoketa, where, as before, the stop of ten minutes' duration was at the water-tank. Here, again, the water was brought to line and the weighing-tank balanced, thus ending the third quarter and beginning the fourth quarter of the test, which in turn ended at Clinton at about 4 :30 in the afternoon upon the same marked tie from which the engines had been started in the morning. This process was persisted in with regularity, the intent being to secure similar conditions for each of the several tests. After the final balance the tank-valves of the experimental loco- motive were opened and the tank drained to allow the inspection of the cooling coil, which inspection, as already stated, preceded every test. Steam was shut off the boiler and the experimental locomotive was pushed into the roundhouse, where men were in waiting to strip it of its jacket and covering. Early in the evening work was com- menced in applying the covering which was to be tested on the follow- ing day, and was continued into the night until finished. It was found impossible to make the running time of all tests. the same. Conditions arose which could not have been anticipated. There were occasional stops due to section gangs and to the presence of other trains. Time which was lost in this way was not made up during the run, the effort being to keep the speed while running constant. It is to be noted also that the running tests actually com- menced before the train was started, that the ten minutes' stop at Maquoketa was a part of the running time, and that the test did not end the moment the engine stopped, that is, there was a certain amount of dead time on all running tests. The facts in detail with reference to this are fully presented in another place. 86. The Coverings Tested. Tests were made in conjunction with the bare boiler (A), with an old covering of wood (B) which was on the boiler when the work began, and with five different forms of manu- factured coverings designated as C, D, E, F, and G respectively. The list included all of those materials now common as insulating materials, 194 LOCOMOTIVE PERFORMANCE. as, for example, magnesia, asbestos, and cellular asbestos board. The thickness in all cases was designed to be the same for all tests, but the practice of covering the surface of a locomotive boiler, with its inequalities, with material in such thickness as will give a smooth exterior surface, leads to variations in the thickness of the covering. In the boiler tested the dome-casing was so small as to allow only a thin layer (from i to f in.) on the barrel. On other portions of the boiler it was intended that the thickness should vary from If in. on the first ring to 1J in. on the slope sheet. An effort was made to have the thickness of all coverings tested the same, and, except in two cases, this result was very nearly attained. As a comparative figure an average thickness of all coverings was obtained by finding the volume of the material used, and dividing this by the area of the surface covered. This process gives the thick- ness which the material would have had if it had been distributed uni- formly over the surface covered. Values thus obtained are as follows: Covering B 1.34 in. = 1$ in. C 1.49in. = li|in. DI 1.45in. = Hf in. " D 2 1.57 in. = Hf in. E ' 1.28in. = l&in. FiandF 2 1.56in. = Hf in. " G 1.49 in. = lf in. In all cases the material as above described was covered by the usual Russia iron lagging. 87. The Tests. Both standing and running tests were made with the experimental boiler bare, and also w^hen protected by six dif- ferent coverings. Tests of two of these were repeated, making alto- gether nine standing tests and nine running tests to be reported. These are designated as follows: A, B, C, DI, D 2 , E. FI, F 2 , and G. "A" represents the test- of the bare boiler. "Di" and "D 2 " are different tests of the same covering, and, similarly, "Fi" and "F 2 " are tests of a single covering. 88. Standing Tests and Results. All standing tests, save one, immediately preceded the corresponding running test. Standing Test C followed the running test. The observed data, and the results which have been derived from them, are presented as Table XLVII. Most of the lines in this table are self-explanatory, but there are a few which demand a word of explanation. RADIATION LOSSES. 195 CO -< d CO O t- 10 CO O co d O CD O t^ -P 05 o 2 I t - o CO CO* b 1 g ^^ 8 rf O i 00 O CO IO 10 o I bo t^ 3 O (M O -< ^H CD -H ^H O e CO GO w CO IT*- IO IO IO OO C~- F- O T* I-H I> (M CO X X CO l> O ^ocoooco 10 IO S 10 S I-H CO O O CD 00 00 O5 O5 00 t- IO 00 O5 O5 ^H O CO (M CO CD 10 10 r co O5 10 CO O 10 00 (M CO O (M rf 00 O t- t- T* GO GO r- O5 <* CO O5 CO O O 00 CO t- 00 O CO 10 1 T* t^ 05 i-H i-H CO 05 00 ^H (M O T* T*H 10 10 10 CO 05 05 00 00 A O * I-H CO GO iO CO (M CO CO U3 GO O5 O (M 00 (M C<1 CO O 00 O CO O O CO GO CO O5 CO 00 t> t 00 (N I 1 00 l-H Tj< IO 00 CO t > - 00 O ^05^10 gH ggggs (M rt* CO t~ iO CO t^ !><# T* O 05 CO CO 00 CO CO CO O O CO CO 00 CD t- IO OO oo r~ t> 05 co * 00 O O OS rjt C^ JH s^sss 71 71 00 GO GO GO GO IO 00 00 CO IO CO O CO O S j IO 05JOCQ G^--; t- GO 00 GO 00 1: V: ! 9 i e -o *6 <3 a. 1st Quarter. . . b. 2d c. 3d " d. 4th " .... e. Total for test. . S ~O o 'XS > -# "S T3 T3 "5 O -H 01 COT* EH C o ^,5 "^S ^ f ^| Designation of test Duration of test, min. Time between initial and final balance. Actual running time, min. I M s i 1 i 1| Average atmospheric temperature, F. - (M n \ * I o CO 198 LOCOMOTIVE PERFORMANCE. IS-fi C5 t> r- co co oo 05 05 CO CO GO r~ i i <*' oi O OS O C5 IO Tf (N t- Tfl t- co co co T? co r- co . 2d " c. 3d " . d. 4th " . e. Total for te g- condensa of water to weighi Orv poun livere barrel bse ted from - to con- hich ver- 150 80R * ~ fl -g 2 8 S b c 5^42 ^ 1!1H| Slil^l ed ed qui o re sp e add bserv an eq been Amount to b the total o vert it int would hav age steam pr Ibs. and the a RADIATION LOSSES. 199 g X CO oo CO CO l-H g i-H o 00* 1 0* i i-H 1 CO GO* 2 05 CD s 05 00* CD 0* i-H i CO CO CO CO 10 oo' d | rH 8 co CO i I' o l-H "CC T3 e 3 j f s & 0) f 5 s c J3 c / / 8 h - 8 / s > d o / P x< a*^ x^ ' .5 !/ cX ^> LX^ a f o - ^ \ 'S ^ ^> *^ ^ Mi lesF erH our 10 20 30 40 50 60 70 80 90 100 FIG. 102. Effect of Speed on Radiation Loss. points are not sufficient to establish with accuracy the complete rela- tionship of radiation and speed, an estimate of real value may be based upon them. Such an estimate is presented in the form of a diagram, Fig. 102. The diagram shows that the bare boiler when at rest radi- ates sufficient heat to condense 6.8 pounds of steam, at 150 pounds pressure per minute, which amount is increased to twenty-eight pounds when the same boiler is driven at a speed of eighty miles an hour. Similar values for the covered boiler are 3.0 pounds and 10.6 pounds respectively. RADIATION LOSSES. 207 Changes in atmospheric temperature would of necessity affect re- sults. Those recorded were obtained in midsummer and all have been corrected for an atmospheric temperature of 80 F. For each 10 reduction in atmospheric temperature below 80 the radiation may be expected to increase 3.5 per cent. For zero degrees temperature the radiation losses recorded in this report should be increased by about 28 per cent. For example, if, when the atmospheric temperature is 80, the conditions are such as result in the condensation of five pounds of steam per minute, when the atmospheric temperature is the condensation will be 5 + 5(.035X80)=5+1.4 = 6.4. From this it appears that very low temperatures are attended by radia- tion losses of considerable magnitude. Changes in steam pressure also would affect results. The ex- periments were conducted under a boiler pressure of 150 pounds by gauge. With an increase of pressure the boiler temperature will be- come higher, and the radiation losses will, as a consequence, be aug- mented. Changes arising from this source, however, are not great. For each ten-pound increase of pressure above the limit of 150 pounds the radiation may be expected to increase by about 1.6 per cent, but this will not apply for pressures much above 200 pounds. A pressure of 200 pounds will involve losses by radiation which are 8 per cent greater than those making up the record of this report. From these considerations it can be shown that with the boiler bare and the locomotive running at eighty miles an hour, under a steam pressure of 200 pounds, with the atmospheric temperature 0, the loss by radiation would be the equivalent of sixty-seven horse-power, while a covered boiler, running under the same conditions of speed, pressure, and atmospheric temperature, would still be subject to a loss of twenty-five horse-power. As a locomotive similar to that tested may be expected to deliver a maximum of 600 horse-power, it is evident that under the extreme conditions just assumed at least 10 per cent of the total power of the engine would be lost in radiation. This is for an uncovered boiler. An application of any of the coverings tested would reduce this maximum loss to about 4 per cent. Finally, it should be remembered that the boiler tested was one of moderate size. Many boilers are now running which present an ex- posed area which is at least 50 per cent greater than that presented by the boiler under test, and it should be evident that the losses from 208 LOCOMOTIVE PERFORMANCE. such large boilers will be greater than those disclosed by the tests under consideration. For boilers of the same general type the loss will prob- ably be proportional to the exposed surface. 95. Conclusions concerning the extent of losses by radiation may be stated as follows : 1. The amount of heat radiated from the boiler of a locomotive in motion is affected by many different conditions, as, for example, by the speed of the locomotive, the temperature of the atmosphere, the direction and velocity of the wind, the steam pressure, and by the pro portion of the whole surface of the boiler which is covered. 2. The application of a suitable covering to 61 per cent of the ex- posed surface of the boiler has resulted in a saving of 62 per cent of the heat radiated from the uncovered boiler. This does not imply that if the whole surface were covered the radiation would be reduced to zero, but rather that the covered portions are those which are most exposed. 3. The radiation from a given boiler is approximately twice as great when the locomotive is running thirty miles an hour as when it is at rest. 4. The loss from a bare boiler of a locomotive running 30 miles an hour is from 4 to 10 per cent of the maximum capacity of the boiler. In summer it will not much exceed the lower limit, in winter it may frequently approach the maximum. An increase of speed to 60 miles will nearly double the loss. 5. The loss from a boiler of a locomotive running 30 miles an hour having a considerable portion of its surface covered in accord with good practice, is from less than 1 to not more than 5 per cent of its maximum power. In summer it will not greatly exceed the lower limit; in winter it may frequently approach the maximum. An in- crease of speed to 60 miles an hour will augment the loss to 1.3 that which occurs at 30 miles. CHAPTER XI. THE FRONT END. 96. Definitions. For the purpose of this discussion the term " front end" refers to all that portion of a locomotive boiler which is beyond the front tube-sheet. It includes the extending shell of the boiler which forms the smoke-box, and in general all mechanism which is therein contained, such as steam- and exhaust-pipes, netting, dia- phragm and draft-pipes. It also includes the stack. The front end as thus defined is to be regarded as an apparatus for doing work, receiving energy from a source of power, and delivering a portion thereof in the form of a specific result. The source of power is the exhaust steam from the cylinders, and the useful work accomplished is represented by the volumes of furnace gases which are delivered against the difference of pressure existing between the smoke-box and the atmosphere. That the power of the jet may be sufficient, it is neces- sary that the engines of the locomotive exhaust against back pressure. The presence of the back pressure tends to lower the cylinder perform- ance, and it is for this reason that designers of front ends have sought to secure the required draft action in return for the least possible back pressure. In other words, the effort has been to increase the ratio of draft to back pressure, which ratio has been defined as the efficiency of the front end. 97. Draft and its Distribution. The office of the front end is to draw atmospheric air into the ash-pan, thence through the grate and fire, to draw the furnace gases through the tubes of the boiler, thence under the diaphragm and into the front end, and to force them out into the atmosphere. In order that this movement may take place, a pres- sure less than that of the atmosphere is maintained in the smoke-box, so that when the locomotive is working there is a constant flow from the atmosphere along the course named and back to the atmosphere again. The difference in pressure between the atmosphere and the 209 210 LOCOMOTIVE PERFORMANCE. smoke-box is spoken of as the draft, and, under normal conditions of running, is represented by from 4" to 10" of water. It has often been asked whether the draft, as observed from gauges attached to the front end, is in any way affected by changing the point of application of the gauge, a question which is not unreasonable in view of the intense activity which characterizes the circulation of gases through the front end. In undertaking a study of the draft action at Purdue, it was early determined to settle this point. Elaborate ap- paratus was prepared which would serve in giving simultaneous obser- vations of the pressure at several different points within the front end, and which by adjustments which could be quickly made, could be arranged to apply to a new series of points. By the use of this appara- tus it was found possible, in a brief interval of time, to explore thor- oughly the condition of pressure within the front end, and to make FIG. 103. Distribution of Draft. of record the pressures at eighty-four different points, the time re- quired being so short as to leave no question as to the constancy of the running conditions. Many different sets of observations thus obtained justified the conclusion that, so far as the space in front of the dia- phragm is concerned, draft-gauges attached at any points will all show identical readings.* The distribution of draft throughout the apparatus intervening be- tween the front end and the ash-pan is a matter of more than ordinary interest. Its value, as determined when the locomotive was operating under three different rates of power, is shown by Fig. 103. Calling * For a detailed description of this work see "American Engineer," May, 1902. p. 131. THE FRONT END. 211 the ash-pan pressure zero and referring to the 40-mile curve of this diagram, the draft in the fire-box, as measured by a gauge attached to a hollow stay-bolt, is 1.5" of water, between the front tube-sheet and the diaphragm (section H) it is 3", and between the diaphragm and the base of the stack (section F) it is 5". The last value is that which is usually measured and referred to as the "draft." The per- centage of the total draft absorbed by various portions of the system is shown by Table LIII. TABLE LIII. PERCENTAGE OF TOTAL DRAFT REQUIRED. Speed, Miles per Hour. To Draw Air into Fire-box. To Draw Smoke and Gases through Tubes. To Draw Smoke and Gases under Diaphragm. 20 22.6 41.1 36.3 30 30.1 33.6 36.3 40 30.4 32.0 37.6 It will be seen that the relative resistance of different portions of the system varies but little under changes in speed. More than one- third of the total draft value is absorbed in drawing the gases past the diaphragm, approximately another third is absorbed in the passage of the tubes, and the remainder is available for drawing air from the ash- pan through the grate and fuel into the fire-box. It is really this latter value only which promotes combustion. Under present conditions of operation, it is thought necessary as a matter of practical importance to have the opening under the diaphragm ocmparatively small, in order that the velocity of the gases may be sufficiently high in passing to clear the front end of cinders. The results show that the cost of this self-clearing action, as measured in the draft absorbed, is so great as to suggest the possibility of a material improvement in design at this point. If the diaphragm could be abandoned, or if the area under it could be materially increased without introducing new difficulties, the efficiency of the front end could be greatly improved. It is often assumed that a locomotive having a wide fire-box and large grate can be operated with a larger exhaust-tip than one having a narrow fire-box and smaller grate and that since the larger grate per- mits a given amount of fuel to be burned at a lower rate of combustion, the work which the steam-jet has to do is diminished. Experience on the road, on the other hand, has shown that differences in grate area do not necessarily involve, or ordinarily permit, changes in the draft 212 LOCOMOTIVE PERFORMANCE. appliances. The reason for this is to be found in the fact that, whether the grate be large or small, two-thirds of the difference in pressure between the front end and the atmosphere is absorbed in moving the air and gases through the tubes and under the diaphragm. The vol- ume of air and gases to be moved is not materially changed by chang- ing the area of the grate, and while the difference in pressure between the atmosphere and the fire-box will be less with the large grate than with the small one, this difference is insignificant in amount when com- pared with the total draft action which is required to stimulate the movement through the remaining portion of the systems. 98. The Action of the Exhaust- jet. A study of the foim and ac- tion of the jet of exhaust steam in the front end of a locomotive was undertaken at the Purdue Laboratory, in cooperation with a commit- tee of the Master Mechanics' Association.* Prior to this time it had been common to assume that the action of the exhaust-jet was similar to that of a pump, that the exhaust from each cylinder-end supplies a ball of steam which fills the stack very much as the piston of a pump fills its cylinder, and which, by virtue of its own momentum, pushes before it a certain volume of the smoke-box gases until it passes out of the top of the stack. In obedience to this theory, experimenters had sought to so design their apparatus that the spread of the jet would be sufficient to fill the stack at its base. A knowledge of the form of the jet was therefore regarded as of prime importance. With this understanding of the general problem the Purdue laboratory entered upon its work. The apparatus employed is shown in part in connection with Fig. 104, which is a drawing to scale, showing a cross-section through the center of the stack of the front end experimented on. In line with the center of the stack and exhaust-pipe, cast-iron sleeves .were fastened to the outside of the smoke-box. Through these were fitted pipes 1, 2, 3, 4, and 5, arranged to slide in and out across the smoke-box, and having their inner ends turned down to a fine tip with sharp edges, and the edges bounding the orifices. The body of each pipe was graduated to tenths of inches, the scale reading from a reference-mark fixed to the sleeve on the outside of the smoke-box. If the zero of the scale were brought under the reference-mark, the inner tip of the pipe would be directly under the center of the stack; that is, directly over * For a full account of the important work accomplished by this committee, see "Proceedings of the American Railway Master Mechanics' Association," Vol. XXIX., 1896. THE FRONT END. 213 FIG. 104. 214 LOCOMOTIVE PERFORMANCE. the center of the exhaust-pipe. It is obvious that if the tip of any pipe be surrounded by the jet of exhaust steam, the velocity of the latter will tend to carry steam through the pipe and to discharge it into the atmosphere outside of the smoke-box. It can be shown also that the force which the steam would exert in its effort to pass the pipe is a function of its velocity; hence, by observing the force or pres- sure, the velocity may be calculated. Pressures were observed by having the outer ends of each sliding-pipe connected by rubber tub- ing with one leg of a manometer or U shaped glass tube, which Avas fastened to the wall of the laboratory. These U tubes were partially filled with mercury, the displacement of which gave the pressure trans- mitted through the tube. If, for example, the tip of any particular pipe were in the jet of steam, its manometer would show pressure greater than that of the atmosphere; if it were withdrawn from the jet its manometer would indicate a pressure less than that of the atmosphere.* Besides these sliding-pipes in the smoke-box, the stack was fitted with three pipes which had plain ends projecting beyond the inner wall about a quarter of an inch. These side orifices were each connected with a manometer. They served to show the extent of pressure or vacuum existing within the stack at points where they were attached. Their exact location is shown by the dimensions in Fig. 104. The normal draft was measured by two different manometers con- nected with the front end, and was permanently registered by a Bristol recording-gauge. Indicators were used to show the steam distribu- tion in the engine cylinders, and a special indicator fitted with a light spring gave a fine record of the back-pressure line. This pressure was also recorded by a Bristol gauge. A Boyer speed-recorder served as a means for maintaining constant speed conditions. Observations were made as follows: Desired conditions of speed and steam pressure having been obtained, the adjustable pipes (1, 2, 3, 4, and 5), Fig. 104, were withdrawn from the smoke-box sufficiently to bring them entirely clear of the exhaust-jet, usually to a distance of 4* : nches from center of jet. Then, upon signal, all manometer gauges were read and all other observations taken, the readings being taken simultaneously. Each sliding-pipe was then moved inward a tenth of an inch and readings repeated, after which they were moved * See "A Gliinpse of the Exhaust-pipe," Proceedings of the Western Railway Club for October, 1895. THE FRONT END. 215 another tenth, and so on until the tip of all the pipes reached the center of the exhaust-jet. The readings thus obtained from the pipes 1, 2, 3, 4, and 5, Fig. 104, were entered upon a half-sized drawing represent- ing a portion of the cross-section of the smoke-box, the position of each entry showing the exact location to which the numerical readings applied. Upon the diagram of pressures thus obtained lines were drawn through points for which the pressure was zero. These lines were assumed to represent the border of the steam-jet. Reduced in scale, they are the full lines given in Figs. 107 to 118, the full signifi- cance of which is to be hereafter considered. Next, lines were con- structed by connecting points showing a pressure of 0.5 inch of mer- cury, and still other lines by connecting points representing a pressure of 1.5 inches of mercury; and, finally, in some cases the curves corre- sponding with 2.5 inches of mercury were located. These lines of equal pressure are those which appear in Figs. 107 to 118, but the numbers given in connection therewith represent the velocity in feet per -second, as calculated from the pressure already noted. The curves are, in fact, equal velocity curves as well as curves of equal pressure. Confining the discussion for the present to matters affecting the action of the jet, it may be said with certainty that the exhaust-jet acts upon the smoke-box gases (1) to induce motion in those portions which immediately surround it, and (2) to enfold and entrain the gases which are thus made to mingle with the substance of the jet itself. The induced action, which, for the jets experimented upon, is by far the most important, may be illustrated by means of Fig. 105. The arrows in this figure represent, approximately, the direction of the cur- rents surrounding the jets. It will be seen that the smoke-box gases tend to move toward the jet, and not toward the base of the stacks, at which point they are to leave the smoke-box. That is, the jet, by virtue of its high velocity and by its contact with surrounding gases, gives motion to particles close about it, and these, moving on with the jet, make room for other particles which are farther away. As the enveloping shell of gas approaches the top of the" stack its velocity decreases and it becomes thinner and thinner, all as shown by Fig. 105. All parts of the jets require gases to work upon, the upper as well as tho lower part. Gauges attached to the side of the stack show a vacuum, because the gases needed for the upper portion of the jet can reach it only by coming in around the jet lower down. In other words, the action of the upper part of the jet induces a vacuum in the lower part 216 LOCOMOTIVE PERFORMANCE. of the stack, just as the action of the jet as a whole induces a vacuum in the smoke-box. It will be shown later that as the amount of work FIG. 105. to be done by the exhaust is increased the jet becomes smaller, thus making room for larger volumes of gas to pass between it and the THE FRONT END. 217 stack; the velocity, both of the jet and of the induced currents, in- creasing. It is of interest in this connection also to note that of the gauges attached to the side of the stack (Fig. 105), a, which is 13 inches above the base, always gave about one and one-half (1.5) times the reduction of pressure which was recorded by the gauge attached to the smoke-box, and that the second gauge, which is 10 inches from the top of the stack, gave approximately six-tenths (0.6) of the value recorded in the smoke-box. That there is some intermingling of the smoke-box gases with the steam of the jet is made evident by the appearance of the combined stream as it issues from the top of the stack. The manner in which this intermingling takes place will be seen from the following consid- erations : Any stream flowing from a nozzle through a resisting medium will have a higher velocity at its center than at its circumference or sides, that is, the particles at the center of the jet move at a higher velocity than those on the outside, the latter being held back by contact with the surrounding gas. The result of the different velocities in the same stream is a wave motion of the individual particles of which the stream is composed. Thus, the path of any one of these particles may be shown by Fig. 106, B, but the exact form and frequency of the loops FIG. 106. will depend upon the relation between these differences in the velocity of particles in different portions of the stream and the actual mean velocity of the jet. If the velocity of the jet is high, and differences for different portions of the cross-section are not great, the loop may disappear, the path appearing as shown by Fig. 106, A. With a still 218 LOCOMOTIVE PERFORMANCE. higher mean velocity and a smaller difference the loops would ap- proach the form shown by Fig. 106, C. The exhaust-jet appears to take this latter form. Measurements to determine its velocity show that particles in the center move much more rapidly than those near the outside, and other measurements to determine the form of the jet as a whole define a boundary which is neither a straight line nor a regu- lar curve, but which agrees closely with the form given by Fig. 106, C. All this shows that the jet is stepped off in nodes, which under given conditions remain fixed in position. This conclusion, based upon meas- urements of the jet, is confirmed by the appearance of the jet as seen in an engine running with the front end open. The jet, when thus viewed, exhibits one or more bright spots which remain in a fixed posi- tion. It is through the wave-like action of the particles making up the steam-jet that the surrounding gases are intermixed with the jet. It is clear that any design of nozzle which will serve to subdivide the stream, or to spread it so as to increase its cross-section, will assist the jet in its effort to entrain the gases, but it is not clear that there is any gain to be realized in such a result. It is possible that as the mixing action is increased the induced action may be diminished, and that the sum total of the effect produced may remain nearly constant. The work which has thus far been done is not conclusive on this point, but the evidence tends to show that the more compact and dense the jet the higher its efficiency. It is certainly clear that for the jets ex- perimented upon the mixing action is hardly more than incidental to the induced action, the latter constituting the influence through which the work of the jet is chiefly accomplished. The similarity existing between the action within the front end of a locomotive and that within an injector will at once suggest itself, but the resemblance is not perfect. In an injector the jet of steam is almost immediately and completely condensed. It loses its iden- tity by intermixing with the water to which it imparts its energy. The mixture is complete before the water-orifice, through which the com- bined stream must pass, is reached. In the front end of a locomotive, on the other hand, the mixing of the jet with the stream of gas upon which it acts proceeds but slowly, and with the apparatus now in use ; it is doubtful if the process is ever completely accomplished. 99. Form and Character of the Jet. In Figs. 107 to 118 the steam- jet is represented by a series of lines. The full or outside lines are assumed to represent the boundary of the jet. They do, in fact, pass through points for which the velocity of the steam and gases is so low THE FRONT END. that their impress upon the bent tubes is only sufficient to balance the pressure of the atmosphere. Each dotted line inside of those already referred to passes through a series of points where the velocities are equal, the value of which in terms of feet per second is given. A comparison of the form and location of similar lines in different dia- grams will show the effect produced by the different combinations of mechanism employed. 100. The Jet as Affected by Changes in Speed of the Locomotive. The three jets shown by Figs. 107 to 109, inclusive, were obtained MILES PER HOUR FIG. 107. MILES PER HOUR 3$ REVOLUTIONS PER MINUTE 188 POUNDS Of .STEAM PER HOUR 81?2 FIG. 108. MILES PCR HOUR 4 REVOLUTIONS PER MINUTE *4 POUNDS Of STEAM PER HOUR 867O FIG. 109. under similar conditions, except that for Fig. 107 the speed was 25 miles per hour; for Fig. 108, 35 miles; and for Fig. 109, 45 miles. Each increment of speed results also in a larger volume of steam de- livered, and, consequently, a greater reduction of pressure. The dia- grams show that the rapidity of the exhaust impulse does not greatly affect the character of the interior of the jet, also that the spread of the jet diminishes as the speed is increased; or, in other words, as the volume of steam passing the nozzle becomes greater, the increased 220 LOCOMOTIVE PERFORMANCE. velocity of both the jet of steam and the body of gas immediately sur- rounding it, together with the reaction of the latter, probably consti- tute the causes which operate to narrow the jet. The velocity curves, which in Fig. 108 are pretty evenly distributed throughout the body of the jet, are in Fig. 109 crowded together, giving evidence in the latter case of a very dense and powerful jet. The results show that the more rapidly moving, though smaller, jet (Fig. 109) is more effective in pro- ducing draft than that which is shown by Fig. 107. 101. The Effect upon the Jet of Changes in the Height of the Bridge. An essential feature of the single exhaust-pipe is the so-called bridge, which maintains a separate steam-passage for each cylin- der for a "portion of the length of the pipe. It is above the top of this bridge only that the steam exhausted from both sides of the loco- motive intermingles. The purpose of the bridge is a twofold one: it prevents the exhaust of one side from blowing through into the ex- haust-passage of the other side, and when properly designed it is the means of making that side from which the strongest stream is passing assist by induction the exhaust from the other side. To determine the best height of bridge exhaust-pipes of four different designs were tested by the Master Mechanics' Committee, all being the same except in the height of the bridge and in the proportions of those parts adjacent thereto. Two nozzles representing the extreme conditions with reference to this detail were employed in experiments designed to determine the character of the jet, with results which are to be seen by comparing Figs. 107 to 109 with Figs. 110 to 112. Figs. 107 and 110 represent identical conditions, except as to the height of the bridge, as do also Figs. 108 and 111, and Figs. 109 and 112. An examination of the velocity curves of these figures will show that for similar condi- tions of running the jet for the two groups is nearly identical. It will be of interest to add that, basing their conclusions upon efficiency tests made at Chicago, the committee concluded that, whenever the bridge was less than 12", some loss of efficiency resulted, though such loss was not great even when the height of the bridge was reduced to 5". They recommend that the bridge be not less than 12". They recommend, also, that where long exhaust-pipes must be employed, the increased length should always be secured by extending that portion of the pipe which is above the bridge. This recommendation in effect contemplates a constant height of bridge for all lengths of pipes. To the discussion with reference to the bridge it will be of interest to add some reference to the "choke." THE FRONT END. 221 That part of the steam-pipe between the bridge and the outside wall, where the contraction of the passage is greatest, has generally been referred to as the choke. An early committee had recom- mended that the area of the choke be 80 per cent of that of the nozzle, and the committee of 1896 did a large amount of work upon pipes having this proportion. Not satisfied, however, with the valid- ity of the conclusions, the later committee extended its research to involve a large range of proportions, with results sustaining the con- fOUNDS OF STEAM PER HOUR 090 FIG. 110. ILES PER HOUR 4* POUNDS OF STEAM PER HOUR (? FIG. 111. FIG. 112. elusions; first, that the area of the choke should not be less than the area of the nozzle; and, second, that whenever, with a given pipe, it is necessary to sharpen the exhaust action, the result should be secured by contracting the nozzle and not by contracting the choke. 102. Jets Formed by a Steady Blast of Steam. Figs. 113 to 115 represent jets which were obtained by blocking the slide-valves clear of their seats, and by slightly opening the throttle. That the results thus obtained might be compared with those presented by the earlier figures, the throttle was so adjusted as to pass similar quantities of 222 LOCOMOTIVE PERFORMANCE. steam. Thus, with the locomotive running at 25 miles per hour (Fig. 107) 6090 pounds of steam were exhausted each hour. The steady blow corresponding is represented by Fig. 113, for which the discharge equaled 6400 pounds per hour. The agreement is sufficiently close for practical purposes. The similarity in the form of the jets serves to fully explain the success of the "blow" in producing draft. A critical comparison of these jets fails to disclose any important difference, except in the POUNDS OF STEAM PER HOU FIG. 113. FIG. 114. FIG. 115. character of the curves bounding their outline and in that of the velocity curves. The lines in the jets produced by blow are more sinuous than in the jets which are sustained by exhaust section. Cor- responding measurements of the draft show that the steady jet pos- sesses substantially the same draft-producing power as the intermittent exhaust, the essential factor being the weight of steam exhausted in a given time. 103. The Form of the Jet as Influenced by Different Tips. At the time when the Master Mechanics' Committee made its investi- THE FRONT END. 223 gation three forms of tips were in common use. These are designated as X, Y, and Z, Fig. 119. The tip X ends in a plain cylindrical portion 2" in length; the tip Y is contracted in the form of a frustrum of a cone; and the tip Z is in the form of a plain cylinder, ending in an abrupt cylindrical contraction. The forms of the jets delivered by these tips were in each case denned by methods already described. It was found that while the form was nearly the same in all cases, the jet delivered by the tip X has the least divergence, and that of the tip FIG. 116. FIG. 117 FIG. 118. Z the greatest. As to draft-producing qualities, the committee was unable to present convincing proof in favor of any one of the forms experimented upon, the results obtained being practically identical. The opinion is expressed, however, that the highest efficiency will be obtained from the jet which is most dense, and for this reason they were inclined to favor the form X. 104. The Form and Efficiency of the Jet as Affected by Bars over the Tip. It is well known that engines which refuse to steam may sometimes be made to do so by bridging the exhaust-nozzle with 224 LOCOMOTIVE PERFORMANCE. a small piece of round iron or by a bar or bars having a knife-edged cross-section .designed to spread the jet and at the same time impede its motion as little as possible. Experiments were therefore made by the committee of 1896, both upon round bars and upon crosses, shown by Fig. 120, with results confirming the experience on the road. The effect of such a bar upon the form of the jet is shown by Fig. 116, it crowds the interior portions outward. The jet is not made materially larger, but the curves representing different velocities are crowded close together on either side . As to the effect of such a device on the efficiency of the jet, it was found that the draft was improved, but in all cases the presence of the bridge so much increased the back pressure that the efficiency of the front end was reduced. In other words, the presence of the bar or cross produced the same effect upon the back pressure as the substitution of a smaller nozzle without the bars. It was found that cross-bars (Fig. 120), not wider than f of an inch, and having the lower portion shaped to a sharp knife-edge, placed on the nozzle, or any distance not greater than one inch above it, increased the back pressure, and that wider bars increased the back pressure at a greater distance from the nozzle. In view of these facts, and also in view of the fact that all evidence goes to show that for highest efficiency the jet should be kept as well compacted as pos- FIG. 119. sible, the practice of splitting it up was not commended. It is believed to be better practice in cases where the draft is unsatisfactory to reduce the diameter of the exhaust-nozzle than to attempt to secure the de- sired result by employing a larger nozzle with a bridge above it. 105. The Form of the Jet as Affected by Stack Proportions. It has already been suggested that the jet adapts itself to the proportions of the stack through which it discharges, and that it shows a disposition to avoid contact with the sides of the stack until very near' its top. The facts in the case are better and more definitely defined by Fig. 117, which shows the form of the jet in the presence of a false lining or choke in the stack, so proportioned as to reduce the effective diameter of the stack from 16", its normal size, to 12". It is evident that the presence of such a device acts as a throttle on the delivery of the combined stream of gases, and that by so doing it produces a ma- THE FRONT END. 225 terial reduction in the velocity of the current within the smoke-box, or points immediately about the jet and below the stack. The re- duced velocity allows the stream to broaden out, but it narrows again, and doubtless in some portions of its cross-section its velocity increases as it approaches the base of the contracted stack. Fig. 118 may be employed as a convenient reference in connection with the two pre- ceding figures. Bridge J Bridge FIG. 120. 106. The Jet as Affected by Cut-off. For the purpose of deter- mining the effect, if any, upon the form of the jet of changes in cut- off, the load at the locomotive draw-bar was held constant, the speed controlled by throttle-opening, and the engine run with the reverse- lever in several different positions, the form of the jet being deter- mined meanwhile by methods already described. The conclusion drawn from these tests is that changes in the cut-off have no effect upon the form and character of jet, except in so far as they affect the quantity of steam discharged. 107. The Stack Problem. Prior to 1902 experiments under service conditions involving front-end appliances were but little concerned with the problems of the stack.* This was due to no lack of appreciation of the importance of the stack, but rather to the complicated nature of the problem, and to the fact that the need of information concerning other details had been more urgent. Previous to this time it frequently happened that one locomotive, having twice the cylinder power and boiler capacity of another, would be fitted with a stack of the same size. In this case the furnace gases and exhaust steam of the larger locomotive were re- * A few years ago there were undertaken at Hanover, Germany, an elaborate series of tests designed to solve the stack problem. The results, which in this country have been known as those of the von Borries-Troske tests, have been widely pub- lished. They were conducted upon a full-sized model front end, steam being sup- plied from a stationary boiler, and air in imitation of the smoke-box gas being ad- mitted to the front end through a restricted opening. While they have undoubtedly served well to guide the practice of German railways, they have not greatly influenced that of our own country. 226 LOCOMOTIVE PERFORMANCE. quired to find their way through an opening which appeared to be not too large for the smaller locomotive. It was, in fact, the practice of some of the important roads to make all stacks of the same diameter, regardless of the size of the locomotive upon which they were to be used. It is evident that in such cases the highest efficiency could only result from a chance combination, and the action in many cases was necessarily far from satisfactory. Moreover, these conditions did not prevail from choice, but rather because the information necessary to guide practice along more logical lines was not available. In view of these facts, the American Engineer of New York City, as represented by Mr. G. M. Basford, its editor, having secured the cooperation of a committee of representative motive-power officials,* agreed to become a patron of the locomotive laboratory of Purdue University, for the purpose of securing information which would aid in the logical design of such apparatus. f 1 08. The Plan of the Tests, as outlined by the laboratory, involved the recognition of the following variable factors : (a) Form of Stack. With reference to this factor, two contours only were provided, one straight and one tapered. (6) Diameter of Stack. Each form of stack was developed into a * This committee consists of Mr. Robert Quayle, Chicago & Northwestern; Mr. A. W. Gibbs, Pennsylvania; Mr. G. R. Henderson, Atchison, Topeka & Santa Fe; Mr. F. H. Clark, representing Mr. F. A. Delano, Chicago, Burlington & Quincy; Mr. W. H. Marshall and Mr. H. F. Ball, Lake Shore & Michigan Southern; Mr. F. M. Whyte, representing Mr. A. M. Waitt, New York Central & Hudson River; Mr. W. S. Morris, Chesapeake & Ohio; Mr. C. H. Quereau, Denver & Rio Grande; Mr. C. A. Seley, Norfolk & Western; Mr. J. E. Sague, American Locomotive Company; Mr. E. M. Herr, who was Assistant Superintendent of Motive Power of the Chicago & Northwestern when the Master Mechanics' Association tests of 1896 were made; and Mr. H. H. Vaughan, who, while not then connected with a railroad, was a recognized authority in matters of locomotive design. f By the arrangement which was entered into, it was agreed that the University should contribute to the proposed work, the equipment of its locomotive-testing laboratory and such portion of the time of its expert staff as could readily be made available; also that the "American Engineer" would meet the cost of special appara- tus, of supplies in excess of those usually required in the routine work of the labora- tory, and of such additional help as might be necessary. The committee already referred to assisted in an advisory capacity. A full description of the work thus organized and of the results drawn therefrom will be found running through the issues of the American Engineer for the year 1902. The work was carried on in connection with the University's locomotive, "Schenectady No. 2." This is an eight-wheeled engine, weighing 100,000 pounds, having a boiler the least diameter of which is 52 inches. THE FRONT END. 227 series of different diameters, the diameters being 9f inches, llf inches, 13f inches, and 15f inches respectively. The dimensions given apply to all portions of the straight stacks above the base, and to the least diameter of the tapered stacks. These diameters were chosen because of their close agreement with the diameters used in the von Borries- Troske experiments. (c) Height of Stack. Each of the eight stacks already described was made in five sections, the upper four sections each being 10 inches in height. This provision made it possible to employ either of the eight stacks in heights varying from 16 J inches to 56 J inches. (d) Exhaust-nozzles. It was not expected that the work should involve any investigation of exhaust-pipes or nozzles, this phase of the draft-appliance problem having been covered as already described. It was assumed, however, that in order to get the maximum efficiency of each different diameter and height of stack, it would be necessary to provide a variable height of pipe. (e) Power of the Locomotive. In the early discussion of the matter the question arose as to whether a combination of apparatus, which would give the highest efficiency when the engine was worked at light power, would prove the most efficient under conditions of heavy power, and that there might be no question arising from this source it was determined to make tests under each combination of apparatus, with the engine running at three different rates of power. It was deter- mined, also, that this could best be accomplished by running all tests under a constant steam pressure, a fixed cut-off, and a .wide-open throttle, the desired variations of power being obtained by varying the speed. The conditions chosen were approximately as follows: Steam pressure, 180 Ibs.; cut-off, one-third stroke; speeds for different power, 20, 30, and 40 miles respectively; the power under these conditions being approximately 260, 370, and 475. The amount of steam generated by the boiler, and discharged from the exhaust-tip, varies nearly in direct proportion to variations in power. 109. Conditions at the Grate. It is well known that the draft, as measured by a reduction of pressure within the front end, is affected by conditions other than those which control the action of the steam- jet. With the force of the jet remaining constant, the draft will vary with every change in furnace condition which serves in any way to affect the freedom with which the air is permitted to move through the grate and fire-box. Thus, all other things remaining constant, the draft is reduced by opening the fire-door, and increased by closing the 228 LOCOMOTIVE PERFORMANCE. ash-pan dampers. Similarly, a thin, clean fire, which offers but little resistance to the passage of air, results in light draft, while a thick, heavy fire, which impedes the movement of air at the grate, increases the draft. This being true, it was deemed essential to provide for constant conditions at the grate. As this could not be done in connection with solid fuels, the experimental engine was equipped for burning oil, and the air-openings into the fire-box so arranged that they would at all times be of fixed dimensions. The degree of ease, therefore, with which air found its way into the furnace was FIG. 121. unaffected by the condition of fire. Sections of the fire-box as arranged for the tests are shown by Fig. 121. no. The Experimental Stacks and Nozzles. The stacks em- ployed in the experiments are shown by Fig. 122. Each stack was made in sections, by the successive removal of which any desired length within the limits shown could be obtained. The several sec- tions were turned to fit one another and were held together by bolts acting upon outside lugs. The nozzles used in connection with the different stacks are shown by Fig. 123, and the details of their con- struction by Fig. 124. When in use the nozzles were fixed upon a short exhaust-pipe, designed in accord with the recommendation of the Master Mechanics' Committee. They consisted of a wrought-iron body, mounted on a cast-iron base, and fitted at the top with a suit- THE FRONT END. 229 able tip. All had the same diameter of tip, namely, 4J inches. Notation, in connection with stacks and nozzles, was employed as set No.3 1-D 1-C DIMENSIONS OF STACKS FIG. 122. forth by Figs. 122 and 123. A No. 7 stack is a straight stack 15| inches in diameter. A No. 7 A stack is a stack 26^ inches long, rUi-^L-ii 4 I g J * U & s 1 ! i -L .i Q N? ; I Oik - i r BmL 1 ? * f J ' -i-f 1 H-- > Unto No. 7 , No. 6 No. 5 No. 4 No. 3 No. 2 No. 1 FIG. 123. while a No. 7 D stack is a stack 56^ inches long. Similarly, a No. 8 A is a taper stack 15f inches in diameter at the choke, and 26 % inches 230 LOCOMOTIVE PERFORMANCE. long. The exhaust-pipes and nozzles were numbered. Thus, No. 1 pipe or nozzle is 5 inches high upon a 10-inch base, while nozzle No. 7 is 35 inches high upon a 10-inch base. While these will hereafter be FIG. 124. referred to as nozzles> they really combine in one design a pipe and nozzle, as will be seen by the drawing representing their construc- tion, Fig. 124. All nozzles were used upon the same base which, as shown, was 10 inches in height. THE FRONT END. 231 in. The Tests. The conditions under which tests were made may be summarized as follows : VASLABLE SPEED SERIES. Constants. Variables. Boiler pressure, 180. } g dg: 2Q 3Q ^ 5 } 23.8 VARIABLE CUT-OFF SERIES. Constants. Variables. Boiler pressure, 180. ~| Throttle, wide open. ', Cut-off, per cent of stroke, Speed, 40 miles per hour. 19, 23.8, 26.9, and 35. 194 revolutions per min. J With the exception of certain conditions, which could be omitted without interfering with the value of a series, tests were run under each condition specified above for each of 40 different stacks and these again were repeated for from two to seven different heights of exhaust- pipes. To these conditions must also be added others applying to an inside or sliding-stack, the results of which will be separately consid- ered. The work as carried out involved 1032 distinct tests, for each of which the locomotive was brought under specific conditions of running, and for each of which a considerable series of observations were made of permanent record. Besides these recorded tests there have been many other trial or rejected tests. In anticipation of a series of tests, the engine was warmed by pre- liminary running. It was then started for a test, and after the throttle and reverse-lever had been brought to their desired positions, the speed of the engine was made to vary by changing its load. When the speed and steam pressure were both that which were required, the signal was given and observations were taken. The engine was then stopped, and such changes in stack or nozzle were made as were necessary for" the next set of observations, after which the process above described was repeated. In this manner a day's work consisted in a succession of short runs, with intervals between to allow the change of equipment. 112. Results. A detailed statement of numerical results, obtained at a speed of 40 miles per hour under a cut-off of 27 per cent, is given in Table LIV While similar tables were developed under the same cut- off for speeds of 20, 30, 50, and 60 miles per hour and under a constant speed of 40 miles per hour, under several different cut-offs, it appears 232 LOCOMOTIVE PERFORMANCE unnecessary to transfer to these pages the whole of so voluminous a record.* 1 13. A Basis of Comparison. A careful study of all results made clear the fact that the relative value of the various devices tested could properly be determined by comparing the draft values resulting from their use. That is, that the best stack was that which, under given conditions, gave the best draft. Similarly, that the best height of nozzle was assumed to be that which gave the best draft. This state\ ment ignores the whole question of back pressure, but in the work described this is justified, because of the fact that the changes made as the experiments progressed were found to produce no measur- able effect upon the back pressure. While, therefore, in Table LIV. values are given for both back pressure and efficiency, the compari- sons which follow are concerned only with the draft, f * The full record will be found in pages of the American Engineer and Railroad Journal, to which references have previously been made. f In outlining the tests it was proposed to base all comparisons upon the effi- ciency of the jet, and efficiency was defined as the ratio of back pressure to draft. The assumption of such a measure is based upon the fact that the result which is sought by the use of any combination of draft appliances is a reduction of pressure within the front end, and that the force effecting such a reduction of pressure is a function of the pressure of steam in the passage between the cylinders and the exhaust- tip. The maintenance of considerable pressures in the exhaust-passages tends to reduce the economic performance of the engine, hence it is desired that the necessity for such pressures be, so far as practicable, avoided. The proposed measure of efficiency takes all this into account, for by its use that arrangement of apparatus which will give a desired reduction of pressure in the front end in return for the least back pres- sure will be the most efficient. Such a conception is perfectly logical. It is not new, but on the contrary is one which has been many times employed in the study of draft appliances. The preceding statement is general in its application. It applies not only to the tests under consideration but to all tests which may be made for the pur- pose of determining the value of this or that draft appliance. It happens, how- ever, that in the experiments under consideration, the exhaust- tip was of the same size for all tests. Furthermore, it appears as one of the significant results obtained from the tests, that a change in the height of the exhaust-pipe does not affect the back pressure by a measurable amount. Consequently, so far as the present study is concerned, the back pressure for any given condition of run- ning appears as a constant; and the efficiency which, in the general case, is a func- tion of both back pressure and draft, is left to depend upon draft alone. All this being true, it appears that effects resulting from changes in the front-end mechan- ism, such as were involved by the experiments under consideration, are quite as well shown by a direct comparison of draft values as by a comparison of efficiency values. Moreover, the draft values involve a single observation made under condi- THE FRONT END. 233 TABLE LIV. FORTY-MILE SERIES. CONSTANTS. Pounds of Steam used. Cut-off. f Miles per hour 40 \R. P. M. 194.4 f Per hour 12988 \ Per minute 216 f in inches 6.4 \ in per cent of stroke 26.9 M.E.P. 52.0 los RESULTS. Smoke-box Smoke-box Pressure. Pressure. y *o ^ > , >' .1 1. . ' $* fj fl 0) ? li d 1 n K %* m " w M 3 fe 3 a m 3 i jpSl |i o W I. II. III. IV. V. VI. I. ii. III. IV. V. VI. 1 2.7 .2 .043 .016 i 2.65 ' .4 .05 .018 2 2.24 .2 .0432 .019 2 2.23 .4 .0504 .023 3 3 2.52 .4 .0503 .02 Base 4 2.25 .3 .0468 .021 1-C 4 2.05 .7 .0612 .03 No. 1 5 2.25 .2 .0432 .019 5 2.13 .8 .0648 .03 6 2.4 .2 .0432 .018 6 2.2 .9 .0684 .031 7 2.55 .0 .036 .014 7 2.4 2.0 .072 .03 1 2.3 .2 .043 .019 1 2.45 1.3 .047 .019 2 2.25 .4 .0504 .022 2 2.25 1.4 .0504 .022 3 2.50 .6 .0577 .023 3 2.4 1.4 .0503 .021 1-A 4 2.3 .7 .061 .027 1-D 4 2.3 1.8 .0648 .028 5 2.22 1.7 .0612 .028 5 2.22 1.9 .0684 .031 6 2.2 1.7 .0612 .028 6 2.2 2.0 .072 .033 7 2.45 1.5 .054 .022 7 2.3 2.1 .0756 .033 1 2.2 .4 .05 .023 1 2.6 2.3 .083 .032 2 2.3 .4 .0504 .022 2 2.27 2.4 .0864 .038 3 2.35 .4 .0503 .021 3 2.35 2.6 .0936 .04 1-B 4 2.25 .8 .065 .029 2-A 4 2.0 2.4 .0864 .043 5 2.21 .9 .0684 .031 5 2.21 2.2 .0792 .036 6 2.4 .9 .0684 .028 6 2.4 2.0 .072 .03 7 2.25 .9 .0684 .03 7 2.45 1.6 .0578 .024 234 LOCOMOTIVE PERFORMANCE. TABLE LIV. (Continued). FORTY-MILE SERIES. Smoke-box Smoke-box Pressure. Pressure. Jb ,, ^ L >> kg - 1 i 1 1 ||l hi III z$ 11 a .2 g jj 6 1 III "fe 111 l! e p 02 ft o w 02 fc O n W I. II. III. IV. V. VI. I. II. III. IV. V. VI. 1 2.4 3.1 .111 .046 1 2.5 2.7 .097 .038 2 2.2 3.2 .1152 .052 2 2.27 2.8 .1008 .044 3 2.37 3.3 .1188 .05 3 2.34 2.7 .097 .045 4 2.2 3.0 .108 .049 3-C 4 2.3 3.1 .112 .049 2-B 5 2.25 2.9 .1044 .046 5 2.23 3.1 .112 .05 6 2.4 2.5 .09 .037 6 2.3 3.0 .108 .047 7 2.4 2.2 .0792 .033 . 7 2.4 2.8 .101 .042 1 2.45 3.5 .126 .051 1 2.6 2.4 .086 .033 2 2.15 3.8 .1368 .064 2 2.23 2.7 .0972 .044 3 2.38 3.7 .1332 .056 3 2.6 3.0 .108 .041 2-C 4 2.3 3.5 .126 .055 3-D 4 2.4 3.0 .108 .045 5 2.25 3.5 .126 .056 5 2.2 3.0 .108 .049 6 2.2 3.1 .1116 .051 6 2.2 3.1 .112 .051 7 2.2 2.7 .0972 .044 7 2.4 3.2 .115 .048 1 2.45 4.0 .144 .059 1 2.3 3.0 .108 .047 2 2.1 4.3 .1548 .074 2 2.3 3.3 .1188 .052 3 2.42 4.0 .144 .059 3 2.35 3.0 .108 .046 2-D 4 2.4 3.9 .1404 .058 4-A 4 2.2 2.9 .104 .047 5 2.21 4.0 .144 .065 5 2.3 2.5 .09 .039 6 2.4 3.4 .122 .051 6 2.2 2.0 .072 .033 7 2.25 3.2 .1152 .051 7 2.4 1.4 .05 .021 1 1 2.2 3.5 .126 .057 2 2.33 2.1 .0756 .032 2 2.35 4.1 .1476 .063 3 3 2.34 3.4 .122 .052 Base 4 2.3 1.9 .068 .030 4-B 4 2.25 3.6 .130 .058 No. 2 5 2.22 1.6 .058 .026 5 2.27 3.2 .152 .051 6 2.2 1.2 .043 .019 6 2.4 2.7 .097 .041 7 2.5 .9 .032 .013 7 2.6 2.1 .076 .029 1 2.6 2.3 .082 .032 1 2.5 4.3 .155 .062 2 2.4 2.6 .094 .039 2 2.3 4.6 .1656 .072 3 2.4 2.65 .096 .039 3 2.28 4.0 .144 .063 3-A 4 2.2 2.6 .094 .043 4-C 4 2.3 4.0 .144 .063 5 2.23 2.4 .086 .038 5 2.34 3.7 .133 .057 6 2.2' 2.1 .076 .034 6 2.4 3.0 .108 .045 7 2.5 1.7 .061 .025 7 2.3 2.6 .094 .041 1 2.5 2.7 .097 .039 1 2.8 5.3 .191 .068 2 2.3 2.6 .094 .041 2 2.27 5.0 .18 .079 3 2.5 2.9 .104 .042 3 2.41 4.4 .158 .065 3-B 4 2.3 2.9 .104 .045 4-D 4 2.4 4.3 .155 .065 5 2.25 2.9 .104 .046 5 2.4 4.2 .151 .063 6 2.2 2.6 .094 .043 6 2.4 3.8 .136 .057 7 2.65 2.4 .087 .033 7 2.35 3.2 .115 .049 THE FRONT END. 235 TABLE LIV. (Continued}. FORTY-MILE SERIES. Smoke-box Smoke-box Pressure. Pressure. i , ^ . .1 , 1 QQ Nozzle. Observed Back Pre sure, Lbs Inches of Water Observed Pounds Ca culated. Efficiency. 1 GQ Nozzle. V 03 s<6;3 3 of jdg Inches of Water Observed Pounds Ca culated. Efficiency. I. II. III. IV. V. VI. I. II. III. IV. V. VI. 1 1 2.27 4.0 .144 .064 2 2.35 2.8 .1008 .043 2 2.33 4.2 .151 .065 3 3 2.2 4.2 .151 .068 Base 4 2.45 1.9 .0684 .028 6-B 4 2.3 3.4 .1224 .053 No. 3 5 5 2.24 3.0 .108 .048 6 2.45 1.1 .0396 .016 6 2.15 2.4 .086 .040 7 2.25 .7 .0252 .011 7 2.25 1.9 .0684 .030 1 2.41 3.2 .115 .048 1 2.1 4.4 .158 .075 2 2.33 3.5 .126 .054 2 2.26 4.6 .1656 .073 3 2.4 3.1 .1116 .046 3 2.2 4.5 .162 .074 5-A 4 2.4 2.8 .1008 .042 6-C 4 2.35 3.9 .1404 .06 5 2.25 2.4 .0864 .038 5 2.27 3.4 .1224 .056 6 2.2 2.0 .072 .033 6 2.2 2.9 .104 .047 7 2.35 1.4 .0503 .021 7 2.25 2.4 .0684 .038 1 2.37 3.2 .115 .049 1 2.1 4.8 .173 .082 2 2.4 3.9 .1402 .058 2 2.25 5.0 .18 .08 3 2.36 3.4 .1221 .052 3 2.3 5.2 .1872 .081 5-B 4 2.45 3.2 .1152 .047 6-D 4 2.4 4.4 .1584 .066 5 2.25 3.2 .1152 .051 5 2.2 4.1 .1476 .067 6 2.3 2.8 .1008 .044 6 2.25 3.3 .1188 .053 7 2.4 2.8 .0792 .033 7 2.5 2.8 .1008 .04 1 2.3 3.2 .115 .05 1 2 2.25 3.6 .1296 .058 2 2.30 2.8 .1 .043 3 2.4 3.6 .130 .054 3 5-C 4 2.4 3.7 .1332 .055 Base 4 2.3 1.7 .061 .027 5 2.25 3.6 .1296 .057 No. 4 5 6 2.25 3.4 .122 .054 6 2.3 .75 .027 .012 7 2.4 3.1 .1116 .046 7 2.35 .5 .018 .008 1 2.45 3.4 .122 .05 1 2.22 3.4 .122 .055 2 2.26 3.8 .1368 .059 2 2.35 3.6 .129 .055 3 2.3 3.6 .13 .056 3 2.13 3.0 .108 .051 5-D 4 2.45 4.0 .144 .054 7-A 4 2.33 2.7 .097 .042 5 2.22 3.9 .1404 .063 5 2.25 2.4 .086 .038 6 2.3 3.8 .1368 .059 6 2.3 1.6 .057 .025 7 2.35 3.5 .126 .054 7 2.3 1.1 .039 .017 1 2.4 3,6 .129 .054 1 2.18 3.6 .129 .059 2 2.3 3.5 .126 .055 2 2.25 4.2 .158 .07 3 2.2 3.4 .122 .055 3 2.30 3.6 .130 .056 6-A 4 2.45 2.7 .0972 .04 7-B 4 2.4 3.4 .122 .051 5 2.28 2.4 .0864 .036 5 2.22 3.0 .108 .049 6 2.4 1.8 .0648 .027 6 2.2 2.4 .086 .039 7 2.4 1.3 .0468 .019 7 2.2 1.9 .068 .031 236 LOCOMOTIVE PERFORMANCE TABLE LIV. (Continued). FORTY-MILE SERIES. Smoke-box Smoke -box Pressure. Pressure. " "e8 0) 03 t 1^ *g fe B i g T3Pn J **" > OT? ?? 1 Nozzle. jit |o os.2 V S 0> 1 W 44 1 0) 1 o hu Jdg |ll Pounds culate Efficien< I. II. III. IV. V. VI. I. II. III. IV. V. VI. 1 2.21 3.6 .129 .058 1 2.4 4.9 .176 .073 2 2.35 4.6 .166 .071 2 2.23 4.7 .169 .075 3 2.37 4.0 .144 .061 3 2.4 4.2 .151 .069 7-C 4 2.35 3.8 .136 .058 8-D 4 2.4 3.6 .130 .054 5 2.22 3.8 .136 .062 5 2.8 3.6 .129 .036 6 2.3 3.2 .1156 .05 6 2.4 2.7 .097 .04 7 2.45 2.9 .104 .043 7 2.25 2.3 .082 .036 1 2.21 3.9 .14 .067 Nor- 2 2.4 4.7 .169 .070 mal, 3 2.3 4.2 .151 .066 Petti- 7-D 4 2.2 4.2 .151 .069 coat 5 2.25 4.2 .151 .067 Pipe 6 2.35 3.7 .133 .057 In * 2.4 3.9 .14 .058 7 2 35 3 5 126 054 Nor 1 2.22 3.5 .126 .057 mal, 2 2.3 3.6 .129 .056 Petti- 3 2.4 3.0 .108 .045 coat 8-A 4' 2.35 2.3 .082 .035 Pipe 5 2.33 1.9 .068 .029 Out * 2.4 3.8 .136 .057 Q 9 QA 1 25 045 02 7 2.35 .9 .032 .014 Slid- 1 2 2.8 2 2 3.6 Q A .129 108 .046 OJ.Q 1 2.25 4.1 .148 .066 ing A 3 2.3 2.7 .097 .042 9 9 A 4 2 AO 8-B 3 4 2 45 2.35 3.5 2.9 .126 .104 .051 .045 Slid- 1 2 2.6 2.2 3.0 3.4 .108 .122 .041 .055 5 2.5 2.4 .086 .035 ing B 3 2.3 3.1 .111 .048 2 4 1 9 068 028 7 2.25 1.4 .05 .022 1 2.6 3.5 .126 .048 OllCl" 2 2 3 Q q 140 AA1 1 1.27 4.5 .162 .071 ing C 3 2^5 4.2 .151 .06 9 2 OK 4 4 158 07 3 2.4 4.0 .144 .06 1 2.4 4.0 .144 .06 8-C 4 2.25 3.4 .122 .054 . " 2 2.2 4.1 .147 .067 5 2.7 3.0 .108 .037 ing -U 3 2.4 4.0 .144 .06 6 2.4 2.4 .086 .036 7 2.2 1.9 .068 .031 * Normal. THE FRONT END. 237 114. The Effect upon Stack Proportions of Changes in Speed and Cut-off. The experiments were made to involve a great variety of conditions with reference to speed and cut-off, in order that the differences in the efficiency of the stack arrangement, which resulted from such changes, might be known. From a careful study of all of this data it can be stated with confidence that, within reasonable limits, no change in the condition of running affects in any marked degree the efficiency of the stack; that a stack and nozzle which are satisfactory at Stack A Height 26!^ Stack B Height 36 'j Stack Diameter. Inches FIG. 125. Straight Stack, Nozzle No. 2. Stack Diameter. Inches FIG. 126. Straight Stack, Nozzle No. 2. one speed will be found satisfactory at all speeds, and if satisfactory under one condition of cut-off they will be found reasonably satisfactory under all conditions of cut-off. This conclusion, which is of the high- est importance, cannot be perfectly demonstrated without the use of an elaborate display of data, but its truth is well illustrated by the similarity of the curves representing the draft obtained under three different speeds by means of several different stacks, Figs. 125 to 132.* tions favorable to accuracy, and, consequently, they supply a better basis for compari- son than efficiency, which depends on two observations, one of which is rather difficult to obtain. * These figures have been chosen to illustrate the effect upon the draft of changes in the proportions of the stack and height of the nozzle. Figs. 125 to 128 present results obtained by using straight stacks of various heights and diameters, in connec- tion with an exhaust tip 2 inches below the center of the boiler Figs. 129 to 132 show results obtained by the use of tapered stacks and an exhaust tip 20 inches above the center of the boiler. The steam pressure, throttle opening, and cut-off were the same for all tests represented by these diagrams. 238 LOCOMOTIVE PERFORMANCE. Curves representing results obtained at different cut-offs present the same characteristics. With a given weight of steam discharged, Stack C Height 43 '4 Height 56Vfc Stack Diameter. Inches FIG. 127. Straight Stack, Nozzle No. 2. Stack Diameter. Inches FIG. 128. Straight Stack, Nozzle No. 2. whether in the heavy exhausts incident to slow speed, or the more rapid impulses which are sent forth at higher speed, the draft result- stack A Height 26 vg StackB Height 36% M.P H 11 '4 13V Stack Diameter. Inches FIG. 129. Tapered Stack, Nozzle No. 7. Stack Diameter. Inches FIG. 130. Tapered Stack, Nozzle No. 7. ing is practically the same. But whenever the weight of steam dis- charged per minute changes, the draft will change. That this is true for changes of speed is made clear by Table LV., in which is given a THE FRONT END. 239 comparison of changes in draft values with corresponding changes in the volume of steam exhausted. Stack C Height Stack D Height 5GJ4 Stack Diameter. Inches FIG. 131. Tapered Stack, Nozzle No. 7. Stack Diameter. Inches FIG. 132. Tapered Stack, Nozzle No. 7. TABLE LV. A COMPARISON OF CHANGES IN DRAFT VALUES WITH CORRESPOND- ING CHANGES IN THE VOLUME OF STEAM EXHAUSTED. Nozzle No. 3. Exhaust-tip on Center Line of Boiler. Stacks. Change in Ratio of Ratio of Speed, Miles. Change in Draft. Change in Steam Used Height. Diameter. Form. per Hour. 864 Taper 20 to 30 .25 .33 26^ Q Taper 30 to 40 .30 .19 36 1 9- Taper 20 to 30 .30 .33 36, 9j Taper 30 to 40 .23 .19 4ft 9i Taper 20 to 30 .15 .33 46^ 9i Taper 30 to 40 .19 .19 5ft 9i Taper 20 to 30 .25 .33 56^ 9i Taper 30 to 40 .14 .19 2ft Hi Taper 20 to 30 .23 .33 2ft Hi Taper 30 to 40 .14 .19 36, Hi Taper 20 to 30 .29 .33 36i Hi Taper 30 to 40 .13 .19 4ft Hi Taper 20 to 30 .24 .33 4ft n\ Taper 30 to 40 .11 .19 ^ Hi Taper 20 to 30 .26 .33 oft Hi Taper 30 to 40 .15 .19 Average. . . .21 .26 Following the first line of this table, it will be seen that when the speed is changed from 20 to 30 miles, the draft is increased in the ratio 240 LOCOMOTIVE PERFORMANCE. of from 1 to 1.25, and the steam exhausted is increased in the ratio of from 1 to 1 .33, which is an approach to agreement. Similar comparisons are shown for various diameters and heights of stacks, the average of all being, for changes in draft, 1.21, and for changes in steam exhausted, 1.26, which makes the check very close. The result of comparisons in- volving other values tends to further confirm the general conclusion. 115. A Review of Best Results. From an inspection of the re- sults of all the tests, the highest draft readings have been selected for each condition of speed and cut-off. These, with the designation of stacks and nozzle employed in securing them, are set down in the columns of Table LVI. Results thus chosen constitute approxi- mately 5 per cent of the whole number obtained, and may be accepted as representing the best results obtainable under any combination of stack and nozzle involved by the experiments. TABLE LVI. STACK AND NOZZLE COMBINATION GIVING BEST RESULTS. Speed, Miles per Hour. Cut-off Notch. Draft, Inches of Water. Nozzle Num- ber. Stack Num- ber. Speed, Miles per Hour. Cut-off Notch. Draft, Inches of Water. Nozzle Num- ber. Stack Num- ber. 20 5 3.4 2 7-D 50 5 5.5 1 8-D 20 5 3.3 1 6-D 50 5 5.4 1 6-D 20 5 3.3 2 4-D 50 5 5.4 2 6-D 20 5 3.2 3 7-D 50 5 5.4 3 8-D 20 5 3. 5 4-D 50 5 5.3 1 4-D 20 5 3. 2 6-D 60 5 6.6 2 4-D 20 5 3. 3 6-D 60 5 6.4 1 8-D 20 5 3. 4 6-D 60 5 6.2 1 6-D 20 5 3. 2 8-D 60 5 6.1 1 4-D 30 5 4.3 2 4-D 60 5 6.0 2 8-D 30 5 4.3 1 6-D 40 3 3.6 2 8-D 30 5 4.1 2 7-D 40 3 3.2 1 8-D 30 5 4.0 2 6-D 40 3 3.1 1 4-D 30 5 4.0 3 7-D 40 3 3.0 3 6-D 30 5 4.0 3 7-D 40 3 3.0 3 8-D 30 5 4.0 1 8-D 40 3 3.0 6-C 30 5 4.0 2 8-D 40 5 5.3 4-D 30 5 4.0 3 8-D 40 5 5.0 8-D 40 5 5.2 3 6-D 40 5 4.8 8-D 40 5 5.0 2 4-D 40 5 4.7 6-D 40 5 5.0 2 6-D 40 5 4.6 8-C 40 5 4.9 1 8-D 40 7 7.8 8-D 40 5 4.8 1 6-D 40 7 7.6 4-D 40 5 4.7 2 7-D 40 7 7.6 6-D 40 5 4.7 2 8-D 40 7 7.6 2 8-D 40 5 4.6 2 4-C 40 7 7.4 2 6-D This table shows at a glance that all the highest draft readings were obtained with the D stacks, which designation embraces those of THE FRONT END. 241 greatest length employed in the experiments (56 J inches). The C stacks, which are 10 inches shorter than the D stacks, appear in the table but three times, and in two of these instances the draft is inferior to that given by the higher stack. Stacks bearing even numbers are tapered, and those bearing odd numbers are straight. It is noteworthy that practically all of the numbers appearing in the table of best results represent tapered stacks. The stack numbers which most frequently appear are 4, 6, and 8, representing a diameter at the choke of 11J, 13f , and 15f inches respectively. Resorting now to a detailed study covering all the various heights and diameters of stacks experimented upon, and dealing first with the straight stacks, the large spots upon Figs. 133 to 136, inclusive, show the best diameters for each height of stack experimented upon, and for each height of nozzle employed. Thus, disregarding for the present the oblique line drawn upon these diagrams, the large spots upon Fig. 133 show the best diameter of stack for each height of nozzle em- ployed, when the height of the stack is limited to 26J inches; those upon Fig. 134 show the best diameter of stack for each height of nozzle employed, when the height of stack is limited to 36J inches. In each case the large spots represent the diameter of the experimental stacks giving the highest draft. The points are not located from curves. Since the experimental stacks varied one from another by steps of 2 inches, the exact diameter represented by a given large spot does not necessarily represent the most desirable diameter for the conditions defined; that we will hereafter proceed to find. Continuing to give attention to the black spots of the diagrams, it will be noticed that some of the larger spots have smaller spots con- nected with them by a horizontal line. In some cases there is a small spot on one side of the larger spot, and in other cases there are smaller spots on both sides. These smaller spots indicate that the next sized stack on one or the other sides, or on both sides of the best experimental stack gave results almost as good as the best. Instead of having a series of points representing the experimental data, we have a series of lines which may be employed to establish a zone of good per- formance. Concerning the width of this zone, it should be noted that the lines span but half the distance between the position of the stack which was best and that which is almost as good; also, that in cases where stacks on one or the other side of the best were not almost as good, no line whatever has been drawn. A review of the diagrams, Figs. 133 to 136, shows at a glance that 242 LOCOMOTIVE PERFORMANCE. the largest straight stack (15f inches), while, perhaps, sufficient in di- ameter for the least height (Fig. 133), was quite insufficient for the lower nozzle position when the stack height was increased beyond 26| inches. This is best shown by Fig. 136. The best results in connec- tion with this stack were obtained for the four lowest positions of the Stack A Height 26'^ Stack B Height 36^ \ ^ ^ ^ \ S \ 1 1 2 1 1 1 I X l s \ \ \ \ 3 ' - ^ \ 2- '-\ \ \ \ i 1 2 1 1 1 I . ^ 1 ! Stack Diameter. Inches FIG. 133. Stack Diameter. Inches FIG. 134. exhaust-pipe. When the tip of the exhaust-pipe was 5 inches above the center line of the boiler, the next smaller stack was almost as good, but for the lower nozzles (3, 2, and 1), the largest stack was in a marked degree better than any which were smaller, justifying the conclusion that if a stack as large as 18 inches in diameter had been tried, the Stack C Height 46 Stack D \ 9 \ \ * ^ x \ "\ So \ \ 1 1 2 i 4 *! 16 1 \ s - Height .56^* \ ^ \ s, \J \ \ 1 1 2 1 I 1 1 1 s Stack-Diameter. Inches FIG. 135. Stack Diameter. Inches FIG. 136. results obtained therefrom would have been given a place on the diagram. Best results obtained from tapered stacks, plotted in a manner already described in considering the action of the straight stack, are presented as Figs. 137 to 140 inclusive. In this case, however, the several diameters of stack experimented upon cover a range sufficiently THE FRONT END. 243 wide to permit the selection of the best stack for all heights of stacks in combination with all heights of nozzles. 1 1 6. Relation of Height to Diameter of Stack. The problem of stack design, as disclosed by the data already presented, is not to be regarded as one requiring a high degree of refinement. The data show Stack A Height 26 > 7~ 5 \ \I \ \ \i \ \ i | 1 1 t \' 1 1 1 8 Stack B Height 36?i Stack Diameter. Inches FIG. 137. \ \ \J 4 \ \ V, \ s \ i 1 2 1 1 1 S Stack Diameter. Inches FIG. 138. that two stacks varying as much as 2 inches in diameter sometimes give results which are almost identical. It happens in some cases, also, that a stack of a diameter which gives maximum results is almost equaled in its performance by a stack 2 inches less, and also by a stack 2 inches greater in diameter. In such a case a variation of 4 inches Stack C Height 46 ! Stack D Height ')f>}4 - i \ 5 \ 1 \ 3 \ 2 > \ l } 1 l 1 1 5 1 S -.N \ 2 | \ \ \ '\ \ 1 1 i r'4 \ i 16 1 S Stack Diameter Inches FIG. 139. Stack Diameter. Inches FIG. 140. in the diameter of the stack appears not to be significant. This is only true, however, with certain heights of stacks in combination with cer- tain heights of nozzles, and, as will be hereafter shown, the occurrence of such cases is more frequent in the case of tapered stacks than in the case of straight stacks. It is evident, therefore, that it will be impracticable to determine within a small fraction of an inch the diameter of stack corresponding 244 LOCOMOTIVE PERFORMANCE. with any given height of stack and position of nozzle, which can be said to accurately represent the experimental results, for it is some- times hard to say whether one or another of two stacks is the better. For the purpose, therefore, of reducing the various discordant elements to order, that a general law may be formulated, no harm will be done by employing a fair degree of liberality in interpreting the data. Straight Stacks. In order that the best relation of diameter and height of stack may be stated, it will at first be necessary to eliminate variations in the height of the nozzles. Comparisons will, therefore, first be based upon results obtained with nozzle No. 3, this being the nozzle for which the tip is on the center line of the boiler. From a study of the plotted spots of Fig. 133, it has been assumed that, when the exhaust-tip is on the center of the boiler, and the height of the stack is 26J inches, the most satisfactory diameter, as disclosed by the data, is 15 inches, and a circle has been struck on the diagram (Fig. 133) to represent this value. Similarly, from Fig. 134, it appears that when the exhaust-tip is on the center of the boiler and the height of the stack is 36^ inches, the most satisfactory diameter is 15.66, and a circle, the location of which corresponds with this value, has been struck on this diagram. In the same manner for the 46J-inch stack, the circle has been struck to represent a diameter of 16.33 inches, and for the 56J- inch stack the circle has been struck to represent a diameter of 17 inches. These, then, are assumed to be the best diameters of stacks for each of the several heights experimented upon, when the exhaust- nozzle is on the center line of the boiler. Stating these facts in the form of equations, in which d is the diam- eter of stack in inches, we have, for the engine experimented upon, the following : For straight stacks 26 inches high: d 1 = 15 = .28X54. For straight stacks 36| inches high : d 2 = 15.66 = .29X54. For straight stacks 46J inches high : d 3 = 16.33 = .30X54. For straight stacks 56 inches high : d 4 = 17 = .31X54. THE FRONT END. 24$ For all of these values the exhaust-tip was on the center of the boiler, and the diameter of the front end of the boiler experimented upon was 54 inches. If, now, we may assume that the data obtained from the engine experimented upon is applicable to engines having boilers of different diameters, and if we may assume also that in applying the data to other engines, we may use the diameter of the boiler as a, unit of measure, then the diameter of stack for any boiler whatsoever which has the exhaust-tip on the center line should be represented by equations in which D, the diameter of the inside of the front end, is substituted for the value 54 in the preceding equations. The result is as follows: For straight stacks 26^ inches high : For straight stacks 36J inches high : d? 2 = 15.66 = . 29 XD. For straight stacks 46 J inches high : d 3 = 16.33 = . 30 XD. For straight stacks 56^ inches high: If, now, an expression can be found which maybe substituted for the coefficient of D, and which will represent the height of stack in inches in each of the four equations, it will be possible to write a single equa- tion in the place of four. That this may be more readily accomplished, the values representing best diameters with which we have been deal- ing were so chosen that, while doing no violence to the experimental data, they will, when plotted in terms of height of stack, fall upon the same straight line, all as shown by Fig. 141. This fact makes it possible to write in simple form a general equation expressing the relation thus defined. Thus, by Fig. 141, it is apparent that when the stack height is zero, the diameter is equal to something over 13 inches (more exactly, 13.23), and the slope of the line connect- ing the several experimental points is such that with each inch height of stack, the diameter increases .00123 inch X 54. We may there- fore write as the coefficient of D, in the four preceding equations, (.246 + .00123#), in which H is the height of the stack in inches. 246 LOCOMOTIVE PERFORMANCE. We may, therefore, write for any straight stack when the exhaust- nozzle is on the center line of the boiler. d being the diameter of the stack in inches when the exhaust-nozzle is on the center line of boiler, H the height of the stack in inches, and D the diameter of the front end of the boiler in inches. Modification in the form of this equation to satisfy the condition arising from varying heights of exhaust-nozzle will be hereafter considered. Diameter of Stack, Inches FIG. 141. Tapered Stacks. The best results attending the use of the tapered stacks of each different height experimented upon, in connection with the seven different heights of exhaust-nozzles, appear in Figs. 137 to 140 inclusive. In these diagrams the experimental results are shown by means of black spots connected by horizontal lines in the manner already described in connection with the straight stack. When two stacks give equally good results, both points are located and the spots connected by a horizontal line; and where a larger or smaller stack gives results almost as good as the best, a line is extended in its direc- tion, terminating in a small spot midway between that representing THE FRONT END. 247 the best stack and the position representing the stack which is almost as good. Proceeding, as in the case of the straight stacks, to locate a repre- sentative point in line with nozzle No. 3, which will fairly represent the experimental data, choice has been made of the diameter 13J inches, and a circle drawn upon all diagrams while at this diameter. It ap- pears, therefore, that an important conclusion to be derived from the experimental data is to the effect that a tapered stack having a least diameter of 13^ inches gives maximum results for all heights of stack between the limits, of 26J and 56^ inches. In other words, unlike the straight stack, the diameter of the tapered stacks does not need to be varied with changes in the height. Stating this fact in the form of an equation, therefore, we have for a tapered stack upon the Purdue engine, the diameter of the boiler of which is 54 inches, the following : d a = 13.5; also d a = .25 X 54 inches. Assuming that the results thus obtained from the experimental engine may be applied to other engines having different diameters of boilers, and using the diameter of the boiler as a unit of measure, we may write for all locomotives, and for all heights of stack where the exhaust-tip is on the center line of the boiler: in which d is the least diameter of the tapered stack when the ex- haust-tip is on the center line of the boiler, and D is the diameter of the front end of the boiler. 117. The Effect of Changes in the Height of the Exhaust-nozzle upon the Diameter of the Stack. For the purpose of passing from the results obtained from the experimental engine to those to be ex- pected from engines having boilers of other diameters, using the boiler diameter as a unit of measure, it has been necessary thus far to deal with conditions for which the parts are symmetrically arranged. It is for this reason that the central position of the nozzle is the only posi- tion which has been employed. We may now consider the influence upon the diameter of the stack resulting from changes in the height of the nozzle. The points represented by the circles (Figs. 133 to 140), and which 248 LOCOMOTIVE PERFORMANCE. have been the basis of equations thus far written, have been so located that it is possible to draw through each of them a straight line, which will fairly represent the best diameter of stack for all heights of nozzles. The oblique line which appears in the several figures may be regarded ,as such a line. It now remains to find an expression for this line which can be added, as a new term, to the equation which has already been deduced, for the purpose of modifying the final results as demanded by the differences in results obtained when changes are made in the height of the nozzle. Straight Stacks. In Figs. 133 to 136 inclusive, representing the straight stacks, the oblique lines representing the relationship between diameter of stack and height of nozzle, as disclosed by the experimental data, have all been drawn at a constant angle. The slope of the line is such that, assuming the effect of the nozzle in position 3 to be zero, the effect upon the diameter of the stack of each inch change in the height of the exhaust-nozzle equals .19 of an inch. It is evident that this correction will effect an increase in the diameter of the stack when the nozzle position is below the center line of the boiler, and a decrease in the diameter of the stack when the exhaust-nozzle is above the center line of the boiler. We may, therefore, write as a new term in the equation, giving the best diameter of a straight stack, in which h is the distance in inches between the center line of the boiler .and the exhaust- tip, the sign preceding this term being positive when h is the distance below the center line, and negative when h is the dis- tance above the center line. Tapered Stacks. By a similar process it may be shown that the effect of changes in the height of nozzle on the diameter of a tapered ;Stack is expressed by .16ft, the sign preceding the term being positive when h is the distance below the center line, and negative when h is the distance above the center line. 118. Equations giving Stack Diameters for any Height of Stack between the Limits of 26 and 56 Inches, and any Height of Nozzle between the Limits of 10 Inches below the Center of the Boiler, and 20 Inches above the Center of the Boiler, and for any Diameter of Front End. Combining the expressions of the two pre- ceding paragraphs, we may have equations giving diameter of stack THE FRONT END. 249 in terms of its height, diameter of front end, and the distance between the center line of the boiler and the top of the exhaust-tip. These several equations obviously are the equations of the oblique lines ap- pearing in the corresponding diagrams, Figs. 133 to 140. They are as follows : For straight stacks: When the exhaust-nozzle is below the center line of the boiler d = (.246 + .00123H)D + .19ft. When the exhaust-nozzle is above the center line of the boiler d = (.246 + .00123#)> - .19ft. When the exhaust-nozzle is on the center line, ft is equal to zero and the equation becomes For tapered stacks: When the nozzle is below the center line of the boiler When the nozzle is above the center line of the boiler d = .25Z)-.16ft. When the nozzle is on the center line of the boiler, ft becomes zero, and d = .25D. In all of these equations d is the diameter of the stack in inches. For the tapered stack it is the least diameter or diameter of choke. H is the height of stack in inches, and for maximum efficiency should always be given as large a value as conditions will admit. D is the diameter of the front end of the boiler in inches, and ft the distance between center line of boiler and the top of the exhaust- tip. If D in the several equations is made equal to 54, the diameter of the front end of the Purdue locomotive, the equations will give results identical with those which are assumed to represent the maximums obtained in the course of the experiments. How far they should be employed in the manner which has been indicated for an engine having a boiler different in size cannot, of course, be stated with certainty, 250 LOCOMOTIVE PERFORMANCE. though it seems clear that, where the conditions surrounding stack and nozzle are similar, they may be depended upon to give satisfactory results for any diameter of front end now in use, or likely soon to come into use. Notwithstanding the short time which has elapsed since the first publication of the equations which have been given, they have already been extensively employed in service, with satisfactory results. The conditions which should be observed in their use is best shown by Figs. 142 and 143. FIG. 142. In this connection, also, it should be noted that it is not claimed that the plain stack and nozzle, as shown, will give better results than some other arrangement, but merely that when the plain stack and nozzle are used the equations will give the best relation of diameter to height which is obtainable. It is this question only that the experi- ments were designed to cover. Whether, for example, as a general proposition, the application of draft- or petticoat-pipes will improve the draft, or whether they will affect the relation of height and diameter of stack as already established, cannot be determined from the present work. THE FRONT END. 251 119. Unavoidable Loss in Draft with Deduction in Height of Stack. The equations already presented, together with the tabulated statements based thereon, are assumed to give the best diameter for a stack of any given height. They cannot be depended upon to give results which will always be satisfactory under conditions which greatly limit or restrict the height. In general, the best draft obtainable from a short stack is inferior to that obtained from a longer stack. The most that can be done when the limit of height has been fixed is to FIG. 143. determine a diameter which, in combination with that height, will give best results. It is this, and this only, that the equations, and the tables based thereon, are assumed to do. The rate at which the draft diminishes with each reduction in height of stack is indicated by Figs. 144 to 147, presenting results plotted in terms of draft and stack height. In explanation of these figures, it should be noted that the stack heights represented are: Base, 16 J inches; A, 26 J inches; B, 36J inches; C, 46J inches; and D 56J inches. Fig. 144 gives results with the smallest tapered stack in com- bination with the lowest nozzle; Fig. 145, those for the largest tapered stack in combination with the lowest nozzle; and Figs. 146 and 147 those for the smallest and largest tapered stacks respectively, for a 252 LOCOMOTIVE PERFORMANCE. series of tests at different cut-offs. In all of these figures it will be seen that there is a marked decrease in the draft obtained for each reduction Nozzle-l S"tack-2 Nozzle-1 Stack-8 Stack Heights FIG. 144. Small Tapered Stack. Stackjtieigh.ts FIG. 145. Large Tapered Stack. in the height of stack. This is the more significant when it is consid- ered that these diagrams represent only tapered stacks, and that the Nozzle- Stack-2 y O >i s -, / c f^ i - ^/ / T t? ^ & ^ J i & 16 I* K- ise -26 f *- -36 t -4<- e -56 [ *- Stack Heights FIG. 146. Small Tapered Stack. Nozzle-1 Stack-8 ^X^ | c&*5 )/ t Vx ^ X^ *> ^ 4 r, ^f "; ^ ^* 1( B; X- se -2f * ~1 K- i 'I * -565^- D Stack Heights } . 147. Large Tapered Stack, diameter of the choke of such stacks is not required to be changed when the stack is varied in height. It is clear, therefore, that these diagrams confirm the statement already made to the effect that a THE FRONT END. 253 short stack, however well designed, must be inferior in Its draft-pro- ducing qualities to a longer stack of good design. There is, in fact, nothing in the relation of diameter to height, or, so far as the investi- gation has proceeded, in the form of the stack itself, which can be accepted as a complete substitute for height. Again, it can be shown that this loss in draft with changes in height of stack does not depend upon the height of the nozzle. Thus, Table LVII. represents a sum- mary of results in connection with all heights of the smallest tapered stack in combination with all heights of nozzle. The table shows that, whatever the height of the nozzle, the loss in draft under the conditions represented in passing from the D, or 56J-inch height, to the A, or 26J- inch height, is a fixed quantity, and under the conditions of the tests referred to is approximately represented by 1.6 inches of water. Thus, for the nozzle No. 1, it is 1.7 inches; for the nozzle No. 3, 1.4 inches; for the nozzle No. 5, 1.8 inches; for the nozzle No. 7, 1.6 inches. TABLE LVII. BEST DRAFT OBTAINABLE UNDER CONDITIONS OF CONSTANT SPEED AND CUT-OFF WITH DIFFERENT HEIGHTS OF STACK IN CONNEC- TION WITH DIFFERENT HEIGHTS OF EXHAUST-NOZZLE. Nozzle. 1 2 3 4 5 6 7 6 g o5 8 g g g fl CH fH c Q g j _^ .. . g J e . ^j 8 | | i 1 1 i i 9) | V i Q 5 Q Q G Q Q Q Q Q Q Q Q Q Stack 2-D. . . 4.0 4.2 4.0 3.9 4.0 3.4 3.2 Stack 2-C. . . 3.5 .5 3.8 .4 3.7 .3 3.5 .4 3.5 .5 3.1 .3 2.7 .5 Stack 2-B. . . 3.1 .4 3.2 .6 3.3 .4 3.0 .5 2.9 .6 2.5 .6 2.2 .5 Stack 2-A. . . 2.3 .8 2.4 .8 2.6 .7 2.4 .6 2.2 .7 2.0 .5 1.6 .6 Total differ- ence 1.7 1.8 1.4 1.5 1.8 1.4 1.6 It is difficult to find a quantitative measure for the loss of draft resulting from reduction in height of stack which will be of general application. The extent of such loss is, however, suggested by the slope of the lines in the figures already referred to, and by the differ- ences in Table LVII. The general conclusion is, however, clear. It is apparent that the shorter stacks are far inferior to those of greater length, and that the maintenance of draft values in connection there- with must necessarily involve the application of additional mechanism, or an increase in the energy of the exhaust-jet. The study here de- 254 LOCOMOTIVE PERFORMANCE scribed does not suffice to indicate what should be the character of such additional mechanism nor the extent of the necessary increase in the energy of the exhaust-jet. It is, however, altogether possible that the adoption of some form of inside stack or of draft-pipes may make good the loss resulting from the diminished length of outside stack. 1 20. Relative Advantage of Straight and Tapered Stacks. But two forms of stacks were employed in the experiments under dis- cussion, one being perfectly cylindrical, and hence referred to as the straight stack; the other having the form best shown by Fig. 122, and generally referred to as the tapered stack. This tapered stack has its least diameter, or choke, at a point 16^ inches from the bottom, and increases in diameter uniformly above this point, the angle of the sides being the same for all stacks, and the divergence being at the rate of 2 inches in diameter for each foot in length. This divergence corres- ponds with that which was found best, as a result of the von Borries- Troske tests. The most noteworthy difference in results obtained from the two contours is to be found in the increased capacity of the tapered stack. Thus, No. 2 stack (9f inches diameter), which is the smallest of the tapered stacks, gives draft values which are two or three times as great as those obtained with No. 1, which is a straight stack of the same diameter. While the results from No. 2 are none of them sufficiently meritorious to warrant a place in the table of best results (Table LVII.), they give a close approach thereunto. The two smaller diameters of straight stacks Nos. 1 and 3 (9f and 11} inches diame- ter) are both far too small to yield results which are to be regarded as satisfactory, the capacity of these stacks being insufficient for the work expected of them. Speaking in rather general terms, it may be said that a tapered stack having a diameter of choke of 9f inches has as great a capacity as a cylindrical stack of 13} inches. Again, only the largest diameter of straight stack No. 7 (15} inches diameter) has a place in the table of best results, whereas all diameters of the tapered stack, excepting the least, find places in the table. It is of interest to note that with the low nozzle, tapered stack 6 B (least diameter, 13f inches; total height, 36J inches) gives identical results with straight stack 1C (diameter, 15} inches; height, 46J inches), the tapered stack being 10 nches lower and 2 inches less in diameter than the straight stack. It has been stated that whereas for best results the diameter of the straight stack must change for every change in height, that of the tap- ered stack remains constant for all heights. In this connection it is THE FRONT END. 255 important to remember that the diameter of the tapered stacks referred to is the diameter of the choke (least diameter). As a matter of fact, if designs based on the equations which have been deduced were to be superimposed, they would show that while the tapered stack, as com- pared with the straight stack, is of lesser diameter at the choke, the diameter of its top would generally exceed that of the straight stack. In general, it would seem that the tapered stack is less susceptible to minor changes of proportion, both of the stack itself and of the sur- rounding mechanism, than the straight stack. Thus, a variation of an inch or two in the diameter of the tapered stack affects the draft less than a similar change in the diameter of the straight stack. Again, the tapered stack is generally less affected by variations in the height of the nozzle, so that altogether it appears quite contrary to the ex- pectation of the experimenter, that the use of the tapered stack gives a greater degree of flexibility in the design of the dependent parts. For these reasons the tapered form is altogether preferable to the straight form. 121. A Summary of Results. The most important conclusions to be drawn from the experiments described may be stated as follows: 1. The jet acts upon the smoke-box gases in two ways; first, by frictional contact, it induces motion in them; and, second, it enfolds and entrains them. 2. The action of the jet upon the smoke-box gases is to draw them to itself, so that the flow within the front end is everywhere toward the jet. 3. The action of the jet is not dependent upon the impulses result- ing from individual exhausts. Draft can be as well produced by a steady flow of steam as by the intermittent action of the exhaust. 4. Draft resulting from the action of the jet is nearly proportional to the weight of steam exhausted per unit of time. It does not depend upon the speed of the engine nor the cut-off of steam from the cylin- ders, except in so far as these affect the weight of steam exhausted. 5. The form of the jet is influenced by the dimensions of the chan- nels through which it is made to pass. Under ordinary conditions, it does not fill the stack until near its top. If the diameter of the stack is changed, that of the jet will also change. 6. All portions of the smoke-box which are in front of the diaphragm have substantially the same pressure; and, consequently, a draft- gauge attached at any point may be depended upon to give a true reading. 256 LOCOMOTIVE PERFORMANCE. 7. The resistance which is offered to the forward movement of the air and gases between the ash-pan and the stack may be divided ap- proximately into three equal parts which are : First, the grate and the coal upon the same; second, the tubes; and, third, the diaphragm. It is significant that the diaphragm is as much of an impediment to draft as the fire upon the grate. 8. The form and proportions of the stack for best results are not required to be changed when the operating conditions of the engine are changed; that is, a stack which is suitable for one speed is good for all speeds, and a stack that is suitable for one cut-off is good for all cut-offs. In future experiments of draft appliances, therefore, results obtained from a single speed and a single cut-off should be deemed satisfactory. 9. Other things remaining unchanged, the draft varies with the weight of steam exhausted per unit of time; if the number of pounds of steam exhausted per minute is doubled, the draft, as measured in inches of water, is doubled; if it is halved, the draft value is halved. 10. As regards the form of outside stacks, either straight or tapered may be used. From a designer's point of view, the tapered is the more flexible: that is, with the tapered stack, the draft is less affected by slight departures from standard dimensions. Incidental reasons, there- fore, make the tapered form preferable. For best results the diameter of a given straight stack should be greater than the least diameter of a tapered stack for the same conditions. The term "tapered stack " used in this and other paragraphs sig- nifies a stack having its least diameter or choke 16J inches from the bottom, and a diameter above this point which increases at the rate of 2 inches for each foot in length. 11. In the case of outside stacks, either straight or tapered in form, the height is an important element. In general, the higher the stack the better will be the draft. 12. The diameter of any stack designed for best results is affected by the height of the exhaust-nozzle. As the nozzle is raised the diam- eter of the stack must be reduced, and as the nozzle is lowered the diameter of the stack must be increased. 13. The diameter of a straight stack designed for best results is affected by the height of the stack. As the stack height is increased, the diameter also must be increased. 14. The diameter of a tapered stack designed for best results, as THE FRONT END. 257 measured at the choke, is not required to be changed when tne stack height is changed. 15. The precise relation between the diameter of front end and the diameter and height of stack for best results is expressed by equations as follows : For straight stacks : When the exhaust-nozzle is below the center line of the boiler 9/> ...... (8) When the exhaust-nozzle is above the center line of the boiler d= (.246 + .00123#)D-. 19fc ..... (9) When the exhaust-nozzle is on the center line h is equal to zero and the last term disappears, and there remains (10) For tapered stacks: When the nozzle is below the center line of the boiler d = .25Z) + .16fc ............ (11) When the nozzle is above the center line of the boiler d = .25D-.16/t ........... (12) When the nozzle is on the center line of the boiler h becomes zero. and d=.25D ............. (13) In all of these equations d is the diameter of the stack in inches For tapered stacks it is the least diameter or diameter of choke. H is the height of stack in inches, and for maximum efficiency should always be given as large a value as conditions will admit. D is the diameter of the front end of the boiler in inches, and h the distance between the center line of the boiler and the top of the exhaust-tip. 122. Later Experiments. The preceding paragraphs present an account of researches upon the locomotive front end up to and in- cluding the year 1902. The work then completed left unsettled several important questions, namely, the best proportions for and the 258 LOCOMOTIVE PERFORMANCE. value of inside stacks, false tops within the front end, and draft- pipes. The equations which had resulted from the previous work were assumed to have been of general application, but their truth when applied to locomotives having short stacks upon boilers of large diameter had not been confirmed by direct experiment. Upon the completion of the work already described, therefore, the American Railway Master Mechanics' Association appointed a committee to co-operate with the American Engineer and Purdue University in doing such work as still remained in determining the proper pro- portions of the several details making up the locomotive front end, and to raise by subscription among railroad companies the funds needed to meet the expense thereof. The following is abstracted from the final report of this committee : * The tests were made in connection with New York Central loco- motive No. 3929 mounted upon the Purdue testing plant. The locomotive was of the Atlantic type having 21 / 'X26 // cylinders and 72" drivers. The inside diameter of the smoke-box was 74", and the height of its normal stack 29". The methods followed were similar to those already described. The tests of outside stacks involved a height of 29 inches, this being the maximum practicable for road conditions upon the loco- motive under test. Stacks of this height were supplied in diameters ranging from 15 to 25 inches by 2-inch steps. In these tests no draft-pipes or netting were employed in the front end; the dia- phragm and exhaust-pipe were the only details present. Under these conditions, with a 29-inch height, the best diameter was found to be 23 inches, though this was not much better than that of 21 inches. The exact arrangement of equipment for the best results is shown by Fig. 148. The notation under this figure and under those which immediately follow gives the draft obtained with a constant back pressure of 3.5 pounds. It will hereafter appear that there are better arrangements than that shown by Fig. 148. The point which is proven is that, assuming a plain outside stack 29 inches high to be used, its diameter for best results is 23 inches, as given. The experiments included inside stacks of four different diam- eters ranging from 15 to 21 inches, a constant outside height of * The report in full will be found in the Proceedings of the Master Mechanics' Association for June, 1906. THE FRONT END. 259 29 inches, and a penetration into the smoke-box of 12, 24, and 36 inches, respectively. The best proportions for this form of stack are shown by Fig. 149 accompanying. Its diameter is 21 inches and its penetration into the smoke-box is 12 inches. Results of nearly the same value were, however, obtained with stacks of smaller diameter having greater penetration. From values obtained it appears that as the degree of penetration increases the diameter of stack should be reduced. The effect is, in fact, of the same nature and degree as that which results from raising the exhaust-tip. It is noteworthy also that these values for the plain inside stack are not materially better than those for the plain outside stack, a fact which was formulated as a conclusion resulting from the work of 1902. DRAFT, 4.54 FIG. 148. DRAFT, 4.71 FIG. 149. DRAFT, 5.06 FIG. 150. It had been planned to fit the front end with three different false tops located at 12, 24, and 36 inches, respectively, from the top of smoke-box, but the presence of the steam-pipes made it difficult to fit the 12-inch top, and as a consequence only those of 25 and 36 inches drop were experimented upon. In each case stacks of different diameters were used, the outside height being always 29 inches. The best results were obtained with a stack 17 inches in diameter having a penetration of 24 inches, all as shown by Fig. 150. The draft resulting therefrom, when the back pressure was 3.5 pounds, was 5.06. In all cases with the false top the 17-inch stacks gave best results. A comparison of results with those obtained from a plain outside stack and from a plain inside stack shows material improvement in draft values. The false top necessarily interferes with free access into the front end, which fact makes it desirable that a way be found in which to secure the results derived from it by means which are more simple. It was suggested that experiments be made to determine the effect upon the plain inside stack of an annular ring or flange which might 260 LOCOMOTIVE PERFORMANCE. be considered as representing a portion of the false top. Responding to this suggestion rings of two diameters were used on 17-inch and 19-inch stacks having a penetration of 24 inches. It was found that the proportions shown by Fig. 151 gave substantially the same results as those obtained with the best arrangement of false top. Believing that the results thus obtained pointed to the desirability of having a broader curve at the base of the stack and that when the proper proportions were understood the best results would be obtained from such a curved surface, the 17-inch stack was fitted with a bell to which, for purposes of experimentation, flanges of various widths were afterward added, with the result that those proportions which appear in Fig. 152 proved most satisfactory. The best draft with the false top was 5.06, with the ring 5.05, and with the bell 4.98 that is, these three arrangements are practically on an equality ^H f 19 rf DRAFT, 5.05 FIG. 151. DRAFT, 4.98 FIG. 152. DRAFT, 4.55 FIG. 153. DRAFT, 4.40 FIG. 154. No other arrangements were experimented upon which gave higher draft values than these. Draft-pipes of various diameters, adjusted to many different vertical positions, were tested in connection with plain stacks of the several diameters available. The elaboration of this phase of the work was very extensive. It was found that for best results the presence of a draft-pipe requires a smaller stack than would be used without it, but that no possible combination of single draft-pipe and stack could be found which would give a better draft than could be -obtained by the use of a properly proportioned stack without the draft-pipe. While the presence of a draft-pipe will improve the draft when the stack is small, it will not do so when the stack is sufficiently large to serve without it. The best proportion and adjustment of single draft-pipe and stack are shown by Fig. 153. Double draft-pipes of various diameters and lengths, and having many different positions within the front ends, all in combination with stacks of different diameters, were included in the experiments THE FRONT END. 261 with results which justify a conclusion similar to that reached with reference to single draft-pipes. Double draft-pipes make a small stack workable, They cannot serve to give a draft equal to that which may be obtained without them, provided the plain stack is suitably proportioned. The arrangements and proportions giving the best results are those shown by Fig. 154. A suggestion as to a standard front end is presented as Fig. 155, BEST ARRANGEMENT OF FRONT END. FlG. 155. which, with the following equations referring thereto, may be accepted as a summary of the conclusions to be drawn from all experiments made. For best results make H and h as great as practicable. Also make d=.2lD + .16h, (14) b = 2d or .5D, P=.32Z), .22D. (15) (16) (17) CHAPTER XII. SUPERHEATING IN THE SMOKE-BOX.* 123. An Experimental Determination. The arrangement of steam-piping on an American locomotive constitutes one of the many ingenious features which have so long served to perpetuate the general characteristics of the early machine. From the time the steam leaves the dome of the boiler until it comes within the influence of the cylin- ders, there is no possible chance by which it can lose any of its heat, and during a portion of its course it actually gains heat. The arrange- ment as a system of piping is perfect. To a limited degree the pipe also serves as one element of a superheater, the smoke-box constituting the other. The extent of the drying or superheating effect upon the steam, which results from having the T head and the two branch pipes within the smoke-box, has long been an open question. Some foreign designers, evidently regarding the thermal gain resulting from such an arrangement as of less consequence than accessibility of parts, have carried the pipe connections outside of the smoke-box; while in this country at least one designer, not satisfied with following the usual practice, has carefully planned the form of his piping to facilitate the transmission of heat through its walls. To throw some light on the extent of the drying or the superheating effect upon the steam while passing these pipes in the smoke-box, a careful test was made upon the Purdue University experimental locomotive. In preparation for the experiment a thermometer was inserted in the T head at A (Fig. 156), another in the middle of one branch of the steam-pipe at B, and a third in the saddle close to the valve-box at C. The locomotive was then run under load with the reverse-lever for- ward and the throttle only partially open, the drop in pressure from the boiler to the pipe caused by throttling being sufficient to superheat all of the steam as it expanded from the pressure of the boiler to that of * The substance of this chapter was contributed to the Railway Review, July 28, 1894. 262 SUPERHEATING IN THE SMOKE-BOX. 263 the pipe. The extent of the superheating of steam at any given pres- sure is determined from its temperature alone. If, in the case under consideration, the steam neither received nor gave up heat in its pas- sage of the pipe, all three thermometers would show the same tempera- ture. A difference in the reading of the thermometers, therefore, must indicate a transmission of heat. The conditions of each test were maintained for a half-hour before any observations were made; the thermometers were then read and FIG. 156. other observations taken simultaneously, at five-minute intervals, for a second half-hour, there being no difficulty in maintaining constant conditions. As affecting the reliability of results it should be said that the thermometers used had a range of from 100 to 200 C., and read to fifths of a degree. They were inserted in long tubes, and at A and B these tubes were protected by allowing steam under the pres- sure of the pipe to flow past them. Fig. 157 shows the arrangement used at A for obtaining the temperature of the steam in the T head. A similar arrangement was used at B. Before the tests were made the thermometers were carefully compared by exposing them, while in the identical tubes used on the locomotive, to saturated steam of about the same temperature as that recorded in tlie experiment. The read- 264 LOCOMOTIVE PERFORMANCE. FIG. 157. SUPERHEATING IN THE SMOKE-BOX. 265 ing of one was accepted as standard, and the errors of the other two determined. Since all results depend upon differences of temperature, a slight error in the actual temperature was not considered important. Only the corrected readings are given. The following is a summary of results for one of the tests: Pressure of steam in boiler by gauge 130 Ibs. Pressure of steam, in steam- pipe by gauge 70 Ibs. Temperature of saturated steam at pressure of steam-pipe. . 315.68 F. ObservetTtemperature of steam at A, Fig. 156 (at T head) . 335.30 F. Observed temperature of steam at B, Fig. 156 (at middle of steam-pipe) 339.65 F. Observed temperature of steam at C, Fig. 156 (at saddle). . . 327.81 F. Observed temperature of gases in smoke-box 700.00 F. Increase of smoke-box temperature over temperature of steam in the boiler 345.00 F. Approximate distance from A to D, measured along center line of pipe 56 in. Size of branch pipe, inside 3 X 6 in. Thickness of walls f in. Approximate time occupied by the steam in passing from A to C 0.1 sec. It will be seen that the temperature of the steam was increased 4.4 in passing from A to B, which is equivalent to a gain of 8.8 in passing through the whole length of the branch pipe. The transfer of a quan- tity of heat represented by an increase of 8.8 in the temperature of superheated steam would affect moist steam by increasing its dryness about 0.5 of 1 per cent. It is believed that for the engine experimented upon, this approaches the maximum benefit which, from a thermal point of view, is to be derived from having the pipes inside the smoke- box. The conditions of the test were varied, and all the work repeated several times with the same general result. When a larger quantity of steam passed the pipe, the smoke-box temperature and the total heat transmitted increased, but the amount of heat transmitted per pound of steam was not materially changed. The figures given are from the test which showed the greatest heating effect. Enlarging the pipes within the smoke-box would have a pronounced effect in increasing the action herein considered, since it would both add to the extent of heating surface and also lengthen the time occupied by the steam in passing the same; but as a practical matter a limit to such enlargement for simple engines is soon reached. It therefore seems unreasonable to expect the steam to be dried to any consider- 266 LOCOMOTIVE PERFORMANCE. able extent in its passage of the smoke-box. Nevertheless, there is some gain, and a little gain is far better than a little loss. It is likely that a two-cylinder compound may reap some material advantage from the heat transmitted to its receiver, for in this case the heating surface may be quite extensive, and the movement of the steam through the receiver comparatively slow. Moreover, the re- ceiver pressure being lower than the pipe pressure of a simple engine, there is a greater difference of temperature between the smoke-box gases and the steam. An interesting result of the test is found in the fact that the ther- mometer in the saddle at C indicated a temperature 7.5 lower than the temperature in the T head at A, and 16.3 lower than the presum- able temperature at the lower end of the steam-pipe at D, so that, from the T head to the cylinder the steam does not gain, but actually loses heat. This effect is to be accounted for in the fact that the mean temperature within the cylinders is much lower than the temperature of the incoming steam, and that this, combined with the effect of radia- tion from the saddle, operates to lower the temperature of the iron which surrounds the steam in its course through the saddle. It is cer- tainly clear that the heat given the steam by the smoke-box is soon taken away again by the saddle, but how much of this cooling effect of the saddle is due to radiation is not shown. Experience with station- ary engines which have no saddles, however, forbids our charging it all to radiation. IV. THE ENGINES. CHAPTER XIII. INDICATOR WORK. 124. Concerning Indicator Work. From the beginning of the investigations of the locomotive laboratory the importance of giving close attention to the work of the indicators employed upon the loco - motive cylinders was recognized. A good indicator, if well cared for and intelligently operated, can be depended upon to give an accurate record of the changing pressure within the cylinder, but the same in- strument, if neglected or if carelessly manipulated, may give results which are worthless. In the work of the locomotive laboratory it was sought not only to have the indicator itself in good condition, but to have its action such as to insure a high degree of accuracy in its use. The reducing mechanism for transmitting the motion of the piston and of the indicator-drum is shown by Fig. 158. The motion of the cross-head is transmitted to the pendulum A through a pin work- ing in a slot of the lever. Another slot in this lever, near its fulcrum, receives and serves to drive a pin mounting a roller fixed to the metal bar B, which slides hi the guides CC. The phi driving the lever A and the pin upon the bar B, which is driven by this lever, can only move in lines parallel with the motion of the cross-head. As, in re- sponse to the motion of the cross-head, the lever A assumes its various angular positions, the leverage with which the driving-pin acts upon it is constantly changing, but similar changes are taking place in the leverage with which the lever A acts upon the pin which is attached to the bar B, hence the ratio of the two arms remains constant, or would do so if the pin and the rollers which they bear were infinitely small. The actual error arises from the measurable diameter of these pins, and is so slight as to be of no practical consequence. The pres- ence of the bar B, having a motion precisely similar with that of the piston and in line with the drum-sheave of the indicator, permits the use of a short cord connection in driving the latter. Neatly made 267 268 LOCOMOTIVE PERFORMANCE. blocks of wood clamped to the bar B serve as points of attachment for the indicator-cards, separate points being employed for each indicator. The connection of the steam end of the indicator with the cylinder also was designed to be as close as possible. This consists of a 3-inch length of f-inch pipe and one f-inch by J-inch elbow, the smaller end of which receives the indicator-cock. The cards obtained by means of this equipment were at once seen to be strikingly different from those obtained from locomotives upon the road, where a considerable length of pipe between the cylinder and indicator appears to be absolutely necessary. The cards from the FIG. 158. Indicator Rigging. laboratory were smoother in their general outline, and the events of the stroke were much more clearly marked upon them. While these differences are entirely creditable to the laboratory, the fact that they existed gave rise to numerous inquiries. Different parties were found to be interested in comparing cards from the laboratory with those obtained from road tests, with no satisfactory results. As an aid to the interpretation of cards obtained upon the road, therefore, it was determined to make a study of the effect of an indicator-pipe upon the INDICATOR WORK. 269 form of the cards. The results of this study are of sufficient interest to warrant some reference to them in this connection, though the fact should be borne in mind that all indicator work, underlying the data of succeeding pages, was obtained under the favorable conditions set forth by Fig. 158. 125. The Effect upon the Diagrams of Long Pipe Connections for Steam-engine Indicators. Experiments were first undertaken upon the locomotive which was equipped with two indicators, as shown by Fig. 159. One of these, hereafter referred to as the cylinder-indica- tor, w T as mounted upon the usual close connection; the other, hereafter referred to as the pipe-indicator, was mounted at the end of such a length of f-inch pipe as was necessary to carry the indicator to the FIG. 159. , top of the steam-chest, as must ordinarily be done in conducting road tests. The pipe was approximately 3^ feet long, but it was bent to form smooth curves, and was carefully covered throughout its entire length. By these means it was thought that the interference resulting from the presence of the pipe would be reduced to a minimum. It was sought not to exaggerate differences in the results occurring through the presence of the pipe, but rather to reduce such differences to mini- mum values. In carrying out the experiments cards were taken simultaneously from both indicators at various speeds. The indicators were then re- versed in their positions, and the observations repeated. The averages of results thus obtained were employed as final values. A comparison of results discloses the fact that the events of the stroke (cut-off, release, and beginning of compression) , as recorded by the pipe-indicator, were all later than similar events as recorded by the cylinder-indicator. The' cards from the pipe-indicator made all changes of pressure appear more 270 LOCOMOTIVE PERFORMANCE. gradual than those from the cylinder-indicator, and the area of the card from the pipe-indicator was greater. Fig. 160 presents two cards taken simultaneously from the indicators in question at a speed of 50 miles an hour, and Table LVIII. presents the differences in power as obtained from the two indicators for all speeds between 25 and 35 miles an hour. TABLE LVIII. ERRORS IN M.E.P. CAUSED BY PIPE CONNECTION SHOWN IN FIG. 159. Speed, Miles per Hour. Speed, Revolutions per Minute. Excess of Power shown by Pipe-indi- cator as Compared with that shown by Cylinder-indicator in Per Cent. 25 134 1.5 30 161 2.1 35 188 2.9 40 215 4.9 45 242 8.4 50 269 14.0 55 296 17.2 126. Experiments upon a Stationary Engine. The significance of the results described in the preceding paragraph was such that a FIG. 160. more elaborate process 01 experimentation was entered upon. As the locomotive was not well adapted to the purposes in hand, the ex- periments were transferred to a Buckeye engine having a cylinder 7f INDICATOR WORK. 271 inches in diameter by 15 inches stroke. Since a description of these later experiments will serve to enforce conclusions which may be drawn from those already described, some reference to them in this connec- tion will be of interest.* The power of this engine was absorbed by an automatic friction- brake, by means of which a very constant load was obtained. The head end of the engine cylinder was tapped with two holes (a and b, Fig. 161), both in the same cross-section, and hence equally exposed to the action of the steam in this end of the cylinder. One of these holes (a) was made to serve for the indicator A, the cock of which was placed as close to the cylinder as possible. The hole b was made to receive one end of a U-shaped pipe, the other end of which entered a coupling fixed in the angle-plate c. The cock of a second indicator, B, was screwed to this coupling. A single system of levers supplied the drum motion for both indicators. The pipe fittings were all half- inch. . A right-and-left coupling at d allowed the U-shaped section, dfb, to be removed at will, and replaced by a similar section of different length. Pipe lengths of 5, 10, and 15 feet were used, length being measured from the outside of the cylinder wall to the end of the coup- ling under the cock of the pipe-indicator. The pipe and fittings were covered first with a wrapping of asbestos board, next with three-eights of an inch of hair felt, and finally with an outside wrapping of cloth. It is to be noted that the bend in the pipe at / is easy, and that there is a continual rise in the pipe in its course from the cylinder to the indicator. Both indicators were always well warmed before cards were taken. A gauge between the throttle and the valve-box was use- ful as an aid in securing constant pressure within the latter. In the tests herein described, however, the boiler pressure was kept constant, as nearly as possible, and the throttle was generally fully open. A pair of new Crosby indicators was set apart for this work, and while it will be shown that the value of the comparisons which were undertaken is not dependent upon a high degree of individual accuracy in the indicators, these instruments, when calibrated under steam, gave results which were nearly identical. The results, which are presented in the form of diagrams (Figs. 162 to 171), were obtained in the following manner: * An account of results obtained from experiments in connection with the Buck- eye engine was first published as a paper before the American Society of Mechanical Engineers, " The Effect upon the Diagrams of Long Pipe Connections for Steam- engine Indicators," Proceedings of the Society, May, 1896. 272 LOCOMOTIVE PERFORMANCE. INDICATOR WORK. 273 The engine having been run for a considerable period, and the desired conditions as to pressure, speed, and cut-off having been obtained, cards were taken simultaneously from the cylinder- and the pipe-indicator. Two pairs of cards (i.e., two from cylinder and two from pipe) were thus taken as rapidly as convenient, after which the position of the indicators was reversed, and the work repeated. There were thus obtained four cylinder-cards and four pipe-cards, one-half of each set having been made by one of the indicators, and one-half by the other. Next, by the use of closely drawn ordinates the eight cylinder-cards were averaged and combined in the form of a single card, and the eight pipe-cards were in the same way combined to form a single pipe-card. The two typical cards thus obtained, superimposed as in Fig. 162, constituted the record of the test. This process was repeated for each of the several conditions under which tests were made. It is proper to add that the accuracy of the indicators used, and the constancy of the conditions maintained, were such as to make each card almost, if not quite, the exact duplicate of the representative of its set. The diagrams presented are full sized, the spring for all being sixty pounds. 127. Different Lengths of Pipe. The effects produced by the use of pipes between the indicator and the engine cylinder, of five, ten, and fifteen feet in length, are shown in Figs. 162, 163, and 164 respec- tively, the speed, steam pressure, and cut-off being constant. As noted upon the figures, the full outline represents the cylinder-card, and the dotted outline the pipe-card. It would seem that, under the conditions stated, the form of cylin- der-cards in the figures referred to should be nearly the same, whereas the figures show them to vary considerably. It will be well to omit, for the present, all discussion concerning the causes of these differences, and to accept the cylinder-card in each case as representing the true conditions within the cylinder. By reference first to Fig. 162 it will be seen that the effect of a five- foot pipe is to make the indicator attached to it a little tardy in its action. Thus, during exhaust, when for a considerable interval of time the change of pressure to be recorded is slight, the lines from the two indicators agree, but during the compression which follows the loss of sensitiveness in the pipe-indicator is made evident by its giving a line which falls below the corresponding line traced by the cylinder- indicator. Similarly, during admission there is an approximate agree- ment, while during the expansion which follows, the lagging of the 274 LOCOMOTIVE PERFORMANCE. pipe-indicator results in a line which is higher than the expansion line given by the cylinder-indicator. As a result of this lagging in the FIVE FOOT PIPE FIG. 162. TEN FOOT PIPE \ Pipe Card V v^ Cylinder Card' FIG. 163. FIFTEEN FOOT PIPE . FIG. 164. .NOTE. The speed (200 revolutions per minute), the steam pressure (80 pounds), and the cut-off (approximately J stroke) were constant for all dia- grams on this page. action of the pipe-indicator, its card is in error in the location and cur- vature of the expansion and compression curves; also in the location of the events of the stroke, and in the area which it presents. The INDICATOR WORK. 275 speed at which these errors are shown to occur is moderate (200 revolu- tions), and the length of pipe attached to the indicator is not greater than is often used. The general effect of a ten-foot length of pipe (Fig. 163) is the same as that of the shorter length, but the lagging action due to the pipe is more pronounced, and all errors are proportionately greater. In this case, also, the admission and exhaust lines fail to agree, the total range of pressure recorded upon the cards being less than the range existing in the cylinder. A still further addition to the length of the pipe brings changes (Fig. 164) into the form of the pipe-card diagram which, while entirely in harmony with those already discussed, are of such magnitude that the form of the card loses some of its characteristic features. The admission and expansion lines are lower and the exhaust line is higher than are the corresponding lines for the true card. Reference to Table LIX. wills how that while cards from pipes of five and ten feet in length present an area greater than that of the true card, the card in question (Fig. 164) from a fifteen-foot length of pipe makes the area less. A comparison of the pipe-cards, Figs. 162, 163, and 164, makes it evident that a pipe of suitable length would result in a diagram some- what similar in form to that shown by Fig. 165 and a pipe still longer would give a card which would be represented by a single line, as AB, Fig. 165. FIG. 165. Various numerical results from Figs. 162, 163, and 164 are exhibited in Table LIX. It is true that the lengths of some of the pipes experimented with are excessive as compared with those commonly used for the connec- tion of indicators, but this fact does not deprive the results of their significance. If pipes of fifteen, ten, and five feet in length will produce the effects shown by Figs. 162, 163, and 164 respectively, it is but reasonable to suppose that pipes of less than five feet in length will produce some effect. And, since the effect of a five-foot pipe is con- siderable, this length must be greatly reduced before the effect ceases to be measurable. 276 LOCOMOTIVE PERFORMANCE. It will be shown later that differences of speed have less effect than would be supposed in modifying the form of the pipe-card; that, even with a speed as low as 100 revolutions per minute, the effect of the pipe is strikingly apparent. It will be shown, also, that the point of cut-off chosen for the whole series now under consideration (Figs. 162, 163, and 164) is not especially favorable for showing the modifying effect of the pipe. These considerations, together with the fact that indicator-pipes of three and four feet in length are not uncommon, all serve to emphasize the practical value of the effects noted. 128. The Form of the Cylinder Diagrams. It may be well at this point to consider the causes tending to change the form of the cylinder diagrams as they appear in the different figures (Figs. 162, 163, and 164), and to consider the evidence which justifies their acceptance as true diagrams. A study of the diagrams and data will make it evident that the differences are due wholly to a change in the length of the pipe. This single change, however, introduces an incidental change (1) in the clearance, (2) in the extent of surface exposed to the action of the steam, and (3) in the velocity of flow in and out of the pipe at the point of its connection with the cylinder. The effect produced by the several pipes upon the clearance of the engine is given below: Cylinder and port clearance, per cent of piston displacement 4.08 Clearance due to 5-ft. pipe, " " " " " 2.69 " " 10-ft. " " " " " " 5.14 " " 15-ft " " " " " " 7.84 Total clearance with 5-ft. pipe in place, per cent of piston displacement 6.77 " 10-ft. " " " " " " " " 9.22 " 15-ft. " " " " " " " " 11.92 The area of surface bounding the clearance space was affected by the pipes as follows: Surface bounding clearance space, no pipe attached 131.4 sq. in. " when 5-ft. pipe was in place 250.2 " " " 10-ft. " " " " 366.1 " " " 15-ft. " " " " 486.0 " " Increased clearance would lower the pressure at the end of com- pression, and would change the curvature of the compression line but it would not make the compression line as it appears in Fig. 164. The larger exposed surface would increase the effect due to the interchange of heat between the steam and the walls inclosing it. If it be assumed that, during the early stages of compression, this inter- INDICATOR WORK. 277 change results in reevaporation, and during the later stages in con- densation, the sum total of the effect would be in line with that re- corded. Such an assumption is reasonable, and such an action may in part account for the change under discussion, but its extent is not likely to be as great as the indicator has recorded. By far the most active agent tending to reduce the curved com- pression line of Fig. 162 to the straight line of Fig. 164 is the movement of steam hi and out of the mouth of the pipe. Thus, when compression begins in the cylinder, the pressure at the end of the pipe is greater than that in the cylinder (see Fig. 164), and steam must flow from the pipe to the cylinder. This current of steam entering the cylinder just when the mixture of steam and water already there is undergoing the early stages of compression helps to augment the cylinder pressure, and to carry the early part of the compression line higher than it would otherwise go. As the process of compression goes on the current hi the pipe is reversed, and the cylinder supplies steam to the pipe, thus causing the curve for this portion of the event to fall lower than it otherwise would. Increased curvature during the early stages and diminished curvature during the later stages result in a line which is approximately straight (Fig. 164). Similar reasoning will account for the rapid drop in pressure after cut-off (cylinder-card, Fig. 164). At the instant of cut-off the cylinder is supplying steam to the pipe. The flow is rapid, and the kinetic energy of the steam causes it to pile up in the pipe; and, although as the stroke advances this pressure is constantly decreasing, the pipe continues throughout expansion to hold a higher pressure than that contained by the cylinder. It is obvious that the pressure is not the same at the two ends of the pipe, except for points indicated by the crossing of the lines, as at a, b, and c (Fig. 164), and that the difference of pressure shown at other points is quite sufficient to account for the pronounced change in the cylinder diagram when pipes of different lengths are used. The existence of these differences hi the form of the cylinder diagram does not in any way affect the results which are here presented. All cylinder-cards may be accepted as true, and the fact that they are not all alike does not diminish their value, but rather emphasizes the importance of this whole subject. 129. The Effect of the Pipe at Different Speeds. The effects thus far discussed are those recorded for a constant speed of 200 revolu- tions per minute. In considering to what extent changes of speed will 278 LOCOMOTIVE PERFORMANCE. modify these results, reference should be made to Figs. 166, 167, and 168, which give a series of results involving a ten-foot pipe for which all condi- tOO REVOLUTIONS PER MINUTE FIG, 166. 200 REVOLUTIONS PER MINUTE FIG. 167. 25Q REVOLUTIONS PER MINUTE PIG. 168. NOTE. The steam pressure (80 pounds), the length of pipe (10 feet) and the cut-off (approximately | stroke), were constant for all diagrams on this page. tions were constant except that of speed. Numerical comparisons may be made from Table LIX. It will be seen that increase of speed pro- duces modifications in the form of the pipe diagrams which, in kind, INDICATOR WORK. 279 are similar to those produced at constant speed by increasing the length of the pipe, but these changes are not great. For example, increasing the speed from 100 to 200 revolutions per minute (Figs. 166 and 167) produces less change than increasing the length of the pipe from five to ten feet (Figs. 162 and 163). The fact that an engine runs slowly, therefore, does not seem to justify the use of an indicator at the end of a considerable length of pipe. Slow running reduces the error; it cannot be depended upon to eliminate it entirely. 130. The Effect of the Pipe at Different Cut-offs. The relative effect of the pipe when the cut-off is changed, other conditions being constant, is shown by Figs. 169, 170, and 171, and numerically by Table LIX . It will be seen that the differences of pressure recorded during ex- pansion by the two indicators (pipe and cylinder) are approximately the same for all cut-offs, but the relative effect of these differences upon the area of the diagram is most pronounced upon the smallest, or shortest, cut-off card. The fact that in Fig. 171 the steam line on the pipe-card rises, while that of the cylinder-card declines, consti- tutes a good illustration of the slowness with which the pressure in the pipe responds to that in the cylinder. TABLE LIX. THE EFFECT OF A PIPE ON THE FORM OF INDICATOR DIAGRAMS. Pressures Steam Con- Apparent Cut-off. Excess ( + ) or Deficiency ( ) shown by Pipe Diagram. sumption per Horse-power per Hour. Excess shown Excess ( + ) or 1 by Pipe Dia- gram. At the Deficiency ( ) as shown 1 At Cut-off. At Release. Beginning of Compression. M.E.P. by Pipe Diagram. 1 J3.S ts ^ o o fl a a d a| S- | s 1 3 T) s T3 g s -0 (T 1 3 1 1 1 I 1 & 1 I J 3.0 11.5 -1.3 T- 2.2 0.0 0.0 -0.1 - 5.6 + 1.2 + 3.7 - 0.1 - 0.5 2 6.8 24.3 -2.8 - 4.8 + 2.1 + 18.0 + 0.2 + 9.2 + 2.7 + 8.5 - 0.4 - 1.7 3 10.0 38.4 -5.1 - 9.8 + 6.5 + 57.7 + 4.0 + 200.0 -1.5 - 5.0 + 11.7 + 45.9 4 4.0 16.6 + 1.1 + 2.0 -0.5 - 3.7 + 0.1 + 6.0 + 3.0 + 8.8 - 2.0 - 7.2 5 7.5 27.7 -2.8 - 4.8 + 2.1 + 18.0 + 0.2 + 9.2 + 2.2 + 6.6 - 0.4 - 1.7 6 6.8 26.7 -3.0 - 5.7 + 2.3 + 24.3 -0.3 - 9.0 + 5.3 + 18.7 - 1.0 - 4.2 7 3.3 25.4 + 3.5 + 6.7 0.0 0.0 0.0 0.0 + 6.5 + 35.3 + 11.5 + 47.9 8 7.0 25.9 -2.8 - 4.8 + 2.1 + 18.0 + 0.2 + 9.2 + 4.2 + 12.8 - 0.4 - 1.7 9 2.3 6.5 + 6.0 + 10.3 + 3.0 + 18.2 + 1.0 + 50.0 + 1.4 + 3.4 - 6.8 -23.6 All percentage values are based on results from cylinder diagrams. For example, in Test No. 4 the pressure at cut-off shown by the pipe-card is 2 per cent in excess of that shown by the cylinder or true card; the pressure at release by the pipe ; card is 3.7 per cent less than by the true card; the pressure at the beginning of compression by the pipe-card is 6 per cent greater than by the true card; and the mean effective pressure by the pipe-card is 8.8 per cent greater than by the true card. 280 LOCOMOTIVE PERFORMANCE. Comparisons have been made from tests run under still other con- ditions, and all conclusions thus reached have been consistent with those presented. This whole plan of work was outlined with the CUT-OFF % STROKE/ FIG. 169. CUT-OFF J/4 STROKE FIG. 170. CUT-OFF y% STROKE PIPE CARD CYLINDER CARD FIG. 171. NOTE. The cut-off as given above is approximate. The steam-pressure (80 pounds), the speed (200 revolutions per minute) and the length of pipe (10 feet), were constant for all diagrams on this page. expectation of securing such data as would permit a complete analysis of the effects produced by a pipe. But the results show that these effects are modified by so many different conditions that their precise character cannot be safely predicted. Even if it were possible to INDICATOR WORK. 281 construct an expression for reducing a distorted pipe diagram to a form which would correctly represent the relation of pressure and volume within the cylinder, the number of its terms would be so great, and its form so complicated, that the expression would have no practical value. 131. Conclusions. The following conclusions constitute a sum- mary of the data already presented : 1. If an indicator is to be relied upon to give a true record of the varying pressures and volumes within an engine cylinder, its connec- tion therewith must be direct and very short. 2. Any pipe connection between an indicator and an engine cylin- der is likely to affect the action of the indicator; under ordinary con- ditions of speed and pressure a very short length of pipe may produce a measurable effect in the diagram, and a length of three feet or more may be sufficient to render the cards valueless, except for rough or approximate work. 3. In general, the effect of the pipe is to retard the pencil action of the indicator attached to it. 4. Other conditions being equal, the effects produced by a pipe between an indicator and an engine cylinder become more pronounced as the speed of the engine is increased. 5. Modifications in the form of the diagram resulting from the presence of a pipe are proportionally greater for short cut-off cards than for those of longer cut-off, other things being equal. 6. Events of the stroke (cut-off, release, beginning of compression) are recorded by an indicator attached to a pipe later than the actual occurrence of the events in the cylinder. 7. As recorded by an indicator attached to a pipe, pressures during the greater part of expansion are higher, and during compression are lower, than the actual pressures existing in the cylinder. 8. The area of diagrams made by an indicator attached to a pipe may be greater or less than the area of the true card, depending upon the length of the pipe; for lengths such as are ordinarily used the area of the pipe-cards will be greater than that of the true cards. 9. Within limits the indicated power of the engine is increased by increasing the length of the indicator-pipe. 10. Conclusions concerning the character of the expansion or compression curves, or concerning changes in the quality of the mix- ture in the cylinder during expansion or compression, are unreliable when based upon cards obtained from indicators attached to the cylinder through the medium of a pipe, even though the pipe is short. CHAPTER XIV. THE EFFECT OF LEAD UPON LOCOMOTIVE PERFORMANCE. 132. Lead. When a valve admits steam to the cylinder before the piston has completed its return stroke it is said to have lead. The amount of lead is the width the steam-port is open when the engine-piston is at the beginning of its stroke. Speaking in very general terms it may be said that the effect of lead is to admit steam to the cylinder before the piston is ready to start on its forward stroke. Its presence tends to insure an abundance of steam behind the piston, at the very beginning of the stroke, to assist in the maintenance of a satisfactory supply throughout admission and to promote the smooth running of machinery, which is likely otherwise to be noisy. Practice in valve-setting has always recognized the value of lead, and it is not often that an engine is run without it. Any admission of steam behind the piston before it has completed its return stroke, however, results in negative work, and for this reason the amount of lead should not be excessive. It appears reasonable to assume that the lead should at all times be such as will insure a complete filling of the clearance space, but that more than this will affect unfavorably the economic performance of the engine. But the amount of lead necessary to give this result can only be determined by the use of the indicator, and as the requirement is likely to vary with changes in speed, or with the amount of compression employed, the 'problem is not a simple one. in establishing the lead for the valves of a locomotive it is usual to make the measurements when the reverse-lever is in its extreme position This determines the lead for the maximum valve travel. Now, it happens tnat the most common forms of locomotive link- motion give an increase of lead as the cut-off is shortened. Thus, with a T^-inch lead in full gear a valve may have as much as f-inch at running cut-off, the exact amount of the increase depending upon 282 EFFECT OF LEAD. 283 the proportions 01 the individual gear. For many years it was the practice to make the lead in full gear anywhere from ^ inch to J inch, the larger amount being commonly employed upon locomotives in higher speed service. These amounts give satisfactory action when the valve is at full travel, and for a long period the practice was not questioned. At about the time the Purdue testing-plant came into existence, it was suggested that the important thing was to have the lead satisfactory for the running cut-offs rather than at full gear, and that this should be accomplished even at some sacrifice in the distribution for other cut-offs. The advent of the Purdue locomo- tive gave some opportunity for experiments along this line. 133. Tests Involving Different Amounts of Lead. Locomotive Schenectady No. 1, as received from its builders, was found to have a A- inch lead for the full travel of the valve, and at short cut- off to give an indicator-card having a loop at the initial end. As the presence of such a loop is always to be accepted as evidence of an excessive lead or compression, it was determined to change the setting of the valves by changing the position of the " go-ahead " eccentric on the axle by an amount sufficient to eliminate the loop. Before proceeding with this, however, measurements were made to determine all events of the stroke, and four tests were run to determine the per- formance of the engine. The change was then made, all measure- ments retaken, and other tests were run. The results of the tests made prior to the change are given in Chapter IV. as Series V, while those obtained after the change constitute Series A. The A setting was, in fact, allowed to stand for all subsequent work with Sche- nectady No. 1, except in the case of Series J and K. The tests with which the present discussion is immediately concerned are eight in number, designated as follows: 15-1-V in comparison with 15-1 -A. 25-1-V " " " 25-1-A. 35-1-V " " " 35-1-A. 55-1-V " " " 55-1-A. 134. Effect of Lead upon the Events of the Stroke. In reducing the lead, the "go-ahead " eccentric was moved backward a distance of about J inch, as measured on the circumference of the axle, or through an angle of approximately four degrees. The lead at full gear, which previous to this change had been ^ inch, was in this man- ner reduced to zero. To secure a measure of the effect of the change 284 LOCOMOTIVE PERFORMANCE. upon the events of the stroke, the engine was run at low speed under a heavy load, with the reverse-lever in each of the several notches from the center to the extreme forward position, for a half-hour, during which time two complete sets of .ndica tor-cards were taken. Cards were taken at low speed and heavy power, defining clearly the events of the stroke. All cards obtained by the process described were worked up to show the per cent of stroke at which the events occurred, and the average values for all cards taken with the reverse-lever in a given position were accepted as representing the events, for that posi- tion. These values are given in Table LX., in which Series V repre- sents conditions before the change, and Series A the conditions after the change. They are also shown graphically by Fig. 172. TABLE LX. EVENTS OF STROKE IN PER CENT MEASURED FROM INDICATOR- * CARDS. Ill Series V. Series A. l|2 Admis- sion. Cut-off. Release. Com- pression. Admis- sion. Cut-off. Release. Com- pression. 1 5.04 25.1 67.6 42.1 3.25 24.7 71.1 33.1 2 3.65 33.3 73.8 31.7 1.7 33.7 77.9 28.6 3 2.6 42.7 78.6 25.4 1.16 45 81.6 23.5 4 1.74 51.4 82.5 20.8 .69 54.1 85.7 17.7 5 1.16 59.3 85.8 17 .41 61.9 89 14.6 6 .74 66 88.5 13.7 .16 69.8 91 11.2 7 .5 71.6 90.7 10.8 .1 74.8 93.1 9.39 8 .21 76.1 92.6 9.06 .14 79.3 94.2 7.5 9 .16 79.8 93.7 7.45 82.9 95.4 6.04 10 .05 83 94.9 6.12 84.1 96.2 5.5 11 85.1 95.8 5.00 86.71 97 4.4 12 87.3 96.6 4.19 88.56 97.5 3.87 13 88.7 97.2 3.52 90.2 98.0 2.77 14 90 97.7 3.02 91.3 98.6 2.19 15 91.1 98.1 2.76 92.7 98.9 2.09 It will be seen that the effect of the change in valve-setting is to retard all the events of the stroke. For long cut-offs these effects are so small as to be negligible, but as the cut-off is shortened they increase in all cases except with reference to the cut-off. From a purely academic point of view these effects appear to be advantageous. For example, with a given cut-off the release is delayed, thus pro- longing the expansion, while a relatively greater delay in the beginning of compression contributes to a free exhaust. With reference to admis- EFFECT OF LEAD. 285 sion, it appears that when the reverse-lever is well forward (tenth to fifteenth notch) neither the original lead of ^ inch nor the absence of it have a sufficient effect on the admission to show on the indicator. The reason for this is probably to be found in the fact that at long cut- offs the travel of the valve is so rapid, when the piston is at the end of its stroke, that the precise time when the port opens is not a matter of great importance. As the reverse-lever is moved toward the center, AV Compression 7 Acl lission t-Off Relea 10 20 30 40 50 ' 60 70 Per Cent of Stroke FIG. 172. Events of Stroke. 90 100 however, the lead increases, and, as will be hereafter shown, the maxi- mum port-opening diminishes, and the effect of the lead upon admis- sion, both relatively and actually, is increased. 135. Effect of Lead upon Valve-travel and Port-opening. The extent of the valve travel for the several positions of the reverse- lever was determined by means of apparatus shown in Fig. 173. This consisted of a stop (a) secured to the valve-stem, a wooden straight-edge (6) secured to the engine frame directly above and parallel to the valve-stem, and a try-square (c) with a blade suffi- c'ently long to have contact with the stop as it moved back and forth with the valve-stem. With the locomotive running under desired conditions the try-square was moved on the straight-edge till the 286 LOCOMOTIVE PERFORMANCE. stop just touched it at one end of its stroke, after which a line show- ing its location was drawn on a piece of paper fastened to the straight- edge. The try-square was then moved to the other end of the straight- edge and the process repeated, after which the actual travel of the FIG. 173. valve was found by measuring the distance between the two lines and subtracting the known width of the stop. The average port-opening was found by suotracting twice the outside ^lap of the valve from the travel of the valve and dividing by two. Table LXI. presents the record of valve travel and the port-openings, as determined in this way for each position of the reverse-lever. TABLE LXI. VALVE TRAVEL AND PORT-OPENING. (From Right Cylinder of Locomotive. ) Valve Travel, Inches. Port-opening, Inches. Notch Forward of Center. Series V. Series A. Series V. Series A. 1 1.94 1.88 .22 .19 2 2.04 1.96 .27 .23 3 2.19 2.13 .34 .31 4 2.37 2.32 .43 .41 5 2.58 2.54 .54 .32 6 2.72 2.81 .61 .65 7 3.07 3.06 .78 .78 8 3.35 3.30 .93 .90 9 3.55 3.62 1.02 1.06 10 3.94 3.90 1.22 .20 11 4.28 4.28 1.39 .39 12 4.59 4.58 1.54 .54 13 4.90 4.91 1.70 .70 14 5.22 5.25 1.86 .87 15 5.57 5.57 2.03 2.03 EFFECT OF LEAD. 287 An examination of the table discloses the fact that at short cut- offs, when the lead is reduced, the valve travel and, consequently, the port-opening are also reduced, though the effect in these particu- lars is not great. The table is perhaps chiefly of interest for the actual values it assigns to the port-openings for the different positions of the reverse-lever. A moment's consideration will show that the lead at short cut-offs does not need to be great to equal the maximum port-opening, in which case the maximum opening of the port occurs before the piston starts on its stroke. With lead at short cut-offs, of a quarter of an inch or more, it is evident that such a condition does occur. The precise manner in which it occurs is best seen from the ellipse (Fig. 191), showing the relative motion of the valve and piston. 136. Effect of Lead upon the Form of Indicator-cards. Repre- sentative cards from the several tests are reproduced as Fig. 174, and have already been referred to. As previously stated, it was the purpose in changing the valve-setting to so reduce the lead that the loop which appears in the cards of Series V should disappear. The results show (cards of Series A) that this was not entirely accom- plished, and that some further reduction might have been made upon the right-hand side, but the change which was made is nevertheless significant in its effect upon the indicator-cards. First to be noticed is the smaller size of the cards of Series A. It is clear that by delaying the opening of the port the amount of steam admitted is reduced and the work per stroke diminished. It must be admitted, also, that the cards of Series A are not so well filled out at the initial end as are those of Series V, and that they are not so well marked by the events of the stroke. These effects are shown by Fig. 175, in which the dotted outline represents a card taken when the lead was considerable, and the full line a card taken after it had been reduced. The reduction of lead effects a reduction in the back- pressure losses, but it also reduces the positive work under the steam and expansion lines. Judging from the cards alone it would appear to be a fair question as to whether any advantage had been gained by reducing the lead; but it must be remembered that a reduction in the size of a card does not necessarily constitute an argument against reduced lead, since the work per stroke can always be increased by moving forward the reverse-lever. The advisability of the change is rather to be judged by the amount of steam consumed per unit power, and to this phase of the matter attention will now be given. SPEED IN MILES PER HOUR CO m CO m m EFFECT OF LEAD. 289 137. Steam Consumption. The steam consumption constitutes the one fact which finally determines the usefulness of any alteration in steam distribution. The results of tests, as set forth in Table Pressure FIG. 175. LXII. and as presented graphically by Fig. 176, show conclusively that the reduction in lead effected a reduction in the amount of steam consumed. It is often assumed that high speeds admit of the use of Speed -Miles Per Hour 10 20 30 40 FIG. 176. Steam Consumption. 50 more lead than is desirable when the speed is low, but the results do not confirm such an assumption. They show that the loss arising from excessive lead is less than when the speed is lower, but they do not in any case show that the larger lead gives the greater economy. From the standpoint of steam consumption per indicated horse- power, there can be no question but that the performance of the engine was improved by so reducing the lead as to entirely remove 290 LOCOMOTIVE PERFORMANCE. the loop from the initial end of the card. The change appears to have been justifiable, even though it resulted in making the initial end of the indicator-card somewhat rounded under certain conditions of run- ning (Fig. 174). A comparative study of the indicator-cards and the record of steam consumption will be found profitable. TABLE LXII. STEAM CONSUMPTION. Steam per I.H.P. Speed, Miles per Hour. Approximate Cut-off, Per Cent. Series V. Series A 15 29.97 28.92 25 28.78 28.06 35 25 27.27 26.93 55 30.62 30.64 138. Lead and Machine Friction. It is shown elsewhere (Chap- ter XIX.) that high initial pressure is attended by heavy losses between the cylinder and the draw-bar in the form of machine friction, from which it is fair to conclude that the friction was greater in the tests of Series V than in those of Series A. If this were actually the case, and if the steam consumption per draw-bar horse-power were plotted, the difference between the two series would be greater than that ob- tained by basing the comparison on the indicated horse-power, that is, greater than represented by Fig. 176. 139. Conclusions. Although the experiments above discussed are somewhat limited in scope, the results justify the following rather general conclusions : 1. Lead should never be so great as to give a loop at the initial end of the indicator-card. 2. If necessary to avoid the loop at running cut-offs, the valves of a locomotive should be set with no lead, and even with negative lead for the full-stroke position of the reverse-lever. 3. Locomotives having their valves set with no lead in full gear will be found to respond to the throttle promptly under starting con- ditions. The valves move so rapidly, when the crank is passing its center, that the port opening will be liberal before the piston is fairly started on its stroke. CHAPTER XV. THE EFFECT OF OUTSIDE LAP UPON LOCOMOTIVE PERFORMANCE. 140. Outside Lap. The term "outside lap/' or "steam lap/' re- fers to the amount by which a slide-valve in its central position over- laps the outside edges of the steam-ports. Thus, in Fig. 177, the out- side lap of the upper valve is J inch, of the middle valve, 1 inch, and of the lower valve, 1J inch. Perhaps the most obvious effect of in- creasing the outside lap is the necessity for increased valve travel, if the cut-off is to remain unchanged. A casual inspection of Fig. 177 will show that a certain movement of the valve with f-inch outside lap may be sufficient to operate the engine, and yet, when applied to the valve having 1^-inch outside lap, prove inadequate to open the steam-ports. To determine the effect upon the power and efficiency of a loco- motive of different amounts of outside lap, three series of tests were run, designated as A, K, and J respectively. The slide-valves em- ployed in each series were alike in every respect, except as to the out- side lap, which varied as follows : Series A, Outside Lap J inch " K, " " 1 " J, " " 1J " Fig. 177 shows all the dimensions. No change was made in the setting of the eccentric or in any other position of the valve-gear, when one valve was interchanged for another having a different amount of outside lap. When increased travel was necessary, it was obtained by carrying the reverse-lever in a notch further from the center of the quadrant. It was sought to have tests run under conditions which would give as nearly as pos- sible identical cut-offs. This result was not perfectly secured, but the degree of success is shown by Table LXIIL, giving the events of the stroke as obtained from the indicator-cards. 291 292 LOCOMOTIVE PERFORMANCE. Series A Outside Lap =*" Inside Lap = &" 15-9-A 15-1-A 35-1- A 55-1-A Series K Outside Lap =1 Inside Lap =& 15-1 1-K 15- 2-K 35- 2-K Series J Outside Lap Inside Lap 15-15-J 15- 4-J 35- 4r-J FIG. 177. TABLE LXIII. EVENTS OF STROKE. Reverse- Actual Events of Stroke. Approxi- mate Cut-off. Outside Lap, Inches. lever, Notches Forward of Center. Cut-off, Release, Compres- sion, Admis- sion, Per Cent. Per Cent. Per Cent. Per Cenv One } 1 23.4 65.9 36.0 4.3 Quarter 2 26.3 69.7 32.8 2.8 li 4 23.3 69.9 31.6 1.3 Three i 9 82.5 95.1 5.7 Quarters i 11 79.5 93.5 7.1 i'i 15 69.4 90.5 10.1 EFFECT OF OUTSIDE LAP. 293 141. Events of Stroke. In order to study the events of the stroke in more detail there was attached to the engine a valve-motion indi- cator, by means of which the diagrams were made, showing the relative motion of the valve and piston. An analysis of one of these valve- motion diagrams * (Fig. 178) which as presented are approximately one fourth their full size, will disclose effects resulting from increased outside lap as follows: The rapidity with which the valve opens the steam-port is increased, resulting in a freer admission. The range of cut-off is decreased. When the cut-off is short the exhaust is hastened, an effect which diminishes as the cut-off is lengthened, and may dis- appear entirely when the reverse-lever is in an extreme position. The amount by which the steam-port is opened to the exhaust is, with short cut-off, increased, the extent of such increase being approxi- mately equal to the increase in the outside lap. This applies, however, only to the shorter cut-offs and to reasonably small outside laps, since, when either of these factors becomes large, the exhaust-port of the valve overtravels the steam-port, and exposes thereby its full width beyond .which it is not safe to go. 142. Changes in the Form of Indicator-cards Resulting from Changes in Outside Lap. The cards for the several tests, as actually obtained, are shown by Figs. 180, 181, and 182. Referring to those taken at a speed of 15 miles an hour (Fig. 180), it will be seen that, as regards cut-off, they are nearly identical, but differ essentially in other respects. The cards made for the valves of greatest lap have a higher steam line and lower exhaust and compression lines than those taken when the laps was less. The result is a card of greater breadth at the initial end, and of greater area for the same cut- off. The same characteristics are to be seen in the cards taken at 35 miles per hour (Fig. 181), though in these cards it is noticeable that the cut-offs are not quite the same, that for which the valve had 1 inch outside lap being greater than for the other two. Cards representing long cut-off (Fig. 182) are hardly to be regarded as comparable, since, owing to the fact that the reverse-lever was in its extreme forward position, the cut-off was regularly diminished as the lap was increased. The effect noted, therefore, with increase in outside lap, is influenced by the corresponding decrease in cut-off. Cards are nevertheless given, since they emphasize the fact that even with the outside lap as great as 1^ inches, the admission is prompt, and the steam line well sustained, when the reverse-lever is nearly in * For a detailed description of a valve-motion diagram, see paragraph 154. 294 LOCOMOTIVE PERFORMANCE. EFFECT OF OUTSIDE LAP. 295 296 LOCOMOTIVE PERFORMANCE. its extreme position. The lower cards of this set (Fig. 182) show the longest cut-off which can be secured with a valve of the proportions given. They well illustrate the difficulty of getting, in connection with large outside lap, a card which shall be full at both ends, a very desirable thing in locomotive service. A comparison of the cards of Figs. 180, 181, and 182 discloses the general effects of increasing the outside lap, or these may be more easily judged by reference to Fig. 183, which is to be regarded as a diagrammatic representation of what actually takes place. In this figure the full line is the normal card, such as may result from f-inch outside lap, and the dotted line is that which results when a valve is used having a greater amount of outside lap. When the cut-off remains constant, the effect upon the form of the card of increasing FIG. 183. Effect of Outside Lap on the Form of an Indicator-card. the outside lap is to raise the steam and expansion lines and to lower the exhaust and compression lines. It gives a freer exhaust, but it reduces the possible range of cut-off. All of these effects may be traced in the actual cards (Figs. 180, 181, and 182), and are what we should expect from a consideration of the valve diagrams. 143. Power Variation. The changes in power, resulting from changes in outside lap, are not of great importance, since, within limits, the power is at all times under the control of the engineer who has but to manipulate the reverse-lever in such a manner as to serve his purpose. For this reason, and because the similar cards are not identical in cut-off, no attempt should be made to compare numerical values representing M.E.P. or power. 144. Steam Consumption. Evidently the important fact is that of steam consumption. If the engine is more efficient under one con- dition of lap than under another, the argument will be strong in favor of adopting that condition. The steam consumption, as shown by the several tests, is shown by Table LXIV. EFFECT OF OUTSIDE LAP. 297 TABLE LXIV. STEAM CONSUMPTION. Speed, Miles per Hour. Approximate Cut-off, Per Cent of Stroke. Actual Cut-off, Per Cent of Stroke Outside Lap, Inches. Steam per I.H.R per Hour. 82.5 i 39.2 15 80 79.5 1 37.8 69.4 li 35.1 23.4 i 28.9 15 25 26.3 i 28.8 23.3 i* 28.5 23.4 I 26.9 35 25 26.3 i 27.8 23.3 li 26.8 The table shows a gain in steam consumption with increased outside lap while running at slow speed and long cut-off, but in drawing any conclusions it must be borne in mind that as the lap was increased the cut-off was materially decreased, a fact in itself sufficient to account for the better economy. At running cut-offs and at ordi- nary speeds there seems to be very little change in the steam con- sumption, whatever the lap. The fact that any increase of outside lap does not increase the economy of the engine, while it does materially reduce the flexibility of the valve-gear in respect to range of cut-off, is sufficient to justify common practice which makes this feature of the locomotive slide- valve comparatively small. CHAPTER XVI. THE EFFECT UPON LOCOMOTIVE PERFORMANCE OF INSIDE CLEARANCE.* 145. Inside Clearance. A slide-valve, which in middle position has its inside edges in line with the inside edges of the steam-ports, over which it is designed to travel, has neither inside lap nor inside clearance. It is sometimes referred to as being "line and line." When the valve is so proportioned as to allow the inside edges of the valve to overlap the inside edges of the steam-ports for mid-posi- tion, then the valve has inside lap; while, on the other hand, if the inside edges of the valve fail to cover the steam-ports, the valve has inside clearance. Thus, in Fig. 184, the first valve shown has ^-inch inside lap; the second, ^-inch inside clearance; and the third, f-inch inside clearance. The effect of changing a valve from inside lap to inside clearance, other things remaining unchanged, is to hasten release and to delay compression, and hence to increase the interval of time during which the exhaust-port is open. It also increases the extent of exhaust- port opening. As a consequence of these effects, the exhaust is made freer and back pressure is reduced, giving an advantage in the operation of the engine, which is greatly desired, and which would be accepted * Acknowledgment for assistance rendered in the investigations of inside clear- ance and outside lap is due to Messrs. Halstead, Robinson, Ede, and Wilson, who, while students, presented the facts which are herein summarized, in the form of the following theses: "The Effect of Increased Inside Clearance upon the Efficiency of Locomotive ' Schenectady,* " by W. G. Halstead, B.M.E., 1897; " The Effect of Increased Out- side Lap upon the Efficiency of Locomotive ' Schenectady,' " by G. P. Robinson, B.M.E., 1897; "The Effect of Changes in Inside Clearance upon the Efficiency of Locomotive 'Schenectady No. 1,'" by S. S. M. Ede, B.M.E., 1903; "The Effect of Changes in Valve Proportion upon the Efficiency and Power of Locomotive ' Schenectady No. 1,' " by A. M. Wilson, M.E., 1903. 298 EFFECT OF INSIDE CLEARANCE. 299 without question if it were not for the assumption that loss of effi- ciency attends the earlier exhaust. Until quite recently it was the practice to run all locomotives with a small amount of inside lap, usually A of an inch. With the advent of the modern engine, how- Series H Steam Lap %" Inside Clearance Series A Steam Lap J Inside Lap Series I" Steam Lap \" . Inside Clearance %' FIG. 184. ever, and especially in response to a demand for a free-running engine at higher speeds, the inside lap has been reduced, and in some cases an J-inch or ^-inch inside clearance has been employed. But as to the wisdom of such practice men best qualified to speak have differ- ences in their opinions. For the study of this subject three valves having the proportion 300 LOCOMOTIVE PERFORMANCE. shown by the three views of Fig. 184 were employed. These valves are identical, excepting as to their inside dimensions. Six tests were run with each set of valves, the conditions of speed and cut-off being those which -are set forth by Fig. 185. All valves were tested with the same setting, the position of eccentrics and reverse-lever being identical for the corresponding tests of the different series. The tests of each series, therefore, were made duplicates of each other, except in so far as the results were effected by the changed proportion of the valves. We may now consider the nature and extent of these effects. 55 -35 ga 25 34 44 Cut-off it v Center 123456789 Reverse Lever Notch * FIG. 185. Tests Run. 146. Maximum Opening of Steam-port into Exhaust. For short cut-offs the travel of the valve is ordinarily insufficient to open the full width of the steam-port into the exhaust. In such cases cutting the valve away to give inside clearance increases the maxi- mum opening into the exhaust by an amount equal to the change. A very material gain in exhaust-opening may thus be secured. For example, the tests which were run with the reverse-lever in the first notch gave a cut-off of 25 per cent. The maximum opening of the steam-port into exhaust under these conditions for the several valves tested were as follows: For valve having ?" inside lap (Series A) f " = /' ' ' clearance (Series H) \" " " (Series I) '/ _ 24'' -32" EFFECT OF INSIDE CLEARANCE. 301 147. Changes in the Events of the Stroke Resulting from Inside Clearance. Release and beginning of compression only are affected. The changes in these, with changes in valve clearance, are most pro- nounced for the shorter cut-offs. Thus, with the reverse-lever in the first notch forward of the center (cut-off, 25 per cent), the relation between clearance, cut-off, and release is shown by Table LXV. TABLE LXV. EVENTS OF THE STROKE, CUT-OFF, 25 PER CENT. Inside Lap or Clearance, Inches. Release, Per Cent of Stroke. Beginning of Compres- sion, Per Cent of Stroke TT2 lap ^ clearance f clearance 71.1 59.0 49.4 33.3 23.9 17.3 The table shows that a change from ,&" inside lap to &" inside clear- ance, a total change of J, hastens the release and delays the beginning of compression by approximately 10 per cent; this for short cut-off. When the cut-off is lengthened, and the travel of the valve is thereby increased, the relative effect upon the events of the stroke of slight changes in the dimensions of the valve itself becomes less pronounced. 148. Changes in the Form of the Indicator-card Resulting from Inside Clearance. Fig. 186 presents cards obtained with the reverse- lever in the first notch (25 per cent cut-off), and at speeds of 15, 35 and 55 miles an hour. The full-lined cards (Series A) were obtained in connection with valves having ^" inside lap, and are assumed as a basis of comparison. The dotted-lined cards on the left (Series H) were obtained with &" inside clearance, and on the right (Series I), with f" inside clearance. Where slight variations in the steam pres- sure for the different tests have appeared, the cards of Series H and I have been reduced to equivalent cards, for which the pressure at cut-off is identical with that of the corresponding card of Series A, thus supplying a logical basis for comparison. Referring to Fig. 186 it appears, first, that with increased inside clearance, the earlier exhaust effects a noticeable reduction in the area of the card under the exhaust line; secondly, that the wider exhaust-port opening and the later exhaust closure are responsible for a significant change in the location and form of the compression line, effecting a material increase in the area of the card over the same; and, third, that the reduced compression tends to diminish the maximum height of the 302 LOCOMOTIVE PERFORMANCE. card, so that, as valve clearance is increased, the maximum pressure is reduced. With a given amount of valve clearance all of these effects become more pronounced as the speed is increased. The per cent of area lost under the exhaust line and gained over the corn- is Miles per Hour (/' 81 Revolutions per Minute ^ 35 Miles per Hour 189 Revolutions per Minute 55 Miles per Hour 297 Revolutions per Minute FIG. 186. Effect of Inside Clearance on the Form of Indicator-cards. pression line respectively, expressed in terms of the area of the full- lined card, is as given in Table LXVI. The values of the Table appear to justify the conclusion that, when the reverse-lever is in running position, the loss in the area of the card, even at slow speeds, is so small as to be negligible, whereas for all but the slowest speeds there is gain, and at high speed the gain is material. The changes in the form of the cards under starting conditions are shown by Fig. 187, in which, as before, the full line represents the card taken in connection with the valve having -fa" inside lap (Series A), and the dotted lines cards representing &" and \" inside clearance EFFECT OF INSIDE CLEARANCE. 303 (Series H and I respectively). These cards represent the tests at the ninth notch (cut-off, 80 per cent) and a speed of 15 miles an hour. They show that the loss of area at the ends of the cards, attending increase of clearance, while actually large, is in part compensated H . d *// FIG. 187. Effect of Inside Clearance Under Starting Conditions. for by the lower exhaust line and, compared with the area of the card, is after all small. TABLE LXVI. REVERSE-LEVER FIRST NOTCH FORWARD OF CENTER, CUT-OFF, 25% OF STROKE. Speed. Clearance, Inches. Area Lost by Early Release in Per Cent. Area Gained by Reduced Back Pressure and Late Compres- sion in Per Cent. Net Gain ( + ) Miles. Revolutions per Minute. or Loss ( ) in Per Cent. 15 81 ft t 3.5 6.0 2.5 .5 -1 -5.5 35 188 ft 1 3.5 5.25 12.75 15.0 + 9.25 + 9.75 55 296 ft t 5.25 8.75 21.0 48.25 + 15.75 + 39.25 149. The Blowing through Effect. The fact that a valve having inside clearance opens both steam-ports to each other as well as to the exhaust-ports (see Fig. 184) has often been regarded as the chief objection to its use. The interval of time during which the two ports are in communication depends not only upon the extent of the clear- ance itself, but upon the travel of the valve and the speed of the engine. An increase either in the travel of the valve or in the speed of the engine diminishes the interval. It is important to note also that the intercommunication of the steam-ports does not lead to losses of live steam. Throughout the interval during which the two steam-ports are in communication 304 LOCOMOTIVE PERFORMANCE. both are seeking to exhaust steam. The only effect, therefore, which can result from the so-called blow-over action is an interference between the exhaust from one end of the cylinder and that from the other end; a portion of the steam from the port which is just opening, having a higher pressure than that which is issuing from the opposite port which has been longer open, follows the exhaust cavity of the valve and enters the opposite port, from which steam would other- wise be flowing (Fig. 188). With this understanding, and with the expectation of seeing the interference manifest itself by a rise in the exhaust line, we may now examine the card for the purpose of ascer- taining to what extent such action really occurs. Most important, of course, is an understanding of the action under FIG. 188. running conditions, the cards for which appear as Fig. 186. Here, it appears, that at a speed of 15 miles an hour, &" clearance produces no noticeable disturbance of the exhaust line, but when the clear- ance is increased to f", an amount in excess of any which has thus far been commonly employed in locomotive service, a decided hump appears. Thus, referring to the dotted outline representing this condition, the exhaust-port did not close until the point a was reached, the sudden rise in back pressure in the vicinity of b being due to an influx of steam from the opposite port which was then exhausting. This, then, is the blowing-through effect. Upon the cards representing the 35-mile tests, the effect gives to the exhaust line only a wave-like appearance; and when the speed has been increased to 55 miles, even this, as a distinct characteristic in the form of the card, nearly or quite disappears. No doubt, however, the action does take place even at the higher speed, and its com- plete suppression would further lower the later part of the exhaust EFFECT OF INSIDE CLEARANCE. 305 line and the compression line, but any harmful effect upon the form of the card is entirely lost in the lower sweep of the line which marks the freedom of the exhaust action as a whole. Those who have feared serious loss in the effect of inside clear- ance under conditions of starting will be reassured by reference to Fig. 187. As would be expected, the effect here is most pronounced, but with the long cut-offs it occurs so near the end of the stroke that it does not greatly reduce the area of the card. Thus, in Fig. 187, in the dotted diagrams, the rise beginning at 6, near the end of the exhaust line, appears to be compression. That it is not the result of compression must be apparent when it is remembered that increased clearance delays compression, and hence the exhaust closure for the cards, which are shown dotted, must be later than for the cards which are shown full. As a matter of fact the exhaust-ports for the dotted cards closed at the points c and d respectively, and consequently most of the rise which occurs at an earlier point in the stroke is the result of steam coming from the opposite steam-port. Obviously the exhaust from the port at the other end began when the return stroke for the end under consideration had reached the points e and / respectively. The conclusion is, therefore, that the fact that inside clearance places the two ports in communication is not significant so far as its effect upon the form of the card is concerned, even though the amount of clearance be as great as f". When the cut-off is late the effect is too near the end of the stroke to be serious, and when the cut-off is early, it is of moment only at slow speed, and even then the increased area of port-opening which the clearance gives serves to quickly dis- sipate the back pressure resulting from the blow-over. 150. Mean Effective Pressure. It has already been shown that increasing the inside clearance will, at speed, increase the area of the card, or, in other words, the mean effective pressure and, consequently, the power of the engine. A measure of such increase for speeds of 55 miles an hour with the reverse-lever in the first notch (cut-off, 25 per cent) is as follows: M.E.P. With valve having &" inside lap (Test 55-1- A) 17. 56 With valve having A" inside clearance (Test 55-1-H) 21.58 With valve having f " inside clearance (Test 55-1-1) 23 . 68 An increase in mean effective pressure, and, consequently, of the power of the engine in the ratio of 17.5 to 23.7, or more than 30 per 30(3 LOCOMOTIVE PERFORMANCE. cent, by merely cutting out the inside of the valve, does, in fact, con- stitute a strong argument in favor of a more general clearance. In urging the force of such an argument, however, it should, of course, be granted that such increase is only possible at high speed; but it is when speeds are high that those effects which are obtainable by means of increased clearance are most desired. While the conclusions of the preceding paragraphs are important and true, too much emphasis should not be given card areas. The fact that clearance helps to make a larger card is but a part of the story, for the effects of a large card may be obtained in various other ways, as, for example, by a change in the position of the reverse- lever. For this reason it is unfair to assume that any device or expe- dient which increases the card is justified. With this in mind may be noted the character of the changes in the form of the card resulting from clearance. The maximum pres- sure, the back pressure, and the pressure during compression are all thereby reduced. Pressures acting upon the piston are practically nowhere increased, but, on the other hand, for the most part reduced. In other words, the increase in effective work is chiefly the result of a diminution in the negative work of the cycle. Herein is the explana- tion of the fact, which all experimenters with inside clearance have observed, namely, that "it makes a free-running engine." But the test by which any apparatus, or system of adjustment, which is designed to improve the distribution of steam within an engine cylinder should be judged, is that of economy. The real ques- tion is whether by its use a given amount of power can be obtained by the expenditure of a smaller amount of steam. If so the de- vice will prove economical, and will also raise the maximum limit of power which the locomotive can deliver. If it cannot it will prove wasteful, and will reduce the maximum power which the locomotive can deliver. The effect, therefore, of inside clearance on steam con- sumption is a matter which should outweigh other considerations. 151. Steam Consumption as Affected by Increased Clearance. A detailed statement of the steam consumption, and of other related facts under the conditions of the several experiments, is set forth in full in Table LXVIL, while a summary of the important facts is presented as Figs. 189 and 190. Fig. 189 shows the steam consumption of the engine at different rates of speed, and proves the truth of the commonly accepted theory, namely, that increased inside clearance results in loss of efficiency at low speeds. For example, changing the EFFECT OF INSIDE CLEARANCE. 307 inside of a valve from &" lap to &" clearance increases the steam consumption at 15 miles an hour one pound per horse-power, and a further increase of inside clearance to f" adds three pounds to the steam consumption. As the speed is increased, however, the differ- 35 30 1 1- 1 h CC \ v ^s r>( j ' Sv - ^ "^-fc X ) M ^ - *^, ^s . ^ ""*, n >- >= ^ ^ ^ ^ ^^ x = ^ ^ 3 -c > ^..** & p;a> & 1 (' 1 1 CUT-OFF 25 PER CENT OF STROKE CURVE C INSIDE CLEARANCE f" " B " ^" A LAP ^-" ?15 CQ '10 10 20 30 40 50 Speed-Miles Per Hour FIG. 189. Inside Clearance and Steam Consumption. ence in steam consumption for the different amounts of inside clear- ance diminishes, until at a speed of 50 miles an hour (270 R.P.M.), or thereabouts, we have the same steam consumption for all cases. For speeds above 50 miles an hour the least steam consumption attends the use of the greatest amount of inside clearance, while the steam 308 LOCOMOTIVE PERFORMANCE. consumption for the valve having no inside clearance increases rap- idly, with increase of speed beyond this limit. Steam per T.R.P. per Hour-Pounds S S 8 ' -8 88? p / / / y / / V / / ^/ ^ / / / C(" X ^^* ^^ ^ / / s / / py x^ s / / A( X *>_ ' x*' SPEED 15 MILES PER. HOUR CURVE C INSIDE CLEARANCE f " B y $' 11 A t* LAP '" 40 60 80 Cut-Off in Per Cent of Stroke 100 FIG. 190. Inside Clearance and Steam Consumption. It has usually been assumed that losses incident to the use of inside clearance are greatest when the cut-off is longest, but the curves of Fig. 190 imply that this is not true. These curves represent the EFFECT OF INSIDE CLEARANCE 309 steam consumption at a constant speed of 15 miles an hour, but with cut-offs varying from 25 to over -80 per cent, and indicate that, as the cut-offs are increased, the steam consumption under all con- ditions of clearance approaches the same value. This tendency is sufficiently pronounced to justify the conclusion that the losses due to clearance are, other things being equal, diminished as the cut-off is increased. TABLE LXVII. STEAM CONSUMPTION AS AFFECTED BY INSIDE CLEARANCE. Speed, Miles per Hour. Cut-off, Per Cent of Stroke. Inside Lap, Inches. Inside Clear- ance, Inches. Steam per H.P.H. Pounds. A 33.34 15 80 & 33.92 ... f 35.39 A 28.92 15 25 :& 30.03 - f 32.96 h . . 26.93 35 25 & 28.9 f 29.36 h 30.61 55 25 A 29.72 ',- .4 f 29.38 The above statement embodies a fair interpretation of the tests under consideration. It would not be safe to generalize therefrom to the extent of assuming that these principles may be applied directly to the performance of all engines, though they may properly be accepted as an indication of what is likely to occur under similar conditions. In this connection it will also be well to remember that the range of clearance used in the experiments extends far beyond the amounts which have thus far been used in practice. Where inside clearance has been resorted to it has not often exceeded J inch, whereas in these experiments a maximum of f inch was employed. In view of this fact it is of special interest to note that even with a clearance as great as the latter amount, it ceases to be disadvantageous when the speed of the locomotive reaches the moderate limit of 50 miles an hour. CHAPTER XVII. LOCOMOTIVE VALVE-GEARS.* 152. The Function of a Valve-gear. The valve-gear of a modern locomotive^ contends with conditions which are difficult to meet. It is designed to so move the valve it drives as to open the port by an amount which, at running cut-off, usually does not exceed three- eighths of an inch, and at speed the entire interval during which any port is open is less than 'a twentieth of a second. If normal steam distribution is to be maintained the valve must move with great precision, since even a slight change in the extent of its travel, or in the time of its action, becomes relatively large when measured by the small port opening and the brief interval during which the port is open. Moreover, a valve of a modern locomotive, weighing with its yoke 150 pounds or more, requires the application of forces of con- siderable magnitude to give it motion, and its action involves more than ten reversals in its motion each second. It is necessary that the gear which actuates such a valve be free from lost motion, and that it be so stiff as to admit of no deformation of its parts. With this understanding of the requirement, we shall do well first to examine that type of valve-gear which is common in American practice, and to review briefly its merits and defects. 153. A Stephenson Valve-gear, which is so familiar to all that it need not be illustrated, as designed to drive a flat valve, contains ten joints having motion when the gear is in action, between the axle and the valve-yoke. Lost motion in any of these will modify the valve action. Accurate fitting and the use of case-hardened parts have, however, made the joint of the link motion reasonably satisfactory, but the average gear leaves much to be desired in the way of stiffness. It is true that great progress has been made in this latter respect. The * Adapted from a paper before the Southern & Southwestern Railway Club, January, 1905. 310 LOCOMOTIVE VALVE-GEARS. 311 modern engine does not slow down when the throttle-opening is increased, as used sometimes to be the case with locomotives having bent and limber eccentric blades; but even in the modern locomotive the spring of parts has not been entirely eliminated. This, however, may be greatly reduced by the adoption of the marine type of link, with double hangers, giving support to the link upon both sides, and having connection with a very heavy reverse-shaft and -arms. 154. What the Stephenson Gear Does. The motion communi- cated to the valve by the Stephenson gear is one which, beginning from rest, increases to a maximum and then gradually diminishes, until at the end of its travel it comes to rest again. The fact that it approaches its point of rest, and upon reversal recedes therefrom with a relatively slow motion, permits the valve to be controlled in its course more easily than would be the case if its changes in velocity were in obedience to a more complicated law. The gradual change in its rate of motion may well be seen in a curve representing the relative movement of valve and piston (Fig. 191). In the diagram of this figure, which is sometimes referred to as a valve ellipse, horizontal distances represent piston displacement, and vertical distances valve displacement. The figure shows two ellipses corresponding with two different positions of the reverse- lever. The distance between the mid-position line and the upper por- tion of either curve indicates the distance which the valve is re- moved from its central position, and distances along the mid-position line represent piston displacement. Thus, in the original diagram from which Fig. 191 was made, the distance along the line AB, from the center line to the curve c, is one and three-fourths inches. If we subtract the outside lap of the valve from this distance the result will be the port-opening. For example, in the present case, the out- side lap is one and one-half inches, and the port-opening on the line AB is one-fourth inch. In the particular case in question the shaded area represents both the portion of the stroke and extent of the port-opening. As many valve ellipses can be made as there are notches in the quadrant holding the reverse-lever. Again, while only the steam-port opening has been referred to, it should be evi- dent, since the displacement of both valve and piston are shown, that the time and extent of the exhaust-port opening can be deter- mined as well- as that of the steam-port. 155. Devices for Increasing the Acceleration of the Valve. The present purpose is less concerned with questions of port-opening 312 LOCOMOTIVE PERFORMANCE. than with the form of the curve representing the motion of the vaive, for it will be well at this point to discuss briefly certain modifications of existing gears designed to accelerate the valve motion when the piston is at the end of its stroke. It is the purpose of this class of de- vice to secure a quick and liberal opening of the port, to hold the valve nearly stationary during admission, and to close the port promptly at cut-off. The mechanical arrangements for securing these results usually consist in the addition of mechanism to existing valve-gears. A number of different gears have from time to time been proposed and used. While the means employed vary greatly in different gears the form of the valve diagram will be much the same for all. In con- trast with that shown by Fig. 191, it will be a figure having the same FIG. 191. Valve-motion Diagram. total length and the same total width, but with ends much flatter indicating the higher velocities of the valve on either side of its points of reversal. It is evident that increased acceleration requires a gear which is stiffer than that which is usually employed. This fact, together with a knowledge of the care which must be exercised in the construction and maintenance of normal gears, suggests some of the difficulties to be met in designing and maintaining any of these accel- erating devices. The conception underlying the accelerated device is good, but the locomotive is an extremely difficult type of engine upon which to employ it. The fact that in locomotive service the valve travel is considerable, and the speeds are oftentimes enormously high, makes it difficult to say just how far a design may be carried to secure such stiffness of gear and increase of wearing surfaces as may be required to perform the work imposed by any of the devices in question. 156. Wiredrawing as a Factor Controlling Valve-gear Design. Another point to be emphasized in connection with the valve ellipse is the small extent of port-opening. The smaller of the two ellipses shown represents a cut-off of 20 per cent of stroke. The greatest width of the shaded portion on a full-sized diagram is one-fourth inch, which is the maximum steam-port opening at this cut-off. LOCOMOTIVE VALVE-GEARS. 313 It is commonly assumed that all wiredrawing in locomotive ser- vice is objectionable, whereas, there is one very useful service which it renders. It is that of assisting in the maintenance of uniform conditions upon the boiler while the engine is operating under widely varying conditions of speed. The significance of this statement will appear if we assume an engine at a speed of 20 miles an hour, with reverse-lever and throttle set to create a demand for steam such as will put a fair working load on the boiler. If now, by a change in grade, the load is diminished sufficiently to permit the speed to increase to 40, 50, or even 60 miles an hour, it will generally not be necessary to change the reverse-lever or the throttle, the wiredrawing action coming into play to prevent the cylinder from taking from the boiler more steam than it can supply, the engine meanwhile continuing at all speeds to work at or near its maximum capacity. This automatic regulation of power through wiredrawing is one which means much, both as a matter of convenience in the operation of the locomotive and as a factor affecting boiler repairs. But there are objections to wiredrawing which are real and some- times serious, arising from the fact that as the process proceeds the cylinder action becomes less efficient. The loss on this account is, however, smaller than is generally supposed. Its value may be seen in the data of Chapter V., or better, perhaps, in a brief summary of facts derived from a locomotive carrying 200 pounds steam pressure, which is as follows: Speed 50, steam per I.H.P.H., 24.9 " 40, " ll I.H.P.H., 23.7 i ( 30, ll " I.H.P.H., 24.6 20, " " I.H.P.H., 25.4 These values correspond with the cards of Fig. 192 for the 28 per cent cut-off. Results for other cut-offs are substantially the same. At 60 miles an hour, were the data available, the consumption would probably rise to 27 pounds. It is a principle in steam engineering that, other things remaining the same, the steam consumption should diminish with increase of speed. This principle would apply in locomotive service were it not for the fact that with each increment of speed there is a change in the distribution of steam. The result is that while the steam consumption diminishes until a speed of 35 or 40 miles an hour is reached, after this point it increases, the losses through 314 LOCOMOTIVE PERFORMANCE. wiredrawing being greater than the thermodynamic gain due to the increase of piston speed. Referring to the figures above, there can be no doubt but that the change from 23.7 pounds to 24.9 pounds is due to wiredrawing. If for this series of tests the steam consump- tion had been obtained for a speed of 60 miles, it would have shown an increase of two pounds, or, say, 10 per cent above the minimum. In conclusion, therefore, with reference to wiredrawing with in- crease of speed, it may be said that its presence is a necessary accom- paniment of valves and gears now common ; that is, it serves a useful purpose in regulating the demands which the cylinders make upon 50 40 Cut-Off Per Cent of Stroke 15 FIG. 192. the boiler, but, at the highest speeds, it is responsible for some increase in steam consumption. 157. Improved Valve-gears. Countless devices have been proposed affecting either the valve or the gear which gives it motion, whereby the card may be made larger than that which results from the normal link-driven valve. A typical improved card is shown at A, Fig. 192. Concerning such devices it should be noted that it is usually assumed, though the assumption is erroneous, that anything which increases the area of an indicator-card is desirable. For example, in Fig. 192, for 20 per cent cut-off and a speed of 40 miles an hour (card A) , the plain out- line is the normal card around which has been drawn a so-called im- proved card. The difference is the shaded area and is presumably the result of the adoption of some new form of gear. Obviously, the shaded LOCOMOTIVE VALVE-GEARS. 315 area represents increase of power. The first mistake that is made concerning the change is that the increase in power results in no expense. Again, while the truth of the preceding statement may be admitted, it is often urged that one may measure pressure and volume represented by two indicator-cards, such as are shown by Fig. 192, and derive therefrom an estimate of the relative amount of steam used per horse-power per hour under the conditions which each represent. Such estimates are, in fact, fairly reliable when made between cards agreeing closely in form, and when all conditions of running are the same, but as a general proposition nothing is more misleading. If there are differences in speed, or in initial or final pressure, or in the number of expansions, the percentage of the total amount of steam used, which is shown by the indicator, will change. Anything which may produce a change in the temperature of the metal of the cylinder at any one point in its cycle is likely to produce changes in the whole cycle. As is well known, a considerable per- centage of the steam drawn from the boiler for each stroke of the engine condenses on entering the cylinder. While the interchange of heat causes some change in the amount of water in the cylinder as the piston proceeds on its course, by far the larger part of the initial condensation continues in the cylinder until the exhaust-port is open, when it flashes into steam and disappears with the exhaust. While the process is a complicated one, and cannot within the limits of this chapter be accurately defined, the fact is that any change in the form of any line bounding an indicator-card has its effect upon the amount of steam which must be admitted to make up the loss due to initial condensation. A change in the cycle remote from the period of admission may have as pronounced an effect on the quantity of steam required as a change occurring during the period of admission itself. There is, in fact, no way to measure the performance of a steam-engine but by a process which determines the actual weight of the feed or the exhaust. Again, a further illustration of the fact that a mere increase in the area of an indicator-card is not significant is to be found in the ease with which such increase of area may be secured. In locomotive practice it is quite unnecessary to adopt a new gear. If. under the conditions prescribed, the normal card A at 20 per cent cut-off (Fig. 192) is not large enough for the work, the reverse- lever may be advanced on its quadrant until the cut-off is 35 per cent, whereupon, in this particular case, the normal card B be- comes equal in size with the card representing an assumed im- 316 LOCOMOTIVE PERFORMANCE. proved gear. The real questions, therefore, may generally be stated as follows: Is the improved card at 20 per cent a better card than the normal card of equal area at 35 per cent cut-off? Will the former yield a horse-power upon the expenditure of less steam than the latter? It is upon this latter statement that the argument rests. No device which seeks to improve the steam distribution in a locomotive can succeed which does not save steam when compared with devices previously in use. In proportion as it saves steam it both increases the efficiency of the engines and increases their maximum power, for since the boiler capacity is limited a pound of steam saved is a pound of steam available for additional services. Turning now to a consideration of the margin upon which those who would improve valve-gears have to work, it must be admitted that it is not large. Results have already been quoted which prove that the locomotive with all its wiredrawing gives a horse-power in return for less than 24 pounds of steam per hour. This is near the minimum. From this performance of a simple locomotive having a normal valve- gear, with its narrow port-openings and wiredrawing effects, we may turn to the Corliss engine, the action of which is generally accepted by all improvers of locomotive valve-gears as a standard of perfection. Such an engine, with its large port-opening and its prompt movement of the valves, can in fact be relied upon to give as good a performance as engines having any other type of valve-gear operating under similar conditions of speed and pressure. Corliss engines, having cylinders which are comparable in size with those of locomotives, and which work under a similar range of pressure, are, however, not common, and hence it is not easy to command data for the proposed compari- son. Generally, simple Corliss engines work under a lower pressure than locomotives. A good performance for a simple Corliss engine exhausting into the atmosphere is that of an 18X48 Harris-Corliss engine, for which the steam consumption was 23.9 pounds per hour.* The steam pressure supplied this engine was only 96 pounds by gauge. On the basis given the engine should, when supplied with steam at 180 pounds, which is the pressure under which the locomo- tive data were obtained, require less than 23 pounds of steam per horse-power per hour. Straining the facts applying to the two classes of engines as widely apart as a knowledge of existing data will pos- sibly permit, we may assume that a Corliss engine, if given the advan- * Peabody's Thermodynamics. LOCOMOTIVE VALVE-GEARS. 317 tage of the high steam pressure and high piston-speed common in locomotive service, may give a horse-power hour on the consumption of two pounds less of steam, or approximately 8 per cent less than the locomotive. This, then, is the margin upon which those who seek to improve the locomotive valve-gear must expect to work. While it is well worth attention, it cannot revolutionize practice. 158. Foreign Valve-gears. Having now considered the defects and merits of the link motion, we may inquire concerning other forms Stationary Link FIG. 193. of gears which, in foreign countries at least, are in common use. In England one occasionally sees the stationary link, as shown by Fig. 193, the link-hanger of which is suspended from a fixed pin, A, while the reverse-shaft is connected with a radius-rod which communicates the motion of the link to the valve-spindle. A modification of this gear is -found in the straight or Allen link, shown by Fig. 194, in FIG. 194. which both the link and the radius-rod have connection with the reverse-shaft, which in reversing causes them to move in opposite directions. The Joy gear (Fig. 195), which has been, and still is, extensively used in English practice, requires no eccentric, but receives its motion directly from the main rod. The reversal of the engine and changes 318 LOCOMOTIVE PERFORMANCE. in the travel of the valve are, in this gear, accomplished by varying the inclination of the curved guide along which a block is forced to travel with each reciprocation of the main rod. The guide remains stationary, except as acted upon by the reverse-lever. Joy Gear FIG. 195. As a last example from foreign practice is presented the typical gear of Continental Europe, namely, the Walschaert gear (Fig. 196). In this gear motion is derived from the cross-head and from a single eccentric, or, where the gear is outside of the frame, from a return- crank. These motions are so combined, by means of a system of FIG. 196. connected rods and a link, as to give the usual motion to the valve. A radius-rod, connecting with the link-block, receives motion from the reverse-shaft in reversing and in changing the travel of the valve. The fact concerning these foreign gears which is to be emphasized is that they all give a motion to the valve, which may be repre- sented by a valve ellipse, such as that shown by Fig. 191. There are minor differences in the character of the resultant motion LOCOMOTIVE VALVE-GEARS. 319 which for the present purpose need not be discussed, the impor- tant fact being that the mechanisms in question are not to be regarded as reform gears. They are not to be contrasted with the Stephenson link motion, but compared with it. They do not exist because they give a better steam distribution than is obtained from the link motion of this country, and, except as may be hereinafter suggested, they are not superior to the gears common in this country. With this understanding of the matter, we may now examine the several gears referred to, and seek to find a reason for their existence in the con- ditions under which they are used. 159. Adaptability of Valve-gears. From that which has already been stated, it may be surmised that the Stephenson link motion, as used upon American locomotives, constitutes a convenient and accept- able gear. There are at least four good reasons why this is so, which may be enumerated as follows : 1. The gear gives a satisfactory distribution of steam. 2. Its design readily adapts itself to conditions involving the use of a rocker between the axle and the valve, and the rocker has until recently been regarded as an essential element in the design of the American locomotive. 3. The accessibility of the forward driving-axle, and the unoccupied space between the frames in American locomotives as a point of attachment for eccentric, has supplied ideal conditions under which to develop the link motion. 4. Other forms of gear which have been used elsewhere have been thought impracticable in American service, because, owing to the small diameter of driving-wheels, which, until recently, have pre- vailed in this country, the mechanism of such gears often extends too near the road-bed for safety. The stationary link (Fig. 193) has proven attractive to the designers of many English engines, because it involves the use of a single eccen- tric, room for which is more easily found upon the inside connected engines common to English practice, where the greater part of the forward axle is occupied by the cranks, than for the two eccentrics common in American practice. The Allen link responds to conditions similar to those which are met by the stationary link, but is a better design, since by it the lines of motion are kept much nearer the lines of force. The Joy valve-gear is distinctively English in its design and in the extent of its use upon locomotives. With the increased size of the inside 320 LOCOMOTIVE PERFORMANCE. connected locomotive, the inside crank of the forward axle gradually de- veloped proportions which made it difficult to find room even for a single eccentric. Confronted with these conditions, the Joy gear naturally makes a strong appeal to the designer. This gear takes its motion from a connecting-rod, and leaves the forward driving-axle to be occupied wholly by the cranks. For locomotive service the Joy gear is perhaps inferior to either of the other motions described, since irregularities in the track materially affect the distribution of steam. For example, when a low joint permits the wheel to drop, the connecting-rod par- takes of its motion and carries it on to the valve-gear and to the valve. The Walschaert gear, which is extensively used in Continental Europe, may be either inside or outside of the frames of the engine. In European practice, where inside cylinders have been much em- ployed, it has generally been outside of the frames. Its design makes a strong appeal to the designer who is forced to go outside of the frame with his valve motion. The gear is one in which the metal for the parts may be well distributed to transmit the stresses which are brought upon them, and for this reason it may constitute a very stiff gear, though its individual parts are relatively very light. For these reasons its use is likely to increase in American practice. 160. The Conclusion to be drawn from the foregoing discussion should not be one of discouragement for those who are interested in improving the locomotive valve-gear. Forms of gears now common have been commended, but this does not imply that better ones may not be devised. The argument is, however, that real and lasting improvement is to be looked for more along mechanical lines than in attempts to improve the character of the motion imparted to the valve. While there is a chance for slight saving in fuel, the real economy which may result from the adoption of a better gear is to be found in its lower maintenance cost, and in the greater certainty of its action, rather than in pounds of coal saved. What is most desired in a valve-gear is a mechanism which, under the adverse conditions of actual locomotive service, will give to the valve that precision of movement it is designed to have. Along these lines there is ample opportunity to improve present practice. CHAPTER XVIII. ACTION OF THE COUNTERBALANCE.* 161. The Problem of Balancing. In the mechanism of a loco- motive the revolving parts at the crank-pins, together with the re- ciprocating parts connected therewith, are balanced more or less completely by the addition of masses, or counterweights, to the drivers. But since the counterweights move in circular paths, it is only the horizontal component of the radial force derived from them which can serve to neutralize the effect of the reciprocating parts; the vertical component of all that portion of the force which applies to reciprocating parts is unbalanced. This unbalanced vertical com- ponent causes the pressure of the driver on the rail to vary with every revolution. Whenever the speed is high it is of considerable magni- tude, and its change in direction is so rapid that the resulting effect upon the rail is not inappropriately called a hammer-blow. Many practical demonstrations have been had of the magnitude of the forces involved. Heavy rails have been kinked, and bridges have been shaken to their fall, all under the action of heavily balanced drivers revolving at high speeds. Such results do not ordinarily follow from the operation of the locomotive, and indeed cannot do so if the locomotive is properly designed and its speed is confined to reasonable limits, but the value of the vertical component of the counterbalance of a modern loco- motive is nevertheless of considerable magnitude. This may be judged from the fact that the reciprocating parts on each side of such an engine exceed 1000 pounds in weight, and that the drive-wheels with which they are connected not infrequently turn 300 revolu- tions a minute. * The facts of this chapter were presented at the New York meeting (Decem- ber, 1894) of the American Society of Mechanical Engineers, forming part of Vol. XVI of the Transactions. 321 322 LOCOMOTIVE PERFORMANCE. Various attempts have been made to so dispose the reciprocating parts of a locomotive that, by their inter-action, a satisfactory degree of balance may be secured. The de Glehn balanced compound con- stitutes a well-known though not the only example of this type, but such engines have not yet come into general use in American practice, nor are they likely soon to do so. In dealing with the mech- anism now common to American locomotives the most which the designer has been able to accomplish in the matter of balancing is in the nature of a compromise. It is found that the mass of the locomotive is sufficient to absorb a portion of the force set up by the motion of the reciprocating parts, and that, as a consequence, the counterweights put in the wheels need not be sufficient entirely to balance these parts. Obviously, the less the weight which is put into the wheel the smaller will be the vertical component which constitutes the objectionable result arising therefrom. An old rule was to pro- vide balance in the wheel for from 75 to 80 per cent of the weight of the reciprocating parts. A later rule of the Master Mechanics' Asso- ciation provides that ^-J-^ part of the weight of the locomotive may be subtracted from the weight of the reciprocating parts, the difference to be the weight to be balanced. Engines designed in accord with these rules are so well balanced horizontally that they do not impart vibrations to the trains they pull, while the excess weight of the coun- terbalance is not sufficient to prove destructive to tracks or structures. The forces which are brought into action by the presence of the counterbalance have been elaborately studied, and their precise effect upon the pressure of contact between wheel and rail have at times been the subject of much discussion. The present chapter describes a series of experiments which were undertaken at the Engineering Laboratory of Purdue University to demonstrate the varying pressure between the revolving driver and the rail of the track. The essential features of these experi- ments consisted in passing a soft iron wire of small diameter under the moving wheel. It was expected that the varying thickness of the wire, which had been subjected to this process, would show the effect tff variation in pressure between the wheel and the track. If the wheel should leave the track entirely a portion of the wire would retain its full diameter; and the real purpose of the experiments, as originally planned, was to determine whether at any speed easily attained the driver would actually rise from the track. 162. Experimental Method. The apparatus employed consisted chiefly of the locomotive Schenectady mounted upon the Purdue ACTION OF THE COUNTERBALANCE. 323 testing-plant. To guide the wire, which was to be fed under the driver, a length of f-inch gas-pipe was secured to the laboratory floor in front of each driver included in the experiment (Fig. 197). Three pipes were thus arranged. A deflector plate was fixed behind the main driver to turn the wire delivered by this wheel away from the rear driver; but, except for this plate, no attempt was made to control the course of the wire after it left the wheel. The wire was of common annealed iron 0.037 inch in diameter. It was prepared by being carefully straightened and cut into lengths of twenty feet; that is, about 3.5 feet longer than the circumference of the drivers, and two incheslonger than the guide-pipe, in which the lengths were to be fed to the wheels. Wires thus prepared were laid in light wooden troughs to preserve them from injury, and a trough thus supplied was placed in line with each guide-pipe (Fig. 197). In conducting the experiments an Guide JPlpe .Wire . ^t Boas FIG. 197. operator at each pipe drew a wire from the trough and passed it into the pipe, until only about two inches of the length remained out- side. From the relative length of guide-tube and wire it was known that the opposite end of the latter was now close to the driver. When desired conditions of speed had been secured and a signal given, a touch of the operator's finger upon the end of the wire was sufficient to start the opposite end under the wheel. The starting of the wire was accomplished without commotion. The man in charge was conscious only of having touched it. To an observer who watched for the wire as it came from the driver it gave the impression of a quivering beam of light, which an instant later became a loosely tangled thread of metal. Or, if one kept his eye upon the wall of the laboratory, against which the wire was allowed to impinge, he saw the whole tangled coil appear instantaneously, and without apparent cause. The initial end of each wire was, in plan, of the outline shown by Fig. 198, from which it would appear that when the wire came under the influence of the wheel's motion the tensional stress upon sections near the end, as at A, exceeded the elastic limit of the material, this stress being required to impart motion to the mass of wire to the 324 LOCOMOTIVE PERFORMANCE. right of A. The weight of the twenty-foot length was about one ounce, and the time occupied in its passage was usually a fifth of a second. These facts will help to show the significance of the speeds used in the experiments. FIG. 198. The speed of the locomotive was noted from a registering-counter, and also by a Boyer speed-recorder, a permanent record being obtained from the latter instrument. To assist in connecting the effect pro- duced on the wire with definite phases of the wheel's motion a nick was made with a sharp chisel across the face of each driver, in line with the counterweight, as at A (Fig. 199). An impression of this nick was sharply defined upon every wire that passed under it. The initial end of the wire could, as has been stated, be determined by an examination; but to leave no doubt as to this matter, and for the purpose of giving a second reference point, one of the wheels was marked with two parallel lines ninety degrees from the first reference line, as atC (Fig. 199). Weight with tcJiich each driver presses on rail when at rest, 14 } OOO Ibs. 297 Ibs, It was found by a comparison of reference marks that distances along the length of the wires could be taken as representing equal distances around the face of the wheel. Thus, the length of each wire being greater than the circumference of the wheel, it would some- times happen that a single wire would receive two impressions from the same reference mark, the distance between the two points thus impressed upon the wire being found equal to the circumference of the wheel. This fact made it easy to connect effects left upon a wire with the wheel positions (crank-angles) producing them. Many of the wires produced by the experiment described have ACTION OF THE COUNTERBALANCE. 325 since been carefully calipered at five-inch intervals; the results plotted, and a smooth curve drawn through the points thus located. Some of the results thus obtained are presented as Figs. 200, 201, and 202, 04 5 10 J5 20 25 30 to'%w%%Wi%- ' / w- %'' : :ft 40 45 50 55 60 65. E Fortcard Driver ItigJit Side 75 80 85 W Bear Driver Bight Side FIG. 200. the points representing the actual thickness of the wires being desig- nated by means of small circles. It will be seen that all diagrams are plotted with reference to definite wheel positions. Scale Zeno** 1 One division-^ 9 e Thickness One division *. gM Side Speed -5.9 miles perTiour-3: yht Side Speed - G3 miles per hour- 337.2 rev. per min. Bear Driver-Right Side Speed-65 mites per hour -347. 9 rev. per min. FlG. 201. 163. The Balance of the Locomotive. Before attempting a dis- cussion of results in detail, it is necessary to consider somewhat briefly the condition of balance of the locomotive experimented upon. The engine, as delivered by its builders, was balanced for the road; but to increase its steadiness in the laboratory weights were afterward 326 LOCOMOTIVE PERFORMANCE. added in equal amounts to the several wheels, until a full horizontal balance had been secured.* The revolving and reciprocating parts which required counterbalancing, exclusive of the crank-pins and Length One division K* One division =.O1* Speed -BS wiles per hour 310.5 revolutions per minute i*+\ Wlieel U i Position Rear Driver - - Riyht Side Rear Driver -- Left Side 5 miles per hour 3^7.9 revolutio Rear Driver -\- Right Side Rear Driver Left Side FIG. 202. crank-pin bosses, which are assumed to be parts of the wheels them- selves, were found to weigh as follows : Piston and piston-rod 297 . Ibs. Cross-head with part of indicator rigging attached 170.5 " Main rod 344. 5 " Side rod. ., 278.0 " Total for one side 1090. Ibs. For complete horizontal balance it was required that the sum of the weights, making up the counterbalance of the two wheels on the side of the engine under consideration, should be equivalent to 1090.0 pounds acting at a radius of one foot. To ascertain the distribution * On January 23, 1894, the plant from which the results herein described were obtained was destroyed by fire. The new plant, now in operation, does not require the locomotive to be in complete horizontal balance. ACTION OF THE COUNTERBALANCE. 327 of balance between the wheels it was necessary to examine them separately. Calculations based upon drawings of the wheel centers gave the following results: Main Wheel. Rear Wheel. Balance cast in rim and between the arms, plus the weights added at the laboratory, all reduced to equivalent weights acting at a radius of 12 inches 744. 1 725. 7 Weight of crank- pin and crank-pin hub to be subtracted 187. 1 179. 1 Net weights available to balance revolving and reciprocating parts acting upon the crank-pins 557 . 546 . 6 The sum of the net weights thus obtained for both wheels (1103.6 pounds) is 13.6 pounds greater than the sum of the actual weights to be balanced. The engine is known to have been in perfect horizontal balance, the experimental methods adopted in securing this condition serving to indicate when the weights were changed even to the extent of a single pound. The calculated weight in each -1 Q f\ wheel is, therefore, assumed to be ^- =6.8 pounds heavier than the 2t weights themselves, and this amount has been subtracted as a correc- tion from the net weights given above, making the Main Wheel. Rear Wheel Corrected net weight of counterbalance, available to bal- ance revolving and reciprocating parts, acting upon the crank-pins 550. 2 539 . 8 The weights of the parts involved, together with certain dimen- sions, are summarized in Fig. 199. Taking the weights of side rod and of main rod, as already given, and considering 0.6 of the weight of the latter as a revolving part, Main Wheel. Rear Wheel. The excess of balance over that required for revolving parts alone is 204. 5 400. 8 which shows 66 per cent of the balance for reciprocating parts to be in the rear wheel. Six different rules for balancing locomotives for the road, reported as being in common use, give weights of counterbalance for the loco- motive in question, as follows: Main Driver. Rear Driver. Rule A (for freight-engines only) 467 260 B (for all classes of service) 462 322 " C " " " " " 547 340 " D " " " " " 570 340 " E " " " " " 573 366 " F " " " " " 588 381 Average of five rules from B to F inclusive 548 350 328 LOCOMOTIVE PERFORMANCE. Compared with these several standards the weights of the coun terbalances in the Purdue engine stand as follows: Main Wheel. Rear Wheel. By Rule A (for freight service only) 17.8% too heavy 107.6% too heavy " B (for all classes of service) 19.1% " " 67.6% " " D " " " " " 3.5% too light 56.9% " " E " " " " " 4.0% " " 47.5% " " " F " " " " " 6.4% " " 41.6% " " By the average of five rules from B to F inclusive ................... 0.4% too heavy 54.2% " It is evident, therefore, that the weight of the counterbalance in the rear wheel, from which most of the results about to be discussed were obtained, is in excess of that allowed by good practice, as expressed by the rules already given. But practice cannot always conform to the law by whic'h it assumes to be governed. It often happens where wheels are of small diameter, and the connections are heavy, as in mogul or consolidation engines, that there is not sufficient room in the main wheel to get in a counterbalance large enough for the revolv- ing parts alone; in this case, therefore, the balance for reciprocating parts of this wheel must be taken by the other coupled wheels, in addition to that which, under the rules, would be counted as properly belonging to them. By this process wheels having revolving parts, which are relatively light, are employed to balance a larger per cent of all the reciprocating weights. Again, almost any eight-wheeled engine, balanced in an approved manner, will, if the coupling-rod is removed, have an excess of balance in the rear wheel greater than that for the engine under consideration; and such engines are not infrequently run while disconnected. These considerations will serve to show that while the total weight of the counterbalances of the Purdue engine is, for reasons already stated, heavier than would be considered necessary for the road, and while at the time of the experiments the weights were not well dis- tributed between the wheels, yet the conditions which existed are not at all rare. Doubtless many wheels are running which carry a greater counterbalance, when compared with the revolving weights to be balanced, than did the rear wheel of the Purdue locomotive. 164. Results. Attention has already been directed to the fact that, in the engine experimented upon, the excess of weight in the counterbalance over that required for the revolving parts alone was much greater for the rear driver than for the main driver. As the ACTION OF THE COUNTERBALANCE. lifting effect is proportional to this excess of weight, it follows that" wires run under the rear driver were likely to show more variation, in thickness than those under the main driver. Results of experi- ments upon this point are shown by Fig. 200, which represents wires obtained at the same instant from the main driver and the rear driver respectively. It will be seen that the wire (7), from the main driver,, shows but slight variation in thickness, notwithstanding the high speed (312 revolutions per minute), and it may be said that no wire was ever obtained frqm this wheel which gave evidence that the wheel had left the track. From mathematical considerations it can be shown that this wheel would not be expected to lift at speeds below 80 miles per hour (428 revolutions per minute), and such speeds are not practicable with wheels of the diameter experimented upon. Passing now to an inspection of wire 77 (Fig. 200), from the rear wheel, which was obtained at the same instant with w r ire 7, it will be seen that there is a jump of the wheel just after the counterbal- ance has passed its highest point, which, when compared with the corresponding movement of the main driver, is very pronounced. Wires from this wheel at higher speeds are shown by Fig. 201. In this figure the full diameter of the wires is in each case shown by a dotted line drawn parallel with the base line. Wire 777, made at 59 miles (316 revolutions), shows that there was an instant in the passage of the wire, corresponding to the point A, when it was barely touched by the wheel. Increasing the speed to 63 miles (337 revolutions) increased the lifting action of the wheel to the extent shown by wire IV (Fig. 201). At the point B the wheel parted contact with this wire, and did not again touch it until the point C was reached, an interval of about 40 inches, the portion of the wire between B and C being entirely round and apparently unaffected by its passage under the wheel. A -further increase of speed gave, as is shown by wire V r a still greater length of full wire, the distance from D to E being very nearly equivalent to a quarter-revolution of the driver. It will be seen that all of these wires (77 to V, Figs. 200 and 201) substantially agree in showing the maximum lifting effect to occur after the counterbalance has passed its highest point, an effect un- doubtedly due to the inertia of the mass to be moved; also in showing that the rise of the wheel from the track is more gradual than its descent. The latter condition follows as a sequence of the first. Portions of the wires not shown on the diagrams do not vary much in thickness. The metal is rolled so thin by the normal pressure 330 LOCOMOTIVE PERFORMANCE. of the wheel that further increments of pressure do not greatly affect it. The wires, therefore, do not emphasize the destructive effect of the variation of wheel pressure when the change is insufficient to lift the wheel from the track. It now remains to mention the effect of certain disturbing ele- ments, which are shown by the experiments, to modify the actual movement of the wheel, other conditions remaining constant. For the rear wheel these disturbing elements are all in the nature of vibrations. The first to be noticed is the rockkig of the engine upon its springs, which motion tends to vary the pressure of the wheel upon the track independently of the action of the counterbalance. At one revolution the effect of the rocking may oppose the action of the counterbalance, and at the next revolution it may supplement the action of the counterbalance in producing a vertical movement of the driver. Again, the effect of the rocking may at a given instant be nil, and the wheel may rise under the action of the counterbalance; but in another instant the effect of the rocking appears, and the path of the wheel while in the air is modified and its time of descent changed. Thus, the existence of this vibration makes it impossible to duplicate wires with certainty, even though the speed is constant. Its effect is well shown by Fig. 202. Wires VI and VII were taken from the rear drivers at the same instant, one from the right side, the other from the left; the speed, therefore, must have been the same for both. The right driver lacked a good deal of leaving its wire, but the left driver was in the air for a tenth of a revolution. Again, wires VIII and IX were made in the same way at a higher speed; and here, while both drivers were off the track, the results are reversed, the right driver giving the greater length of full w r ire. It will also be seen from the diagrams that not only is the extent of the vertical movement of the driver modified by the rocking of the engine, but the position of the wheel when such motion occurs is changed. It is evident, therefore, that this movement of the engine upon its springs will prove a serious difficulty whenever an attempt is made to predict as to the precise movement of the center of gravity of the driver, whether the method of investigation be mathematical or experimental. There appears, also, to be a vibration of parts, as, for example, of the wheel as a whole, these vibrations being of small amplitude. Evidence of the presence of such vibration is shown by the location of points on the diagrams of wires, Figs. 200 to 202, which points repre- sent the thickness of the wires as found by measurement. Referring ACTION OF THE COUNTERBALANCE. 331 especially to wires 7 and II (Fig. 200) it will be seen that the actual thickness of the wire alternately increases and diminishes with every point. The time involved in passing from one high point to another (a distance of ten inches) was about 0.01 of a second. This vibration may be traced on other diagrams; its amplitude is from two to four thousandths of an inch only. Whether the process of introducing the wire starts, or has any connection with this vibration, the experi- ment does not show. A third class of vibrations is made apparent by a duplication upon the wire of the reference mark on the wheel. As has been stated a light nick from a sharp chisel was made across the face of the wheel to serve as a reference mark. This nick leaves a clear-cut projection upon the wire. But at high speeds the single nick across the face of the wheel leaves two projections upon the wire, showing that after making one impression the surface of the wheel must for an instant have actually cleared the wire and then impressed itself a second time. The distance between these projections on the wires varies somewhat, but is usually about an eighth of an* inch, which represents a time interval between the two impressions of about 0.008 of a second. The contact between wheel and track is, therefore, not continuous, but is a succession of exceedingly rapid impacts. These vibrations cannot affect the wheel as a whole; they are doubt- less due to the elasticity of the materials, and involve only the parts immediately about the point of contact. 165. Conclusions. The results of the experiments appear to justify the following conclusions: 1. Wheels balanced according to usual rules (which require all revolving parts, and from 40 to 80 per cent of all reciprocating parts, to be balanced, the counterbalance for the reciprocating parts to be distributed equally among the several wheels connected) are not likely to leave the track through the action of the counterbalance, and cannot do so unless the speed is excessive. 2. A wheel which, when at rest, presses upon the rail with a force of 14,000 pounds, and which carries a counterbalance 400 pounds in excess of that required for its revolving parts alone, may be expected to leave the track through the action of the counterbalance whenever its speed exceeds 310 revolutions per minute. 3. When a wheel is lifted, through the action of its counterbal- ance, its rise is comparatively slow, and its descent rapid. The maxi- mum lift occurs after the counterbalance has passed its highest point. 332 LOCOMOTIVE PERFORMANCE. 4. The rocking of the engine on its springs may assist or oppose 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. 5. The contact of the moving wheel with the track is not con- tinuous, even for those portions of the revolution where the pressure is greatest, but is a rapid succession of impacts. CHAPTER XIX. MACHINE FRICTION. 166. A Statement of the Problem. The power which is developed in the cylinders of a locomotive is the sum of that which is delivered at the draw-bar and the losses which occur between the cylinders and the draw-bar. These losses, in the case of a locomotive upon the road, include the frictional resistance of the machinery, the resistance of the atmosphere, and the rolling and journal friction of the loco- motive-truck and tender. Upon a testing-plant the only loss which occurs is that due to the factor first named. An interesting problem, therefore, which at once presented itself with the advent of the loco- motive testing-plant, was a determination of the amount of power absorbed by the machinery under different conditions of running. There had been no opportunity previous to the establishment of the Purdue plant to make such a determination in connection with loco- motives, and even in stationary practice the difficulties attending the measurement of high power when delivered from a fly-wheel has operated to limit available data to plants which are comparatively small in size. The process which must be followed to determine the machine friction of any locomotive under load is one requiring accurate obser- vations. This grows out of the fact that the value sought appears as a difference between two quantities which are relatively large. The two quantities involved are cylinder power and draw-bar power. The value of these may be ten times as great as their difference, and hence an error of one per cent in a determination of the indicated power, or of the draw-bar stress, may mean an error of ten per cent in the value of the difference which is the factor sought. It was with a full appreciation of the significance of this statement that the experiments to determine the machine friction of a locomotive were undertaken. 333 334 LOCOMOTIVE PERFORMANCE. 167. Methods. Every efficiency test of the Purdue locomotive is expected to yield all data necessary to a determination of machine friction. The present discussion, however, deals chiefly with tests which involved a short run designed to give in a few minutes' time such data only as is required to determine the frictional loss. Each such test was run under prescribed conditions of speed, steam pressure, cut-off, and throttle-opening. The engine having been warmed by preliminary running, was brought under the conditions prescribed for the tests, after which, upon signal, all observations were simul- taneously taken. These observations were four times repeated at intervals of four minutes, during which time the operation of the engine was maintained as nearly constant as possible. Each test, therefore, yielde 16 indicator-cards and such observed data as are given in the exhibit of observed results on page 335. This exhibit is a typical data sheet for such a test. In the conduct of this work four indicators were used, one of which was closely connected to each end of each cylinder. An absolute drum motion, connected by short cords, was in regular use upon the locomotive. Great care was employed throughout the tests in ques- tion in keeping the indicators in perfect order, in working up the cards, and in checking numerical calculations based thereon. Because of these precautions it was felt that the indicated power of the engine was obtained with a degree of refinement not often equalled in loco- motive work. A correct measurement of the power delivered at the draw-bar was found to be a matter of greater difficulty. Accuracy at this point involved not only the design of a traction dynamometer, but also a study of the effect of the mounting mechanism upon the locomotive as a whole. So serious were these difficulties that several years passed before accuracy was obtained in measuring the actual tractive power exerted by the locomotive. A brief description of the nature of these difficulties will be of interest. 168. Difficulties Encountered in Measuring Draw-bar Stresses. When the experimental locomotive was first installed in 1891, it was provided with a traction dynamometer made up of a very simple system of levers, which required a dash-pot of considerable size to steady them. It was soon found that the delicacy of this apparatus was insufficient to permit a high degree of accuracy in the determina- tion of draw-bar stresses. In consequence of this conviction, when, in the rebuilding of the laboratory after the disastrous fire of 1894, MACHINE FRICTION. 335 RUNNING CONDITIONS. Speed, miles per hour (approximate) 25 Steam pressure (approximate) 135 Reverse-lever, notch from center forward 3 Cut-off, per cent of stroke 45 Throttle Wide open. OBSERVED RESULTS. Gong. Time of Gong. Revolution Counter. Pressure. Dynamom- eter. Engine Position- 30 Sec- onds Before. 30 Sec- onds After. Differ- ence. Boiler. Dry- pipe. 1 2 3 4 3.20 3.24 3.28 3.32 2097 2637 3154 3716 2223 2760 3304 3837 126 123 150 121 136 126 126 128 132 124 124 125 5840 5850 5690 5860 + 1.5 -fl.5 + 1.5 + 1.5 Totals 520 516 505 23240 5810 + 6 + 1.5 Averag* 33 . 130 129 126 DERIVED RESULTS. Mean effective pressure (each value the average of four cards). Right head end 57.03 ' ' crank end 57 . 11 Left head end 61 . 69 ' ' crank end 61 . 33 Indicated horse-power. Right head end. 102.28 ' ' crank end. 99.25 Left head end 110.70 * ' crank end 106. 18 Total 418.41 Friction, horse-power. Total I.H.P. 418.41 Dynamometer horse-power= dynamometer pullxR.P-M. X. 0004956. ... . 374.32* Frictional horse-power 44. 09* Frictional resistance. Mean effective pressure equivalent to friction=F.H.P. -r- (.0542XR.P.M.) 6.25* Draw-bar pull equivalent to friction=M.E.P.Xl09.419. . . 683.8* * No correction applied for engine position. 336 LOCOMOTIVE PERFORMANCE. the opportunity came to improve the plant, it was determined to supply a dynamometer of high quality. An order was subsequently given to the William Sellers Company for an Emery machine, con- sisting of a hydraulic support, suitably mounted to receive the draw- bar, and a separate scale-case, all of which was duly received, and has since been used. This machine, while capable of measuring stresses as high as 30,000 pounds, is so sensitive that the pressure of one's finger upon the front bolster of the locomotive will produce sufficient motion in the whole mass of the locomotive to give an indication upon the dynamometer scale. Having possession of so delicate an instru- ment, it was with high expectation that tests to determine the engine friction were entered upon early in October of the first year of the reestablished plant. After carrying out a rather elaborate series of tests upon the plan already outlined, it was found that the results were unsatisfactory in that they appeared to follow no law, and because many of the values obtained indicated negative friction, the draw-bar power being greater than that of the cylinders. In entering upon the work at the beginning of a new school year in the fall of 1895, it was decided that the failure of the previous October had been due to the faulty position of the locomotive drivers upon their supporting wheels, and hence the true position of the drivers upon the supporting wheels, both when the engine was at rest and in motion, was first determined. The importance of this will be seen from the following considera- tions : With the locomotive at rest, and so located that a line joining the center of each driver with the center of its supporting wheel is vertical (Fig. 203), no force need be exerted along the line of the 00 Line of Draw-ba FIG. 203. draw-bar to hold the engine in its position upon its supporting wheels. Moreover, with the engine in this position and in motion the indica- tion of the draw-bar dynamometer should serve as a correct measure of the force transmitted from the drivers to the supporting wheels, MACHINE FRICTION. 337 that is, of the tractive force exerted by the locomotive. In the event, however, that the drivers are ahead of their neutral position on the supporting wheels, or back of the same (Fig. 204), it is evident that when the engine is at rest force must be exerted along the line of the draw-bar to hold it in position, and that with the engine in motion in either of these positions the indication of the draw-bar dynamometer will not measure the force transmitted from the drivers to the sup- porting wheels, unless corrected for the initial condition of stress in the draw-bar. It is obvious, therefore, that if comparisons are to be made between the power developed in the cylinders and that delivered FIG. 204. at the draw-bar the position of the drivers upon the supporting wheels must be neutral, or their position must be SD accurately known that correction may be applied. The significance of all this was appreciated from the first, though the difficulties in the problem were not anticipated. When the loco- motive had been installed upon the plant it had been carefully ad- justed to its position while cold. The failure to get consistent records at the draw-bar suggested that the spacing of the driving-axles might change sufficiently with the increased temperature of the frame resulting from working conditions to account for the difficulties encountered. Consequently, when the work was next undertaken (September, 1895), the wheels were lined up when the engine and boiler of the locomotive were heated to working temperature. More- over the effect of machine stresses was also taken into account, all measurements being repeated with steam in the cylinders, first with 338 LOCOMOTIVE PERFORMANCE. the reverse-lever in forward motion, and afterward with the reverse- lever in backward motion, the supporting wheels being prevented from turning by means of the friction-brakes. This process was repeated for various positions of the crank. The completion of this programme of preliminary work failed to disclose any gross errors in the position which the engine had occupied in the previous tests, and trial runs to determine values for engine friction proved no more satisfactory than those which had resulted from previous attempts. Meanwhile, the locomotive was being occupied with work not depending upon accurate records of draw-bar stress. Up to this time the traction dynamometer, an expensive machine selected especially because of the reputation of its type for accuracy, had been accepted as a standard of measure No one had doubted its accuracy. But the experiences already recounted, in combination with other events at the laboratory, required that some effort be made to check its indications. This was finally done by moving an Olsen testing-machine to the locomotive laboratory, by disconnecting the dynamometer from the locomotive, and by mounting the hydraulic head of the dynamometer, with all attached parts, upon the testing- machine in such a manner that a load could be imposed upon it in the testing-machine through the same spindle which had previously connected with the locomotive draw-bar. By these means it was possible to apply to the dynamometer any desired load as measured by the beam of the testing-machine, and to determine the indication which the dynamometer scale gave in response to these loads. The degree of accuracy attending the work, of course, was limited to the accuracy of the Olsen machine, but this proved quite sufficient for the purpose. The results of comparison proved that the indications of the dynamometer were too high and singularly irregular. These results having been confirmed by a representative of the manufacturer, the whole machine was shipped to the factory for cor- rection. It was there recalibrated, and in February, 1896, was restored to its place in the laboratory. Soon after attempts were again made to secure a measurement of the machine friction of the locomotive, but the results, while never showing negative friction and while much more consistent than those previously obtained, were still too discor- dant to be entirely satisfactory, whereupon nothing remained but to again give attention to the location of the wheels of the locomotive. That this might be more carefully investigated than on previous occasions a telltale was arranged to connect between the locomotive MACHINE FRICTION. 339 frame at a point midway between the two drivers and the wall of the laboratory (Fig. 205). It consisted of a small pointer, suitably mounted, and so proportioned as to magnify the actual movement of the locomotive approximately ten times. The end of the pointer passed over a stationary scale graduated from an arbitrary zero by divisions of equal value, each of which represented an actual move- ment of the engine- frame of an .0543 of inch. The value of this arrange- ment was to be found in the fact that any position of the pointer indi- cated at all times a certain relative position of drivers and supporting wheels. By its aid the problem of locating the drivers upon the supporting wheels resolved itself into the finding of a "position cor- rection" corresponding to the different positions of this pointer. y F A <> @ ,11 > 7" - Point in Locomotive Secured to Frame Midway laboratory "Wall between Drivers FIG. 205. As it was found that correct values for draw-bar stress could not be read directly from the dynamometer, an attempt was made to determine the amount of the position correction by direct weighings. The method of procedure was to move the locomotive by means of .a jack-screw, until the pointer stood at zero, and with an initial load on the dynamometer to balance the scale-beam. Next, without changing the initial load, the locomotive was moved by means of .the jack until the pointer stood at each of the several different division marks on the forward scale, and thence back to zero, the dynamometer being balanced and readings taken for each position both forward and back. It was expected that when the dynamometer readings thus obtained were plotted, they would give a line which would determine the correction to be applied to the dynamometer rea-dings when the engine was working in any of the given divisions experimented upon. The results, however, did not justify the pains, since the values ob- tained showed great irregularities, and, while the experiment was conducted with great care, was carried out with many variations, 340 LOCOMOTIVE PERFORMANCE. and was many times repeated, it did not suffice to give a satisfactory position correction. he conclusion reached was to the effect that minor forces were brought into play whenever the engine was moved, and that the presence of these prevented the machine from behaving in the same way in every different position. It was thought, how- ever, that these would disappear if similar tests could be made when the engine was in motion. To accomplish this involved the following program : With the locomotive in any given position and running ahead, the pull registered by the dynamometer will be D F =M F -F+x, ....... (18) where Mp =pull corresponding to the M.E.P. in the cylinders, engine running ahead; F = pull equivalent to the machine friction; x = pull due to engine position, which is the correction sought. With the locomotive in the same position, running aback under the same conditions of speed, cut-off, and steam pressure as when running ahead, it may be assumed that the machine friction will be the same, and consequently that the draw-bar pull will be DB=M B -F-X, ....... (19) where MB =pull corresponding to the M.E.P. in the cylinders, engine running back; F and x have the same values as above. Solving these two equations for x, we have It will be seen that equation 20 makes the value of x, which is the correction for position, depend upon M and D. Numerical values for the former are obtained directly from the indicator-cards, and for the latter by direct observation from the traction dynamometer. The only assumption involved by the equation is that the friction of the engine is the same, when the same conditions of speed, cut-off, and initial pressure in the cylinders are maintained, whether the engine is running ahead or aback. The value of x, if considered positive when MACHINE FRICTION. 341 the engine is running ahead, must be considered negative when the e gine is running aback. In the application of this program a series of position tests were made in March, 1898, with results which were too much at variance one with another to be accepted as satisfactory. A month later it was determined to repeat the work with all of the friction-brakes removed from the axles of the supporting wheels, so that the power developed should approach closely the friction-load, whereupon it was found that with the engine running ahead there was compression in the draw-bar. An investigation of the cause of this unexpected result revealed the fact that the friction of the front trucks on the rails was too great to allow a free movement of the locomotive forward and back. Previous to this time but little attention had been given to the trucks, the belief being that the vibration of the engine in motion would be sufficient to overcome any resistance which they might offer. A careful investigation at this time, however, revealed the fact that the rails which were mounted on wood had yielded under the pressure and vibration of the truck, becoming slightly deformed under each wheel. This fact having been determined, the truck was removed, the truck wheels, which had been cast in a contracting chill, were ground truly round, the wooden support for the rails was removed, and the work rebuilt in masonry and capped with metal. The rails were planed on their upper surface and drawfiled. The effect of these improvements was marked, and the results obtained from posi- tion tests which followed, when compared by means of equation 20, gave evidence for the first time of a definite law. So much improvement had, in fact, been effected by bettering the mounting of the truck, that it was finally decided to entirely elim- inate friction from this source by its entire removal. That this might be accomplished, the front of the engine was suspended from the roof of the laboratory, the construction of the building being such as readily to admit of such a plan (Fig. 206). Two rods, 17 feet in length, connected the front bolster with their point of support. With the engine at zero position upon the supporting wheels, these rods were made to stand vertically. Thus supported, the front of the engine was as free to move in any direction as a pendulum of equal weight. With the engine thus suspended, tests were run to determine the indicated power- and the tractive power for each position of the indi- cator (Fig. 205). By the manipulation of the turn-buckle connecting 342 LOCOMOTIVE PERFORMANCE. between the draw-bar and dynamometer, the pointer of the telltale was allowed to occupy in succession the several positions along its scale, and in each position observations were made in triplicate. This was done with the engine running both ahead and aback. The reverse- lever was in all cases in its extreme position. The speed was as nearly as possible the same for each test, and the load was only that due Roof Beam FIG. 206. Method Used to Determine Position of Engine. to the friction of the supporting-wheel axles. A summarized state- ment of the results of these tests appears as Table LXVIIL, and the value of the position correction, as obtained by plotting these values, is shown by Fig. 207. The straight line of this diagram represents the average value of the readings given in Table LXVIIL The equa- tion for this line is a = -199 4- 44 y, (21) where x = pull due to engine position ( + if tension, if compression); y = position of locomotive as shown by pointer (+ if forward, if back of the arbitrary zero). MACHINE FRICTION. 343 The point at which the mounting mechanism has no effect on the draw-bar is given as +4.52. 250 \ 200 a '7 a 150 n B+ 100 r 2 ^ CH x * c ) y Pos tun 4 of Engi 3 ne as SI own by Poi Itl'l Positior of' Kny 3 ne as SI 4 S\ lovvixoy 5 Pointer . -1 ack froi nC ?nt t + : "or\ /ard fro 11 C >ute J/C 3 a / 100 ( x / ) 150 | x S 3 / ( Xc >|'S / 250 ^ x ' II ^ ^ 300 / ' C ) B ^ 350 a j y / 400 / > / FIG. 207. Draw-bar Pull Due to Engine Position. When the locomotive is in the position corresponding with the zero position of the pointer, which, it will be remembered, was arbit- rarily located, y will equal zero and X--199. 344 LOCOMOTIVE PERFORMANCE. TABLE LXVIII. INDICATED AND TRACTIVE POWER FOR DIFFERENT POSITIONS OF THE DRIVERS ON SUPPORTING WHEELS. Engine Running Ahead. Engine Running Aback. .o'-s'e-i 41 1 ons per 2* |S i; ffjj OJ 1 ons per 5J J 1 p to" si H ||| ine Posit aic Sum dVII) + : 1. Compre j 33 3 fij .2 J "3 fj j lf r^ *3 ^ ttC ^ C *O ^ a Jg i| ffe R 1 J o -^ S-SjS 6s' PI! H i 3 ; m gjf laiilel^l j- s j| -2< ' i 31 KS - 3 Jftj^. If JH* Q al tf.2 . 3 l^-rf S a g Q^GoQG_;35 i 1 IF tb 3l il lS- 11 H laii 1 "gJS"O J3 isSs 02 02 PH PH 02 * PH PH PH i. II. III. IV. V. VI. VII. VIII. IX. X. XI. XII. 29 54 2552 2245 307 25 53 2151 2203 -52 -179 + 2 27 55 2405 2168 237 25 52 2146 2154 -8 -123 + 4 25 61 2155 2016 139 26 51 2225 2162 63 - 38 +8 29 48 2544 2532 12 27 49 2315 2149 166 + 77 + 4 27 53 2300 2210 90 27 48 2264 2168 96 + 3 + 2 28 51 2397 2209 188 24 58 1925 1913 12 - 88 27 50 2368 2097 271 22 59 1848 1932 -84 -178 -2 26 52 2276 1903 373 25 52 2129 2345 -216 -295 -4 24 56 2097 1657 440 25 52 2183 2490 -307 -374 -6 23 53 2071 1595 476 25 51 2193 2516 -323 -400 -4 23 56 2042 1565 423 26 50 2249 2560 -311 -367 -2 23 57 1968 1582 386 25 49 2212 2456 -244 -315 23 56 2051 1765 286 23 49 2295 2448 -153 -220 That is, under these conditions, 199 pounds must be subtracted from the observed reading of the dynamometer; corrections for other positions may, of course, be read directly from the diagram (Fig. 207), or found by means of equation 21. It will be evident, also, that having the data which appears in Table LXVIII., the value of the correction for any position of the loco- motive may easily be calculated by means of equation 20. The values given are: In Col. IV, M F ] in Col. V, Dp] in Col. IX, M B ', and in Col. X, DB of this equation. The application of this equation to the data of any given position will give a result consistent with that which appears in the same line under Col. XII. With reference to the accuracy of the foregoing conclusions, it should be noted that data was obtained for all positions of the loco- motive from 1.5 to .25, a range which extends far beyond the limits of Table LXVIII. and Fig. 207. The extreme values, while of use in MACHINE FRICTION. 345- determining the slope of the plotted line, are omitted as unnecessary to the present discussion. Finally, with reference to the long and laborious process which has; been described, the fact should be emphasized that most of the diffi- culties encountered were of a kind which, in the light of past experi- ence, should have been easily avoided. It would be wholly wrong to conclude that a locomotive testing plant, constructed along the lines developed at Purdue, is defective as an instrument for measuring the tractive power of a locomotive, or that accuracy of observed data is obtained at the expense of unusual effort. It should be clear that defects in the rating of the dynamometer, which for so long a time proved so perplexing, should not be charged against the principle underlying the plant. The importance of one rule of practice, however r is emphasized by the experience in question. It is to the effect that, for very accurate work, too much reliance should not be placed upon vibration as a means of overcoming truck friction. The rail surface should be in perfect form. If the same locomotive is to be operated on a plant for a considerable time, it will be necessary frequently to inspect the rails, to make sure that corrosion or side movement of wheels has not formed incipient grooves under the tread of each wheel. If care is given this matter, and if the drivers are made plumb over the supporting wheels by careful measurement, the record of the dynamometer will be sufficiently accurate for every practical purpose, even though the stresses to be measured are relatively light, 169. Friction Tests and their Results. Having obtained a dyna- mometer which could be depended upon, and having determined the engine position with such care as to permit the measurement of draw- bar stresses with a high degree of accuracy, a series of tests to deter- mine machine friction were undertaken. In proceeding with the- work, the locomotive having been warmed by preliminary running, was brought under conditions for which information was desired r Upon signal, indicator-cards were taken, the draw-bar pull ascer- tained, and all running conditions observed. All observations were taken simultaneously, and were three times repeated at intervals of four minutes, after which other conditions of running were sought and another test was entered upon. Tables LXIX. and LXX. which in effect constitute a single table, give in each line the average of the three observations making up one test and the calculated values obtained therefrom. The running conditions are set forth by Cols. 2 to 7 the power developed in the cylinders and at the draw-bar by 346 LOCOMOTIVE PERFORMANCE. . TABLE LXIX. RESULTS OF FRICTION TESTS. J a 9 * 1 1 2 3 4 6 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 1 Speed. Cut-off. * Pressure M.E.P. Miles per Hour. 2 R.P.M. Notch For- ward from Center. Per Cent of Stroke. Boiler. Dry- pipe. Right. Left. Aver- age. H.E. C.E. 9 H.E. 10 C.E. 3 4 5 6 7 8 11 12 47.35 42.36 38.83 37.30 34.43 31.00 27.29 23.33 14.35 15.04 25.16 25.70 35.64 35.47 46.68 56.83 14.29 24.94 35.96 46.48 58.06 77.22 80.94 135.4 138.3 191.8 190.9 251.2 305.8 76.89 134.2 193.5 250.6 312.4 25 25 25 25 25 25 25 25 132 128 129 130 132 128 132 131 128 124 124 127 126 125 129 127 42.87 37.30 34.05 33.51 30.64 25.98 23.17 18.42 51.77 44.98 42.67 40.71 36.86 33.33 28.30 24.34 46.95 42.17 38.98 35.75 .34.33 30.71 27.69 24.49 47.82 44.98 39.64 39.24 35.91 34.30 30.00 26.09 2 2 2 2 2 35 35 35 35 35 45 45 129 124 126 125 123 131 125 135 132 125 119 121 121 122 128 118 57.16 46.84 38.79 32.04 25.04 63.74 53.03 45.23 37.61 29.75 61.49 51.62 44.66 37.57 30.64 61.85 52. 7i 46.78 38.83 31.74 61.06 51.07 43.86 26.51 29.29 66.73 54.28 25.48 36.70 137.1 197.5 3 3 9 9 2 2 2 2 61.78 49.85 66.82 54.68 69.65 56.10 68.68 56.50 14.38 14.74 77.40 79.30 191.6 191.9 193.1 191.5 80 80 81 73 65.74 60.06 68.26 62.49 70.44 62.17 72.13 63.35 69.14 62.02 35.61 35.66 35.88 35.59 35 35 35 35 148 126 95 76 104 100 95 76 31.23 30.66 27.77 21.35 32.86 33.00 32.77 33.61 34.69 34.47 35.27 30.78 37.26 35.75 31.80 24.99 44.16 45.80 44.57 45.45 43.79 41.22 41.57 39.34 37.05 35.35 31.56 22.91 40.00 41.30 40.05 41.02 41.24 39.73 41.14 36.37 38.48 35.93 33.61 25.10 36.00 34.42 31.18 23.59 39.67 40.83 40.15 40.87 40.34 38.80 39.83 35.77 24.16 25.09 25.18 25.46 25.27 25.18 25.46 26.48 130.0 135.0 135.5 137.0 136.0 135.5 137.0 142.5 1 1 1 1 1 1 1 1 25 25 25 25 25 25 25 25 127 129 124 126 131 129 132 130 124 125 122 123 129 127 130 124 41.64 43.24 43.23 43.40 41.46 39.78 41.35 36. 5f Cols. 9 to 18, and the friction loss as expressed in various terms by Cols. 19 to 22 The dynamometer pull equivalent to the indicated horse-power (Col. 18), depends upon the relation of cylinder dimensions to diameter of drivers. It assumes no loss of power in transmission from the cylinder to draw-bar and is expressed by the equation Ibs. M.E.P. in each cylinder = .009139 times Ibs. pull at the draw-bar. MACHINE FRICTION. 347 TABLE LXX. RESULTS OF FRICTION TESTS (Continued). Number. I.H.P. Dynamometer. Friction in Terms of Actual Read- ing. Correc- tion (to be added). Dyna- mometer Reading (cor- rected). D.H.P. Dyna- mometer Pull Equiva- lent to I.H.P. I.H.P. Per Cent of I.H.P. Pounds Dyna- mometer Pull. Pounds? M.E.P. 1 13 14 15 16 17 18 19 20 21 22 1 2 3 4 5 6 7 8 198.1 185.7 285.0 285.6 357.8 321.3 371.4 396.1 4,392 3,936 3,472 3,449 3,015 2,681 2,292 1,799 166 133 168 133 169 141 194 205 4,558 4,069 3,640 3,582 3,182 2,822 2,486 2,004 174.5 163.2 244.4 245 .4 302.4 267.0 309.5 303.9 5,177 4,630 4,245 4,169 3,764 3,396 2,983 2,614 23.6 22.5 40.6 40.2 55.4 54.3 61.9 92.2 11.95 12.12 14.25 13.07 15.46 16.90 16.66 23.31 619 561 605 587 582 574 497 610 5.66 5.13 5.53 5.40 5.33 5.24 4.49 5.57 9 1 2 3 254.4 371.4 460.0 495.6 497.1 6,045 4,965 4,188 3,402 2,524 141 151 155 184 184 6,186 5,116 4,343 3,586 2,708 235.6 340.1 416.6 445.3 419.1 6,678 5,585 4,796 3,992 3,212 18.8 31.3 43.4 50.3 78.0 7.38 8.47 9.43 10.12 15.65 492 469 453 406 504 4.50 4.28 4.14 3.71 4.61 4 5 495.9 582.0 6,768 5,424 130 134 6,898 5,558 468.8 544.4 7,297 5,939 27.1 37.6 5.49 7.31 400 381 3.66 3.48 1.30 2.08 3.84 3.98 4.41 4.35 6 7 8 9 1 290.1 266.6 7,284 6,416 136 141 7,420 6,557 284.6 257.7 7,562 6,784 5.5 8.9 1.89 3.33 142 227 420 435 483 476 373.9 358.0 327.5 245.1 3,349 3,193 2,812 1,978 167 135 128 130 3,516 3,328 2,940 2,108 333.9 316.6 281.3 199.9 3,936 3,763 3,423 2,584 40.0 41.4 46.2 45.2 10.70 11.62 14.11 18.42 2 3 4 5 6 7 8 9 285.4 298.5 294.8 303.2 296.9 284.9 295.7 276.1 3,732 3,617 3,592 3,580 3,662 3,472 3,650 3,225 156 130 149 144 140 137 133 130 3,888 3,747 3,741 3,724 3,802 3,609 3,783 3,355 250.5 250.7 251.2 252.8 256.2 242.4 256.8 236.9 4,430 4,462 4,390 4,466 4,405 4,242 4,356 3,910 34.9 47.8 43.6 50.4 40.7 42.5 38.9 39.2 12.23 16.01 14.79 16.62 13.71 14.91 13.12 14.19 542 715 649 742 603 633 573 555 4.95 6.53 5.93 6.78 5.51 5.7$ 5.24 5.07 170. A Comparison of Results. The factors, upon which the loss* of power by friction in the machinery of a locomotive depends, con- cern chiefly the lubrication of the moving parts. If a viscous oil' is used in any or all of the journals it should be expected that the- frictional loss will be greater than when a more limpid oil is used. Again, with a given lubricant in service, the coefficient of friction; 348 LOCOMOTIVE PERFORMANCE. is affected by the temperature of the lubricant and of the metallic parts surrounding. Within limits which are rather wide the friction 'diminishes as the temperature of the lubricant is increased. In view >f these facts, therefore, attempts to measure frictional loss should not be expected to give results which are free from inconsistencies, isince it is obviously impracticable to control the temperature of all the journals of a locomotive, or even the precise rate at which lubri- cant is supplied them. The most that can be attempted is to secure results under a sufficient number of different working conditions to fairly represent the ordinary running conditions of the locomotive. Minor inconsistencies between individual results must be expected. The preceding statements will naturally suggest the fact that the operation of a locomotive, when first started, is attended by greater frictional loss than after its parts have become well warmed by running. The truth of this statement was demonstrated by actual test. After being at rest for a period of twenty-four hours the Purdue locomotive was started and brought to a speed of 25 miles an hour, with the reverse-lever in the first notch forward of center and the throttle fully open. As soon as possible after starting, observations necessary to a determination of the frictional loss were made, and were thereafter repeated at five-minute intervals for a period of thirty minutes, after which the interval between observations was increased to ten minutes, and still later it was further increased to twenty minutes. At the end of the first ten minutes the friction, as expressed in pounds M.E.P., was 6J; after twenty minutes it had fallen to 5.29, after which the friction fluctuated by small amounts both above and below this value throughout the remainder of the run, which was continued for 110 minutes. Referring to Tables LXIX. and LXX. it will be seen that all tests II = = l o a jnojj jad eanK ' HH O CO (N lO lO O CO CO CO il a a^nmj\[ jad suoy^n|OAa^j HH HH ^ o o GO OS GO 0^ UOpBUSlsaQ ^JO^BJOq'B r [ HH hH ir/2.,2 ? ? 4 lO O lO co co co HH -H W !, 1 2 3 4 5 6 7 .8 9 10 11 12 13 14 Equivalent Evaporation per Square Foot of Heating Surface per Hour FIG. 215. Evaporative Performance. coal. Thus, suppose that at a certain speed and load the boiler, as at present designed, is required to evaporate 11.8 pounds of water per square foot of heating-surface per hour, its evaporative efficiency will be that represented by the point a. For the present purpose it may be assumed that an increase of 20 per cent in the weight of the boiler will give a proportional increase in the extent of it? heating- surface, the pressure for which it is designed remaining unchanged. 370 LOCOMOTIVE PERFORMANCE. On this assumption the boiler with 20 per cent increase in weight would at the same power be compelled to evaporate only 9.7 pounds of water per square foot of heating-surface, and its efficiency would rise to that represented by the point c, an increase in -efficiency of about 11 per cent. This increase in efficiency is due wholly to the increased capacity of the boiler. Numerical values, based on the assumption that increase in capacity will be proportional to increase of weight, are given in Table LXXIX., from which it will be seen that any increase in the size of the boiler can be depended upon to yield a definite return in the improved performance of the locomotive. TABLE LXXIX. SAVING IN FUEL BY USING A BOILER THE CAPACITY OF WHICH IS GREATER THAN AN ASSUMED NORMAL BOILER. Pounds of Water required to be Evaporated per Square Foot of Heat- ing-surface per Hour in a Normal Boiler. Percentage Saving in Fuel by Using a Boiler the Capacity of which is Greater than the Normal Boiler by 5, 10, 15, and 20 Per Cent Respectively. 5 Per Cent. 10 Percent. 15 Per Cent. 20 Per Cent. I. II. III. IV. V. 5 0.8 1.5 2.2 2.9 6 1.0 1.9 2.8 3.7 7 1.2 2.3 3.4 4.5 8 1.5 2.9 4.1 5.3 9 1.8 3.4 4.8 6.1 10 2.0 3.9 5.5 7.1 11 2.3 4.5 6.4 8.2 12 2.7 5.1 7.4 9.4 13 3.0 5.7 8.4 10.7 14 3.4 6.4 9.4 12.0 15 3.8 7.4 10.6 13.6 183. Summary. The preceding paragraphs all refer to a simple cylindered locomotive using saturated steam. They give evidence to the fact that locomotive efficiency can be increased either by employing a higher pressure or a larger boiler, and some data are presented tending to show the rate of change in each case. The problem in design which is here touched upon may be stated as follows: When there is opportunity to increase the weight of a locomotive boiler of a given class, will it be better to make a stronger boiler that higher steam pressure may be carried, or a larger boiler that the rate of evaporation may be reduced? In the one case the advantage will appear in the improved performance of the cylinders, and in the EFFECT OF HIGH STEAM PRESSURES. 371 other in the higher evaporative efficiency of the boiler. The variables in the problem have each been considered, and an attempt has been made to estimate their value, but its complete solution, which will fix limits and give a numerical measure of the benefits to be derived as the result of definite changes along either line, must await the accumulation of a greater array of facts than are at present available.* * Since the preparation of this manuscript, there has been completed at the Purdue Laboratory (Aug. 1906), under the patronage of the Carnegie Institution, an elaborate series of tests to determine the value of high steam-pressures in loco- motive service. Values whicb fairly represent the consumption of steam ID pounds per horse-power hour under different pressures are as follows: Boiler- Steam per I H. P. pressure. per Hour. 120 29.1 140 27.7 160 26.6 180 '. 26.0 200 25.5 220 25.1 240 24.7 Tests were run under the several pressures given at different speeds and cut-offs. The values quoted do not represent the minimum steam consumption nor the maxi- mum, but are those which have been determined from a full analysis of all data. They may be accepted as fairly representative of the performance of the Purdue locomotive under normal conditions of running. A summarized statement covering the work and the results derived therefrom is as follows : 1. The results apply only to practice involving single-expansion locomotives using saturated steam. 2. Tests have been made to determine the performance of a typical locomotive when operating under a variety of conditions with reference to speed, power, and steam-pressure. 3. The rate of change in efficiency resulting from changes in steam-pressure has been established by the results of carefully conducted tests. They show that the higher the pressure the smaller the possible gain resulting from a given increment of pressure. An increase of pressure from 160 to 200 pounds results in a saving of 1.1 pounds of steam per horse-power hour, while a similar change from 200 to 240 pounds improves the performance only to the extent of .8 of a pound per horse -power hour. 4. The improvement in performance with increase of pressure will in service depend upon the degree of perfection attending the maintenance of the locomotive. The values quoted in the preceding paragraph assume a high order of maintenance. If this is lacking, it may easily happen that the saving which is anticipated through the adoption of higher pressure will entirely disappear. 5. The difficulties to be met in the maintenance both of boiler and cylinders increase with increase of pressure. 372 LOCOMOTIVE PERFORMANCE. 6. A simple locomotive using saturated steam will render good and efficient ser- vice when designed to work under a pressure as low as 160 pounds; under most favor- able conditions no real advantage is gained by designing for pressures greater than 200 pounds. 7. Wherever the water which must be used in boilers contains foaming or scale- making admixtures, the pressure for best results should not exceed 180 pounds; where feed- water is exceptionally bad, it will be found advantageous to fix the maxi- mum below this limit. 8. As the scale of pressure is ascended, an opportunity to further increase the weight of a locomotive should in many cases find expression in the design of a boiler of increased capacity rather than one for higher pressures. 9. For the development of a given power, any increase in boiler capacity brings its return in improved performance without adding to the cost of maintenance, or opening any new avenues for incidental losses. As a means to improvement, it is more certain than that which is offered by increase of pressure. 10. From the simple standpoint of efficiency, and neglecting all questions of maintenance above 180 pounds, it is better to utilize any allowable increase in weight by providing a larger boiler rather than to provide a stronger boiler to permit higher pressures. 11. Pressures designated in the preceding paragraphs are to be accepted as run- ning pressures. They are not necessarily those at which safety-valves open. 12. The preceding statements justify the conclusion that steam-pressures in American locomotive service have already been carried to a limit which, for best results, should be accepted as maximum. CHAPTER XXII. CONCERNING DIAMETER OF DRIVING-WHEELS. 184. Practice with Reference to Wheel Diameters. Prior to the renascence of the American locomotive, which had its beginning in 1894-95, the practice of this country was committed to the use of driving-wheels of comparatively small diameter. With wheels thus proportioned, locomotives were very effective at slow and moderate speeds, but at high speeds the rotation reached limits which were not equaled in the practice of other countries. Under this practice the driving-wheels of fast passenger locomotives were driven in regulai and ordinary service to a speed of between 300 and 400 revolutions a minute; while a rule much employed by a prominent builder was to make the diameter of the drivers in inches equal to the speed in miles per hour for which the locomotive was designed, a rule which contemplates a speed of rotation of 336 revolutions per minute. With the upbuilding of the modern American locomotive, driving- wheels for all classes of service have been materially enlarged, so that in ordinary service the speed of rotation is now not so high as for- merly, notwithstanding which fact an analysis of the general question will not be without interest. 185. A Study Based upon Observed Facts. From the data and discussion of Chapter V., it is evident that whenever the speed of rotation exceeds the critical speed a loss of efficiency results. If, therefore, high train speeds are demanded, the diameters of the driving- wheels should, for best results, be so proportioned as to give the desired rate of travel without exceeding the critical speed of rotation. The following illustration will serve to show how this principle may be worked out hi the case of Schenectady No. 1. The proportions of the locomotive are such that, neglecting fric- tion, every pound of mean effective pressure exerted in the cylinders will produce a draw-bar pull of 109.4 pounds. It has already been 373 374 LOCOMOTIVE PERFORMANCE. shown (Chapter V., paragraph 36) that the power of this locomo- tive becomes maximum when the speed reaches 188 revolutions per minute. Examining the conditions, first, with reference to a cut-off of 35 per cent, it will be found that the mean effective pressure at this speed is 42.4 pounds (Test 35-2- A), which is equivalent to a pull at the draw-bar of 42.4X109.4 = 4639 pounds. If, now, it is required to operate this locomotive at a speed of 55 miles per hour, the revolutions must increase from 188 to 296 per minute, which will cause the mean effective pressure to drop to 27.4 pounds (Test 55-2- A). This makes the draw-bar pull at 55 miles 27.4X109.4=2997 pounds. Suppose, now, that instead of increasing the speed of rotation from 188 to 296, the diameter of the driving-wheels were increased from 63 inches, the present diameter, to 99 inches, a diameter which will permit a speed of 55 miles an hour at 188 revolutions per minute. With these new proportions the locomotive would give but 69.4 pounds pull at the draw-bar for each pound mean effective pressure. But a speed of 55 miles would now involve only 188 revolutions per minute, and the mean effective pressure would be 42.4 pounds (Test 35-2- A), which would give a pull at the draw-bar of 42.4X69.4 = 2943 pounds, a result which is practically identical with that obtained with the smaller wheels and a higher speed of rotation. In this case, therefore, there has been no loss or gain, so far as power is concerned when operating at a train speed of 55 miles an hour, by the substitution of 99-inch drivers for the 63-inch, which are normal to the engine. Since, however, the engine is more economical in its use of steam at the critical speed than when the revolutions are greater, there would be some gain in mechanical action through the use of the larger driving-wheels, since, at 296 revolutions, the cylinders require 32 pounds per horse-power hour, while at 188 revolutions they require but 26.3 pounds, a gain of 21 per cent. A similar comparison based upon a cut-off of 25 per cent gives results which are as follows : CONCERNING DIAMETER OF DRIVING-WHEELS. 375 The mean effective pressure under this cut-off is, at 188 revolu- tions, 29.6 pounds, and at 296 revolutions, 18.3 pounds. When the train speed is 55 miles, the draw-bar stress with the present 63-inch driving-wheels will be 18.3X109.4 = 2002 pounds. Whereas, if the driving-wheels were increased to 99 inches the draw- bar stress would be 29.6X69.4 = 2054 pounds, which is a gain in draw-bar stress in favor of the larger wheels. In this case, also, there would be an increase of economy resulting from the use of the larger wheels, the steam consumption under the con- ditions imposed by the small wheels being 30.6 pounds, and under those imposed by the larger but 27, a gain of 13 per cent. This analysis would seem to establish the fact, within limits that are pretty well defined, the draw-bar pull at speed is not reduced by increasing the diameter of the wheels, while the cylinder action is made more efficient. Since, in the case of a locomotive, anything which saves steam may at the limit be utilized in producing more power, the benefits in increased power to be derived from an increase in the diameter of driving-wheels is drawn from a twofold source. 186. A Recapitulation of the facts of the analysis which has been given is presented in Table LXXX. It is admitted that there are mechanical difficulties to be over- come before wheels of very large diameter can be used, also that conditions of service requ re some sacrifice of efficiency at speed to insure satisfactory performance in starting, and doubtless in many classes of service such considerations fully justify the use of the smaller wheels. The purpose is not to condemn practice, but to show what proportions are desirable for operation at speed. Finally, in this connection it is to be noted that the modern loco- motive has been given wheels which, while yet too small for highest efficiency at the rates of speed at which many locomotives are driven, are much larger than those which were used in the early ; 90's, when Schenectady No. 1 was built. These large-wheeled engines have proved economical in the use of water and coal. 376 LOCOMOTIVE PERFORMANCE. TABLE LXXX.* SHOWING CERTAIN RESULTS OBSERVED IN CONNECTION WITH LOCOMOTIVE SCHENECTADY, AND SIMILAR RESULTS DEDUCED FROM DATA GIVEN ON THE SUPPOSITION THAT THE DIAMETER OF ITS DRIVING-WHEELS HAD BEEN INCREASED IN THE RATIO OF 35 TO 55. SPEED CONSTANT AT 55 MILES AN HOUR. THROT- TLE FULLY OPEN. Present Drivers, 63-inch Diameter. Proposed Drivers, 99-inch Diameter. Revolutions per minute Approximate speed in miles per hour Indicated horse- power (Table I): 6-inch cut-off 8-inch cut-off Tractive force, pounds: 6-inch cut-off 8-inch cut-off Steam per indicated horse-power per hour (Table III): 6-inch cut-off 8-inch cut-off Coal per indicated horse-power per hour (Table IV): 6-inch cut-off 8-inch cut-off Gain or loss in indicated horse-power resulting from use of 99-inch drivers in place of 63-inch drivers for speed of 55 miles an hour: 6-inch cut-off 8-inch cut-off Decrease in steam consumption resulting from use of 99- inch drivers in place of 63-inch drivers for speed of 55 miles an hour: 6-inch cut-off 8-inch cut-off Decrease in coal consumption resulting from the use of 99-inch drivers in place of 63-inch drivers for speed of 55 miles an hour: 6-inch cut-off 8-inch cut-off. . 296 55 292 438 1995 2987 188 55 298 431 2054 2943 30.6 32.0 5.12 6.03 26.9 26.28 4.18 4.54 Gain 2.9 per cent Loss 1.4 per cent 12 per cent 18 per cent 18 per cent 23 per cent * This table and the arguments based thereon were first presented to the Western Railway Club, May, 1896. CHAPTER XXIII. ATMOSPHERIC RESISTANCE TO THE MOTION OF RAILWAY TRAINS. 187. Atmospheric Resistance. The resistance which must be overcome by a moving train arises from several causes, as, for ex- ample, from the rolling friction of wheel on rail, the effect of gradients and curvatures in the track, the necessity of producing accelerations in the speed, the friction of journals, and the resistance of the atmos- phere. The work which must be done to overcome the effect of grades and to produce accelerations in speed can be accurately determined, and the value of journal and rolling friction, when considered apart from complicating conditions, is already somewhat definitely known, but the available evidence concerning atmospheric resistance is con- tradictory and the result of its application uncertain. This fact makes of interest certain experiments, which were conducted at the Engineering Laboratory of Purdue University during the school year 1895-96 by Professor H. C. Solberg, then a graduate student in the laboratory.* The conditions under which experiments were made were assumed to be similar to those surrounding a train moving through still air, and the object of the experiments has been to disclose the value of forces resulting from the resistance offered by a quiescent atmosphere * An account of these experiments was first published as a paper before the West- ern Railway Club, April, 1896. The principal references given were: "A Study of Atmospheric Resistance to the Motion of Railway Trains." A thesis by H. C. Solberg, B.S., M.E., '96. "A Study of Air-currents in a Rectangular Conduit." A thesis by Augustus C. Spiker, B.S.,'96. "An Investigation of the Air-currents about a Moving Car or Train of Cars."' A thesis by Norman E. Gee, B.S., '96. 377 378 LOCOMOTIVE PERFORMANCE. to the forward movement of trains through it. No attempt was made to consider the effects resulting from oblique or other winds. 188. The Plan of the Experiments involved a rectangular con- duit, within which a current of air having any desired velocity could be maintained. Within this conduit, and exposed to the action of the air-currents, small dummy or model cars were mounted. Each model was connected by means of a sensitive dynamometer, with a suitable base so arranged as to indicate the value of any force tending to displace it in the direction of its length. A single model, or any number of models placed in order, as in a train, could be employed in any given experiment, the effect of the wind upon each car being always shown by the indication of its attached dynamometer. It is evident that, as a matter of principle, it is not material whether the model is at rest and the air is moved past it, or the air still and the model moved through it; that is, if the velocity of movement is the same in each case the value of the reaction between the wind and model will be the same. While the effect of the actual conditions surrounding the apparatus employed will be carefully reviewed in another paragraph (see paragraph 203), it may for the present be assumed that effects observed under the conditions of the experi- ments are the same as would have been observed had the model cars been caused to move through still air. Before proceeding the details of the apparatus employed should be briefly examined. 189. Conduit. The conduit in which the flow of air was main- tained for the experiments is in the form of a rectangular tube 20X20 inches in section and 60 feet in length. A cross-section is shown by Fig. 216. The lower face is of solid wood; the upper, also of wood, is pierced at intervals of six feet by good-sized openings, through which one may reach into the interior. These openings are closed by tight-fitting covers. The side faces of the conduit consist of large panels of glass set in wooden frames. The glass sides expose to view the whole interior of the conduit, so that both the position of the model cars and the reading of their dynamometers can readily be seen by the observer on the outside. The conduit is practically air-tight, the joints between glass and wood being covered with glued strips of paper. The interior surfaces also are unbroken from end to end, and, where of wood, are made so smooth by shellac as to offer but slight resistance to the passage of air through the tube. 190. Air-supply. The conduit is connected at one end with a No. 60 Sturtevant blower, the opposite end being open to the labora- ATMOSPHERIC RESISTANCE TO MOTION OF TRAINS. 379 tory. The whole apparatus being in one room, the duty of the blower is simply that of circulating the air of the room through the tube, forcing it in at one end, and allowing it to discharge at the other. The blower is of sufficient power to produce air-currents in the con- duit having a velocity of 100 miles an hour. Man-Hole Cover Ol CROSS SECTION OF BOX FIG. 216. FIG. 217. 191. The Determination of the Velocity of the Air-currents. The velocity of the moving air within the conduit was determined by use of instruments in the form of Pitot's tubes.* These were * A simple form of Pilot tube is shown by Fig. 217. It consists of two small tubes, having ends inserted into the flowing stream, the velocity of which it is desired to measure. The end of one tube is shaped to face the flow of the stream, while that of the other is normal to the flow. The exposed ends of the tubes connect with a U-shaped glass tube, partially filled with water or other liquid. It will be seen that as both sides of the U tube are in connection with the flowing stream, the difference in the height of liquid columns in the U tube cannot be due to the pressure of the flowing stream, but must result from the motion of the stream. The relation between the displacement of the gauge and the velocity of the flowing stream is expressed by the equation when v is the velocity of the stream in feet per second, and h the difference in height of the gauge columns in feet, measured in terms of a substance having the same density with the one whose velocity is to be determined. Many experimenters have from time to time testified as to the accuracy of this method of measuring velocities. It has long been used by physicists and meteor- ologists, and Professor W. S. Robinson, who has recently employed it extensively in determining the flow of natural gas in pipes, states that he was able to check his results with those obtained from meters with a satisfactory degree of certainty. In the present experiments water was used in the U tubes, and the relation 380 LOCOMOTIVE PERFORMANCE. made up of two brass tubes, arranged within a larger tube or jacket, all being cemented together by resin, which filled the interior of the jacket around the smaller tubes. The interior diameter of the small tubes was a sixteenth of an inch, and the diameter of the jacket-tube somewhat less than a half-inch, while the length of the combination was such as made it possible to reach from the ex- !^X s~t terior to 'any portion of the interior of the conduit. This portion of the apparatus is shown by Fig. 218. When in use the tip end, a, of the gauge was in- serted into the current through holes bored in the top planking, a cork bushing lining the hole, and making tight the joint between the wood and the gauge. Each of the two small brass tubes making up a gauge was then connected by rubber tubing with one side of a glass U tube fixed to a suitable scale outside of the conduit. The U tubes were sealed with water, from the displacement of which . B the velocity of the air passing the tips of the gauge was determined. Five such gauges with all their connections are shown in place in the conduit by Fig. 219. , .. , The several gauges employed were subjected to a careful examination, involving a series of simul- taneous observations in connection with a systematic interchange of position, to determine whether all could be depended upon to give like indications when the conditions were the same. Another preliminary to the main investigation w r as that of deter- mining the relative velocity of the stream of air at different points in the cross-section of the conduit. This was done by dividing the ,-, r IG. between the density of water and air is such as to make a column of water one inch high the equivalent of a column of air 68.37 feet high. The equation therefore becomes v 2 = (2 X 32.2 X 68.37 )h=U03h, where v is the velocity in feet per second and h is the head in inches of water. If, therefore, the U tube of a gauge used in the experiment showed a displacement of an inch and a half, the velocity of the air passing the tips of gauge in feet per second was assumed to be v=\/4403X 1.5= 81.2. ATMOSPHERIC RESISTANCE TO MOTION OF TRAINS. 381 cross-section into twenty-five or more imaginary squares, and by ob- serving the velocities at the center of several of them at the same in- stant, after which some of the gauges were changed to other squares and the process repeated, the observations for each set of readings overlapping those of the preceding set as a check on the constancy of conditions. A number of typical diagrams resulting from this process are presented as Fig. 220. They show velocity of the cur- FIG. 219. rent in miles per hour for different portions of the cross-section of the conduit. That there might be no uncertainty, also, as to the character of the flowing current of air, the cross-section of the stream was care- fully examined at many points throughout the length of the conduit, and as a result the following conclusions were reached: 1. That while considerable unevenness of flow was observed near the initial end of the conduit the eddies disappeared at a distance of 35 feet from the initial end, and from this point to a point near the 382 LOCOMOTIVE PERFORMANCE. discharge end of the conduit the flow was found to follow lines which were approximately straight. 2. That the glass surfaces forming the sides of the conduit offered less resistance to the movement of the air than the wooden surfaces forming the top and bottom. 3. That the lowest velocities were found, as would be expected, in the corners of the conduit, that is, where the sides joined with the top and bottom. Wood 1 21 23 24 23 21 25 29 29 29 25 25 29 29 29 25 25 28 29 29 25 21 23 25 23 21 Wood 47 53 53 53 46 54 61 61 61 54 55 61 61 61 55 55 61 61 61 54 46 52 53 52 47 Wood Wood Wood 80 95 95 96 81 96 100 100 100 96 96 100 100 100 96 96 100 100 100 96 81 95 95 95 81 Wood FIG. 220. Relative Velocities in Different Portions of the Conduit. 4. That there was a comparatively large vein in the interior of the stream, all portions of which flowed with practically the same velocity. The experiments which are to be described made use of that por- tion of the stream which was most free from eddies, and which was least influenced by the walls of the conduit". 192. The Model Cars. Having obtained means for making a ATMOSPHERIC RESISTANCE TO MOTION OF TRAINS. 383 breeze of satisfactory quality and for determining its velocity, the next and last step concerned the model cars which were to be exposed to its influence. To facilitate the description these model cars will hereafter be referred to as models. These were & the size of an assumed standard box car, the body of the model extending down- ward and occupying the space which in an actual car intervenes SECTION ON A-3 SECTION ON C-D FIG. 221. between the sills and the rails. Each model was 12^r inches long, 3J inches wide, and 4 inches high. Its form may be more perfectly apprehended by reference to the drawing (Fig. 221). The painted tin body of the model was fitted over a wooden base supported by four leg-pieces of light, hard-rolled sheet brass S, which in turn were securely fastened to a suitable foundation. 384 LOCOMOTIVE PERFORMANCE. The length and lightness of these legs, or springs, allowed the car to be displaced longitudinally, under the action of the slightest force, and they were at the same time so proportioned as to resist all ten- dency to motion in other directions. Between the body of the car and its foundation, also, and entirely independent of the springs already referred to, was a system of levers, the purpose of which was to multiply any longitudinal displacement to which the model might be subject. These levers were made of thin metal, the several parts being soldered to each other. All motion, consequently, was within the elastic limit of the parts affected. There were no loose joints. The whole arrangement proved to be both sensitive and reliable. The least pressure upon the car would result in a movement of the pointer, and the pointer would promptly return to its zero when the force producing the displacement had ceased to act. Excessive vibrations of the pointer were prevented by a vertical fin which could be made to dip into light oil contained in a suitable pan beneath. That no part of the dynamometer might be directly affected by the currents of air acting upon the model, the mechanism was en- tirely enclosed in the foundation, a portion of the surface of which was of glass through which the movement of the pointer could be observed. The degree of refinement attending the action of these dynamom- eter cars will be appreciated when it is said that, while the actual movement of the car was always slight, the leverage was such that an inch and a quarter movement of the pointer was readily obtained. The springs for a number of cars were made so flexible as to give an inch movement of the pointer under the force of one ounce acting upon the end of the model. Two models, however, to serve at the ends of trains, were provided with much stiffer springs. The foundations were longer than the models, so that when arranged in trains the proper spacing of models would be observed, while the foundation would present an unbroken surface. Sections of blank foundation, also, were supplied in front and rear of train. A draw- ing showing a train of three cars arranged in the conduit is shown by Fig. 222, and a front view of a train taken from within the conduit is given as Fig. 223. 193. Observations. With the desired number of models arranged as a train, and with a single Pitot tube located at A (Fig. 222), which alone was used in determining velocities in the conduit, the experi- ments proceeded about as follows : The blower engine was started and ATMOSPHERIC RESISTANCE TO MOTION OF TRAINS. 385 allowed to run at a slow speed for a sufficient time to secure con- stancy of conditions with the conduit, after which readings were taken simultaneously from the gauge A and the dynamometers of the several cars composing the train. These observations were repeated at intervals of thirty seconds until five readings had been taken, when the averages of the five sucessive readings were brought forward to a condensed log of observation. When one set of readings had been taken, the speed of the blower was increased, and all observations made for the new conditions. In this manner the work was advanced with each length of train, the velocities of the air-currents varying from 20 miles per hour to something over 100 miles per hour. No effort was made to obtain definite conditions of air-velocity, the Car Car Car Cat Car Arrangement of Pitots Tubes and Cars in Ho JB -<^_ Car Car Car Car Car * THE IRON TRADE REVIEW FIG. 222. Side View of Train in Conduit. object being to have a constant flow, and to observe accurately what were the precise values by which the conditions were defined. The work extended through several weeks, and the care taken throughout its progress was such that the data are remarkably con- sistent. Tables LXXXI. to LXXXVL, which are given herewith, are somewhat changed in form, and are much condensed from the very elaborate presentation by Professor Solberg. 194. Observed and Calculated Results. The significance of the seven different headings appearing in the several tables accompany- ing may be explained as follows : I. Gauge Displacement, Inches of Water. Values in this column are results obtained from direct readings of the U tube gauges con- nected with the Pitot tubes. They represent pressures measured in inches of water, due to the velocity of the air-currents within the conduit. 386 LOCOMOTIVE PERFORMANCE. II. Pressure Equivalent of Gauge Displacement in Pounds per Square Foot. These values are calculated directly from those of the preceding column, on the assumption that a cubic foot of water weighs 62.3 pounds. Thus, a column of water an inch high is the equivalent of 1X62.3 12 = 5.19 pounds per square foot, FIG. 223. so that the values of Col. II. are equal to those given in Col. I. tiplied by 5.19. The values given in this column will be found useful for purposes of comparison, since wind pressures are usually meas- ured in pounds per square foot. III. Pressure by Gauge, Multiplied by the Area of the Cross-section of the Model in Square Feet. In general, this would be expected to give the force with which the wind would act upon the end of a model. The cross-section of each model was equal to 14 square inches, or ATMOSPHERIC RESISTANCE TO MOTION OF TRAINS. 387 TABLE LXXXI. ONE MODEL. Number of Test. Gauge Displace- ment, Inches of Water. gag <5 a Ija . ifil w*s *sP !" iPls l&! ?tual Force Tending to Displace Model, as Shown by Model Dynamometer in Pounds. il 2 ~ ^.fe5 $^6 loB^ s'srl o MO 13 ^.SSAH> Velocity Calculated from Reading of Gauge. Feet per Second. Miles per Hour. 3 (5 i. II. III. IV. V. VI. VII. 1 0.3 1.6 .15 .094 .61 36 25 2 0.8 4.2 .40 .219 .54 60 41 3 2.6 13.5 1.31 .625 .48 107 73 4 3.8 19.7 1.91 .906 .47 130 88 5 5.0 26.0 2.52 1.250 .49 149 102 TABLE LXXXII. TWO MODELS fa If! llr? Velocity Cal- . 3 > a CS O 3 & s 03 (H . -H * ^ o ^ r- O Q Jj ^. O GO 3JJJ3 K II ' ' jjj-3 O lO O O >^ [2^ t^ CO CO 71 -M -GO ^ ""I "~I ^ "~1 "S-s g" C3 O O O O pf is ||| II 5 i a |J s| j "T^ d ^^ O CO GO .a S |i *^"o O Gi CO O OC g| * -- . - ;. * * 1 1-1 -v~ g* 1 * t^* CO OS O TJ CO ^H O O Ol CO -I ^ SQS fe|| jlJ'oJ O U5 O * O Jll GTS CM (N o * JM go ^ co o oc i lj a.nmbg m ppoj^ JO UOI^OaS-SSOJQ JO |_3 CO i-H O ^H C-^ -aoi: idsiQ aSn^o jo co oo co rn c:' qua] ^Ambg aanssajj ^ co 7J ^ M jo saqou! o o (M o uiaoBjdsia aS-n^o ^ N ^ W ^n N ..,, ATMOSPHERIC RESISTANCE TO MOTION OF TRAINS. 389 ift s w H H > ] H i-H O5 Oi OS 1O O CO 05 O I 1 W i>1 02 b* K H ^ ^ I S 2 2 ft S2 .si fa o M CO OS i 1 fH 10 CO 01 ^ 53 10 CO ^H f-H ^ Gk .2f> - E| . fi o 5 2 Fourth Model. O5 O O GO OS Ji si ^ H 4! S ^ 3 3 o o o o o 1 "QI i Second Model. lO O O i G<> g lla tf^ 3 g S5 S; t g 3 % ,a i| s| t- rt< O G<1 CO O5 ^ CO 00 CO CO t^ i 1 TjH CO S^ |.2 11 5"S s| T# CD 10 t- Oi 10 i i ir Oi O i CN Fourth Model. O-l ^ CO C^ O (3 >> -SQ 1 QO !> O O G^ 1 ;f f 1 2 | 1 H o ^ cs 10 co f ( CO IO 1~- CO o o o o o "08 * tiCQ U Tfl Ci O CO co CO CO IO CO CO CM ^ t- O5 O ^ j; aj^nbg ui ppoj^ JO UOl^09S-SSOJQ JO Bajy Aq paiidiijnj^ 83nBQ c( aanssaaj h ^ K d H H CO CO CO (N (M CO i 1 O5 TtH IT- -H i-H (N C w M cc ^t tt t^t Ig I OOOOOOOCC< I ^JSJ^S 1 I OOOO< I oooooooooo: puooag 1 oooopoooooppop i -*^>OO5ICOO5COOCOJOOO^^ ISJTJ amssajj ?o jo aonssaaj - N (N M CC ATMOSPHERIC RESISTANCE TO MOTION OF TRAINS. 391 *3- s 8Ay -AuaAix IIV j GO CO lO CO O? -A^'uaXL | "-J puooag i Cl 1T3 * I I 1 Oi i 1 O5 ; I T^H "^ T^ -^ co co TJ s.s .= S ii ^ WB r^* GO T^iooo OOOO puooag O O OOO-H jo uo^oas-ssoj jo paiidDpift A"q aanssaaj 8O _ i _i cl 83 oo spunoj '^ ^r) jo aanssajj CO -H CM B J W t3 W a ?Oi O^OiiOCM -H I 1 l-H (N CO S Or-ioosiooo^ocsoo ^H i 1 CO "* CO O -H O, O WE, CO 00 O^COOOCOOS J. o5 2> I EH h | ^ OS iO II i^e I 2 ?2 i3 S !'"(2 I ^ M O iO I CO |g I CO OS 1> a i N 8 (M OO CO i (lOlOCOt^- 410 LOCOMOTIVE PERFORMANCE. e 8 J--J o o o T* 05 CO OS O5 i < CO * CO I i I 5 (M co 1-1 "* ^ co ce II o o T( r- CO O i-H (N oo co co r- t i i i I CO (N l> O 05 O GO i I I GO I o (M O C5 IO i i ^ O5 CO CO OO <~H C^l CO 33 S-< II O CO 1-4 co co cq CO U5 00 p- 1 "<* CO in I 2 JIM i s CHAPTER XXIV. A GENERALIZATION CONCERNING LOCOMOTIVE PERFORMANCE. 212. Application of Data. From the data derived from Sehe- nectady No. 1 it is possible to make estimates concerning the per- formance of locomotives in general, and to construct equations expressing performance with reference to several important functions. 213. Boiler Performance. Dealing first with the matter of evap- orative power, the following facts are to be noted : The records obtained from the Purdue locomotive disclose several tests for which the evap- oration is above 12 pounds of water per foot of heating-surface per hour, and the maximum record for the boiler is 14.6. The results were obtained with Brazil block coal, which is of a light and rather friable character. While they were secured at the expense of very hard firing, it is probable that, with a superior grade of fuel, higher rates of evaporation could have been had, and that a possible maxi- mum for this boiler need not be lower than 15 pounds. In accepting this limit it should be considered a possible maximum merely. It is at least three pounds higher than the practical maximum which e greatly increased beyond the normal maximum, but this is done at the sacrifice of fire condition or of water level. Power obtained l>y such means is the result of abnormal development, and is not a subject for consideration in connection with the present discussion. From these considerations it is proposed to accept 12 pounds of water per foot of heating-surface per hour, as a fair measure of high performance for the Purdue locomotive under the ordinary conditions of the road, and as a close approach to the maximum evaporative power of all locomotives. With this understanding the value, 12 pounds of water per foot of heating-surface, will be accepted as a measure, by use of which the power of any locomotive may be predicted. It is 411 412 LOCOMOTIVE PERFORMANCE, certainly one which can be accepted for all boilers of similar design. Probably, also, it will apply to all ordinary boilers in locomotive service, though this is not settled beyond doubt. It has been suggested that boilers having large grates may be easily forced to higher limits than those given, while boilers having a large extent of heating-sur- face and small grates may have difficulty in working up to the limit of 12 pounds. This argument is not without force, though probably it has less significance than would at first appear. Grate areas have an important influence on the efficiency of a boiler, but it is not clear that they can operate greatly to increase the power. If, in each of two boilers, one having a wide grate and the other a narrow grate, the same amount of heat is liberated, the heat in passing the tubes should in each case produce the same evaporation. On the other hand, the advantage of the large grate as a power producer appears in its ability to withstand forcing, and it must be admitted that an effort to secure maximum power in a locomotive is likely to resolve itself into a fuel-burning contest. In such a contest the larger grate is likely to have some advantage, but, for reasons stated, differences on this account will not be great. On the other hand, very large locomotives, if hand-fired by a single man, can be made to work up to this limit of 12 pounds of water per foot of heating-surface per hour only with great diffi- culty and usually for intervals of time which are brief. In this case the limit upon the output of power is found not in the propor- tions of the locomotive, but in the strength and endurance of the fireman. The capacity of the locomotive is present, and may be utilized whenever means are found for feeding it. This inherent capacity is the factor for which a value is sought. Finally, it should be said that the measure proposed cannot be a precise one. It is based on actual evaporation, which is necessarily influenced by changes in steam pressure and in temperature of feed, but it is evident that the requirements of the present process are not such as to make it necessary to take these into account. From these considerations it will appear that the measure pro- posed is not likely .to pass unchallenged; but, as one studies the problem, he will gain confidence in its value. For the present pur- pose, therefore, it will be assumed that the performance of all boilers may be predicted from the known extent of their heating-surface; that the measure proposed is not a measure of maximum perform- ance, but is a close approximation thereto; and that the measure GENERALIZATION. 413 is a fairly representative maximum for ordinary conditions of service. The measure expressed in the form of an equation is Water evaporated per hour = 12 (feet of heating-surface in boiler), or E = 12H. For example, a modern passenger locomotive having 3500 feet of heating-surface should, when working continuously at its maximum power, deliver per hour # = 12X3500 = 42,000 pounds of steam. 214. Cylinder Performance. Having now a measure of the weight of steam which a locomotive boiler may be depended upon to deliver, we may next make inquiry concerning the degree of economy attending the consumption of steam by the cylinders. The steam consumption per horse-power for the Purdue engine, for all speeds and cut-offs within its range of action, under a full throttle, has already been presented (Fig. 66, Chapter V). Assuming that this engine was designed to work at speeds varying from 25 to 55 miles an hour, and at loads necessitating a cut-off of from 6 to 10 inches, there are required from 26.28 to 32 pounds of steam per horse- power per hour. The maximum limit of 32 pounds is exceptional. A single value which fairly represents the actual performance of this 17X 24 engine, working under 130 pounds steam pressure, is 28 pounds, and this value will for the present be accepted as representing the normal performance of a modern engine under its usual range of action. It is not the minimum value, for this is from 2 to 3 pounds lower; nor is it the maximum, for the maximum may greatly exceed the measure stated. But it is, as has been stated, a value which represents the average consumption under such range of action as is normal to ordinary service. We are now prepared to express cylinder power in terms of heat- ing-surface. Thus, if 28 pounds of steam are required each hour for the development of a horse-power, and if each foot of heating-surface yields 12 pounds of steam per hour, then as many feet of heating- surface will be required for one horse-power as 28 will contain 12, or 2J. That is, , . , , total pounds of water evaporated per hour Cylinder horse-power = pounds consumed per horse-power per hour _ 12 X square feet of heating-surface in boiler ~~w = 0.43 (square feet of heating-surface). 414 LOCOMOTIVE PERFORMANCE. Or, calling the cylinder power C.H.P., and the square feet of heating- surface in the boiler H } we have C.H.P. =0.43# (42) For example, the maximum cylinder power under continuous conditions of operation of a modern passenger locomotive having 3500 feet of heating-surface is C.H.P. = 0.43 X 3500 = 1500 (approximately) . As this equation is one of great significance, it should be noted that the process leading up to its development has presented several steps, some of which have been subject to qualification. It is evident that the accuracy of the concluding statement cannot be greater than that of the parts of which it is composed. It does not apply With strict accuracy to any particular engine, or to any specific con- dition of running. It is merely an approximate measure of the maxi- mum power which will be developed by a fairly representative modern engine under ordinary conditions of service. Its use is justifiable because of certain interesting and useful comparisons which it permits. 215. Draw-bar Pull. When the horse-power and speed of a locomotive are known its draw-bar stress may be easily calculated, assuming no loss in transmission from the cylinder to the draw-bar. Thus, for any locomotive, Work developed in the cylinders = the work given out at the draw- bar. Therefore, Cylinder horse-power X 33,000X60 = pounds stress in* draw-bar X dis- tance passed over in feet in one hour = stress in draw-bar X speed of train in miles per hourX 5280. Transposing, horse-power X 33, 000 X 60 Pounds draw-bar stress = speed of train in miles per hour X 5280' This is a general expression, true of any locomotive. Now, the horse-power of the typical engine has been shown to be C.H.P. = GENERALIZA TION. 415 0.43 X feet of heating-surface, and substituting this value for horse- power in the preceding equation. 0.43X 33, 000 X 60 X feet of heating-surface rounds draw-bar stress = rrr r- : : n -, 528UX speed oi tram in miles per hour cylinder horse-power speed in miles per hour = 375^. (43) This equation is assumed to be of general application. Applying it to the modern passenger locomotive employed in the previous illustration, having 3500 feet of heating-surface and developing 1500 horse-power, it appears that for a speed of 10 miles an hour such a locomotive will exert at the draw-bar a pull of 56,250 pounds, and at a speed of 50 miles an hour but = 11,250 pounds. o(J The draw-bar pull for all speeds for such a locomotive is given by Fig. 226. This diagram, applying only to an engine of the dimensions stated, represents its maximum tractive effort, assuming that all the energy of the cylinders appears as a stress on the draw-bar. So far as the mathematical relations involved are concerned, the curve may be extended in either direction indefinitely, but practical con- siderations establish limits. The limit of its extent upward is found when the tractive force equals the adhesion of the locomotive, and the limit of its extent to the right is reached when the speed of revo- lution reaches a fixed maximum. Assuming the weight on drivers to be 96,000 pounds, and the adhesion to be one-fourth this amount, the maximum pull which the drivers will transmit is 23,750 pounds; that is, on the line AB, Fig. 226. The maximum speed is somewhat more difficult to fix. Assuming that the locomotive has 79-inch drivers, and that they may be allowed to travel as many miles per hour as they are inches in diameter, the limit of speed will obviously be 79 miles, cutting off the diagram along the line CD, Fig. 226. 416 LOCOMOTIVE PERFORMANCE. Within limits thus defined the curve is a perfect definition of the theoretical maximum pulling-power of the engine under consideration. It shows that the engine may be started from rest at its maximum tractive force of 23,750 pounds, and may continue to exert this force until a speed of 24 miles an hour is reached. At this speed (point B) the power, which at lower speeds has been less than that which the Speed in Miles per Hour. FIG. 226. Draw-bar Pull as Affected by Speed. cylinders are capable of exerting, becomes maximum at 1500 horse. Beyond this point each increment of speed is attended by a loss in tractive force, until at the maximum speed of 79 miles it becomes reduced to 7000 pounds. It is worthy of especial note that the reduc- tion in draw-bar stress with increased speed is not due to any reduc- tion in the power of the engine, but is in response to a physical con- dition which must always prevail; namely, that, when the available power is constant, the force in action must diminish as the velocity increases.* * The proportions chosen for purpose of illustration in this and the preceding examples are those of a New York Central Atlantic type passenger locomotive. GENERALIZA TION. 417 216. Losses between Cylinder and Draw-bar. Thus far in the discussion it has been assumed that no loss occurs in transmitting the power of the cylinder to the draw-bar. As a matter of fact, losses occur which are due to the following causes : I. Machine friction of the engine, including the rolling resistance of drivers. II. Rolling resistance of trucks and tender. III. Resistance which the atmosphere offers to the head of the train. There may be other resistances as, for example, the flange friction due to side-winds, which may result in loss of power, but these are exceptional and are constantly varying in value. For this reason they have no place in a consideration of the general case with which we are now concerned. A general equation for machine friction may be written either in terms of mean effective pressure or of draw-bar stress. In anticipa- tion of such an equation we may write Z) = Diameter of drivers, feet; d= " " cylinders, inches; L = Length of stroke, feet; N = Number of strokes in one revolution; P = Mean effective pressure; t = Draw-bar stress. Remembering, also, that with no loss in transmission, The foot-pounds of work 1 f The foot-pounds of work given out at the draw-bar [ = -j developed in the cylin- in one revolution J I der in one revolution, it is evident that '*!? 4 or which will give the draw-bar stress whenever the mean effective pressure and the necessary dimensions of the locomotive are known. If, therefore, the value of P be made that which is just sufficient to keep the engine in motion against its own friction, its corresponding 418 LOCOMOTIVE PERFORMANCE. value will be draw-bar stress equivalent to the machine friction. From conclusions elsewhere presented (Chapter XIX) it appears that for the Purdue locomotive Schenectady No. 1, the frictional resistance may be represented by a mean effective pressure of 3.8 pounds, and that this value remains fairly constant for all speeds and cut-offs. Whether the higher pressures of the more modern engine operate to increase this value cannot definitely be determined, but the indica- tion is that the increase, if any, would be slight. In the absence of more definite information, we can do no better than to accept the value stated as of general application. If, therefore, the value 3.8 be sub- stituted for P in the preceding equation, and if it is remembered that for simple locomotives N equals 4, the frictional resistance in pounds stress at the draw-bar may be written (44) which as an approximate measure may be accepted as true of any simple locomotive. For the Purdue locomotive the machine friction is tl = 3 ' 8 TIT = 418 P unds > while for a modern Atlantic type passenger-engine, previously employed as an illustration, having cylinders 21" by 26" and drivers 79", 441 V 2 1 fi ti = 3.8 ;V - 548 pounds. o.b This force must be exerted at the draw-bar to keep in motion the pistons, cross-heads, valves, rods, and all connected machine parts, including the axles and flanges of coupled wheels. The sig- nificance of this statement is shown by the line CD (Fig. 227). It is independent of speed. The resistance to rolling, offered by trucks and tender, cannot be much different from that of an equal weight of train following the tender. This is often expressed by the formula f = 2 + 1/67, ........ (45) in which t is the resistance of each ton weight, and V is the velocity in miles per hour. This equation may be accepted as of general GENERALIZATION. 419 application. To apply it the value of the rolling load must be known. The rolling load for the modern passenger-engine, hitherto employed as an illustration, exclusive of that carried by drivers, the resistance of which has been included with the machine friction, is as follows: Weight on truck 42,000 " " trailing wheel 48,300 ' ' " tender in working order 110,900 Total.. . 201,200 2^000 *2 1-,000 NX 550 riction Speed (S) in Miles per Hour FIG. 227. Losses between Cylinders and Draw-bar. The formula shows that the tractive force necessary to overcome this rolling load of practically 100 tons, at a speed of ten miles an hour, is 366 pounds, and, for a speed of 50 miles an hour, 1033 pounds, values for all speeds being shown by the line EF, Fig. 227. Experiments to determine the resistance offered by the atmos- phere to the head of a train (Chapter XXIII) show that the progress of the locomotive and tender is resisted by a force approximately 420 LOCOMOTIVE PERFORMANCE. ten times greater than that which acts upon an intermediate car of the same train, the formula being (46) where t% is the tractive force in pounds necessary to overcome the resistance offered by still air to the progress of a locomotive at the head of a train, and V the velocity of the train in miles per hour. The formula makes the resistance which the atmosphere offers to the motion of a locomotive, at a speed of 10 miles an hour, 11 pounds; and at a speed of 50 miles, 275 pounds, and for all speeds as shown by the curve GH, Fig. 227. We have now obtained a general measure of the draw-bar stress equivalent to the power developed in the engine cylinders, assuming no loss in transmission, and have applied the same to modern pas- senger locomotives (Fig. 226); also values for the several losses which in service occur between the cylinder and the draw-bar (Fig. 227). If, now, we apply these latter measures as corrections to the former, the results should be a measure of the forces actually developed at the draw-bar. The effect of this process, as applied to the modern locomotive, is most easily made apparent by a graphical process, in which the curves of Fig. 227 are combined with that of Fig. 226. The result is shown by Fig. 228. In this figure the curve AB is the same as that given in Fig. 226. It is the draw-bar pull as calculated from cylinder work, disregarding engine friction and all other inci- dental losses. Applying to this curve the correction for machine friction which is denned by the curve CD, Fig. 227, we shall need to measure down on each ordinate a distance from the curve AB of 550 pounds. The result is the line CD, Fig. 228. Similarly, to make a further correction for the resistance of the rolling load, values from the curve EF, Fig. 227, are laid off downward from the curve CD, resulting in the line EF, Fig. 228. And, finally, to correct for wind resistance, values from the curve GH, Fig. 227, are laid off downward from the curve EF, resulting in the curve GH, Fig. 228. In the completed diagram the area between the first curve AB and the second CD represents frictional losses; that between the second CD e,nd third EF the rolling resistance of trucks and tenders ; that between the third EF and the fourth GH atmospheric resist- ance; and that between the fourth GH and the horizontal axes the forces which are available at the tender coupler for useful work in drawing a train. It will be seen that, at a speed of 25 miles an GENERALIZA TION. 421 hour, the stress at the draw-bar is 21,500 pounds; while, when the speed is increased to 75 miles an hour, the draw-bar stress is reduced to 5000 pounds. Some insight as to the effect of proportions upon performance at speed may be had by comparing the percentage of the cylinder effort, which is lost when the speed is low, with that which is lost when the speed is high. Thus, at a speed of 25 miles an hour, the line UV 25;000 20:000 15;000 10,000 -5400 20 40 GO Y 80 I 100 I'JO Speed in Miles per-Hour. FIG. 228. A Characteristic Diagram of Locomotive Performance. is only 7.5 per cent of the line UX, while at a speed of 75 miles an hour the line ZW is 40 per cent of the line ZY. ' The relation of lost work to total work developed in the two cases is proportional to the length of the lines referred to. 217. The Application of Results to several Typical Locomotives. The values shown by Figs. 227 and 228 are applicable to the New York Central Atlantic type passenger-engine only, but similar rela- tionships may readily be developed which will be of general applica- tion. The general case can best be developed by means of the equa- tions of the several curves involved, which are given together in 422 LOCOMOTIVE PERFORMANCE. paragraph 218. The result of their application to four well-known engines is shown by Fig. 229. A study of the diagrams of Fig. 229 will disclose some of the advan- tages which attend the use of heavy and powerful engines. Thus, 10 20 30 40 50 60 70 80 90 100 FIG. 229. Characteristic Diagrams of Several Typical Locomotives. comparing the performance of the Purdue locomotive, a representative engine of the year 1895, with that of the modern New York Central Atlantic type, it appears that, at a speed of 60 miles an hour, the actual draw-bar stress of the former is reduced to below 2000 pounds, while that of the latter is not less than 7000 pounds. At 70 miles an hour the pull of the Purdue engine drops to a thousand pounds, GENERAL1ZA TION. 423 whereas that of the New York Central engine is five times this amount. These are results due to the greater power available, and to the fact that the losses between cylinder and draw-bar are relatively less for heavier engines than for lighter ones. Some of the effects of different diameters of drivers are to be seen by a comparison of the diagrams of the Pittsburgh, Bessemer & Lake Erie engine and the Lake Shore & Michigan Southern engine. The former, with its 54-inch drivers and its enormous adhesive weight, gives at slow speed a draw-bar pull of not less than 55,000 pounds, which is more than 50 per cent in excess of the maximum pull of the Lake Shore engine. The small-wheeled engine is able to develop the full power of its cylinders when a speed of 10 miles an hour is reached, whereas this result is not secured by the large-wheeled engine until a speed of 18 miles an hour is attained. Correcting for losses between the cylinder and draw-bar, it appears that, for a speed of 40 miles an hour, the tractive effort of the two engines becomes equal. Again, assuming the maximum speed in miles per hour to equal the- diameter of the drivers in inches, the Pittsburgh, Bessemer & Lake Erie engine reaches its limit at 54 miles, whereas the Lake Shore engine continues to run until 80 miles are reached. Actual difference in respect to maximum speed would be greater than that which is shown; for, while the Lake Shore engine, with its large wheels and proportionately lighter reciprocating parts, may readily attain a speed approaching 400 revolutions, such a speed would be impossible with the smaller wheels and proportionally heavier reciprocating parts of the Pittsburgh, Bessemer & Lake Erie engine. The basis of com- parison in this respect is, therefore, more favorable to the small- wheeled engine than it should be. If the maximum speed limit of the small-wheeled engine be fixed at 40 miles an hour, its entire range of action is represented by the area of the diagram to the left of the 40-mile line, which area is to be compared with the full area of the diagram of the large-wheeled engine. It will be seen that increasing the diameter of the drivers greatly increases the possible range of action, and, further, as it tends to diminish fnctional losses, it serves in sustaining the draw-bar pull as the speed is increased. 218. A Summary of Results, expressed in the form of general equations applicable to any locomotive, is as follows: Draw-bar stress based on cylinder performance, I.H.P.=C.H.P.=0.43#, . . , (42) 424 LOCOMOTIVE PERFORMANCE. ......... (43) 161^ ........ (47) Corrections to draw-bar stress to be applied to cover loss in trans- mission from cylinder to draw-bar are as follows : (44) (45) .. . (46) Draw-bar stress corrected for the several losses is as follows: Ti-f-f,- 1613.8^ ........ (48) (49) 2 . . . . (50) In these equations the characters employed signify as follows: I. H. P. = indicated horse-power; C.H. P. = cylinder horse-power, not necessarily obtained by means of an indicator, but estimated to be equivalent to power thus obtained; f/ = square feet of heating-surface in boiler; V = speed in miles per hour; PF = tons weight of rolling load, including that on trucks and tender, but excluding that on drivers; w=tons weight carried by drivers; d= diameter of piston in inches; L = length of stroke in feet; D = diameter of drivers in feet; f = tractive force in pounds equivalent to the power developed in the cvlinders, assuming no loss in transmission: GENERALIZATION. 425 t\= tractive force in pounds equivalent to the machine friction, including the rolling friction of drivers ; t 2 = tractive force in pounds equivalent to the resistance of the rolling load, including that of trucks and tender, but ex- cluding that of drivers; * 3 = tractive force in pounds equivalent to the resistance offered by still air to the movement of an engine at the head of a train, bein^ the same factor which is defined as A in sec- tion (6) of paragraph 211. TI= tractive force in pounds equivalent to cylinder power cor- rected for machine friction. The application of the equa- tion for t\ to the New York Central engine gives the line CD, Fig. 228; T 2 = tractive force in pounds equivalent to cylinder power cor- rected for machine friction and resistance of rolling load. The application of this equation for T 2 to the New York Central engine gives the line EF, Fig. 228; TS= tractive force in pounds equivalent to cylinder power cor- rected for all losses between cylinders and draw-bar, in- cluding machine friction, resistance of rolling load, and atmospheric resistance. The application of the equation for T 3 to the New York Central engine gives the line GH, Fig. 228; T 3 must never be greater than 1 / 4 w. By solving for 7" 3 , the net draw-bar stress for any locomotive, at any speed, may be found. The facts which are required to be known are, area of heating-surface, diameter and stroke of pistons, diameter of drivers, the weight of the rolling load, and the weight on drivers. No value for T 3 thus obtained should be considered practicable if greater than one-fourth the weight on drivers. f UH1VERSITY 1 '-, , ItvSS INDEX. A PAGE Actual evaporation 127 << " , table giving 128 ' ' trains, atmospheric resistance to 399 Air-supply for atmospheric resistance tests 378 Alden friction brakes, the 13 " " " , drawing of the 14 Allen link, drawing of 317 Ash-pan, refuse caught in 181 ' ' , table showing refuse caught in 182 Atmospheric resistance to the motion of railway trains 377 " " , plan of experiments concerning 378 11 " , conclusions concerning 406 Axles, the supporting 12 ' ' , drawing of the supporting 13 B Back pressure, chart showing 140 Balancing, the problem of 321 Blower, the Sturtevant 11 ' ' for waste gases, illustration of 12 Bio wing -thro ugh effect, the 303 Boiler, the 125 ' ' , dimensions of 52 " , performance of 115, 411 ' ' , effect of thick firing upon performance of 167 * ' , table showing performance of 86 " , power of 132 * ' , efficiency of 156 * ' , thermal efficiency of 137 " , drawing of '. 56, 189 * ' , chart showing pressure of ' 127 * ' , conclusions with reference to performance of 153 427 428 INDEX. PAGE Brakes, the Alden friction 13 ' ' , drawing of tne Alden friction 14 ' ' , work absorbe by 21 " , " performed by 21 Branch-pipe, drawing of 54 Buckeye engine, arrangement of pipes and indicators on 272 C Capacity, pressure vs 367 Carnegie Institution, tests under the patronage of the 371 Choice of a locomotive, conditions controlling 46 Clearance as affected by different lengths of indicator-pipes 276 1 ' " affecting events of stroke, inside 301 " " " indicator-cards, inside 301 tl il " steam consumption, increased 306 1 ' , diagram showing steam consumption and inside 30 1 ' , excessive 304 , range of 309 Coal, analysis of. 82 ' ' , composition of 132 " , kind of 82 ' ' and combustible 132 ' ' , consumption of 134, 141 ' ' per I.H.P. per hour, diagram showing 119 " " D.H.P. " " , " " 120 " " mile run, diagram showing 122 Coefficient of friction, definition of apparent 16 Combustible, coal and 132 , rate of 137, 138 Combustion and coal, table showing 134 ' ' , high rates of. 156 ' ' , losses due to incomplete 163 Computation, methods of 76 Conduit used in experiments concerning atmospheric resistance 378 ' ' employed in atmospheric resistance tests, view within the 386 Constants for locomotive 49 Counterbalance, weights of 327 ' ' , action of the 321 ' ' , conclusions concerning 331 Course followed by locomotive, drawing showing 4 Cover, drawing of the valve-box and 58 Coverings, tests of boiler. 193 ".'"..'*, efficiency of boiler 203 Critical speed, definition of 112 Cross-head, drawing of 60 Cut-off as affecting form of indicator-cards, speed and 103 " mean effective pressure, speed and 106 INDEX. 429 PAGE Cylinder and saddle, drawing of 57 ' ' condensation 112 1 ' diagrams, form of 276 ' ' and draw-bar, losses between 417 1 { performance 413 ' ' heads, drawing of 59 Cylinders, table showing thermal action within 115 D Data covering forty-four efficiency tests 71 " , selection of 124 ' ' , application of 411 ' ' , observed and calculated 160 Deflector-plate, drawing of netting and 55 Departure of Schenectady No. 1, photograpn showing 39 Derived relations 150 Dimensions of locomotive, summary of 50 " ''boiler 52 " " stacks, drawings showing 229 " " nozzles, drawing " 229 Distribution of draft 209 " " " , diagram showing 210 Draft 137 ' ' , table showing 138 " , chart " 139 " , distribution of 209 1 ' , diagram showing distribution of 210 " , table showing percentage of 211 ' ' , diagrams showing values of 237 1 ' , table giving changes in 239 ' ' , unavoidable loss in 251 Draw-bar, performance at the 118 11 , diagram showing pull at the 120 ' ' , difficulties encountered in measuring stresses at the 334 " pull 414 ' ' , losses between cylinder and 417 ' ' stress, equations for , 423 Drawings of locomotive 49 Driving-wheels, diameter of 373 Dry-pipe, drawing of 54 Dynamometer, the traction 18 ' ' , drawing of the locomotive 19 ' ' , illustration of the 22 " , the Emery 27 " horse-power, drawing showing 121 " car, drawing of model '...... 383 430 INDEX. E PAGE Eccentric strap, drawing of W 4 rod, drawing of 64 ' ' , drawing of 65 Efficiency of boiler, power and 148 of the boiler, thermal 137 as affected by quality of fuel i 13 " , diagram showing rate of evaporation and 144 Emery dynamometer, the 27 Engineering laboratories at Purdue, growth of 1 ' ' laboratory, 1891, illustration of the 5 ' ' , ground plan of 5 " prior to January, 1894, illustration of 24 " " , burning of 24 " " , plan of the reconstructed 38 Engine, experiments upon a stationary 270 { ' performance, table showing 90 Equivalent evaporation, quality of steam and 129 per hour, diagram showing 116 " " , table showing 130 Evaporation, actual 127 , diagram showing rate of 144 , equivalent 129 , table giving actual 128 , table giving equivalent 130 Evaporative efficiency, diagram showing 117 performance as affected by increased power 142 " " , table giving 14? Events of stroke 105 . " " " , table showing 106 " " " as affected by lead 283 " " " " " " inside clearance 301 Excess air, losses due to 163 Excessive clearance 304 Exhaust, character of 145 " fan, the 23 ' ' pipe and nozzle, drawing of 53, 230 " J3t, action of 212 " " , illustration of apparatus used in exploring 213 " nozzles 227 " ' ' , drawing showing dimensions of 229 ' ' nozzle as affecting stack, changes in 247 Experimental train, illustration of head of 187 ' ' boiler and its equipment, the 188 " , drawing of. 189 INDEX. 431 Feed-water log 71 Fire, condition of 167 Fire-box, drawing of 228 Fireman, influence of 171 First model of a train, forces to be resisted by 394 First plant, behavior of mounting mechanism of 20 ' ' " , work of the 24 Forces acting throughout the length of a model train, distribution of 395 Forces and velocity of current acting upon models of a train, relation of resulting 396 Freight-cars, atmospheric resistance offered to trains of 404 Frict^n-brakes, the .^ 7 , the Alden * 13 ' ' , drawing of the 14 ' ' , illustration of 16 Friction horse-power, diagram showing 121 ' ' tests, results of 347 Front end, the 209 " " , committee on .- 226 " , " , conditions at the 227 ' ' losses as affected by spark losses 159 H Heat radiated from a locomotive boiler 185 Heat radiated, amount of . 185 432 INDEX. PAGE Heating surface, definition of 65 " " , losses along 163 ' f value of sparks, the 178 " " " , diagram showing 179 High steam-pressures, thermal advantages of 364 Higher pressures, arguments concerning 365 Horse-power, indicated 107 ' ' , diagram showing dynamometer 121 " , " " friction 121 I Increased clearance as affecting steam consumption 306 Indicator record ,, . . . 74 " cards 354 " "as affected by changes in speed and cut-off, form of 103 " " , diagram showing effect of speed and cut-off on form of. ... 104 ' " , effect of lead upon 287 " "as affected by changes in lap 293 " " " " " inside clearance 301 Indicated horse-power 107 " ' ' , diagram showing 109 Indicator rigging, drawing of 268 Indicator work, conclusions with reference to 281 Indicators on Buckeye engine, arrangement of 272 Influence of the fireman 171 Inside clearance, effect of 298 " tl as affecting indicator-cards 301 " " " " events of stroke 301 " ' ' and steam consumption, diagram showing 308 Items appearing in tables, definition of 76 * Jet, form and character of exhaust 218 Jet as affected by changes in speed, the 219 " " " " "bridge,the 220 " " " " cut-off, the 225 " " <' " stack, the 224 " " " " bars over the tip, the 223 " " influenced by tips, the 222 '* formed by a steady blast of steam 221 ' ' , drawing showing 219 , " " form of 216 Joy gear, the 317 " '/ , drawing of the 318 INDEX. 433 Laboratory, 1891, illustration of the engineering 5 ' ' , 1894, drawing of the engineering 5 ' ' , ground plan of the engineering 5 ' ' prior to January, 1894, illustration of 24 ' ' , burning of engineering 24 ' ' , plan of the reconstructed engineering 38 Laboratories at Purdue, growth of engineering 1 Lap, definition of outside 291 "' '" steam. . . " 291 Last model of a train, atmospheric forces to be resisted by 395 Lead, definition of 282 ' ' , determination of 282 ' ' , tests involving different amounts of 283 " , effect of 282 ' ' as affecting indicator-cards 287 11 " " valve-travel and port-opening 285 ' ' and machine friction 290 ' ' , conclusions concerning 290 Links, drawing of 63 Link, " . " Allen 317 ", " "stationary 317 Link-block, drawing of 63 Locomotive, choice of 46 " , arrival of first 2 1 ' , course followed by first 4 " in the laboratory, illustration of the 6 ' ' , elevation of mounting of 8 " , plan of mounting of 10 ' ' dynamometer, drawing of the 19 " , sale of the 39 ' ' , specifications for the 49 ' ' , arrival of second 41 ' ' , constants for 49 " , dimensions of 50 ' ' , drawings of 49 " , balancing of 325 " , power of 227 " , reciprocating parts of a 322 " and tender, resistance offered to 4C3 " , application of deduced results to a typical 421 " kboratory 1894, drawing of the 37 " " " ; illustration of the 38 " operation. . 42 under conditions other than those of the track 42 " performance ' 99, 117 434 IXDEX. PAGE! Locomotive performance, generalization concerning 411 ' ' , results concerning 423 , illustration of the second 40 testing-plant, considerations leading to 1 " , establishment of 1 " " , growth of interest in 43, 44 " " after the fire, illustration of the 25 " " , the second 25 ' ' valve-gears 310 Losses along the heating-surface 163 ' ' due to incomplete combustion 162 " by radiation. . . _ 185 M Machine friction 333 " , lead and 290 " ' ' , conclusions concerning 349 Maximum power dependent upon efficiency 119 Mean effective pressure as affected by inside clearance 305 " speed and cut-off 106 1 ' , diagram showing 108 Miller, extract from letter of Robert 41 Model dynamometer car, drawing of 383 ' * cars employed in tests to determine atmospheric resistance 382 " of a train, forces acting upon the first 394 " " " " , forces acting upon the second 395 " " " " , forces acting upon the last 395 " train, distribution of forces acting throughout the length of a 395 Models between the second and last car of a train, forces acting upon 395 " , forces acting upon trains of three, five, ten, and twenty-five 394 Mounting, plan of locomotive 6 ' ' , elevation of locomotive 8 ' ' mechanism, behavior of the 20 N Netting and deflector-plate, drawing of 55 New plant, work with the 34 ' ' testing-plants 45 ' ' wheel foundation, the 25 Nozzles, experimental 228 O Observers employed in testing 69 Oil circulation, drawing of the arrangement for 15 Outside lap, definition of 291 " ' ' as affecting indicator-cards 293 INDEX. 435 Passenger-cars, atmospheric resistance offered to trains of. 404 Pipe, differe.it lengths of indicator 273 Piston, drawing of 60 Piston-rod, drawing of 60 Pitot tube, drawing of 379 Port-opening, effect of lead upon 285 ' ' , determination of 286 Power and efficiency of boiler. . . . ., 148 " of boiler 132 " variation 296 1 ' developed by boiler, table showing 133 Pressure vs. capacity 367 Pressure as affected by inside clearance, mean effective 305 " as affected by speed and cut-off, mean effective 106 tl on bra es, photograph showing mechanism for controlling 22 Pressures, tests at different 366 Pull at draw-bar, diagram showing 120 Purdue University, opening of 1 ' ' , interest in the work of 44 Q Quality of fuel as affecting efficiency 148 " " steam 129 " " " , table showing 130 R Radiation 204 ' ' loss, diagram showing effect of speed on 206 " losses 185 " upon the road 186 ' ' , coal required to maintain 205 ' ' , table showing power lost by 205 ' ' , conclusions concerning loss by 208 Reciprocating parts of a locomotive 322 Reconstructed laboratory, plan of the 38 Refuse caught in ash-pan 181 " , table showing 182 Results of thirty-five efficiency tests 168 " , interpretation of 169 Revolution counters, illustration showing 22 Reverse-lever, drawing of 66 ' ' -shaft, drawing of 65 Rocker, drawing of 63 436 INDEX. PAGE Running log 70 " tests 196 " " , table giving results of 197 . S Saddle, drawing of cylinder and 57 Second locomotive, illustration of 40 Second model of a train, forces acting upon the 395 Second testing-plant, the 25 * ' " , elevation of 26 . " , floor plan of 33 " " , interior view of the 35 " " , plan of the 28 " " , section of 32 Schenectady No. 1, specifications for .- . 47 " ", arrival of 2 " " , elevation of 46 ' ' " in a new role, illustration showing 39 " . " , illustration of 3, 40 ' ' " , illustration showing the departure of 39 " " , sale of 39 Setting of valves of Schenectady No. 1 52 " " " in connection with tests to define cylinder performance. . . 103 Smoke-box, superheating in the 262 ' ' , temperature of 137 ' ' , table giving temperature of 138 Sparks 173 Spark discharge, drawing showing density of 183 Sparks, heating value of 178 ' ' , diagram showing heating value of 179 " , losses of 173 ' ' passing out of stack, drawings showing 183 " , sample of 184 " . size of 183 ' ' , volume of 179 Spark losses as affecting grate losses 159 " ' ( , diagram showing 177 " , table showing 180 " trap, the 174 " , diagram of 174 Specifications for locomotive 47 Speed and cut-off as affecting form of indicator-cards 103 " " " mean effective pressure 106 Stack, problem of 225 " , plan of 175 ' ' , form of ,_ 226 INDEX. 437 Stack, diameter of 226 ' ' , height of 227 ' ' , relation of height to diameter of 243 ' ' , effect of different proportions of 237 1 ' , equations giving diameter of 248 ' ' as affected by changes in exhaust-nozzle 247 Stacks, experimental 228 . " , drawings showing dimensions of 229 ' ' , straight 244 ' ' , equations for straight 244 ' ' , tapered 246 ' ' , equations for tapered 247 ' ' , relative advantage of straight and tapered 254 Standing tests to determine radiation losses 194 ' ' " , table showing results of 195 Stationary engine, experiments involving different lengths of indicator-pipes upon a 270 ' ' link, drawing of 317 Steam, consumption of 110 ' ' , dryness of 131 " accounted for by indicator, percentage of 113 ' ' , quality of 129 ' ' , table showing quality of 73, 130 ' ' accounted for by indicator at cut-off, percentage of 114 ' ' consumption as affected by lead 289 " " " " " outside lap 296 ' ' increased clearance 306 " " " throttling 360 ' ' , table showing 290 , diagram showing inside clearance and 308 ' ' engine indicators 269 ' ' lap, definition of 291 ' ' passages, areas of 53 ' ' passage areas, diagram showing 68 ' ' port, maximum opening of 300 ' ' pressure, table showing drop in 107 Stephenson valve-gear 310 gear does, what the 311 Straight stacks 244 { ' , drawing showing proportions for. . 251 ' ' and tapered stacks, relative advantage of 254 Summary-sheet 75 Superheating in the smoke-box 262 Superstructure, the 31 Supporting axles, drawing of the 13 ' ' wheels, the 7 " " , movement of the 9 438 INDEX. T PAGE Tapered stacks. . 246 ' ' ' ' , drawing showing proportions for 250 " " , relative advantag3 of straight and tapered 254 Tender, resistance offere by atmosphere t locomotive and 403 Testing plant, consid: rations leading to design of 1 " ' ' -, elevation of the second 26 ' ' ' ' , establishment of locomotive 1 " " , the first 4 " , work of the first 24 " " immediately after the fire, illustration of 25 " " , the second . ' 25 " " , plan of the second 28 " " , sectiDn of the second 32 tl " , floor plan of second 33 ' ' plants, new 45 ' ' , observers employed in 69 " , method of 69 Tests, the 102 ' ' , conditions of the. 175 " , outline of 168 " to determine heat losses by radiation upon the road, plan of 187 " , movements during the radiation 191 1 ' of coverings 194 " , front-end 231 ' ' involving different amounts of lead 283 " , results of 168 Thermal units, performance in terms of 135 11 , table sho'wing 13G ' ' efficiency of the boiler 137 Thermometer-cup, drawing of 264 Three, five, ten, and twenty-five models in experiments upon atmospheric resistance 394 Throttle, drawing of 67 ' ' lever, drawing of 66 ' ' pipe, drawing of 67 Throttling as affecting steam consumption 360 , effect of 353 tests 353 Track, the 190 Train, illustration of head of experimental 187 ' ' , resistance offered to any 405 Two models in experiments upon atmospheric resistance 394 V Valve, acceleration of the 311 ' ' box and cover, drawing of & INDEX. 439 PAGE Valve diagrams ; . 292 " ellipse 311 * ' , drawing of 62 ' ' gear, a Stephenson 310 " " design as affected by wire-drawing 312 ' ' gears, adaptability of 319 " " , determination of lead for 282 ' ' " , improved 314 " " , locomotive 310 * ' " , conclusions concerning 320 4 ' motion diagram 312 * ' travel, device for measuring ,. 286 " " , effect of lead upon 285 ' ' rod, drawing of steam-chest 62 ' ' yoke, drawing of '. 62 Valves, proportions of 52 " , the setting of 103 ' ' , table showing setting of 52 Velocity, relation of force and 396 Volume of sparks 179 von Borries-Troske tests, the 225 W Walschaert gear, the 318 " " , drawing of the 318 Water, diagram showing evaporation of 142 Weight of sparks passing from the stack 176 Wheel diameters, practice concerning 373 ' ' foundation, the new 25 Wire-drawing 106 as affecting valve-gear design 312 Work absorbed by brakes! 21 " with the new plant ' 34 THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO SO CENTS ON THE FOURTH DAY AND TO $1.OO ON THE SEVENTH DAY OVERDUE. 26 1940 |l I LQTO BEC. OR. NOV 1. LD 21-95t-7,'37|