A HANDBOOK OF TESTING MATERIALS A HANDBOOK OF TESTING C. A. M. SMITH, M. Sc. (ENG.), A.M.I. MECH. E., A.M.I.E.E., LATE K.N., ASST. PROFESSOR, EAST LONDON COLLEGE (UNIVERSITY OF LONDON) ; AUTHOR OF " SUCTION GAS PLANTS." MATERIALS NEW YORK D. VAN NOSTRAND COMPANY 23, MURRAY, AND 27, WARREN STREETS 191 1 PREFACE All technical colleges, and many of the large engineering works, now possess a laboratory furnished with apparatus for the testing of materials. It is hoped that some portion of this book may be of service to engineers engaged in practice ; it is, however, written primarily for the college student, although a great engineer has said that one of the attractions of our work is that we are always students in engineering. If, at times* the reader thinks that the descriptions of apparatus, or of tests, are somewhat detailed, he should remember that, although he may be quite familiar with a machine or process, other readers, possibly, have not had the same opportunities. The chief object of the author has been to interest engineers in experimental work. The experience gained in four colleges, and a works testing department, has led to the belief that the great importance of experimental work is not sufficiently recognised. It is hoped that students may be stimulated to look up, both before and after they have made experiments, the description of the apparatus and tests men- tioned. The diagrams will perhaps be useful for sketching purposes. The methods of teaching engineering vary in the different centres. No attempt is made in this book 'to standardise such methods each instructor is the best judge of what suits the plant and the students. A list of experiments is given in Chapter XIL, as it may be useful to those who are commencing to organise a laboratory course. The illustrations have been made especially for the book. Attention has been given to the scheming, as well as the actual drawing and reproduction of diagrams, in the hope that the details of the apparatus described may be made clear. An effort has been made to avoid catalogue illus- trations. In some places especially in the chapter on 372 J SO vi PREFACE alternating stress machines the diagrams used in the Pro- ceedings of Societies and technical papers have been replaced by others drawn with the object of showing principles rather than details of construction. In the Appendices there are included discussions on certain researches, which have been included for the advanced student. Some of these researches are so recent that they have not been previously described in a text- book on materials. Further experiments may be suggested to the reader by a study of the methods used and results obtained. The number of instructors in college laboratories are frequently insufficient for their arduous duties. It is sug- gested that the book will assist the instructors and students by reason of the various experiments, methods, and data recorded. It is advisable to take a group of not more than ten students and explain the apparatus to be used to them. It is usually inconvenient to take notes during such demon- strations, and the contents of the book may therefore be useful to the student for reference purposes. During the conduct of the ordinary laboratory work the most suitable number to be engaged on one experiment is (usually) three. Suggestions are given in the Introduction concerning the record of such work. The author begs to thank numerous friends who have dis- cussed the contents of the book for their assistance. His assistant, Mr. V. C. Davies, B.Sc., has not only carefully read through the proofs, but made many useful suggestions con- cerning the contents of the book. Mr. E. J. Surman, B.Sc., has also helped to revise the proofs. It is therefore hoped that the book is free from misprints, etc., but the author would be glad to hear of any that are noticed by readers. The following have kindly granted permission for the extracts from their publications, viz. : editors of technical journals mentioned, British Engineering Standards Committee, Institu- tion of Civil Engineers, Institution of Mechanical Engineers, Physical Society, etc. C. A. M. S. LONDOX, June, 1910. CONTENTS CHAP. PACiE I. INTRODUCTION . 1 II. GENERAL PROPERTIES OF MATERIALS 7 III. MACHINES FOR TENSION, COMPRESSION, AND BENDING TESTS . 16 IV. STRAIN-MEASURING INSTRUMENTS 51 V. METHODS AND RESULTS OF TESTS ON MATERIALS ... 83 VI. TORSION TESTING 116 VII. IMPACT AND HARDNESS TESTS 137 VIII. SHEAR AND MISCELLANEOUS TESTS 155 IX. ALTERNATING STRESS TESTS . 168 X. THE TESTING OF CEMENTS, REINFORCED CONCRETE, AND STONES 189 XI. THE TESTING OF TIMBER. . . . . . . ,207. XII. EXPERIMENTS IN COLLEGE LABORATORIES . . . .213 APPENDIX I. STANDARD RESULTS OF TESTS ON THE STRENGTH OF MATERIALS 227 II. ADMIRALTY RULES FOR TESTING MATERIALS FOR MACHINERY 234 ,, III. RESEARCHES ON COMBINED STRESS . . . 236 IV. HEAT TREATMENT OF STEELS .... 260 BIBLIOGRAPHY 265 USEFUL CONSTANTS 270 INDEX . 277 LIST OF ILLUSTRATIONS PIG. PAGE Testing Machine for Testing up to 300 Tons in the University of Birmingham . ... . . Frontispiece 1. Curves showing Effect of varying Percentages of Carbon in Steel on the Tenacity, Hardness and Ductility of the Material 1 1 2. Diagrammatic Sketch of Vertical Type Testing Machine . . 18 3. Diagrammatic Sketch of Horizontal Type Testing Machine . 19 4. Wicksteed's Testing Machine . . . . .20 5. Sir Alexander Kennedy's Testing Machine . . . .22 6. Diagrammatic Outline showing Principle of Werder Machine . 24 7. Side Elevation of 300-ton Testing Machine in Birmingham University . . . . . . . ... 25 8. End View of 300-ton Birmingham Testing Machine , f . 26 9. Details of Poise Gear and Levers on Birmingham 300 -ton Machine . . . . . . , . . .27 10. Details of Main Earn on Birmingham 300-ton Machine . . 29 11. Amsler Testing Machine for Short Compression Specimens . 35 12. Riehle Testing Machine 37 13. General Arrangement for Girder Testing . . . : .. . 38 14. Apparatus for Testing Deflections of Beams . . . . 40 15. Change Speed Mechanism in Poise Gear of Riehle Testing Machine . . . . ... . 42 16. Details of Knife-Edges, Shackle, and Grip . . 45 17. Ball Seating for Tension Specimens with Screwed Ends . . 46 18. Ball Seating for Specimen in Compression . . 47 19. Method of Loading Compression Specimen . . . . 48 20. Unsuccessful Method of Loading Compression Specimen . 48 21. Prof. Lilley's Method of Testing Hollow Struts ... 48 22. Diagram showing Effect of Wear of Knife-Edge in Machines of the Wicksteed Type . . 49 23. Ewing's Extensometer ... . .52 24. Diagrammatic Outline of Latest Type Ewing's Extensometer . 53 25. Unwin's Extensometer ...... .55 26. Marten's Pointer Extensometer . . 56 27. Ashcroft's Extensometer ... .56 28. Kennedy's Extensometer . 57 29. Stromeyer's Rolling Pin Extensometer . . 58 30. Stromeyer's Rolling Pin Type Extensometer (as used on finished structures) 59 x LIST OF ILLUSTRATIONS FIG. PAGE 31. Cambridge Extensometer 60 32. Bauschinger's Mirror Extensometer ..... 61 33. Marten's Mirror Extensometer 62 34. Detail of Mirror Attachment, Marten's Extensometer . . 63 35. Stromeyer's Optical Extensometer ...... 64 36. Strips for Sphingometer 65 37. Section through Sphingometer Strip Holder .... 66 38. Strip Carrier and Specimen Grip for Sphingometer ... 67 39. Calibration Curve for Sphingometer Strip .... 69 40. The Sphingometer, fitted with Torsion and Tension Strips . 70 41. Un win's Stress-Strain Recorder 73 42. Wicksteed's Stress-Strain Recorder ...... 74 43. Hemiing's Portable Autographic Stress-Strain Recorder . . 76 44. Kennedy's Automatic Stress-Strain Recorder .... 78 45. Autographic Stress-Strain Recorder as attached to a Single Lever Machine ......... 79 46. Autographic Diagram taken with Apparatus shown in Fig. 45 80 47. Autographic Stress-Strain Recorder with Double Pencil Gear giving Two Scales for the Extension 81 48. Standard Tension Specimen Plate 83 49. Typical Tension Specimen Bar ... . .83 50. Screwed End for Tension or Compression Specimen ... 84 51. Ultimate Tension Test on Muntz Metal ..... 86 52. Tension Test on Aluminium . . . . . . .86 53. Distribution of Extension ....... 89 54. Typical Compression Specimen . . ... 90 55. Short Ductile Specimen in Compression . ... 90 56. Ball and Socket Joint ... ... 90 57. Real and Apparent Stress-Strain Curves in Compression (from single observations) ........ 92 58. Appearance of Ductile Compression. Specimen at Failure . 93 59. Shear Stress on Brittle Material in Compression . 93 60. Fracture of Ductile Material in Tension 94 61. Fracture of Cast-Iron in Compression 94 62. Resolution of Forces in a Specimen subjected to Tension . 95 63. Mild Steel Plates in Tension (Autographic Diagram) . .100 64. Curves of Real and Apparent Stress . . . . .101 65. Autographic Diagram of Mild Steel in Tension, showing Effect of removing temporary 102 66. Effect of Time and Low Heat Treatment on Mild Steel in Tension 103 67. Effect of Time and Boiling on Mild Steel Specimen in Com- pression . . . . . . . . . .104 68. Curves showing Effect of Heat Treatment on Bessemer Steel 105 69. Curves showing Mechanical Hysteresis ..... 106 70. Mild Steel in Tension 108 LIST OF ILLUSTEATIONS xi FIG. PAGE 71. Sphingometer Test on Muntz Metal . . . . .113 72. Cast -Iron under Torsion . . . . . . .116 73. Fracture of Cast-Iron Hollow Specimen in Torsion . . 117 74. Torsion Test on Mild Steel (Autographic Diagram) . 118 75. Stresses induced in a Bar subjected to Pure Torsion . . 119 76. Bailey Torsion Testing Machine . . . . . . 120 77. Hand-Torsion Testing Machine . . 123 78. Diagram of Torsion Meter . . . . . .124 79. Effect of Overstrain in Torsion (Mild Steel Specimen) . .125 80. Effect of Overstrain on Mild Steel in Torsion . ^ . . 126 81. Effect of Time and Boiling on Mild Steel .... 126 82. Effect of Overstrain and Boiling on Mild Steel in Torsion . ] 27 83. Effect of Overstrain and Boiling on Mild Steel in Torsion . 127 84. Pure Torsion Test on Aluminium showing Time Effect . .128 85. Torsion Tests on Aluminium . . . ...'.. . 129 86. Torsion Tests on Copper . . . . ... 129 87. Effect of Overstrain on Copper . . . ... . . 130 88.' Torsion Tests on Muntz Metal . . . . 130 89. Torsion Tests on Muntz Metal . . . . '.bi . 131 90. Eectangular Block under Shear Stress . j .132 91. Circular Specimen in Torsion . ,. ; . . ' . '.' . 132 92. Apparatus for Torsion Experiments on Wires . ; . > . 134 93. Apparatus for Testing Torsional Vibrations of Wires . .136 94. Impact Testing Machine '. . . .139 95. Impact Testing Machine, repeated Hammer Blows .. . 143 96. Unwhrs Apparatus for Hardness Tests . . - ... . 145 97. Brinell's Machine ..'.-. .148 98. Apparatus for Shearing Tests . . . ... .156 99. Apparatus for Punching Tests . . . f . . 158 100. Punching Test op Mild Steel . . . . ... 159 101. Apparatus used for Testing Steel Balls . . . . . 161 102. Steel Ball Tests .162 103. Small Beam Testing Machine for Cast-iron . . . .164 104. Work done in Breaking Specimen 166 105. Load-Extension Curves showing Effect of Annealing on Mild Steel 167 106. Wohler's Alternating Torsion Testing Machine . . .169 107. Mechanism of Wohler's Tension Alternating Stress Machine . 171 108. Wohler's Variable Bending Stress Machine . . . .172 109. Wohler's Machine for Eepeat Bending in opposite directions . 173 110. Diagram showing Nature and Stress-Cycle of the Wohler Test 175 111. Diagram showing Nature and Stress-Strain Lines in Arnold's Test 175 112. Diagram of J. H. Smith's Alternating Stress Machine . .179 xii LIST OF ILLUSTRATIONS PTG . PAGE 113. Diagram of J. H. Smith's Alternating Stress Machine . . 180 114. J. H. Smith's Alternating Stress Machine . . . .181 115. Curve for Eationalisation of Alternating Stress Experiments 183 116. Sankey's Hand Bending Machine . . . . .185 117. Autographic Diagram from Hand Bending Machine . . 186 118. Dimensions of Standard Briquette (British Standard Specifi- cation) .... . . 190 119. Setting Needle for Cement (British Engineering Standards Committee's suggestion) . . . . . . .191 120. Le Chatellier Soundness Testing Instrument (British Standard Specification) . . . . - . . . . . 192 121. Cement Testing Machine . ' . . . ' . . .193 122. Machine for Compression Tests of Stones and Cements . .194 123. Tests on Cement with Brinell's Ball Test .? . . . 200 124. Amsler-Laffon Beam Testing Machine . ... . .204 125. Section of Specimen under Compound Stress . J 236 126. W. Scoble's Method of deciding Yield Point . . 242 127. Distribution of Stress in Specimen subjected to Non- Axial Loading 244 128. Arrangement of Loading for Combined Compression and Torsion Experiments . . . . . . . . 245 129. General View of Combined Tension and Torsion Apparatus, showing the Specimen with Torsion Bars and Pulleys . . 246 130. Curve showing Method of Plotting Results of Compound Stress Experiments ........ 250 131. Diagram to show Method of Loading in a Combined Bending and Torsion Test on Dr. Coker's Apparatus .... 254 132. Professor Coker's Apparatus for Combined Bending and Torsion . ' . . . '. . . . . " . ., ' .... 256 133. Curve of Correction Factor for Professor Coker's Combined Bending and Torsion Machine . . . . . .257 134. Professor Coker's Torsiometer for Measuring Strains in a Combined Bending and Torsion Test 258 PLATE. I. Fracture of Specimen tested as Cast. Micrographs of Cast Steel before and after annealing. Magnified 28*5 dia- meters. Specimen tested when annealed (no fracture) To face p. 10 II. Fractures of Wrought-Iron and Mild Steel in Tension. Fractures of Cast-Iron in Compression. Fracture of Cast- iron Specimens in Pure Torsion . . . To face p. 87 III. Fracture of Various Materials in Tension . . To face p. 87 IV. Fracture of Cast-Iron Specimens in Combined Torsion and Bending To face p. 251 A HANDBOOK OF TESTING MATERIALS CHAPTEK I INTRODUCTION Theory and Practice. It is possible to learn a great deal of the science which underlies all engineering work by making tests and experiments. There is a continuous use, and adjust- ment to correct perspective, of results obtained by theory and practice. The work of testing materials is very largely of a practical nature. For commercial routine tests little is required beyond a knowledge of arithmetic and a skill in the manipulation of certain machines and measuring appliances. An engineering training is necessary if the correct deductions are to be made from the test. It must, however, be under- stood that commercial figures, although of great value to engineers, are not always the only results to be sought in making tests, and in commercial work it frequently happens that new tests and apparatus must be devised. The properties of the materials used in structures and machines are of great importance. A knowledge of these properties can be best obtained by conducting a series of experiments upon specimens or samples of the various materials. In making these tests, the difference between work done in an engineering and a physics laboratory will at once become apparent. Although there is a certain T.M. B 2 A HANDBOOK OF TESTING MATERIALS similarity between the methods employed, yet there will be noticed in the materials used and tested by engineers a variation in the properties of different samples of the same material. The object, therefore, of the experiments made upon steel, iron, copper, alloys, etc., is to find average values, rather than rigidly exact numerical results. The properties of materials used by engineers vary owing to many causes. If we take a pound of almost any gas and measure its temperature, pressure, and volume, we can estimate with accuracy the behaviour of the gas, if it is to be compressed, under certain conditions, to a new pressure and volume. If, however, we take a pound of steel we cannot forecast with any accuracy the behaviour of the material under load if we know nothing more about it than the fact that it is called steel. It is of some assistance if we know the chemical composition of the steel, but, even with such information, we cannot estimate with accuracy the load which the material will carry at the point of fracture. Experience has taught engineers that the only satisfactory method of estimating suitable loads in design works is to subject samples of the actual material to be used, to tests, which reproduce, as nearly as is possible, the conditions met with in actual practice. Thus, if a quantity of steel is made for a purpose in which the material is subjected to a direct pull, the most satisfactory way of deciding the maximum value of the pull to be allowed in practice is to test samples of the material under similar conditions, and obtain a record of the physical properties of these samples. Of course it is possible that all of the steel from which the samples have been selected will not behave in a similar fashion to the specimens tested. In general, however, we can obtain suffi- cient data to assist us to estimate with considerable certainty the safe loads to which the materials may be subjected. In building a bridge, roof or other structure, or machine, there are two possible errors which may be made owing to a lack of knowledge of the materials employed. Insufficient material may be used, in which case there will probably be a INTRODUCTION 3 collapse. Too much material may be used, in which case the structure or machine will cost more than is necessary. Although generally it is not so obvious, yet the fault of the engineer who uses more material than is necessary is as great as that of the engineer who uses too little. All safe structures and machines use more material than is theoretically neces- sary, because a certain factor of safety, arrived at from the result of experience, is employed. The safe load must not, however, be reckoned too high or too low. The most satis- factory method of estimating the safe load is to test one or more specimens of the material to be used. From a com- mercial point of view that is the purpose of the science of the testing of materials. There are, however, other reasons why this work should be undertaken. These are outlined below. The Testing of Materials Laboratory. Instruction in the properties of materials is given in the laboratory for the following definite objects : (1) To demonstrate the behaviour of various materials under stress. (2) To establish clear con- ceptions as to the meaning of fundamental terms of the engineer's vocabulary, such as, yield-point, ultimate strength, modulus of elasticity, and shear modulus. (3) To make the student familiar with the methods by which materials are tested to obtain such numerical results as will enable their properties to be recorded and compared with other materials. (4) To fix in the memory a few average results for the materials more commonly used in engineering work. (5) To lay the foundation for a certain habit of thought, invaluable to the engineer. This might be called cultivating the habit of testing laws and materials (wherever it is possible to do so) instead of accepting them from authorities. (6) To give the student practice in writing reports of work done by himself. (7) To undertake original investigations (for publication) calculated to advance the knowledge of the strength and elasticity of materials. The laboratory experiments, as a means of illustration, must go hand-in-hand with theoretical instruction. The latter will place before the student underlying principles. It will assist B 2 4 A HANDBOOK OF TESTING MATERIALS him to understand, in their correct perspective, the maze of experimental facts which he will obtain during his own personal investigations. It will aid in developing habits of clear and discerning thought. It frequently happens that students are eager and willing to conduct tests, but put off the working out of the results. No greater fallacy exists than that the test is completed and everything known when the actual experiment is finished. Numerical results should, whenever it is convenient to do so, be calculated in the laboratory. The slide-rule is sufficiently accurate. In many cases approximate results are sufficient. If that precaution is not taken it may happen that an after- noon's work will be wasted because of some fault of the material, apparatus, or observer, which would have been dis- covered immediately the first test was worked out. The amount of time on each experiment varies with the individual student. It is usually necessary to spend a good deal of time upon any subject of which it is desired to have a full knowledge. The student should not consider it irksome or unnecessary to repeat experiments. While a multiplicity of experiments may be very attractive especially when they are varied yet it is possible that labour which appears to be mere drudgery may be of the greatest educational value. There is a great deal of what people who possess no enthusiasm for their work call drudgery in the life of the engineer. " Staying power " is often the best characteristic, in the laboratory or in the works of a reliable man. The object of the laboratory is not to pump into the student's mind a large amount of information, but to cultivate in that mind the ability to reason, to plan and scheme out experiments. In order, however, to thoroughly understand underlying principles, routine work must be done. Originality will be useless unless combined with a knowledge of fundamental laws. Discipline is essential in the laboratory, as elsewhere, and the student must subordinate his own ideas until the opportunity presents itself for their presentation. INTRODUCTION o That laboratory instruction is essential is now no longer doubted. The equipment in all of the centres of higher education in this country is growing each year ; in some it might almost be called magnificent. The whole tendency of the work is to arouse interest. It appeals to the eye and the sense of touch in a way impossible to obtain in the class-room. In the University of London Engineering -Examinations (for internal students) marks are given for the laboratory work. There is a growing tendency in the same direction in the provinces. Each of the students, who takes the examination in the strength of materials, presents a book containing a report of the tests which he has made. It might be an advan- tage if the laboratory rough note-book were also inspected. In any case, the student's ability is judged from his record of laboratory instruction. It is advisable, therefore, for this to be neatly compiled. At the end of this book is a chapter dealing with the experiments which are usually undertaken for the purposes of these examinations. Those actually performed must of necessity depend upon the equipment of the particular laboratory in which the student is working and the time which he is able to give to this branch of study. Possibly in some laboratories opportunities will occur for the more advanced students to undertake special investiga- tions. It is suggested that some of the researches alluded to in this book may serve as a guide to that end. The Report of the Test. The writing up of the report of the experiment is by no means the least important portion of the work. Unfortunately, very few people express their thoughts in an intelligible fashion. There are men of great technical ability whose writings do not always contain clear expression of their thoughts. It is to be hoped that the twentieth century system of training for engineers will alter that in time. Clear expression should be the natural outcome of clear thinking. Every engineering student will be called upon, at some period in his career, to furnish a report upon a technical subject. It is probable that the report will be read 6 A HANDBOOK OF TESTING MATEEIALS by those who know little or nothing of technical work. The client may be, for instance, a solicitor, banker, financier, or a journalist. He will be influenced by the manner in which the report is written. Most of us who have to write upon technical subjects live to regret that we had but little training in the .art of composition in our student days. Lucidity is the mental lubricant which so many of us lack. In writing up the report of any test there is an opportunity for neatness and clearness of expression. It should be com- posed as if for a client who knows nothing whatever about the subject. There is a great temptation to occupy considerable space with a dissertation upon this subject of writing the report. It is with difficulty resisted. Let the student remember that a shop labourer can put a piece of material in the testing machine and break it. It requires a competent engineer to conduct, and report upon, a test such as is needed for commercial purposes. CHAPTER II GENERAL PROPERTIES OP MATERIALS Ductile and Brittle Materials. The materials used in engineering work are roughly divided into two classes, viz., ductile and brittle. It is difficult to say whether some materials should be considered as ductile or brittle, as there is no decided line of demarcation. The past history of the material will affect its properties. Ductility may be defined as the ease with which a metal can be elongated into a wire by being drawn through the gradually diminishing holes of the wire-drawer's plate. In general, for testing purposes, we call a material ductile when it stretches perceptibly before fracture during a tension test. A gauge of this property called ductility is supplied if we calculate the percentage elongation of the material, and also the percentage reduction in area of the fracture, after the test is completed. We can group together a number of metals which are ductile. Steel, the material most generally used in engineering work, varies considerably in composition and in mechanical proper- ties. There is the steel which can be used as tie-rods or crank- shafts, and there is the steel used for castings. The problem is more complicated each year by the discovery and introduc- tion of steel alloys. Again, steel and its alloys are extremely sensitive to mechanical working and temperature treatment. The physical properties may be completely changed by heating the material to a certain temperature and cooling slowly or suddenly. In general, we may say that the steel which has a low carbon constituent is ductile. The following metals are usually classified as ductile, viz., gold, silver, platinum, iron, 1 1 Certain types of iron (including steel in the general term iron). 8 A HANDBOOK OF TESTING MATERIALS nickel, copper, palladium, aluminium, zinc, tin, and lead. Metals which stretch imperceptibly before fracture during a tension test are said to be brittle. In general such metals are very strong in compression, weak in tension, and unsuitable to withstand shock. The properties of the new alloys (which are constantly being discovered) are in many cases very remarkable. It is advisable, when testing any material, to obtain some rough idea before commencing the test as to its mechanical properties, otherwise damage to the instruments or shackles may result. Chemical Analysis. This is a branch of the subject which the average college student has little or no time to investigate. The skill and experience necessary for an exact and reliable analysis of steel or cast-iron, which are by far the most important materials, can only be gained by long specialisation. The only part of the subject which it is necessary for the engineer to know is how to prepare a sample for submitting to the chemist. This should be done as follows : Preparing Samples. These should be in the form of fine turnings or drillings, but not dust. They must be perfectly free from all oil and other foreign substances. Take a bar specimen and commence drilling in one side with a flat angled drill about ^-inch diameter. When the top J inch has been removed, shake out the drillings and prepare to make a proper sample. With a ^-inch drill, bore into the specimen to a depth not exceeding \ inch. Repeat this in various parts of the bar, and collect together about 2 or 3 ounces of fine drillings. These should be at once placed in a clean test-tube or sample bottle, and handed on to the chemist who will carry out the analysis. The usual iron and steel analysis will give as percentages : carbon, silicon, sulphur, phosphorus and manganese. Structure of Materials. In general, the appearance of the fractured surface of a bar of metal is an index of its character. An easy way of obtaining some idea of the properties of the material is by nicking a bar on one side with a chisel, gripping GENERAL PROPERTIES OF MATERIALS 9 it in a vice, and breaking it with a hammer. The brittleness or toughness, as well as the general strength of the material, is indicated by the angle through which it bends and the force required to break it. The appearance of the fracture should be noticed in all tests, and it is usually advisable to record it. The bar is said to be crystalline when made up of visible crystals, either coarse or fine. When the crystals are so fine that they cannot be noticed as crystals by the eye, the fracture is said to be granular. Sometimes the grains are so minute as to be called silky as is the case with a tool steel properly hardened. Wrought-iron usually shows a fracture which has the appearance of the material breaking piecemeal, owing to the imperfect adhesion between the numerous microscopic slag-fibres and the iron and the great difference in the tough- ness of the two materials. The structure is called fibrous. Generally speaking, a coarsely crystalline or granular metal has less satisfactory working properties than one of the same class in which the fracture is finer. Incidentally, it may be noted that the size of the grain often largely depends upon the temperature at which the metal was cast, and upon the subsequent thermal and mechanical treatment. The Microstructure of Materials. Although the examina- tion of metals under a microscope is a matter for the metal- lurgist rather than the engineer, this branch of the testing of materials has become of such importance during recent years that it is essential that the engineer engaged in the testing of materials should at least be able to follow the methods employed in this branch of metallography, and be able to judge something of the properties of a material from micro- photographs prepared by an expert in that particular branch. Like large numbers of other members of the mineral kingdom, the metals have a crystalline structure, formed when the metal solidifies or is subjected to certain mechanical treatment. In the impure state of the ordinary materials of construction, this crystalline condition is changed to give a peculiar but characteristic appearance to the material, when examined under the microscope. This depends on the exact 10 A HANDBOOK OF TESTING MATERIALS nature of the impurities and on the previous mechanical and heat treatment to which it has been subjected. This appear- ance is termed the microstructure. The specimens to be examined are cut from the material to a convenient size suitable for holding in the fingers. The piece is then brought to a smooth surface by machining, grinding, or filling and smoothing up on several grades of emery cloth glued to a smooth block of hard wood. The finishing process is performed by polishing on a revolving polishing disc covered with chamois leather to which a small amount of jewellers' rouge has been applied. The specimen is then " etched " with a dilute acid solution or some other liquid which will attack the metal and preferably stain some parts different from others. When mounted on a microscope and illuminated with light directed normal to its surface the microstructure is clearly exhibited by a magni- fication of from 30 diameters upwards. When necessary for the purpose of keeping records photographs can be taken of the image by attaching a special form of camera to the eye- piece. The examination is most important in connection with iron and steel, as it not only gives a rough idea of its composition, but also gives to the expert a good idea of the heat treatment which the steel has undergone ; a point which it is obvious would be difficult, if not impossible, to determine by chemical analysis. Thus, in the case of iron containing 0*9 per cent, of carbon, if this steel is properly and carefully annealed a uniform product results known as pearlite, and when examined under the microscope is observed to consist of either uniform streaks of dark and light lines or similar granules, and derives its name from the play of colour on its surface, causing a slight resemblance to mother-of-pearl. When this is observed in the case of the particular steel mentioned it shows careful annealing. On the other hand, if this same sample was raised to a bright cherry red, and suddenly quenched in a freezing mixture of ice and salt, i.e., an exaggerated case of bad heat treatment or a hardening process, the structure observed Fracture of specimen tested as cast. Micrographs of Cast Steel before and after annealing. Magnified 28'5 diameters. Specimen tested when annealed (no fracture). PLATE I. GENERAL PEOPERTIES OF MATERIALS 11 is known as martensite, and takes the form of interlacing needles quite different from pearlite. It might be mentioned that in the particular case we have chosen the martensite is known sometimes as hardenite, and is observed as the principal constituent of hardened steel. Percentage of Carbon J4-0.000 110,000 CO c. Q. CO .O 80,000 g 0-5 1-0 1-5 20 ^ \ ^' ov-/. ^ x / \ ^ +* \ N \ / ,\' 60J. 407. 207. c ,^ 3 C Ferrite \ \ '$' *d *?' \ \ \ \ Zt /* t * 6 '' ! " \ ' ^' ' \ ^/ ^ ^ \ \ ^ ' / \ ($50000 / \ %, X X . " x< x>v -^ 20000 ^^ 5 10 15 20 25 30 / 100 95 90 85 80 75 70 "/ FIG. 1. Curves showing Effect of varying Percentages of Carbon in Steel on the Tenacity, Hardness and Ductility of the Material. The reason for these differences is largely due to the fact that (again referring to the particular case chosen) when cooled slowly the carbon and some of the iron combine to form what is considered a definite compound known by its appearance in the microscope as cementite, or chemically as iron carbide (Fe 8 C), and containing 6*67 per cent, of carbon and pure iron, known in microstructure phraseology as ferrite. On the other hand, when^cooled suddenly the steel 12 A HANDBOOK OF TESTING MATERIALS forms unsegregated pearlite, a combination of iron and carbon containing 0*9 per cent, of carbon which has crystallised without separating into carbide and ferrite. In this connec- tion Howe has given a very instructive diagram showing the relation between the varying percentages of cementite and ferrite, such as occur in annealed specimens and the physical properties of the material. It should be noted that unless the specimen is annealed, while, of course, the percentage of carbon will remain the same, the percentage of cementite will be less. Fig. 1, taken from Howe's work, shows the effect on tenacity, hardness and ductility, of different percentages of carbon. When the percentage of carbon is large (as in grey pig-iron) very slow cooling causes the carbon to separate into crystalline flakes, distinguished as graphite, and is observed in the microscope obtruding among the crystal grains of ferrite. The presence of sulphur, phosphorus, manganese, and silicon can likewise be recognised by the characteristic appearance which they give to the structure depending on the heat treat- ment. Marked changes are likewise produced in the non- ferrous metals, and annealed or hardened brass can be distinguished at a glance by the entire change of its crystalline nature. Plate L, prepared from drawings and photographs kindly loaned by Dr. J. Arnold, of Sheffield University, will illustrate in a striking fashion the remarkable change brought about in cast steel by annealing. The top half (as cast) illustrates martensitic structure, while the lower half (annealed) illus- trates pearlite structure. The change in mechanical properties will be readily seen from the drawings of a specimen, one bent before and the other after annealing. Mechanical Properties. The results of different tests made upon various materials show some remarkable contrasts. It is of advantage that the materials do not all possess the same properties, and it is obvious that while one material may be the most suitable for one purpose, it is quite unsuitable for another. GENERAL PROPERTIES OF MATERIALS 13 It is important to know, for any material, (a) the load at which it ceases to be elastic ; (b) the maximum load before or at the point of fracture. In commercial testing (a) is usually taken as that load at which a large amount of stretch takes place and the beam drops. This point is usually called the yield point. In scientific testing there is some difficulty con- cerning this matter of elastic breakdown, and there are in use three terms which require explanation. They are : (1) The Limit of Proportionality. This is usually taken as that point on the stress strain diagram where the increase in stretch is not exactly proportional to the increase in load. Obviously the exact point at which this takes place will depend upon the sensitiveness of the extensometer used to detect the stretch. (2) The Elastic Limit. If, when the weight is run back and there is no load on the specimen, the extensometer pointer returns to its original position at the starting of the test, there is said to be no permanent set upon the material. When such a permanent set is recorded we have reached the elastic limit. In many alloys there is such a gradual change from the elastic to the plastic state that it is essential to have some other criterion than the drop of the beam. In which case it is usual to record the load at which the permanent set is a certain definite amount, such as, say, T J-Q part of an inch. (3) The Yield Point. This is usually taken as the load at which the drop of the beam is first noted. Mild, or low carbon, steel usually shows all three points quite distinctly. At the critical point where elastic breakdown first takes place, however, the influence of time on the amount of stretch is so very marked that it seems probable that, under ideal conditions of loading and a perfectly homogeneous material, all three points would coincide. In the practical use of materials the three characteristics elastic limit (or yield point), breaking load, and extensibility- are of the first importance. It is not sufficient to know what weight a bar of metal will withstand without rupture ; it is of 14 A HANDBOOK OF TESTING MATERIALS the utmost importance to ascertain what load it will bear without sensible distortion. A metal which must be shaped under a hammer may with advantage have a low elastic limit, so long as its extensibility is sufficient. Such a metal is usually tough. Hard steel is very strong, but its extensibility is slight, and the hardest kinds are liable to be broken under a sudden blow. The Various Tests. Although it is most usual to determine the tensile strength of a material, it is being gradually recog- nised that other tests are also essential. If a material is to be subjected to a push stress, then a compression test is the most satisfactory. Similarly, if a material is to be used in a shaft, where it is subjected to a twist, a torsion test should be made. At present there is no law for all materials which connects the strength in tension with that of the material in compression and torsion. It is quite possible to imagine a bar of metal which almost resembles a bundle of straws. It would be per- fectly satisfactory if tested in tension, but fail utterly if subjected to torque or compression. It often happens that materials are subjected to alternating stresses of push and pull. In which case the tests which reveal the properties suitable for such work are the alternating stress tests described later. Unfortunately such tests usually take a considerable time to carry out, and are therefore not generally conducted for commercial purposes, although their importance is now fully recognised. Similarly, materials are in practice frequently subjected to combined loadings, such as torsion and thrust in a propeller shaft, and it is necessary to determine the loads which it is possible to safely carry. Tests made, however, under such conditions would be too expensive for commercial purposes, and it is left for researches to determine if static tests will suffice to allow us to estimate the loads to be carried. Again, it happens on railways that the steel lines are sub- jected to heavy blows or to suddenly applied loads. In tension, compression bending, or torsion tests (or static tests, as they are called) the load should be applied very gradually and GENERAL PROPERTIES OF MATERIALS 15 steadily (any irregularity in the time of loading, or even irregularity of the speed at which the load is applied, will affect the results ; hence the importance of adding equal increments of load during equal increments of time). But frequently materials (such as those used for rails or armour plate) are subjected to impact. A special form of testing machine is used for such tests. Therefore the reader will see that various special tests are described for certain materials, and while the actual com- mercial value of such tests may be a matter for discussion, there can be no two opinions as to their advantage for a complete experimental study of the properties of materials used in engineering work. CHAPTEK III MACHINES FOR TENSION, COMPRESSION, AND BENDING TESTS The Usual Machine. It is the usual practice to build a tensile testing machine so that it can be readily adapted to testing materials in compression or bending. Although this " omnibus " testing machine is of great advantage in a com- mercial laboratory, since it saves capital outlay, yet it is some- times awkward in a college laboratory, because only one group of students can be at work doing one of the tests. Although the usual size of testing machine in colleges is from 50 tons to 100 tons, it is possible that for most of the experimental work done by the student a 10-ton machine is large enough, in which case there may be sufficient capital available for two or even three machines. However, if the reader grasps the principles of the " omnibus " machine, he will understand those used only for one of the tests. The above tests are enumerated, but it will be seen later that the " omnibus " testing machine is also used for other special tests. Tensile Testing Machines. The easiest way of obtaining the tensile strength of any material would be to take a rod of that material, suspend it in a vertical position from one end, and hang weights on the other end until it breaks. The aggregate of the weights suspended at that moment from the free end would form the breaking load for that particular bar. The cross-sectional area of the bar, being known, the break- ing stress for that material could easily be deduced. In a similar manner the compressive or bending strength could be obtained. But this simple and crude method of obtaining the strength of materials could only be used where the bar tested had an extremely small cross-section, or else when the material itself was very weak. With ordinary materials, such as are MACHINES FOE TENSION, COMPKESSION, ETC. 17 used in engineering practice, this method could not be adopted, as, with even small sizes of test bar, the weights required would be so large, and the labour of handling them so great, as to render its use in ordinary workshops or labora- tories a practical impossibility. This difficulty, then, is over- come by the use of levers interposed between the point of action of the weight and the specimen whose strength we wish to determine. The interposition of this lever, or set of levers, makes the use of a smaller weight possible, and gives consequent facility in handling the machine, even when large powers are to be employed. We will now sketch out a very simple form of testing machine. The specimen of the material to be tested consists of a truly turned cylindrical bar, one end of which is gripped in a pair of jaws fixed to the framework of the machine, and the other attached to the short end of a lever. On the long arm of this lever a weight is allowed to slide, 1 so that its distance from the fulcrum can be varied at will. We com- mence with the weight near the fulcrum, so that little or no pull is exerted on the specimen. The weight is then run out slowly towards the end of the arm, the load being thus gradually applied until the specimen breaks. The breaking load is then equal to the weight on the long arm, multiplied by its leverage, or by the ratio between the long and short arms of the lever. But here we encounter another difficulty. If the specimen always remained of the same length, the lever would always remain of the same effective length, as it would in all cases be horizontal. But on the application of a load, the specimen stretches, more or less, according to the nature of the material that is being tested. The lever thus assumes an oblique position and tends to bend, as well as to stretch the material. It is necessary, then, to take up the stretch in the material so as to keep the lever in a horizontal position. This is usually done either by means of a hydraulic ram or 1 In some machines the weight is fixed in position but variable in amount. The load is applied by adding weights at the end of the long arm of the lever. The same result is obtained, but the other method is usually employed as the application of the load is thereby made more uniform and gradual. T.M. C 18 A HANDBOOK OF TESTING MATERIALS a screw. It requires no great effort of imagination, now that we have reached this point, to assume that the hydraulic ram or screw produces the pull on the test bar, while the move- ment of the weight along the lever gives us a method of measuring the load so produced. There are two principal types of machine used for testing Movable Jockey Weight FIG. 2. Diagrammatic Sketch of Vertical Type Testing Machine. materials in tension, compression, or bending. They are termed vertical or horizontal machines, according as the specimen or test bar is placed in a vertical or horizontal position. A diagrammatic sketch of a vertical testing machine is shown in Fig. 2. The specimen or test bar A is held in a vertical position by means of grips, a detailed description of which will be given later. The load is applied by means of a hydraulic ram C, which works in a cylinder B. Water or oil MACHINES FOE TENSION, COMPRESSION, ETC. under great pressure is forced into the cylinder B and causes the ram C to move downwards. This tendency, however, is re- pressed by the specimen A, the tenacity of which opposes the load. Owing to the friction, of the cup leathers, which are employed to prevent the water and oil from leaking past the rarn, the pressure of the oil or water in the cylinder does not give us a true measure of the load on the specimen. The upper end of A is accordingly fastened by means of similar grips to the shorter arm F of a lever whose fulcrum is fixed at D. On applying the load, the downward tendency of F is counterbalanced by running the movable jockey-weight E along the longer arm of the lever (usually called the "stillion"). The length of the " stillion " is usually so calibrated that from the position of the weight along this arm we can at once read off the load on the specimen. The horizontal type of tester is shown in Fig. 3. In principle it is precisely the same as the vertical machine, only the arrangement of the details being different. Here the specimen A is supported in a horizontal position and the load applied as c 2 20 A HANDBOOK OF TESTING MATERIALS MACHINES FOB TENSION, COMPRESSION, ETC. 21 before by a hydraulic cylinder B and ram C. The lever in this case, however, instead of being straight, is in the form of a " bell crank," as it is now desired to balance the horizontal pull on the specimen by the vertical action of the weight E. Wicksteed Testing Machine. We now pass on to consider some of the more important, and at the same time more complicated, machines used for testing the strength of the stronger materials (such as iron, steel, timber, etc.), in tension, compression, and bending. One machine can, as a rule, be adapted to test all these properties of a material, though, of course, the grips used and the kinds of specimen employed differ in each case. The principles upon which the Wicksteed machine works are indicated in Fig. 4. Water or oil is first admitted into the compressing cylinder. This consists of a hydraulic cylinder whose ram is moved backwards and forwards by means of a screw. It may thus be described as a screw-driven pump, in which the water attains a very high pressure (2,000 to 3,000 Ibs. per square inch). This water under pressure is then admitted through a valve into the hydraulic cylinder, which is usually secured in an inverted position to the base of the main frame of the machine, so that, to all intents and purposes, it forms an integral part of the latter. Here it acts downwards on the ram, the pressure being transmitted through the lower crosshead to the adjusting screws, and thence to the upper crosshead which carries the lower grip for the test piece, and usually slides on some part of the framework of the machine. The upper grip is con- nected by means of a shackle to the beam through a knife edge. The beam itself usually consists of two wrought-iron plates kept apart by brackets at intervals, though sometimes it is made of cast-iron, which gives greater rigidity. Between the plates two steel cylinders are fixed, having grooves into which the hardened steel knife edges fit. On one of these the beam rests, the knife edge in that case being downwards. On the other is hung the upper shackle which supports the test piece. The jockey weight is usually moved along the beam by 22 A HANDBOOK OF TESTING MATERIALS plate, and is means of a screw which traverses the whole length of the beam. The weight also carries a vernier which moves along a scale fixed to the beam, so that the position of the weight on the beam can be observed with great exacti- tude. The centre of gravity of the jockey weight is arranged to be as nearly as possible in an exact line joining the two knife edges, as if this were not so the effect would be that of a bent lever, and the leverage would alter slightly as the beam tilted up or down. The motion of the beam in this direction is, how- ever, limited by stops at the free end. The object of the ad- justing screws is to enable the distance between the grips to be altered to suit the length of the specimen to be tested. The machine is here shown fitted up for a tensile test. If, how- ever, a compression test is required, the lower grip is replaced by a square A similar plate is fixed at some distance below this connected to the same point on the beam as the upper MACHINES FOR TENSION, COMPRESSION, ETC. 23 shackle in the tensile test. Consequently, what was the lower shackle in the latter test is now the upper one, and vice versa. If now the cylindrical test piece is placed between these two plates, the direction of stress will be in the reverse direction, and consequently the material will be compressed. The com- pression is transmitted to the beam in the same way as before The Kennedy Machine (Fig. 5). This is a well-known testing machine of the horizontal type. The ram of the hydraulic cylinder A is connected to a cast-iron sliding frame B, which carries an adjustable crosshead C in which are fixed the grips for receiving one end of the test bar. The other crosshead D is fixed between two screwed bars E which traverse the whole length of the machine, one above and one below the hydraulic cylinder. If a tensile test is to be performed the specimen is gripped between these two cross- heads. To adjust the distance between them the key which fastens the movable crosshead C to the sliding frame is drawn out, the crosshead moved to the required distance, and the key then replaced. The frame has keyways cut in it at intervals of about 4 inches to allow this crosshead to be fixed in any position. For compression tests a flat plate is provided at that end (F) of the sliding frame which is nearest to the ram, and a similar plate on the fixed crosshead D, the specimen being gripped between them. The screwed rods which transmit the stresses to the levers are thus always in tension, the sliding frame being similarly invariably in compression. The tension is transmitted from the rods to a back crosshead G, which, through knife-edges, actuates two side levers H. The power being applied to these levers at a point above the fulcrum gives the effect of a bell -crank lever, so that the direction of the forces is changed from horizontal to vertical. This is invariably the case with horizontal testing machines, as we wish to balance the forces by means of a sliding weight, the direction of whose action is, of course, vertical. The outer ends of these side levers are connected by means of tension rods K to the short arm of the beam, the forces being measured by running out a jockey weight in the usual way. 24 A HANDBOOK OF TESTING MATEEIALS The following are the dimensions of a 50-ton Kennedy machine : Diameter of ram, 16'15 inches. Maximum jockey weight, J ton. Arms of bell-crank lever, 3 inches and 24 inches. Arms of beam, 8 inches and 100 inches. q Q -j .*. Total leverage of machine X - . . Capacity of machine 50 tons. Since the scale is 100 inches long, every 2 inches of travel represents an increment in loading of 1 ton. The sliding weight is, however, composed of cast-iron discs which are removable, so that for smaller powers greater accuracy can be obtained. The accumulator, which supplies the hydraulic ^Spirit Level Specimen /Hydraulic Cylinder Load FIG. 6. Diagrammatic Outline showing Principle of Werder Machine. cylinder, is capable of giving a pressure of 2,000 Ibs. per square inch. It is fed by three throw pumps, 1^ inches in diameter with 3-inch stroke. These are driven through gear- ing by a three-phase electric motor running at 880 revolutions per minute. These figures are given with the intention of conveying to the reader some idea of the actual size of the principal details and auxiliary apparatus of a machine of this capacity. The Werder Machine. Although not used, except in a modified form, in England, one of the most convenient and accurate machines is that devised by Werder. Fig. 6 indi- cates the general scheme of mechanism employed, the action of which is obvious. It will be seen that by this arrangement of levers it is possible to place the whole of the controlling gear at one end. Hence, specimens of any desired length can be tested merely by extending the frame. The usual length MACHINES FOR TENSION, COMPRESSION, ETC. for which these machines are built is 30 feet. Arrangements are made by which specimens can be tested up to this length in both tension and compression. Machines on this pattern have been built for the Government testing laboratories at Berlin and Munich, for the polytechnic schools at Zurich and Vienna, and for several manufactories and railways. It was on a machine of this type that most of the famous researches of Bauschinger were carried out. It should be clearly under- stood that the diagrammatic scheme illustrated in Fig. 6 is only intended to illustrate the general principle employed. For details of the methods employed for utilising this prin- ciple in actual practice, the student should refer to the original treatise (in the German language) given below. 1 300-ton Machine. The Civil Engineering Department of the University of Birmingham has the privilege of possessing a very large testing machine. It was desired to carry out work 1 Mittlieilungvn a. d. Mechanixch- techniischen Laboratorium in Miinchen, Heft 1 und 3. Mascliine zum, Priifen der Festigkeit der Materialien und Instruments zum Messen der Gestaltsrerdnderung der Proleltorper. Miinchen, 1882. 26 A HANDBOOK OF TESTING MATERIALS on actual structures. The smallest capacity of a machine for this class of work was fixed at 300 tons. There was some difficulty in preparing a specification for such a large size machine. Eventually Messrs. W. and T. Avery, Ltd., Soho Foundry, Birmingham, were commissioned to prepare and submit designs for such a machine, and a machine of 700,000 Ibs. capacity was finally constructed by them and installed in the laboratory. FIG. 8. End View of 300-ton Birmingham Testing Machine. The following details will give the reader an idea of the general design : Leading dimensions: Maximum length for tension, 28 feet; maximum length for compression, 30 feet ; span for bending, 20 feet. The over-all dimensions of the machine are : Total length, f>7 feet 9 inches; maximum height, 13 feet 3 inches; width varying from 7 feet to 21 feet. The wedge-grips can take specimens 3-f inches in diameter, or 6 inches by 2 inches flat. a 28 A HANDBOOK OF TESTING MATERIALS General arrangement : The machine is of the horizontal type, with the ram at one end and the lever system and operating platform at the other. Figs. 7 and 8, on pages 25 and 26, show the general arrangement, the former being a front elevation, and the latter an end elevation. In Fig. 7 the massive cast-iron bed X X of the machine is ter- minated at each end by a rigid vertical standard. The tops of the standards are connected by two horizontal columns Y Y, spanning the distance between them without intermediate support. Thus, for the whole of the length of the machine occupied by a specimen under test, there is a clear space of 2 feet 9 inches at the sides and a clear space of 3 feet 9 inches visible from the platform on the top. The latter is not only important because it enables the operator to watch the specimen, but also because it enables heavy specimens to be lowered into the machine by the overhead traveller. To one vertical standard is fixed the hydraulic cylinder, and to the other the lever system. Fig 9, on p. 27, is a diagrammatic view of the steelyard mechanism, and Fig. 10, on p. 29, of the main and return rams. Straining mechanism : The machine is operated by hydraulic pressure, two supplies of water being available : the town pressure, of about 100 Ibs. per square inch, for preliminary operations and adjustment ; and the accumulator supply, of 1,000 Ibs. per square inch. The ram L is 2 feet 8 inches in diameter, giving, with the low-pressure supply, a total thrust of 35 tons, and with the 1,000 Ibs. per square inch the full 700,000 Ibs. The stroke of the ram is 5 feet 6 inches. The main cylinder is bolted to the standard, and thus forms part of the frame of the machine, and the main ram is hollow and forms a cylinder moving over a fixed subsidiary ram M 1 foot 8 inches in diameter. The object of this arrangement is to provide for the return of the main ram after a test. Water from the low pressure supply is admitted behind the fixed rani, and the main ram is driven back into its cylinder. To the head of the main ram are fixed four racks N sliding in grooves in the frame of the machine, and notched throughout tuny ujrqdy LUOJJ. pun oj_ ui -bs J9cf -qi 59 iimyy UMOI ;j() A HANDBOOK OF TESTING MATEETALS their length to permit of the travelling crosshead No. 2 being keyed to them in any desired position, according to the length of the specimen under test. The travelling crosshead is moved by means of gearing, terminating in a handle at each side. Thus the load on the ram is transmitted to the sliding racks, thence to the crosshead 2 fixed to them, and from the cross- head through a spherical seating to the specimen. The other end of the specimen is held by one of the floating crossheads 1 and 3, according to the character of the test. Between crosshead 2 and crosshead 3 the specimen is in compression ; or if the bending beam is in position, the load on the beam is transmitted to crosshead 3. Between crosshead 2 and cross- head 1 the specimen is in tension. The floating crossheads 1 and 3 are suspended on knife-edges on the top of the frame, and are connected together by four tension rods P, the whole arrangement forming a floating frame. When crosshead 3 is in use i.e., in bending and compression tests these rods transmit the load to crosshead 1, and hence to the lever system. The connecting link between the floating frame and lever system is the main link B B, terminating in a steel bearing-block Q Q, which engages with the upper knife-edge of the first or bell-crank lever. Weighing or recording mechanism : This consists of a main bell-crank lever A A, with its principal knife-edge, 5 feet long, engaging with a plate fitted in a recess in the bed. The load is brought on to a similar knife-edge (forming with the main knife-edge the short vertical arm of the bell-crank) directly by the main link B B. The second lever C C, Fig. 8, and the final lever or steelyard D D run at right angles to the axis of the machine. The former has its fulcrum on a short column E fixed to the bed of the machine, and the load from the end of the long arm of the bell-crank lever is transmitted to it through side links G G and a cross-shackle F. The steel- yard has its fulcrum on a special column H on the right hand of the machine. The steelyard carries seven poises, each of which at the end of its run represents 100,000 Ibs. on the specimen. Thus, whatever the total load, 1 inch on the steel- MACHINES FOR TENSION, COMPRESSION, ETC. 31 yard scale always represents the same increment of load. At the same time, there are no loose weights to put on or take off. The poises can be made to slide over or back along the steel- yard by hand, or they may be put into gear with a screw, and moved by hand- wheels J. The machine is operated from a raised platform at the steelyard end ; the supply and exhaust pipes are brought to an operating valve on this platform, and the load can be taken on or off, the steelyard balanced and the load recorded by a single operator on this platform, who is also able to see the specimen under test. It would have been a great loss to the equipment had this unique machine been omitted. The great amount of research work possible with it is obvious to every engineer. The author is indebted to Professors Stephen Dixon and F. H. Hummel for the above particulars and drawings, and to the editor of Engineering for the photograph of the machine given in the frontispiece. Avery Testing Machine. The following specification of a 100-ton machine is supplied by this firm and describes its construction in great detail. A novel feature is the adoption of double acting in the hydraulic cylinder, the other points being very similar to those of the Kennedy machine. This is included as an example of how such a specification should be drawn up as well as with the object of familiarising the reader with the principles of design adopted by this firm. SPECIFICATION OF HORIZONTAL TESTING MACHINE. To test specimens in tension, compression, bending and shearing. Maximum testing capacity . . . . 100 tons. Longest specimen in tension 12 feet. Longest specimen in compression . . . .15 feet. Longest beam for transverse test . . 15 feet. This machine is designed specially for the testing up to 100 tons strain of full-size members to destruction. It consists generally of a hydraulic cylinder and ram having a stroke of 3 feet 6 inches, forming the straining portion, and a system of compound levers forming the weighing or recording portion. The hydraulic supply is derived from the town main supply. 32 A HANDBOOK OF TESTING MATERIALS The straining portion of the machine consists of a heavy section double- acting hydraulic cylinder of special design. The main cylinder, which is a heavy casting, has a large hollow ram. which slides over a smaller stationary ram. The working pressure may vary from 850 Ibs. to 1,120 Ibs. per square inch and the diameter of the larger ram is of such dimensions that the tester will give 100 tons strain with the minimum pressure. The cylinder is accurately bored, and the rams are both turned. The cylinder is provided with the usual hydraulic leathers, secured in position by a turned and faced cast-iron cover bolted to the cylinder flange. The main ram is fitted into a massive cast-steel head, into which the turned and screwed ends of the straining racks are secured by means of turned and screwed nuts. These straining racks are four in number, and are of mild steel and run the whole length of the bed. They are for the greater part of their length rectangular in section, terminating in round ends where they pass through the ram head. Machined slots are pro- vided at intervals of 9 inches to receive the main crosshead. Four nuts are screwed on to the straining racks where they pass through the cylinder flange, and these allow of the ram being locked at any position of its stroke, allowing specimens to be placed under a strain for an extended period. The main ram is copper-plated to prevent corrosion. The hydraulic cylinder is turned outside and fitted in a heavy casting, which is bored to receive it. This casting stands upon the ground, and to it are bolted the four cast-iron columns in which the straining racks slide. These columns are planed between the castings supporting the hydraulic cylinder and the casting against which the pull upon the main lever fulcrum is received. They each have a machined groove or slot to act as guides to the straining racks, and their ends are all faced. A substantial cast-steel crosshead runs upon wheels, and is arranged to be capable of being placed in any position in the straining racks ; machined slots are provided at the top and bottom of the crosshead into which loose keys are inserted to connect the same to the straining racks. This crosshead is made hollow to receive on the one side the special holders for the tension test and on the other side the adjustable plattens for testing columns in compression. The holders for the tension test are inserted at the back of the cross- head and have spherical seatings to give true alignment to the specimens. These holders will allow of any of the recognised forms of specimen being tested. Headed specimens are secured in split collars, while other types of specimens are secured in hardened steel wedge grips. The plattens for testing columns are also provided with spherical seatings. The weighing or indicating portion consists of a very heavily -constructed main lever, the sides of which are mild steel plates. Hardened steel knife-edges are fitted into these plates, and these backed by means of machined mild steel blocks acting as distance pieces. The knife-edges MACHINES FOE TENSION, COMPEESSION, ETC. 33 are of the best quality steel hardened, and are arranged to be 1 inch in length for every 5 tons of strain. The bearings are all of hardened steel, the fulcrum bearing being fitted into a cast-iron pedestal on the main casting. The weight of the lever is supported by means of wrought-iron links, with hardened steel bearings, from a cast-iron cross-beam resting upon short columns which are bolted down to the cast-iron base plate. The main knife-edges of the levers bear in a hardened steel bearing- block contained in a projecting verge which is bolted to a massive cross- head secured to the main tension rods. Two other crossheads are attached to these rods, and the whole floating frame of three crossheads and two tension-rods are suspended from cast-iron pedestals upon the main frame by means of wrought-iron links having hardened steel bearings. The steelyard is a massive iron casting, fitted with hardened steel knife- edges arranged to be 1 inch in length for every 5 tons of strain. Its fulcrum knife-edge rests upon a hardened steel bearing fitted into a cast- iron pedestal bolted down to the main -column. The steelyard is traversed by three poises, which are connected or disconnected at will to a central screw. When tests below 25 tons are required the first poise is used indepen- dently of the other two and gives readings from zero up to 2o tons, the 1-ton mark being 8 inches apart and the sub-divisions being TfiW of a ton. For tests between 25 and 50 tons the first and second poises are coupled together, and these now give readings from zero up to 50 tons, the 1-ton marks being 4 inches apart, the sub-divisions being TTRH7 f a ton. For tests between and 50 and 100 tons all three poises are coupled together, and readings from zero up to 100 tons are obtained, the 1-ton marks being 2 inches apart, and the sub-divisions in this case being J^TJ- f a ton. All the finer sub -divisions are obtained by means of verniers upon the poises. The poises are propelled by a hand- wheel at the front of the tester, this being connected by means of gearing to the central screw. This arrangement of poises secures that neither of the poises need to be removed from the steelyard. A carrier and buffer spring are provided to lessen the shock upon the steelyard when the specimen breaks. The steelyard is connected to the main lever by means of wrought- iron links having hardened steel bearings. Tension test. Seven pairs of hardened steel wedge grips are supplied with the machine to enable flat and round specimens to be tested in tension. The maximum length of specimen in tension is 12 feet, and this specimen can be tested without the removal of the cross-beam. Compression test. Columns up to 15 feet in length and of a maximum section of 12 inches diameter can be tested without removing the cross- beam used in the bending test. 34 A HANDBOOK OF TESTING MATERIALS Bending or transverse test. This test is made by means of a heavy steel cross-girder, consisting of rolled steel joists having steel plates riveted to their flanges. A cast-iron grooved track is bolted to the plates of the cross-girder, and this is traversed by two pedestals, each containing a hardened steel hemi-cylindrical bearing block. These pedestals can be quickly adjusted to any desired span to a maximum of 15 feet, by the movement of hand-wheels at the ends of the cross-girder. A presser foot which has two hemi-cylindrical bearing blocks is fitted into the movable crosshead in place of the platten for the compression test, and this bears centrally on the beam being tested. The cross-beam is arranged to be easily removed. To facilitate this a carriage, running free upon wheels and having side rollers, is provided. To centralise the beam, projecting pieces upon it are allowed to clip at each side of a projecting boss upon the end crosshead. Shearing test. This apparatus consists of two cast-iron portions which are guided and slide in one another. One portion is bolted to the casting at the front of the steel cross-beam, and to this the specimen is clamped by means of cross-plates and set pins. The other casting envelops the specimen, against which it is forced by means of the straining crosshead of the machine. Both portions of the shearing apparatus are provided with hardened steel tools to give the shear. The whole apparatus is strongly constructed and well finished. An absolutely pure shearing test is given by means of this apparatus. 1 The general design of the machine is arranged to give easy access to the ram head for the removal of the hydraulic leather packing. The straining racks are always in tension and are secured from deflection due to their own weights by the insertion of machined strips. The floating frame is suspended by means of links having hardened steel bearings. Amsler Testing Machine (Fig. 11). This machine differs from any that we have hitherto mentioned in the fact that no beam or weight is used to measure the load on the specimen. Great care is taken in machining the cylinder A and its ram, to ensure a perfect fit between the two, so that no cup leather, or other packing, is necessary to prevent leakage. The friction at this point being eliminated, the pressure inside the cylinder may be taken as a measure of the load on the test bar. Oil is forced from the compressing cylinder B into the hydraulic 1 This statement, made by the makers, must be understood to mean pure shear as far as it is possible to obtain it without the very special means described in the section on compound stress. MACHINES FOR TENSION, COMPRESSION, ETC. 35 cylinder A, where it does work on the specimen. The pressure from A is transmitted by means of a pipe D to the cylinder C. As this pressure is necessarily very high, no mercury column of convenient height can be utilised at this point to record it. Accordingly it is made to act on the ram E which actuates the plunger G. This combination forms a reducer (which is exactly the opposite of an intensifier), for, since the rod and plunger E and G are in equilibrium, it follows that the total FIG. 11. Amsler Testing Machine for Short Compression Specimens. downward pressure on E is equal to the total upward pressure on G. Now, since G is much larger than E, the pressure per square inch on G must therefore be much less than that on E ; in fact, the pressures per square inch in C and F will be inversely as the areas of E and G. The plunger G presses downwards on a layer of oil, the pressure of which is balanced by that of a column of mercury H, which is calibrated to read directly the load on the specimen. In some cases, however, the maximum power of the machine is not required, so that greater accuracy in reading the loads on the specimen is desirable. The ratio of reduction is then D 2 36 A HANDBOOK OF TESTING MATERIALS made less by the following device : Oil is pumped into the chamber F by means of the hand pump K from the reservoir L. In this way the plunger G is raised until the sleeve M comes into contact with the ring N and raises it off its seating. The oil in C is now pressing not only on the ram E, but also on the annulus N, which surrounds it, thus enlarging the effective area of E. This, of course, has the effect of raising the total load on the top piston and, consequently, that on the bottom one G. Thus the pressure in F is more nearly equal to that in C than was before the case ; the mercury column H is made longer, and thus more sensitive to the rise in pressure in the main cylinder A. In fact, the arrangement is equivalent to an enlargement of the scale of loading. To eliminate friction as far as possible in the reducer, the plunger and piston are given a backward and forward rotation by means of linkwork actuated from the main compressor. The specimen to be tested is placed between the two tables and P, adjustment for length being effected by the screw and hand- wheel Q. Richie* Testing Machine (Fig. 12). In the Riehle machine we again revert to the jockey weight and lever as a method of measuring the force exerted on the test bar. The loading mechanism in this case, however, consists, not of a cylinder and ram, as in previous machines, but of two strong screws D and E, which pass vertically up the machine. These screws are rotated (through gearing which permits of several speeds being used) by an electric motor, and pass freely through the bed of the frame. The specimen A is held by clips to the framework B and the table C. The latter is moved either upwards or downwards by rotating the screws D and E, on which it fits like a nut. The whole of the framework B rests on a system of levers to which the load on the specimen is transmitted through knife-edges at F and G. The final balance is effected by the jockey weight H, which moves on the graduated stillion arm K. It can easily be shown that the pull P in the connecting rod M L does not depend 011 the relative loads on F and G, MACHINES FOE TENSION, COMPEESSION, ETC. c O-4 60 but merely on the total load on the knife-edges. Conse- quently, if the specimen is un symmetrically placed in the grips, the accuracy of the machine will not be affected. In machines of this type as actually built there are several 38 A HANDBOOK OF TESTING MATERIALS special features. Placed between the main driving pulley and the vertical screws is a system of change gears, controlled by three levers and a hand-wheel. One lever works the reverse while each of the other controllers give two speeds, and by varying the combination of these four levers we can get eight speeds forward or reverse (i.e., tension or compression). The top speed is generally only used for setting the machine. For TbSnilior FIG. 13. General Arrangement for Girder Testing. rapid commercial testing it is also desirable to fit change speed pulleys on the driving shaft, or use a variable speed motor. By this means the machine can be set or changed over from a tension test to a compression test in a few minutes without any manual labour on the part of the experimenter. These machines are almost invariably fitted with autographic apparatus, and the whole can be made entirely automatic by the means described on page 42. One excellent feature of this type of machine is that, as the experimenter stands at the levers, specimen, poise reading, autographic diagram, and all MACHINES FOE TENSION, COMPRESSION, ETC. 39 the controlling levers can be seen or regulated without moving his position. It may be added that, since to develop the full advantages of this machine the whole is run as fast as possible, there is, of course, a liability for an inexperienced experimenter to smash some of the gears by changing speed carelessly. In fact, considerable experience is necessary to manipulate everything in the best possible manner. It is sometimes urged against this type of multilevel* machine that wear and friction are likely to cause greater inaccuracies than in single lever machines. Arrangement of Machines for Bending Tests. All the machines hitherto described can be used for tension, com- pression, or bending tests. Fig. 13 shows the arrangement of a vertical testing machine for testing girders, etc., in bending. Connected with the stillion knife-edge by means of four tie rods A and B, is a table C, with a flat machined face. Two knife- edges D and E are bolted to this table in such a way that their distance apart can be adjusted to suit any size of specimen within the limits of the machine. These knife-edges often rest on spherical seats, which adjust themselves as the load is applied, so that the supporting forces are exactly vertical in direction. The specimen to be tested rests symmetrically across these knife-edges, and is loaded at the centre by means of a knife-edge F. The ram L is forced downwards by the water in the hydraulic cylinder, and the load transmitted by means of tension bars H and K to the knife-edge F. Horizontal machines can be adapted to take specimens for bending in a similar manner. They are not, however, so convenient for testing large specimens as the vertical type of machine, as in the latter the girder rests by its own weight on the supporting knife-edges, while in the horizontal machine it has to be supported in position by the central load. Bailey Transverse Testing Machine. A simple machine for testing the transverse strength of a beam is that made by W. H. Bailey & Co., Ltd., which has a capacity of 40 cwt. The load is applied, as in the case of ordinary testing machines, MACHINES FOE TENSION, COMPEESSION, ETC. 41 by means of a lever and sliding weight. The test bar is sup- ported at the ends in two blocks, which have knife-edges pointing downwards as the load is applied in an upward direction. The load is applied through a knife-edge at the centre of the specimen, and is produced by sliding the jockey- weight along a graduated lever. The lever itself is counter- poised, as in the case of other testing machines, so that it is only the sliding weight which loads the test bar. As the specimen bends, the lever is prevented from assuming an inclined position by means of a screw and hand-wheel, which take up the movement of the bar. In making these tests it is obvious that the bars must be accurately placed and the load applied centrally. Keep's Testing Machine. This is used to test small bars about J inch in thickness, and from the tests so made the quality of cast-iron is determined. It is constructed to trace a diagram of the behaviour of the bar while under test. The load is applied in a similar manner to the Bailey machine, but the lever is not kept floating, being allowed to go down with the bending of the bar. A pencil arm is attached to the centre of the bar and a sheet of paper placed in the holder behind it. The movement of the bar is magnified five times, so that the actual deflection is more easily measured. The paper holder is moved in a horizontal position as the jockey weight moves along the lever, the two being connected together by means of cords. In this way a curve of deflection against load is obtained for each specimen, and the behaviour of the bar under transverse loading is thus automatically recorded. The Deflection of Beams. The apparatus shown in Fig. 14 is for the purpose of directly measuring the deflection of small beams under a given loading. The frame A carries two knife- edges B B, whose distance apart is adjustable to suit beams of different lengths. These knife-edges support the beam C which is to be tested by applying a load D through a knife- edge at G. The beam also carries at this point a frame and cross-wire E, the position of which can be observed with great accuracy by means of a telescope F. The load is applied by 42 A HANDBOOK OF TESTING MATEEIALS To Man drive adding weights at D, and the deflection for each load observed by noting the movement of the cross-wire at E against a suitably graduated scale. In this way the accuracy of the laws for theoretically determining the deflection of a beam can be experimentally checked by means of the very simple apparatus just .described. In some cases the cross-wire is placed inside the telescope, and a scale graduated to T Jo of an inch is fixed vertically to the centre of the beam, a reading being taken after the application of each load. The beam may also be fixed as a cantilever by bringing the two knife-edges closely together and -TO poise actuat/ng gear rigidly gripping the end of the beam between them. An old lathe bed can be used for the frame- work of the apparatus. Automatic Testing Machines. The general principle underlying machines fitted with some appliance for auto- matically keeping the poise balanced is as follows : The load is applied at a constant rate in the case of screw machines, such as the Kiehle, by driving the screws at a constant speed, or in the case of hydraulic machines, such as the Wicksteed, by applying pressure at a constant rate to the hydraulic cylinder. The poise weight is likewise run along the beam at a speed which is adjusted to such a value as to tend to run out at a slightly faster rate than is necessary to keep the machine balanced. At some point in the driving mechanism of this latter arrangement is an electric clutch, which is put in or out of gear according as a contact worked by the movement of the poise arm is made or broken. FIG. 15. Change Speed Mechanism in Poise Gear of Eiehld Testing Machine. MACHINES FOE TENSION, COMPEESSION, ETC. 43 Thus, in the arrangement sometimes fitted to the Riehle machines, an electric contact is arranged so that when the beam rises it makes contact, actuates the clutch, and drives the weight along the beam until balance is restored, when the circuit is broken and the poise weight remains stationary until the increase in the load again raises the beam and starts the poise driving-gear again. This latter mechanism is driven by the same source of power as that which drives the screw gear for applying the load. The relative speed of the load-applying gear and the poise-controlling mechanism can be varied by means of the three-disc mechanism shown in Fig. 15. The disc B is driven at constant speed, while the disc A can be moved by hand along its shaft and thus vary the relative speed of A and B. It is desirable to adjust this velocity so as to keep the arm balanced as near as possible, independent of the electric control, as otherwise the contact makes and breaks in rapid succession, and as the circuit is of necessity highly inductive, violent arcing takes place, especially if the current is supplied from the lighting mains through a suitable resistance, as is sometimes done. In any case, a certain amount of trouble is generally expe- rienced from this latter cause, and some skill is required in order to reduce it as much as possible. It will be observed that the mechanism shown in Fig. 15 is reversible, and consequently as soon as the maximum load is reached it is desirable to put this over to a fairly high speed of reverse and actuate the clutch mechanism by hand. No doubt even this latter could be controlled electrically by placing a contact on the bottom stop and changing over from one stop to the other at the same time as the mechanism of Fig. 15 is reversed, but such is not usually done, and the time between maximum load and breaking load is usually very short, and an experienced operator generally prefers to control this portion himself. A very fine example of an automatic testing machine is that installed in the Northampton Polytechnic Institute, built 44 A HANDBOOK OF TESTING MATEEIALS by Messrs. J. Buckton & Co., of Leeds, to the requirements of Mr. C. E. Larard. 1 In this machine the poise weight is driven by a separate electric motor, the speed of which can be varied over a wide range by an adjustable field resistance. The motor drives the poise through an electric clutch, but differs from the Riehle control, inasmuch as the fall of the beam does not break the circuit, but short circuits the clutch coils, an arrangement which prevents a great deal of the destructive arcing previously mentioned. A novel and important addition is likewise made by which the instant the clutch is short circuited a brake is applied to the driving shaft, thereby preventing any over-running of the poise weight. This clutch is held off by means of an electro-magnet, which is released simultaneously with the clutch. The straining motion is applied by hydraulic power from an accumulator, and the speed can consequently be adjusted to a nicety by the control valve. It is stated that in practice it is found that the speed of straining and working the poise weight can be so relatively adjusted as to be almost independent of the electric control. An important addition to this particular machine is the provision whereby the poise weight can be made either 1,000 or 2,000 Ibs., and a further load can be added at the end of the beam, increasing the maximum load to 150,000 Ibs., while another special attachment enables torsion loads of 400,000 inch-lbs. twisting moment to be carried out on short specimens. Shackles. The design of the grips for holding the specimen in the testing machine is a very important point. They should be designed to give as nearly as possible a perfectly axial stress. If the pull is not central, bending of the specimen will result, and the strength obtained by experiment will not be the true tensile or compressive strength of the material. The wedge principle is mostly employed for tension shackles, to prevent the specimen from slipping when the load is applied. 1 See paper in Proc. Inst. Mech. Eng.. July. 1!M)7. MACHINES FOR TENSION, COMPEESSION, ETC. 4o 46 A HANDBOOK OF TESTING MATERIALS The principle of these is exemplified by the " Wicksteed " grips, shown in Fig. 16. The flat test bar is gripped between two wedge grips which have serrated faces, like a file, to pre- vent the specimen from slipping. The grips themselves rest in a seating, whose inner sides are inclined to the same angle as the outer sides of the wedge. The whole is secured in a round cast-iron plate, which is fixed by tension bars to the beam on which the jockey weight slides. It will readily be seen that, the greater the pull on the specimen, the more firmly will the grips wedge them- selves in the seating, and the more tightly will they hold the specimen. Another type of specimen is enlarged to form an eye at each end. Pins are passed through the hole at each end to secure it to the forked shackle. In this way the test piece is free to move in one plane to ensure axial loading. The method most often used, however, enables the test bar to move in any direc- tion, so that bending the bar is almost an impossibility. This 3L 17. Ball Seating for Tension Specimens with Screwed Ends. spherical or ball ends. seat. If the bar is merely cast and not machined, it may be made with The grips, which are each made in two parts, have spherical seatings on them, between which the ball ends of the specimen rest. As the load is applied, these seatings allow the test bar to adjust itself with its axis in the direction of the pull. When the specimen is machined, the ends are usually enlarged, the inside end of the enlarged part being spherical in shape, so that it is self-adjusting in the clips ; or the ends of the specimen can be screwed and fitted in hemispherical seats as shown in Fig. 17. MACHINES FOE TENSION, COMPRESSION, ETC. 47 Compression Shackles are easier to design, merely consist- ing in most cases of two flat parallel plates between which the accurately faced ends of the cylindrical test piece are gripped as the load comes on. In some cases, especially where long specimens are used, one of these plates is formed with a spherical back, which adjusts itself in parallel with the other. Where great accuracy is required, as in research work, it is necessary to use seatings which shall ensure, as far as possible, central loading. Figs. 18 and 19 show two methods of loading ordinary compression specimens. The method shown in Fig. 20 has been tried by the author, and found entirely unsuccessful. Fig. 21 illustrates the method employed by Prof. Lilley in loading hollow struts during some of his classic experiments, mention of which will be found in the bibliography. Calibration of Vertical Machines of the Wicksteed Type. Machines of this type can be readily tested for accuracy in the following manner. To test for wear of knife-edge and to estimate width of edge : (1) Open out the shackle heads to the maximum distance apart as if for a compression test. (2) Carefully ascertain if any parts of the shackles which move relatively to one another are in contact, as the friction produced by this means may be considerable. (3) When everything is free, set the beam indicator so as to be right over one side of the slot, and then get beam in equilibrium with the beam end near the top stop. Alft FIG. 18. Ball Seating for Specimen in Compression. 48 A HANDBOOK OF TESTING MATEEIALS FIG. 19. Method of Loading Compres- sion Specimen. Y///////////////// ///////// FIG. 20. Unsuccessful Method of Loading Compression Specimen. Sectional Elevation. Plan. FIG. 21. Prof. Lilley's Method of Testing Hollow Struts. MACHINES FOE TENSION, COMPEESSION, ETC. 49 (4) When all is balanced set vernier accurately to zero, and again adjust for equilibrium if necessary. (5) Add weights to the lower shackle up to half a ton or more. In this connection it will be found convenient to first put on the cross-beam used in beam tests, as in a 50-ton Wicksteed, this weighs over 400 Ibs., besides making a con- venient support for the other weights. (6) Again restore equilibrium and note reading on scale. (7) Shift the beam indicator over to other end of slot, set FIG. 22. Diagram showing Effect of Wear of Knife-Edge in Machines of the Wicksteed Type. for equilibrium with beam nearly on bottom stop; note reading. (8) Eemove load ; adjust to equilibrium ; note reading. It will be seen that we thus obtain (a) True load on shackle. (b) Eeading of machine with beam up. (c) Eeading of machine with beam down. Record of Calibration Test. In a certain test on a 50-ton Wicksteed machine the following results were obtained : (a) 1485-6 Ibs. = '663 tons ; (6) '657 tons ; (c) '660 tons. The mean of (b) and (c) = '6585. T.M. E 50 A HANDBOOK OF TESTING MATEKIALS Hence percentage error at low loads = X 100 low. The percentage error would probably be much less at higher load. The difference in reading with the beam up and the beam down is due to the width of the knife-edge. The correct distance of the knife-edge D from the fulcrum is 3 inches, hence the width CB is of the nature 660- -657 ,,_^K)3 _ 6685 --6685 X 3 ' Exaggerating the width as in Fig 22, we see that when the AC beam is up the leverage is ^, and when the beam is down AB' the leverage is :r- This is as small as can be reasonably be expected. To Check the Weight of the Balance-weight. (a) Hang some known weight, say 56 Ibs., on the longi- tudinal scale near the end of the beam. (b) Balance the machine and adjust vernier to zero. (c) Move the weight through a definite number of scale divisions (preferable 40) and again balance the machine. (d) Note how far the balance-weight has been moved. In a certain test on the same machine as above when the 56-lbs. weight was moved through 40 divisions (from 50 tons to 10 tons) the balance- weight was moved as near 1 division as could be read on the machine. Since moving 56 Ibs. through 40 divisions is equivalent to moving 1 ton through 1 division, the balance-weight was certainly within *1 per cent of 1 ton. CHAPTEK IV STRAIN-MEASURING INSTRUMENTS WHENEVER a body is subjected to a load, or stress of any kind, some strain or deformation of shape is sure to result. If this strain occurs in finished structures, it may have incon- venient or even dangerous results, so that we desire to know the actual effect produced by stressing a material to a certain degree in changing its shape. Moreover; if we can find out the separate effect of each increment of load, the stress strain curve drawn from such readings may give us some useful information about the material that we propose to use. The most usual case is that of tension. Let us assume, then, that we are about to test a bar of the material in question, with a tensile load between certain limits. The immediate effect of the application of this load will be to stretch the bar by an amount which varies with the intensity of the load. We wish to measure the amount of this stretching, which is, after all, very small in proportion to the length of the bar. The easiest way, of course, is to centre-punch the bar at a point near each end, and measure the distance between them with a pair of dividers after the application of each increment in the load. Another method would be to screw two collars to the test bar, at a distance apart, fixed by the length of a standard rod. As the bar began to stretch, a wedge gauge could be introduced into the space between the end of the standard rod and the bottom collar, the extension thus being measured with very fair accuracy. But these methods are quite inadequate to measure the very small extensions which take place at stresses below the elastic limit of the specimen. These it is most valuable for us o know, as it is only at stresses below the elastic limit that any material can be used. For the purpose of measuring these E 2 52 A HANDBOOK OF TESTING MATEEIALS minute extensions, some form of " extensometer " must be used. In these instruments the stretching of the specimen is magnified either by mechanical or optical means, until the slightest extension can be measured with an accuracy ranging from JODOO f an i ncn ^ n ^ ne former case to jooiooo ^ an inch in some forms of optical instruments. Screw f*\ icrom cKc,t\ FiG.^23. Ewing's Extensometer. E wing's Extensometer. A diagram of this instrument is shown in Fig. 23. The framework of this extensometer, which is indicated by thick black lines in the diagram, is secured to the specimen by screws at A and B. The distance between them, generally called the gauge length, is usually eight inches. As the test bar stretches, the distance A B becomes greater, and STRAIN-MEASURING INSTRUMENTS as the end H is fixed in position relative to the frame, it will readily be seen that the movement of G will be double that of B. The movement of G is observed by the passage of a cross hair C along a graduated scale in the eye-piece of the telescope D, which is capable of adjustment by means of the screw E. &%>. . c^afl i Cross Wire Microscope FIG. 24. Diagrammatic Outline of Latest Type Ewing's Extensometer. The graduated scale is calibrated by means of the micrometer screw F, which is turned through a known distance, and the movement of C observed. The actual value of each scale division can thus be obtained in a very simple manner, the accuracy being about ^oioo ^ an inch. In the latest type of this important instrument some modifications have been made in the general arrangement, although not in the principle. Fig. 24 shows this latest form. 54 A HANDBOOK OF TESTING MATERIALS The object sighted is one side of a wire stretched hori- zontally across a hole in the bar K and illuminated by a small mirror behind. The distances C P and C Q are in this instance equal, with the effect that the movement of the sighted mark is double the extension of the rod. The length of the microscope is adjusted so that one turn of the screw causes the mark to pass over 50 scale divisions in the eye-piece. This adjustment should be tested with an extensometer as mounted on the specimen, and, if need be, the length of the microscope tube can be altered by drawing out or in the portion carrying the eye-piece. A complete revolution of the screw L, which has a pitch of /^th of an inch, should cause a displacement of the mark through 50 divisions of the eye -piece scale. Eeadings are taken to tenths of a scale division, so that this displacement corresponds to 500 units. Each unit then means g^ou inch, in the extension of a test-piece. The scale engraved in the eye -piece of the microscope has 140 divisions each corresponding to ^V^ inch of extension, and by estimation of tenths of a division readings are taken to -soiioo inch. The screw L further serves to bring the sighted mark to a convenient point on the micrometer scale, and also to bring the mark back if the strain is so large as to carry it out of the field of view ; thus, a single turn of the screw adds 500 units to the range shown on the micrometer scale. In dealing with elastic strains there is no need for this, as the range of the scale is itself sufficient to include them, but it is useful when observations are being made on the behaviour of metals as the elastic limit is passed. In other forms of this instrument the micrometer is dispensed with, the position of the telescope relative to the frame being fixed. The scale in the eye-piece of the telescope is then so graduated that its divisions represent some definite fraction of an inch. Unwin's Extensometer. In Prof. Unwin's instrument, shown in Fig. 25, two tee-brackets are fixed to the specimen a gauge length apart. To each of these brackets a spirit level STRAIN-MEASURING INSTRUMENTS 55 is attached, so that they can always be kept exactly in a horizontal position. To the lower bracket, in addition, is clamped the measuring pillar C, which carries within it a fine screw D, with a micrometer head E. This screw has FIG. 25. Un win's Extensometer. 50 threads per inch, and, since the micrometer has 200 divisions, an accuracy of joioo f an i ncn * s obtainable. When about to take a reading the lower bracket is first levelled by the adjusting screw F, and then the upper bracket is levelled by the micrometer. The difference between this micrometer reading and the previous one gives us the caange in length of the specimen. 56 A HANDBOOK OF TESTING MATERIALS Marten's Extensometer. Fig. 26 is a diagrammatic sketch of Marten's instrument. An arm A with a point at its lower end is clamped to the specimen by elastic bands or springs at B and C. Between the top end and the specimen there is a r - EiG. 26. Marten's Pointer Extensometer. FIG. 27. Ashcroft's Extensometer. small diamond-shaped lever E. Any movement of A relative to the specimen, such as is caused by an extension, tilts the piece E, thus moving a long pointer D over a fixed scale F. When the magnification is 50, readings can be conveniently taken to ^ J n mm. Ashcroft's Extensometer. An instrument described by Ashcroft is shown in Fig. 27. STRAIN-MEASURING INSTRUMENTS 57 The upper end of the knife-edge B is rigidly connected to the upper end of the specimen whilst A is clamped to the lower. A and B fit into small notches in the lever D, and consequently a greatly magnified movement of the relative FIG. 28. Kennedy's Exteusometer. motion of A and B is shown by a pointer D on a fixed scale C. Kennedy's Extensometer. This is almost identical in principle with that of Martens. Fig. 28 shows the instru- ment in outline, and is practically self-explanatory. Beading can be taken to the nearest TU Joo or 20000 i ncn - Stromeyer's Rolling Pin Type Strain Indicator. This type 58 A HANDBOOK OF TESTING MATERIALS of instrument is shown diagrammatically in Fig. 29 adapted for the measurement of strains in testing-machine specimens. A modified form, but involving the same principle, has been largely used for the measurement of strains in finished structures such as bridges. It will be seen that in the instrument illustrated in Fig. 29 two flat strips A B and C D are held together by springs E and F. Between them is a small roller G consisting of a piece of circular wire which has been carefully prepared so as to be as nearly an exact circular section as is possible. FIG. 29. Stromeyer's Boiling Pin Extensometer. The ends C and B are clamped to two points of the specimen which move relatively when the specimen extends. The motion of A B over C D causes the roller G to turn, which movement is read by observing the movement of a pointer H over a fixed scale J. Kesults have been obtained in which each scale division represents joioo inch, and a further estimation could be made. In the design of instrument used on existing structures, shown in Fig. 30, A is the roller moving between a fixed plate C and a moving one B. B is connected to the upper gauge clamp by a piece of annealed wire F of the same material as the structure. F is kept tight by means of a spring E. The gauge length is STRAIN-MEASURING INSTRUMENTS between the screws H H and G G. The rolling pins are made of hardened cast-steel. They are attached to light straw pointers by means of paper envelopes and sealing wax. The papers are cut after fixing, so that the pointer is balanced. These rolling pin pointers can easily be renewed or re-fixed with a warmed pair of pincers, and this is generally necessary when sudden strains, which appear to act like blows, are recorded by the instruments. In such cases the straw pointers will generally break off. It is not advisable to replace them by wire pointers, for then even small but sudden strains have a powerful effect and loosen the rolling pin, which defect is not always noticeable and may lead to errors. The Cambridge Extenso- meter. This instrument, the general scheme of which is shown in Fig. 31, consists of two separate parts, each of which is separately attached to the test-piece A by hard conical points. The steel rods carrying these points slide in geometric slides, and after being driven gently in centre punch marks at P and P l are clamped in position. Both parts of the instrument should FIG. 30. Stromeyer's Rolling Pin Type Extensometer (as used on finished structures). 60 A HANDBOOK OF TESTING MATERIALS be capable of rotating quite freely about the points, but there must be no backlash. The lower piece carries a micrometer screw fitted with a hardened steel point B, and a divided head C. It also carries a vertical arm D, at the top of which is a hardened steel knife-edge. The upper and lower pieces work together about this knife-edge, the balance weight serving to keep the two parts in contact. A nickel-plated Balance Weight Graduated Scale^ FIG. 31. Cambridge Extensometer. flexible steel tongue F, forming a continuation of the upper piece, is carried over the micrometer point B. This tongue acts as a lever, magnifying the extension of the specimen, so that the movement of the steel tongue to or away from the point B is five times the actual extension of the specimen. To take a reading the thin steel tongue F is caused to vibrate, and the divided head then turned till the point B just touches the hard steel knife-edge on the tongue as it vibrates to and fro. This has proved to be the most delicate method of setting the micrometer screw, as the noise produced and the 62 A HANDBOOK OF TESTING MATERIALS fact that the vibrations are quickly damped out indicate to jouo mm - ^ ne instant when the screw is touching the tongue. This instrument is, according to the National Physical Laboratory's report, reliable to about one-thousandth part of a millimetre. Optical Instruments. Greater sensitiveness than that given by any of the foregoing instruments can in general only be attained by some optical appliance. There has been rather a prejudice against the use of optical instruments in this country, probably on account of difficulties in focussing, etc. This is, however, less noticeable now than formerly. These optical instruments were first used on the Continent. Their dh FIG. 33. Marten's Mirror Extensometer. great advantage is the employment of a weightless lever a beam of light for magnification purposes. Bauschinger's Apparatus. Fig. 32 shows an instrument designed by Bauschinger. a a and b b are the points held in by suitable clips to the gauge points of the specimen. The connection between the two parts of the instrument, one of which is fixed at a and the other at 6, is a rolling one formed by a caoutchouc roller c moving over a light spring d. Hence, as the two gauge points move relative to one another, the roller c is turned in a similar manner to that in Stromeyer's instrument. The twisting of the roller moves a small mirror e, the angle of movement being observed by means of a telescope at/. Readings can be taken to T Q Juo mm - It will be observed that readings are taken on both sides, and to Bauschinger belongs the credit of first realising the necessity of measuring strains in more than one plane, an idea which has since been extended by the author to three planes at 120 apart. STKAIN-MEASUBING INSTRUMENTS Marten's Mirror Extensometer. In another form of optical instrument devised by Martens, the rollers used by Bauschinger are replaced by small diamond-shaped knife-edges. Figs. 33 and 34 illustrate this instrument. Like Bauschinger's instru- ment, it measures strains on two sides. A A are two flat plates having knife-edges at E. They are held together by springs B on each side, and press on knife-edges at C. Fig. 34 shows another view of the diamond pieces, and the method of carrying the mirrors. M is the mirror pivoted in the frame P by small set screws at N N. On the other side a set screw presses against the mirror, whose motion is resisted by a small spring D. This allows the inclination of the mirror to be adjusted. E is a weight to balance the mirror, and T a pointer which, when set in contact with the bar S, indicates that the knife-edge Q is correctly at right angles with the bar S. Morrow's Extensometer. Yet another form of optical instrument is one devised by Dr. Morrow, which embodies the same principle as Marten's but is so arranged that the image of a scale held some distance away can be viewed in the mirror by means of a telescope. A second stationary mirror reflects the zero mark. This instrument is said to read to j^oJooo f an inch, having a magnification of about 3,000. Stromeyer's Optical Extensometer. An extensometer of great delicacy was designed by Mr. C. E. Stromeyer in the eighties, originally for the measurement of local strains in metal structures, but the particular design illustrated was r. 1 i r ^p ^D N N M { FIG. 34. Detail of Mirror Attachment, Marten's Extensometer. A HANDBOOK OF TESTING MATERIALS specially designed for measuring the cross contraction of test pieces in order to obtain a direct measurement of Poisson's Batio. 1 It depends on the important principle of the in- terference of light. If white light is projected on to a small piece of dark glass, the reflected ray can be con- sidered as made up of two parts, one reflected from the outer surface and one from the back surface. Since one has travelled slightly farther than the other, the two rays will not be exactly in phase, the consequence being that they interfere and in the case of white light the reflected ray would be split up into coloured bands, while in the case of yellow sodium light, which is used with this instrument, alternate bands of dark and light are observed when the reflected ray is seen in a suitable form of telescope. If before being reflected by the dark glass the light suffers a previous reflection at another surface and the two reflecting surfaces are moved relative to one another, the dark inter- ference bands will appear to move along the second surface past a line which can be scratched on the surface. The FIG. 35. Stromeyer's Optical Extensometer. 1 See p. 134. STKAIN-MEASUEING INSTRUMENTS 65 Washers for fasten- ing. Twisted Strip. \M Mirror. distance apart of the interference bands can be calculated and also the relative movement of the two reflecting surfaces for a given movement of the bands past the fixed line. Fig. 35 shows a diagrammatic sketch of the method used to carry out this principle in practice. T is the section of the test piece which is pressed against the point on the frame F by the screw S. G is the dark glass which, as soon as T contracts, is pulled away from the glass prism P by means of the four helical springs 1 Z Z which surround the columns C C and which are firmly secured to the frames F 2 F 3 . The latter carry the adjustable glass prism P, which is so shaped that the ray of sodium light LI does not coincide with its reflected ray L 2 . The source of light was a sodium-tinted Bunsen flame, while L 2 was observed through a telescope. The inclination of the rays of light in the narrow space between the prism P and the dark glass G was carefully measured, and found to be 19, so that each interference band, as seen in the reflected yellow light, ought to represent a distance of "0000109 inches. That is to say, a relative movement of the two frames of this amount would cause a movement of the dark bands equal to their distance apart. Careful measurements with the fine screw S, however, showed the move- ment to represent "0000120 inches, or 10 per cent. more. The Sphingometer. This instrument may be employed to measure strains in one, two or more planes. Calibration is made for each test, if it is desired to note the elastic con- stants ; if it is only desired to note the load at certain marked 1 There are two other springs and columns not shown. T.M. F Twisted strip. Washers for fasten- ing. FIG. 36. Strips for Sphingometer. 66 A HANDBOOK OP TESTING MATERIALS points, calibration is unnecessary. The adaptation of the instrument for meaburement of strains in one or two planes will be fairly obvious if the type used for measurement in three planes is described. The results obtained by experiment show that, with ordinary R --- Twisted, Strip. M --- Mirror. K --- Micrometer. L U N N- _ Sliding bush to \^J r^? which strip is fastened. _ Set screws. _ Feather key. -Spring. FIG. 37. Section through Sphingometer Strip Holder. methods of testing, there is a considerable variation in the stress on the bar. It is suggested that this is due to the fact that the load never passes directly through the axis of the specimen. It is therefore misleading to divide the total load registered on the testing machine by the area of the specimen in order to estimate the maximum stress on the material. STEAIN-MEASUEING INSTEUMENTS 67 To avoid this error the author has devised a special form of extensometer for the purpose of measuring the strain on a specimen, in tension or compression, in three planes. The general principle on which the instrument depends is that of the twisted strip used by Professors Ayrton and Perry, who have shown that the angular rotation of a mirror fastened . FIG. 38. Strip Carrier and Specimen Grip for Sphingometer. to a strip, as described below, is proportional to the extension (or shortening) of the length oFthe strip. A piece of phosphor bronze, about 7 inches long, J inch wide, and 0*004 inch thick, is used (Fig. 36). The material is divided into two lengths. At the centre an attachment is made so that a mirror (preferably of about 1 metre focus) may be fastened. One-half of the strip is then wound as a right-handed spiral, and the other half is wound as a left- handed spiral. At the two ends the strip is soldered to a thin F 2 68 A HANDBOOK OF TESTING MATERIALS strip of metal, which has drilled through it a hole for a set screw. By means of this set screw at each end the strip is attached to a tube in which it is carried. By fastening this tube to a frame with two set screws, to secure the frame to the specimen, an extension of the length of the specimen between the two fastening screws can be measured. During the author's experiments it was found that different strains were recorded on the same specimen if the points of attachment were moved round the bar. The instrument has, therefore, been arranged to take measurements in three planes, three tubes of the form shown in Fig. 37 being used, which were clamped to the carrier (Fig. 38) by means of the V-blocks V. The carrier grips the specimen centrally by the set screws F, which are screwed in so as to space the strips equi-distantly from one another and from the centre. A lamp is arranged so that a ray of light is reflected by the mirror on to a scale usually placed about 1 metre from the mirror. For calibrating the strip a known extension is given by rotat- ing a micrometer head. The effect of this is that the beam of light passes across, say, 100 scale divisions, when the micro- meter head shows that the strip has been extended, say, TifW P ar t f an inch. In which case each scale division is clearly jodooo P ar ^ f an inch. When the instrument is in use, T and T' are rigidly clamped in position to the upper and lower points of con- nection to the specimen. From the figure (37) it will be seen that when the micrometer head K is turned round, the bush L is forced down, compressing the spring N and shortening the strip E, E. This causes the mirror to rotate. If it is desired to change the direction of the beam of light across the scale during calibration, the micrometer head is turned round in either direction, the spring N forcing the bush L up when K is unscrewed. It will be seen that the carriers are built up of three castings, C, which are interchangeable. They are screwed on to distance pieces, D. By altering the position of the points, at which the castings are fastened to D, the size of STRAIN-MEASURING INSTRUMENTS 69 the triangle can be altered at will. This, therefore, gives the desired flexibility as regards the diameter of the speci- men. Flexibility concerning the length between the gauge- points of the specimen is secured by altering the position of the sphingometer casing in the V-blocks. The limiting dis- tance is the thickness of one of the castings 0, which is about half an inch. In order to extend the length between the gauge-points beyond that shown in Fig. 37, a longer i SCALE DEFLECT/OMS FIG. 39. Calibration Curve for Sphingometer Strip. tube is used at T', and this can be extended indefinitely. It is, therefore, obvious that the distance between the gauge-points may be any length required from half an inch upwards. Experience has proved that in calibration the most satis- factory method is to work across the scale in sections. This removes any error due to the non-curvature of the scale. It makes it also quite a simple matter to read direct the actual extension ; Fig. 3 ( J shows the method employed. It 70 A HANDBOOK OF TESTING MATEEIALS will be seen that the readings are plotted against each other on squared paper. The sphingometer is generally used for tension or com- pression tests, but by the addition of another strip and casing, A.. Specimen,. C .Carriers. D . Sphingometer strips . T *Torslon . M; Micrometer heads. Z -. Z/Y> mirror. FIG. 40. The Sphingometer fitted with Torsion and Tension Strips (for description of Torsion Attachment, see page 124). torsion tests can also be carried out by the instrument. This torsion sphingometer is on a plane perpendicular to the plane of the tension strips. The illumination for this torsion mirror is obtained from the same lamp as that used for the tension mirrors. In Fig. 40 is shown an isometric projection of the instrument. The torsion fitting is explained fully later. It has_been found advantageous to use gauges in order STKAIN-MEASUKINa INSTRUMENTS 71 to insure that the two carriers are parallel. When the instrument is, for the first time, put on a specimen, the procedure is as follows : Steel distance pieces are placed in the V-blocks, made to hold the tubes in which are carried the strips. A gauge is then inserted at each corner of the carriers. This gauge can, of course, be altered for varying lengths of specimens. When the instrument is changed over from one specimen to another of a smaller or greater dia- meter it is convenient to use a gauge to ensure that all of the three fastening screws are at equal distance from the carrier frame. If it is desired to remove the instrument from one specimen to another one of the same gauge length, all that is necessary is to remove the sphingometer tubes and insert the steel distance pieces. The relative positions of the carriers cannot then change, and the framework is quite rigid. One of the most important facts noticed by the use of the instrument is the variation in strain which accompanies a tensile or compressive stress. Bending, probably due to eccentric loading, occurs always. It is to be expected that when a specimen is in compression there will be bending. In order, therefore, to emphasise the importance of this fact of variation of strain the author will confine his remarks to the tension tests. The problem of axial loading for a tension specimen is more difficult than would at first sight appear. A spherical seat (Fig. 17) may be used to secure alignment, but whatever precautions be taken to make the spherical seatings an accurate fit, it is doubtful whether they do pull into line when once the load is applied. In any case, it has been found from the results of tests that more uniform stress distribution exists when spherical seats are used than when the specimen is held in the ordinary wedge grips. The unequal distribution of stress upon the specimen is the most determining cause of the unequal strains recorded during the experiments. It is the strain readings just pre- vious to elastic failure which are most important. From 72 A HANDBOOK OF TESTING MATERIAL these we are able to deduce the maximum stress upon the material, and compare it with the mean stress. To show this clearly the author may mention a test, carried out by him, in which a specimen of mild steel showed from such deductions that a maximum stress of more than 13 tons per square inch was really on the specimen, while the mean stress recorded was 4'5 tons. This was an extreme case, but it would have been passed as normal under the usual conditions of recording extensions. In other words, it would have been recorded that the material had a load of 4*5 tons at elastic failure, from which the average stress is 37,980 Ibs., whereas really the maximum stress was practically three times this amount. The results of a test are given on p. 113. Autographic Recorders. When a material is tested in tension or compression, certain strain-measuring instruments, to be described later, are of value for recording the stretch of the specimen under load, so long as it is elastic. There is, however, a critical load at which the strain ceases to be even approximately proportional to the stress, and after this load is passed the instruments useful for noting stretch during the elastic period are too sensitive. There are two methods of procedure which may then be followed if the stretch, after the material is no longer elastic, is to be noted. An observer may take certain measurements of the distance between the gauge points, the load at the instant the reading is taken being carefully noted. Or an arrangement may be fitted so that both the load and the stretch are automatically recorded. It is usually the practice to fit the recorder to the specimen before the test commences. The scale of the diagram on which the strain is thus automatically traced is important. If, as is usual, it does not exceed ten times that of the actual strain, the value of the diagram is in the fact that it gives a continuous record of the relationship of stress and strain after the material is no longer elastic. It is not sufficiently accurate for obtaining the value of coefficients within the period of elasticity. The following description of some STKAIN-MEASUKING INSTEUMENTS 73 ingenious autographic recorders will show how these diagrams are obtained. Unwin's Stress-Strain Recorder. This apparatus, as its name indicates, automatically draws a stress-strain, or rather a load-extension diagram, without any readings being taken PIG. 41. Unwin's Stress-Strain Recorder. at all. This is done by using a rotating drum, as in the case of the steam-engine indicator. It is necessary to give the pencil two motions at right angles : one motion proportional to the load put on the specimen, and the other proportional to the extension produced by that load. To obtain the latter of these two motions, two clamps are fixed to the specimen, a gauge length apart, to each of which is fixed a pulley A A. Similar pulleys are fixed at B and C. A thin wire-cord, to A HANDBOOK OF TESTING MATERIALS D C u j u 1 1 / r which the pencil D is attached, passes round these pulleys, in the manner shown in Fig. 41, and teminates in a weight W STKAIN-MEASUKING INSTEUMENTS 75 to keep all taut. Thus, by this arrangement, any extension of the specimen is reproduced to double the scale by the vertical movement of the pencil. Any total movement of the specimen, due to slipping in the grips, does not affect the pencil, as the wire between the specimen and the drum is made long enough to prevent this. The motion proportional to the load, at right angles to the extension movement, is obtained by rotating the drum E, which gives essentially the same result as moving the pencil horizontally in the opposite direction. This is obtained from the shaft which moves the weight along the beam, and so actuates the load. As this shaft rotates, it turns the pulley G which drives the pulley H by means of a wire driving-cord F. The worm-wheel K is rotated by means of a shaft and worm from H, and so the rotation of the drum E is proportional to the movement of H, and hence to the load on the specimen. Wicksteed's Apparatus. The apparatus shown in Fig. 42 is also used for this purpose. Two arms A and B rest against projections fastened to the gauge-points. The arm B is supported in a horizontal position by means of levers and a balance weight H. The continuation of the arm A is bent downwards, and is guided in a vertical direction by a bearing E. This part of the arm is also provided with a rack C, which gears with the pinion D. Consequently, when the specimen stretches, the arm A moves relatively to B, and so rotates the pinion D. This rotation is transmitted through bevel gearing to the drum F, whose motion is thus proportional to the extension of the test bar. The screwed shaft K, which moves the load along the stillion, rotates at the same time the bevel gear L. The motion is thereby transmitted to the threaded spindle H, which causes the nut G which carries the pencil to move upwards. The vertical motion then in this case is proportional to the load, and the horizontal one to the extension. As it is only the relative motion of A and B that causes the pinion D to rotate, any total displacement of the specimen will not affect the movement of the drum. 76 A HANDBOOK OF TESTING MATEEIALS Henning's Stress-Strain Recorder (Fig. 43). As the exten- sion of the test bar previous to the elastic limit is very small, and after that in the case of ductile materials comparatively PIG. 43. Henning's Portable Autographic Stress-Strain Eecorder. large, the scale of the load-extension curve must be a large one if the former is to be visible. The apparatus here described makes it possible to reproduce the first part of the diagram to a large scale, the extensions after the elastic limit STRAIN-MEASUEING INSTRUMENTS 77 being automatically reproduced to a smaller one. It consists of two hinged frames A and B, one of which is provided with vertical rods N, and the other with tubes N N' into which these fit, so that one frame steadies the other. The length of these rods is such as to keep the frames, at the beginning of the test, a fixed distance apart, though at the same time they are free to move axially when the specimen extends. The frames themselves are provided with spring-cushioned bushes C, and are hinged by taper plugs e and/. The bushes C are allowed to move backwards and forwards by means of the springs D. Through these bushes pass the screws H, which have hardened ends shaped like knife-edges, and are used to fasten the frames to the test bar. The lower frame A carries the drum G, on which the paper is fixed for recording the test. The frame A also carries a parallel motion, similar to that on steam-engine indicators. This mechanism rests on the bar K, carrying two tubes L, which slide on two rods L' screwed into the frame A at a. It is operated by means of a connecting rod d by means of which the relative motion between A and B is thus transmitted to the pencil on an enlarged scale. The wheel F is also supported on A by the arm g l , to which it is con- nected by a link and screw, so that it can swing to and from the marking point at will. The drum G is rotated by a string which is wrapped round it, one end being connected to the travelling load, and the other to a weight which keeps it tight. In using this apparatus during a test, it is necessary to keep the lever absolutely balanced, as if this is not done the increments of load on the diagram will not coincide with the actual loads on the specimen. After the yield point is reached, the extensions become greater, so that it is necessary to reproduce them to a smaller scale. The hooked rod P is then so adjusted that at this point it automatically arrests the parallel motion, causing it to slide on the rods L', so that all subsequent extensions are measured full size. The extensions previous to this are recorded to a scale of five times full size. When the test piece breaks, the instrument divides into two parts, the rods N and 78 A HANDBOOK OF TESTING MATEKIALS tubes L simply sliding out of the tubes N' and rods L', while the parallel motion is suspended from frame B by means of the connecting rod. Should it be desired to use longer specimens it is only necessary to use a longer connecting rod. For com- pression tests a shorter connecting rod is used, so that the marking point will stand at the top of the drum at the beginning of the test, as its subsequent movement will then be downwards in direction. Kennedy's Autographic Method. In most autographic stress-strain recorders the stress recording movements depends on the movement of the poise weight. Now it is obvious that unless the poise is absolutely maintained the PIG. 44. Kennedy's Automatic Stress-Strain Eecorder. diagram produced will not be strictly correct, hence attempts have been made to render the autographic apparatus entirely independent of the operator. Probably the instrument which best fulfills this condition is that devised by Prof. Kennedy. Fig. 44 illustrates in outline the method employed. A is the specimen to be tested, and this is connected by suitable grips to a second and larger bar of mild steel B. The latter is so chosen that at the load at which A will break B is well within that stress below which Hooke's law is exactly fulfilled. Attached to B is a " rolling pin " extensometer (see page 58), and the pointer C is arranged to trace out a line on a smoked glass screen D. Now it is obvious that if B is well within the elastic limit, the movement of the pointer C will be exactly proportional to the extension of B, which is in turn propor- tional to the load on both B and A. The screen D is attached to a fine wire passing over a pulley E and back to the other STEAIN-MEASUEING INSTEUMENTS 79 end of the specimen under test. As A extends it is obvious that the movement of the screen longitudinally will be pro- portional to this extension, and hence the curve traced out by the pointer C will be a true stress-strain curve independent of any other adjustments to the testing machine. There is one rather unfortunate drawback to this apparatus, and that is Helical Spring FIG. 45. Autographic Stress-Strain Eecorder as attached to a Single Lever Machine. the fact that the movement of C will be in a circular arc. It would seem that, in spite of various attempts, this difficulty cannot be overcome without so increasing friction and inertia as to affect the accuracy more than is desirable. The Wicksteed Recorder. One of the most successful methods of obtaining an autographic diagram or stress-strain curve from single lever or even multiple lever machines is to place the poise weight at a position beyond the point of maximum load and then raise the end of the beam slightly 80 A HANDBOOK OF TESTING MATERIALS by means of a helical spring fixed to a support at the top. It is then obvious that, if a specimen be fixed in the machine while the poise weight is just supported by the spring, the load on the specimen will be zero. If now we begin to load '/3 rpr D Area *i Mild Steel Specimen Extension FIG. 46. Autographic Diagram, taken with Apparatus shown in Fig. 45. the specimen by introducing water into the hydraulic cylinder or by any other means according to the particular design of machine employed it is obvious that we tend to raise the poise beam. Any upward movement of the poise lever, however, shortens the helical spring, and hence reduces the load carried by it. Hence, the unbalanced load must be taken STBAIN-MEASUEING INSTRUMENTS 81 T.M. 82 A HANDBOOK OF TESTING MATEEIALS up by the specimen. Now it is obvious that since the load carried by the spring is directly proportional to its extension, the load carried by the specimen is directly proportional to the upward movement of the beam. It is now a simple matter to arrange that this upward movement shall move a pencil over a .drum while the latter is given a rotary motion by means of an attachment to the specimen. Messrs. Buckton have taken out patents for applying this principle by various methods to their machines, of which Fig. 45 can be taken as typical. Fig. 46 is an autographic diagram taken on one of these machines fitted with an apparatus similar to that shown in Fig. 45. Double Autographic Attachment. Fig. 47 illustrates yet another method of obtaining stress-strain diagrams. It will be seen that there is the usual autographic drum driven by the movement of the poise weight, while there are two pencil gears. Both are worked through a system of compound levers from two fixed points on the specimen, the relative movement of which determines the motion of the pencil gear. One system of levers is arranged to give a much magnified movement for the extension of the specimen, and records extension below the elastic limit. The second system of levers is such as will allow the whole extension of the specimen up to fracture to be recorded. It will be seen that there are a number of knife-edges at each centre, so that the leverage can be varied to suit different specimens having different elastic properties. It is, of course, necessary to disconnect the first system of levers when past the elastic limit, otherwise further movement would smash the pencil gear. A typical pair of stress-strain curves are shown. CHAPTEK V METHODS AND KESULTS OF TESTS ON MATERIALS Tension Specimens. There are many varieties of tension specimens, the size and shape varying according to the method of testing and type of machine to be employed. The kind of _ p > f o _ t v^ i _/~ i i fe i f ~~\ \ PIG. 48. Standard Tension Specimen Plate. specimen most usually met with for commercial tests is cut from the rolled metal, the shape being shown in Fig. 48. The cross-section is generally about 2 inches by three- eighths, the ends, which are clamped in the grips being wider FIG. 49. Typical Tension Specimen Bar. so that the specimen is less liable to break outside the gauge length. For bar tests, round specimens are employed, the ends being gripped in V-shaped grips. Such a specimen of normal proportion is illustrated in Fig. 49. Where greater accuracy is G 2 84 A HANDBOOK OF TESTING MATERIALS required the ends of the specimen are threaded, this portion screwing into corresponding holders, which in turn are fastened to the shackles by ball and socket joints. This gives greater freedom in lining up, and enables the load to be applied more nearly axially to the specimen. Fig. 50 shows the type of end used with the grip shown in Fig. 17. The actual length on the specimen over which the test is made is called the "gauge length." This is most usually 8 or 10 inches on specimens of the size already illustrated, but of course tests are often carried out on specimens of much greater length than this. Standard Test Pieces. The following are the recommendations of the Engineer- ing Standards Committee for Sizes of Standard Test Pieces : For Plates and other Structural Material (Test piece A, Fig. 48). In all cases L to be approximately 18". ,, P not less than 9". ,, ,, G ,, 8 . Width W for p IG 50 _ Screwed Thickness over f" maximum width =1J". End for Tension or from f" " =2". Compression Speci- under f =2j". men. For Bars, Rods, and Stays. (Test piece under 1 inch diameter) (Test piece B). Gauge length to be no less than eight times the diameter, and if provided with enlarged ends to be parallel for not less then nine times the reduced diameter. (Test piece over 1 inch diameter) (Test piece F). Gauge length not less than four times the diameter, and if provided with enlarged ends to be parallel for a length not less than four and a half times the reduced diameter. METHODS AND EESULTS OF TESTS ON MATERIALS 85 For Tyres, Axles, Forgings, Castings, Etc. (Test piece C.) Diameter, -564" (J sq. in.). Parallel for not less than 2 Gauge length, 2". (Test piece D.) Diameter, '798" ( sq. in.). Parallel for not less than 3 Gauge length, 3". I! tit O (T) O *- ^ tO _ :S -g fl H 5 g '& Should a rather larger test piece than C or D be desirable, the following should be adapted : (Test piece E) Diameter *977 in. (f sq. in.). Parallel for not less than 4 inches. Gauge length 3j inches. Test of a Ductile Material. In a test of this kind there are certain recognised observations which should be taken. These are as follows : 1. The load at which rupture occurs, thus giving the break- ing stress. 2. The load at which yielding occurs, giving the yield point. 3. The elongation in the gauge length, giving the percentage elongation in the stipulated gauge length. 4. The reduction in the cross-sectional area, giving the percentage reduction in area. Observations before the Experiment. To enable these observations to be carried out the gauge length should be carefully marked off. Measurements should also be taken of the diameter of the specimen if round, or the breadth and thickness if flat. To obtain an accurate value, those measure- ments should be taken in three or more places, and the mean of these readings taken. The specimen should also be marked with some distinctive letter, number, or sign, so that it may be easily recognisable for future reference. This mark should be made outside the gauge length, preferably on one end of the specimen. The Test. The specimen is then placed in the machine, and the load gradually applied. When a certain load is reached, it is observed that the specimen suddenly begins to elongate very appreciably with the further addition of little or 86 A HANDBOOK OF TESTING MATEEIALS no load. The point at which this phenomenon occurs is known as the yield point, and it is said that the " elastic limit " has been reached. There is no difficulty in recognis- ing this stage when reached, as the beam suddenly drops rapidly, and in consequence the pump has to be worked correspondingly fast to keep the beam floating. When the yielding stage is passed, the specimen continues to stretch appreciably as the load is increased. As the breaking point is approached the specimen begins to draw out at its weakest section, where rupture will finally occur. In consequence of this decreasing cross-sectional area, the load ULTIMATE TENSION TESTS AL I (Q) AL '(+) AL J/ijV EXTENSION - INCHES FIG. 51. Ultimate Tension Test on Muntz Metal. EXTENSION - INCHES FIG. 52. Tension Tests on Aluminium. has to be run back, the maximum load which the specimen maintained being noted. When the specimen is exhibiting these qualities, it is said to be in the " plastic stage." Eupture may occur without again increasing the load, though in some cases the load has to be run forward again before the specimen finally breaks. The characteristic properties of a material after elastic breakdown, and before fracture, are exhibited by the stress- strain curve, many examples of which will be found in this book. Such curves are, of course, either obtained auto- graphically or from observations plotted from the readings of the load (read on machine), and the extension (read on the extensometer) in the manner previously described. Figs. 51 Fractures of Wrought-Iron and Mild Steel in Tension. Fractures of Cast-Iron in Compression. Fractures of Cast-iron Specimens in Pure Torsion. PLATE II. PLATE III. METHODS AND RESULTS OF TESTS ON MATERIALS 87 and 52 show typical examples obtained by the author during tests on non-ferrous metals. Observations after the Test. After the specimen is removed from the machine the dimensions at the point of rupture should be taken, so that the reduced sectional area may be calculated. The two halves of the specimen should also be fitted as accurately as possible together, and the new distance between the two gauge points measured. Results. If Mean original cross-sectional area of specimen = AI. Gauge length of specimen = LI. Cross-sectional area at point of rupture = A 2 . Extended gauge length = Lg. Yielding load on specimen (i.e. the load at yielding point) =" Wi. Ultimate load on specimen = W2. Then the percentage elongation 2 ~ ^lOO. The percentage reduction in area 2 ~ ^lOO. Stress at yield point on original area -r- 1 . Ultimate stress or breaking stress on contracted area -p A2 Usually the units employed are the inch and the pound, and the stresses are therefore calculated in pounds per square inch. Characteristics of Rupture. Mild steel and good wrought- iron show much contraction at the point of rupture. Stronger steels are less ductile. Brass is a very ductile material, and exhibits a silky section when broken (see Plates II. and III.). Distribution of Extension. Let the gauge length of the specimen be divided into equal divisions before testing. After rupture has taken place the extension in each division is measured, and plotted as ordinates on a base line representing the equal divisions on the specimen. Fig. 53 shows the curve obtained from a ductile material. The distribution of extension 88 A HANDBOOK OF TESTING MATERIALS w S ^ o g S J< S I 3 B s fl 1 M % "3 1 ^^O^O^^n p-j wlEs-^ i c , * | < PH .= g S w S 6 The table gives the results obtained on five different materials, and may be taken as typical of the method of recording commercial tension tests. It will be seen that the loads usually recorded are those noted at the yield point and the maximum load. The other measurements made will be seen from the table. cS 3 5 i 03 O 1 Reduction of Area. Zl 8 8 8 8 I 1 1 8O CO O CO CO Ol Ci C/5 r t CN CO i 1 CM Reduced Area. 1 V9M* |l 5 1 CO CO 1 1 3 1 I Extension. 21 s s s ON CN CN 3 CO ^J 1 S =5 g s g5 Ai *1 -- N / X -x^ 7 "^*-- __ 2 , ! PIG. 53. Distribution of Extension. due to imperfect clamping in the shackles, as this would materially distort the result. The best way to ensure perfect alignment is to employ the ball and socket tension joint already described. TYPICAL EESULT CAST-IRON. Mark. Diameter. Area. Maximum Load. Maximum Stress. 1 0.1. -750 -442 5-94 tons. 13'43 tons per sq. in Compression Tests. The specimens for compression tests are generally short cylinders, a very usual size being 1J inches long by f inch in diameter, as shown in Fig. 54. Whatever may be the size of specimen chosen, however, it should be taken as a general rule that when only the ultimate strength is to be determined the length should not 90 A HANDBOOK OF TESTING MATEEIALS exceed two to three times the diameter. If this ratio of length to diameter be exceeded, bending of the specimen may occur, so that its failure may be due, not to com- pression alone, but to a combination of compression and bending. For purposes, other than ultimate resistance tests, specimens of 1 inch diameter and 10 or 20 inches in length FIG. 55. Short Ductile Specimen in Compression. FlG. 54. Typical Compression Speci- men. FIG. 56. Ball and Socket Joint. may be used. Great care must be taken to ensure axial loading. The Test. The machine is fitted up for compression tests, and the specimen, after its length and diameter have been carefully measured, is inserted between hardened steel plates. It is preferable to have one of these plates fitted with a ball joint (Fig. 56) as then the pressure is distributed evenly over the two faces of the specimen. The load is applied, and increased by equal increments, the length of the specimen together with its mean diameter being measured in each case. METHODS AND EESULTS OF TESTS ON MATEEIALS 91 This mean diameter is a somewhat indeterminate quantity. An easier method of obtaining this dimension is as follows : In Fig. 55 let ai and li be the mean diameter and length respectively of the specimen before testing. Let a 2 and Z 2 be the new dimensions after a certain load is applied. Then since the volume of the specimen remains approximately constant, Hence 2 Consequently all that it is necessary to do after each new load is applied is to measure the new length, from which the new mean diameter can be obtained from the above simple equation. The length of the specimen can be measured by a scale and vernier without removing the specimen from the machine. Owing to the fact that the specimen continues to compress for some time after each load is applied, it is important that a stipulated and definite time should elapse between the application of each new load and the corresponding reading of the new length. A typical table of readings is given below : SPECIMEN MILD STEEL. Dimensions Diameter, f '' ; Length, Load in Ibs. Total Compression. New Length, li in inches. New Area, , = Real Stress Load,, -^-Ibs.persq.in. 000 1-500 4-12 6,000 006 494 444 13,520 12,000 013 487 446 26,900 15,000 016 484 447 33,550 18,000 057 443 459 39,220 24,000 098 402 472 50,750 30,000 167 333 497 60,350 36,000 251 1 -249 531 67,800 42,000 343 1-157 573 73,400 51,000 478 1-022 648 78,750 57,000 523 0-977 678 84,200 66,000 662 0-838 791 83,500 92 A HANDBOOK OF TESTING MATERIALS Curves plotted from these figures are shown in Fig. 57. It will be observed that yielding takes place at a load of about 15,500 Ibs., giving a stress on the specimen of 34,500 Ibs. per sq. in. Appearance of Specimen. As the tests on the mild steel MlL-D 5TEEL.. TEST o INCHES. FIG. 57. Real and Apparent Stress-Strain Curves in Compression (from single observations). proceeds the specimen is seen to bulge, finally assuming a barrel-shaped form, and if the load is increased to a sufficiently large extent, cracks and seams, approximately parallel to the axis of the specimen, appear as seen in Fig. 58. Wrought-iron would show very much the same character- istics, except that the cracks would probably be more pronounced. METHODS AND EESULTS OF TESTS ON MATERIALS 93 Cast-iron. If cast- iron be tested in com- pression, failure will occur in a very different way from that of wrought-iron or mild steel,since the specimen fails by shearing. To thoroughly understand this, it is necessary to inquire into the manner in which the shear stress varies on planes in- clined to the axis through which the pressure acts. FIG. 58. Appearance of Ductile Compression Specimen at Failure. Let a = cross-sectional area of the specimen. Let/ = intensity of stress on planes normal to axis. Eesolve this stress into forces parallel and per- pendicular to plane A, B (Fig. 59). Then shear force on plane AB = f.a. cos 6. B But area of plane AB = rf I ^< a sin 6 . ' . Shearing stress on plane AB = /* cos e = a sin 6 f cos 6 sin 6. From which it is seen that the shear stress is a maximum when 6 = 45. FIG. 59. Shear Stress on Cast-iron. 94 A HANDBOOK OF TESTING MATERIALS Theoretically, therefore, fracture should take place along a plane inclined at 45 to the axis. This is found to be so, or very nearly so, in practice, any deviation from theory being FIG. 60. Frac- ture of Ductile Material in^Ten- sion. FIG. 6 1 . Fracture of Cast-iron in Compression. probably due to non-uniformity in loading, local variation in the material, or possibly to internal friction. Brittle hard materials all behave in this way, while ductile plastic metals fail similarly to mild steel. Cast-Iron in Compression. If the fracture of a ductile specimen, which has been broken by application of a direct METHODS AND BESULTS OF TESTS ON MATEEIALS 95 pull, be examined, it will be found that failure has not been the result of tearing across a plane section perpendicular to the axis, but that shearing has taken place along the surface of a cone of semi-vertical angle, approximately j (Fig. 60). If a specimen of the same material be taken and tested in compression it will be found that, if the speci- men is too short for buckling to take place, there will be no definite fracture (Fig. 58). A possible explanation is that, when the specimen is short enough for buckling not to occur it is too short for shearing to take place in one fracture. There will therefore be an internal crumbling along numerous shearing planes, which in a plastic material may not exhibit itself as a separation of the specimen into so many pieces. In a brittle material (Fig. 61) the shearing fracture produced by direct compressive stress is very marked. So far, then, we may say that in at least ductile materials, directly stressed, shearing is the governing factor. Let us examine how such a stress arises and what its magnitude will be. Consider a specimen of unit-cross-section subjected to a steady pull of P Ibs. (Fig. 62). Then on any section perpendicular to the axis there will exist a tensile stress numerically equal to P. On this plane there will be no tangential stress since P can have no component at right angles to itself. On a plane FIG. 62. Eesolution of Forces in a Specimen sub- jected to Tension. 96 A HANDBOOK OF TESTING MATERIALS section inclined to the axis there will, however, be in general two stresses, one normal and the other along the plane of the section. Consider the equilibrium of one of the portions cut off by the plane of section. We have along the axis a stress = P, along the plane a shearing stress, say, q, and perpen- dicular to it a normal stress, say r. These three must equilibrate. Eesolving vertically we have r . sin 0+ -3 cos 6=P or r-\-q cot 0=P ... (1) sin sin 6 Eesolving horizontally we have cos e=-2 sin GOT q=r cot B ... (2) sin & sin whence r = P sin 2 <9 (3) and q = P sin cos 6 ...... (4) p This last equation may be written as q= ^ sin 20. The maximum value of sin 20 is 1 and occurs when 0=^. Hence we see that a direct stress induces shear on planes inclined to the axis and that this shear stress is a maximum on planes inclined at 45 to the axis, or their envelopes (cones of semi-vertical angle ^). Its value is then half the maximum direct stress. Therefore if no disturbing factors enter into the question a directly stressed specimen should, if the shearing stress is the determining factor, shear along a surface inclined at 45 to the axis. The symmetrical surface fulfilling this condition is a cone of semi-vertical angle =2- This fracture allows separation in the case of a tension specimen, but not so in the case of compression. A specimen fractured by direct push therefore separates along a plane surface. It has already been indicated that the measurement of this angle of yield for a ductile specimen, subjected to compression, presents very considerable difficulties. In the case of tension the difficulties, though less manifest, are none the less real. The conical fracture is very apparent, and the measurement METHODS AND EESULTS OF TESTS ON MATERIALS 97 of the angle is easy ; it must, however, be remembered that the angle which can be measured gives only an approximate indication of the form the conical surface assumed when the material yielded, and was virtually destroyed. When the yield point is passed the material offers no permanent resistance to the application of the force ; it is, in effect, a liquid, and the increase of the force merely shortens the time occupied in viscous flow, and until actual surface separation must take place. The plastic drawing out will deform the surface along which yield takes place and which probably also forms the surface of separation. In a ductile material it is, therefore, as has been seen, a matter of no little difficulty to determine exactly the surface along which yielding occurs. In a brittle material under compression the problem is considerably simpler, and its consideration may throw some light on what really happens within a stressed material. In a specimen of cast iron the fracture is perfectly definite, and, as no plastic flow occurs, it may be assumed that the surface of yield coin- cides with the surface of separation. Therefore if the maximum value of the shear stress is the only determining factor, the fracture should take place along a plane inclined to the axis at exactly 45. This is never found to occur ; the angle 6 is consistently less than 45, and a reason has to be sought. Let us examine the following results obtained on a specimen of cast iron. Diameter '727 inch. Length = 2*125 inches. Breaking load = 19'25 tons. 33. Maximum shear stress on plane where 6 = 45 = 23'2 tons per sq. inch. Direct stress on this plane = 23'2 tons per sq. inch. Shear stress on plane of fracture = 21*2 tons per sq. inch. Direct stress on plane of fracture = 13'7 tons per sq. inch. T.M. H 98 A HANDBOOK OF TESTING MATERIALS It will be seen that fracture took place along a surface where the shear stress was 2 tons per sq. inch below, or, roughly, 10 per cent, lower, than the maximum. It will, however, be noticed, that the normal stress on the breaking section was considerably lower than that on the section where the shear stress was a maximum. This immediately suggests the fact that a direct push between two surfaces increases the resistance to their shearing or sliding over one another. The effect, in fact, is very akin to friction, and a theory, usually known as Navier's theory, has been developed on these lines, and is as follows. Let the true shear resistance, when no normal stress is exerted between the surfaces, be f. Then if r is the normal stress on the section, we may suppose that the actual shearing resistance offered (q) is of the form q=f+n r ; but q = P sin d cos 0, and r P sin 2 0. /. / = P (cos 6 sin d - p sin 2 <9). Fracture will take place across the section where / is a maximum, i.e., where -j^=0. du df ^jj==P cos Q (cos d n sin 0)+P sin (sin 6 /* cos 0} = 0. . ' . cos 2 sin 2 2 n sin 6 cos 0. i.e., /x=cot 20. If 4> Bangle of friction, then tan =cot 20. 20+>=~, and =j-* Taking the particular case where 6 33, we have = 24 or ju = *45 approximate. Whether this theory is at ail justifiable is for future research to determine. The following values have been obtained from compression tests on cast-iron specimens. It will be a good exercise for the student to measure the angles of fracture and evaluate the co-efficient /x. It should be remembered that discrepancies may be due to (1) non-homogeneous material ; (2) the fact that yield point and ultimate fracture loads are not neces- sarily co-incident. This may account for the varying results recorded in the following table. METHODS AND RESULTS OF TESTS ON MATERIALS 99 TABLE II. COMPRESSION TESTS ox CAST-!ROX. D in inches. Liu inches. W in tons. *-I in tons per sq. inch. Q in degrees. / fin tons ' per sq. inch. 916 1-662 37-09 56-3 29 625 15-6 752 1-470 22-15 49-9 34 404 16-8 966 1-675 41-65 56-8 33 445 18-5 904 807 26-89 41-9 30 577 12-1 815 760 28-61 55-0 31 532 16-5 943 652 27-33 39-1 28 675 10-4 816 762 26-92 51-5 28 675 13-6 961 663 33-^6 46-1 28 675 12-2 Autographic Diagrams. An autographic diagram gives much fuller information as to the behaviour of a metal under test than does the mere breaking test. Not only does it give all the results obtainable from a test to rupture, but in addition the extension of the bar at any load can be easily obtained from it, together with the work expended in breaking the bar. Before taking the diagram the necessary measurements of gauge length and cross-sectional area of the specimen are made. The specimen is clamped in the grips and the autographic apparatus fitted. The load is then run on slowly, the beam being kept floating by means of the pump. Great care should be exercised to prevent the beam touching the stops, otherwise the diagram produced will not be a true one. This difficulty will be particularly pronounced at the yield point and near the point of rupture. At the yield point, owing to the comparatively rapid change in length of the specimen, the pump has to be worked continuously with little or no increase in load. Nearing the rupture point with ductile materials the load has actually to be run back owing to the rapidly diminishing cross-sectional area of the specimen. It was stated previously that the load was run on and the beam kept in equilibrium by means of the pump. This, in a sense, is hardly true when the pump is worked from the shop H 2 100 A HANDBOOK OF TESTING MATERIALS shafting. When such is the case it can only be operated at some definite speed, so that the beam is kept floating by the rate at which the load is applied. Fig. 63 shows a typical diagram for mild steel plates. If the cross-sectional dimensions of the specimen be 3 fi 2 - o Q J ^ V MILD STEEL PLATES TENSION TEST 05 25 JO -;5 -20 TxTT/S/S/OA/ - INCHES FIG. 63. Mild Steel Plates in Tension (Autographic Diagram). measured at definite loads during the tests, a means is given of determining the actual stresses on the reduced section of the bar. A second curve plotted from results obtained thus is shown in Fig. 64. Raising the Yield Point. The autographic diagram affords a very convenient method of verifying the statements regard- ing the raising of the yield point by a process of repeated loading. The phenomenon is as follows : Let the load be METHODS AXD RESULTS TES'JJS j^T/MljBJSiALS 101 carried a little beyond the initial elastic limit, and allowed to remain so for a certain time. On removing the load, and then again running it on, the specimen will be found to yield not MILD STEEL Autographic Diagram. Mean D!a.=-90B" Gauge L ength -8* Elongation inches. FIG. 64. Curves of Eeal and Apparent Stress. at the initial or primitive elastic limit, but at about the load originally applied to the specimen. By such a process the elastic limit may be raised until finally it coincides with the breaking load of the specimen. This is shown in Fig. 65. 102 4-8 VNI)BQOK OF TESTING MATERIALS This phenomenon of raising the yield point by repeat loadings is but one of a number which can be conveniently investigated by the student. For a full discussion of the change of elastic properties by mechanical and heat treatment the student should refer to proceedings of scientific and technical societies. It will suffice to give here a few typical examples together with Extension - Inches FIG. 65. Autographic Diagram of Mild Steel in Tension, showing Effect of removing temporary. the corresponding stress strain curves obtained by the author and his students. The Elastic Range. The theory was enunciated by Bauschinger that a change in the elastic limit follows the extension of a specimen, and that if mild steel has an elastic limit of, say, 13 tons in a tension test, and the same value in a compression test, then if the material be overstrained in tension until the new elastic limit is raised to, say 16 tons, the olaetic limit in compression will be 10 tons. In other words, METHODS AND EESULTS OF TESTS ON MATERIALS 103 the elastic range of the material is 26 tons. In July, 1908, the author made tests. Stress distribution was not taken into account, and the results are therefore not within an accuracy of 5 per cent. The first elastic limit was 13 tons. By increasing the tension elastic limit 2*7 tons, the compression elastic limit was lowered 2'5 tons. A full account of these 25 20 Specimen HTS I. Tension Test. A. Primary Test. B. 35mi'ns.afterA. C. 7 days after B. D. 20 m ins. after C. L^/lfier 10 m/ns. at 100" C, F. Se/ays after E. G. After /O mins. at 100" C. FIG. 66. Effect of Time and Low Heat Treatment on Mild Steel in Tension. tests considered unsatisfactory for various reasons set forth are to be found in the Journal of the Institution of Junior Engineers, July, 1909. Mr. Leonard Bairstow, shortly afterwards, published in Yol. OCX., Series A, of the Philosophical Transactions of the Royal Society a valuable contribution. He noticed that fatigue was able to produce slow yielding whenever the compressive and tensile stresses were not equal, even though the maximum stress applied was considerably below the yield stress. 104 A HANDBOOK OF TESTING MATERIALS Time Effect. If a mild-steel specimen be tested up to just beyond its elastic limit, and, after removing the load, allowed to stand, it will be found to recover its elastic properties with time, and on again testing, the elastic break-down point will be found to be raised. Effect of Low Temperature Heat Treatment. This same recovery of elastic properties is even more marked if the A. Primary Test B. 3 Weeks after A C . dfter 15 mins. at 100 C FIG. 67. Effect of Time and Boiling on Mild Steel Specimen in Compression. specimen be boiled at 100 C. for a short time. Both time and heat treatment effects are shown in Fig. 66. The specimen was a sample of high tenacity steel of remarkably uniform properties. The data in connection with the method of treatment is given on the curve. A similar phenomenon is observed both in compression and tension tests. Fig. 67 shows a typical curve obtained with mild steel in compression. Effect of High Temperature Heat Treatment. This phenomenon is discussed in Appendix IV. The curves shown in Fig. 68 are those obtained from the same material but subjected to various high temperature heat treatments. It \L 30 BN 30 BA 50 300 00 150 200 250 EXTENS/ON (UNIT = -0001 INCH) FIG. 68. Curves showing Effect of Heat Treatment on Bessemer Steel. 350 106 A HANDBOOK OF TESTING MATEEIALS will be observed that there is a great change in the elastic properties of the Bessemer steel thus treated. Eeferences to the tests will be found on p. 260. Mechanical Hysteresis. If a specimen be loaded up past a certain point (in general below the elastic limit) and then 2,000 80 120 160 200 240 280 I - ORIGINAL STRESS-STRAIN CURVE.DEC. 7TH 1901* II - IMMEDIATELY AFTER NO. 1. II - JANUARY 15TH 19O2 IV .FEBRUARY 23RD 19O2 FIG. 69. Curves showing Mechanical Hysteresis, the load be decreased and readings of extension be taken both during increasing and decreasing load we get the phenomenon known as mechanical hysteresis. That is to say, the two curves for increasing and decreasing load do not coincide but enclose area as shown in Fig. 69. Professor Coker has made many interesting researches on this phenomenon, 1 and 1 On the effect of low temperature on the recovery of overstrained iron and steel. Physical Review, Vol. XV., 1902. METHODS AND EESITLTS OF TESTS ON MATEKIALS 107 the curves reproduced are from some of his results obtained on mild steel. His object was to determine the effect on mechanical hysteresis of time. It will be seen on comparing the curve in Fig. 69, and the data given immediately below it, that the effect of time is to increase the hysteresis. This hysteresis is probably due to friction between the particles of the metal moving relative to one another under strain. Further data as to the effect of time, etc., will be found on consulting the proceedings of the engineering institutions. Some curves are also given in Chapter VI. showing these phenomena as exhibited with torsion specimens. Modulus 'of Elasticity, E. The modulus of elasticity, or Young's Modulus is the ratio of stress to strain within the elastic limit. Suppose a specimen be tested to the elastic limit and by some means a curve is obtained with the extensions of the specimen as abscissae, and the loads causing these extensions as the ordinates. load x , -r, stress sectional area of bar Now E=-r = - strain extension original length load original length "extension sectional area . ' . E = slope of the load-strain line multiplied by a constant. From which simple equation E can be found. The load-strain, or as it is somewhat loosely called, the stress-strain line, is obtained by some form of extensometer. For such work round specimens are almost invariably used, as they can be turned and measured very accurately. The Experiment to Find the Yalue of E. The specimen is marked off for the reception of the extensometer, the gauge length and cross-sectional area being accurately determined. The extensometer is then clamped on to the specimen, the latter being placed in the testing machine, and a small load applied to steady the whole. The load is then run on in equal increments, measurements of the extensions of the specimen being taken in each case by the extensometer. 108 A HANDBOOK OF TESTING MATERIALS When the maximum load is reached, the load is reduced by equal decrements, readings of the extensometer being taken as before. A curve is then plotted between the loads and extensions, both for ascending and descending values of the load, Such curves for mild steel are shown in Fig. 70. The curves are straight lines, proving that within the elastic range the stress is proportional to the strain. MILD STEEL. Gftuce LENGTH * Seer. &f*&* 7" ool -ooa -c EXTENSION IN INCHES FIG. 70. Mild Steel in Tension. The slope of curves = tan 6 = . = 1,345 inch ton units. The cross-sectional area of the specimen = "768 sq. in. The gauge length = 8 inches. '.' E = l, 845X^ = 14,000 tons per sq. in., or 31,400,000 * / uo Ibs. per sq. in. METHODS AND EESULTS OF TESTS ON MATERIALS 109 The following tests were made with the Ewing extensometer, and the tables indicate how results of such tests should be set out. Specimen I. Brass Rod. Observations. Distance between gauge marks on specimen, 8 inches. Leverage of machine, 22*4. Value of 1 division on microscope scale, 0*0002. Diameter of specimen, '3755, '875, '375 inches. Mean diameter of specimen, '375 inches. Load. Lbs. Ktading on Scale. (Loading.) Readings on Scale. (Unloading.) Extensions. Inches. 10 20 30 40 50 60 70 80 90 95 21 26 32 36 42 47 52 57 62 68 21 26 31 36 41 46 51 56 61 68 0010 0011 0009 0011 0010 0010 0010 0010 0013 85 Totals. 0094 Specimen II. Hard Steel Rod. Observations. Distance between gauge marks on specimen, 8 inches. Leverage of machine, 22*4. Value of 1 division on microscope scale, 0'0002 inches. Diameter of specimen, '358, *356, *356 inches. Mean diameter, '357 inches. 110 A HANDBOOK OF TESTING MATERIALS Load. Lbs. Reading on Scale. (Loading.) Reading on Scale. (Unloading.) Extensions. Inches. 10 20 30 40 50 60 70 80 90 95 100 28 31 34 37 40 43 46 49 52 53-5 55 28 31 34 37 40 43 46 49 52 53-5 55 0006 0006 0006 0006 0008 0006 0006 0006 0003 0003 90 Totals. 0054 Specimen III. Mild Steel. Observations. Distance between gauge marks on specimen, 8 inches. Leverage of machine, 22*4 inches. Value of 1 division on microscope scale, 0'0002. Diameter of specimen, '374, '374, '374 inches. Mean diameter, '374 inches. Load. Lbs. Reading on Scale. (Loading.) Readinc on Scale. (Unloading.) Extensions. Inches. 10 3-9 3 '75 00055 20 4-2 4-00 00055 30 4-4 4-35 00070 40 4-8 4-65 00060 50 5-1 4-95 00045 60 5-25 5*25 00040 70 5-4 5-5 00060 80 5-7 5-8 00050 90 6-0 6-0 00070 100 6-35 6-35 90 Totals. 00505 METHODS AND KESULTS OF TESTS ON MATERIALS 111 Specimen IY. Wrought-Iron Rod. Observations. Distance between gauge marks on specimen, 8 inches. Leverage of machine, 22'4. Value of 1 division on microscope scale, 0*0002 inches. Diameter of specimen, '381, '389, *385 inches. Mean diameter, *385 inches. Load. Lbs. Reading on Scale. (Loading.) Reading on Scale. (Unloading.) Extensions. Inches. 10 28 28 0005 20 31 30 0005 30 33 33 0006 40 36 36 .0004 50 38 38 0006 60 41 41 0005 70 43-5 43-o 0005 80 46 46 0005 90 48-5 48-5 80 Totals. 0041 COLLECTED BESULTS OF EXPERIMENT TO FIND THE VALUE OF YOUNG'S MODULUS. Specimen Number. Load x Leverage. Area. Length. Exten- ' sion. Strain. Stress. E. Ibs. sq. in. inches. inches. Ibs. sq. in. Ibs. sq. in. 1 . 1,904 Ill 8 0094 00118 17,190 14-55 X 10 6 2 2,017 101 8 0054 00063 20,000 29-4 X 10 6 3 . 2,017 no 8 .00505 00063 18,310 29-1 X 10 6 4 . 1,791 117 8 0041 00051 15,350 30-1 X 10 6 Determination of the Modulus of Elasticity by Bending. The value of the modulus of elasticity can also be obtained by the method of bending. Suppose, for example, it is required to find E, for a wrought-iron girder. The testing machine is arranged so that the girder to be tested is placed across the knife-edges, which are fixed a definite distance apart. The load is then applied in the usual way to the specimen by the central knife-edge, the maximum deflection being read by a microscope and scale affixed to the girder. 112 A HANDBOOK OF TESTING MATERIALS ii I ShH^-g 'i -s s "o ^ h rl a * . CH -+3 ^ ill *-: METHODS AND EESULTS OF TESTS ON MATERIALS 113 c 5 ^ . |1* CO C^ Ci ^ ^-JH CO CO 1^ c -M S O >O O5 O CO S g 1 - 1 x Is O > i ^i o-i co co ^< Tf Th o 'O o co co i~~ t^- !jg , . T- 1-^ CO CO O CM 'O GO i I -P GO i 1 >O O5 CO GO 3 ^= x iOCOCOCOtr"'.t^r-GOC>0 w g B ll| oSg^^^gs^S^SS^i^Si > "1 BQ - 5 s W CO i r 3 | i A A s i 11 J-r-(l-TflG^'rHC^:OTftOl OOOi' IfM't* J) M i i i i 1 7^7 ^ 4 " p< oi >o co i i i GO r- Hi +a 53 Ci CO 05 >0 t^ C<1 .>0 CO CO CO C ^OOCXDt OOi'MCOOrHCiCOGOTtiOl^ fl 1 & i s 8 C> fl J^ 6 o H.S s H S ^ PJ 3 C g js s. .2.2 .2 M a) c' .2 *2 -S i III PH ^ ^ > ^ IL'I '^ SS M |- S 2 Tl | ^ 1 III ,1 ^_^ o OCDCO CN >O O 1^- O5 O C*O i 55 i CO S ^ 5 a 1 M 'ari Ho' M W C3 1 1 1 1 1 1 1 1 1 1 1 o O P O ^3 T-H g-gSi cec^THoeoocooooopoooo ^^OCOOCOCOCOCO^THl-OCOCOO^CO fc ! O 3 00 2 ir 1 1 H B 00 -3 II CH c III ^ M 8^SS83aS88 M 8 1 i. OQ S- S 3 W CO EH ^1 bb .2 5 r- ( i I rH i I i-H II 1 1 1 1 'S P *l ~? M ** 9 - 1 O^^CO^OOCOCOCO^^GOGOCOOiOS C M. 114 A HANDBOOK OF TESTING MATERIALS A succession of readings are taken for loads varying from zero to a load well within the elastic limit, and a curve plotted with the deflections as abscissae, and the loads as ordinates. The deflection of a bar supported at the ends and loaded centrally is given by the expression W/ 3 y = Deflection. W Load applied. I = Distance between knife-edges. E = Modulus of elasticity. I = Moment of inertia of the cross-section of the girder, from which ~7'48I I 3 ~4ST X s ^P e load-deflection curve. The cross-sectional dimensions of the girder are taken, and the moment of inertia calculated. A simple method of obtaining a diagram of the section is to smear the latter with red lead, and obtain an impression of it on a piece of paper. The distance between the knife-edges is known, and the slope of the load-deflection line measured, and consequently E can be calculated. Approximate values of E for different materials: Material. E. Lbs. per sq. in. 1 Cast-iron 17,000,000 Wrought-iron bars 29,000,000 Steel boiler plates . 30,000,000 Steel plate (mild) . 31,000,000 Cast-steel (untempered) . Copper rolled plate 30,000,000 15,000,000 Brass .... 13,500,000 Gun -metal or bronze 13,500,000 Phosphor bronze 14,000,000 Wood (pine) . 1,600,000 Wood (oak) . ... 1,450,000 1 Unwin's Machine Design, Part I, METHODS AND EESULTS OF TESTS OX MATERIALS 115 The values of the moduli must necessarily depend on the qualities of the materials, so that the values given above must only be taken as approximate. Testing with the Sphingometer. As explained briefly in Chapter IV., page 66, the Sphingometer can be used, not only to determine direct extensions, but also to determine the stress distribution in the specimen. Normally one measures the extension in three planes at 120 and the mean value of these three measurements is used for the calculation of E. The Table on page 113 shows how the readings are set out and calculated. If the readings of each strip be plotted separately we obtain the irregular curves shown in Fig. 71. The mean curve, however, in the case of an elastic specimen is straight. This latter line gives all the necessary information for obtaining the value of E, etc., while the separate curves are used to determine stress distribution in the manner explained in Appendix III., page 249, where a test on mild steel is worked out. The readings taken in three planes demonstrate that the load does not pass through the axis of the specimen. It is especially interesting to test a specimen, and obtain readings with the instrument, with Yee-grips in the testing machine. i 2 CHAPTEK VI TORSION TESTING Brittle Materials in Torsion. When brittle materials, such as cast-iron, are subjected to torsion, a fracture, usually almost perfect in form, results. It is inclined roughly, at an angle of 45 to the axis of the specimen, .and makes a complete revolution of the bar, the junction of the ends of the spiral being approximately a straight line. The theoretical line of fracture is illustrated in Fig. 72. Fig. 73 shows a r\ 1 1 I / v i r v^ 1 f----> j i / FIG. 72. Cast-iron under Torsion. drawing of a hollow specimen fractured in torsion, and photo- graphs of actual fractures are shown in Plate III. It can be readily seen why rupture occurs in this peculiar way if an elementary square on the surface of the specimen be con- sidered . The torque on the specimen introduces shear forces q (Fig. 72) on the faces of the elementary square as shown. But as equilibrium is maintained, it is evident that there must be equal and opposite shear forces q on the other two faces It can now be readily proved that this brings into action a tensile stress of equal intensity on the face inclined to the others at 45. Now, cast-iron is weaker in tension than in compression or TORSION TESTING 11' shear, and will consequently give way along that surface where the stress is tensile. If this tensile stress be calculated, it will be found to approximate closely to the breaking stress found by pure tension. The calculation is performed as follows : Twisting moment = q Z. Where q = shear stress produced Z = modulus of the section. But Z for a round bar ~YQ~ 16XTM but tensile stress f t q 16XTM The following results were obtained from hollow cast-iron specimens : FIG. 73. Fracture of Cast - Iron Hollow Specimen in Torsion. No. of Specimen. Internal Diameter. External Diameter. Torque. ; Tensile Stress. Internal Ibs. Lbs. persq.in. Angle of Fracture. Internal. Angle of Fracture. External. I 1 904 1-289 6,020 19,000 45 49i 2 892 1-125 2,920 17,500 45 49 3 990 1-249 3,180 13,850 47 45 Ductile Materials. The behaviour of ductile materials in torsion is very different from that of brittle materials. The specimen twists considerably and fracture, being due to shear, takes place in a plane approximately perpendicular to the axis. Fig. 74 shows a load-strain curve plotted for mild steel. As the angle through which the specimen twists is large, it can be obtained sufficiently accurately, as explained later 1 , by observing the number of turns of the hand- wheel actuating the torque. Then, knowing the number of teeth on the wheels brought into play, the angle of twist can 1 See description of the Bailey machine, page 119. 118 A HANDBOOK OF TESTING MATERIALS be calculated. The actual readings of the mercury column are plotted as ordinates. Then since the torque arm in this case is 4 inches long, the torque on the specimen can be obtained by multiplying the mercury column readings by four. It will be observed that the diagram is not dissimilar to those obtained from tension or compression tests. There is first the elastic period, giving a straight line, then a distinct TEST M/LO STEEL 5PC, 650' FIG. 74. Torsion Test on Mild Steel (Autographic Diagram). yield point, followed by a curve of the shape usually met with in autographic diagrams. Mild steel being a very ductile material twists considerably before rupture, and the curve shows a total twist of 1250 degrees. Wrought-iron exhibits the same characteristics, though not to so large an extent, although the cracks and markings are much more noticeable. Gunmetal breaks more quickly, the surface presenting a a very uneven and blotched appearance, owing to the effects TORSION TESTING 119 of compression in some places and tension in others, as shown in Fig. 75. Tests on cast-steel show that its behaviour is intermediate between that of cast-iron and the ductile metals, approaching more nearly to either extreme according as the steel is hard or soft. TORSION TESTING MACHINES. The " Bailey " Torsion Machine (Fig. 76). The specimen consists of a cylindrical bar with enlarged ends, which are either of square section or are fitted into square caps by means of keys. The object of the test is to find the resistance of the bar to torsion or twisting, so that it is not necessary to grip the bar tightly in the clips, but only to prevent it from rotating in them. In fact, it is necessary to allow the bar to slide a little longi- tudinally, as when twisted it becomes rather shorter than before. Two views of the machine are shown in Fig. 76. The end AI of the bar is twisted by means of the hand-wheel B, which turns the worm-wheel C. This in its turn rotates the spur-wheel D, to which the grip AI is rigidly attached. The other end of the bar, held in the clip A 2 , attempts to turn with AI, and with that intent pulls at the lever E, which is connected by the tie rod Gr to the mercury diaphragm 1 F. The pressure on this diaphragm 1 In the usual type of machine this diaphragm, is made of rubber- Prof. Hummel states that he finds a great improvement by substituting a thin brass diaphragm in this machine. PIG. To. Stresses induced in a Bar sub- jected to Pure Torsion. 120 A HANDBOOK OF TESTING MATERIALS TORSION TESTING 121 causes the mercury to rise in a column H, which balances the pressure due to the tie rod G. Thus the height of the column H gives a measure of the torque or twisting force exerted on the specimen. The mercury column is calibrated to give the tension in the tie road G, and, since the length of the arm E is known, the torque on the specimen is easily found. "Thurston" Torsion Machine. In this machine the load is applied in the same way, but the method of measurement is different. A pendulum weight is 'affixed to the free end of the test bar, and, as the load is applied, the test bar is able to move this pendulum through a distance proportional to the load. The pendulum, therefore, is caused to move a pointer along its quadrant, the latter being calibrated to read the load on the specimen. When the bar breaks, the pendulum swings back into a vertical position, but the pointer remains in the position that it had assumed just before the specimen broke. The final position of the pointer, then, gives us the breaking torque on the test bar. "Avery" Torsion Machine. The Avery machine also has its load applied in the same manner as the Bailey and Thurston machines, viz., by a worm and worm-wheel, but, again, the method of measurement is different. The stress is indicated by means of a system of weighing levers, similar to those in the Kiehle testing machine, being finally measured by running out a poise along a graduated steelyard. In the 15,000-inch Ib. machine there are three of these poises, each weighing 60 Ibs. The first indicates up to 5,000 inch Ibs., when each scale division on the steelyard represents J Ib. When two poises are coupled together readings can be taken up to 10,000 inch Ibs., and when the whole 180 Ibs. are run along the scale they give a total capacity of 15,000 inch Ibs., with readings of 5 inch Ibs. per scale division. A vertical scale is sometimes fixed on the front end of the steelyard, and a telescope to the frame of the machine. By this means we can adjust the lever with great accuracy until it rests in a perfectly horizontal position. This machine will take square bars up to J inch side, or rectangular specimens up 122 A HANDBOOK OF TESTING MATERIALS to a maximum size of 1 inch X f inch. The bracket on which the straining gear is fixed is capable of movement to admit specimens of a maximum length of 15 inches. The shortening of the specimen under the torsional load is provided for by the insertion of hardened steel rollers. The actual strain on the specimen is observed by fastening indicating arms to its two ends a gauge length apart. These arms are in line with each other at the start, but as the. load is applied one end gets twisted more* than the other, so that the angle between them at a definite load gives us the torsional strain at that load. This machine is also arranged to measure a torsional stress applied in the reverse direction. The main torsion lever T is keyed on a sleeve which is free to revolve in ball bearings. An intermediate lever E is arranged within this main lever, and is pivoted at a point G between the axis of the specimen and that of the tension rod which transmits the stress to the steelyard. When the torsion is applied in a clockwise direc- tion, the knife-edge C of the main lever pulls up the left hand end of the intermediate lever, and so depresses the end that is attached to the tension rod. When the stress is applied in a contra-clockwise direction, the knife-edge K raises the point E of the intermediate lever, and again pulls the tension rod downward. So that, in whatever direction the torsional load is exerted, the short end of the steelyard is always pulled downward, balance being restored by running the poise along the arm. The leverages are so arranged that the load on the specimen, in either direction, can be read directly on the same scale. The same makers also manufacture a testing machine to give results in tension and torsion simultaneously, so that the effect of the combined stresses can be read off in one machine. It consists practically of a hydraulic tension machine and a hand-power torsion machine on the same bedplate. The principles upon which the combined machines act are similar to those already given for the separate machines. A detailed account of the tension-torsion tester TORSION TESTING 123 Tension Rod- ft \ Fig. 77. Hand-Torsion Testing Machine. 124 A HANDBOOK OF TESTING MATERIALS is given on page 251. The two steelyards, one for each kind of stress, are placed so that neither interferes with the working of the other, but are near enough together to enable the readings to be taken by one man. The Torsion Sphingometer. The principle of the twisted strip has been applied to the instrument which is used for recording torsion strains. To each of the carriers " C ' : (Fig. 40) is fastened an arm carrying a " V " block. The sphingometer tube is now placed at right angles to the axis of the specimen. A 45 mirror is placed vertically under the mirror of the sphingometer strip, and is so hinged that it will move in two planes. This latter arrangement makes it easy to arrange for illumi- nating both mirrors. Any move- ment of the carriers "C"due to a torque will cause the sphingometer strip to extend. This extension is magni- fied and recorded in the usual manner. The method of calibration is by using a gauge to measure the perpendicular distance of the strip from the circumference of the specimen. Then since the specimen is circular, the distance of the centre of the strip from the centre of the specimen is also known. The proof of this is as follows : Let B C (Fig. 78) represent the strip of length Z, and let A be the centre of the specimen. Then from the figure we have / 2 i=c 2 +Z> 2 2fcc cos A. Whence 2Z2cirpen ADI. TORSlOn TfcST A... Primara fesf. B . . . After 10 mSofcS aT lOcTC C... Imrpe&iafelg afferD. ...4daQ3 after C. _.. ImnTcSiofelu after D. fr..... After 10 mirufea ar FIG. 81. Effect of Time and Boiling on Mild Steel. TOESIOX TESTING 127 TORSION TESTS SSff ct.e./itHi t d- Primary Test e - Mer 10 minutes at 100'C h - After raising elastic limit by overstrain I - AHer 10 minules at 100'C Tw/sr FIG. 82. ---Effect of Overstrain and Boiling on Mild Steel in Torsion. Specimen 35? TOR5IOrtT5T. D... C...A(IerlOrr?ir?aar 100" C PIG. 83. Effect of Overstrain and Boiling on Mild Steel in Torsion. 128 A HANDBOOK OF TESTING MATERIALS another phenomenon in connection with the elastic break- down of materials. If after passing the elastic limit the load be sustained constant for some time, the twist will be found to increase. In Fig. 84 the time of taking the different readings is placed in juxtaposition with the plotted point. It TORSION TEST ALIa Torque arm - 38 2 inches TW/ST FIG. 84. Pure Torsion Test on Aluminium showing Time Effect. will be seen that at first the " slip " is very small even over a considerable time, while the effect becomes gradually more marked until complete breakdown occurs. A full discussion of this gradual breakdown will be found in a paper read by the author before the Iron and Steel Institute. 1 Fig. 85 shows the effect of boiling in water on the recovery of elastic properties in the case of aluminium. 1 " The Elastic Breakdown of Certain Steels," Journal of the Iron and Steel Institute, May, 1910. TOESION TESTING 129 eol TORS/ON TESTS AL4 6w C- After /O minuses at /OOC Torque arm = 38 2 inches. TW/ST FIG. 85. Torsion Tests on Aluminium. Figs. 86 and 87, on copper specimens, will show that normal annealed high conductivity copper has no very marked elastic limit, the curve gradually bending over. TORS/ON TESTS HCJ'a AND HC4'a Tw/sr FIG. 86. Torsion Tests on Copper. T.M. 130 A HANDBOOK OF TESTING MATERIALS roffs/ott rtsrs Hcs'donde /* Primary Tetf A After overstraining Torque arm * J8Z /acfes TWIST FIG. 87. Effect of Overstrain on Copper. After overstraining beyond the elastic limit, a hardening effect takes place, which causes the material to give a curve more closely allied to other materials. The curve obtained TORS /ON TESTS MS a **o b a Primary Test b /mmediate/y offer ~ Tw/sr FIG. 88. Torsion Tests on Muntz Metal. with the primary test in Fig. 87 is the one starting from a point farthest to the left. Table IV. and the curves in Fig. 88 show the arrangement of a test on Muntz metal. TOESION TESTING 131 TORS/ON TESTS MS b *HD C C after /O minufes af /00C Tw/sr FIG. 89. Torsion Tests on Muntz Metal. TABLE IV. TORSIOX TEST ON Muxrz METAL (M5a). (FOE, CURVES, SEE FIG, 88.) Diameter of specimen = 1*000 inch. Perpendicular distance of strip from centre of specimen = 3*72 inches. Calibration of strip, 1 scale division = 0'0001080 inch. Load in Ibs. Scale Readings. Scale Differences. T*n of a Radian. 20 -193 40 66 127 0-368 50 2 191 0-554 60 +61 254 0-737 70 + 125 318 0-923 80 + 193 386 1-120 85 +228 421 1-220 90 +265 458 1-330 90 -30 Inst*. reset. 95 + 10 498 1-445 100 +53 541 1-570 105 + 100 588 1-707 110 + 152 640 1-857 115 +215 703 2-040 On examining curve SSa, it will be seen that the curve ceases to be straight where load is 85 Ibs., whence for the shear stress at elastic limit, the moment of the couple K 2 132 A HANDBOOK OF TESTING MATERIALS is TxL (in test M5a, T 85 Ibs. and L=38'2 inches) we i ft have rr:=TLx FO, where q is maximum shear stress. ml 3 85 X 38-2 X 16 Whence q = 7T X 13 = 16550 Ibs. per square inch. /(? -1 >- FIG. 90. Kectangular Block under Shear Stress. The curve S8b is a test on the same specimen immediately after test M5#, plotted in Fig. 88. Fig. 89 shows the effect of overstrain and subsequent boiling on Muntz metal. FIG. 91. Circular Specimen in Torsion. Modulus of Rigidity C. The modulus of rigidity C _ shear stress in Ibs. per sq. in._^ ~ shear strain per inch length ~~ 5 ' 5 is the strain between two planes an inch apart (Fig. 90). The simplest method of finding C for any material is by applying a torque to a round specimen and observing the angle oi twist over a definite gauge length. This is quite TOKSION TESTING 133 easily done in the Bailey machine, a torsionometer being used to measure the angle of twist. A succession of readings of twisting moment and angle of twist are thus obtained, both for ascending and descending values of the twisting moment, and a curve plotted with the angles of twist as abscissae, and the twisting moments as ordinates. Careful measurements are also made of the diameter of the specimen and the gauge length. Let Fig. 91 represent the specimen of gauge length L, and the radius r. Let $ = angle of strain in circular measure, and 9 = angle of twist in circular measure. Then C=| ... (1) where q shear stress produced. But twisting moment, TM = q Z. where Z polar modulus of the section. TM . * . o = . /^ And since 0, not , is measured by the torsionometer, _r0 Substitute in (1) , TM L L . *. C ^--^-X-^^-x slope of stress-strain curve. All these quantities are known, and consequently C can be calculated. Approximate values of C for different materials. Material. C. Lbs. per sq. in. Cast-iron 6,300,000 Wrought-iron bars . . . 10,500,000 Steel boiler plate . . . 13,500,000 Cast steel (untempered) . . 12,000,000 Copper rolled plate . . . 5,600,000 Phosphor bronze . . I 5,250,000 134 A HANDBOOK OF TESTING MATEEIALS Poisson's Ratio. When a specimen is loaded, there is a direct strain tensile or compressive, according to the nature FIG. 92. -.-Apparatus for Torsion Experiments on Wires. of the test and also a lateral strain, the sign of this strain heing the reverse of the direct one. The ratio of the direct strain to lateral strain is known as Poisson's Eatio, and is usually designated by the letter in. As the lateral strain is very small, it is exceedingly difficult to TOESION TESTING 135 measure it accurately by direct method. Hence the most usual way is to obtain the moduli of elasticity and rigidity in the manner already stated, and calculate m from the following equation : m= E-2C The following table gives mean values of m for some of the more usual metals : Metal. Cast-iron Wrought-iron Steel 3-7 3-6 3-25 Brass 3-0 Copper . . . . . j 2'6 Torsional Experiments on Wire. The value of the shear modulus can be determined in the case of wires by two methods (a) by static deflection and (b) by torsion vibration. Fig. 92 shows the first method in diagrammatic outline. A wire is first pulled taut by being fixed at its lower end and attached to a tightening screw at the top. To the wire is attached a light brass drum carrying a scale divided in degrees. A light cord is passed once or twice round the drum and over two fixed pulleys. To the ends of the cords are attached small scale pans or weight hooks. Equal weights are placed in the scale pans, and the twist of the wire observed by reading the move- ment of the circular scale relative to a fixed pointer. The twisting movement is, of course, the product of the weight in one scale pan into the diameter of the drum. C is obtained from the formula n _oS4M/, W where M is the twisting movement in Ibs. inches, I the length of wire in inches, the deflection in degrees, and d the diameter of the wire. 136 A HANDBOOK OF TESTING MATEEIALS Fig. 93 shows the very simple form of apparatus used to determine C by torsional vibrations. The wire is fixed as before, and carries near its lower extremity a light brass tube. Four other pieces of tube are provided, each exactly one quarter of the length of the long tube, and made to slip inside it. Two of the short pieces are filled with lead, and two are empty. Two values of the time of torsional vibration are obtained one when the two empty tubes are at the extremity of the long tube, and the other when this order is reversed and the loaded tubes are at the extremities and the empty ones at the centre. It can be shown that D FIG. 93. Apparatus for Testing Torsional Vibrations of Wires. Where C is modulus of torsional rigidity. I is length of wire in inches. d diameter of wire in inches. mi mass of each of the tubes filled with lead in Ibs. w 2 mass of each of the empty brass tubes in Ibs. x half the length of the long tube. ti and t% the observed times of torsional vibrations, with loaded tubes outside and inside respectively. CHAPTEE VII IMPACT AND HARDNESS TESTS THE object of an impact test is to obtain the efficiency of the material on test, when it has to be introduced into a machine or structure where it will be subjected to shocks or suddenly applied loads. The importance of this branch of testing work is well illustrated in a paper on " Impact Tests " read before the Institute of Mechanical Engineers in 1904 03^ Messrs. Seaton and Jude, where an approximate analysis of the stresses to which the steel parts of a reciprocating steam engine is given as follows: Constant tension .... 3*91 per cent. Constant tension and compression (range from to a maximum) . 1-80 Constant tension and shock . . 48'80 Alternating tension and compression with shock 2*81 Kepeat tension (from a constant to a maximum) with shock . . 36'00 ,, Miscellaneous and doubtful I'll Total . 100-00 "It will therefore be seen that 87'6 per cent, of the whole of the engine's stresses are more or less due to shock, whilst pure tension stresses form an insignificant percentage of the total stress." It is furthermore stated that if other machines were analysed in a similar manner, nine out of ten would be found to be working under similar conditions. AYery Machine. The machine consists of two A frames of 138 A HANDBOOK OF TESTING MATEEIALS cast-iron, between which swings a pendulum consisting of a steel tube, terminating in a cast-iron head. The specimen, consisting of a small piece of metal f of an inch wide and T 3 6 - inch in thickness, is placed in a vertical position in a vice at the base of the instrument. The pendulum is then raised" to the required height and secured by the releasing trigger carried on the curved arm of the frame. The height to which the pendulum is raised, and consequently its potential energy, is indicated by a pointer on the quadrant at the top of the frame. When the trigger is released, allowing the pendulum to fall and strike the specimen, the indicating pointer engages with a loose registering finger, and carries the latter forward with it. Now if no specimen were present, the pendulum would rise to an equal height on the other side, no energy being absorbed. But when a specimen is broken, the pendulum does not rise so high ; its travel on the other side of the vertical, being inversely proportional to the amount of work done in breaking the specimen. Hence the loose registering finger is carried over to the farthest point reached by the pendulum after breaking the specimen, and stays there, thus indicating directly the amount of work done in breaking the test piece. The maximum capacity of this machine is 23 ft. Ibs. Tensile Impact Tester. The following is a simple machine for examining the properties of materials under tensile impact. The specimen, consisting, let us say, of a long piece of steel wire, is gripped at the upper end in a block which slides upon two or four vertical pillars. The lower end is attached to a heavy hammer-block which keeps the specimen or wire in tension. The whole system is hoisted to some known height up the vertical pillars, and then allowed to fall. In falling, the block which holds the top end of the specimen is suddenly arrested by a stop, fixed to the supporting pillars, while the hammer-block at the lower end is still unsupported. The energy contained in the hammer is arranged to be more than sufficient to rupture the test bar or wire. The actual energy required to just break the specimen is obtained as follows. IMPACT AND HARDNESS TESTS 139 60 m a & 140 A HANDBOOK OF TESTING MATEEIALS A rotating drum is provided, and the hammer has a pencil attached to it, which, on falling, describes a time-displace- ment curve. The surface speed of the drum is found by holding against it a vibrating tuning fork, which has a known period. This being known, the velocity of the hammer at any point in the descent can be obtained from this curve. We can therefore find the energy contained in the hammer at the time of rupture by computing the velocity just after the moment of rupture. We can obtain the total energy in the hammer at the start from the height of fall, and therefore, by subtraction, we can find the energy required to break the specimen. A machine of identically the same principle, but some- what different in detail, has been recently described in a paper before the Institution of Mechanical Engineers by Messrs. Blount, Kirkaldy, and Sankey. 1 This machine, which is constructed for breaking specimens of mild steel 0'357 inch diameter, is arranged in a building so as to have an available fall of forty feet. The anvil is of an exceptionally rigid and heavy design consisting of two castings, each weighing 400 Ibs., securely bolted to the top of four rolled steel joists, which in turn are bedded in concrete. The effective weight is 2,000 Ibs. The test piece connects together the "tup" and arresting cross-head, and the whole is allowed to fall freely until the arresting piece strikes the anvil, when the specimen is broken and the tup continues to fall. The tup can be varied in weight 2 Ibs. at a time between 10 Ibs. and 20 Ibs. Electric contacts are broken by the falling weight (1) at the moment of release, (2) on striking the anvil, and (3) after the tup has fallen 10 feet after fracture of speci- men. By means of a Morse telegraph " inker," a pendulum half-second contact maker, and a vibrating tuning fork, the making and breaking of the electric contacts works the usual cronographic method, so that the times between the standard points given above can be ascertained to within about "005. 1 " Comparison of the Tensile, Impact-Tensile, and Repeated-Bendin Methods of Testing Steel." Proc. Inst. Mech. Eng., May 27th, 1910. IMPACT AND HAEDNESS TESTS 141 The time between points (1) and (2) merely serves as a check, while the time between (2) and (3) serves to calculate the energy remaining in the tup after fracture of the specimen. The full expression for the energy of breaking is Energy absorbed == W | H-i(|-f where H is the height of free fall before striking anvil ; h is the height of free fall after striking anvil, i.e., between the anvil contact and the bottom contact ; W is the weight of the tup ; and t is the time-interval between the anvil and bottom contact. It was found by the experimenters previously mentioned that the energy absorbed per cubic inch in an impact tensile test, divided by the elongation, gave a measure of the breaking stress in an ordinary tensile static test. The strength given by the impact test averaged about 1*2 times that given in a static test. It is worth noting that the authors of the above mentioned papers say, " The energy absorbed per cubic inch does not vary greatly with the various types of steel ; it is approximately 50 per cent, more than that obtained by the static tensile test, and is also no definite criterion of the type of steel." National Physical Laboratory Apparatus. 1 The machine consists of a cast-iron anvil and a tup, each supported by four pieces of steel strip inch wide by - inch thick and about 12 feet long. The anvil has two heavy bosses on the sides, through which passes two pieces of round steel bar. These can be adjusted to protrude any distance towards the middle of the anvil, and are locked by means of set-screws. The ends of these bars are cut away to hold the knife- edges, against which rests the specimen, kept at the right height by adjustable supports. The tup is provided with a steel knife-edge, adjustable outwards so as to just touch the specimen when the tup and the anvil are at rest. 1 Described in Proc. Inst. Mech. Eng., 1905, p. 886. 142 A HANDBOOK OF TESTING MATEEIALS From the back of both tup and anvil a string is carried over a pulley near the roof, with a small weight attached (just sufficient to keep the string taut). The rise of these weights is a measure of the height through which the tup or anvil is raised. The actual heights of the tup and anvil corresponding to the observed motions of the small weights on the strings is obtained by separate experiment. The anvil weighs about 60 Ibs. and the tup about 47 Ibs. The specimens used in these tests were 5 inches long by f inch square, and were notched on the tension side with a small V groove. The knife-edges were placed 4^ inches apart. The method of test is as follows : The specimen is placed in the anvil, and the tup tied back at the desired height by a piece of thin string. The tup is then released by severing the string with a sharp knife, and an observer notes the height to which the anvil is forced, while a second TABLE Y. NICKEL ALLOYS FORGED AND COOLED FROM 800 C. (1,472 P.). Shock Tests. Hardness Tests. Ni. Alloy. Fall of 46-7 Ibs. Hammer. Energy Absorbed. Bending Angle. Indentation in l5 \ T?f inch (Unwin's Test). BrineU's Ball Test. Hardness Number. Relative Hardness. Indenta- tion Method. Swedish Load Load 1-5 tons. 2 -5 tons. Iron = 1. Inches. Inch Ibs. Degrees. A 13-23 451 18-0 7-2 15-0 202 1-6 B 13-05 428 17-0 6-4 14'5 207 1-8 C 13-67 454 16-5 7-0 19-5 212 1-6 D 13-92 460 15-5 6-0 12-3 217 1-5 E 13-67 217 Broken 4-2 8-7 321 O. F 13-67 105 Broken 2-5 5-5 532 , O. G 14-15 230 Broken 2-5 5-7 578 2-2 O. H 14-17 436 7*5 3-2 6-2 555 2-2 J 13-33 432 14-5 5-0 10.3 293 2-0 K 13-77 452 28-0 16 40 131 1*2 IMPACT AND HABDNESS TESTS 143 observer notes the height to which the tup swings after the blow. The work given as that required to deform or break the specimen is the difference between the kinetic energy of the system before and after the blow, calculated from the heights to which the masses are raised. The preceding table from the Seventh Keport of the Alloys Eesearch Committee l shows some results obtained on the nickel alloys described in that paper. For purpose of comparison hardness tests taken by means of (a) penetration of a hardened steel point, (b) Unwin's tests, (c) Brinell's ball test (see page 147), are given below. FIG. 95. Impact Testing Machine, repeated Hammer Blows. Seaton and Jude's Impact Testing Machine. This is a simple machine and consists merely in arranging a weight attached to guides and arranged to fall through a definite distance on to a notched specimen held in a fixed anvil. A very usual size of specimen is 4 inches X J inch x J inch with a notch about J inch deep in the centre. This is held in the anvil so as to form a beam supported at the ends with about 3 inches span. The total falling weight is about 7 Ibs., and is arranged to fall a distance of, say, 2 feet. After each blow the specimen is reversed and the blows continued until fracture occurs. A Repeat Impact Testing Machine, which is a modified form of that originally designed by Dr. Stanton, is shown in 1 Proc. Inst. Mech. Eng., 1905, p. 880. 144 A HANDBOOK OF TESTING MATEEIALS diagrammatic outline in Fig. 95. A is a hammer head carried at the end of a lever, which is pivoted at B. The power is supplied by a belt 00, which drives a pulley H, which is in turn attached to a crank J. As this crank revolves it works a connecting rod M, whose motion is guided by a grooved pulley K. The end of the rod is thus caused to describe an irregular elliptical path, the amplitude of which varies with the position of the pulley K, which is adjustable. At N is a catch so arranged that as the end of the rod ascends it engages with this catch and raises the hammer to a height depending on the position of L. The connecting rod then moves forward and the hammer descends freely. The speci- men, shown in section at D, is placed across the knife edges and gripped in a chuck attached to a chain pulley F. This later is connected to a second chain pulley G on the main driving shaft, the two wheels being geared 2 to 1. By these means the hammer is caused to strike the specimen twice in every revolution, the cycle being repeated about 100 times a minute. L is arranged to move along a graduated scale showing directly the fall of the hammer. The specimen under test is usually about J inch in diameter, with a groove turned at its centre to ensure its fracture there. The knife edges are cut slightly hollow, and a spring holds one end of the specimen in place. The chuck hold is so arranged that it does not take any portion of the hammer blow, all of which comes on the knife edges. Although not shown in the diagram, there is an intervening mechanism between the chuck and the pulley F, by which the former does not have a uniform rotation, but a spring and trigger arrangement causes the specimen to have a step-by- step motion. That is to say, the specimen receives a blow, twists suddenly through 180 degrees, and remains stationary until another blow is struck, when, after the hammer has lifted, it again turns through 180 degrees. A revolution counter records the number of blows struck. When fracture occurs, the specimen falls away, the hammer head continuing to fall first works an electric switch which shuts down the IMPACT AND HAEDNESS TESTS 145 driving motor and then comes to rest on a steel stop pin. This machine is made by the Cambridge Scientific Instru- ment Company. PLAN , E> A O EDGE Ic. & Vor <2 eclge D E LEVATiON P FIG. 96. Un win's Apparatus for Hardness Tests. Hardness Tests. Hardness tests may be divided into two chief classes, (a) scratch tests, (b) indentation tests. In the first of these methods a loaded diamond is pressed on the polished surface of the material, and the latter pulled so as to T.M. L 146 A HANDBOOK OF TESTING MATEEIALS cause the diamond to make a scratch. Standardisation may be taken either by basing the hardness factor or " number," as it is generally termed, on the load necessary to make a scratch of standard width (the latter being measured by a microscope), or on the width of scratch made by a standard load. This method is specially adapted for extremely hard materials, such as rock, etc., and in such cases indentation tests cannot be satisfactorily applied. On the other hand, it is probable that for the testing of ordinary engineering materials indentation tests are more reliable and easier to standardise. In the case of indentation tests a body of some standard form, such as a knife-edge of standard angle, or a ball of known diameter, is pressed into the material by a definite load. The hardness number is in general based on the volume of the indentation produced by a standard load. One apparatus used in connection with the indentation test is shown in Fig. 96. It consists of a strong framework A, with open sides, accurately bored to receive a ram C. The knife-edge D, which is of square section, rests against a true face on the bottom of the ram. To obtain a truly axial load a ball and socket joint is provided between B and C. The test piece, resting on the bottom of the frame A, is indented by the knife-edge D, care being taken that the indentation is not so deep as to cause the metal to spread at the sides. The depth of the groove thus formed is measured by the scale and vernier E as in the punching test. A series of observations of indentation and load are taken, from which a constant is deduced which is the measure of hardness. The test piece is usually about f inch square and 2j inches long, and the load is applied by placing the apparatus in an ordinary testing machine and applying a compressive load. It is necessary to standardise the machine by measuring the amount of compression which the machine itself undergoes apart from the indentation of the specimen. Prof. Unwin has IMPACT AND HAKDNESS TESTS 147 shown that a relation between indentation and load takes place according to the formula where i is the depth of the indentation in inches, p the pressure per inch of width of knife-edge producing the indentation, and C a constant for the material, termed the hardness number. Readings should be taken at varying loads, and the mean value of C calculated from the equation log C=l*2 log p log i. The following are some values obtained by Unwin : Metal. Value of C. Cast steel, normal. ..... 554'0 Mild steel, normal. . . . . . 148*5 Copper, annealed . . . . , . - 62*0 ,, unannealed . . . . 105*2 Brass, No. 1 . . . . * . 221*0 No. 2 ..... . . 246*0 Aluminium, squirted . . ...'". . . 41*8 alloy, cast ..... 103*5 Lead, cast . . . . . 4*2 Zinc, cast . . . . . . . 40*8 Messrs. Calvert and Johnson used, instead of a knife-edge, a small truncated cone, and took as the measure o,f hardness, the weight which would indent the metal to a depth of 3 J milimetres in half an hour. In some United States tests the volume of indentation produced by a pyramidal point loaded with a weight of 10,000 Ibs. was used to measure the hardness of the material. Brinell's Ball Test. Since in all hardness tests in which the indenting tool has a sharp point or edge the latter is likely to become blunt, it is obvious that results may become depen- dent on the condition of the tool, and for this reason it is probable that a special ball offers the best form for making the indentation. This is the method employed by Brinell. L 2 U8 A HANDBOOK OF TESTING MATERIALS The test consists in pressing a hardened steel ball into the flat surface of the specimen under a known pressure and measuring the volume or curved area of the impression. Fig. 97 shows one form of the apparatus employed. FIG. 97. Brinell's Machine. Brinell has taken as a basis of comparison Hardness number Tofad Curved area of impression' But curved area=2*rf r Where I) is diameter of impression and r radius of ball. Whence H.N.=- IMPACT AND HAKDNESS TESTS 149 The number obtained by this formula will vary somewhat for different value of P and r, hence Brinell has fixed the standard by taking r= 5 millimetres, and P= 3,000 kilogrammes, D being in millimetres. It has been shown by Benedicks of Upsala that for balls of different radius a constant value is obtained by multiplying the above hardness number by \/r. Or Benedick's hardness number=(Brinell's H.N.) 5 \A. Various types of machines for applying the Brinell test have been devised, notably by Guillery and by Brinell himself. The following table, given by Prof. Warren, of the University of Sydney, gives a ready means of determin- ing the hardness number from the diameter of the depression : TABLE VI. KELATION BETWEEN THE DIAMETER OF THE IMPRESSION AND BRINELL' s HARDNESS NUMBER. Hardness Number for the Hardness Number for the Diameter of Pressure in kgs. Diameter of Pressure in kgs. Impression. Impression. Mm. 1 Mm. 3,000 500 3,000 500 2-00 946 158 2'75 495 83 2-05 898 150 2-80 477 80 2-10 857 143 2-85 460 77 2-15 817 136 2-90 444 74 2-20 782 130 2-95 430 73 2-25 744 124 2-30 713 119 3-00 418 70 2-35 683 114 3-05 402 67 2-40 652 109 3-10 387 65 2-45 627 105 3-15 375 63 2-50 600 100 3-20 364 61 2-55 578 96 3-25 351 59 2-60 555 93 3-30 340 57 2-65 532 89 3-35 332 55 2-70 512 86 3-40 321 54 150 A HANDBOOK OF TESTING MATERIALS TABLE VI. Continued. Diameter of Impression. Mm. Hardness Number for the . Pressure in kgs. Diameter of Impression. Mm. Hardness Number for the Pressure in kgs. 3,000 500 3,000 500 3-45 311 52 5-20 131 2118 3 -50 302 50 5-25 128 21-5 3 '55 293 49 5-30 126 21*0 3-60 286 48 5-35 124 20-6 3-65 277 46 5-40 121 20-1 3-70 269 45 5 '45 118 19-7 3-75 262 44 5-50 116 19-3 3-80 255 43 5.55 114 19-0 3-85 248 41 5-60 112 18-6 3-90 241 40 5-65 109 18-2 3-95 235 39 5-70 107 17'8 5*75 105 17'5 4-00 228 38 5-80 103 17-2 4-05 223 37 5*85 101 16-9 4-10 217 36 5-90 99 16-6 4-15 212 35 5-95 97 16-2 4-20 207 34-5 4-25 202 33-6 6-00 95 15-9 4-30 196 32-6 6-05 94 15-6 4-35 192 32-0 6-10 92 15-3 4-40 187 31-2 6-15 90 15-1 4-45 183 30-4 6-20 89 14-8 4-50 172 29-7 6-25 87 14-5 4'55 174 29-1 6-30 86 14-3 4-60 170 28-4 6-35 84 14-0 4-65 166 27'8 6-40 82 13-8 4-70 163 27-2 6-45 81 13-5 4-75 159 26-5 6-50 80 13-3 4-80 156 25-9 6-55 79 13-1 4-85 153 25-4 6-60 77 12-8 4-90 149 24-9 6-65 76 12-6 4-95 146 24-4 6-70 74 12-4 6-75 73 12-2 5-00 143 23-8 6-80 71-5 11-9 5-05 140 23-3 6-85 70 11-7 5-10 137 22-8 6-90 69 11-5 5-15 134 22-3 6-95 68 11-3 Professor Warren gives some interesting figures showing the relation between the tensile strength and the hardness as obtained by Brinell's ball test. IMPACT AND HAEDNESS TESTS TABLE VII. 151 Tensile Strength. Tons per sq. in. Brinell's Hardness Number. Ratio of the Tensile Strength to Hardness Number. 28-9 170 170 30-2 149 203 25-3 141 180 2,3-6 140 183 27-0 140 193 25-8 140 184 24-9 137 182 25-6 134 191 Mean 26'7 143-9 186 The above figures were obtained on structural steel, the ball being 10 millimetres diameter ; pressure 3,000 kilogrammes. It will be observed that the value obtained by dividing the actual tensile strength by the Brinell's hardness number is fairly constant for the same quality of steel. In the case of axle steel the following was obtained TABLE VIII. Tensile Strength. Tons per sq. in. Brinell's Hardness Number. Ratio of the Tensile Strength to Hardness Number. 34-0 36-5 33-9,5 167 168 205 202 216 165 Mean 34-8 180 194 Dillner, of Stockholm, has investigated this same relation between tensile strength and hardness number, and found with only a mean error of 3'3 per cent, that the tensile strength in tons per sq. in. for steels having a hardness number below 175 could be obtained by multiplying the 152 A HANDBOOK OF TESTING MATERIALS hardness number by 0*230 when the indentation was made transversely to the direction of rolling, or 0*225 when made in the direction of rolling. The " Scratch " Test is used by Prof. Turner in his " scleor- meter." This consists of a perfectly balanced lever, which has a diamond point at one end. The whole arrangement can be moved from side to side, so as to make a scratch on the smooth surface of the metal to be tested. The load on the point can be varied by adding weights to a small scale pan on the lever, and the weight in grams acting on the diamond point required to produce a standard scratch is used as a measure of the hardness of the metal. In connection with hardness tests an important communi- cation upon the subject was compiled by Prof. Thomas Turner, 1 M.Sc., from which the following observations are obtained. He compared four methods, viz.: (1) Turner's sclerometer; (2) Shore's scleroscope ; (3) Brinell's test; (4) Keep's test. It may be mentioned that method (2) involves the use of a new patent, viz., the scleroscope. In this instrument there is a small cylinder of steel with a hardened point. This is allowed to fall on the smooth surface of the metal to be tested, and the hardness of the material is taken by measuring the rebound. Prof. Turner says : " Each form of test has its advantages and its limitations- The sclerometer is cheap, portable, and easily applied, but it is not applicable to materials which do not possess a fairly smooth reflecting surface, and the standard scratch is only definitely recognised after some experience. The Shore test is simple, rapid, and definite for materials for which it is suited, and appears likely to have an important future. But further information is yet needed as to the exact property which is measured by this form of test. As shown by De Ereminville, the result obtained varies somewhat with the size and thick- ness of the sample, while if the test piece is supported on a soft material such as a plasticine the results are valueless. It may, however, be pointed out that indiarubber gives a 1 Journal of the Iron and Steel Institute, 1909. IMPACT AND HAEDNESS TESTS 153 rebound of 23, which is equal to that of mild steel ; while I have found light soft pine-wood give a rebound of 40, which is nearly double as great as that of grey cast-iron. Curiously enough hard wood, like teak, gives a rebound of about 12, while some samples are considerably lower than this. As illustrating the influence of the support, a sample of excep- tionally hard rolled copper about -^ of an inch in thickness, when supported on a block of hard steel and tested with the blunt or " magnified " hammer supplied, gave a value of 30, which was increased to 34 when the copper was supported on wood. A sample of brass only gave a value of 17, and yet this brass would scratch the copper, while the copper would not scratch the brass. From these results it is evident that the Shore test is only applicable to a certain class of -sub- stances. It appears to test what may be termed the " elastic hardness," and gives high results with metals in the " worked hard " or ecroui condition ; values which are not fully con- firmed by the tool or by the sclerometer. My tests appear to show that good results are, however, obtained with glass and with porcelain, as well, of course, as with most metals. The Brinell test is specially useful for constructive material; it is easily applied and definite, and is now of all hardness tests the one most employed. It appears to give satisfactory results with wood, but cannot be applied to very brittle materials such as glass, or to hard minerals. Keep's test is specially suited for castings of all kinds, as it records not merely the surface hardness, but also that of the whole thickness, and gives indications of blowholes, hard streaks, and spongy places. Obviously it can only be applied to materials the hardness of which is less than that of hardened steel." Curiously enough, however, it may be noted that all four methods give comparative results for all pure metals in their normal condition. The following results, obtained from the experiments of Prof. Turner, will give an idea of the closeness of agreement and the actual values of various materials. Other tests of hardness are grinding, machining and 154 A HANDBOOK OF TESTING MATERIALS drilling tests, but the methods most used are those described above. TABLE IX. HARDNESS SCALES COMPARED. Metal. Turner. (Sclerometer.) Scleroscope. Brinell. Keep. 6. Lead . . 1-0 1-0 1-0 Tin . . . . . 2-5 3-0 2-5 Zinc 6-0 7-0 7-5 Copper, soft . ,, hard 8-0 8-0 12-0 12-0 Angle Softest iron . 15-0 . 14-5 varies Mild steel . 21-0 22-0 16-24 from Soft cast-iron Eail steel . . 21-24 24-0 24-0 27-0 24-0 26-35 to 90. Hard cast-iron 36-0 40-0 35-0 Hard white iron . 72-0 70-0 75-0 Hardened steel . 95-0 93-0 Combined Static and Shock Tests. At a recent meeting of the Physical Society (October, 1910) Mr. Kogers presented a paper describing the results of tests made upon steel specimens subjected to stresses caused (simultaneously) by bending and shock, or impact. It appears that the results would be easier to interpret if the static load were either tension or compression. The work of Mr. Eogers is important, as the author believes it to be the first published on combined static and impact tests. It should suggest the lines of further interesting experiments. CHAPTEK VIII SHEAR AND MISCELLANEOUS TESTS DIRECT shear or " rivet " tests, that is to say shear tests as distinct from torsion tests, are not very frequently carried out, owing to the difficulty of ensuring pure shear without hending. When such experiments are made they are generally performed in an ordinary " omnibus " machine arranged for tension or compression, use being made of a special form of shackle. There are many forms of the latter, and such will readily suggest themselves to the student as modifications of the single example taken. Shearing Shackles. This apparatus, shown in Fig. 98, con- sists of two rectangular blocks A and B, which are bolted together by the four bolts C, D, E, F. Down the centre of these blocks is cut a rectangular channel G, in which a rectangular piece H is allowed to slide. The three discs K, L, and M fit accurately in a hole which is bored through A and B, and are fixed in position by the hollow nuts N and 0. The specimen is supplied in the form of a cylindrical bar, which accurately fits inside the holes in K, L and M. When the bar is in position the load is applied to H, and the specimen is sheared off at the joints between K and L and L and M, the strength in double shear thus being ascertained. It is important that the faces of K, L and M should be perfectly true, as if there is any space between them the material is not in pure shear, but partly in bending. Shear Tests. With this apparatus tests on materials can be carried out in either single or double shear. The bearing surface should always be kept as large as possible, so that when making tests in single shear, the specimen should be 156 A HANDBOOK OF TESTING MATEEIALS pushed so far through the middle die that it just misses fouling the end die. Z i id J uJ To avoid errors due to bending the specimen should be a very good fit in the dies, and for the same reason there should be no clearance between the latter when in position. SHEAE AND MISCELLANEOUS TESTS 157 The following results were obtained for mild steel, wrought- iron, and cast-iron in both double and single shear : TABLE X. Material. Method of Shear. Diameter. Inches. Area. Sq. in. Load. Tons. Shear Stress, Tons, per sq. in. Mild steel Double 875 601 21-66 18-02 Single 875 601 10-90 18-14 Wrought -iron Double 877 606 20-97 17-32 Single 877 606 10-38 17-20 Cast-iron Double 871 596 15-12 12-70 Single 874 600 7-48 12-47 As will be seen from the table, the strength of a material in double shear is about twice that in single shear. In addition to thus testing in single and double shear some interesting and useful results can be obtained by testing in shear specimens cut from the same bar as those specimens tested in tension. From such results valuable information can be gained as to what proportion of the tensile stress is to be taken for a corresponding safety factor in shear. The following table gives actual results : TABLE XI. Material. Breaking Tension. Shear Stress. Ratio. Tons per sq. in. Best Staffordshire iron . 24-41 19-35 793 Yorkshire iron 23-13 18-02 781 Mild steel bar ... 28-64 20-04 700 Steel boiler plate . 27-17 18-68 688 Spring steel .... 48-10 30-84 642 Copper plate 14-33 9-96 695 It is probable that, with such tests, a pure shear is never obtained because of the complication due to bending. The most satisfactory method of obtaining pure shear is that of 158 A HANDBOOK OF TESTING MATERIALS SHEAR AND MISCELLANEOUS TESTS 159 using a hollow specimen for a torsion test. At the same time the results obtained of this " rivet shear " test are valuable, as they are obtained under conditions such as take place in practice. Punching Tester (Fig. 99). The apparatus consists essentially of a steel cylinder A which slides up and down in a cylindrical guide B. The test piece, consisting of a plate of the required material, is secured tightly between the bottom 5000 10 Q * :> 3000 a t 2ooo c __j JOOO PUNCHING TESTS. MILO STEEL OBSERVED. -2 //V FIG. 100. Punching Test on Mild Steel. of B and the base plate C by means of the bolts D, E. The punch is fastened to the cylinder A, while the die fits in a corresponding recess in the base plate C. The load is applied vertically downwards on the cylinder A, and the punch thus driven through the test plate. The strain of the metal can be read for any given load by means of a scale fitted to A, and a vernier attached to the guide B. The tests can also be automatically recorded by the following arrangement. The downward movement of A, and consequently the yielding of the specimen, is transmitted by means of a rack and pinion, through a stretched cord to a drum G, which thus rotates in 160 A HANDBOOK OF TESTING MATERIALS proportion to the strain of the test plate. The pencil is moved along the axis of the drum by means of a cord or wire from the loading mechanism, its movement being proportional to the load on A. Hence the diagram recorded is a stress-strain curve, and the work done in punching out the plate can be obtained from it in the usual way. Punching Tests. By means of this apparatus, load-strain diagrams can be obtained either autographically or by plotting simultaneous readings of load and strain, the latter being obtained by means of the vernier. Load calibration of the autographic diagram is performed by observing the maximum load which is reached during the test. This load will correspond with the highest point on the curve, and thus the load scale can be determined. The strain is, of course, obtained knowing the thickness of the plate being punched. Fig. 100 shows diagrams obtained by each of these two methods on the same plate of mild steel with the same punch. These diagrams afford a simple method of finding the work done in punching the hole, equalling, as it does, merely the area under the curve to the correct scale. Eesistance to punching is practically a shearing resistance. Then if d= diameter of hole, and =thickness of plate Shearing resistance of plate per sq. in. maximum load irdt ' For comparison these quantities have been calculated for the two curves, Fig. 100. Mild Steel Specimen. j Diameter of punch =*870 inches. ( Thickness of plate='359 ,, From autographic diagram. Maximum load = 43, 800 Ibs. per sq. in. Work done in punching hole =700 ft. Ibs. , r . P 43,800 Maximum punching resistance of plate = =-= =44,750 Ibs. per sq. in. SHEAR AND MISCELLANEOUS TESTS Diagram from observed readings. Work done in punching=725 ft. Ibs. AT 45,200 Maximum punching resistance : ~ 161 "7rX'87X-359 =46,100 Ibs. per sq. in. Tests of Steel Balls. The testing of steel balls in a satisfactory manner presents some difficulty, as it is not an easy matter to obtain a surface which will not become indented Hardened Steel ^Hardened Steel Sectional Elevabion . Side View. FIG. 101. Apparatus used for Testing Steel Balls. before the pressure becomes sufficiently great to crush the ball. Up to sizes of about f inch diameter hardened steel plates, between which to crush the ball, can be used with moderate success, but with balls of larger diameter than this, the plates either indent, or if sufficiently hard, actually crack. A method of testing the larger sizes of balls at the East London College, which has been found to answer very successfully, is roughly as follows. The ball to be tested is supported between two other balls, these in turn exactly fitting T.M. M 162 A HANDBOOK OF TESTING MATEKLALS into turned recesses in hardened steel blocks. With such an arrangement the centre ball, i.e., the one to be tested, breaks first, almost without exception, and in addition a point contact is obtained. Fig. 101 shows the apparatus employed. CRUSHING LOAD IN TONTS. ;o 2o .30 <*o So Area ABCD AreaDEEG This probably only applies to mild steel and wrought-iron, but it would be worth the student's while to see if he can get any relation which remains fairly constant for different changes of state. 1 Zustandsdnderungen der Metalle infolge von Festigkeitslea?ispruchimgen. A. Marten, Preuss. Akad. Wiss. Berlin, Site. Ber. 11, pp. 209220, 19io. CHAPTER IX ALTERNATING STRESS TESTS Alternating Stress Machines. Many structures or parts of machines are subject to alternating stresses, that is to say the load on them is constantly being increased or decreased, and in many cases reversed. The diagonal bracing in the centre of a bridge, a steam-engine connecting rod, or a railway- carriage axle are typical examples. It is important, therefore, that materials subject to such stresses should be tested under similar conditions, and the following machines have been devised to that end : The pioneer of this type of test was Wohler, who, in 1871, published valuable researches with reference to the effect of alternating stress upon the strength of materials. Other prominent workers in the same field have been Prof. J. 0. Arnold, Captain Sankey, and Dr. Stanton, all of whom have invented machines capable of performing alternating stress tests. Such tests are useful for experimental investigations, and are of advantage in some commercial cases. The static test will probably always remain the chief method used in commercial work. It is, however, desirable that the value of alternating stress experiments, especially when applied to new steel and other alloys, should not be underrated. Wohler's Machine for Testing Alternating Torsional Stresses. This is illustrated in Fig. 106. The test bar A is a simple cylindrical bar with enlarged ends, held in suitable chucks and supported by two bearings B, C. To the end of the mandril passing through the bearing C is attached a lever D, which carries knife-edges at its end. H is a vertical rod attached at its lower end to a horizontal lever P, and having at E and F adjustable collars which bear on the 170 A HANDBOOK OF TESTING MATEEIALS knife-edges. The lever P, which is pivoted at G, is moved up and down by the connecting rod Q, and arrangements are made by which this latter lever can be put rapidly in and out of gear ; also by means of a slotted link in P the stroke of the latter can be varied. It will thus be seen that the vertical oscillations of Q are transformed to twisting oscillations of the specimen A. The reaction of the torque thus transmitted is taken by means of the system of levers J Ji, K K b etc., to the springs L LI. By adjusting the nuts at E and F the stroke of D can be varied so as to give equal stresses in opposite directions. N Niand M MI are then adjusted, so that the contacts at B and RI are alternately lifted against the spring by the reaction of the torque given to the specimen. Different stresses are obtained by varying the stroke of P. Wbhler's Tension Alternating Stress Machine. Fig. 107 shows in diagrammatic outline the machine used by Wohler in his tension alternating stress experiments. The specimen A is held in suitable shackles B and C. The low shackle C is made of adjustable height by means of a screwed tail rod, which can be fixed in different positions relative to the frame of the machine by means of the nuts shown. The upper shackle B is carried on knife-edges on the end of a beam C, which is supported, also on knife-edges, at D. Q is a vertical oscillating connecting rod, and by means of a slot and pin can be made to give a variable oscillating stroke to the lever E. The end of the lever C carries a link attached to the horizontal beam J, whose other end is connected to a similar lever and spring used in the torsional machine. The oscillations of the lever E are transmitted by means of a pin working in a slot F to the spring H, which consequently transmits a varying downward pull on the lever J, and hence causes a variable stress on the specimen A. It is obvious that the pull on the specimen is proportionate to the tension of the spring L necessary to keep the lever K from rising. To set the machine for giving a stress alternating between two fixed limits, the spring L is adjusted by means of the nut M, so that the minimum stress would make the lever K rise. ALTERNATING STRESS TESTS 171 172 A HANDBOOK OF TESTING MATERIALS The nut P and the right- and left-handed coupling nut G are then adjusted so that K is just lifted. The spring L is now set to give the maximum stress, and the stroke of the lever E adjusted until the maximum stress just raises the lever K. All is now ready for a test. Wohler's Bending Stress Machines. Wohler designed two machines for testing specimens with an alternating bending stress. One shown diagrummatically in Fig. 108 is arranged for varying the stress between a fixed maximum and minimum, FIG. 108. Wohler's Variable Bending Stress Machine. while that shown in Fig. 109 bends the specimen rapidly in opposite directions, so that at any particular point the stress alternates between a compressive and tensile stress of equal intensity. Eeferring to Fig. 108, A is the specimen (a beam) supported on knife-edges at 1) and E. One of these latter I) is fixed, while the other E is attached by a link to the same lever and spring arrangement which we have seen in the other alternat- ing stress machines. The load is applied through the usual oscillating connecting rod and lever B to the knife-edge M. The lever B is attached to the vertical rod K by a pin working in a seat, so that no upward push can be given to the beam. ALTERNATING STEESS TESTS 173 To set the machine for a required variation, the spring G is adjusted by the nut L to such a load that the beam F is lifted by the minimum load to be applied to the beam. The screw J, which is rigidly held by the fixed cross-beam H, is then screwed down until F justs lifts. G is now tightened up by the nut L to the maximum load, and the stroke of the lever B adjusted until the lever F is just lifted at the end of the stroke. All is now ready for a test. Adjustments in the length of K can be made by means of a turn-buckle, while the stroke of B PIG. 109. Wohler's Machine for Eepeat Bending in opposite directions. is varied by moving the attachment pin on C along a slotted link. In the second bending stress machine shown in Fig. 109 two specimens are used, A and AI; these are fixed by driving into two chucks placed at either end of a mandril C, running in bearings D and DI, and carrying a pulley B. The specimens are first turned in a lathe, so that they shall be perfectly straight, and then bearings are attached to them at E and EI of such a kind that a downward pull can be exerted at these points by means of two spring balances F and FI. The pull of these, which were of a special kind, shown diagrammatically in the figure, could be varied by adjusting the nuts at G and Gi. It is obvious that when the pulley B is rotated the specimens will be subjected to a bending which is equivalent to being 174 A HANDBOOK OF TESTING MATERIALS rapidly bent backwards and forwards in opposite directions. The load, in fact, is similar to that acting on the axles of a railway coach when running. TABLE XIY. WOHLER'S EXPERIMENTS ON BARS SUBJECTED TO REPETITIONS OF TRANSVERSE STRESS (ROTATING BARS) BETWEEN EQUAL AND OPPOSITE LIMITS OF STRESS. SET I. No. of Bar. Material. Stress applied in tons per sq. in. Range of Stress in tons per sq. in. No. of Repetitions before Fracture. Maximum. Minimum. 1 1 + 15-3 -15-3 30-6 56,430 2 14-3 14-3 28-6 99,000 3 13-4 13-4 26-8 183,145 4 Iron for 12-4 12-4 24-8 479,490 5 axles, 11-5 11-5 23-0 909,840 6 Phoenix Co. 10-5 10-5 21-0 3,632,588 7 9-6 9-6 19-2 4.917,992 8 8-6 8-6 17-2 19,186,791 9 7-6 7'6 15-2 132,250,000! SET II, 24 +23-9 -23-9 47-8 2,375 25 22-9 22-9 45-8 4,986 26 21-9 21-9 43-8 11,636 27 Homo- 18-2 18-2 36-4 31,586 28 geneous 16-3 16-3 32-6 94,311 29 iron. 14-3 14-3 28-6 161,262 30 13-4 13-4 26-8 464,786 31 12-4 12-4 24-8 636,500 32 11-5 11-5 23-0 3,930,150 SET III. 33 . +20-1 -20-1 40-2 55,100 34 17-2 17-2 34-4 127,775 35 36 37 38 Krupp's - cast-steel - axles. 16-3 15-3 15-3 15-3 16-3 15-3 15-3 15-3 32-6 30-6 30-6 30-6 797,525 642,675 1, 665,5*0 3,114,160 39 H-3 14-3 28-6 4,163,375 40 ' * 14-3 14-3 28-6 45,050,640 1 Not broken. ALTERNATING STRESS TESTS 175 Fig. 110 shows the nature of the stress-cycle in Wohler's machine (Fig. 109). In the diagram are marked the points of maximum skin stress tension and compression, or + and viz., MT S and MC S . At Z and Z 1 , the stress is zero. Prof. Arnold l (in a paper from which the author has obtained much of the follow- ing information) quotes three sets of Wohler's tests (p. 174). The first set has reference to wrought-iron of average statical maximum stress, 21*3 tons per sq. in. ; the second set, on material called FlG. 110. Diagram showing nature and Stress-Cycle of the Wohler Test. " homogeneous iron," was probably mild steel, or ingot iron, having a mean maximum stress of 28'2 tons per sq. in. The third set was probably made upon crucible steel of 0*6 per cent, carbonj with a mean maximum stress of 46'4 tons per sq. in. Arnold's Machine. Prof. Arnold conceived the idea of bringing Wohler's method into practical works use by greatly reducing the time occupied in making a test. He proposed to stress the material just beyond its elastic limit, and so reduce the time MT? / Z-\ taken to make a test from hours to seconds. In Fig. Ill the nature of Arnold's test is shown. B is the test piece, | inch diameter and about 6 inches long. It is gripped in the dieD, and the stress is applied by the strokes of the slotted 1 Trans. Inst. N. A., 1909 FIG. 111. Diagram showing Nature and Stress-Strain Lines in Arnold's Test. 176 A HANDBOOK OF TESTING MATERIALS plunger PP. A rate of alternation of 650 per minute is adopted. The diagrams show the alternation of the stress-strain lines in this test. Z is the fixed zero of stress. Theoretically the Wohler test is perfect, while Prof. Arnold's is wrong; but never- theless it gives valuable results concerning properties which Wohler's tests do not reveal. No matter how dangerously brittle steel may be, from chemical or physical causes, if such treatment has produced a high elastic limit, the Wohler test shows the material safe if stressed well short of that limit. As a matter of fact, such steel suddenly ruptures, sooner or later, under stresses theoretically quite safe. As an instance of this, results of some tests made on several steels by Mr. J. E. Stead, F.R.S., are given. There was a series of three sets with ascending phosphorus up to 0'5 per cent. the last-named being most undesirable for engine parts. The analysis showed everything as being practically identical, except the phos- phorus, which varies as shown in the table giving the results of static tests. TABLE XV. EESULTS OF STATIC TESTS. Steel No. Phosphorus. Per cent. Yield Point. Tons per sq. in. Maximum Stress. Tons per sq. in. Elongation. Per cent, on 6 ins. Reduction of Area. Per cent. 1 041 20-4 33-1 23-0 52-0 2 302 25-4 39-9 23-0 45-3 3 509 32-0 44-0 20-0 45-3 The Wohler tests made by Mr. Stead give the results set forth in the table below, the stresses being + and 15 tons per sq. in., i.e., a range of 30 tons. TABLE XVI. EESULTS OF WOHLER TESTS. Steel No. Phosphorus. Per cent. Reversals of Stress endured. Ratio of Resistance to Alternating Stress. 1 2 3 041 302 509 61,000 167,300 651,000 1-0 2-7 10-6 ALTERNATING STRESS TESTS 17 Prof. Arnold's tests made by him at Sheffield University in complete ignorance of the nature of the steels registered the figures embodied in the table below : TABLE XVII. PROF. ARNOLD'S ALTERNATING STRESS TESTS. Steel No. Phosphorus. Per cent. Alternations endured in Test No. (Under standard conditions.) Ratio of Resistance to Alternating Stress. 1. O 4 Mean. 1 2 3 041 302 509 258 188 72 284 212 128 272 200 100 100 37 Speaking in round numbers, the Wohler test indicated that a mild steel containing 0'5 per cent, phosphorus was, with equal stresses, ten times as capable of resisting alternating stress as a steel containing 0'04 per cent, phosphorus. Prof. Arnold's test, on the other hand, indicated that a steel con- taining 0*5 per cent, phosphorus had about one-third the endurance of a steel containing 0'04 phosphorus. The curves of the yield points of the two alternating tests of the three steels show that the Wohler curve is similar in type to that registered by the yield point or apparent elastic limit, Professor Arnold's curve being in an opposite direction and indicating what is well known to be the mechanical effect of phosphorus on steel. The evidence supplied above seems to show that for prac- tical purposes Prof. Arnold's test is the more useful. Stanton's Alternating- Stress Testing Machine. The prin- ciple of this machine is that of employing a rotating crank to cause periodic motion of a reciprocating mass by means of a connecting rod, the specimen under test forming the con- nection between the reciprocating mass and the crosshead. This device has been employed by Prof. Osborne Eeynoids in the testing machine at Owens College, which is of the vertical liype with a single balanced crank. In the National Physical Laboratory machine there are four cranks operating four T.M. N 178 A HANDBOOK OF TESTING MATEEIALS specimens, the motion of the specimens being in a horizontal plane. By this means the balancing of the machine is made independent of the ratio of the crank-arm to the connecting rod, so that a length of crank-arm has been adopted which enables experiments to be made at moderately low speeds, that is, from 600 to 1,000 revolutions per minute. Although this arrangement causes the motion of the specimens to deviate from the simple harmonic law, the effects on the stresses set up in the specimens are sufficiently small in value to be neglected, so that the maximum tensile force on the specimens may be taken as and the maximum compressive force on the specimens as where W = weight of mass attached to end of specimen in Ibs. ; r = radius of crank pin ; ic = mean angular velocity of crank-shaft ; I = length of the connecting rod. It will be observed that the maximum tensile stress is 1*4 times the value of the maximum compressive stress, which is approximately the ratio of the stresses in the piston-rod of an ordinary reciprocating steam engine. The form of specimen adopted is the same as in Eeynolds and Smith's experiments, consisting of a 5- inch bar screw cut ^ 3 T -inch Whit worth, and turned down in the centre to a diameter of J inch for a length of J inch. Great care has to be taken in the preparation of the specimens to ensure a gradual change of section in the turned-down part, as the effect of a change of section on the resistance of the specimen is much more marked in the case of tests under alternating stresses than in statical tests. Prof. J. H. Smith's Alternating Tension and Com- pression Stress Machine. 1 This machine, called by its 1 Engineering, Vol. LXXXVIIL, p. 105. ALTERNATING STRESS TESTS 179 inventor a fatigue testing machine, was designed by Dr. J. H. Smith in 1904, but subsequently modified to its present form. In this machine, as in Stanton's, the load is produced by the centrifugal force of a rotating mass. Fig. 112 shows the general scheme of the apparatus employed. If we have two masses fixed to the ends of revolving arms which are pivoted at A and B, it is obvious that when the weights are at the top FIG. 112. Diagram of J. H. Smith's Alternating Stress Machine. of their motion they will pull the arm AB upwards with a force expressed by the usual formula 2WV' 2 gr If now AB is rigidly attached to a vertical spindle passing through guides at C and D, and if we place a specimen F between the end of the spindle and a fixed mass E, then the specimen will be subjected to the force expressed above in compression. When the weights are at the bottom of their revolution the force on the specimen will be reversed. Suppose now we attach a spring H to the bottom of the spindle and N 2 180 A HANDBOOK OF TESTING MATERIALS tighten this up so as to exert a force of S Ibs. weight. Then the load on the specimen will vary between , S and - -fS. fir By making S greater than 2WV* gr the load on the specimen is always tensile. In the latest form of apparatus used by Dr. Smith, the apparatus is double, two specimens being under test at one FIG. 113. Diagram of J. II. Smith's Alternating Stress Machine. time. Fig. 113 shows this arrangement, and it will be observed how two other masses are introduced for the purpose of balancing. P is a pulley, and C a counter. The specimens generally used are small, being only J inch in diameter and | inch long, but by introducing a special form of chuck lengths up to 4 inches can be used. Fig. 114 gives an idea of the scheme as actually carried out in practice. The part shown is, as was explained previously, duplicated, and the whole machine runs in an oil bath. Eef erring to Fig. 114, there is, of course, no rigid connection at the point A, the arrangement consisting of a small slider working in guides. Otherwise part of the forces due to the ALTERNATING STRESS TESTS 181 revolving weights would be taken up by the driving shaft. The reader is referred to the article in Engineerinq, (*) FIG. 114. J. II. Smith's Alternating Stress Machine. mentioned previously, for a full description, together with particulars of the very ingenious optical strain recorder and stress-strain oscillograph used. 182 A HANDBOOK OF TESTING MATERIALS The following are some results obtained by Dr. Smith on this type of machine : TABLE XVIII. Oscillatory weight, 12-42 Ibs. Diameter of specimen, -2409 inch. Set A. Mild Steel. Revolutions per minute. Maximum or Tensile Stress per sq. in. Minimum or Compression Stress per sq. in. Range of Stress per Sq. ill. Number of Reversals before Rupture. Tons. Tons. Tons. Annealed . Unannealed 2,126 2,122 7-99 8-03 7-11 7-17 15-1 15-2 248,700 226,500 A Ring-Shaped Specimen. Dr. Stanton has very ingeni- ously devised an apparatus in which the specimen used is a ring. The latest annual report (1910) from the National Physical Laboratory contains an account of some high, frequency experiments. The number of reversals has been 2,200 per minute, and the results are very satisfactory for hard steels (above 3 per cent, carbon). In the case of the softer steels and iron, indentations which weaken the ring occur. Dr. Stanton has successfully checked the results from his ring method against tests made with a Wohler machine running at the same frequency. Dr. Stanton adheres to the opinion that the range of stress is not altered by the speed, and that neither iron nor high -carbon steel are worse at high speeds than low speeds. Dr. Smith, of Belfast, and others are opposed to this opinion. Since various independent investigators are at work on the subject we may soon hope for a definite decision and numerical data on the effect of frequency upon the range. C. A. M. Smith's Alternating Torsion Machine. Wohler and other investigators have designed machines to work between + and loads as already described. A machine has been designed by the author to do the same type of work, but the alterations are produced by loading a shaft so that a torque is applied. The principle of the machine is the ALTEENATING STEESS TESTS 183 stopping and starting of a fly-wheel. The power is supplied by a small motor. Experiments have not yet been made with the apparatus, which is still in the course of construction. The chief practical difficulties appear to be those concerned Static Breaking + h D 007. i iz Mian oFZO F 50 2 2 Tear oF8 30 20 Ten ion Zero St ress Con ore: s/or 9 / 10 20 30 4-0 C PIG. 115. Curve for Eationalisation of Alternating Stress Experiments. with obtaining balance so that there shall not be undue vibration at high frequencies. There is a minor difficulty of securing specimens to the grips, which would not occur but for the special design for other purposes of the specimens used. Repeat Stresses. Prof. Goodman has given a method of determining graphically the allowable maximum stress for a 184 A HANDBOOK OF TESTING MATERIALS given variation of stress, the statical breaking stress being given. Draw two ordinates A B (Fig. 115), C D set off D to represent the statical breaking stress, and through put in through E F. Then if H k is the minimum stress on the member, H I is the maximum load which can be repeated E at 45 as a convenient slope, the horizontal scale being immaterial. Bisect D and join indefinitely without breaking the bar. The stress scale is marked in per- centage of the static breaking load. The points indicate the results of experiments by Wohler, Spangenberg, and Bauschinger, and it will be observed that theory and experiment coincide as closely as can reasonably be expected. Expressed algebraically we have, when designing a member which will be subjected to both a steady load W min . and a fluctuating load (W max . W min .), the equivalent static load. * W = W min: + 2 (W max . - W min .) The plus is used when both the loads act together, i.e., when both are tension or both compression, and the minus when they act against one another. Sankey's Hand Bending Machine. This machine has been specially designed for the rapid testing of material, and is in a convenient form for use in works. The principle on which it is based is to bend the test piece backwards and forwards until it is broken, the bending effort being measured by the deflection of a spring. A device is fitted for automatically recording not only this bending effort for each bend, but also the number of times the specimen can be bent without rupture, as well as the total energy required to break it. A diagram is obtained, the form of which shows at a glance the quality of the material. The machine gives most of the information needed in the workshop as regards the strength of the material in respect of static stresses, and hence compares not unfavourably with the more lengthy and expensive tensile tests ; but, in addition, it exhibits in a striking manner what, for want of a better ALTEENATING STEESS TESTS 185 word, may be called the " leatheriness " of the material, or, in other words, its power to resist the effect of alternations of stress, that is to say, " fatigue." One bend is denned as bending the test piece from the extreme position on the right to the extreme position on the left, or from the extreme position on the left to the extreme position on the right. The machine is illustrated in Fig. 116, and consists of a small bed-plate, arranged to bolt down to a bench, at one corner of FIG. 116. Sankey's Hand Bending Machine. which there is a grip A for securing one end of a flat steel spring B. The other end of the spring is fitted with a special grip C for holding one end of the test piece D. The other end of the test piece is fixed into a handle E, about 3 feet long, by means of which it is bent backwards and forwards through the " standard "~ angle. An indicator F is provided to show this standard angle. Alongside of the spring, and fixed to the bed- plate, there is a horizontal drum G to carry the recording paper, and the pencil H has a horizontal motion actuated by the motion of the grip C and conveyed by the steel wires L and M and the multiplying pulley N, the wires being kept taut by the spring box 0. The zero line is in the middle of the paper, and the pencil H moves in one direction when the bending 186 A HANDBOOK OF TESTING MATEEIALS is from right to left, and in the opposite direction when it is from left to right. The drum has a ratchet wheel K with a detent (not shown) worked by the motion of the pencil carrier. The result of the combined motion of the pencil and of the drum is to produce an autographic diagram such as shown in Fig. 117. Obviously the greater the stiff- ness of the test piece the more the flat spring B will have to be bent before its resistance is equal to the resistance to bending of the test piece. Hence the motion of the pencil is proportional to the effort required to bend the test piece. The test piece is properly secured in the handle E (Fig. 116) by means of the set screw, it is then inserted into the grip C, j FIG. 117. Autographic Diagram from Hand-bending Machine. and the free length (If inches) is adjusted by means of a gauge provided for the purpose, after which the grip C is tightened. The first bend is taken to the left until the mark on the handle coincides with the pointer indicating the " standard " angle. The bending is then reversed, and the test piece is bent until the mark on the handle coincides with the second pointer. The bending is again reversed, and so on until the specimen breaks. The point at which the test piece breaks should be noted in decimals of one bend, which are marked on the indicator. The " standard " angle is so fixed that the distance travelled along the arc of the circle 1 foot radius from the point of bending the test piece is 1*60 ft. (this angle is 91|). Hence by multiplying the bending effort (in Ibs.) by 1/6 the energy (in ft. Ibs.) required to make a complete bend is found. ALTERNATING STRESS TESTS 187 Generally, the stronger the material the less the number of bends it will endure, and approximately it may be taken that, in the case of mild steel, the bending effort of the first bend is proportional to the yield stress in tension of the material. It has been found by experience that with the standard test piece (f inch diameter), one half of the bending effort in ft. Ibs. is nearly equal to the yield stress in tons per sq. in. This rule is only approximate, but it will give a fair idea of the strength of the material as expressed in the .ordinary way. The number of bends is proportional to the ductility of the material, and experiment shows that this number is approximately proportional to the elongation multiplied by the reduction of area in a tensile test. The area of the autographic diagram represents the energy required to break the test piece. The recording gear has been so proportioned that 1 sq. in. of this diagram is equiva- lent to 400 ft. Ibs. The area in question can be obtained by means of a planimeter, but it can also be approximately arrived at by estimating the average bending effort, and multi- plying by 1/6 times the number of bends. 1/6 times the number of bends must be taken, because, as already pointed out, the arc swept by the point of application of the bending effort (in ft. Ibs.) is 1/60 foot for each bend. This energy figure gives valuable information as to the quality of the material, and for machinery steel should not be less than 2,500 ft. Ibs., but for the steel used in petrol engines and the like a higher figure is desirable, say 3,500 to 4,000 ft. Ibs. Many trials show that with steel in a normal condition the first bending effort is always distinctly less than the second bending effort (see Fig. 117). But if the steel has been artificially stiffened by drawing or hammering, the first bending effort is the greatest (see Fig. 117). In fact, the effect of the bending is to undo the artificial stiffening. This is a valuable and unique property of this testing machine. Fractures. The following is a short list of the fractures, 188 A HANDBOOK OF TESTING MATEEIALS with this machine, obtained with steels of good quality, and the probable inferences to be drawn therefrom : FRACTURES. PROBABLE INFERENCES. Silky "... Mild steel, nickel steel. Granular ... ... ... Medium carbon steel. Fine crystalline ... ... High carbon steel. Granular and crystalline . . . Mild and medium carbon steels when overheated. Coarse crystalline... ... Mild and medium carbon steels when overheated, and then hammered at too low a temperature. Effect of Speed. Some interesting results have been obtained by Mr. E. M. Eden 1 on a machine of the rotating beam type, in which the specimen is subjected to a bending moment and no shear. Five materials were tested at speeds of about 300, 600 and 1,300 v.p.m. Within the range of the experiments the endurance is independent of the speed at which the machine is run. The apparatus is comparatively simple and seems suitable for college laboratories. The dis- agreement of various experimenters using different methods of obtaining alternating stress shows that the subject is not yet exhausted. Work has also been done by Bairstow 2 and Howard. 3 1 Univ. of Durham Phil. Soc.,Proc. 190910. 2 Phil. Trans. Royal Society, December, 1909. :i International Association for Testing Materials, 1909. CHAPTER X THE TESTING OF CEMENTS, REINFORCED CONCRETE, AND STONES AFTER mild steel, the above are probably the most im- portant materials of construction ; all of them are very variable in their properties, and much depends on the method of testing. In the case of Portland cement we have given an outline of the methods laid down by the Engineering Standards Committee, and as so much depends on method, it is important to follow this method as closely as possible where comparative results are desired. It will be noted that even the rate of loading may considerably affect the results. Table 21 gives an idea of results obtained at different rates of loading. In the case of reinforced concrete, only those results obtained on full size constructional members or pieces of work can be considered as giving reliable results, hence we have refrained from giving results of tests except such as give what may be considered the fundamental properties, i.e., the co-efficient of elasticity, the adhesive force of iron bars, and the crushing strength, together with an account of a full size test on a floor. THE TESTING OF CEMENTS AND CONCRETES. The standard method of testing Portland cement is laid down by trie British Engineering Standards Committee ] as follows : Fineness and Sieves. The cement shall be ground to comply with the following degrees of fineness, viz. : Residue on sieve 76 X 76 meshes per sq. in. not to exceed 3 per cent. 1 The Committee's actual specification (Report No. 12. Revised August, 1910 British Standand Specification for Portland Cement) should be consulted for tests intended to comply in detail with their recommendations. 190 A HANDBOOK OF TESTING MATERIALS X--0- 06 approximately 10'approximattly 3 00' Eesidue on sieve 180 X 180 meshes per sq. in. not to exceed 18 per cent. [Sieves as per British Standard Specification.] Specific Gravity. Not less than 3*15 when fresh burnt and ground, and not less than 3*10 if cement has been ground for not less than four weeks. Chemical Composition. [See British Standard Specifi- cation]. Mode of Gauging. The cement shall be mixed with such a proportion of water that after filling into the mould the mixture shall be plastic. Fresh water to be used at a tempera- ture between 58 and 64 F. DIMENSIONS OF BRIQUETTE A suitable form of mould designed to give a form of briquette to the dimensions shown in Fig. 118 to be filled with the cement without mechanical ramming and allowed to stand on a non- porous plate until the cement sets. As soon as the briq- uette can be removed without injury, this should be done, and the briquette kept in a damp atmosphere for twenty- four hours, after which it should be kept in a bath of clean fresh water between 58 and 64 F., and allowed to remain there until breaking, the water to be changed every seven days. Testing. Twelve briquettes should be prepared for each test six to be broken after seven days, and six after twenty- eight days. The usual type of machine for testing the speci- mens is described below. When testing with standard sand, the latter to be obtained from Leighton Buzzard, and accord- ing to the British Standard Specification. The Committee lay down the following tensile strengths as the minimum allowable. Neat Test. (Average of six specimens.) 7 days from gauging . . 400 Ibs. per sq. in. FIG. 118. Dimensions of Standard Briquette (British Standard Speci- fication). CEMENTS, EEINFOECED CONCRETE, AND STONES 191 Sand Test. (Average of six specimens) (3 parts sand, 1 cement.) 7 days from gauging . .150 Ibs. per sq. in. Setting Time. Cement is said to be set when on gently Weight 300 grammes FIG. 119. Setting Needle for Cement (British Engineering Standards Committee's suggestion). applying the "needle" of the instrument illustrated in Fig. 119 no impression is made in the surface. The follow- ing times of setting define the terms slow, medium, and quick setting : Quick. Final setting time not less than 10 nor more than 30 minutes. 192 A HANDBOOK OF TESTING MATEEIALS Medium. Final setting time not less than 30 nor more than 120 minutes. Slow. Final setting time not less than 120 no more than 300 minutes. Soundness. The standard method of testing for soundness is by means of what is known as the " Chatellier test," using the apparatus illustrated in Fig. 120. The small brass mould is filled with neat cement and placed in a bath of water at a temperature of 58 to 64 F. for twenty-four hours, the two open ends being covered with glass plates. At the end of this Spring Brass suitable Metal about m in thickness FIG. 120. Le Chatellier Soundness Testing Instrument (British Standard Specification). time the distance between the pointers is measured, and the mould placed in water at 58 to 64 F. which is brought to boiling point in twenty-five to thirty minutes and kept boiling for six hours. After cooling, the distance between the points is again measured. This distance will be found to have increased, and it is laid down that this expansion must in no case exceed 10 millimetres after twenty-four hours aeration, or, if the above test has failed, 5 millimetres after seven days aeration. Cement Testing Machine. The simple form of machine, shown in Fig. 121, is used for testing cement and concrete in tension. As the strength of these materials in this direction is very small, the machine used is of a correspondingly small size, and is much simpler than those used for testing metallic specimens. The cement is first moulded into the form of a briquette, such as is shown in the figure, and is such that the CEMENTS, EEINFOECED CONCRETE, AND STONES 193 smallest section of the specimen is 1 inch square. The briquette is held in well-greased jaws of the form shown in Fig. 121. The lower of these is capable of being moved vertically by means of a hand-wheel and screw which are not shown in the diagram. The upper jaw is connected to a point B on the lever A C, whose fulcrum A is attached to the frame of the machine G. The outer end C is connected by a rod C D to the short arm of the lever D F, whose fulcrum is at E. The length of C D is made adjustable so that the lever D F may be in a horizontal position at the beginning of the test. FIG. 121. Cement Testing Machine. The load is applied by means of a variable weight H which is fixed to the end F of the lever D F. In one type of this machine the weight H consists of an iron pan into which lead shot is poured at a constant rate, so that the load uniformly increases by 100 Ibs. in every 12 seconds. When the specimen breaks the pan of shot drops on to a movable lever which it depresses, thereby automatically cutting off the supply of shot. The shot is then weighed and the strength of the cement calculated by means of the known leverages in the machine. The Bailey machine for cement testing has only a single lever, from the short end of which the specimen is gripped in the same way as in the machine just described. The long end, however, supports a long cylindrical vessel into which water is flowing from a tank above the lever during tests. The T.M. o 194 A HANDBOOK OF TESTING MATEEIALS load is thus applied as before, at a uniform rate. When the specimen breaks the downward movement of the long arm of the lever is utilised to automatically cut off the supply of water. A scale is provided on the water vessel, and is so divided that the height of water in the vessel gives the breaking load of the specimen as a direct reading. FIG. 122. Machine for Compression Tests of Stones and Cements. Compression. The resistance of concrete to compression is usually determined by crushing cubes of 4, 6, 8, or 12-inch sides at some stated age. The strength per sq. in. will in general decrease with the size of the cube. The crushing load may be obtained in any ordinary testing machine provided with compression plates or, where much of this class of work has to be done, on a machine specially built for cement testing. These special machines in general consist of CEMENTS, EEINFOECED CONCEETE, AND STONES 195 an ordinary hydraulic press, but worked with oil or glycerine, in which the piston friction is so far reduced that a gauge attached to the main cylinder may be made to read the crushing load direct. This gauge must be so constructed that after the material fails the pointer will still indicate the maximum pressure, as failure generally takes place suddenly and without previous cracking of the material. Messrs. Bailey make such machines, capable of exerting a crushing load of 12, 60, and 150 tons respectively. It is most important in the testing of cements, concretes, stones, etc., that the bedding should be absolutely even, as otherwise splitting at the highest corner will take place. To ensure this even bedding plaster of Paris is generally used. The compression plates are first cleaned and then slightly oiled. A thin paste of plaster is then put on the lower plate about a quarter of an inch thick. The block is carefully bedded on to this and another quarter of an inch put on top of the block. A small amount of pressure is then applied, and the machine allowed to stand until the plaster is set, say, in about five to ten minutes. Loading should be applied slowly and evenly till fracture occurs, which in homogeneous material tends to take place by shear at 45, thus forming cones in the case of cylinders, or pyramids in the case of cubes. Both the tensile and compressive strength are very variable, especially the former. The chief conditions which determine the strength are : 1. Proportion of ingredients. 2. Quality of ingredients. 3. Amount of water used 4. The method and amount of mixing. 5. Amount of consolidation effected. 6. The form of the piece. 7. Atmospheric conditions during hardening. 8. Time after gauging. 9. Manner and speed of applying load. The following tables of experimental results will illustrate the effect of the above : o 2 196 A HANDBOOK OF TESTING MATEEIALS p I t ed puug s^a^d g puug s^JBd ^ 85 GO ** C5 CO ^M 1 (N i I 1 1 1 1 C5 GO 1 (M -" O ! - CO CO l>- GO OO5i-HCcecio^^H 1C CO Ci tC O O r-H (M T^TT 's t- ti t- cc co I ^ Fine Per Pas eiimbs -in 9HX). sapH '031 'ON 9A9I8 CO -*i 1 w Icb IGO ' b aiBubs -in 0900. sapH 001 'OR 8A8ig i^t CO l> ?O O O S O C5 CO 'COCOCOGOCOCOOS A-l i-l rl Ditto 3 months 590 617 622 640 5 ) 12 3 months 445 467 487 507 510 Natural Neat cement 7 days 150 169 186 202 Ditto 3 months 309 351 363 378 390 5 1-2 3 months 255 299 327 329 354 TABLE XXII. EFFECT OF PROPORTION" OF SAND. H and E are two samples of Portland Cement. Sand used, Eiver Sand, "Point-aux-Pins." Tensile Strength, Ibs. per sq. in. Parts Sand to 1 Cement by Weight. Proportionate Strength, Two years if 1 2-100. 6 Months. 2 Years. H R II R Mean. 2 512 504 534 548 541 100 3 390 335 363 355 359 66 4-09 295 261 296 288 292 54 6 175 144 191 174 182 35 8 113 96 132 132 132 24 10 64 74 104 116 110 20 1 The rate of loading specified by the B.E.S.C. for testing Portland cement is 100 Ibs. in 12 seconds, i.e., 500 Ibs. per minute. 198 A HANDBOOK OF TESTING MATEBIALS l w .J c> "^ m ^ - Si - 1 1 1 1 1 1 3 1 1 8 a= 1 i portional <^sO Ci 00 23 , 1 = 8- 1 i * 00 rH rH -H CO ** t> 1C * CO CO 1C (M C rH QC CO O IIS 00 OS 00 I (N (M M (M IM rH I 1 1-1 (M rH rH 11 5 1 1C CO CO OD rH rH S SI! 1 O GO 1C o GO CO CO CO CO CO (M - rH GO 1C OS -* ic m co 1C CO Ci CO 33 CO CO CO CO CO > 1 .2 1 S^ g 5 g, 3 ^ I 1 | 4 1 s 1 s n .2 1^ fl c3 .ll| d I c3 3 o ce s rig 8 o o d 'o Y " ^ 3 ^3 g JFO ce *-< V I 8 O l J| "S I .2 s - "cu S * o o> fjli 1 "o 0. S d 1 | o, '" ^ ^ O CEMENTS, EEINFOECED CONCEETE, AND STONES 199 * 1 i 1 00 1 1 1 t 1 ! I i portional O b- >O '^>O OS OS 1 1 1 i o os os bb CO !>'*& 05 CO i-H ^ s to 00 O CO r 3 ^d 9 II * H M OJO < O CO OS 1 ^. ^ r 71 ' ' "* i-H d H 8 >3 ^S j GO a j3 ^ j f is * S bjo a S 60 s 1 1 00 o U Hi 1 .2 ill M 0> K ^ 1 'o S S O" S s o* -+J 6 f .2 o> ^ *o d o a> S Q p . r-3 ^ o c3 bo 1 i 8 1, d s| 111 W O 1 w H OQ O flJ -*J OD O 1 a o ^ x"5 So ^O 1 &K* Szj (2 1 1 a> g S-S o ^ Q 02 1 ^T^ D s S cs "o a 200 A HANDBOOK OF TESTING MATEEIALS The Application of BrinelFs Ball Test to Cement Testing. La Revue de Metallurgie and Le Genie Civil have recently published articles dealing with the testing of cement and concretes of all kinds by means of Brinell's Ball test. A rough and ready method of a similar nature has long been 20 19 .18 5 3.10 159 0-079 0-083 0-22 5 5 3.25 ,, 130 0-071 0-077 0-193 55 5 3.45 ,, 100 0-055 0-063 0-161 11, , 7.45 a.m. 100 0-063 0-067 0-165 12, 2.45 p.m. 100 0-067 0-071 0-173 55 > 3.0 72 0-043 0-055 0-157 55 5 3.15 42 0-043 0-047 0-126 55 1 3.45 ,, 13-5 0-027 0-037 0-095 55 5 4.15 0-016 1 0-023 0-063 13, , 9.0 a.m. 0-016 1 0-012 0-051 14, , 12 noon 0-016 1 0-008 0-043 15, , 11.50a.m. 0-008 o-ooo 0-043 1 1 Instrument has probably been disturbed. span and the other two at the distance of 2 feet 9 inches from the supports. Loading was effected by means of bricks and Portland cement in bags. The results of the tests are sum- marised in the subjoined table, from which it will be seen that the maximum deflection under double the normal superload of 84 Ibs. per sq. foot was only 0*139 inch, calculated over the central 14-foot portion of the beam, or j^to f tne span. No records were taken of the settlement at the supporting walls, so it would not be fair to regard the maximum deflection of 0'22 inch as having taken place in the floor itself. But 206 A HANDBOOK OF TESTING MATERIALS adopting that value for the sake of argument, it will be seen that the proportion is only joVo f the entire span, or less than one-half the proportion generally allowed by architects. Another point worthy of special note is that on removal of the loading the floor returned practically to its original form, thereby demonstrating the perfect elasticity of the construction. Stones, Bricks, etc. These materials are tested in com- pression in a similar manner to that described for cements. The following table indicates some results of such materials taken from Popplewell's " Testing of Materials of Construc- tion " : TABLE XXVII. CRUSHING STRENGTH OF VARIOUS STONES, ETC. Material. Authority. Crushing Strength. Tons per sq. ft. Lbs. per sq. in. Granite, Aberdeen Grey . Unwin. 1,412 22,000 . Eed . M 1,614 25,100 Basalt, Penmaenmawr . Fail-bairn. 1,086 16,850 Sandstone, York Grit Unwin. 712 11,050 ,, Eed Mansfield 609 9,560 Eed Alton . _ . 309 4,800 Limestone, White Italian Marble ..... Eennie. 1,400 21,800 Limestone, Portland * . . Unwin: 516 8,020 ,, Purbeck . . Eennie. 587 9,110 ,, An caster . ':.. [ Eoyal 150 2,330 ( Com. ,, Bramham Moor . y > 380 5,900 Bricks, London stock, average Unwin. 140 2,180 ,, Leicester, wire cut, average . 290 4,500 ,, Staffordshire, Common Blue . " 400 6,210 CHAPTEE XI THE TESTING OF TIMBER Tension Tests. Considerable difficulty is encountered in the testing of timber in tension owing to the tendency of the material to crush in the grips or to shear. The portion of the specimen which enters the grips should be large in proportion to the breaking section, and should be extended for some distance out of the grips before being gradually reduced. If it is difficult to develop the full tensile strength of timber in a testing machine, it is still more difficult to do so in structures built of this material ; failure invariably taking place by shearing or splitting at both connections. For this reason tension tests in timber can have little more than academic interest except in a few isolated cases. Timber is furthermore extremely variable in its properties, and, like all other materials, it is of vital importance in cases of important structures that strength calculations should be based on tests specially carried out on actual samples of the material employed. The amount of moisture, for instance, will greatly affect the results. In the adjoining tables, however, we give some standard results which will indicate what may be expected in testing. Compression Tests. Unlike metallic materials, the full compressive strength is frequently developed in timber struts, and the true crushing load becomes of importance. Tests are generally carried out on cubes or short cylinders, which should be as large as the testing machine employed will conveniently take. It is sometimes useful to embe'd the ends on a sheet of millboard, especially if the ends are rough. Shear and Cross-bending Tests. Timber is in the majority of cases only stressed to its full capacity when employed as 208 A HANDBOOK OF TESTING MATEEIALS beams, and hence the most important tests are those in which the material is tested in the same manner. "Whenever possible full-size beams should be employed and tested in a similar manner to iron girders, except that the ordinary knife-edges should be prevented from penetrating the fibres by placing a piece of iron plate between the knife-edge and the wood. If the depth is at all great compared with the length, failure will invariably take place by shear along the length of the beam, and hence the shear strength becomes of consider- able importance. Shear strength can be deduced either from beam tests carried out so as to cause the material to fail in this manner, or by direct experiment. In all practical struc- tures shear failure will take place along the fibres, the strength across the grain being much more than along the grain. With a beam supported at the ends and loaded centrally, the maximum bending moment is WL 4"' " W " being the central load and " L " the span. But bending moment M = SZ, where S = stress in the outermost fibres, and Z = modulus of the section. With wood, however, it is usual to introduce a constant into this equation : . . M=KSZ, WL but M:=- , bd? And " Z " for a rectangular beam where 1= width of beam, and d= depth of beam. WL For samples of the same wood " S " should be approximately constant. THE TESTING OF TIMBEE Then uniting all the constants 209 Then K=WL bd*' This expression should remain fairly constant for samples of the same wood, and consequently is often used com- mercially as a standard of comparison. TABLE XXVIII. TESTS OF TIMBER IN TENSION AND COMPRESSION, BY MR. T. LASLETT. Carried out on tension specimens 2 inches square, 30 inches long, and compression specimens about 1, 2, 3 and 4 inch cube; crushed in direction of fibre. Kind of Timber. Sp. Grav. Ult. Resist, in Ibs. per sq. in. (Tension.) Ult. Resist, in Ibs. per sq. in. (Compression.) English Oak (unseasoned) 0-858 3,837 4,900 ,, ,, (seasoned) . . 0-893 7,571 7,480 French Oak . . . . 0-976 8,102 7,950 Dantzic Oak . . . . 0-838 4,217 7,480 American White Oak 0-969 7,021 6,070 American Oak (Baltimore) 0-762 3,832 5,890 African Oak (Teak) 0-971 7,052 Teak, Moulmein . , . 0-777 3,301 5,730 Iron Wood, Burmah 1-176 9,656 11,670 Chow, Borneo .... 1-134 7,199 12,590 Greenheart, Guian . ' . 1-141 8,820 14,420 Sabien, Cuba . . ' ; . 0-917 5,558 8,470 Mahogany, Spanish ,, Honduras 0-765 0-659 3,791 2,998 6,400 6,380 ,, Mexican 0-655 3,427 5,600 Eucalyptus, Australian . Tewart . . T.. . ."- 1-169 10,284 9,350 Mahogany . ; > ". 0-996 2,940 7,170 Iron-bark ,-',". . 1-150 8,377 10,300 Blue Gum .... 1-049 6,048 6,900 Ash, English .... 0-750 3,780 6,970 Ash, Canadian 0-588 5,495 5,490 Beech ..... 0-705 4,853 Elm, English .... 0-642 5,460 5,780 Eock Elm, Canadian 0-748 9,182 8,580 T.M. 210 A HANDBOOK OF TESTING MATEEIALS TABLE XXVIII. -Continued. Kind of Timber. Sp. Grav. Ult. Resist, in Ibs. per sq. in. (Tension.) Ult. Resist, in. Ibs. per sq. in. (Compression.) Hornbeam, England 0-819 6,405 8,310 Fir, Dantzic . . . . 0-603 3,231 6,940 ,, Eiga .... 0-553 4,051 5,240 Spruce, Canadian 0-484 3,934 4,850 Larch, Kussia . . 0-649 4,203 5,820 Cedar, Cuba .... 0-469 2,870 4,480 Red Pine, Canada . . . 0-553 2,705 5,690 Yellow Pine, Canada 0-551 2,759 4,210 Pitch Pine, American 0-659 4,666 6,470 Kauri Pipe, New Zealand 0-544 4,040 6,430 Hatfield's experiments.* Georgia Pine, American . Locust, American . White Oak, American Spruce, American . White Pine, American . Hemlock . ... 16,000 24,800 19,500 19,500 12,000 8,700 \ * These experiments were carried out on specimens only 0'35 inch round, such a size being far too small. TABLE XXIX. LAWSA'S TESTS OF AMERICAN TIMBERS, 12 FEET AND 2 FEET LONG; FAILURE BY DIRECT CRUSHING. Name of Timber. Sectional Area in sq. in. Ult. Strength in Ibs. per sq. in. Coeff. of Elasticity, Ibs. per sq. in. Yellow Pine . 42 to 102 4,544 1,996,351 White Oak . 32 to 93 3,470 1,398,908 Old and seasoned White Oak . 28 to 87 3,957 1,817,539 THE TESTING OF TIMBER 211 TABLE XXX. KIRKALDY'S EXPERIMENTS ON BEAMS. Description of Timber. Breadth and Depth in inches. Span in feet. Modulus of Rupture in Ibs. per sq. in. Modulus of Elasticity, Ibs. per sq. in. Pitch Pine ( from IMSX 11-30 ( to 13-10X13-10 12 7,626 Daiitzic Eir (from 10-00x12-00 ( to 13-25x14-38 from 8 to 12 4,581 >i (from 2-50x10-10 ! to 3-00 X 10-10 j 10 3,726 571,760 Baltic Oak. 6-4x16-00 10 7,686 Baltic Eed . (from 11-72x11-82 ( to 11-77x11-86) 12 4,890 English Oak (from 4-55x12-00) ( to 4-58X12-00 10 9,762 St. Petersburg . ( 3-09x11-07 ( 3-08X11-02 13 8,187 2,446,000 St. Petersburg 1st Yellow . (from 2-75 X 8'7o ( to 3-00 X 8-75 10 8,556 1,677,500 St. Petersburg 2nd Yellow . ( from 2-87 X 8'75 \ to 2-99 X 8-75 10 6,918 1,396,700 Archangel . (from 3-00x11-06 ( to 3-09x11-02 13 6,738 2,014,300 Archangel Deal 2A . 3-00x3-00 10 6,252 2,043,000 Swedish (from 3-08X11-07 ( to 4-10X 9-10 10 to 13 5,663 1,838,300 Swedish SS ( from 3-00 X 9'10 ( to 3-15 X 9-10 10 6,258 1,149,600 Swedish DDD . (from 2-93 X 8'75 1 to 2-95 X 8-75) 10 6,978 1,528,700 The Modulus of Elasticity is generally deduced from beam tests, but in employing figures thus obtained it should be remembered that time has considerable effect on the elastic properties of timber, and tests extended over long periods have shown that, roughly speaking, the permanent deflection attained after six months or more may be taken as at least twice the value obtained on tests of short duration. Tables XXIX. and XXX., taken from different sources, give results obtained by the above methods. p 2 212 A HANDBOOK OF TESTING MATEEIALS TABLE XXXI. SHE AR STRENGTH OF TIMBER. Kind of Timber. Shearing Strength, Ibs. per sq. in. Authority. Maximum. Minimum. Ash . 700 458 Watertown Arsenal Yellow Birch , 815 563 tests White Maple 647 367 EedOak . 999 726 White Oak . 966 752 White Pine . 366 267 Yellow Pine . 415 286 Spruce . . , 374 253 Whitewood . 406 382 Eed Fir A . , 517 146 Ha tt.* B . 273 173 . 395 74 Longleaf Pine (Georgia) 291 247 * These results were obtained as the resistance to splitting due to longitudinal shear under cross bending. Ligno- Concrete. The author has recently made some rough tests 1 on concrete reinforced with wood. Interesting results will be obtained if hard and soft woods are used. Com- parisons may also be made with concrete reinforced with steel. Further tests may be devised by varying the shape of a wooden framework round which is placed the concrete. Enginyering, 1910. CHAPTEE XII EXPERIMENTS IN COLLEGE LABORATORIES ONE of the regular parts of every engineering student's course of study is the carrying out of tests in a materials testing laboratory, and the following is suggested as a suitable and systematic series of experiments. It is assumed that the student has already completed a first-year's course in applied mechanics, and is familiar with the use of verniers, micro- meters, microscopes, etc. Few students will have the opportunity to carry out all of the tests mentioned, and much must, of course, depend on the resources of the particular laboratory in which the student is working. It is not even suggested that the experiments should be carried out strictly in the order given, although an attempt has been made to arrange them as far as possible in the usual order in which they should be performed. The general instructions to the student should be carefully read and, subject to the dis- cretion of the particular professor under whom the student is working, adhered to faithfully. It is hoped that the tables prepared for " setting " these experiments will be found useful by both students and demonstrators. They are intended to show at a glance the work which has already been done by the student, and suggesting fresh experiments to be performed. GENERAL INSTRUCTIONS TO STUDENT. Note-Books. Each student should be provided with a note-book for entering records of all laboratory tests and illustrative sketches. In any case this book should have good paper and be provided with stiff binding. The following method has been found excellent for students preparing for a degree in engineering. 214 A HANDBOOK OF TESTING MATERIALS All notes and descriptions should be on separate sheets of thick foolscap ; curves on foolscap size sheets of squared paper ; diagrams on drawing paper and photographs pasted on same. All these can be bound up at the end of the session at quite a small cost. Needless to say, blank pages should be left for further .additions. Such a book, even apart from examinations, is an excellent record of a student's work and neatness, suitable for showing to a prospective employer. In all cases a good wide margin should be left at the side of the page ; when separate sheets are used an allowance for binding is necessary. On no account should different subjects (such as heat engines and materials) be mixed in the same note- book. Before starting an experiment the student should make a sketch of the apparatus to be used, employing diagrammatic sketches rather than scale drawings. He should then calculate approximate data so as to know what to expect during the experiment. Thus in the case of the determination of Young's modulus he should measure the specimen, and look up in the tables the stress at the elastic limit of the material. He will then be able to form an estimate of the load which the specimen can safely withstand without damaging the instru- ment. This must be well within the calculated elastic limit of the material. In marking out specimens centrepops should be light, as in testing to destruction it is possible that they may have a considerable effect on the ultimate strength. The greatest care should be taken with instruments and apparatus of all kinds. Instruments, of precision, such as extensometers, cannot be handled too carefully. Before testing the student should enter in his rough book the date, and as far as possible every dimension which can possibly affect the result. Nothing is more annoying than to find, after a long and careful experiment lasting over, perhaps, two or three days, that some vital dimension was not taken at the beginning, and as a consequence the whole test spoiled. During the test every reading should be entered in the rough EXPERIMENTS IN COLLEGE LABOEATORIES 215 book directly it has been taken ; never trust the memory more than a few minutes when carrying out scientific work. . As soon after the finish of the test as is possible a full report of the experiment and apparatus used should be written up in the recording note-book stating: (a) object of test; (b) apparatus employed, including details, with sketches, if necessary, of such parts as grips, shape of specimen, etc. ; (c) order of making observations ; (d) record of observations ; (e) deduced results showing carefully how such were obtained; (/) graphical representation of results when possible ; (g) comparison with standard laws or results to be verified. Wherever possible, photographs of apparatus, specimens, fractures, etc., should be pasted into the note-book. In describing experiments it cannot be too strongly impressed that language is " the first tool of the mind." Great care should be taken by the student in expressing clearly and in suitable words any work upon which he has to write a report. Inaccuracy of expression is as great a source of error as inaccuracy of observation. Huxley's theory of style was " to say that which has to be said in such language that you can stand cross-examination on each word." EXPERIMENTS SUITABLE FOE COLLEGE LABORATORIES. On Wires and Springs. 1. To Determine the Relation between Load and Extension of a Spring. The apparatus is usually found set up, and consists simply of a spring attached to a hook at the top and provided with a scale pan or similar contrivance at the bottom. A vernier moving over a scale gives the deflection. Load with increasing weights so as to get ten or more readings, and plot a curve showing the above relation. It should come out a perfectly straight line. Deduce the " slope " of same and note. 2. To Determine Relation between Load and Compression of a Spring. This is carried out in an exactly similar manner to experiment 1, and similar results deduced. 3. Extension of a Short Wire to Determine Young's Modulus. See page 165 for general description. Care should 216 A HANDBOOK OF TESTING MATEEIALS be taken not to overload the wire. Take as many readings as possible, plot curve as in experiment 1, deduce Young's modulus from slope of curve and the dimensions of the wire. This experiment may be repeated for wires of different material. 4. Extension of a Long Wire. Some laboratories are provided with a very long wire (say 90 feet or more) running on pulleys down the laboratory. Readings as in experiment 3, but only scale and vernier necessary for extension. Take readings with increasing load and decreasing load ; plot curves for both sets of readings. The curves will not coincide owing to friction of pulleys, hence take mean of the two sets and deduce value of as before. 5. Wire Stressed to Fracture on Autographic Apparatus. (See description of this experiment on p. 166.) It is not desirable to ink in autographic diagrams if they are drawn by a pencil apparatus, but, if faint, it is allowable to dot with a sharp pencil or a pen along the line traced out. Copies may be taken with tracing paper. 6. 1 Value of C by Torsional Deflection of a Wire. (See p. 135.) Take increasing and decreasing loads as before and obtain mean value to eliminate friction. Value of C deduced from the formula n _584 Ml 0D 4 ' where M is twisting moment in Ibs. inches, / the length in inches, D the diameter of wire in inches, and 6 the twist in M degrees. Obtain the mean value of -^ by plotting a curve showing relation between them. This should come out to a straight line. 7. 1 Value of C by Torsional Vibrations of a Wire. (See p. 136.) Value of C deduced from 0=^ gd* \_ t where mi, ???2, #, etc., have the values given on p. 136. 1 If same kind of wire is employed for experiments 3, 6, and 7, the value of oisson's ratio Poisson's ratio. Poisson's ratio should be deduced from formula E = - " ' where - is n n EXPEEIMENTS IN COLLEGE LABOKATOKIES 217 8. Spring Tested in Extension for Value of C, and 9. Spring Tested in Compression for Yalue of C. These experiments can be carried out in a similar manner to experi- ments 1 and 2, but for large springs either a special apparatus is employed, or in the case of 9 this can be performed in an ordinary testing machine arranged for compression. Deflec- tion can be taken in the latter case by measuring with an internal micrometer between the compression plates. . ~ 2-55 D 2 LW C is obtained from the formula C= -^ , O Ci where D is mean diameter of coils, L total length of spring if pulled out to a plain rod (approx. nirD), d is W diameter of wire, and -r- relation between load and deflection o obtained by plotting a curve. EXPERIMENTS WITH TESTING MACHINES (TENSION). 10. Testing Small Specimen for Yield Point and Fracture in a Small Testing Machine. Some such machine as the Bailey machine can be employed (see p. 121). As many materials as the student has time and opportunity to test should be employed; in any case (a), 1 (c), and (e). 11. Calculating Mechanical Advantage of a Large Testing Machine and Checking for Sensitiveness and Accuracy. See p. 47 for such a test fully worked out. 12. Yield Point of Full Size Specimen by drop of beam. This will also give experience in working a machine and setting up specimens (see p. 85). Materials suggested, (a), (b), (e), and (g). 13. Fracture of Various Materials. This involves the marking out of specimens, setting up, determination of yield point and measurement of elongation per cent., reduction of area, etc. (see pp. 86, 87, and 89). Materials suggested, (a), (b), (c), (d), (e), (/), (0),and(fc). 14. Full Commercial Test. Test, say, half a dozen 1 These letters refer to mild steel (), wrought iron (J), cast iron (c), copper (d), brass (e), gun metal (/), aluminium (rolled) (#), tool steel (/), and (c) 27 Alternate tension and com- pression . . 28 Deflection of beams. 29 Breaking of beams . 30 Elastic curve of beams . 31 Determination of C (torsion test) ... EXPERIMENTS IN COLLEGE LABORATORIES 225 TABLE XXXII. Continued. Experi- ment No. Description of Experiment. Set by. Date. Examined by. Date. 32 Fracture of torsion specimen . 33 Effect of flaws and surface markings (torsion) 34 Impact tests . . 35 Shear tests . . . 36 Punching tests . . . 37 Cold bending tests . 38 Hammering tests . , . , 39 Rough microstructure examina- tion ' 1 . . . 40 Repeat stresses , . ' . 41 Combined bending and torsion 42 Combined torsion and direct stresses .... 43 Tests on balls . 44 Hardness tests . , . ' . 45 Comparison of tensile strength and hardness .... 46 Testing thick cylinder 47 Timber test in tension T.M. 226 A HANDBOOK OF TESTING MATERIALS TABLE XXXII. Continued. Experi- ment No. Description of Experiment. Set by. Date. Examined by. Date. 48 Timber test -crushing test 49 Shear test on short beams 50 Testing of long timber beams . I * 51 Testing of long timber struts . 52 Specific gravity of cement 53 Time of setting test . 54 Tensile tests with cements, etc. 55 Compression tests ,, ,, 56 Soundness tests of cements 57 Compressiye tests on stones . 58 Tests on ferro- concrete beams . 59 Tests on wood-concrete beams . APPENDIX I. STANDARD RESULTS OF TESTS ON THE STRENGTH OF MATERIALS. THE following tables, taken from various sources, give the results obtained on various materials and will indicate the kind of results to be expected in practice. It will be found on comparison of results from different sources that there are often wide differences ; from which it will be seen that in all important machines and structures, where the material is to be used in the most economical manner, it is essential that samples and specimens of the actual material employed should be tested and under as near the conditions of use as is possible in the testing laboratory. TESTS IN TENSION, TORSION, AND SHEAR ON THE CHIEF MATERIALS OF CONSTRUCTION. 1 TABLE XXXIII. TENSION. Material. Specimen. Elastic Limit. Lbs. per sq. in. Breaking Stress. Lbs. per sq. in. Wrought iron, Nether- ( ton Crown best. ( 1 2 31,970 35,550 47,950 48,850 Bessemer steel. 1 2 69,760 70,700 116,930 108,550 Crucible steel. 1 2 67,500 71,680 113,020 120,680 Rivet steel. I 2 40.000 40,190 65,500 62,630 Crown rivet iron. j 1 2 37,500 37,970 56,700 55,300 Cast iron, skin on. 1 2 3 28,310 22,140 26,380 Cast steel (cut from ( casting). ( 1 2 38,350 38,870 85,700 84,850 1 Proc. Inst. Civ. Eng., vol. xc., pp. 396 407. Q 2 228 APPENDIX I. TABLE XXXIII. TENSION C v on// //<< /. Material. Specimen. Elastic Limit. Lbs. per sq. in. Breaking Stress. Lbs. per sq. in. Cast steel, in compres- j stow. ( 1 2 39,010 39,860 Siemens-Martin steel, j 1 9 37.060 35,760 57,500 57,900 Wrought iron, S.C. ( Crown. ( 1 2 38,260 38,490 54,330 55,690 Muntz metal bar. 1 2 25,000 25,000 57,500 56,540 Gunmetal, Cu. 64 parts, ( Sn. 8, Zn. 2 parts. ( 1 > 17,500 15,000 29,070 32,380 TABLE XXXIV. TORSION. Material. Specimen. Elastic Limit, Lbs. per sq. in. Breaking Stress. Lbs. per sq. in. "W rought iron, Nether- \ ton Crown best. j 1 2 3 20,560 18,700 18,900 57,800 54,900 56,600 Bessemer steel. 1 2 3 46,400 45,400 44,500 101,000 99,460 99,550 Crucible steel. 1 2 3 43,100 43,600 43,300 97,900 90,000 96,800 Landore rivet steel. 1 2 3 37,200 23,200 22,400 78,700 66,840 67,100 Netherton Crown rivet j iron. | 1 2 3 22,950 21,600 25,200 64,700 64,700 64,600 APPENDIX I. 229 TABLE XXXV. TORSION Continued. Material. Specimen. Elastic Limit. Lbs. per sq. in. Breaking Stress. Lbs. per sq. in. Cast steel. 1 2 3 24,300 23,500 22,400 78,200 78,250 77,000 Siemens Martin steel, j 1 2 3 24,200 21,800 21,800 65,300 63,000 60,600 Wrought iron, S.C. \ Crown. | 1 2 3 22,950 22,100 23,700 67,400 62,700 68,400 Muntz metal. 1 2 3 19,700 19,200 19,700 59,000 57,600 59,000 Gun-metal, Cu. 64, ( Zn. 2, and Sn. 8 parts. ) 1 2 3 12,100 12,100 12,100 33,800 , 36,500 36,200 Cast iron ; Turned. 1 ( part Dalmelington, 6 j parts best scrap. ( Highest Lowest Mean of Six 40,100 28,450 33,040 Cast iron, skin on. 1 > 3 46,800 36,600 38,500 Cast iron ; Turned, 1 <' part Sunder] and, 3 J parts best scrap. \ Highest Lowest Mean of Six 41,800 33,900 38,200 Cast iron. Skin on. < Highest Lowest Mean of Five 37,500 32,150 34,330 230 APPENDIX I. TABLE XXXYI. SHEAR on " RIVET" TESTS. Material. Specimen. Shear Strength. Lbs. per sq. in. Material. Specimen. Shear Strength. Lbs. per sq. in. Wrought iron, / Netherton \ Crown / best Highest Lowest Mean of 6 44,350 40,000 42,050 Wrought- i iron, S.C. Crown Highest Lowest Mean of 5 46,950 46,030 46,510 Muntz ( metal | Highest Lowest Mean of 6 42,860 40,270 42,000 Bessemer \ steel | Highest Lowest Mean of 6 81,910 76,780 78,880 Gunmetal, / Cu. 64, Sn. 8, j Zn. 2 parts Highest Lowest Mean of 6 30,650 22,630 27,960 Crucible I steel | Highest Lowest Mean of 6 76,200 73,500 74,500 Landore rivet ) steel f Highest Lowest Mean of 4 53,180 50,540 51,570 Cast-iron. r Turned. \ Dalmeling- / ton, 1. Best \ scrap 5 / parts Highest Lowest Mean of 11 13,860 10,280 11,860 Netherton ( Crown rivet iron Highest Lowest Mean of G 48,800 47,160 47,870 C.I. Turned. Sunder- ( land, 1. Best scrap, f 3 parts Highest Lowest Mean of 12 ! 13,740 9,280 11,420 Cast steel Highest Lowest Mean of 5 63,520 58,720 60,160 Siemens- i Martin steel i Highest Lowest Mean of 5 48,300 45,850 46,910 Ditto, ditto. ( Skin on j Highest Lowest Mean of 6 13,920 6,740 8,810 APPENDIX I. 231 8 SSfloS8 .8 3 2i~ 5 i iTiiigTi I 1 1 issss MI- 7 ^^^ *- 5-g . J3J - fl |8g w a 1 1 1 1 3 -^a 5 8 2 33 a 1 1 1 a 1 ^ = =i i i i i i i i i i ill i il ^s| II -- i- i in 1 i! 1 1 |l t t-5OO>f~ 2fsl = Eel - S rf g = = & i 1 - H " "So - Sfeo2" ^ fcC_ -g - 03 ^fe o SIS 1 ill Si. g s (MIMC^ | rtrfof .p' s s ^^ gSs'gg tf'sS |!.g ^'S oT 2.S- o p -1 it feOO p H ^i astic Limi 3. per sq. i iil 1 1 1 lift *S "?N ifelfe Irs 1 ss ' S ^ C^ C 1 * o BJ M S, 2 Drawn, 7,713,000 (WO Annealed, 7,555,000 (WO ^ "^" ^'0 |1^5 1 rt si ? Mirror, 8,792,000 (WO Crystal, 5,830,000 (WO fill Drawn, 2,564,000 (W.) Annealed, 2,457,000 (W.) - s | i a 1 6 1 O "o O 1 APPENDIX I. 233 |5' ? = 2 S >^ N - x ^** 5 ft IS 1 Pg^g ofco" of 1 ^ d.S _ Is 4 I s i IS, CO CO i l| i-i rH a* 5 ^ tc fcjj s''^ B 1 & '5 1 ^ s 1 b ^ () 3? "*. '"- ** o "3 i I 3 1 2 i = O sj II tc I?* pq | 55 8 cC S o" & ^S- -3 ll CO 5 2 ISII II i I 1 coVfo^ ->| o p s" CO 1 s IS l?li r &~~^' ^ |pgg i il isli I|ll _0 5 " s 1 . Fraction of Tension Yield Fra Com o? Yfe O X o >.. JL a FIG. 130. Curve showing Method of Plotting Results of Compound Stress Experiments. Mr. Mason's Experiments, At the University of Liverpool Mr. William Mason, M.Sc., has made experiments upon tubes subjected to compression and internal or external fluid pressure. An ingenious apparatus for holding the tubes in tension was also designed. Mr. Mason recognised the great necessity for, and difficulty of, obtaining axial loading in compression. There was also very considerable trouble in making reliable tests under simultaneous axial and hoop compression. However, most of these were overcome. The experiments recorded in Fracture of Cast-iron Specimens in Combined Torsion and Bendin< PLATE IV. APPENDIX III. 251 Mr. Mason's paper " show an approximate agreement between the maximum shear-stress at the yield-point in compression and the yield- point in pure shear, the mean difference in the tests of annealed specimens being about 3 per cent. " Mr. Mason concludes by saying : "It appears, then, that mild steel in direct compression yields by shearing ; and to a first approximation that the value of this shear-stress is indepen- dent of any normal compressive stress on the planes of the slide." It is worth noting that the Report of the Steel Committee of Civil Engineers, as far back as 1870, included results which show very close agreement between the yield-stress in tension and compression of steel and wrought -iron bars. Guest's Law. It is suggested that the bulk of the evidence furnished by these experiments proves that Guest's law for the failure of ductile materials is accurate enough for design purposes. Brittle Materials. Mr. Walter Scoble and Prof. Goodman (of the University of Leeds) have made experiments upon brittle materials such as cast iron and hard tool steel. They find that at failure the maximum principal stress is constant. Fracture is the most satisfactory criterion of strength for a brittle material. The table on page 252, from Mr. Scoble's latest tests, supplies evidence to justify his conclusions. It is from a paper recently (1910) presented to the Physical Society. Plate IV. shows the type of fractures obtained. Prof. Goodman has made experiments and has published the following table, showing the relation between the angle of fracture and the principal stresses for cast-iron bars I- TABLE XLIII. Twisting Moment. Lbs. in. Bending Moment. Lbs. in. Equivalent Twisting Moment. Modulus of Rupture. Tons per sq. in. Angle of Fracture. Actual. Calculated. Zero 2,300 4,600 25-5 777 1,925 4,000 26-7 12 11 1,170 2,240 J,750 27-1 14 14 1,228 2,255 4,820 23-1 17 15 1,308 2,128 4,628 24-0 19 16 2,606 1,375 4,320 20-8 33 31 2,644 766 3,520 16-2 38 37 3,084 Zero 3,084 16'0 43 45 Pure shear. 13-0 QO \ Mean of ,, tension. 11-5 ()0 numer- ; ous tests. COMBINED STRESS TESTING MACHINE. A machine has been erected at the Glasgow and West of Scotland Technical College designed to give combined tests in tension and torsion. 252 APPENDIX III. I !.!> P |"U5 I'M Ifc Ifc O II tf O 6 6 o""" 6 oK s 00 .III g g oa o o o O O 1C t, .. ^ o o as o o 888 o o o O GO OS o o o o o o o o 88 S8 OS O tC (M ^1 Oi C^ CO ^o *o i CO CO r-H --' J -' "i^l 1C t- GO r-H -H 00 C^ 1C GO ~t- I CO ^ C 1C OS Minimum Principal Stress. Lbs. per tq. in. |oo 1 i? ^ o co~ GO" * 1^. 1 1 ss r-^CM^ (M t^ 1 1 o QC_-*_ CO l^ 1 1 g cs^ f-^ 10,800 o o To" M o o 1 -' 7 .5 'G ^ 000 S8S CO (N O o o o o o o (M -H C. 8 11 o o r o GO 1C ** 5 CM iM QO t- OO ggg OO t^ GO CO iC CO OS t^ CO CO >C CO 05 1C iQ CO t- 1C * ||| o o s o o o 2R - 2 ^H ^ t.o o - I ( t^- CO ce O' O 1^ ^S 02 l> to t>- "* ^ 1C t- CO C^J 1C * r- "5 J Ll* "g o5 - O CO 000 Sol 8 8 o -^ -f O i CO O i^H t- CO ^ I- 1 - 1C GO 5 % t - 1 1 1 ( s CO 1C |.s loo ooq: 05 GO t^ }l, where 6 and ^ are the alterations of angle at the ends and 11 is the length of the specimen under observation. Since the bending is uniform 8 <$> and no correction is necessary. Bending in any other plane can be resolved into components in the vertical and horizontal planes, and therefore falls under the preceding cases. In order to effect the adjustment required, both the wire and the microscope slide in adjustable tubes provided with graduated scales, and the movement to bring the wire into focus, is divided between them. To check the setting of the wire in the central position it is convenient to apply a uniform bending moment, and then to observe if any change takes place in the reading. The position for no change in the reading can be found in a few seconds. In experiments where the bending moment is constant and the twisting moment is varied, no adjustment is practically required during the elastic life of the specimen ; and even when the bending moment is variable the adjustment is practically negligible, as the length of the specimen under test is only a few inches. The instrument is used for observations of the angular change due to bending by adjusting the wire in the horizontal plane passing through the axis of the specimen, and at a fixed' distance away from the central plane, as shown in Fig. 134. Thus if the wire is at a distance x from the central plane, and the specimen is subjected to a uniform bending moment, the reading will be (l+x) Q(lx} 0=2x6, and this is a measure of the angular change between the ends, since any further corrections are negligible for elastic strains. The instrument may therefore be used for measuring strains due to bending or twisting, and the single calibration required for both sets of readings is effected when the instrument is in position on the specimen. Further Researches. Although the laws for the failure of mild steel and very brittle materials (cast iron and hardened tool steel) are now established, it yet remains to be shown whether other materials fail by these laws. The effect of repeated stress upon ductile materials, and its bearing upon failure under combined stress, also requires investigation. Th;re are numerous original experiments for those having the requisite facilities. s 2 APPENDIX IV. HEAT TREATMENT OF STEELS. IT is impossible, in this book, to give a full account of the work done on heat treatment of various steels. However, Prof. A. McWilliam and Mr. E. J. Barnes, of the University of Sheffield, have recently published 1 a very complete study of the effect of heat treatment on Bessemer steels, and the following facts, abstracted from their paper, will serve as a guide for work, either upon similar material or some other types of steel. The tests were made on ordinary commercial English acid Bessemer steels of carbon content varying from O'lO per cent, to 0'86 per cent. The steels were received and treated in the form of forged or rolled bars 1 inch round, and either as received or after the treatment detailed were all tested in tensile, and as far as possible also under Dr. Arnold's alternating stress test, and were examined under the microscope. Treatments with Distinguishing Letters. Treatment. Letter. As received B Normalised. 950 0. for 20 minutes and cooled in air . BN . Annealed. Slowly heated up to 950 C. ; kept at 950 C. for about 35 hours ; very slowly cooled down in furnace BA - Quenched from 850 C. in water and tempered at 400 C. BY 500 C. BX 600 C. BZ 700 C. BW " ,, 900 C. ,, ,, 600 C. BJ" 700 C. BK - ,, 950 C. 700 C. BH Some typical curves of this series were given in Fig. 68, page 105. Methods of Experiment. All steels were in the form of 1-inch round bars, and were sawn into lengths of about 11 inches. Some of the harder steels showed a peculiar structure after being sawn off, the sawn surface exhibiting a curious pattern of intersecting elliptical elevations about inch broad. These gave no indications of their presence in the polished and etched micro -sections. 1 Iron and Steel Institute, May, 1909. APPENDIX IV. 261 Normalising. The pieces were placed in a Fletcher gas muffle at about 750 C., slowly raised in about half-an-hour to 950 0., and so maintained for twenty minutes. They were then taken out, reared on end on firebricks, and allowed to cool in the air. Annealing. Annealing was carried out in a coal-fired reverberatory furnace, according to the details described in the general table of treatments. Quenching. The bars were heated in a Brayshaw salt bath furnace, the temperature of which was simultaneously determined by a platinum resistance pyrometer with Whipple recorder and by a Paul Pt. Pt. 10 per cent. Ir. thermo-couple, the latter calibrated by means of sulphur (444 C.) and silver (962 0.). When the temperature was kept steady these two pyrometers were in remarkably close agreement, but, as was to be expected, on a falling temperature the readings of the former were somewhat higher, and on a rising temperature a similar amount lower than those of the latter. The pieces were put in the bath when it had attained a temperature of about 50 C. lower than the quenching tempera- ture desired; the heat was gradually raised to the temperature and maintained for fifteen minutes, when they were rapidly withdrawn and instantly quenched in pure Sheffield water at 15 to 20 C. The molten salts entirely prevented scaling, and thus the quenching was as efficient as the temperature used, compared with the size of the bars, would admit. As much that is misleading has been published on hardening, the subject is considered shortly here as it is involved in the reasons for choosing a hardening temperature of 850 C. for the majority of their quenchings. In ordinary hardening it is essential that the piece of steel should be heated to a temperature such that when quenched it will throw off its scale, or " shale " as the hardener calls it. Tf any of this scale remains firmly adhering, whether from the nature of the steel or the temperature used, then it acts as a blanket over the part of the steel it covers, prevents efficient quenching, and the part underneath this scale will be soft. A hardening temperature which results in the steel properly shaling is also in general a temperature that will give efficient quenching ; but when steels are heated, as in a salt bath, so that no scaling takes place, an efficient quenching temperature seems to be a function of the dimensions of the cross-section of the steel, for steels of the same composition and previous treatment. Even when bars were put into the salt bath with the original rolling-mill scale on them the molten salt seemed to remove it during the heating process. Experience with hardening steels of various sections led the authors to consider 800 to 850 C. a suitable range, and as some of their steels were very mild they decided on the higher limit, 850 C., for their preliminary series of tests. This was fixed only in consultation on hardening work done, and not from the study of the previous work of others which had been read as it came out. The previous work was re-read only immediately before writing the paper, so the coincidence, of the 262 APPENDIX IV. choice of 850 C. with the opinions of such a vigorous and reliable worker as Wahlberg after reviewing his own elaborate series, done from an entirely different standpoint, is of great interest. Tempering. With one exception the tempering was done by putting the bars in a lead bath at the required temperature, and maintaining the bath at a constant temperature for fifteen minutes, when the bars were removed and cooled in the air. In carrying out this treatment it is necessary to remember that samples of steel float in molten lead like wood in water, and that some efficient means must be adopted for pressing them down into the liquid lead. The tray of the Brayshaw furnace weighted with a billet on the top part of the frame outside the bath was used with success. The exception mentioned above arose as follows : It was desired to save the trouble of holding the pieces down in the bath whenever possible, and hence as the Brayshaw mixture of one molecular weight of potassium chloride (74'5) to one molecular weight of sodium chloride (58'5), or in the proportion of about 2| Ibs. to 2 Ibs., melts at about 650 C., it was thought that the tempering at 700 could be easily done in this bath. He who taketh short cuts seeketh trouble. The bath was steadied at the exact temperature, and the pieces put in and kept fifteen minutes at 700 C., but on endeavouring to get them out it was found that with the whole furnace cooling, the top had become pasty, and ultimately set before they could be removed ; so that it was necessary to put on a little more gas, and, after re-melting the bath, keep agitating the whole right up to the top and thus get the bars removed. The total operation took forty-five minutes, during which the pieces had been down to about 600 C. and then up again to 700 0. Unless the recently described potassium bichromate and potassium chloride mixture, which melts at 360 C., proves to be non-oxidising to steel immersed in it, the authors are likely to use a metallic bath only for future temperings. The tensile test-pieces were all turned to 0'564 diameter and 2 inches parallel, and the tensile piece was turned as near as possible to one end, so that after fracture the other end was long enough to turn to the standard alternating stress test-piece, namely f inch diameter by 6 inches long. The micro-sections were cut off the unstrained part, of the short end of the tensile test-piece. Thus the alternating stress tests and micro- sections were not off duplicate bars put through the same series of operations, but by the methods adopted represent tests actually off the same piece. The composition of one of the steels 1 with regard to carbon and manganese is given on page 263, the carbon being determined by com- bustion and checked. 1 The original paper contains records of several steels, but one type only is given here, APPENDIX IV. 263 TABLE XLV. SHOWING EFFECT OF HEAT TREATMENT ON STEEL OF CARBON 0*10 PER CENT. AND MANGANESE 0*56 PER CENT. Treatment. Mark. Yield Point. Tons per sq. in. Maximum Stress. Tons per sq. in. Elonga- tion. Per cent, on 2 in. Reduction of Area. Per cent. Dr. Arnold's Alternating Stress Test. As received 10 B 19-1 25-9 37-1 63-4 Normalised 10 BN 18-5 24-8 37-4 59-8 Annealed . 10 BA 9-6 21-0 43-8 72-0 352 850 C. water and 400 C. air 10 BY 17-9 27 4 39-0 72-9 850 C. water and 500 C. air 10 BX 17-6 26-0 38-0 74-7 336 850 C. water and 600 C. air 10 BZ 19-4 26-0 38-5 71-6 326 850 C. water and 700 C. air 10 BW 16-7 26-1 40-0 74-5 331 950 C. water and 700 C. air 10 BH 21-1 28-4 35-0 67-5 239 TABLE XL VI. THE EFFECT OF HEAT TREATMENT ON STEEL OF CARBON 0'29 PER CENT. AND MANGANESE 0'92 PER CENT. Mark. Yield Point. Tons per sq. in. Maximum Stress. Tons per sq. in. Elongation per cent, on 2 in. Reduction of Area per cent. Dr. Arnold's Alternating Stress Test. 30 B 26-6 40-9 25-0 46-8 322 30 BN 25-7 40-8 26-3 63*5 329 30 BA 21-5 37-1 26-5 4S-C 296 30 BW 37*0 45-7 25'j o7'8 202 30 BJ 50-9 57'5 17-3 48-6 173 30 BK 46-2 54-3 19-5 50-8 184 The table containing the 30 B results shows the very marked influence of the higher quenching temperatures, namely, 900 for 30 BJ and 30 BK, even after these have been tempered at 600 and 700 respectively ; for example, the high reduction in area, 51 per cent., for a 54-ton steel with yield point 85 per cent, of the maximum stress. 264 APPENDIX IV. As showing the effect of heat treatment on other than Bessemer steels the following table from a paper on "High Tension Steels," l by Mr. Percy Longmuir, B.Met., read before the Iron and Steel Institute, is instructive : TABLE XLVII. 5 PER CENT. XICKEL STEEL. No. Treatment. Elastic Limit. Tons per sq. in. Maximum Stress. Tons per sq. in. Elongation per cent, on 2 in. Reduction of Area per cent. 23 As received .... 47-6 5696 13-5 20-0 24 Air cooled from 800 C. 34-8 49-08 18-0 38-0 2o Heated to 1 ,000 0., quenched in oil at 1,000 C., tem- j pered at 490 . 04-8 76-00 12-5 43-6 20 Heated to 1, 000 C., quenched in oil at 1,000 C., tem- pered at 490 . 68-4 70-92 10-0 37-6 27 Heated to 1, 000 C., quenched in oil at 900 C., tempered at 490 .... 66'8 70-08 12 'a 4,3-2 28 Heated to 1, 000 C., quenched Not in oil at 9o() C., tempered de- at 490 .... tected 82-56 1-.3 0-4 29 Heated to 800 C., quenched in oil at 800 C., tempered at 490 . 05-2 78-40 3-0 2-5 The above results will probably inspire the reader to conduct similar tests under various conditions. The authors of the above paper do not give details concerning the method of gripping the specimens, etc., and it is therefore not possible to estimate within what measure of accuracy some of the jigures come. That they form a valuable guide and refer- ence cannot be doubted. The enormous amount of work entailed in such a research can only be appreciate 1 by those who have attempted similar experiments. Captain H. B. Sankey and Mr. J. Kent-Smith published a paper' 2 entitled " Heat Treatment Experiments with Chrome-Vanadium Steel." The reader who wishes to do original work on this subject would do well to consult the journals of the scientific institutions or the pages of "Science Abstracts" for the last ten years. At the same time there is no need for the student who seeks to do new laboratory experiments to hesitate because someone else has done the same thing. 1 Proceedings, Iron and Steel Institute, May, 1909. 2 Transactions of the Institution of Mechanical Engineers, 1904. BIBLIOGRAPHY. APPLIED MECHANICS TEXT BOOKS CONTAINING GENEEAL EEFEEENCES TO THE STRENGTH OF MATERIALS AND TESTING. Applied Mechanics. EANKINE. (Griffin, London.) Applied Mechanics. COTTERILL. (Macmillan, London.) Mechanics of Engineering. CHURCH. (Wiley, New York.) Mechanics Applied to Engineering. GOODMAN. (Longmans, London.) Applied Mechanics. LANZA. (Wiley, New York.) Applied Mechanics. ALEXANDER and THOMPSON. (Macmillan, London.) Applied Mechanics. PERRY. (Cassell, London.) Applied Mechanics and Mechanical Engineering (five vols.). JAMIESON. (Griffin, London.) Mechanical Engineering. CARPENTER. (Wiley, New York.) Experimental Mechanics. BALL. (Macmillan, London.) Applied Mechanics. D. A. Low. (Longmans, Green & Co., London.) Experimental Engineering. E. C. CARPENTER. (Chapman & Hall. 1905.) Theory of Structures and Strength of Materials. II. T. BOVEY. (John Wiley & Sons, New York.) Chain Cables and Chains. THOS. W. TRAILL, M.Inst.C.E. (Crosby, Lockwood & Co.) Strength and Properties of Materials. (W. G. Kirkcaldy, London.) Mechanical Engineering Materials. E. C. E. MARKS. (Technical Publishing Co.) Strength of Materials and Structures. SLR J. ANDERSON. (Longmans, Green & Co., 1892.) Design of Structures. S. ANGLIN. (Chas. Griffin & Co.) Bridge Construction. PROF. FIDLER, M.E.C.E. (Chas. Griffin & Co.) Testing Materials (two vols.). MARTENS and HENNING. (Wiley, New York.) Mechanics of Materials. MERRIMAN. (Wiley, New York.) Theory of Structures and Strength of Materials. BOVEY. (Wiley, New York.) Testing of Materials. UNWIN. (Longmans, London.) The Strength of Materials. EWING. (Cambridge University Press.) Strength of Materials. MORLEY. (Longmans, London.) Elasticity and Eesistance of Materials. BURR. (Wiley, New York.) 266 BIBLIOGEAPHY Materials of Construction. JOHNSON. (Wiley, New York.) Materials of Engineering (three vols.) THURSTON. (Wiley, New York.) A History of the Theory of Elasticity and of the Strength of Materials. TODHUNTER and PEARSON. (Cambridge University Press.) Strength and Elasticity of Structural Members. WOODS. (Arnold, London.) Testing and Strength of Materials. POPPLEWELL. (Scientific Publishing Co., Manchester.) MISCELLANEOUS PAPERS, ETC., ON TESTING, AND SUB- JECTS CONNECTED WITH TESTING. Naval Accidents (Microstructure, &c.). THOS. ANDREWS, F.B.S. (Engineering, Dec. 2nd, 1904, pp. 737 et. seq.} The Adoption of Standard Forms of Test-Pieces for Bars and Plates, by WILLIAM HACKNEY, B.Sc., Assoc.M.Inst.C.E. (Proc. Inst. C. E., vol. Ixxvi., part ii.) The Shape of Compression Test-Pieces. A. MARTENS. (Ibid., vol. cxxvii., p. 400.) Forms of Tensile Test-Pieces. S. L. H. APPLEBY. (Ibid., vol. cxviii., p. 395.) The Practical Strength of Beams. (Proc. Inst. C. E., vol. Ixii., p. 251.) Tensile Tests of Iron and Steel Bars. PETER D. BENNETT. (Prcc. Inst. M. E., 1886.) Testing Some Specimens of Malleable Cast Iron. A. G. ASHCROFT. (Proc. Inst, C. E., vol. cxvii., 1894.) The Resistance of Materials under Impact. MANSFIELD MERRIMAN. (Ibid., vol. cxxii., 1895.) A New Indentation Test for Determining the Hardness of Metals, by W. C. UNWIN, B.Sc., F.R.S. (Ibid., vol. cxxix., 1897.) Testing the Strength of Materials. A. H. JAMESON, M.Sc. (Ibid., vol. cxxxiv., 1898.) Notes on the Endurance of Steel Bars Subject to Repetitions of Torsional Tests, by E. G. COKER, B.A., B.Sc. (Ibid., vol. cxxxv., 1899.) The Relation of the Constants of the Elongation Equation to Contraction of Area. PROF. ELLIOTT. (Ibid., vol. clviii., 1904.) The Fatigue of Metals. A. N. KEMP. (Engineering Review, 1904.) A Throw -testing Machine for Reversals of Mean Stress, by Profs. OSBORNE REYNOLDS, F.R.S. , and J. H. SMITH, M.Sc., etc. (Phil. Trans, of the Roy. Soc. of London, vol. 199, 1902.) Heat Treatment Experiments with Chrome-Vanadium Steel, by CAPT. H. R. SANKEY and Mr. J. K. SMITH. Wohler's Experiments and Results. (" Zeitschrift fur Bauwesen,"< Berlin, 1870 ; and Engineering, vol. ix., 1871.) BIBLIOGEAPHY 267 Fatigue of Materials. A. N. KEMP. (Engineering Revieiv, Sept., 1904.) Launhardt's Experiments. (" Zeitschrift des Architecten und Ingenieur- Yereins," Hanover, 1873.) Further Wohler Experiments. SPANGENBERG. (" Zeitschrift fiir Bauwesen," 1874.) Experiments on Iron and Steel (Alternating Stresses). SIR BENJAMIN BAKER, M.Inst.O.E. (American Soc. Mech. Engineers, 1886.) Variation of the Elastic Limits of Materials. BAUSCHINGER. (Mitt- heilungen aus dem Mech. Techn. Laboratorium in Munchen, 1886.) Fracture of Metals under repeated Alternations of Stresses. PROF. EWING and J. C. W. HUMPHREY. (Eoy. Soc., 1903.) Der Einfluss von Ungleichmassigkeiten in Querschnitte des prismatischen Theiles eines Probestabes. (" Zeitsch. Yer. Deutsch. Ing.," vol. xlvii., p. 426, 1903.) Zustandsanderungen derMetalle infolge vonFestigkeitsbeanspruchungen, by A. MARTENS. (Preuss. Akad. Wiss. Berlin, Sitz, Ber. ii., 1910.) Wicksteed Machine. See Proc. Inst. Mech. Eng., 1907 (Electrical control), and Proc. Inst. Mech. Eng., 1882 and 1891 (older type). Extensometers. (Eeport of British Assoc., 1896.) Measurement of Strains. MORROW. (Proceedings Inst. Mech. Engineers, 1904.) Tensile Strength of Open Hearth Steel. H. H. CAMPBELL. (" Science Abstracts," Jan. 1905.) Widmonstatten Figures in Steel Castings (Microstructure, &c.) J. O. ARNOLD and A. McWiLLiAM. (Nature, Nov. 10th, 1904.) Cement Exhibits at the St. Louis Exhibition. (Engineering News, Nov. 10th, 1904.) Strength and Testing of Timber. T. HUDSON BEARE. (Engineering, Dec. 9th, 1904.) Yanadium Steels. L. GUILLET. ("Eevue de Metallurgie," Oct., 1904.) Tests on Concrete Steel Beams. J. J. KAHN. (Engineering Record, Oct. 8th, 1904.) Punching as a Method of Testing. L. BACLE. (Soc. d'Encouragement, Bull. vol. cvi., Nov. 1904.) Special Industrial Steels. H. LE CHATALIER. (Soc. d'Encouragement ; Revue de Metallurgie, Nov. 1904.) Impact Tests of Steel. A. E. SEATON and A. JUDE. (Engineering, Nov. 25th, 1905.) Materials which Eetard the Setting of Portland Cement. E. C. CAR- PENTER. (Engineering Record, Dec. 31st, 1904; Engineering News, Jan. 5th, 1905.) Eussian Standards for Portland Cement. (Eiga-Ind. Zeit, vol. xxx., p. 189, 1904.) Toughness of Steel after Working at a Blue Heat. C. FREMONT. (" Comptes Eendus," Dec. 12th, 1904.) 268 BIBLIOGRAPHY Properties of Hard- Silicon Steel (Soc. d' Encouragement, vol. w\.,Rei-uede Metallurgie, Dec. 1904.) Drop Testing Machine. W. T. M. Goss. (Pages Weekly, Jan. 13th, 1905.) Impact Tests of Steel. (Proc. Inst. Mech. Eng., pp. 1135-1168, July, 1904.) Heat Treatment of Chrome -Vanadium Steel. (Ibid,., pp. 1235-1282, July, 1904.) Hard-drawn Copper Wire. T. B. DOOLITTLE. (Harvard Engineering Journal, pp. 133-134, Nov. 1905.) Hard-drawn Copper Wire. T. BOLTOX. (Electrical Review, vol. lx., pp. 131-133, Jan. 25th, 1907.) Specification of Hard-drawn Copper Wire. (Electric Railway Journal, vol. xxxiv., pp. 181-183, July 31st, 1909.) A Heat Treatment Study of Bessemer Steels. ANDREW McWiLLiAM and ERNEST J. BARNES. (Iron and Steel Institute, May 20th, 1909.) The Elastic Breakdown of Materials, subjected to Compound Stress. L. B. TURNER, B.A. (Engineering, Feb. 5th, 1909.) Notes on Tests for Hardness. PROF. T. TURNER, M.Sc. (Iron and Steel Institute, May 14th, 1909.) Ductile Materials under Stress. C. A. M. SMITH, M.Sc. (Junior In- stitution of Engineers, May, 1909.) Some Experiments on Impact. J. E. SEARS, Jun. (Engineering, April 30th, 1909 ; May 7th, 1909.) The Collapse of Tubes under External Stress. S. E. SLOCUM, (Engineer- ing, Jan. 8th, 1909.) Guest's Law of Combined Stress. C. A. M. SMITH, M.Sc. (Engineer- ing, April 23rd, 1909.) Comparison of Tensile Impact-Tensile and Repeated-Bending Methods of Testing Steel. (Inst. Mech. Eng., May 27th, 1910 ; Engineering, June 3rd, 1910.) Research on the Hardening of Carbon and Low Tungsten Tool Steels. SHIPLY N. BRAYSHAW. (Inst. Mech. Eng., April 15th, 1910; Engineering, April 22nd, 1910.) Mechanical Tests of Insulator Porcelain and Glass. R. P. CLARKSON. (Electrical World, vol. Ivi., pp. 25-27 ; July 7th, 1910.) Anomalous Effects on First Loading a Wire, and some Effects of Bending Overstrain in Soft Iron Wires. A. I. STEVENS, M.A., B.Sc. (Phil. Mag., April, 1910.) The Breakdown of Tubes under Combined Stress. J. J. GUEST. (Phil. Mag., 1900.) Experiments on Combined Stress. PROF. HANCOCK. (PA/7. Mag., 1906.) Ductile Material under Combined Stress. W. A. SCOBLE, B.Sc. (Phil. Mag., 1906.) Brittle Material under Combined Stress. W. A. SCOBLE, B.Sc. Mag., 1906.) BIBLIOGEAPHY 269 Strength of Pipes and Cylinders. C. A. M. SMITH. (Engineering, March, 1909.) Elastic Breakdown of Non-Ferrous Metals. (Proc. Inst. Metals; and Engineering, 1909.) A Method of Detecting the Bending of Columns. C. A. M. SMITH, M.Sc. (Inst. Mech. Eng., 1908.) Mild Steel Tubes in Compression and under Combined Stress. W. MASON, M.Sc. (Proc. Inst. Mech. Eng., 1910.) Elastic Breakdown of Certain Steels. C. A. M. SMITH, M.Sc. (Journal Iron and Steel Institute, 1910.) Compound Stress Experiments. (Proc. Inst. Mech. Eng., 1910.) Beports of Alloys Research Committees. (Proc. Inst. Mech. Eng., 1891 onwards.) (Post- graduates about to commence a research should carefully look through Journals mentioned above before commencing experiments.] The Elastic Limit of Manganese and other Bronzes. J. A. CAPP. (Amer. Soc. Mech. Eng., 1910.) Some Experiments on Solid Steel Bars under Combined Stress. C. A. M. SMITH, M.Sc. (Engineering, Aug., 1909.) The Stresses in a Thick Hollow Cylinder subjected to Internal Pressure. L. B. TURNER, B.A. (Trans. Cambridge Phil. Soc., Sept. 1910.) Further Tests of Brittle Materials under Combined Stress. W. A. SCOBLE, B.Sc. (Phil. Mag., June, 1910.) The Design of Struts. W. E. LILLY, M.A., D.Sc. (Engineering, Jan. 1908.) The Behaviour of Ductile Material under Torsional Strain with Restora- tion of Elasticity at Low Temperatures. C. E. LARARD, A.M.I.C.E. (Proc. Inst. C.E., vol. 179, p. 3, 1910.) An Experimental Investigation into the Flow of Rocks. F. D. ADAMS and E. G. COKER. (Amer. Journ. Sci., June, 1910.) The Elastic Properties of Platinum -Iridium Wires. K. E. GUTHE and L. P. LIEG. (Physical Review, May, 1910.) Tesfs of Concrete Columns made under Building Conditions. H. C. BERRY. (Engineering Record, Feb., 1910.) Untersuchung eines im Betriebe geplatzten Siedevohrs. E. HEYN and 0. BAUER. (Konigl. Materialpriifungsamt.) Berechnung Zylindrischev Druckfedeon auf Sicherheit gegen seitliches Ausknicken. E. HURLBRINK. (Zeitschr. Vereines. Deutsch., Jan., 1910.) USEFUL CONSTANTS. 1 Inch=2iv40 millimetres. 1 Gallon = -1604 cubic foot=10 Ib. of water at 62 F. Weight of 1 Ib. in London=445,000 dynes. One pound avoirdupois =7 000 grains=453'6 grammes. One cubic foot of water weighs 62 -3 Ib. A column of water 2'3 feet high corresponds to a pressure of 1 Ib. per in. Absolute temp., t=0 C.+273 or 0F.-f460. One radian =57 '30 degrees. To convert common into Napierian logarithms, multiply by 2-3026. The base of the Napierian logarithms is e=2'7183. The value of g at London =32 '182 feet per sec. per sec. USEFUL CONSTANTS 271 LOGARITHMS. 1 2 3 4 5 6 7 8 9 10 0000 0043 0086 0128 0170 0212 0253 0294 0334 0374 4 9 13 17 4 8 12 16 20 26 30 34 38 24 28 82 37 11 0414 0453 0492 0531 0569 0607 0645 0682 0719 0755 4 8 12 15 4 7 11 15 19 19 23 27 31 35 22 26 30 33 12 0792 0828 0864 0899 0934 0969 1004 1038 1072 1106 3 7 11 14 3 7 10 14 18 17 21 25 28 32 20 24 27 31 13 1139 1173 1206 1239 1271 1303 1335 1367 1399 1430 3 7 10 13 3 7 10 12 16 16 20 23 26 30 19 22 25 29 14 1461 1492 1523 1553 1584 1614 1644 1673 1703 1732 3 6 9 12 3 6 9 12 15 15 18 21 24 28 17 20 23 26 15 17-61 1700 1818 1847 1875 1903 1931 1959 1987 2014 3 6 9 11 3 5 8 11 14 14 17 20 23 26 16392225 16 2041 2068 2095 2122 2148 2175 2201 2227 2253 2279 3 5 8 11 3 5 8 10 14 13 16 19 22 24 15 18 21 23 17 2304 2330 2355 2380 2405 2430 2455 2480 2504 2529 3 5 8 10 2 5 7 10 13 12 15 18 20 23 15 17 19 22 18 2553 2577 2601 2625 2648 2672 2695 2718 2742 2765 2579 2579 12 11 14 16 39 21 14161821 19 2788 2810 2833 2856 2878 2900 2923 2945 2967 2989 2479 2468 11 11 13 16 18 20 13 15 17 19 20 3010 3032 3054 3075 3096 3118 3139 3160 3181 3201 2468 11 13 15 17 19 21 22 23 24 3222 3424 3(517 3802 3243 3444 3636 3820 3263 3464 3655 3838 3284 3483 3674 3856 3304 3502 3C92 3874 3324 3522 3711 3892 3345 3541 3729 3909 3365 3560 3747 8927 3385 3579 3766 3945 3404 3598 3784 3962 2468 2468 2467 2457 10 10 9 9 121416 18 12 14 15 17 11 131517 11 12 14 16 25 3979 3997 4014 4031 4048 4065 4082 4099 4116 4133 2357 9 10 12 14 15 26 27 28 29 4150 4314 4472 4624 4166 4330 4487 4639 4183 4346 4502 4654 4200 4362 4518 4669 4216 4378 4533 4683 4232 4393 4548 4698 4249 4409 4564 4713 4265 4425 4579 4728 4281 4440 4594 4742 4298 4456 4609 4757 2357 2356 2356 1346 8 8 8 7 10 11 13 15 9 11 13 14 9 11 12 14 9 10 12 13 30 4771 4786 4800 4814 4829 4843 4857 4871 4886 4900 1346 7 9 10 11 13 31 32 33 34 4914 5051 5185 5315 4928 5065 5198 53-28 4942 5079 5211 5340 4955 5092 5224 5353 4969 5105 5237 5366 4983 5119 5250 5378 4997 5132 5263 5391 5011 5145 5276 5403 5024 5159 5289 5416 5038 5172 5302 5428 1346 1345 1345 1345 7 7 6 6 8 10 11 12 8 9 11 12 8 9 10 12 8 91011 35 5441 5453 5465 5478 5490 5502 5514 5527 5539 5551 1245 6 7 9 10 11 36 37 38 39 5563 5682 5798 5911 5575 5694 5809 5922 5587 5705 5821 5933 5599 5717 5832 5944 5611 5729 5843 5955 5623 5740 5855 5966 635 5752 5866 5977 5647 5763 5877 5988 5658 5775 5888 5999 5670 57S6 5899 6010 1245 1235 1235 1234 6 8 6 5 7 81011 7 8 910 7 8 9 10 7 8 910 40 6021 6031 6042 6053 6064 6075 6085 6096 6107 6117 1234 5 6 8 910 41 42 43 44 6128 6232 6335 6435 6138 6243 6345 6444 6149 6253 6355 6454 6160 6263 6365 6464 6170 6274 6375 6474 6180 6284 6385 6484 6191 6294 6395 6493 6201 6304 6405 6503 6212 63*14 6415 6513 6222 6325 6425 6522 1234 1234 1234 1234 5 5 5 5 6789 6789 6789 6789 45 6532 6542 6551 6561 6571 6580 6590 6599 6609 6618 1234 5 6789 46 47 48 49 6628 6721 6812 6902 6637 6730 6821 6911 6646 6739 6830 6920 6656 6749 6839 C92S 6665 6758 6848 6937 6675 6767 6857 6946 6684 6776 6866 6955 6693 6785 6875 6964 6702 6794 6884 6972 6712 6803 6893 6981 1234 1234 1234 1234 5 5 4 4 6778 5678 5678 5678 50 6990 0998 7007 7016 7024 7033 7042 7050 7059 7067 1233 4 5678 272 USEFUL CONSTANTS LOGARITHMS. 1 2 3 4 5 6 7 8 9 1234 5 6789 51 52 53 54 7076 7160 7243 7324 7084 7168 7251 7332 7093 7177 7259 7340 7101 7185 7267 7348 7110 7193 7275 7356 7118 7202 7284 7364 7126 7210 7292 7372 7135 7218 7300 7380 7143 7226 7308 7388 7152 73K 7396 1233 1223 1223 1223 4 4 4 4 5678 5677 5667 5667 55 7404 7412 7419 7427 7435 7443 7451 7459 7466 7474 1223 4 5567 56 57 58 59 7482 7559 7634 7709 7490 7566 7642 7716 7497 7574 7649 7723 7505 7582 7657 7731 7513 7589 7664 7738 7520 7597 7672 7745 7528 7604 7679 7752 7536 7612 7686 7760 7543 7619 7694 7767 7551 7627 7701 7774 1223 1223 1123 1123 4 4 4 4 5567 5567 4567 4567 60 7782 7789 7796 7803 7810 7818 7825 7832 7831) 7846 1123 4 456 6 61 62 63 64 7853 7924 7993 8062 7860 7931 8000 8069 7868 7938 8007 8075 7875 7945 8014 8082 7882 7932 8021 8089 7889 7959 8028 8096 7896 7966 8035 8102 7903 7973 8041 8109 7910 7980 8048 8116 7917 7987 8055 8122 1123 1123 1123 1123 4 3 3 3 4 5 C) 6 4566 4556 4556 65 8129 8136 8142 8149 8156 8162 8169 8176 8182 8189 1123 3 4556 66 67 68 69 8195 8261 8325 8388 8202 8267 8331 8395 8209 8274 8338 8401 8215 8280 8344 8407 8222 8287 8351 8414 8228 8293 8357 8420 8235 8299 8363 8426 8241 8306 8370 8432 8248 8312 8376 8439 8254 8319 8382 8445 1123 1123 1123 1122 3 3 3 3 4556 4550 4456 4456 70 8451 8457 8463 8470 8476 8482 8488 8494 8500 8506 1122 3 4 4 5 6 71 72 73 74 8513 8573 8633 8692 8519 8579 8639 8698 8525 8585 8645 8704 8531 8591 8651 8710 8537 8597 8657 8716 8543 8603 8633 8722 8549 8609 8669 8727 8555 8615 8675 8733 8561 8621 8681 8739 8567 8627 8686 8745 1122 1122 1122 1122 3 a 3 3 4455 4455 4 4 5 5 4 4 r. r, 75 8751 8756 8762 8768 8774 8779 8785 8791 8797 8802 1122 3 3 4 5 5 76 77 78 79 85:08 8865 8921 8976 8814 8871 8927 8982 8820 Sb76 8932 8987 8825 8882 8938 8993 8831 8887 8943 8998 8837 8893 8949 9004 8842 8899 8954 9009 8848 8904 8960 9015 8854 8910 8965 9020 8859 8915 8971 9025 1122 1122 1122 1122 3 3 3 3 3455 3445 3445 344:. 80 9031 9036 9042 9047 9053 9058 9063 9069 9074 9079 1122 3 3445 81 82 83 84 9085 9138 9191 9243 9090 9143 9196 9248 9096 9149 9201 9253 9101 9154 9206 9258 9106 9159 9212 9263 9112 9165 9217 9269 9117 9170 9222 9274 9122 9175 9227 9279 9128 9180 9232 9284 9133 9186 9238 9289 1122 1122 1122 1122 3 3 3 3 3445 3445 3445 3445 85 9294 9299 9304 9309 9315 9320 9325 9330 9335 9340 1122 3 3445 86 87 88 89 9345 9395 9445 9494 9350 9400 9450 9499 9355 9405 9455 9504 9360 9410 9460 9509 9365 9415 9465 9513 9370 9420 9469 9518 9375 9425 9474 9523 9380 9430 9479 9528 9385 9435 9484 9533 9390 9440 9489 9538 1122 0112 0112 0112 3 2 2 2 3 4 4 5 3344 3344 3344 90 9542 9547 9552 9557 9562 9566 9571 9576 9581 9586 0112 2 3344 91 92 93 94 9590 9638 9685 9731 9595 9643 9689 9736 9600 9647 9694 9741 9605 9652 9699 9745 9609 9657 9703 9750 9614 9661 9708 9754 9619 9666 9713 9759 9624 9671 9717 9763 9628 9675 9722 9768 9633 9680 9727 9773 0112 0112 0112 0112 2 2 2 3344 3344 3344 3344 95 9777 9782 97-6 9791 9795 9800 9805 9809 9814 9818 0112 2 3344 96 97 98 99 9823 9868 9912 9956 9827 9872 9917 9961 9832 9877 9^21 9965 9836 9881 9926 9969 9841 9886 9930 9974 9845 9890 9934 9978 9850 0894 9939 9983 9854 9899 9943 9987 9859 9903 9948 9991 9863 9908 9952 9996 0112 0112 0112 0112 2 2 o 3344 3344 3 3 4 4 3334 USEFUL CONSTANTS ANTILOGARITHMS. 273 g 00 1000 1002 1005 1007 1009 1012 1014 1016 1019 1021 0011 1 1222 01 02 03 04 1023 1047 1072 1096 1026 1050 1074 1099 1028 1052 1076 1102 1030 1054 1079 1104 1033 1057 1081 1107 1035 1059 1084 1109 1038 1062 1086 1112 1040 1064 1089 1114 1042 1067 1091 1117 1045 1069 1094 1119 0011 0011 0011 0111 1 1 1 1 1222 1222 1222 2222 03 1122 1125 1127 1130 1132 1135 1138 1140 1143 1146 0111 1 2222 06 07 08 09 1148 1175 1202 1230 1151 1178 1205 1233 1153 1180 1208 1236 1156 1183 1211 1239 1159 1186 1213 1242 1161 1189 1216 1245 1164 1191 1219 1247 1167 1194 1222 1250 1169 1197 1225 1253 1172 1199 1227 1256 0111 0111 0111 0111 1 1 1 1 2222 2222 2223 2223 10 1259 1262 1265 1268 1271 1274 1276 1279 1282 1285 0111 1 2223 11 12 13 14 1288 1318 1349 1380 1291 1321 1352 1384 1294 1324 1355 1387 1297 1327 1358 1390 1300 1330 1361 1393 1303 1334 1365 1396 1306 1337 1368 1400 1309 1340 1371 1403 1312 1343 1374 1406 1315 1346 1377 1409 0111 0111 0111 0111 2 2 2 2 2223 2223 2233 2233 15 1413 1416 1419 1422 1426 1429 1432 1435 1439 1442 0111 2 2233 16 17 18 19 1445 1479 1514 1549 1449 1483 1517 1552 1452 1486 1521 1556 1455 1489 1524 1560 1459 1493 1528 1563 1462 1496 1531 1567 1466 1500 1535 1570 1469 1503 1538 1574 1472 1507 1542 1578 1476 1510 1545 1581 0111 0111 0111 0111 2 2 2 2 2233 2233 2233 2333 20 1585 1589 1592 1596 1600 1603 1607 1611 1614 1618 0111 2 2333 21 22 23 24 1622 1660 1698 1738 1626 1663 1702 1742 1629 1667 1706 1746 1633 1671 1710 1750 1637 1675 1714 1754 1641 1679 1718 1758 1644 1683 1722 1762 1648 1687 1726 1766 1652 1690 1730 1770 1656 1694 1734 1774 0112 0112 0112 0112 2 2 2 2 2333 2333 2334 2334 23 1778 1782 1786 1791 1795 1799 1803 1807 1811 1816 0112 2 2334 26 27 28 29 1820 1862 1905 1950 1824 1866 1910 1954 1828 1871 1914 1959 1832 1875 1919 1963 1837 1879 1923 1968 1841 1884 1928 1972 1845 1888 1932 1977 1849 1892 1936 1982 1854 1897 1941 1986 1858 1901 1945 1991 0112 0112 0112 0112 2 2 2 2 3334 3334 3344 3344 30 1995 2000 2004 2009 2014 2018 2023 2028 2032 2037 0112 2 3344 31 32 33 34 2042 2089 2138 2188 2046 2094 2143 2193 2051 2099 2148 2198 2056 2104 2153 2203 2061 2109 2158 2208 2065 2113 2163 2213 2070 2118 2168 2218 2075 2123 2173 2223 2080 2128 2178 2228 2084 2133 2183 2234 0112 0112 0112 1122 2 3344 3344 3344 3445 35 2239 2244 2249 2254 2259 2265 2270 2275 2280 2286 1122 3 3445 36 37 38 39 2291 2344 2399 2455 2296 2350 2404 2460 2301 2355 2410 2466 2307 2360 2415 2472 2312 2366 2421 2477 2317 2371 2427 2483 2323 2377 2432 2489 2328 2382 2438 2495 2333 2388 2443 2500 2339 2393 2449 2506 1122 1122 1122 1122 3 3 3 3 3445 3445 3445 3455 40 2512 2518 2523 2529 2535 2541 2547 2553 2559 2564 1122 3 4455 41 42 43 44 2570 2630 2692 2754 2576 2636 2698 2761 2582 2642 2704 2767 2588 2649 2710 2770 2594 2655 2716 2780 2600 2661 2723 2786 2606 2667 2729 2793 2612 2673 2735 2799 2618 2679 2742 2805 2624 2885 2748 2812 1122 1122 1123 1123 3 3 3 3 4455 4456 4456 4456 43 2818 2825 2831 2838 2844 2851 2858 2864 2871 2877 1123 3 4556 46 47 48 49 2884 2951 3020 3090 2891 2958 3027 3097 2897 2965 3034 3105 2904 2972 3041 3112 2911 2979 3048 3119 2917 2985 3055 3126 2924 2992 3062 3133 2931 2999 3069 3141 2938 3006 3076 3148 2944 3013 3083 3155 1123 1123 1123 1123 3 3 4 4 4556 4556 4566 4566 T.M. 274 USEFUL CONSTANTS ANTILOGARITHMS. 6789 50 3162 3170 3177 3184 3192 3199 3206 3214 3221 3228 1123 4 4 f> 6 7 51 52 53 51 3226 3311 3388 3467 3243 3319 3396 3475 3251 3327 3404 3483 8268 3334 3412 3491 3266 3342 3420 3499 3273 3350 3428 3508 3281 3357 3436 3516 3289 3365 3443 3524 3296 3373 3451 3532 6304. 3381 3459 3540 1223 1223 1223 1223 4 4 4 4 5567 5567 5667 5 6 6 7 55 3548 3556 3565 3573 3581 3589 3597 3606 3614 3622 1223 4 5677 56 57 58 59 3631 3715 3802 3890 3639 3724 3811 3899 3648 3733 3819 3908 3656 3741 3828 3917 3664 3750 3837 3926 3673 3758 3846 3936 3681 3767 3855 3945 3690 3776 3864 3954 3698 3784 3873 3963 3707 3793 3882 3972 1233 1233 1234 1234 4 4 4 5 5678 5678 5678 5678 60 3981 3990 3999 4009 4018 4027 4036 4046 4055 4064 1234 5 6678 61 62 63 64 4074 4169 4266 4365 4083 4178 4276 4375 4093 4188 4285 4385 4102 4198 4295 4395 4111 4207 4305 4406 4121 4217 4315 4416 4130 4227 4325 4426 4140 4236 4335 4436 4150 4246 4345 4446 4159 4256 4355 4457 1234 1234 1234 1234 5 5 5 5 6789 6789 6789 6789 65 4467 4477 4487 4498 4508 4519 4529 4539 4550 4560 1234 5 6789 66 67 68 69 4571 4677 4786 4898 4581 4688 4797 4909 4592 4699 4808 4920 4603 4710 4819 4932 4613 4721 4831 4943 4624 4732 4842 4955 4634 4742 4853 4966 4645 4753 4864 4977 4656 4764 4875 4989 4667 4775 4887 5000 1234 1234 1234 1235 5 5 6 6 6 7 9 10 7 8 9 10 7 8 9 10 7 8 9 10 70 5012 5023 5035 5047 5058 5070 5082 5093 5105 5117 1245 6 7 8 9 11 71 72 73 74 5129 5248 5370 5495 5140 5260 5383 5508 5152 5272 5395 5521 5164 5284 5408 5534 5176 5297 5420 5546 5188 5309 5433 5559 5200 5321 5445 5572 5212 5333 5458 5585 5224 5346 5470 5598 5236 5358 5483 5610 1245 1245 1345 1345 6 6 6 6 7 8 10 11 7 9 10 11 8 9 10 11 8 9 10 12 75 5623 5636 5649 5662 5675 5689 5702 5715 5728 5741 1345 7 8 9 10 12 76 77 78 79 5754 5888 6026 6166 5768 5902 6039 6180 5781 5916 6053 6194 5794 5929 6067 6209 5808 5943 6081 6223 5821 5957 6095 6237 5834 5970 6109 6252 5848 5984 6124 6266 5861 5998 6138 6281 5875 6012 6152 6295 1345 1345 1346 1346 7 7 7 7 8 9 11 12 8 10 11 12 8 10 11 13 9 10 11 13 80 6310 6324 6339 6353 6368 6383 6397 6412 6427 6442 1346 7 9 10 12 13 81 82 83 84 6457 6607 6761 6918 6471 6622 6776 6934 6486 6637 6792 6950 6501 6653 6808 6966 6516 6668 6823 6982 6531 6683 6839 6998 6546 6699 6855 7015 6561 6714 6871 7031 6577 6730 6887 7047 6592 6745 6902 7063 2356 2356 2356 2356 8 8 8 8 9 11 12 14 9 11 12 14 9 11 13 14 10 11 13 15 85 7079 7096 7112 7129 7145 7161 7178 7194 7211 7228 2357 8 10 12 13 15 86 87 88 89 7244 7413 7586 7762 7261 7430 7603 7780 7278 7447 7621 7798 7295 7464 7638 7816 7311 7482 7656 7S34 7328 7499 7674 7852 7345 7516 7691 7870 7362 7534 7709 7889 7379 7551 7727 7907 7396 7568 7745 7925 2357 2357 2457 2457 8 9 9 9 10 12 13 15 10 12 14 16 11 12 14 16 11 13 14 16 90 7943 7962 7980 7998 8017 8035 8054 8072 8091 8110 2467 9 11 13 15 17 91 92 93 94 8128 8318 8511 8710 8147 8337 8531 8730 8166 8356 8551 8750 8185 8375 8570 S770 8204 8395 8590 8790 8222 8414 8610 8810 8241 8433 8630 8831 8260 8453 8650 8851 8279 8472 8670 8872 8299 8492 8690 8892 2468 2468 2468 2468 9 10 10 10 11 13 15 17 12 14 15 17 12 14 16 18 12 14 16 18 95 8913 8933 8954 8974 8995 9016 9036 9057 9078 9099 2468 10 12 15 17 19 96 97 98 99 9120 9333 0550 9772 9141 9354 9572 9795 9162 9376 9594 9817 9183 9397 9616 9840 9204 9419 9638 9863 9226 9441 9661 9886 9247 9462 9683 9908 9268 9484 9705 9931 9290 9506 9727 9954 9311 9528 9750 9977 2468 2479 2479 2579 11 11 11 11 13 15 17 19 13 15 17 20 13 16 18 20 14 16 18 20 USEFUL CONSTANTS 275 Angle. Chord. Sine. Tangent. Co-tangent. 90 57-2900 28-6363 19-0811 14-3007 Cosine. 1-5708 1-5533 1-5359 1-5184 1-5010 De- grees. 1 3 4 Radians. 0175 0349 0524 0698 1 9998 9994 9986 9976 1-414 90 017 035 052 070 0175 0349 0523 0698 0175 0349 0524 0699 1-402 1-389 1-377 1-364 89 88 87 86 5 0873 087 105 122 140 157 0872 0875 11-4301 9 5144 8-1443 7-1154 6-3138 9962 1-351 1-338 1-325 1-312 1-299 1-4835 1-4661 1-4486 1-4312 1-4137 85 6 7 8 9 ' 1047 1222 1396 1571 1045 1219 1392 1564 1051 1228 1405 1584 9945 9925 9903 9877 84 83 82 81 10 1745 174 1736 1763 5-6713 9848 1-286 1-3963 80 11 12 13 14 1920 2094 2269 2443 192 209 226 244 261 1908 2079 2250 2419 1944 2126 2309 2493 5-1446 4-7046 4-3315 4-0108 9816 9781 9744 9703 1-272 1-259 1-245 1-231 1-3788 1-3614 1-3439 1-3265 79 78 77 76 15 2618 2588 2679 37321 9659 1-218 1-3090 75 16 17 IS 19 2793 2967 3142 3316 278 296 313 330 347 2756 2924 3090 3256 3420 2867 3057 3249 3443 3-4874 3-2709 3-0777 2-9042 9613 9563 9511 9455 1-204 1-190 1-176 1-161 1-2915 1-2741 1-2566 1-2392 74 73 72 71 70 69 68 67 66 65 64 63 62 61 60 20 3491 3665 3840 4014 4189 4363 3640 27475 9397 1-147 1-2217 21 22 24 25 26 27 28 29 364 382 399 416 3584 3746 3907 4067 3839 4040 4245 4452 2-6051 2-4751 2-3559 2-2460 9336 9272 9205 9135 1-133 1-118 1-104 1-089 1-2043 1-1868 1-1694 1-1519 1-1345 1-1170 1-0996 1-0821 1-0647 433 4226 4663 2-1445 9063 8988 8910 8829 8746 1-075 4538 4712 4887 5061 5236 450 467 484 501 4384 4540 4695 4848 4877 5095 5317 5543 2-0503 1-9626 1-8807 1-8040 1-060 1-045 1-030 1-015 30 518 5000 5774 1-7321 8660 1-000 1-0172 31 32 33 34 5411 5585 5760 5934 6109 534 551 568 585 5150 5299 5446 5592 6009 6249 6494 6745 1-6643 1-6003 1-5399 1-4826 8572 8480 8387 8290 985 970 954 939 1-0297 1-0123 9948 9774 59 58 57 56 55 35 601 5736 7002 1-4281 8192 923 9599 36 37 38 39 6283 6458 6632 6S07 618 635 651 668 5878 6018 6157 6293 7265 7536 7813 8098 1-3764 1-3270 1-2799 1-2349 8090 7986 7880 7771 908 892 877 861 9425 9250 9076 8901 54 53 52 51 40 6981 684 6428 8391 1-1918 7660 845 8727 50 41 42 43 44 7156 7330 7505 7679 700 717 733 749 6561 6691 6820 6947 8693 9004 9325 9657 1-1504 1-1106 1-0724 1-0355 7547 7431 7314 7193 829 813 797 781 '8552 8378 8203 8029 49 48 47 46 45 7854 765 7071 Cosine. 1-0000 roooo Tangent. 7071 Sine. 765 7854 Radians. 45 Co-tangent. Chord. De- grees. Angle. T 2 INDEX A. ACCURACY of large testing machine, test for, 47, 217 Admiralty specification for cast iron, 163 ; for machinery, 264 Alloys, new, 8 Alloys Eesearch Committee, 143 Alternate tension and compression, 220 Alternating stress machines, 1 68 ; Arnold's, 175 ; J. H. Smith's, 178 ; Stanton's, 177 Alternating torsion machine, C. A. M. Smith's, 182 Aluminium, 8 ; tension tests on, 86 ; effect of boiling on elastic properties of, 128 Amsler testing machine, 34 Amsler - Laffon beam testing machine, 203 Angle of twist, 117 Annealing, 12 Antimony, 235 Appearance of compression speci- men, 92 Arnold, Dr. J., 12, 168, 175, 177, 260; alternating stress machine, 175 Ashcroft's extensometer, 56 Attachment, double autographic, 82 Autographic recorders, 72 ; method, Kennedy's, 78; diagram, 80; attachment, double, 82 ; dia- gram, test to fracture with, 218 Automatic testing machine.-;, 41 A very, Messrs. W. and T., Ltd., 26, 254 ; testing machine, speci- fication of, 31 ; torsion machine, 121 ; impact machine, 137 Axial loading, the problem of, 46 71 Ayrton, Prof., 67 13. BAILEY transverse testing ma- chine, 39 ; torsion machine, 119; cement testing machine, 193; cement crushing machine, 195 ; testing machine, experi- ments with, 217 Bairstow, Mr. Leonard, 103, 188 Balance weight, to check the weight of the, 50 Balls, crushing tests of, 221 ; Brinell's test on, 147 Barnes, Mr. E. J., 260 Bauschinger, 25, 102, 181 ; instru- ment, 62 Beam testing machine, 164 Beams, deflection of, 39 ; breaking of, 220 ; testing of, 89, 220 ; shearing tests on short, 221 Bedding, effect of, 219 Bending tests, arrangement of a machine for, 39 stress machines, 172; standard angle of, 186; effect of speed on, 188 ; and torsion combined, 221, 251 Benedicks of Upsala, 149 278 INDEX Berlin, 25 Bessemer steel, heat treatment of, 105, 260 Bibliography, 265 Birmingham, University of, 25 Blount, Kirkaldy and Sankey, Messrs., 140 Boiler, combined stress in a, 236 ; factor of safety of a, 237 Brass, 238 ; annealed, 12 ; rupture of, 87 ; rod test of a, 109 Brayshaw salt bath furnace, 261 Breakdown, elastic, 13 Bricks, testing of, 206 Briquettes, standard, 190 Brinell's ball test, 143 ; for cement, 200 Brittle materials in torsion, 116 ; criterion of strength for, 251 Brittleness, 9 Buckton & Co., Messrs., 44, 82 Burr, Prof., 166 C. ' ' C," determination of , 2 1 7 ; values of, 133, 216 Calculation of stresses, Guest's, 240 Calibration of vertical machines, 47 Calvert and Johnson, Messrs., 147 Cambridge extensometer, 59 Cambridge Scientific Instrument Co., the, 145 Carbide, iron, 11 Carbon, 8 Cast iron, compression of, 94; in shear, 157 ; roller test, 163 ; rules for, 235 ; combined bend- ing and torsion of, 251 Cast steel, torsion of, 119 Castings for machinery, steel, 234 Cement testing machine, 192; crushing machine, Bailey, 198 ; Brinell's ball test for, 200 Cementite, 11 Centres of higher education, 4 Chatellier cement test, 192, 222 Chemical analysis of steel, 8 Chrome-vanadium steel, heat treat- ment of, 264 City and Guilds Technical College, 257 Civil Engineers, Steel Committee of, 251 ; Institution of, 204, 227 Clutch, electric, 42 Coker, Prof., 106, 125, 254 ; instru- ment for measuring torsional and bending strains, 257 College laboratories, experiments in, 213 Combined stress test worked out, 249 ; bending and torsion, fracture of cast iron in, 251 ; testing machine, 251 ; bending and torsion machine, Coker's, 254 Commercial test, full, 5, 85, 217 Composition, the art of, 6 Compound stress, 236 et seq. Compression shackles, 47 ; tests, 89, 219; specimen, appearance of, 92; of cast iron, 94; of concrete, 194, 222 ; of short specimens, 219 ; and torsion, alternate, 220 Concrete, determining conditions of strength of, 195199; tests of, 222 Connecting rods, rules for, 234 Constants, useful, 268273 Copper, 8, 235, 238 ; torsion tests of, 129 ; for pipes, 235 Crank-shafts, rules for, 234 ; com- bined bending and twisting of, 236 Cross-bending tests, 164 Crystalline structure, 9 ; effects of annealing on, 10 INDEX 279 D. DE FREMINVILLE, 152 Deflection of beams, 41 Diagram, automatic, 80 ; stress strain, 13 Dillner, 151 Discipline, 4 Dixon, Prof. Stephpn, 31 Doubling tests, 164 Ductile materials in torsion, 117 ; tests of, 85, 219 Ductility, 4 ; of steel, 11 E. "E," values of, 114 East London College, 161, 254 Education, centres of higher, 4 Elastic breakdown, 13 ; limit, 86, 242 ; range, 102 ; limit, deter- mination of by extensometer, 218 ; curve, obtaining, 220 ; limit and yield point, 243 Elasticity, modulus of, 3, 107 ; of materials, 232 Electric clutch, 42 Elongation, percentage, 7 Engineering, 31, 181, 236, 241 Engineering Standards Committee, 84, 189 Equipment, 4 Swing's extensometer, 52, 109 Examinations, 4 Extensibility, 13, 14 Extension, distribution of, 87 ; of a long wire, 216 Extensometer, 52,218 ; Ewing's, 52, 109 ; Unwin's, 54 ; Marten's, 56 ; Ashcroft's, 56 ; Kennedy's, 57 ; the Cambridge, 59 ; Marten's mirror, 63 ; Morrow's, 63 ; Stro- meyer's optical, 63 ; " rolling pin," 78 F. FACTOR of safety, 3 ; of a boiler, 237 Failure by shear, 93, 95 Fatigue, effect of, 103 ; testing machine, 179 Ferrite, 11 Ferro-concrete beams, testing of, 203 Flaws and surface markings, effects of, 220 Fletcher gas-muffle, 261 Floor tests, ferro- concrete, 204 Folding tests, 164 Forgings for machinery, steel, 234 Fracture, appearance of, 9 ; of a specimen, 13 ; of cement by crushing, 195 ; testing small specimens for, 217; of various materials, 8689, 217 Friction, effect of internal, 94, 107 Furnace, Brayshaw salt bath, 261 G. GAS-MUFFLE, Fletcher, 261 Gauge-length, the, 84 Genie Civil, Le, 200 Glasgow and West of Scotland Technical College, 251 Gold, 7 Goodman, Prof., 183, 251 Graphite, 12 Grips, Wicksteed, 46 Grosvenor Square, floor tests at, 205 Guest, Mr. J. J., 237; law, 237 tt seq. ; tests on steel tubes, 238 Guillery, 149 Gun-metal, torsion of, 118; rules for, 234 H. HAMMERING tests, 221, 234 Hancock, Prof., 241245 280 INDEX Hand bending machine, Sankey's, 184 Hard steel rod, test of a, 109 Hardening, time effect on, 218 ; of Bessemer steel, 261 Hardenite, 11 Hardness of steel, 11; tests, 145; factor, 146 ; numbers, 147 ; scales compared, 154 ; tests on con- crete, 222 ; number and tensile strength, relation between, 222 Heat treatment of iron, 12 ; effect of, 102 ; low-temperature, 104 ; high- temperature, 104 ; on Bessemer steel, 260; of vana- dium-chrome steel, 264 Ilenning's stress strain recorder, 76 Homogeneity, 247 Hooke's law, 78 Howard, 188 Howe, 12 Hummel, Prof., 31, 119 Hysteresis, mechanical, 106; time effect on, 107 I. IMPACT machine, Avery, 137 Impact tester, tensile, 138; testing machine, Seaton and Jude's, 143 ; testing machine, repeat, 143 ; tests, object of, 137, 220 India-rubber, hardness of, 153 Institute, the Iron and Steel, 128, 152, 245, 260, 264 Institution of Junior Engineers, 103 ; of Mechanical Engineers, 137, 140, 141, 143, 249, 267; of Naval Architects, 175 ; of Civil Engineers, 204, 227 Instruments of precision, handling of, 214 Interference of light, 64 International Association for Test- ing Materials, 188 Iron, 7 ; carbide, 1 1 Iron and Steel Institute, 128, 152, 245, 260, 264 J. JUDE, Messrs. Seaton and, 137 K. KEEP'S hardness test, 152 ; test- ing machine, 41 Kennedy, Prof., 78; testing ma- chine, 23 ; exteiisometer, 57 ; autographic method, 78 Kirkaldy, Sankey and Blouiit, Messrs., 40 ; tests on beams, 211 Knife-edge, wear of, 47 L. LABORATORY, the testing of materials, 3 Larard, Mr. 0. E, 44 Laslett, Mr. T., 209 Lead, 8, 235 Leatheriness of a material, 185 Light, interference of, 64 Lilley, Prof., 47 Liverpool, University of, 250 Limit of proportionality, 13 ; elastic, 13 Loading, axial, 46 ; result of non- axial, 244 London, University of, 4 Longitudinal shear in timber, 208 Longmuir, Mr. P., 264 Low heat treatment, effect of, 104 ; effect of in torsion, 126 M. MACHINE Co., Coventry, the, 162 Machinery, steel castings for, 234 ; forgings for, 234 INDEX 281 Manchester School of Technology, the, 204 Manganese, 8 Marking out specimens, 214 Marten, Prof., 166 ; extensometer, 56 ; mirror extensometer, 63 Martensite, 11 Mason, Mr., 250 Materials, structure of, 8 ; micro - structure of, 9 ; of construction, tests on the chief, 227230 ; properties of, 231 ; strength and elasticity of, 232 Maximum stress in a tensile test, 72 ; shear stress, 237 Me William, Prof. A., 260 Mean diameter, determination of, 91 Mechanical properties, 7 ; treat- ment, effect of, 102 ; advantage of a large testing machine, 47, 217 Metallography, 9 Micrograph of cast steel, 13 Microstructure, 9, 221 Mild steel plates in tension, 100 ; test of, 110; in shear, 157; punching test on, 160 Mirror extensometer, Marten's, 63 Modulus of elasticity, 3, 107; shear, 3 ; by bending, 111 ; of rigidity, 132 ; for wood, 211 Morrow's extensometer, 63 Mother-of-pearl, 10 Munich, 25 Muntz metal, ultimate tension tests on, 86, 113, 131 N. NATIONAL Physical Laboratory, report, 62 ; impact apparatus, 141; alternating stress machine, 177 Naval brass, rules for, 235 Navier's theory, 98 Nickel, impact tests on, 142 ; steel, Hancock's tests on, 245 Non-axial loading, results of, 244, 247 Non-ductile materials, tests of, 89 Northampton Polytechnic Institute, 43 Notched specimens, impact tests on, 220 Notches, effect of, 219 Note-books, 213 0. " OMNIBUS " testing machine, 16 Optical instruments, 62 ; extenso- meter, Stromeyer's, 63 Osborne Eeynolds, Prof., 177 Osciltograph, stress strain, 181 Overstrain, effect of boiling on, 104, 219 ; effect of, 125, 218 Owens College, 177 P. PALLADIUM, 8 Pearlite, 10 Percentage elongation, 7 Permanent set, 13 Perry, Prof., 67 Philosophical Magazine, 241 Phosphorus, 8 Physical Review, 106 Physical Society, the, 251 Pig iron, grey, 12 Pinewood, hardness of, 153 Pipes, copper for, 235 Piston rods, rules for, 234 Plastic stage, the, 86 Platinum, 7 Poisson's ratio, 63, 134 ; values of, 135, 216 Polytechnic Institute, Northamp- ton, 43 282 INDEX Popple well, Mr. W. C., 204 Portland cement, standard test of, 189 Potassium chloride, 262 ; bichro- mate, 262 Propeller shafts, rules for, 234; analysis of cuttings from, 235 Properties, mechanical, 8 ; of materials, 231 Proportionality, limit of, 13 Punching tester, 159 Pyrometer, platinum resistance, 261 0. QUENCHING of Bessemer steel, 261 E. EATIO of maximum to mean stress, 243 Eead and Macdonald, Messrs., 204 Eecorder, the Wicksteed, 79 ; autographic, 72 Eeduction in area, percentage, 7 Eeinforced concrete, tests of, 202 Eepeat impact testing machine, 143 ; stresses, 183, 221 Eeport of the test, 5, 215 Revue de Metallurgie, La, 200 Eiehle testing machine, 36,43, 121 Eigidity, modulus of, 132 Eing-shaped specimen, 182 Eivet shear, 159 Eoller tests, 163 Eolling-pin strain indicator, 57 Eough shop tests, 163 Eoyal Society, Philosophical Trans- actions of, 103, 188 Eupture, characteristics of, 87 S. SAFETY, factor of, 3; of a boiler, factor of, 237 Samples, preparing, 8 Sankey, Messrs. Blount, Kirkaldy and, 140 Sankey, Captain, 168, 264 Sankey's hand bending machine, 184 " Science Abstracts," 264 Sclerometer, 152 Scleroscope, Shore's, 152 Scoble, Mr. Walter, 241, 245, 251 ; results obtained by, 243 Scratch test, the, 152 Seaton and Jude's impact testing machine, 143 Sensitiveness of large testing machine, testing the, 47, 217 Set, permanent, 13 Setting test, 222 Shackles, design of, 44 ; compres- sion, 47 Shafts, rules for crank and propeller, 234 Shear, modulus, 3, 135 ; failure by, 93, 95 ; tests, double and single, 221 ; on the chief materials of construction, tests in, 227 230 ; stress, maximum, 237 Shearing tests on short beams, 221; tests, 155; shackles, 155; stress, constancy of, 240 Sheffield University, 12, 177 Shock, 8 Silicon, 8 Silver, 7, 261 Smith, Prof. J. H., 178 Smith, 0. A. M., 182 Sodium, light, 64 ; chloride, 262 Spangenburg, 181 Specific gravity test, 222 Specimen, mild steel, 80 ; tension, 83 ; making out the, 214 Spherical seat, 46 Sphingometer, the, 65, 247 ; test- ing with the, 115; the torsion, 124 Spring, relation between load and extension of a, 215; com- pression of a, 215; determina- tion of " " with a, 217 INDEX 283 Standard test pieces, 84 ; angle of bending, 186 Standards Committee, British Engineering, 84 Stanton, Dr., 143, 168, 182 ; alter- nating stress machine, 177 Stead, Mr. J. B., 176 Steel, 7, 238 ; physical properties of, 7 ; hardened, 1 1 ; tenacity of, 11 ; balls, tests of, 161 ; castings for machinery, 234; forgings, 234 ; tubes, Guest's tests on, 239 ; Committee of Civil Engineers, 251 ; chrome- vanadium, 264 Stephen Dixon, Prof., 31 Stillion, 19 Stones, testing of, 206 Strain indicator, Stromeyer's, 57 Strength, ultimate, 13; of mate- rials, 3,232; of concrete,, deter- mining conditions of, 195 199 Stress, compound, 236 et seq. ; failure by shear, 93, 95 Stresses, repeat, 181, 221; calcula- tion of, Guest's, 240 Stress strain diagram, 13 ; re- corder, Unwin's, 73; Wick- steed's, 75 ; Henning's, 76 Stromeyer's strain indicator, 57 ; optical extensometer, 63 Structure of materials, 8; crystal- line, 9 Struts, tests of, 47, 219 Sulphur, 261 Surface markings, effect of, 220 Sydney, University of, 149, 202 T. TEAK, hardness of, 153 Tempering of Bessemer steel, 262 Tenacity of steel, 11 Tensile impact tester, 138 ; tests of concrete, 222; strength and hardness number, relation ' be- tween, 222 Tension specimens, 83 ; and com- pression, alternate, 220 ; on the chief materials of construction, tests in, 227230; stresses, alternating, 170 Test, report of the, 5 ; pieces standard, 84 ; of a ductile material, 85 ; of a non-ductile material, 89 ; compression, 89 Tests, the various, 14 Testing machine, the "Omnibus," 16 ; types of, 18 ; vertical, 18 ; horizontal, 19 ; Wicksteed, 21; Kennedy, 23 ; Werder, 24 ; at Birmingham University, 25 31 ; Avery, 31; Amsler, 34; Biehle, 36, 43 ; Bailey transverse, 39 ; Keep's, 41 ; automatic, 41; beam, 64 ; cement, 192 ; Bailey cement, 193; experiments with, 217; combined stress, 251 Thermo-couple, Paul, 261 Thick cylinder, to test a, 222 Thuvston torsion machine, 121 Timber, tension tests of, 207, 221 ; compression tests of, 207 ; bending tests of, 207 ; shear strength of, 212; longitudinal shear in, 208 ; tests in crush- ing, 221 ; beams, testing of long, 221 Time effect, 104, 245, 247; in torsion, 128; on wood beams, 211 ; on hardening, 218 Tin, 8, 235 Torsion machine, the Bailey, 119; machine, the Avery, 121 ; sphin- gometer, 124 ; experiments on wire, 135 ; specimens, fracture of, 220 ; combined bending and, 221 ; on the chief materials of construction, tests in, 227 230 ; and bending strains, Prof. Coker's instrument for measur- ing, 257 Torsion -meter, 238 284 INDEX Torsional vibrations, 36, 216; stresses, alternating, 168 Toughness, 9 Transactions of the Institute of Naval Architects, 175 Tubes, combined stress on, 238 Turner, Prof., 152 Twist, angle of, 117 U. ULTIMATE strength, 3, 237 University of London, 4 ; of Shef- field, 12, 177, 260 ; of Sydney, 149, 202; of Durham Philosophical Society Proceedings, 188 ; of Liverpool, 250 Unwin, Prof., 146 Unwin's extensometer, 54 ; stress strain recorder, 73 ; hardness tests, 143 Upsala, Benedicks of, 149 Useful constants, 268 tt seq. V. VANADIUM-CHROME steel, heat treatment of, 264 Vibrations, torsional, 36, 216 Vienna, 25 W. WAHLBERG, 262 Warren, Prof., 149, 202 Wear of knife-edge, 47 Werder testing machine, 24 Whipple recorder, 261 White metal, rules for, 235 Wicksteed testing machine, 21 ; grips, 46 ; stress strain recorder, 75 ; recorder, the, 79 Wire, torsion experiments on, 135, 216 Wires, tests on, 165, 216 Wohler, 168, 181 Wood, tests on, 207 ; shear strength of, 212 Wooden beams, time effect on, 211 Work done in fracturing a speci- men, 166 Wrought iron, fracture of, 9 ; rod, test of, 111 ; in shear, 157 Y. YIELD-POINT, 13; by drop of the beam, 217 ; stress, 237 ; raising the, 100; testing a small speci- men for, 217; Scoble's deter- mination of, 242 ; and elastic limit, 243 ; determination of, 245 Young's modulus, determination of, 218; by the extension of a short wire, 165, 215 Z. Zmc, 8, 235 Zurich, 25 BRADBURY, AGNEW, & CO. LD., PRINTERS, LONDON AND TONBRIDGE. VAN NOSTRAND'S "Westminster Series Bound in uniform style. Fully Illustrated. Price $2.00 net each. The Volumes in the " Westminster" Series have been designed to meet the growing demand for books on practical subjects ; to bring within the ken of the non-technical reader an accurate knowledge of manufacturing processes and the practical appli- cation of modern science to industries. Each volume is written by an expert to the end that practical readers and all who are engaged in the numerous allied branches of the engineering and technical trades may have reliable works of reference. The series provides for a class not hitherto reached in published works. The volumes can be easily read by the general public, and make excellent handbooks at a moderate price for the student. The series is well suited to public libraries and will be found valuable for libraries in engineering shops and factories. D. VAN NOSTRAND COMPANY Publishers and Booksellers 23 Murray and 27 Warren Streets, New York. THE "WESTMINSTER" SERIES Goal. By JAMES TONGE, M.I.M.E., F.G.S., etc. (Lecturer on Mining at Victoria University, Manchester). With 46 Illustrations, many of them showing the Fossils found in the Coal Measures. LIST OF CONTENTS : History. Occurrence. Mode of Formation of Coal Seams. Fossils of the Coal Measures. Botany of the Coal-Measure Plants. Coalfields of the British Isles. Foreign Coalfields. The Classification of Coals. The Valuation of Coal. Foreign Coals and their Values. Uses of Coal. The Production of Heat from Coal. Waste of Coal. The Preparation of Coal for the Market. Coaling Stations of the World. Index. This book on a momentous subject is provided for the general reader who wishes accurate knowledge of Coal, its origin, position and extent, and its economical utilization and application. Iron and Steel. By J. H. STANSBIE, B.Sc. (Lond.), F.I.C. With 86 Illustrations. LIST OF CONTENTS: Introductory. Iron Ores. Combustible and other materials used in Iron and Steel Manufacture. Primitive Methods of Iron and Steel Production. Pig Iron and its Manu- facture. The Refining of Pig Iron in Small Charges. Crucible and W T eld Steel. The Bessemer Process. The Open Hearth Process. Mechanical Treatment of Iron and Steel. Physical and Mechanical Properties of Iron and Steel. Iron and Steel under the Microscope. Heat Treatment of Iron and Steel. Elec- tric Smelting. Special Steels. Index. The aim of this book is to give a comprehensive view of the modern aspects of iron and steel, together with a sufficient account of its his- tory to enable the reader to follow its march of progress. The methods of producing varieties of the metal suitable to the requirements of the engineer, foundryman and mechanician are described so that the worker may learn the history of the material he is handling. Natural Sources of Power. By ROBERT S. BALL, B.Sc., A.M.Inst.C.E. With 104 Diagrams and Illustrations. CONTENTS : Preface. Units with Metric Equivalents and Abbre- viations. Length and Distance. Surface and Area. Volumes. Weights or Measures. Pressures. Linear Velocities, Angular Velocities. Acceleration. Energy. Power. Introductory Water Power and Methods of Measuring. Application of Water Power to the Propulsion of Machinery. The Hydraulic Turbine. Various Types of Turbine. Construction of Water Power Plants. Water Power Installations. The Regulation of Turbines. Wind Pressure, Velocity, and Methods of Measuring. The Application of Wind Power to Industry. The Modern Windmill. Con- structional Details. Power of Modern Windmills. Appendices A,B,C Index. Two departments of Engineering and their applications to industry form the subject of this volume : the " natural " sources of water ( 2 ) THE "WESTMINSTER" SERIES and wind power which supply mechanical energy without any inter- mediate stage of transformation. Most people will be surprised at the extent to which these natural power producers are used. The widespread application of water power is generally known, but it is interesting to learn that the demand for windmills was never so great as it is to-day, and there are signs of abnormal expansion in the direc- tion of their useful application in the great agricultural countries of the world. Though primarily of importance to the engineer, this work will be of great interest to every manufacturer who in economizing his means of power production can take the natural forces that lie to his hand and harness them in his service. The author is the son of Sir Robert Ball, the eminent mathematician and astronomer. Liquid and Gaseous Fuels, and the Part they play in Modern Power Production. By Professor VIVIAN B. LEWES, F.I.C., F.C.S., Prof, of Chemistry, Royal Naval College, Greenwich. With 54 Illustrations. LIST OF CONTENTS : Lavoisier's Discovery of the Nature of Com- bustion, etc. The Cycle of Animal and Vegetable Life. Method of determining Calorific Value. The Discovery of Petroleum in America. Oil Lamps, etc. The History of Coal Gas. Calorific Value of Coal Gas and its Constituents. The History of Water Gas. Incomplete Combustion. Comparison of the Thermal Values of our Fuels, etc. Appendix. Bibliography. Index. The subject of this book has, during the last decade, assumed such importance that it is hoped this account of the history and develop- ment of the use of various forms of combustible liquids and gases for the generation of energy may do some service in its advancement. Electric Power and Traction* By F. H. DAVIES, A.M.I.E.E. With 66 Illustrations. LIST OF CONTENTS : Introduction. The Generation and Distri- bution of Power. The Electric Motor. The Application of Electric Power. Electric Power in Collieries. Electric Power in Engineering Workshops. Electric Power in Textile Factories. Electric Power in the Printing Trade. Electric Power at Sea. Electric Power on Canals. Electric Traction. The Overhead System and Track Work. The Conduit System. The Surface Contact System. Car Building and Equipment. Electric Rail- ways. Glossary. Index. The majority of the allied trades that cluster round the business of electrical engineering are connected in some way or other with its power and traction branches. To members of such trades and callings, to whom some knowledge of applied electrical engineering is desirable if not strictly essential, the book is particularly intended to appeal. It deals almost entirely with practical matters, and enters to some extent into those commercial considerations which in the long run must overrule all others. (3 ) THE "WESTMINSTER" SERIES Town Gas and its Uses for the Production of Light, Heat, and Motive Power, By W. H. Y. WEBBER, C.E. With 71 Illustrations. LIST OF CONTENTS : The Nature and Properties of Town Gas. The History and Manufacture of Town Gas. The Bye-Products of Coal Gas Manufacture. Gas Lights and Lighting. Practical Gas Lighting. The Cost of Gas Lighting. Heating and Warm- ing by Gas. Cooking by Gas. The Healthfulness and Safety of Gas in all its uses. Town Gas for Power Generation, including Private Electricity Supply. The Legal Relations of Gas Sup- pliers, Consumers, and the Public. Index. The " country," as opposed to the " town," has been denned ''as " the parts beyond the gas lamps." This book provides accurate knowledge regarding the manufacture and supply of town gas and its uses for domestic and industrial purposes. Few people realize the extent to which this great industry can be utilized. The author has produced a volume which will instruct and interest the generally well informed but not technically instructed reader. Electro-Metallurgy* By J. B. C. KERSHAW, F.I.C. With 61 Illustrations. CONTENTS : Introduction and Historical Survey. Aluminium. Production. Details of Processes and Works. Costs. Utiliza- tion. Future of the Metal. Bullion and Gold. Silver Refining Process. Gold Refining Processes. Gold Extraction Processes. Calcium Carbide and Acetylene Gas. The Carbide Furnace and Process. Production. Utilization. Carborundum. Details of Manufacture. Properties and Uses. Copper. Copper Refin- ing. Descriptions of Refineries. Costs. Properties and Utiliza- tion. The Elmore and similar Processes. Electrolytic Extrac- tion Processes. Electro-Metallurgical Concentration Processes. Ferro-alloys. Descriptions of Works. Utilization. Glass and Quartz Glass. Graphite. Details of Process. Utilization. Iron and Steel. Descriptions of Furnaces and Processes. Yields and Costs. Comparative Costs. Lead. The Salom Process. The Betts Refining Process. The Betts Reduction Process. White Lead Pro- cesses. Miscellaneous Products. Calcium. Carbon Bisulphide. Carbon Tetra-Chloride. Diamantine. Magnesium. Phosphorus. Silicon and its Compounds. Nickel. Wet Processes. Dry Processes. Sodium. Descriptions of Cells and Processes. Tin. Alkaline Processes for Tin Stripping. Acid Processes for Tin i Stripping. Salt Processes for Tin Stripping. Zinc. Wet Pro- cesses. Dry Processes. Electro-Thermal Processes. Electro- Galvanizing. Glossary. Name Index. The subject of this volume, the branch of metallurgy which deals with the extraction and refining of metals by aid of electricity, is becoming of great importance. The author gives a brief and clear account of the industrial developments of electro-metallurgy, in lan- guage that can be understood by those whose acquaintance with either ( 4 ) THE ' WESTMINSTER ' SERIES chemical or electrical science may be but slight. It is a thoroughly practical work descriptive of apparatus and processes, and commends itself to all practical men engaged in metallurgical operations, as well as to business men, financiers, and investors. Radio-Telegraphy* By C. C. F. MONCKTON, M.I.E.E. With 173 Diagrams and Illustrations. CONTENTS : Preface. Electric Phenomena. Electric Vibrations. Electro-Magnetic Waves. Modified Hertz Waves used in Radio - Telegraphy. Apparatus used for Charging the Oscillator. The Electric Oscillator : Methods of Arrangement, Practical Details. The Receiver : Methods of Arrangement, The Detecting Ap- paratus, and other details. Measurements in Radio-Telegraphy. The Experimental Station at Elmers End : Lodge-Muirhead System. Radio - Telegraph Station at Nauen : Telefunken System. Station at Lyngby : Poulsen System. The Lodge- Muirhead System, the Marconi System, Telefunken System, and Poulsen System. Portable Stations. Radio-Telephony. Ap- pendices : The Morse Alphabet. Electrical Units used in this Book. International Control of Radio-Telegraphy. Index. The startling discovery twelve years ago of what is popularly known as Wireless Telegraphy has received many no less startling additions since then. The official name now given to this branch of electrical practice is Radio-Telegraphy. The subject has now reached a thor- oughly practicable stage, and this book presents it in clear, concise form. The various services for which Radio-Telegraphy is or may be used are indicated by the author. Every stage of the subject is illustrated by diagrams or photographs of apparatus, so that, while an elementary knowledge of electricity is presupposed, the bearings of the subject can be grasped by every reader. No subject is fraught with so many possibilities of development for the future relationships of the peoples of the world. India-Rubber and its Manufacture, with Chapters on Gutta-Percha and Balata. By H. L. TERRY, F.I.C., Assoc.Inst.M.M. With Illustrations. LIST OF CONTENTS : Preface. Introduction : Historical and General. Raw Rubber. Botanical Origin. Tapping the Trees. Coagulation. Principal Raw Rubbers of Commerce. Pseudo- Rubbers. Congo Rubber. General Considerations. Chemical and Physical Properties. Vulcanization. India-rubber Planta- tions. India-rubber Substitutes. Reclaimed Rubber. Washing and Drying of Raw Rubber. Compounding of Rubber. Rubber Solvents and their Recovery. Rubber Solution. Fine Cut Sheet and Articles made therefrom. Elastic Thread. Mechanical Rubber Goods. Sundry Rubber Articles. India-rubber Proofed Textures. Tyres. India-rubber Boots and Shoes. Rubber for Insulated Wires. Vulcanite Contracts for India-rubber Goods. ( 5 ) THE 'WESTMINSTER" SERIES The Testing of Rubber Goods. Gutta-Percha. Balata. Biblio- graphy. Index. Tells all about a material which has grown immensely in com- mercial importance in recent years. It has been expressly written for the general reader and for the technologist in other branches of industry. Glass Manufacture. By WALTER ROSENHAIN, Superin- tendent of the Department of Metallurgy in the National Physical Laboratory, late Scientific Adviser in the Glass Works of Messrs. Chance Bros, and Co. With Illustra- tions. CONTENTS : Preface. Definitions. Physical and Chemical Qualities. Mechanical, Thermal, and Electrical Properties. Transparency and Colour. Raw materials of manufacture. Crucibles and Furnaces for Fusion. Process of Fusion. Processes used in Working of Glass. Bottle. Blown and Pressed. Rolled or Plate. Sheet and Crown. Coloured. Optical Glass : Nature and Properties, Manufacture. Miscellaneous Products. Ap- pendix. Bibliography of Glass Manufacture. Index. This volume is for users of glass, and makes no claim to be an ade- quate guide or help to those engaged in glass manufacture itself. For this reason the account of manufacturing processes has been kept as non-technical as possible. In describing each process the object in view has been to give an insight into the rationale of each step, so far as it is known or understood, from the point of view of principles and methods rather than as mere rule of thumb description of manu- facturing manipulations. The processes described are, with the exception of those described as obsolete, to the author's definite know ledge, in commercial use at the present time. Precious Stones. By W. GOODCHILD, M.B., B.Ch. With 42 Illustrations. With a Chapter on Artificial Stones. By ROBERT DYKES. LIST OF CONTENTS : Introductory and Historical. Genesis of Precious Stones. Physical Properties. The Cutting and Polish- ing of Gems. Imitation Gems and the Artificial Production of Precious Stones. The Diamond. Fluor Spar and the Forms of Silica. Corundum, including Ruby and Sapphire. Spinel and Chrysoberyl. The Carbonates and the Felspars. The Pyroxene and Amphibole Groups. Beryl, Cordierite, Lapis Lazuli and the Garnets. Olivine, Topaz, Tourmaline and other Silicates. Phos- phates, Sulphates, and Carbon Compounds. An admirable guide to a fascinating subject. ( 6 ) THE ' WESTMINSTER " SERIES Patents, Designs and Trade Marks : The Law and Commercial Usage* By KENNETH R. SWAN, B.A. (Oxon.), of the Inner Temple, Barrister-at-Law. CONTENTS : Table of Cases Cited Part I. Letters Patent. Intro- duction. General. Historical. I., II., III. Invention, Novelty, Subject Matter, and Utility the Essentials of Patentable Invention. IV. Specification. V. Construction of Specification. VI. Who May Apply for a Patent. VII. Application and Grant. VIII. Opposition. IX. Patent Rights. Legal Value. Commercial Value. X. Amendment. XI. Infringement of Patent. XII. Action for Infringement. XIII. Action to Restrain Threats. XIV. Negotiation of Patents by Sale and Licence. XV. Limita- tions on Patent Right. XVI. Revocation. XVII. Prolonga- tion. XVIII. Miscellaneous. XIX. Foreign Patents. XX. Foreign Patent Laws : United States of America. Germany. France. Table of Cost, etc., of Foreign Patents. APPENDIX A. i. Table of Forms and Fees. 2. Cost of Obtaining a British Patent. 3. Convention Countries. Part II. Copyright in Design. Introduction. I. Registrable Designs. II. Registra- tion. III. Marking. IV. Infringement. APPENDIX B. i. Table of Forms and Fees. 2. Classification of Goods. Part III. Trade Marks. Introduction. I. Meaning of Trade Mark. II. Qualification for Registration. III. Restrictions on Regis- tration. IV. Registration. V. Effect of Registration. VI. Miscellaneous. APPENDIX C. Table of Forms and Fees. INDICES. i. Patents. 2. Designs. 3. Trade Marks. This is the first book on the subject since the New Patents Act. Its aim is not only to present the existing law accurately and as fully as possible, but also to cast it in a form readily comprehensible to the layman unfamiliar with legal phraseology. It will be of value to those engaged in trades and industries where a knowledge of the patenting of inventions and the registration of trade marks is important. Full information is given regarding patents in foreign countries. The Book; Its History and Development* By CYRIL DAVENPORT, V.D., F.S.A. With 7 Plates and 126 Figures in the text. LIST OF CONTENTS : Early Records. Rolls, Books and Book bindings. Paper. Printing. Illustrations. Miscellanea. Leathers. The Ornamentation of Leather Bookbindings without Gold. The Ornamentation of Leather Bookbindings with Gold. Bibliography. Index. The romance of the Book and its development from the rude inscrip- tions on stone to the magnificent de Luxe tomes of to-day have never been so excellently discoursed upon as in this volume. The history of the Book is the history of the preservation of human thought. This work should be in the possession of every book lover. (7) Van NostrandV Westminster" Series LIST OF NEW AND FORTHCOMING VOLUMES. The Gas Engine. By Captain RYALL SANKEY, M.I.M.E Timber. By J. R. BATERDEN, A.M.I.C.E. Steam Engines. ByJ. T. ROSSITER, M.I.E.E., A.M.I.M.E. Electric Lamps. By MAURICE SOLOMON, A.C.G.L, A.M.LE.E. The Railway Locomotive. By VAUGHAN PENDRED, M.I.Mech.E. Pumps and Pumping Machinery. By JAMES W. ROSSITER, A.M.I.M.E. Workshop Practice. By Professor G. F. CHAR- NOCK, A.M.I.C.E., M.I.M.E. Textiles and their Manufacture. By ALDRED BAR- KER, M.Sc. The Precious Metals. By THOMAS K. ROSE, D.Sc., of the Royal Mint. Photography. By ALFRED WATKINS, Past Presi- dent of the Photographic Convention. Commercial Paints and Painting. By A. S. JEN- NINGS, Hon. Consulting Examiner, City and Guilds of London Institute. Decorative Glass Processes. By A. L. DUTHIE. Brewing and Distilling. By JAMES GRANT, F.C.S. Wood Pulp and Its Applications. By C. F. CROSS, E. J. BE VAN and R. W. SINDALL. The Manufacture of Paper. By R. W. SINDALL. Wood Working Machinery. By STAFFORD RAN- SOME. D. VAN NOSTRAND COMPANY 'Publishers and Booksellers 23 MURRAY AND 27 WARREN STREETS, NEW YORK. 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. MAR 5 1936 OCT 8 l * ft/fl LD 21-100m-7,'33 3689 UNIVERSITY OF CALIFORNIA LIBRARY