It is generally assumed that a constant pro portion exists between stresses and strains. This relation is expressed by the term "limit of proportionality." and represents the limit be low which each increase of a given amount in the load results in the same change in the length of the specimen. It agrees very closely with the elastic limit, but it is not considered an important factor in engineering calculations, while on the other hand the coefficient or mod ulus of elasticity, based upon the theory of the proportionality between stresses and strains, and obtained by dividing the stress on each unit of area of the section by the resultant strain in each unit of length, of a specimen, is extensively used in calculations pertaining to the deflection of beams and springs. The coefficient of elas ticity for all grades of steel is practically the same, about 29,000,000, and very delicate meas urements are required for its determination, so much so that the values obtained by different observers vary considerably, and places its exact value very much in doubt. For similar reasons, the determination of the exact value of the elas tic limit of a metal is a very difficult matter, and as a rule the values stated as the elastic limit, in the reports of tests of specimens, sent out from rolling mills, is really the "yield point" of the metal, or the point at which the specimen suddenly elongates without any increase of the load, and although it is not theoretically as cor rect a gauge of the property of the metal as the true elastic limit, it is really a more re liable guide to the manufacturer and user, on account of the ease and accuracy with which it is determined by simple practical methods, and the fact that practice structures are not intentionally designed to sus tain a stress as high as the true elastic limit.
Relative to the ultimate or breaking strength of the metal, the tests show that the speci men continues to elongate under a decreas ing load after the maximum load has been applied, so that owing to the reduced sectional area the stress under which the specimen finally breaks is generally much less than the maxi mum, and represents the °tenacity," rather than the ultimate breaking strength of the metal. High elongation indicates a metal of good mal leable and ductile quality, and if produced under a cold test shows that the metal is especially suitable for boiler plate, rivets, etc. Another element considered in the tensile test is the character and appearance of the fracture. It often takes a form described as °cup" or °half cup,° in which one of the ends is wholly or in part concave, while the other is correspondingly convex. In other cases, especially with flat bars, the fractured ends are quite flat and smooth, but with their surfaces making an angle of 45 degrees with the length of the bar, instead of a right angle, as in the case of the former. In hard steel the surfaces present a rough crystalline appearance, while in soft steel they have the glossy appearance of woven silk, or bundles of fine silk fibres. Although the char acter of the fracture is always noted in connec tion with the other data derived from the test, it is seldom made an important requirement in specifications owing to the impossibility of de termining a fixed standard of comparison. Compression tests are made by the same ma chine in slightly different arrangement so that the hydraulic power is pressure instead of pull.
Cold bending tests are useful for the pur pose of determining the ability of steel, or any other metal, to withstand severe distortion, when the metal is cold, under such operations as punching, shearing, flanging and riveting, usu ally employed in the manufacture of bridge members and structural shapes in general. The test does not require an expensive or elaborate plant for its application, and is, therefore, avail able where more complete tests are impracti cable. On the other hand, however, it is very
difficult to deduce accurate conclusions from the data obtained, owing to a lack of proper clas sification and standards of comparison. In making the test, the metal in each case is bent over upon itself through an angle of 180 degrees. The inner radius of the bend varies from 0, in steel with a tenacity less than 62,000 pounds per square inch, to the of the specimen" in steel with a tenacity not greater than 70,000 pounds per square inch. The specimen is re quired to stand the bending without fracturing on the outside of the bend. The test for hard ness is applied by pressing or driving a hard ened steel ball into the material so as to pro duce a dent. The diameter of the ball used is commonly 10 millimeters and the standard pres sure used is 3,000 kilograms. The diameter of the indentation is then measured. Two in dentations are usually made not far apart on the specimen being tested. The formula ap 61 1 plied is x ; this giving the hard 71(n 1.0718 ness number in the Brinell scale. It has been objected to these ball tests that the tendency of the pressure is to raise up a cup-shaped rim of steel around the ball and above the level of the test specimen, and that the diameter meas ured will thus be of the raised cup. To obviate this a delicate instrument has been devised to measure the depth of the indentation below the level of the tested surface, and from this to calculate the true diameter. Another method of applying this ball test is by a falling ham mer which carries the ball upon its face. The standard weight of the hammer is 1.76 kilo grams, and the drop may be any distance up to 90 centimeters. Several tests are made at different falls, and the results correlated. A variation of this test substitutes a cone-shaped point with an angle of 90 degrees for the spherical ball. This also is applied by a stand ard pressure or by the falling hammer, and a closer accuracy is claimed. As determined by these methods the Brinell hardnes% numbers for the commoner grades of steel are as follows: Mild steel 120.
Boiler plate 128.
Rail steel 183.
High carbon steel 236.
Cast iron 213.
Another method of testing steel of great accuracy in the hands of an expert is the *spark test? This is applied by pressing the sample of steel for an instant firmly against a sharp emery wheel revolving 7,000 feet per minute. To the educated eye the kind of sparks emitted indicate with great certainty not only the proportion of carbon in the sample, but the several alloys with tungsten, manganese or chromium.
Other tests are made for endurance of steel samples, especially that used in the automobile industry. Such steels are tested for fatigue, being subjected again and again times without number to the same kind of strain, and also to shocks which they must safely withstand in the ordinary use of a car. These tests are ap plied by automatic machines acting unceasingly for many hours. Another test of great import ance in automobile steel is the abrasion test for wear.
In some cases torsional and shearing tests are made, but these stresses are not commonly expected. The impact test is more often em ployed, applied by means of a heavy, swinging pendulum, or by an oft-repeated drop of a weight upon a bar supported at the ends. The tests of finished steel are similar to those made in the mills during the process of manufacture, with the exception of the methods employed in the preparation of the test specimens. These are cut directly from the finished product in the form of rectangular pieces 10 inches in length, in the larger kinds of work, while small bars are generally tested full size. Consult Bullen, D. H., 'Steel and Its Heat Treatment' (New York 1918).