STEEL, TESTING OF. Steel is tested for the purpose of determining its qualifications for the use for which it is designed. The test may not be simply as to its strength, but also to obtain a measure of several other important qualities such as malleability, ductility and hard ness, which gives it the first rank as a material for structural purposes. Of the three usual tests — tensile, compressive and transverse or bend ing (see TESTING MACHINES), the tensile test gives the simplest and most accurate data from which those properties of steel most important in structural work may be most readily deter mined. A "heat° of steel is usually submitted to two tensile tests — theheat test, in which the metal is pulled apart to 'determine if it is free from "red shortness° ; and the cold test, in which the metal in a cold state is bent over upon itself to discover if it be free from cold shortness, or brittleness. At the mill the speci men to be tested is prepared so as to indicate the general properties of the metal, and its suit ability to the purpose for which it is required, and upon the resulting- data is based the treat ment of the metal during the subsequent oper ations in the mill and at the forge. In selecting the specimen, great care is taken. to secure one that is an average of the heat, so as to obtain uniform results, as nearly as possible, from the operations of heating, rolling, forging and cool ing. Very often apparently inconsequential dif ferences in the methods employed to select and prepare the test specimen give very mislead ing if not absolutely erroneous information. Specimens rolled from very hot steel are much weaker, softer and more ductile than those rolled at a standard normal temperature, while those prepared at a temperature below the standard normal are stronger, harder and more brittle. Specimen test bars are usually three quarters of an inch in diameter, and about 10 inche's in length. Before a test, the exact diam eter of the specimen is measured to within 1,000th of an inch, by means of a micrometer caliper, and about 12 inches of its length near the middle is divided by light scratches or by centre-punch marks at intervals of an inch, from which the teduction in the area of the cross-sec tion, and the amount of stretch under the pull of the load applied, is measured. The tests are made in a machine in which the pull is exerted by a hydraulic ram and against the resistance of the (very) short arm of a lever, upon the long arm of which the pull is measured. In the testing machine, the effect produced on the specimen by the gradually increased load varies greatly for different qualities of metal. Steel containing a large •amount of carbon, of the quality generally used for the manufacture of springs, stretches slightly and uniformly up to the breaking point. In the case of softer steel, at the beginning as the load is gradually in creased the metal stretches uniformly for a little while, but the period is much shorter than that of high carbon (harder) steel, then it stretches very rapidly for a few seconds, with out any appreciable increase of the load, until it is apparently on the point of breaking, when it partially recovers its strength, and stretches slightly but uniformly as the load is gradually increased to the maximum. In th'e case of hard
steel, the metal ruptures under the maximum load, but the soft steel continues to stretch for a little while beyond that point under a decreas ing load, with a great reduction in the sectional area of the specimen. To determine the amount of elongation produced by the test, the frac tured ends of the two pieces of the broken bar are put together and the increase in the lengths of the original inch spaces marked on the bar are measured. In very hard steel, the amount of elongation is very small, but even the hard est and most brittle varieties undergo a meas urable change of length. In soft steel it is very great, varying from 25 to 30 per cent of the original length. The behavior of a specimen in the testing machine when subjected to a gradu ally increasing load is studied by means of a stress-strain diagram which consists of two sets of parallel lines intersecting each other at right angles. The horizontal lines represent the strains in pounds per square inch of the sec tional area of the specimen, and the vertical lines represent the amount of elongation of the specimen at the rate of 0.01 per inch of the original length. When the data obtained by the test are plotted upon the diagram, the behavior of the specimen is indicated by a characteristic curve. Assuming an original length of six inches (between the grips of the machine) for a specimen the behavior of soft steel under a gradually increasing load may be briefly sum marized as follows: Up to a load of 40,000 pounds to the square inch the elongation is very slight, about 0.01 of an inch in a length of six inches. From 40,000 to 43,000 pounds the elongation is more rapid, but the total amount is only 0.02 of an inch. From this point the metal stretches very rapidly with no increase of load until the elongation•amounts to 0.15 of an inch, then it apparently regains some of its strength and stretches slowly and uniformly until the elongation amounts to 1.96 inch, or 33 per cent of the original length, under a maxi mum load of 63,000 pounds to the square inch. An analysis of these results shows that it would require a load greater than 40,000 pounds per square inch to induce a permanent set in the metal, and is, therefore, the value representing its "elastic limit," or the limiting stress below which there is practically no change in the original length of the metal, no matter how often that stress may be applied or removed.