Testing machines may be divided broadly into three classes according to the manner in which the applied load is measured: (1) lever balance; (2) hydraulic press gage; (3) spring balance. In type (1) the load is applied by means of a screw-press or a hydraulic press but is measured with a lever balance as in the ordinary platform scales. Fig. 2 (a), after Martens, is a schematic representation of this Experiment shows that if a tangential force acts on top of a block fastened at the base, as in Fig. I, the displacement of the top is within certain limits proportional to the force, i.e. F «.l. Using q for the shear stress we may write this in the form q= G tan 0.
In nearly all cases p is so small as to be very closely equal to its tangent where G is the modulus of shear or modulus of rigidity; it depends on the material but in many cases not upon the stress. q is the angle of shear or the shear strain; it must be measured in radians. Typical values of G in pounds per square inch are: Steel, 12,000,000; cast iron, 6,000,000; stone, 1,800,00; wood (across grain), 400,000.
All materials are compressible. If a cubical block is subjected to equal compressive stresses p on all six faces experiment shows that the volumetric strain, Aviv, varies directly as D. i.e.
p KA v where K is the bulk modulus and depends only on the material. As proved in the article ELASTICITY the four constants E, G, K and m are related as follows: m E G (1+ m) =3 K (m-2).
Testing Machines.—It would seem to be comparatively simple to find experimentally the safe stress a body can carry. Yet it is not simple because it is difficult to test for one kind of stress without having some other kind present; furthermore it is not definitely known what stress or combination of stresses is the ultimate cause of failure. The precise meaning of the word ((strength" is not fully understood; that is, we do not know upon what properties of a material its strength depends. The best we can do in the physical testing of materials is to find by direct and, if possible, imitative ex periments the values of the elastic constants and the stresses p and q that enter the formulas derived in the rational mechanics of perfectly elastic bodies. Shear stresses are best found by experiments on circular shafts in torsion; tensile tests are made by subjecting a test piece of specified size and shape to direct pull. As both kinds of tests are made in about the same way we shall, on account of lack of space, describe only the latter. Testing machines dif fer in the devices used for applying and measur ing the loads; they are fully discussed in John son, J. B., The Materials of Construction' (1919) ; Martens, Adolf, 'Handbook of Testing Materials' (translated by Henning, 1899); Unwin, 'Testing of Materials' (1910).
type; it shows only the bare principle and omits the actual arrangement of levers, supporting frame, method of gripping and applying the loud, measuring the extension and details of design. The machines made by Riehle and by Olsen are of this kind; so are those made by Wicksteed in England, Martens in Germany and Marie in France, except that the lever system is attached to the upper shackle or clamp. The Olsen at the Bureau of Standards
in Pittsburgh is the most powerful machine in the world, having a capacity of 10,000,000 pounds compression; it is used for tests on full size columns and masonry piers. In type (2) one of the shackles is connected to the dia phragm of a hydraulic chamber the pressure in which is read on a gage. The load is ap plied at the other shackle either mechanically or hydrostatically. Fig. 2 (b) shows the es sentials of an Emery machine. The Emery, designed at the Watertown arsenal in 1879 for the United States government, differs from other types in having flexible steel plate-fulcra instead of knife edges; the one at the Bureau of Standards in Washington can take specimens 35 feet long and subject them to a push of 2,300,000 pounds or a pull of half as much. The Martens machine is similar to Fig. 2(b) only the diaphragm is so arranged as to com press the water in the gage chamber; in the Whitworth a gage is attached directly to the hydraulic press that applies the load. Type (3) is represented by the Dalby weigh-bar ma chine. (Consult Dalby, 'An Optical Load Ex tension Indicator,' in Proceedings Royal So ciety, Vol. LXX XVI, 1912; (Load Extension Diagrams', ib., Vol. 88, 1913; The Engineer 1917; pp. 422, 453, 468). Dalby describes it thus: ((The instrument is self-contained and its principal element is a hollow bar of fine steel about a foot long. This bar, called the weigh bar, hangs by one end from the upper shackle of a vertical testing machine. The other end is coupled to the specimen bar to be broken, and the lower end of the specimen bar is se cured in the lower shackle of the testing machine. . . . The cross-sections of the bars are proportioned so that the pull applied through them is ultimately able to break the specimen bar without loading the weigh-bar beyond its elastic limit. . . . The stretch of the weigh bar is used to determine the load,* and is meas ured by the movement of a spot of light recorded on a photographic plate. The spot moves horizontally in proportion to the load on the weigh-bar and vertically in proportion to the stretch of the specimen bar; the resultant locus on the photographic plate is the load stretch diagram of the specimen which is being tested. The idea used by Dalby is not new but he has employed it very skilfully. Actual spring balances were in existence long before in small machines for testing paper, cloth and wire; in 1890 Martens designed a 50-ton ma chine having a steel bar instead of a spring and Kennedy and Ashcroft had employed a similar device in 1886. Dalby's form is superior to the earlier ones because the inertia of the moving parts is reduced to a very small minimum and the recording, being photographic, does away with the unavoidable friction of pencil re corders. Diagrams may be taken in less than two seconds from start to break.