This section is from the book "Safe Building", by Louis De Coppet Berg. Also available from Amazon: Code Check: An Illustrated Guide to Building a Safe House.
In iron and steel the principal tests resorted to are for tensional strength, elasticity, ductility and limit of elasticity. These will be explained presently. The nature of the material can also be more or less ascertained by fracture and bending. It is almost impossible, however, to distinguish mild steel from wrought-iron excepting by tests for strength, and to ascertain whether or no they can be hard-ened or tempered, as already explained. If specimens of iron or steel on fracture show a close, silky, fibrous grain, with few bright or crystaline specks the metal is good. If there are many crystals in the fracture, and particularly if they are coarse and large, and there are great "blotches" of color, or other evident impurities, the metal is poor. The color of the fibres is grey, that of crystals of a bright silvery aspect. It must be always remembered, however, to get the fracture by planing, or by slow bending, when cold, otherwise crystals will be formed while testing.
Plane off test specimens-
1A skilful moulder will arrange vent-holes and spaces so in the mould at to avoid dirt and blow-holes, no matter what the shape of the piece.
Specimens are frequently nicked and then bent cold until fractured, in such cases there will usually be a large percentage of crystals close to the nicked part, but none in the other half. All iron should be required to be free from all visible seams, blisters, buckles, cinder spots or imperfect edges.
If iron or steel pieces on fracture show the seams where the different layers have been piled and rolled together, that is, if the seams open perceptibly and are very marked, the metal is badly rolled or united. If these seams can be distinctly seen by the naked eye before fracture, the quality is even poorer. If iron or steel for a few hours be immersed in or wetted with some strong acid, such as either nitric, muriatic or sulphuric acid, the parts between the fibres will be eaten away, and the latter will be distinctly exposed. The architect can then readily judge of the nature of the rolling or working, whether the layers (or piles) are thoroughly worked together, and whether the fibres are close and all thoroughly interlaced.
Good wrought-iron should be capable of being bent double, without fracture, while cold, over a cylinder of diameter equal to twice the thickness of the piece. The iron should not crack under this test. If nicked, and then bent double suddenly by a blow from a sledge hammer the fracture should show but a small proportion of crystals and these nearer the nicked edge.
Beams, channels, angles and tees should be capable of having their flanges or legs bent, that is rolled up longitudinally and away from the web or other leg without opening the inside joint between web and flanges, or between legs. In practice, however, the iron usually furnished by the mills will not stand so severe a test. Its quality will be in proportion to its ability to stand this severe test.
Figrure 161 shows a few illustrations of such bending tests, the dotted lines showing outlines of original sections, before bending.
Tests for mild steel should be the same as for wrought-iron. Ductility and elasticity are the same practically as stretching. If the material has high ductility its cross-section will be diminished greatly, when stretched, before breaking. If it has great elasticity it will stretch greatly, that is, become much longer before breaking.
Wrought-iron should be allowed a stretch = 1/225 of an inch for each foot of length of specimen if subjected to a tensional strain of 10000 pounds, per square inch. Every important tensional piece should be tested up to 10000 pounds, per square inch. Those elongating more should be condemned. On removing the strain the piece should recover its original length exactly.
Ductility and Elasticity.
Test pieces should be strained up to 18000 pounds, per square inch, for compression and should then recover their original thickness, they should not crush completely under 36000 pounds, per square inch. Compression test-pieces must, of course, be short blocks or cubes. They should stand at least 25000 pounds, per square inch, in tension without losing their capacity for recovering their original length, and should not tear apart under less than 50000 pounds tension, per square inch for small sections; if sections or plates are large they should stand at least 4G000 pounds tension.
Safe strains in testing.
For steel the above limits must be varied, according to whether mild or hard steel is desired.
All metals, theoretically change their dimensions under the slightest strain. That is, they become longer under tensional strains, and shorter under compressive strains. When the strain is removed the material is supposed to recover its original length. This, however, is true only up to a certain strain ; when this is exceeded the material loses its power to recover and remains permanently elongated or shortened. This permanent elongation or shortening is called the "permanent set" the strain (per square inch) which produces it is called the "elastic limit." This elastic limit varies in different materials, but is approximately at about one-half of the ultimate strain, or strain per square inch that will crush or tear apart the material.