The edges should be struck with a light hammer. If the blow make a slight impression, the iron is probably of good quality, provided it be uniform throughout.
If fragments fly off and no sensible indentation be made, the iron is hard and brittle.
Air bubbles are a common and dangerous source of weakness. They should be searched for by tapping the surface of the casting all over with the hammer. Bubbles, or flaws, filled in with sand from the mould, or purposely stopped with loam, cause a dulness in the sound which leads to their detection.
The exterior surface should be smooth and clear. The edges of the casting should be sharp and perfect.
An uneven or wavy surface indicates unequal shrinkage, caused by want of uniformity in the texture of the iron.
Cast-Iron Pipes should be straight, true in section, square on the ends and in the sockets, the metal of equal thickness throughout. They should be proved under a hydraulic pressure of four or five times the working head. The sockets of small pipes should be especially examined, to see if they are free from honeycomb. The core nails are sometimes left in and hammered up. They are, however, objectionable, as they render the pipe liable to break at the points where they occur.1
For small girders and other castings intended to carry weight, it is usual to test a certain proportion of the number supplied by loading them till they break, and noting the weight under which they give way.
For large castings this system of testing would be too expensive. Small bars are therefore cast from the same metal and at the same time as the castings, and these are tested to fracture by a weight applied at the centre.
Some engineers require that the test bar should be cast with the main casting, and not broken from it until they have seen it.
The test bars are usually about 3 feet 6 inches long, 2 inches deep, and 1 inch wide, with a clear bearing of 3 feet.
The test weight varies, according to the opinion of the engineer, from 16 to 35 cwt.
It is important, however, to ascertain not only the weight that will break the test bar, but also the amount of deflection that will occur before fracture.
The reason for this is that a very hard iron will often bear a considerable cross strain when it is steadily applied, though it would be so brittle as to be unfit for any position in which it would be liable to slight vibration or shocks of any kind.
With regard to this point Mr. Matheson says : -
"A strength capable of enduring 25 cwt. on the test bar without fracture should be the minimum quality allowed even for short and heavy columns; but for other purposes a load of from 28 to 30 cwt., and a deflection of 5/16 inch, should be demanded.
"The deflection will vary from .3 to .5 inch.
"There is no difficulty in getting such iron, and higher qualities can be given if necessary, breaking strains of 30 to 35 cwt. being obtainable with judicious mixtures of the best kinds of iron; and in testing such iron it will generally be found that some of the bars will endure as much as 38 cwt."1
Mr. Stoney points out "a singular fact that there is an excess of about 16 per cent in the weight that a 2-inch by 1-inch test bar will support when cast on edge and proved as cast, over that which it will support when proved with the underside as cast placed at the top as proved, and 8 per cent over the weight which the same test bar will support if cast on its side or end, and proved on edge.
Dr. Pole has pointed out that small cast bars do not give a fair indication of the strength of larger castings run at the same time, for the reasons stated at page 302, in the paragraph headed Size of Section.
The cast-iron sleepers for the Great Indian Peninsula Railway were tested by a falling weight; and test bars, of the ordinary form, cast at the same time, were broken by cross strain; others, having a central section one inch square, were broken by tension.