The hot-water test is similar to the last, but the water is main tained at a temperature of from 195° to 200° F.
The boiling test consists in immersing the specimen in cold water immediately after mixing and gradually raising the temperature of the water to the boiling point, continuing the boiling for three hours.
For an emergency test, the specimen may be prepared as before, and after setting may be held under a steam cock of a boiler and live steam discharged upon it.
The results of accelerated tests must not be accepted too literally, but should be interpreted in the light of judgment and experience.
The cracking or contortion of the specimen (sometimes called "blowing"), is due to the hydration and consequent expansion of the lime or magnesia. If the effect is due to lime, the cement can be im proved by exposure to the air, thus allowing the free lime to slake. This treatment is called "cooling the cement". The presence of uncombined magnesia is more harmful than that of lime.
Some idea of the quality of a cement may be gained by exposing to the air a small cake of neat cement mortar and observing its color. "A good Portland cement should be uniform bluish gray throughout, yellowish blotches indicate poor cement": Tests of soundness should not only be carefully conducted, but should extend over considerable time. Occasionally cement is found which seems to meet the usual tests for soundness, strength, etc., and yet after considerable time loses all coherence and falls to pieces.
Strength. The strength is usually determined by submitting a specimen of known cross-section (generally one square inch) to a tensile strain. The reason for adopting a tensile test is that since even the weakest cement cannot be crushed, in ordinary practice, by direct compression, and since .cement is not used in places where cross strain is brought to bear upon it, torsion being out of the.ques tion, the only valuable results can be derived from tests for tensile strength. In case of a cracking wall the strain is that of tension due to the difference of the direction of the strain caused by the sinking of one part of the wall.
In comparing different brands of cement great care must be exercised to see that the same kind and quality of sand is used in each case, as difference in the sand will cause difference in the results. To obviate this a standard sand is generally used. This consists of crushed quartz of such a degree of fineness that it will all pass a No. 20 sieve (400 meshes to the square inch; wire No. 28 Stubbs' gauge) and be caught on a No. 30 sieve (900 meshes to the square inch; wire No. 31 Stubbs' gauge).
Valuable and probably as reliable comparative tests can be made with the sand which is to be used for making the mortar in the pro posed work. Specimens of neat cement are also used for testing, they can be handled sooner and will show less variation than speci mens composed of cement and sand. • The cement is prepared for testing by being formed into a stiff paste by the addition of just sufficient water for this purpose. When sand is to be added, the exact proportions should be carefully deter mined by weight and thoroughly and intimately mixed with the cement in a dry state before the water is added; and, so far as possible, all the water that is necessary to produce the desired consistency should be added at once and thereafter the manipulation with the spatula or trowel should be rapid and thorough. The mortar so obtained is filled into a mould of the form and dimensions shown in Fig. 3. These moulds are usually of iron or brass. Wooden moulds, if well oiled to prevenf their absorbing water, answer the purpose well for temporary use, but speedily become unfit for accurate work. In filling the mould care must be exercised to complete the filling before incipient setting begins.
The moulds while being charged and manipulated, should be laid on glass, slate, or some other non-absorbent material. The specimen, now called the "briquette", should be removed from the mould as soon as it is hard enough to stand it, without breaking.