The greater the number tested, the more reliable will be the mean of the results; but the greater the number, the more time required to make the test. The greater the skill of the operator, the more uniform the results, and hence the fewer briquettes required to obtain any particular degree of accuracy. The accuracy of a series of tests is expressed numerically by the probable error.* "A skilled operator should be able to obtain results having a probable error for a single determination of not more than 4 per cent of the
"An expert working under good conditions may expect to obtain an extreme variation between the results in a set of ten briquettes not exceeding 20 per cent of the mean, and a maximum variation from the mean not exceeding *12 per cent."$ The inexperienced person should make preliminary tests to de termine his degree of uniformity, and one should not attempt to make tests upon which to accept or reject cement' until he can at least approximately approach the above limits. Some expert operators always make ten briquettes for each test, while others make only three to five. If the total number of briquettes made of any one sample of cement requires a larger batch of mortar than can be conveniently mixed at one time, the briquettes for each period should be taken equally from the different batches.
During the first 24 hours the test specimens should be kept in a moist chamber or under a damp cloth to prevent them from drying out. The moist chamber is usually either a soapstone or slate box having a shallow receptacle at the bottom for holding water, or a wooden box with metal lining and inside of that a lining of felt which is kept wet. When cloth is used, it should be kept from coming into direct contact with the briquettes by means of a wire screen or some similar device; and care should be taken to keep the cloth uniformly moist, either by immersing its ends in water or inverting a pan over the cloth and the briquette moulds.
The damp cloth is in more common use than the moist chamber, but the latter is much the better, since it is nearly impossible to prevent the cloth from drying out unequally.
After 24 hours in moist air, the briquettes to be broken at one day should be taken from the chamber and immediately tested; and those to be broken later should be immersed in water at about 70° Fahr. The volume of the water should be at least four or five times the volume of the immersed briquettes; and the water should be renewed about every seven days or there should be a gentle current through the storage tank all the time. In no case should the water be allowed to become "stale" or alkaline by the absorption of lime and salts from the briquettes, as "stale" water may reduce the strength of the briquettes nearly one half.
The briquettes should be labelled or numbered to preserve their identity. The briquettes may be marked with a soft lead pencil or be stamped with steel dies. Neat-cement briquettes may be stamped with steel dies, as may also sand briquettes provided a thin layer of neat cement is spread over one end in which to stamp the number.
Since in many cases it is impracticable to extend the tests over a longer time, it has become customary to break the briquettes at one and seven days. This practice, together with a demand for high tensile strength, has led manufacturers to increase the proportion of lime in their cements to the highest pos sible limit, which brings them near the danger-line of unsoundness. A high strength at 1 or 7 days is usually followed by a decrease in strength at 28 days. Steadily increasing strength at long periods is better proof of good quality than high results during the first few days. The German standard test recognizes only breaks at 28 days. In all cases the time is counted from the instant of adding the water when mixing the briquette. The briquettes should be tested as soon as taken from the water, as drying out materially lowers their strength.
There are three types in common use. In one the weight is applied by a stream of shot, which runs from a reservoir into a pail suspended at the end of the steelyard arm; when the briquette breaks the arm falls, automatically cutting off the flow of shot. In the second type, a heavy weight is slowly drawn along a graduated beam by a cord wound on a wheel turned by the operator. The third type is simply a spring balance. The first form is the most compact, the most rapid, and the most common; the second is the most accurate; and the third is the cheapest and most portable. Each type is made by each manufacturer of cement testing appliances.
The Clips. The most important part of the testing machine is the clips, by means of which the stress is applied to the briquette. There are four important conditions to be fulfilled. 1. The form must be such as to grasp the briquette on four symmetrical surfaces. 2. The surface of contact must be large enough to prevent the briquette from being crushed between the points of contact. 3. The clip must turn without appreciable friction when under stress. 4. The clip must not spread appreciably while sub jected to the maximum load.
Fig. 3 shows three types
in common use. Type 1 is the form recommended by a Committee of the American Society of Civil Engineers in 1904. The surface of contact between the clip and the briquette is } inch wide, and the distance between the centers of the surfaces of contact of a clip is 1} inch. The points of contact should accurately fit the briquette, for if the pressure is not uni formly distributed over this surface the concentrated pressure creates a tendency to break in the clip. At best a considerable pro portion of the briquettes do break in the clips; and several devices have been introduced to prevent it, but none is reasonably success ful. Type 2 is a roller-bearing clip, which permits a rolling contact with the briquette, and has a rear projection, or sometimes a solid back, on the clip to aid in inserting the briquette centrally. Type 3 is a self-adjusting roller-bearing clip so arranged that the briquette may receive the same pressure on both sides as it adjusts itself to the pull.