74. In Table 5 are given some of the results obtained by the writer which will serve to show what variations may exist in sieves which have been selected from a considerable number offered for use.

Table 6 gives the data available concerning certain sieves that have been used or recommended in this country and elsewhere.

Table 6. Sizes Of Openings In Sieves Recommended Or In Use

75. The time sifting should be continued will depend on the fineness of the meshes, the diameter of the sieve, the amount of cement taken, and the manner of sifting ; it will also depend upon the fineness of the cement, as well as its nature, and its condition as to dryness. But, although some care is necessary concerning these points, very large variations in results due to variations in the time the sifting is continued may easily be avoided. The diameter of the sieve is usually made greater for the finer meshes, but this is not always the case. It is a common practice in America to use one-tenth of a pound of cement in testing the fineness, using a scale weighing in ten-thousandths of a pound. Where the metric system is in use (and it may well be adopted in a cement laboratory), 100 grams of cement are usually taken.

76. M. H. LeChatelier recommends a sieve having 900 meshes per sq. cm., of wire 0.15 mm. diameter, giving holes 0.18 mm. (.0072 inch) square. He prefers machine screening, but says that for current tests it might be sufficient to screen by hand for ten minutes with a sieve three decimeters (about 12 inches) in diameter.

Table 7 is taken from experiments made by M. Durand-Claye and M. Candlot, and shows what differences may arise from varying the length of time that a sample is screened. The cements used were not the same in the two cases, but the sieves had each 5,000 meshes per sq. cm. (about 180 per linear inch), and 100 grams of cement were taken in each case. Had a coarse sieve been used, the differences would have been much less after the same lengths of time.

Table 7. Fineness. Mechanical And Hand Sifting Compared

77. Table 8 gives the results obtained by the author in - sifting several samples. The No. 80 sieve was about 6 1/2 inches in diameter, and Nos. 120 and 200 about 5 1/2 inches. One hundred grams of cement were taken in each case, and the sieve was shaken vigorously by hand. It is seen that coarse samples require less time for sifting than fine samples, and that natural cements require a longer time than Portlands. With the No. 80 sieve, five minutes usually suffices to obtain the fineness, and with the No. 120 sieve, but little cement usually passed after the sifting had continued ten minutes, though with one brand of natural, Gn, it appears that the true fineness would not be indicated by sifting less than 20 minutes. With the No. 200 sieve 20 minutes is usually required, and in the case of two samples of natural cement, a still longer time appears to be necessary.

Table 8. Effect Of Time Of Sifting On The Result Obtained In Testing Fineness

78. Conclusions

Until there is a proper standard in the United States concerning sieves and methods of sifting, the best that can be done is to select, from the sieves that manufacturers have to offer, those which appear to be most nearly uniform in size of mesh, and then actually determine the size of the holes. This may be done by counting, under the magnifying glass, the number of meshes per inch each way, and determining the size of wire with a micrometer wire gage.

As to the time sifting should be continued, one can easily find by trial the time required in using a given sieve in order to confine the error within given limits. A fine natural cement should be selected to determine this, as such a cement requires the longest sifting. Care should be taken that the cement is well dried before making the test for fineness. It will be found that for sieves having holes between .003 inch and .004 inch square (sieves approximating 170 to 200 meshes per linear inch) 20 to 30 minutes are required, while for sieves having holes .007 to .009 inch square (approximately 70 to 100 meshes per linear inch) from five to ten minutes will usually suffice.

79. Specifications For Fineness

The following table has been compiled to show what are considered reasonable requirements for fineness. In most specifications there is the usual indetermination concerning the sizes of holes in the sieves.

Table 9. Requirements As To Fineness