Analysis of the results of checking gauges and instruments

Conclusions

Rational values of the permissible errors can be established only by analyzing the probable checking scrap. The latter is determined by technical and economic considerations which include: a) the permissibility of a certain small percentage of unsatisfactory instruments remaining in circulation, b) the permissibility of condemning a certain percentage of serviceable instruments, c) the cost of a checking method which reduces to the required small dimensions or eliminates scrap.

No general relation between the permissible errors in the tested and reference measures or instruments is possible.

A rational and reliable selection of permissible errors and methods of checking is only possible when sufficient data is available on the error distribution laws in the manufactured measures and instruments. The provision of such data should be one of the primary objects in testing standard instruments.

Everything stated above refers to large scale testing of mass produced measures and measuring instruments when the aggregate of both the tested and reference measures and measuring instruments is sufficiently large. Above methods of calculations, however, are applicable to small groups of reference measures and instruments. In that case, however, it is necessary to use the discrete distribution of probable error densities which corresponds to the given number of instruments and their errors. Such a case could occur in evaluating checking results in any laboratory with a small number of reference measures or instruments. Such an evaluation is, of course, only possible if the errors of the available measures are determined with sufficient accuracy by more accurate methods than those normally used in checking these measures.