Is MIL-STD-781B A Good Reliability Test Specification? Part II

Conspicuous by its absence in Part I was any reference to the field-laboratory numerical reliability relationship. This subject was intentionally avoided, since it is complex enough to be discussed separately. This paper provides further support to Part I, and finds that MIL-STD-781B is a good reliability test specification; it ought not to be replaced by the proposed MIL-STD-781C. Minor changes to MIL-STD-781B are desirable, particularly those which will make the laboratory definition of relevant failure include hardware discrepancies that are now excluded. For example, selected failures during unofficial test time ought to be counted. This paper also provides further support to the Part I conclusion that the Air Force laboratory definition of failure is not compatible with the Air Force field definition of failure.