Method to reduce sample size in accelerated reliability verification tests

摘要

A method for determining a sample size W required for accelerated testing of a product includes the steps of selecting a Reliability Goal R as appropriate for the product, selecting a Confidence Level CL appropriate for the accuracy required from the results of the accelerated testing, selecting the number of testing cycles N.sub.t defining the accelerated testing period, calculating the Sample Size W for the accelerated testing as ##EQU1## and then testing the W product samples for the N.sub.t testing cycles to validate the required Reliability when no test failures are observed over the N.sub.t testing cycles. A method for determining the Number of Cycles N.sub.t required for accelerated testing of a product having a service lifetime is also described. In either of the above methods, if any failures are observed in the N.sub.t testing cycles/time, then the number of testing cycles/time may be extended to at least 2 N.sub.t and a new Confidence Level is calculated ##EQU2## The Reliability Goal R for the product design is validated if the new Confidence Levels CL.sub.NEW is greater than a CL.sub.min value specified as a minimum confidence level required for the accelerated testing method.