SRC Forum - Message Replies
Forum: Reliability & Maintainability Questions and Answers
Topic: Reliability & Maintainability Questions and Answers
Topic Posted by: Reliability & Maintainability Forum
Organization: System Reliability Center
Date Posted: Mon Aug 31 12:47:36 US/Eastern 1998
Posted by: J dudley
Date posted: Tue Mar 20 10:36:37 US/Eastern 2001
Subject: Testability sample sizes
Sample sizes for testability demostrations MIL hdbk471 seems to be very a little miss leading as towards the sample size required to test for T demos.Could some body explain thier own theory on the methods published in this standard. We belive that 10 percent of the fault catalog should be demostrated.
Subject: Testability Sample Sizes
Reply Posted by: Bruce Dudley
Organization: Reliability Analysis Center
Date Posted: Mon Mar 26 15:17:32 US/Eastern 2001
The testability demonstration procedures in MIL-HDBK-471 are generally based on the log normal distribution, which could be challenged as the results are usually go or no go answers. The resulting sample size from this procedure seems to fall some where between 30 and 100 faults as result of the distribution. To determine how the exact statistics were developed, one would need to obtain the technical reports that were developed back the 1970 and 1980 time frame that support the handbook. I do not know where one would go to obtain this information. Regarding your desire to use 10 percent of the faults could be reasonable, but lacks confidence. I would suggest that the sample size should be related to the one shot statistics using binomial values as the test are really find or not find conditions. On the RAC home page you can find this tool, one shot statistical demonstrations, in our tool box.