Downloaded from https://www.everyspec.com MIL-HDBK-816 NOTICE 1 TABLE XI. Step by step procedure for calculating end-point limits. Step Check-off item 1. Select device type and manufacturers. 2. Obtain characterization data and documentation. 3. Select desired lot acceptance probability, confidence level, and sample size for sample size series tests. 4. Select parameters to be tested and test conditions. 5. Begin data examination. 6. Check data for sufficiency. 7. Check for outlying devices. 8. Check for outlying lots. 9. Discard outliers. 10. Check data for abrupt failures. 11. If abrupt failures have a significant likelihood, parameter end point limits usually cannot be calculated. 12.
Check for unexpected functional dependencies. 13. Check for systematic or nonrandom effects. 14. Check for correlations with respect to time of manufacture. 15. Check for unexpected magnitudes of lot-to-lot variations. 16. Check for bad data points. 17. Correct or discard bad data points. 18. Check data for sufficiency. 19. Recast the data to the specified conditions. 20. Determine within-lot and lot-to-lot variations. 21. If within-lot variations are larger than lot-to-lot, use case 1 analysis method to calculate the RHA end-point limits (RHEPL). 22. If within-lot variations are smaller than lot-to-lot, use case 2 analysis method to calculate the RHEPL. 23.
Incorporate the RHEPL and sample size series test requirements into the device specification. SUPERSEDES PAGE 26 OF MIL-HDBK-816 26