SRC Forum  Message Replies
Forum: Reliability & Maintainability Questions and AnswersTopic: Reliability & Maintainability Questions and Answers
Topic Posted by: Reliability & Maintainability Forum
(src_forum@alionscience.com
)
Organization: System Reliability Center
Date Posted: Mon Aug 31 12:47:36 US/Eastern 1998
Original Message:
Posted by: Floyd Kreuze
Date posted: Tue May 16 11:31:37 US/Eastern 2000
Subject: Lamp Accelerated Testing
Message: I have a question about accelerated life testing of incandescent lamps. Using failure data points from testing at an accelerated (higher than normal use) voltage level, if a Weibull plot were used to determine mean life (with both failures and suspensions), what would be a good average recommended exponent (acceleration factor for lamps) when the Inverse Power Law is used to determine mean life at normal use voltage.
Reply:
Subject: Determination of N for Inverse Power Law
Reply Posted by: Jack Farrell
(jfarrell@alionscience.com
)
Organization: Reliability Analysis Center
Date Posted: Wed May 17 14:05:28 US/Eastern 2000
Message: The Inverse Power Law requires two or more mean life data points to determine the exponent N. Typically, the user already knows the mean life at normal use and conducts a test to determine mean life at some accelerated level. Then using the Inverse Power Law the acceleration factor due to the stress level can be determined. If I understand your question you have an accelerated stress mean life and are seeking a typical value of N to use to determine a normal stress mean life. This is not a realistic situation, as there are too many variables that affect the value of N. Some of these variables include the unit material, construction practices, screening tests, date of manufacture, and numerous others. If determining a mean life at normal stress is impractical due to the length of time required, you could run another accelerated stress test at a different level. From this test determine the mean life, insure the Weibull shape (beta) matches your previous data and then using these two data points determine N. Using this value, you can now estimate the mean life at a normal stress level. The assumption (and risk) here is that the beta for your mean life at normal stress is the same as your accelerated stress data.
