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: Michelle Hackmann
Date posted: Wed May 12 17:42:11 US/Eastern 1999
Subject: MTBF improvements thru accel life tests
I used a Reliability Software package to evaluate a complex
telecommunication system. The MTBF results were lower than expected. Are there accelerated life tests or anything else that may be implemented in order to improve the MTBF number?
Subject: Accelerated Life Tests
Reply Posted by: Bruce Dudley
Organization: Reliability Analysis Center
Date Posted: Tue May 18 13:33:05 US/Eastern 1999
There are lots of accelerated tests that one can conduct if you know the specific failure mechanisms that are primary in your system. High temperature, high humidity testing is a well known accelerated test along with temperature cycling. Another consideration is the use of vibration stressing. All of these tests will accelerate a number of failure mechanisms, the problem is to develop a stress test plan that does not exceed the electrical and mechanical design or part limitation. MIL-HDBK-344 (Environmental Stress Screening) has several test parameters that can be developed to stress parts, assemblies, or units. Combinations of these stresses can be used to accelerate failure mechanisms and to remove defects from the population. The 1998 Proceedings for Reliability and Maintainability Symposium has a paper by Hank Carouso that is titled "A Fundamental Overview of Accelerated Testing Analytic Models" that is a excellent source of what accelerated tests can be utilized. Of course, any testing must be followed by root cause failure analysis and corrective actions. Redesigns are necessary if one wants to improve the reliability of the item.
Subject: Accelerated Testing
Reply Posted by: Dave White
Date Posted: Tue May 18 20:46:00 US/Eastern 1999
There are quite a few accelerated tests you could try, elevated temperatures, humidity, vibrations. All of these will provide some acceleration. The hard part is, you must resist the temptation to quantify that acceleration. Remember - a year in Seattle is not equal to a year in Pheonix, so which "year" are you simmulating in the test chamber?
Of course I'm assuming that some components of your system are located outside in the messy real world. The situation improves somewhat if the system is in a nice climate controlled base station.