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: Clay Davis
Date posted: Tue Aug 15 2:50:17 US/Eastern 2000
Subject: MTBF Calculations (Customer Complaint)
Several of our computers have an MTBF of 50,000 hrs. However, a major customer states that if they purchase 100 systems 9 will fail in the first 6 months and another 9 will fail by the end of the first year.
Agreed this is numerically accurate (1/(50,000*365*24*100)) - however this does not seem right. I tried to explain that not every failure will cause a system failure (ie a capacitor) but they did not agree without us being able to prove these parts are truly redundant (which they are probably not).
Thoughts? There has to be something that I am missing. Thank you for your help. --Clay
Subject: Computer MTBF Calcualtions
Reply Posted by: Bruce Dudley
Organization: Reliability Analysis Center
Date Posted: Tue Aug 22 11:29:17 US/Eastern 2000
Your assessment of the number of failures expected is based on the assumption that the failure events are random in nature. If the part failure rates are decreasing as the units get fixed, then the basic problems may be early defects which could be eliminated via some type of stress testing. If the problem is one of determining the effect of failure, then you must perform some type of falure mode analysis (FMEA) to determine the impact or lack of impact of failure. Some capacitors that are used as EMI filters could fail and performance in normal conditions would continue after failure of the component. An FMEA analysis could be an effective tool for improving the expected reliability, i.e. finding and fixing problems, and for convincing your customer that the unit is better than advertized.