SRC Forum - Message Replies


Forum: Reliability & Maintainability Questions and Answers

Topic: 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

Back to message list Topics List About this forum
Original Message:

Posted by: Xiangping,Wu (7815@notesmail.huawei.com.cn )
Organization:Huawei Tech.Ltd.,China
Date posted: Mon Jan 3 3:14:14 US/Eastern 2000
Subject: Reliability Prediction
Message:
I did a reliability prediction several days before.This system consists of about 10~15 PCBs,and there are average 500 components on each PCB.I used 217F and Bellcore respectively.MTBFs are 5800h and 40000h.Is that reasonable? After that,I saw a reference about similar system of other company(a famous company).In that reference,they claim their system'MTBF is 1000 years(failure rate is 111 fits).Oh my god!Is that possible?Or am I wrong?


Reply:

Subject: Reliability Prediction
Reply Posted by: bruce Dudley (bdudley@alionscience.com )
Organization: Reliability Analysis Center
Date Posted: Wed Jan 5 14:14:21 US/Eastern 2000
Message:
A reliability prediction of 8 to 9 million hours mean-time-between failure (MTBF) for a 5,000 to 7,500 part system is much better than one should expect given the defect history of most electronic parts. This sounds like a unsubstantiated advertisement for a product. Considering systems of this complexity, such as a desktop CPU, have only shown operating MTBFs of 80,000 to 90,000 hours in relative benign environments. This figure relates to your prediction of 40,000 hours but not the 8 million hour figure. I believe that if you investigate the 8 million hour figure, the company will state that no random defects are expected therefore; only a wearout failure mechanism is accounted. Most electronic part wearout failure mechanisms occur after millions of cycles or hours at benign conditions. Since no random defects need be counted, the time to the first failure is the wearout time of the most significant failure mechanism, resulting in this figure of merit.


Back to message list Topics List About this forum
Reply to this message