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: mark schofield (mschofield@wvintenltd.com )
Organization:W Vinten
Date posted: Fri Dec 15 4:54:10 US/Eastern 2000
Subject: Benefits of 883 grade components
Message:
MIL-STD-883 is a parts qualification and screening program. A fundamental requirement of any screening program is that its effectiveness is measured otherwise we don't know if we are actually doing useful work or simply destroying good product. Is there any data available that substantiates the benefits of the 883 screening process or has it simply become a means to an end: "We don't know what benefits it provides but we have to do it to meet the approval". Like many of us, I am stuck in the military spec versus commercial grade component trap and need to justify decisions made as to whether or not the components we select are fit for purpose.


Reply:

Subject: MIL-STD 883 Screeing
Reply Posted by: Bruce Dudley (bdudley@alionscience.com )
Organization: Reliability Analysis Center
Date Posted: Tue Dec 19 12:37:21 US/Eastern 2000
Message:
Screening of microcircuits is an established process that has been performed for at least 30 years with the intent of improving the reliability by removing defects due to manufacturing processes or materials. In a Motorola paper for the IEEE (1996 Proceeding Annual Reliability and Maintainability Symposium) by Sut-Mui Tang using actual burn-in test data on over a million devices, the defect rate was shown to be 36 parts per million at a 95% confidence. This value is better than the expected rate of 60 part per million predicted by D. L. Crook in his paper Evolution of VLSI Engineering (1990 IEEE/IRPS). In a Honeywell paper by J. W. Fink, ATS Plastic IC screening of over one million devices shows electrical fallout of 1.5% and mechanical fallout of 1.02%. Dan Quearry , Navy NSWC Crane, wrote a paper called Environmental Evaluation of Plastic ICís for Military Application proves through accelerated testing that preconditioning improves the reliability of these tests. The preconditioning involved temperature cycling, temperature shock, life test, salt fog and HAST testing. All of these papers seem to say that some level of screening should be performed so that the defects are precipitated before operation. Many vendors perform screen test at the chip level to produce a quality product without specific requirements. These are usually high volume established production lines. In this case extra testing might not be required, however, in the case of small lots on new or limited production runs, extra testing could be necessary. The selection of the tests should be prioritized to eliminate the expected failure mode without degrading the reliability of the device due to overstress conditions. The summary answer to your question is, if you can live with 36 part per million defects then screening is not necessary for your product, however if this is not acceptable, then some form of MIL-STD 883 or equivalent commercial standard should be applied.


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