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: Wed Jul 7 14:16:55 US/Eastern 1999
Subject: Component Reliability (Belcore vs Supplier Data)
I am starting to use the Belcore (Part Stress / Method 1 Case 3) to predict the reliability for various components (esp ICs) but I can also use the supplier's FIT data from their internal demonstration test. Does anybody have any experience on which which method would provide a better result?
Subject: Prediction methods versus vendor data
Reply Posted by: (firstname.lastname@example.org
Organization: Reliability Analysis Center
Date Posted: Fri Jul 16 9:04:46 US/Eastern 1999
In general, prediction methodologies such as Bellcore will provide a more accurate prediction of field reliability than simply using vendor life test data. There are two basic reasons for this. First, handbook prediction models are usually based on vast amounts of field data for part types used in representative applications, and not sometimes unrealistic test scenarios. Second, most all vendor test data is of a highly accelerated nature which is then extrapolated back to typical operating conditions using a derived activation energy, Ea, and the Arrhenius relationship. The problem with this is that the activation energy derived is usually a relatively large number representing only a single failure mechanism. Using this value to extrapolate back to typical operating conditions often results in an unrealistically low predicted failure rate because it does not include other competing failure mechanisms that are also active at lower temperatures.