With the drive for digitalization and the Internet of Things the expression “Content is King”, coined by Bill Gates, has now been replaced by “Data is King”. This is especially TRUE in functional safety where the safety and reliability predictions generated from a Failure Modes, Effects & Diagnostics Analysis (FMEDA) are only as good as the failure rate data used to calculate them.
A new whitepaper available from exida discusses what makes for an effective component reliability database and the importance of using good data. It highlights how the limitations of current component reliability handbooks and databases can lead to inaccurate failure predictions (pessimistic or optimistic), missing of important failure modes during design, and to customers creating potentially dangerous designs.
The FMEDA technique, which was developed in the late 1980’s by engineers now working for exida, is a staple of functional safety engineering because of its effectiveness predicting product safety and reliability. However the “Achilles Heel,” the big weakness in any FMEDA, is the component reliability database that is used in the calculations.
Criteria for an Effective Component Reliability Database for use in FMEDA
To overcome this weakness, an effective component reliability database should meet the following criteria:
- Contains data for all components used in the product design
- Documents component failure modes and distributions (not just 50%/50% assumptions)
- Documents component useful life
- Provides failure rates for various design profiles and operating environments
- Is updated on a regular basis to keep up with changes in technology (impact of Moore’s Law)
- Supports IEC 61508 Route 2H architectural constraints
- Includes the impact of analog drift in components
- Includes the impact of integrated circuit soft error rates (SER)
- Data can be loaded easily into an FMEDA tool (e.g., automatically), to minimize the chance of human error and reduce the time it takes to perform the calculations
- Data is calibrated / validated against field failure data to ensure accuracy.
We Did the Homework So You Don’t Have to
We know that if the database does not contain all the components used in your design, then engineers will need to search the internet for the missing information, ending up with data of questionable pedigree. If the database does not contain all necessary failure rate information (based on operating environment, failure modes / distributions, or useful life), then assumptions will need to be made.
Unfortunately, bad assumptions lead to bad results. These are just some of the reasons why exida has focused significant resources on the development of a comprehensive, calibrated CRD, available as part of OEMx V2.0. If you’d like to learn more about the exida CRD and its associated FMEDA tool (FMEDAx), then please reach out to us via the website.