How do we improve our understanding of Reliability and Expected Failure Rate

Reliability Engineers are always expected to know everything about a productÂ’s expected failure rate before any field data available. If failure rate turns out to be much higher than expected or required, the reliability engineers become the focus for attention and often are made the scapegoat.

Other engineers and management simply assume a small sample size and short test time will produce a magical result !

However, help is at hand for the over stressed reliability engineers !

Accelerated Testing is generally used to make some form of reliability prediction but failures under stress testing are not the only failure reasons in field for a component/product. Hence another parameter (maturity level) is required to define possible fieldfailures in field due to range of other stresses;

- ESD related
- Over voltage
- Function at extreme temperatures
- Combined Voltage / Temp stress failures
Etc

Focusing on Design Maturity measurement during development stage testing is a vital contributor to understanding the future product failure performance and the maturity measurement will almost always have a direct correlation to the future field failure level. It also acts as an excellent KPI for RD management to monitor progress of design as it passes through each review stage or milestone point.

Measuring Design Maturity is a concept that has been used for many years , but finding suitable application of the technique is the key to success, several key stages are involved in this;

Designing a strong Design Quality Assurance (DQA) and building in wide range of combined function/stress tests and range of different reliability stress tests to MAXIMISE defect detection

Applying Weighted Test Score for each specific DQA test based on its ability to detect possible failures, with more harsh or extreme tests carrying higher weighted scores

Calculating total score possible for DQA Test Programme, can be as small as 1500 for basic electronic product such as an LCD TV up to 5000 for a more complex product

Monitoring failures and scoring each failure type that is not truly demonstrated as resolved by solution being retested under same conditions as original failure
This is done by assessing SEVERITY level of the defect and multiplying by the RESOLUTION STATUS

Cumulating all score loss for all defects not truly resolved

Calculating Design Maturity as a Ratio of Total Score Achieved (Max DQA Test Score – Defect Score Losses) / Max DQA Test Score

This method has worked extremely well for clients of Reliability Solutions and helped them manage and improve Design Quality / Reliability in a controlled and expedient manner.

It provides simple product benchmarking of different designs and forms a key component of the overall correlation modelling for design test results and expected field failure rates BEFORE mass production and BEFORE field return data is made available