Reliability prediction and Modelling Real Life Failure rates from Accelerated testing

Reliability Solutions have set up Customised Prediction Techniques for a range of its clients to ensure Accelerated Reliability Stress Test data can be used to predict Early Life and Longer Life Reliability levels.

Sequential Stress tests are applied in different forms to maximise defect detection capability for all types of electronic products . Correlation with Field data can be made to ensure Prediction Models developed are robust and accurate.

The standard method of predicting Reliability is using a Mil-Std approach which rarely provides any reasonably accurate prediction of expected Field Failure Rate, hence a different approach is required to split prediction into 2 failure periods :

  • Early Life Period
  • Longer Term Useful Life Period

Early Life Failures are predominantly due to Process Type issues and weak components, with Longer Life Failures being more Design and Component Related. Hence Totally Different Test Programmes are required to evaluate the Full Life Reliability Performance of any electronic product.

Traditional MTTF (Mean Time To Failure) testing at elevated temperature is still widely used to try and predict Longer Life Reliability and is fairly effective, but when combined with Humidity and Increased Voltage levels, MTTF testing and ability to induce weaknesses to fail is greatly enhanced.

Reliability Solutions has also developed quite unique methods to optimise Accelerated Stress Testing for Early Life simulation and predict the Early Life Failure Rate. This enables a very accurate prediction of expected Early Life Reliability which is a separate failure distribution to that of Longer / Useful Life period.

In summary the Reliability Solutions model will allow engineers to focus on the 2 key Failure Distributions in a products life cycle and make accurate predictions of expected fail rates and also warranty performance.