Reliability Assessment for Commercial DC/DC Converter Considering Stress Variance with Weibull Regression Analysis Method

As commercial DC/DC converters are widely used in space application, their failures can cause the electronic system to work abnormally and lead to severe safety accidents. To predetermine time to failure and generate substantial safety, it is necessary to conduct an accelerated life test and assess reliability in advance. This paper has applied high-temperature accelerated tests to predict DC/DC converters reliability and lifetime under practical temperature conditions considering field stress variation. First, periodic performance parameters measurements, such as output voltage and efficiency over high temperature tests are performed at 100, 120 and 140oC. Additionally, the Arrhenius-Weibull model is developed and the Weibull regression analysis method for time-censored failure data is proposed. The estimation results are compared with maximum likelihood estimation (MLE) and failure-censored approximation method, demonstrating the effectiveness of the regression analysis method. Finally, reliability analysis is established considering variations in field temperature conditions. The whole procedure is illustrated by the time-censored testing results of the ISL8225M series DC/DC converter.