Power-related failure mechanisms in the analysis of wireless system availability

Many systems (e.g, wireless systems, routers, etc.) are designed to shut down at high operating temperatures to protect the electronics from overheating. Setting the shutdown threshold high would reduce the frequency of shutdown events, while increasing the system failure rate due to higher component junction temperatures. Setting the threshold low would reverse the relative downtime contributions from the two competing shutdown mechanisms. This work proposes a method to find an optimal threshold by minimizing the sum of the downtimes contributed by system shutdown and hot component failures. The system selected for this study is a CDMA fixed wireless system. The downtime due to component failures is estimated using a proportional hazards model with a time-dependent junction temperature as the covariate. The junction temperature of a RF output transistor in the power amplifier is used to compute the probability of shutdown. Combining the downtimes from shutdown and component failures, we demonstrate that an optimal threshold can be found by minimizing the total system downtime. The proposed method is useful for choosing a shutdown threshold at the design stage of a system.