Methodology for connector reliability analysis

Connector suppliers as well as end users in the industry today lack a unified approach to environmental testing. In addition, the criterion on which a connector failure is based appears to be arbitrary. Moreover, the typical sample sizes used do not provide for an adequate prediction of the low failure rates required for high-end systems. This paper describes the rationale and methodology in which to perform a system level reliability analysis for a connector system. IBM Rochester uses four main environmental sequences to stress connectors. Each stress sequence is designed to create failures based on a specific set of failure mechanisms. Also, each stress has its own acceleration factors, which can be tailored to the specific application.