We have experienced a failure during life testing of an electronic control unit, 10 units were continuously powered and monitored at a fixed ambient temperature of 95degC. After 8.5 weeks of a 10 week test one unit exhibited a failure which we believe is attributable to a transient short circuit within the 8 layer printed circuit board. The PCB has 95V rails in specific areas and the failure appears to have been caused by high voltage getting onto a signal line. So far we have been unable to locate the failure site as the leakage path disappeared during the course of investigation. We suspect a conductive filament type mechanism, (CFF), but are told this is normally associated with formation of copper dendrites under high humidity conditions. Has anyone experienced failures under similar conditions ? Are there any suggestions on the failure mechanism ? What corrective actions are applicable ? Best Regards Stephen Lane Lucas Automotive Electronics [log in to unmask]