Special applications of the voting model for bridging faults

It is shown that the most accurate bridging fault model, the voting model, could be generalized. In particular, details on how to handle the Byzantine general´s problem, IC fabrication process variations, complex gates, and transistor size variations are presented. An example using BDDs to represent the voting model and to compare it to other fault models is presented. The relationship between bridging faults and delay faults is discussed, demonstrating that delay tests make poor bridging fault tests. It is concluded that the voting model is a logic-level model which accurately models real faults in real circuits