A novel behavioral testability measure

In this paper a new approach, called BETA, for computing testability is presented. The approach is based on analyzing the circuit´s behavioral description: the control flow graph (CFG). Based on path analysis, testability measures are derived. Unlike a traditional testability measure which computes testability, our approach also derives the exact sequence for justifying and propagating the contents of each variable when possible. Variable classification is used to diagnose every register´s controllability and observability and to classify them into several groups. Based upon this classification, guidance is provided to the test generator for each group. Variable classification is also useful in pointing out hard-to-control and hard-to-observe areas of a circuit. For fully controllable and observable variables, the best path sequence to justify and propagate these registers is generated. This approach has been implemented in a computer program and applied to several examples. All these results are verified by a CFG-based test generator and proven to be successful