A new tool for random testability evaluation using simulation and formal proof

A set of tools is described, allowing one to compute random testability measurement for combinational circuits, based on a black box worst case hypothesis. These tools provide enough information to allow circuit modification, in order to meet a prescribed testability value. The efficiency of these tools is due to the use of a statistical method combined with formal proof mechanisms. The random testability of the complete ISCAS benchmark of combinational circuits is computed. For the least testable circuits, a few modifications, guided by the testability measurements, are shown to be sufficient to make them randomly testable.<>