On effective criterion of path selection for delay testing

Since a logic circuit often has too many paths to test delay all paths in the circuit, it is necessary for path delay testing to limit the number of paths to be tested. Paths to be tested should be ones with large delay that are more likely to cause a fault. In addition, a test set for the paths are required to detect as many as possible other models of faults. In this paper, we investigate criteria of path selection for path delay testing. We first define two typical criteria to be investigated here, and then experimentally show the features of paths selected with each criterion, with respect to fault coverage of other delay fault models. From our experiments, we observe that test patterns for the longest paths cannot cover many other faults such as gate delay faults or segment delay faults.