W e propose a coverage metric and a two-pass test generation method f o r path delay faults in combinational logic circuits. The coverage is measured f o r each line with a rising and a falling transition. However, the test criterion is different from that of the slow-to-rise and slow-to-fallIn this paper, we define a rising line delay test that sensitizes the longest sensitizable path passing through the target line producing a rising transition on it. Similarly, a falling line delay test is defined. The definition of "longest" can be appropriately chosen. For example, in the simplest case, it can be the path with largest number of gates. Alternatively, gates can be weighted by their nominal delays. However, once the path is selected, the test generation is independent of gate delays. The criterion of delay test through the longest path has been used for diagnosis [5].The coverage is measured for all lines with two possible transitions. Thus, the maximum number of faults (or tests) is twice the number of lines. Yet, test criterion is similar to path delay fault, and not like gate or transition delay fault. In general, a test will cover several lines. This coverage methodology can also be applied to the reported methods that extract sensitizable paths [2, 31.An iterative approach for generating a robust test was first proposed by Park and Mercer [SI. They devised an approzzmate method where the search space of test generation process is biased to find a test along a path whose propagation delay is greater than or equal to a predefined threshold value. Our approach is to use an exact method for generating a test for the longest robustly testable path through each line.
Two-Pass Test GenerationFinding the longest sensitizable and robustly testable path through a given delay fault site is an NP-hard problem [8]. We first attempt to find a robust test for the longest structural path through a line. If the path is not sensitizable, then we try to find a robust test for the next longest structural path, and so forth, until a test for the longest sensitizable path is found. Given enough resources this method guarantees a test for the longest sensitizable path through the line if such a test exists.The first pass of our two-pass test generation strat-