Fault set partition for efficient width compression

In this paper, we present a technique for reducing the test application time of the counter-based pseudo-exhaustive built-in self-testing(BIST). The proposed technique is based on the width compression method and the divide-and-conquer strategy. More formally, we divide the target fault set in k groups(subsets)and apply the width compression method in respect to each fault subset. By selecting k(the number of the groups of the fault set partition), this technique allows a trade-off between test application time and area overhead to be achieved. The synthesis procedure and an example for design of a self-testing test pattern generator using a configurable LFSR is presented. The experimental results show that in the critical cases the proposed technique achieves complete fault coverage of the stuck-at faults in much smaller test application time than the previously published counter-based exhaustive BIST techniques. We study different options to apply the proposed technique based on a test-per-clock(non-scan)BIST scheme. The resultant low-overthead BIST solutions allow at-speed testing and complete fault coverage of the single stuck-at faults to be achieved in reasonable test application time.