On methods to match a test pattern generator to acircuit-under-test

摘要

Autonomous circuits such as linear feedback shift registersn(LFSRs) and cellular automats are used as low-cost test patternngenerators for circuits testable by pseudo-random patterns. Wendemonstrate that different LFSRs of the same degree, started fromndifferent initial states, may yield significantly different faultncoverages and test lengths when used as test pattern generators for angiven circuit, especially when the circuit has faults which are hard tondetect by a practical number of pseudo-random patterns. Methods tontailor an LFSR to a circuit-under-test are proposed, that attempt tonselect the most effective LFSR and initial state for the circuit. Thenfirst method is based on a learning process that can be applied directlynto certain types of circuits. The learning process is also used tonestablish a collection of (primitive and nonprimitive) LFSRs and initialnstates, effective for arbitrary circuits. This collection can then benused as a starting point for a genetic optimization procedure aimed atnimproving the selected LFSR and initial state. The use of an LFSR thatncan apply complemented as well as uncomplemented test patterns is shownnto significantly improve the fault coverage, at the cost of a small areanoverhead. Experimental results demonstrate the applicability of thenproposed approaches to stuck-at faults and to transition faults