On-chip test embedding for multi-weighted random LFSRs

We present for the first time a systematic approach for partitioning deterministic test set into subsets so that multiple weight-sets, one weight-set per subset, are generated for efficient Weighted Random LFSR Test Pattern Generation. The basic partitioning criterion is the maximum Hamming distance between any two test patterns in the same set. The number of test patterns within each subset is also taken into consideration. The proposed tools make use of optimal partitioning algorithms. Experimental results clearly indicate the effectiveness of the proposed scheme.