Improved sequential ATPG using functional observation information and new justification methods

Abstract

Sequential ATPG (Automatic Test Pattern Generation) is a very desirable CAD tool, but to date, the site and complexity of circuits for which sequential ATPG could be performed has been limited. We have discovered a method for collecting functional information which makes fault observation significantly easier. We also propose a new method for state justification which is a combination of function-based methods and structure-based methods. Our sequential ATPG system deals with circuits without a reset state or a synchronizing sequence, and the experimental results show that the proposed method achieves significant improvements over existing sequential ATPG methods. >