Optimized synthesis of self-testable finite state machines

Abstract

A synthesis procedure for self-testable finite state machines is presented. Testability comes under consideration when the behavioral description of the circuit is being transformed into a structural description. To this end, a novel state encoding algorithm, as well as a modified self-test architecture, is developed. Experimental results show that this approach leads to a significant reduction of hardware overhead. Self-testing circuits generally employ linear feedback shift registers for pattern generation. The impact of choosing a particular feedback polynomial on the state encoding is discussed.