Cellular automata circuits for built-in self-test

Abstract

Results are presented for a variation on a built-in self-test (BIST) technique based upon a distributed pseudorandom number generator derived from a one-dimensional cellular automata (CA) array. These cellular automata logic block observation (CALBO) circuits provide an alternative to conventional design for testability circuitry such as built-in logic block observation (BILBO) as a direct consequence of reduced cross-correlation between the bit streams which are used as inputs to the logic unit under test. The issue of generating probabilistically weighted test patterns for use in built-in self test is also addressed. The methodology presented considers the suitability of incorporating structures based on cellular automata, a strategy which, in general, improves test pattern quality. Thus, CA-based structures quality as attractive candidates for use in weighted test pattern generator design. The analysis involved in determining and statistically evaluating these potential models is discussed, and is compared with that for previous as well as statistically independent models. Relevant signature analysis properties for elementary one-dimensional cellular automata are also discussed. It is found that cellular automata with cyclic-group rules provide signature analysis properties comparable to those of the linear feedback shift register. The results presented here are based upon simulation.