Analysis of Specified Bit Handling Capability of Combinational Expander Networks

Test compression schemes based on combinational expander networks have become very popular in recent times. The idea behind these schemes is to use m bits from the tester to produce N(m < N) bits for the internal scan chains of the circuit under test. In this paper we address the general problem of designing combinational expander networks with N outputs which guarantee that any S specified bits can be justified at the expander output. By analyzing the constraints imposed on the output space of such a network we derive formulae that provide the minimum value of m (and consequently a maximum value of the amount of compression that can be achieved). We then show that a subclass of one of the state-of-the-art combinational expander designs (XPAND) being currently used in several industrial designs achieves the maximum amount of compression possible.