Abstract: Aimed at improving test accuracy of analog VLSI circuits,an approach based on digital signal processing (DSP) is presented.Sub-band filtering to test response signals is achieved with a filter bank implemented by Cosine modulation.Energy computing and correlation analysis follow sequences acquired from sub-band filtering.Consequently,digital signature extraction to analog response signal is executed.Experiments to 19 faults in an international Benchmark circuit show that energy computation of a sub-band sequence is applicable for diagnosing hard faults,and correlation analysis has high effectiveness for both hard faults and soft faults.It is also shown that the fault resolution of the approach presented in this paper is much higher than the one reported in Ref.7］.

Key words: testing of analog circuits, fault diagnosis, sub-band filtering, Cosine modulation