Cost-driven statistical analysis for selection of alternative measurements of analog circuits

In this paper, we introduce a new method for improving the use of Alternative Measurements Strategy for analog circuit testing. The goal is to reduce the test cost by selecting the cheapest set of measurements that do not bring redundant information about the state of the circuit, without reducing test coverage. The proposed method consists in finding a subset of a given set of parameters which explains the dispersion of the circuit performances. The efficiency of the proposed approach is demonstrated on a critical example using continuous-time A/D conversion.