Indirect and adaptive test of analogue circuits based on preselected steady-state response measures

Abstract

Alternate testing techniques have been progressively adopted as a promising solution due to their effectiveness against classical specification-based test methods. This work presents a built-in test system, which adaptively generates an indirect digital signature characterising the circuit under test, which is later used to diagnose the actual performances relying on a statistical dictionary-based diagnosis method. The system is composed of an integrated digital signature generator and a digital control and acquisition subsystem. The signature generator is based on a converter architecture, in which the analogue range can be adapted to the magnitude of the indirect measure using the well known information of the fault-free circuit. The digital subsystem controls the proposed architecture and stores the digital codes sent by the integrated digital signature generator. The digital signature generator has been designed and fabricated in an industrial 65 nm complementary metal- oxide-semiconductor technology from STMicroelectronics, whereas the digital control and acquisition subsystem have been prototyped in a field-programmable gate array. The fabricated system has been used to test a biquad filter affected by parametric variations. Successful experimental results are reported showing the capabilities of the proposed adaptive test system to diagnose circuit performances with discrepancies as low as 1% of the actual real value.