Testing is a critical factor for modern large-scale mixed-mode circuits. Strategies for mitigating test cost and duration include moving significant parts of the test hardware on-chip. This paper presents a novel low-overhead approach for design for test and built-in self-test of analog and mixed-mode blocks, derived from the oscillation-based test framework. The latter is enhanced by the use of complex oscillation regimes, improving fault coverage and enabling forms of parametric or specification-based testing. This technique, initially proposed targeting large subsystems such as A/D converters, is here illustrated at a much finer granularity, considering its application to analog-filter stages, and also proving its suitability to backfit existing designs. The simple case of a switched-capacitor second-order bandpass stage is used for illustration discussing how deviations from nominal gain, central frequency, and quality factor can be detected from measurements not requiring A/D stages. A sample design is validated by simulations run at the layout level, including Monte Carlo analysis and simulations based on random fault injections
Complex oscillation-based test and its application to analog filters
PARESCHI, Fabio;SETTI, Gianluca;
2010
Abstract
Testing is a critical factor for modern large-scale mixed-mode circuits. Strategies for mitigating test cost and duration include moving significant parts of the test hardware on-chip. This paper presents a novel low-overhead approach for design for test and built-in self-test of analog and mixed-mode blocks, derived from the oscillation-based test framework. The latter is enhanced by the use of complex oscillation regimes, improving fault coverage and enabling forms of parametric or specification-based testing. This technique, initially proposed targeting large subsystems such as A/D converters, is here illustrated at a much finer granularity, considering its application to analog-filter stages, and also proving its suitability to backfit existing designs. The simple case of a switched-capacitor second-order bandpass stage is used for illustration discussing how deviations from nominal gain, central frequency, and quality factor can be detected from measurements not requiring A/D stages. A sample design is validated by simulations run at the layout level, including Monte Carlo analysis and simulations based on random fault injectionsI documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
