Authors: Thomas Nirmaier, Infineon Technologies AG, DE; Manuel Harrant, Infineon Technologies AG, DE; Georg Pelz, Infineon Technologies AG, DE
Zusammenfassung:
Abstract
Automotive Power Devices range from low pin-count devices, like power switches, up to complex system control SOCs for the airbag and brake system, that may contain multiple μC cores embedded with multiple sensor and communication interfaces, often all integrated on a single die. A strong trend towards the integration of digital functionality on power electronics devices is visible due to ever increasing monitoring and safety requirements. The combination of analog power electronics and increasing digital functionality leads to an explosion of the verification space that cannot be covered by traditional directed testing anymore.
In order to debug the wide range of mixed-signal functionality we propose to extend Constrained Random Verification methods, which are well known in digital verification.
These methods are based on the principle “do anything the specification allows” and “do anything the application could possibly do”. In this paper we show how to extend constrained-random digital verification to mixed-signal functionality, including test coverage for embedded switches, regulators, transceivers, sensor interfaces and other mixed-signal circuitry. There is no single approach that fits all needs, but several methods need to be applied for mixed-signal instead, including test pattern generation using non-stationary Markov processes to address random-resistance, random-power-up tests, postprocessing on captured digital or analog results from ATE or other equipment. We show several case studies from devices of different complexity, starting from power switches, up to complex airbag control-system devices and how results link to requirements engineering.
7th IEEE International Workshop on Silicon Debug and Diagnosis
Publication Date: 2011/09/22
Location of Publication: International Test Conference (ITC), Anaheim, California, USA
Keywords: Automotive; Analogue/Mixed Signal Design; Verification
