Ensuring the quality of a circuit implies ensuring the quality of test. Despite the fact that performance-based testing has been the golden standard for Analog, Mixed-Signal and RF test for decades, high-reliability markets like automotive have found that functional test leaves some potential defects undetected that can produce in-field failure. There is thus a push towards defect-oriented testing which, in turn, calls for an efficient defect simulation framework. This paper presents a statistical adaptive defect simulation based on likelihood-weighted random sampling to evaluate the quality of AMS-RF tests in terms of defect coverage and fault escape. The adaptive loop takes a decision at each new defect simulation on whether it is more efficient to assess the defect coverage or the fault escape rate of the test under evaluation, as a function of the desired targets for these two metrics. Several decision criteria are proposed and validated by simulation of a complete IC for different tests.

确保电路质量意味着确保测试质量。尽管几十年来基于性能的测试一直是模拟，混合信号和RF测试的黄金标准，但汽车等高可靠性市场发现，功能测试遗漏了一些潜在的缺陷，可能导致现场故障。因此，存在着对面向缺陷的测试的需求，这反过来又需要一种有效的缺陷仿真框架。本文提出了一种基于似然加权随机抽样的统计自适应缺陷仿真，以从缺陷覆盖率和故障逃逸角度评估AMS-RF测试的质量。自适应回路会在每次新的缺陷模拟时做出决策，以评估评估覆盖范围或评估测试的故障逃逸率是否更有效，根据这两个指标的预期目标。提出了几种决策标准，并通过针对不同测试的完整IC的仿真进行了验证。