The use of alternate tests in addition to specificationbased measurements is achieving more recognition in industry due to the higher coverage that they provide. The fault and yield coverages of these tests depend on how the pass/fail test decision is made. In this paper, we address the critical issue of accurate test threshold determination for these alternate tests. We propose to post-process the given set of sensitive and linearly independent measurements to synthesize a new set of measurements based on which the pass/fail decision is made. A novel methodology for post processing the measurement results called measurement synthesis is presented. Simulation results show that test effectiveness can be greatly enhanced by measurement synthesis.
IntroductionTraditionally, analog circuits are tested for parametric variations by verifying the functional specifications of the circuit. In the past, there has been a lot of effort towards reducing the specification testing time by eliminating unnecessary specification tests and ordering them in an optimal way [1-4]. Nevertheless, specification-based testing is still time consuming and expensive. It is reported in [5] that in mixed-signal circuits, 95% of the test cost is expended in testing the analog parts, while the digital counterparts account for only 5% of the overall test cost. In the recent past, researchers have proposed various methods for designing alternate tests for detecting catastrophic and parametric failures in analog circuits. These alternate testing schemes can be broadly classified into DC (static), AC (steady state) and transient (dynamic) testing. Below, we summarize recent work in alternate test design.In [6], the authors propose adding extra tests during wafer probe so that many obviously faulty circuits can be eliminated from the production line at an early stage. Marlett and Abraham [7], and Devarayanadurg and Soma [8] give algorithms for computing DC tests for detecting catastrophic faults. However, parametric faults caused by global process variations like mask misalignment and line width variations, resulting in degradation of circuit performance, are hard to detect using static DC tests. Nagi et. al. in [9], present an algorithm for test frequency selection for AC test using behavioral level fault models and fast fault simulation. In [10], the authors present a test generation algorithm for detecting single and multiple faults based on circuit sensitivity computation. In [11], this is further extended to generate test frequencies for catastrophic faults. However, the catastrophic faults considered are limited to opens and shorts in the passive components. In [12], the authors present an automated tool for sensitivity analysis and test generation called LIMSoft. In this paper they generalize the test generation algorithm presented in [11] to include shorts between any two arbitrary nodes in the circuit under test (CUT). In [13], Abderrahman et. al. solve the test frequency selection problem by formulating it as a minimax optimization problem....