Accurate Diagnosis of Cell Internal Defects with Multiple Excitation and Propagation Conditions

Cell-aware fault models are created to model multiple classes of cell-internal defects such as cell-internal shorts and bridges. These cell-aware fault models significantly aid in improving the test quality and diagnosability of the failing Integrated Circuits (IC). Cell-aware fault models are typically created to minimize the number of faults modeled ensuring that all the cell-aware defects are covered with the minimal set of faults created. This is a good approach for ATPG but poses challenges for diagnosis accuracy and resolution. This paper talks about improving accuracy, and resolution of logic diagnostics, by proposing a novel technique to improve the results of logic diagnostics by combining simulation results of multiple cell aware faults to accurately identify the cell aware defect behavior and physical location. Experimental results on industrial design using thousands of failing samples indicated ~14 % accuracy improvement with the proposed approach.