SET: Substrate Extraction Tool for Predicting Substrate Noise in SOI Smart Power ICs

Substrate noise propagation is a critical issue in smart power SOI design, as the switching noise generated by power circuits can couple with sensitive analog circuitry causing malfunctions. This work presents SET, a lightweight tool for extracting SPICE-like netlists to provide a rapid estimation of aggressor-victim coupling through the substrate. By implementing a local mesh refinement, SET reduces the simulation time while preserving accuracy. Validation on elementary test structures shows good agreement with commercial tools in terms of spreading resistance and substrate coupling, making SET a practical solution for early-stage architecture exploration and design definition.