Measurement Uncertainty Propagation in Transistor Model Parameters via Polynomial Chaos Expansion

We present an analysis of the propagation of measurement uncertainty in microwave transistor nonlinear models. As a case study, we focus on residual calibration uncertainty and its effect on modeled nonlinear capacitances extracted from small-signal microwave measurements. We evaluate the uncertainty by means of the polynomial chaos expansion (PCE) method and compare the results with the NIST Microwave Uncertainty Framework, which enables both sensitivity and Monte Carlo (MC) analyses for uncertainty quantification in microwave measurements. We demonstrate that, for the considered application, PCE provides results in agreement with classical MC simulations but with a significant reduction of the computational effort.