Corrections for nonlinear vector network analyzer measurements using a stochastic multi-line/reflect method

A new 16-term statistical calibration has been developed for the correction of the vector network analyzer (VNA) data. The method uses multiple measurements of generic transmission line and reflection standards. Using a functional model of the system and transmission line standards, we apply a nonlinear least-squares estimator to simultaneously optimize the correction terms in the measurement model and the propagation constant. The method provide estimates of the uncertainty on each of the parameters using the final Jacobian. This paper shows for the first time an application of the new calibration to a commercial nonlinear VNA, plus quantitative statements regarding the quality of the parameters.