Covariance-Matrix-Based Uncertainty Analysis for NVNA Measurements

This paper studies the systematic uncertainties of nonlinear vector network analyzer (NVNA) measurements and presents a covariance-matrix-based method to analyze the uncertainty propagation from calibration standards and hardware to error model terms and effective measurements. Analytic deductions of the Jacobians for frequency-domain quantities are proposed, dealing with the complex NVNA measurements and calibration functions. An NVNA two-port-calibration uncertainty is put forward as a new kind of systematic uncertainty consideration for this vector-correction-based instrument. Experiments on our prototype NVNA show the necessity of using covariance-matrix-based uncertainty analysis, particularly for polyharmonic measurements, and reveal some features of NVNA uncertainty analysis due to the calibration techniques.