Over-determined calibration schemes for RF network analysers employing generalised distance regression

This paper presents over-determined calibration schemes for vector network analysers (i.e. schemes involving more than the traditional three calibration standards). Generalised distance regression is used to obtain estimates for the three complex-valued calibration coefficients of the network analyser from the indicated and 'true' complex-valued reflection coefficients of the standards. Account is taken of the uncertainties associated with the indicated and 'true' reflection coefficients of the standards to establish the uncertainty in the calibration coefficients and subsequently the uncertainty in the reflection coefficient measurements made using the calibrated network analyser. Typical results are presented to demonstrate the performance of these new calibration schemes compared with more traditional methods.