An optimal vector-network-analyzer calibration algorithm

We present an iterative algorithm for calibrating vector network analyzers based on orthogonal distance regression. The algorithm features a robust, yet efficient, search algorithm, an error analysis that includes both random and systematic errors, a full covariance matrix relating calibration and measurement errors, 95% coverage factors, and an easy-to-use user interface that supports a wide variety of calibration standards. We also discuss evidence that the algorithm outperforms the MultiCal software package in the presence of measurement errors and accurately estimates the uncertainty of its results.