Vector network analyzer verification algorithms using model with local spectrally selective components

This paper presents verification methods for a calibrated vector network analyzer. The methods are based on a special measurement technique and a model simulating residual errors. Each residual signal has its own unknown frequency properties and known parameters (distance). The properties of the signals are extracted from joint estimation algorithms. Four experiments were performed and verified the proposed methods. Software implemented this methods is also developed.