A generalised approach to the propagation of uncertainty in complex S-parameter measurements

This paper presents a generalised method for evaluating the effects of the uncertainty in complex S-parameter measurements on other, related, measurement quantities. The method utilises random numbers to simulate distributions for the measured S-parameters. These distributions are then passed through the measurement model to establish distributions in the output quantities (i.e. the measurands). Uncertainty estimates for the measurands are then obtained from the output distributions. This method finds particular application where the measurement model exhibits significant nonlinearity and/or is too complicated to allow more conventional approaches to be applied. Examples of two such instances are given in the paper.