The evaluation of the uncertainty associated with comparison loss in microwave power meter calibration

A model of comparison loss in microwave power meter calibration is considered in the case when the standard power meter used for calibration and the signal generator to which this meter and the meter to be calibrated are connected are reflectionless. The uncertainty associated with an estimate of the model output quantity, viz., the ratio of the power absorbed by the meter being calibrated and that absorbed by the standard, is to be evaluated given estimates of the model input quantities, the standard uncertainties associated with those estimates, and a correlation coefficient relating to the estimates. This problem is addressed using three approaches, viz.: (1) the GUM uncertainty framework; (2) the propagation of distributions, implemented using a Monte Carlo method; and (3) analytical or semi-analytical. Circumstances in which approach: (1) can provide invalid solutions, whereas (2) provides sound numerical solutions, are identified