Features of measurement uncertainty evaluation during calibration of digital ohmmeters

Abstract

The scheme for transferring the size of the unit of resistance during the calibration of digital ohmmeters at direct current is considered. The procedure for the measurement uncertainty evaluation is described: recording the measurement model and its refinement, evaluation of input and measured values, evaluation of standard uncertainties of the input and measured values, evaluation of the expanded uncertainty. The refined model includes the dependence of the resistance of the reference resistor on temperature and a correction to the drift of the resistance value of the reference resistor since its last calibration. To evaluate the expanded uncertainty, the kurtosis method was used. An uncertainty budget has been made, including the kurtosis of input and measured values. The use of the Excel package makes it possible to implement, based on this budget, a program for automation of measurement uncertainty calculations. An example of the measurement uncertainty evaluation during the calibration of a digital ohmmeter of type 2318 at a point of 1 mOhm using an electrical resistance coil R310 with an accuracy class of 0.01 is considered. The influence of nonlinearity of the measurement model on the estimates of the numerical value of the measurand and its combined standard uncertainty is studied. To verify the results, the distribution law of the measurand was modelled by the Monte Carlo method. An algorithm for determining the expanded uncertainty using the NIST Uncertainty Machine web application for the missing confidence level of 0.9545 is proposed. The comparison of the results of the measurement uncertainty evaluation by the kurtosis and Monte Carlo methods has shown their good agreement.