The Method of Instantaneous Comparisons Applied to the Calibration of SPRTs in Liquid Nitrogen and Liquid Argon

Abstract

Temperature drift is one of the most significant sources of uncertainty in the calibration of thermometers by comparison. Measuring the resistances, \(R_{1}\), etc., by ratio to the resistance of a (similar) reference thermometer, \(R_\mathrm{ref}\), rather than to a fixed standard resistor, \(R_{\mathrm{s}}\), allows the effect of drift to be overcome. The ratios, \(R_{1}/R_\mathrm{ref}\), etc., represent “instantaneous comparisons” of the two thermometers. They are practically independent of temperature, so the exact temperature of the comparison (and the actual value of \(R_\mathrm{ref}\) itself) is not of first importance. The key points are that the \(R_{1}/R_\mathrm{ref}\), etc. are measured accurately, and that \(R_\mathrm{ref}\) is accurately related to temperature. This paper describes the method and gives results of its application to the comparison of standard platinum resistance thermometers in liquid nitrogen and liquid argon, with repeatabilities of less than or about 0.1 mK. The method is not new, but it appears not to have been widely used or discussed in the literature. It has potential for application in all comparisons of like thermometers, in liquid baths, metal blocks, and furnaces, as well as in testing the temperature uniformity of the baths, etc., and investigations such as into thermometer stability or hysteresis. Some comparisons in conditions of rapid drift, where thermometer response times become important, are also presented.