Lowest uncertainty contributions in temperature calibrations by comparison

The paper deals with improvement and achievement of the lowest uncertainty contribution in calibration of thermometers by comparison on a basis of evaluation of high precision calibration baths encountering influential metrological parameters, which are typically not evaluated. By this method practical limits of calibration methods by comparisons are achieved. The efforts are sensible, if all error sources in the entire system are of a matching magnitude. A typical set-up for calibration of thermometers by comparison consists of a calibration bath or a furnace with a calibration medium, a reference thermometer, a measuring device and a thermometer under calibration. According to the type of a reference thermometer used and a thermometer being calibrated there are some typical and some particular uncertainty contributions. These contributions involve all the uncertainties, which occur during the calibration process. In general there are four main uncertainty contributions because of repeatability, inhomogeneity and instability of a calibration bath or a furnace, reference thermometer and a measuring device. Especially an uncertainty of a calibration bath or a furnace represents a large contribution in a total uncertainty calculation of a calibrated thermometer.