< back to main site

Publications

Strategies for minimising the uncertainty of the SPRT self-heating correction.

Veltcheva, R I; Pearce, J V; da Silva, R*; Machin, G; Rusby, R L (2013) Strategies for minimising the uncertainty of the SPRT self-heating correction. AIP Conf. Proc., 1552. pp. 433-438.

Full text not available from this repository.

Abstract

Self-heating of standard platinum resistance thermometers (SPRTs) can be problematic for measurements which require the very highest accuracy. The convention is to correct for the effect by taking measurements at two different measuring currents and then performing a linear extrapolation (in power) to obtain the resistance at zero current. Following Batagelj et al [1], various methods for improving the extrapolation are here explored by: a) optimising the choice of the two currents which should be used, and b) investigating whether measuring at more than two currents would reduce the uncertainty of the extrapolation, and c) if so, how many and what extrapolation scheme should be used. We present a series of measurements to examine these questions using three SPRTs in three environments, namely in cells at the triple points of mercury and water, and the freezing point of zinc. The standard deviations of the measurements at the 10 available measuring currents, for each of the 3 SPRTs and 3 fixed points, were used to quantify the uncertainty of the extrapolations to zero current for all the pairs of currents. Extrapolation schemes using more than two currents have been investigated using a Monte Carlo method. It is found that however many currents are used, they should be approximately equally spaced in current, not power.

Item Type: Article
Keywords: SPRT, platinum resistance thermometer, zero current, self-heating
Subjects: Engineering Measurements
Engineering Measurements > Thermal
Identification number/DOI: 10.1063/1.4821390
Last Modified: 02 Feb 2018 13:14
URI: http://eprintspublications.npl.co.uk/id/eprint/5954

Actions (login required)

View Item View Item