Comments on: Calibrating a platinum resistance thermometer at a fixed point

By: Hans Liedberg

Hans Liedberg — Thu, 30 Jul 2015 15:53:12 +0000

Thanks for the valuable comments! Yes, the measurement at the lower current is more significant, as it contributes more to the combined uncertainty. If you report zero-power resistance, R0, the formula is R0 = R1 – (R2 – R1)*i1^2/(i2^2 – i1^2). The sensitivity to uncertainty in R1 is dR0/dR1 = 1 + i1^2/(i2^2 – i1^2). For i2 = 1.41*i1, this yields dR0/dR1 = 2, or, for i2 = 2*i1, dR0/dR1 = 4/3. dR0/dR2 = i1^2/(i2^2 – i1^2), so, dR0/dR2 = 1 for i2 = 1.41*i1 and dR0/dR2 = 1/3 for i2 = 2*i1. In other words, R1 is twice or four times more important, thereby warranting the greater number of repeated measurements.

By: Edward Kubicki

Edward Kubicki — Tue, 18 Nov 2014 00:40:22 +0000

Have you any comments to make on the following article “The optimization of self-heating corrections in resistance thermometry” Metrologia, 50, page 345-353, (2013) -by J.V. Pearce et al. ? – which basically states that the best method (with 25 ohm SPRT) to extrapolate to zero current is to test at only 2 currents (any more is wasteful) of I(1.0mA) & I/2(0.5mA) and spend 8X the time measuring at the lower current(0.5mA) compared with the higher current(1.0mA).
Your article – 6. Freeze :c) Self-heating of the PRT – uses the commonly accepted method of I & I/1.4142.