11
Introduction and Specications
Specications
is converted to temperature by dividing 0.03 ppm by 1.0 × 106 and then multiplying by 1.02. The result is then
divided by WT90 sensitivity, dW/dT, at 0.01 °C, which is 0.004. The nal result is 0.000008 °C. After multiply-
ing by WT90 (1.612), the uncertainty of the RTPW resistance ratio, when applied to 157 °C, is 0.000013 °C.
2.2.5.3.5 Measurement Noise at 0.01 °C
During measurement of the RTPW, it is observed that the standard error of the mean is 0.0000018 Ω. To
convert this value into temperature, divide by the resistance sensitivity (dR/dT) of the SPRT at 0.01 °C. dR/dT
at 0.01 °C is 0.1 W/°C (see tip below). The result is 0.000018 °C. Multiplying by WT90 (1.612) yields 0.000029
°C.
Tip: Most SPRT calibration reports list the Temperature versus W values of the SPRT in a table.
Typically, the dT/dW value at each temperature will be included in the same table. dR/dT can be
calculated by inverting dT/dW and multiplying by the RTPW of the SPRT. Also, dW/dT can be calculated
by simply inverting dT/dW.
2.2.5.3.6 Uncertainty of the TPW Cell
Uncertainty of the temperature of the triple-point of water cell must also be included. For this example, the
standard uncertainty of the triple-point of water cell is 0.000050 °C. Multiplying by WT90 (1.612) results in
0.000081 °C.
2.2.5.3.7 Reference Resistor Drift
Possible drift of the 25 W reference resistor between the TPW measurement and the 157 °C measurement must
be accounted for. To reduce the possible error, both measurements should be taken in close proximity in time.
For this example, the 24-hour stability specication will be used. This requires that both measurements are
taken within the same 24-hour period. The standard uncertainty due to drift of the reference resistor is 0.125
ppm. This is converted to temperature by dividing 0.125 ppm by 1.0 × 106 and then multiplying by 1.612. The
result is then divided by WT90 sensitivity (dW/dT) at 157 °C which is 0.0038. The result is 0.000053 °C.
2.2.5.3.8 Combining the Uncertainties
At this point, all of the uncertainties can be combined by root-sum-square (RSS) since they are uncorrelated.
Even though the 1595A resistance ratio accuracy is used twice in the calculation, both measurements are con-
sidered uncorrelated.
This RSS sum produces a combined standard uncertainty of 0.000115 °C. Multiplying by the coverage factor
(k = 2) results in a total expanded uncertainty of 0.000230 °C.
In this example, it is assumed the SPRT did not drift between the measurements at the TPW and 157 °C. It
may be necessary to add an additional uncertainty that accounts for SPRT drift.
2.2.5.3.9 Measuring With Calibration Report RTPW
In this example, the SPRT is monitored by periodically measuring its RTPW in a TPW cell but the original
RTPW from the SPRT calibration report is entered in the 1595A, not the measured RTPW. This requires a dif-
ferent set of specications to be used in the measurement uncertainty.
The uncertainty of this measurement is based on four uncorrelated uncertainties. These uncertainties are:
Resistance accuracy of the 1595A when measuring the SPRT at 157 °C
Measurement noise at 157 °C
Uncertainty of the SPRT resistance at the triple-point of water
Drift of the RTPW of the SPRT
2.2.5.3.10 Resistance Accuracy at 157 °C
The resistance of the SPRT at 157 °C is 41.1 Ω. Using the 25 Ω internal reference resistor, the 1595A one-
year resistance standard uncertainty is 2.5 ppm. This uncertainty, in terms of temperature, is calculated by rst
dividing 2.5 ppm by 1.0 × 106 then multiplying by 41.1 Ω. The result is then divided by the sensitivity (dR/dT)