SPRT to make sure it is accurate. One way to do this is to compare its
measurements to those of a second SPRT.
7.
Check system accuracy with a fixed-point cell.
An excellent way to
verify the accuracy of the entire system, including the 1575/1590, refer-
ence resistors, and SPRT, is to check measurements using a fixed-point
cell. The most accurate, useful, and economical fixed point is the triple
point of water. Other fixed points such as the gallium melting point, mer-
cury triple point, and tin freezing point can also be used for verification.
8.
Check accuracy after calibrating a PRT.
After calibrating a PRT com-
pare measurements using the PRT and its new characterization coeffi-
cients with your reference SPRT at each calibration point.
9.
Be aware of other problems that can affect accuracy.
Be careful of
other potential problems that can affect accuracy. When using baths
make sure probes are immersed adequately to prevent stem conduction
error. When comparing or calibrating one probe against another it is im-
portant that the bath have good temperature stability and uniformity.
Make sure probe wires have good low-resistance electrical connections.
Avoid excessive electromagnetic interference. Keep leads as short as
possible and use shielding if necessary. Prevent abrupt or extreme
changes in ambient temperature around the 1575/1590 and wire connec-
tions. Operate the 1575/1590 in a stable environment free from signifi-
cant drafts.
6.2.2
Evaluating Temperature Uncertainty
The 1575/1590 Thermometer Readout is often used for calibrating PRTs and in
other applications where precision temperature measurements are critical. It is
important to know how much uncertainty the 1575/1590 contributes to total
system uncertainty. The specification section of this manual (Section 3.1) pro-
vides typical temperature uncertainty figures for some particular conditions.
The actual uncertainty of your measurements may differ. A significant factor
that affects uncertainty is the sensitivity of the probe. The greater the sensitiv-
ity, the lower the uncertainty. For SPRTs sensitivity generally decreases as tem-
perature increases. As a result, uncertainty will be greater at higher
temperatures. To determine the exact uncertainty at a given temperature take
the uncertainty of the resistance measurement (ohms) and divide by the slope
of the resistance-temperature curve (ohms per degree) at that temperature. As
an example, suppose the 1590 and a 25.5
Ω
SPRT are used to measure a tem-
perature of 419°C. The SPRT’s resistance at this temperature is about 65.45
Ω
and its slope is 0.08916
Ω
/°C. The resistance measurement uncertainty is 6
ppm x 65.45
Ω
= 0.00039
Ω
so the temperature uncertainty of the 1590 is
0.00039
Ω
/ 0.08916
Ω
/°C = 0.0044°C.
Realize that this is only the contribution of the 1590 to total uncertainty. The
SPRT or sensor contributes additional uncertainty. External reference resistors,
if used, must also be considered. With comparison calibrations using baths or
dry-wells you must take into account uncertainty caused by temperature varia-
1575/1590 Thermometer Readout
User’s Guide
32
