Specifications

Section 3. CR3000 Measurement Details
ranges. Temperature difference measurements made outside of the reference
temperature range should be made by obtaining the actual temperatures
referenced to a junction within the reference temperature range and subtracting
one temperature from the other. Table 3.4-3 gives the reference temperature
ranges covered and the limits of error in the linearizations within these ranges.
Two sources of error arise when the reference temperature is out of range. The
most significant error is in the calculated compensation voltage, however error
is also created in the temperature difference calculated from the thermocouple
output. For example, suppose the reference temperature for a measurement on
a type T thermocouple is 300
o
C. The compensation voltage calculated by the
CR3000 corresponds to a temperature of 272.6
o
C, a -27.4
o
C error. The type
T thermocouple with the measuring junction at 290
o
C and reference at 300
o
C
would output -578.7
µV; using the reference temperature of 272.6
o
C, the
CR3000 calculates a temperature difference of -10.2
o
C, a -0.2
o
C error. The
temperature calculated by the CR3000 would be 262.4
o
C, 27.6
o
C low.
TABLE 3.4-4. Reference Temperature Compensation Range and
Polynomial Error Relative to NIST Standards
TC
Type Range
o
C Limits of Error
o
C
T -100 to 100 ± 0.001
J -150 to 296 ± 0.005
E -150 to 206 ± 0.005
K -50 to 100 ± 0.01
Error Summary
The magnitude of the errors described in the previous sections illustrate that
the greatest sources of error in a thermocouple temperature measurement with
the CR3000 are likely to be due to the limits of error on the thermocouple wire
and in the reference temperature. Errors in the thermocouple and reference
temperature linearizations are extremely small, and error in the voltage
measurement is negligible.
To illustrate the relative magnitude of these errors in the environmental range,
we will take a worst case situation where all errors are maximum and additive.
A temperature of 45
o
C is measured with a type T (copper-constantan)
thermocouple, using the
±20 mV range. The nominal accuracy on this range is
1 µV (0.01% of 10 mV) which at 45
o
C changes the temperature by 0.012
o
C.
The reference thermistor is 20
o
C but is indicating 20.15
o
C, and the terminal
that the thermocouple is connected to is 0.05
o
C cooler than the reference
thermistor.
3-15