Campbell Scientific CR1000 Measurement and Control System - Manual (Page 310)

Section 8. Operation
310
Table 66. Limits of Error on CR1000 Thermocouple Polynomials
TC
Type Range °C
Limits of Error °C
Relative to NIST
Standards
-130 to 200 ±0.005
200 to 1000 ±0.02
K -50 to 1372
-50 to 950 ±0.01
950 to 1372 ±0.04
8.1.4.1.7 Reference-Junction Error
Thermocouple instructions TCDiff() and TCSe() include the parameter TRef to
incorporate the reference-junction temperature into the measurement. A reference-
junction compensation voltage is computed from
TRef as part of the thermocouple
instruction, based on the temperature difference between the reference junction
and 0°C. The polynomials used to determine the reference-junction compensation
voltage do not cover the entire thermocouple range, as illustrated in tables Limits
of Error on CR1000 Thermocouple Polynomials
(p. 309)
and Reference-
Temperature Compensation Range and Polynomial Error
(p. 310).
Substantial errors
in the reference junction compensation voltage will result if the reference-junction
temperature is outside of the polynomial-fit ranges given.
The reference-junction temperature measurement can come from a PanelTemp()
instruction or from any other temperature measurement of the reference junction.
The standard and extended (-XT) operating ranges for the CR1000 are -25 to
50°C and -55 to 85°C, respectively. These ranges also apply to the reference-
junction temperature measurement using PanelTemp().
Two sources of error arise when the reference temperature is out of the
polynomial-fit range. The most significant error is in the calculated compensation
voltage; however, a small error is also created by non-linearities in the Seebeck
coefficient.
Table 67. Reference-Temperature Compensation Range and Error
TC Type Range °C Limits of Error °C
1
T -100 to 100 ± 0.001
E -150 to 206 ± 0.005
J -150 to 296 ± 0.005
K -50 to 100 ± 0.01
1 Relative to ITS-90 Standard in NIST Monograph 175
8.1.4.1.8 Thermocouple Error Summary
Errors in the thermocouple- and reference-temperature linearizations are
extremely small, and error in the voltage measurement is negligible.