Chapter 2
SC-2070 Board
© National Instruments Corporation
2-23
SC-207X Series User Manual
Sources of Error
When making thermocouple measurements with the SC-2070 board and an MIO-16 board, the possible
sources of error are compensation, linearization, measurement, and thermocouple wire errors.
Compensation error can arise from two sources: inaccuracy of the temperature sensor and temperature
differences between the sensor and the screw terminals. The sensor used on the SC-2070 board is
specified to be accurate to
±
1
°
C. Minimize temperature differences between the sensor and the screw
terminals by keeping the SC-2070 board away from drafts, heaters, and warm equipment.
Linearization error is a consequence of the polynomials being approximations of the true thermocouple
output. The linearization error is dependent on the degree of polynomial used. Table 2-4 lists the
linearization errors for the NBS polynomials.
Measurement error is the result of inaccuracies in the MIO-16. These inaccuracies include gain and
offset. If the board is properly calibrated, the offset error should be zeroed out. The only remaining
error is a gain error of
±
0.08% of full range (see the MIO-16 specifications). If the input range is
±
10 V and the gain is 500, gain error contributes 0.0008 x 20 mV, or 16
µ
V of error. If the Seebeck
coefficient of a thermocouple is 32
µ
V/
°
C, this measurement error adds 0.5
°
C of uncertainty to the
measurement. For best results, use a well-calibrated MIO-16 board so that offsets can be neglected.
Offset error can be eliminated, however, by grounding one channel on the SC-2070 board and
measuring the voltage. This value, the offset of the MIO-16, can then be subtracted in software from
all other readings.
Thermocouple wire error is the result of inconsistencies in the thermocouple manufacturing process.
These inconsistencies, or nonhomogeneities, are the result of defects or impurities in the thermocouple
wire. The errors vary widely depending on the thermocouple type and even the wire gauge used, but a
value of
±
2
°
C is typical.
For best results, use an average of about 100 readings. When using an average, typical accuracies are
about
±
2
°
C.
Thermocouple Measurement Accuracies
Table 2-5 lists the expected thermocouple accuracies in degrees Celsius, subject to the following
conditions:
•
The MIO-16 board is correctly calibrated.
•
The temperature of the screw terminals equals the temperature of the board (no gradients on the
board).
The uncertainties listed apply at either 0
°
C (Type J, K, E, T) or 600
°
C (Type S, R). The linearization
errors of the NBS polynomials in Table 2-4 , measurement error of a calibrated MIO-16, and the
±
1
°
C
cold-junction sensor error are included. Thermocouple wire error is neglected because of a
dependence on the conditions listed previously. Finally, these uncertainties are for the gains listed with
a
±
10 V input range.
Summary of Contents for SC-207 Series
Page 15: ......
