E1C08 09/14/2010
14:54:1 Page 347
Data-Acquisition Considerations
Measuring temperatures by using thermocouples connected to a data acquisition systems is common
practice. However, the characteristics of the thermocouple, including the need for a reference or cold
junction and the low signal voltages produced, complicate its use. Nevertheless, with a little
attention and realistic expectation of achievable accuracy, the systems are quite acceptable for most
monitoring and moderate accuracy measurements.
Once the appropriate thermocouple type is selected, attention must be given to the cold junction
compensation method. The two connection points of the thermocouple to the data acquisition
system (DAS) board form two new thermocouple connections. Use of external cold junction
methods between the thermocouple and the board eliminates this problem, but more frequently the
thermocouple is connected directly to the board and uses built-in electronic cold junction
compensation. This is usually accomplished by using a separate thermistor sensor, which measures
the temperature at the system connection point to determine the cold junction error, and providing an
appropriate bias voltage correction either directly or through software. An important consideration
is that the internal correction method has a typical uncertainty of the order of 0.5
to 1.5
C and, as a
systematic error, this error is directly passed on to the measurement as an offset.
These boards also may use internal polynomial interpolation for converting measured voltage
into temperature. If not, this can be programmed into the data-reduction scheme by using, for
example, the information of Table 8.7. Nonetheless, this introduces a ‘‘linearization’’ error, which is
a function of thermocouple material and temperature range and typically specified with the DAS
board.
Thermocouples are often used in harsh, industrial environments with significant electro-magnetic
interference (EMI) and radiofrequency (rf) noise sources. Thermocouple wire pairs should be twisted
to reduce noise. Also, a differential-ended connection is preferred between the thermocouple and the
DAS board. In this arrangement, though, the thermocouple becomes an isolated voltage source,
meaning there is no longer a direct ground path keeping the input within its common mode range. As a
consequence, a common complaint is that the measured signal may drift or suddenly jump in the level
of its output. Usually, this interference behavior is eliminated by placing a 10- to 100-kV resistor
between the low terminal of the input and low-level ground.
Cu
Cu
Measuring
junctions
Reference
junctions
1
2
Potentiometer
Copper connecting
wires
Thermocouple
wires
A
A
N
A
B
B
B
Figure 8.24 Parallel arrangement of thermocouples for sensing the average temperature of the measuring
junctions. (From Benedict, R. P. Fundamentals of Temperature, Pressure and Flow Measurements, 3rd ed.
Copyright # 1984 by John Wiley and Sons, New York. Reprinted by permission.)
8.5 Thermoelectric Temperature Measurement 347
14:54:1 Page 347
Data-Acquisition Considerations
Measuring temperatures by using thermocouples connected to a data acquisition systems is common
practice. However, the characteristics of the thermocouple, including the need for a reference or cold
junction and the low signal voltages produced, complicate its use. Nevertheless, with a little
attention and realistic expectation of achievable accuracy, the systems are quite acceptable for most
monitoring and moderate accuracy measurements.
Once the appropriate thermocouple type is selected, attention must be given to the cold junction
compensation method. The two connection points of the thermocouple to the data acquisition
system (DAS) board form two new thermocouple connections. Use of external cold junction
methods between the thermocouple and the board eliminates this problem, but more frequently the
thermocouple is connected directly to the board and uses built-in electronic cold junction
compensation. This is usually accomplished by using a separate thermistor sensor, which measures
the temperature at the system connection point to determine the cold junction error, and providing an
appropriate bias voltage correction either directly or through software. An important consideration
is that the internal correction method has a typical uncertainty of the order of 0.5
to 1.5
C and, as a
systematic error, this error is directly passed on to the measurement as an offset.
These boards also may use internal polynomial interpolation for converting measured voltage
into temperature. If not, this can be programmed into the data-reduction scheme by using, for
example, the information of Table 8.7. Nonetheless, this introduces a ‘‘linearization’’ error, which is
a function of thermocouple material and temperature range and typically specified with the DAS
board.
Thermocouples are often used in harsh, industrial environments with significant electro-magnetic
interference (EMI) and radiofrequency (rf) noise sources. Thermocouple wire pairs should be twisted
to reduce noise. Also, a differential-ended connection is preferred between the thermocouple and the
DAS board. In this arrangement, though, the thermocouple becomes an isolated voltage source,
meaning there is no longer a direct ground path keeping the input within its common mode range. As a
consequence, a common complaint is that the measured signal may drift or suddenly jump in the level
of its output. Usually, this interference behavior is eliminated by placing a 10- to 100-kV resistor
between the low terminal of the input and low-level ground.
Cu
Cu
Measuring
junctions
Reference
junctions
1
2
Potentiometer
Copper connecting
wires
Thermocouple
wires
A
A
N
A
B
B
B
Figure 8.24 Parallel arrangement of thermocouples for sensing the average temperature of the measuring
junctions. (From Benedict, R. P. Fundamentals of Temperature, Pressure and Flow Measurements, 3rd ed.
Copyright # 1984 by John Wiley and Sons, New York. Reprinted by permission.)
8.5 Thermoelectric Temperature Measurement 347