Campbell Scientific CR10X Measurement and Control System - Manual (Page 208)

SECTION 13. CR10X MEASUREMENTS
13-16
ERROR SUMMARY
The magnitude of the errors described in the
previous sections illustrate that the greatest sources
of error in a thermocouple temperature
measurement are likely to be due to the limits of
error on the thermocouple wire and in the reference
temperature determined with the built-in thermistor.
Errors in the thermocouple and reference
temperature polynomials 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
°C
is measured with a type T (copper-constantan)
thermocouple, using the
± 2.5 mV range. The
nominal accuracy on this range is 2.5 µV (0.1%
of 2.5 mV), which at 45
°C changes the
temperature by 0.06 o C. The RTD is 25 °C but is
indicating 25.3
°C, and the terminal that the
thermocouple is connected to is 0.3
°C cooler
than the RTD.
TABLE 13.4-4. Example of Errors in
Thermocouple Temperature
Source Error °C % of Total Error
1 °C 1% Slope
Error Error
Reference Temp. 0.6 36.1 69.6
TC Output
ANSI 1.0 60.1
0.01 x 20 o C 0.2 23.2
Voltage
Measurement
0.06 3.6 7.0
Reference
Linearization
0.001 0.1 0.1
Output
Linearization
0.001 0.1 0.1
Total Error
With ANSI error 1.662 100
Assuming 1% 0.862 100
slope error
13.4.2 USE OF EXTERNAL REFERENCE
JUNCTION OR JUNCTION BOX
An external junction box is often used to facilitate
connections and to reduce the expense of
thermocouple wire when the temperature
measurements are to be made at a distance from
the CR10X. In most situations, it is preferable to
make the box the reference junction, in which case
its temperature is measured and used as the
reference for the thermocouples; copper wires are
run from the box to the CR10X (Section 3.4).
Alternatively, the junction box can be used to couple
extension grade thermocouple wire to the
thermocouples being used for measurement, and
the CR10X panel used as the reference junction.
Extension grade thermocouple wire has a smaller
temperature range than standard thermocouple wire,
but meets the same limits of error within that range.
The only situation where it would be necessary to
use extension grade wire instead of an external
measuring junction is where the junction box
temperature is outside the range of reference
junction compensation provided by the CR10X. This
is only a factor when using type K thermocouples,
where the upper limit of the reference compensation
linearization is 100
°C and the upper limit of the
extension grade wire is 200
°C. With the other types
of thermocouples, the reference compensation
range equals or is greater than the extension wire
range. In any case, errors can arise if temperature
gradients exist within the junction box.
Figure 13.4-4 illustrates a typical junction box.
Terminal strips will be a different metal than the
thermocouple wire. Thus, if a temperature gradient
exists between A and A' or B and B', the junction box
will act as another thermocouple in series, creating
an error in the voltage measured by the CR10X.
This thermoelectric offset voltage is a factor whether
or not the junction box is used for the reference. It
can be minimized by making the thermal conduction
between the two points large and the distance small.
The best solution in the case where extension grade
wire is being connected to thermocouple wire would
be to use connectors which clamped the two wires in
contact with each other.
An external reference junction box must be
constructed so that the entire terminal area is very
close to the same temperature. This is necessary so
that a valid reference temperature can be measured,
and to avoid a thermoelectric offset voltage which
will be induced if the terminals at which the
thermocouple leads are connected (points A and B in
Figure 13.4-2) are at different temperatures. The
box should contain elements of high thermal
conductivity, which will act to rapidly remove any
thermal gradients to which the box is subjected. It is
not necessary to design a constant temperature box;
it is desirable that the box respond slowly to external
temperature fluctuations.