Page 156 - Mechanical Engineers' Handbook (Volume 2)

2 Thermocouples 145

difference across that length when the thermocouple is in service. The cold-worked region
is often very short, which makes it difficult to detect.
A similar drop in calibration was reported by Bentley and Morgan for PtRh versus Pt
10
thermocouples in response to cold work introduced by handling. They reported a drop of
about 0.4% in the cold-worked region. Annealing at 200 C did not entirely remove the drift
in the PtRh leg. It is difficult to identify this problem by subsequent recalibration, since the
region of partially degraded material may be placed in a uniform temperature zone during
recalibration and hence play no part in generating the signal under calibration conditions. If
a thermocouple is suspected of being inhomogeneous, it should be tested for homogeneity
along its entire length. If the test shows that the wire is homogeneous, then no calibration
is required, since the used portion of the wire is the same as the unused portion. If the wire
is not homogeneous, no recalibration can be of value because it will be impossible to place
the temperature gradient in the same location for calibration as it was for service. This
situation is described in Figs. 10–12.
Assume that the thermocouple was exposed to the unfavorable environment only near
the hot end, as shown in Fig. 10, and that both wires became less active as a result of the
reaction. The output will drop, as shown. If this defective thermocouple were placed in a
usual calibration facility, all of the affected material would be in the region of uniform
temperature. The emf would be generated entirely by the material near the entrance of the
furnace, which was never changed—the wires between points 1 and 2 and between 6 and 7
in Fig. 11. Under such conditions, a perfectly normal signal would be developed.
Figure 12 illustrates a test for homogeneity that can identify a defective thermocouple.
To test a thermocouple for inhomogeneity, clamp the junction at a constant temperature and

Figure 10 Temperature distribution along the
thermocouple, in service, and the resulting output
after deterioration of the wires. (Reproduced
from Ref. 2, with permission.)