The modified thermocouple is illustrated in Fig. 2. The position
of the cold junction was chosen arbitrarily as an example; the position
can be selected by the experimenter for convenience of measuring
its temperature. If possible the cold junction should be
positioned in an ice bath to eliminate the need for an additional
temperature sensor.
The first change from the Heichal et al. design was the addition
of the Teflon coating on the wire, which is simply the standard
insulator sold with the wire (Omega product number TFCI-010-
XX). The Teflon coating was left on as otherwise electrical isolation
between the wire and the substrate could not be achieved consistently
with cement alone. It should be noted that while Teflon’s
melting temperature is 327 C, it is only rated to 260 C and the
thermocouple may not be suitable for temperatures above this.
The ceramic cement (Omega product number CC HIGH TEMP) is
the same type used by Heichal et al.
The second significant modification was the addition of a
thick copper layer applied by a wire arc spray coating technique.
The copper performs the same function as it did with the
Furuhama and Enomoto thermocouple [4], that being to conduct
heat away from the tip of the constantan wire to the aluminum
substrate. Thus, the copper helps to equalize the temperature of
the constantan and the aluminum substrate, allowing the thermocouple
to measure true temperatures much more accurately.
The negative consequence of the copper layer is that it delays
the thermocouple’s response time. However, as will be shown
in the Finite Element Analysis section, this delay is small and
acceptable when considering the accuracy improvements the
copper layer provides. The copper also forms the electrical
connection between the aluminum and the constantan. The
thickness of the junction was optimized using FEA as described
in the next section.