What are Thermocouples?

Thermocouple

According to the Seebeck effect (also called the thermoelectric effect), a temperature

gradient along a conductor induces a voltage across the conductor. However,

measuring this voltage involves attaching a second conductor to the h ot end, which will

see its own voltage induced and cancel out the voltage we wished to measure.

Fortunately, the relationship between temperature difference and voltage varies with

materials, so by using a second conductor made of a different material, a small voltage

difference can be found at the ends. This device, a two-terminal element consisting of

two different metal wires joined at the end, is called a thermocouple.

The effect is weak, typically 10’s of microvolts / ○C, but can be amplified to a more

useful range. Thermocouples operate over a very wide range of temperatures, are

inexpensive (just two pieces of wire), and can stand up to harsh conditions.

You can
make your own thermocouple by joining two wires of dissimilar metals. If the wire is

thin, they can respond to temperature changes very quickly.

Thermocouples measure the difference in voltage between two points. To know the

absolute temperature at a given point, one must know the temperature at the other

point. Typically we measure the temperature at the hot junction at the tip of the

thermocouple where the two metal wires are joined.

The other end is called the cold

junction. We must know or measure the temperature at the cold junction, either by

immersing the cold junction in a liquid bath of a known temperature or by using a

thermistor or other thermal sensor to measure the temperature at the cold junction. The

temperature at the cold junction should be converted to an equivalent voltage, then

added to the measured voltage. The sum shou ld then be converted back to find the ctual temperature at the hot junction. This is called cold junction compensation. If a

system has multiple thermocouples, they can all share a single cold junction for

simplicity.

The relationship between the temperature difference and the thermoelectric voltage is

somewhat nonlinear. It is often described by a look-up table rather than an analytic

function.

Thermocouples produce a very small output voltage that must be mightly amplified to

reach the full scale of a typical DAQ. Moreover, the output impedance is high, so the

amplifier must have very high input impedance. Instrumentation amplifiers are well-

suited to this job. Figure 2 shows an instrumentation amplifier connected to a

thermocouple. The circuit produces an output of

 

100 1 out ref

k

V V V V

R

            

Remember that Vout should be positive. If the thermocouple may be measuring

temperatures colder than the cold junction, it is helpful to set Vref > 0 so the output isn’t

clipped.

R

5 or 3.3 V

Vref

Vout Thermocouple

Instrumentation Amp

Figure 2: Instrumentation Amplifier for measuring thermocouple

For example, a type-E thermocouple has one conductor made of nickel-chromium and

one of copper-nickel. It produces a voltage difference of about 59 V / ○C. The accuracy

is 1.0 – 1.7 ○C.