This week, I wanted to start off with something that everyone has seen;
thermocouples. Nearly every physical phenomenon in this universe is
affected by temperature. Therefore, the measurement of temperature is
paramount to almost any scientific or industrial (think process
control!) endeavor. Thermocouples are widely used, but not always well
Here we go...
“Thermocouples: How, what, when, where, why”
This will be a two-part part discussion in order to keep the mailing
short. We’ll end this topic next week. This week we will answer
questions, How, What When Where Why, and next week the letter will be
in the form of an FAQ with questions that we see all the time.
What is a Thermocouple?
A thermocouple is a temperature transducer created when two dissimilar
metals are connected electrically. The electrical potential created is
a function of the composition of the materials that are connected and
of the temperature differential between the electrical ‘junction’
and the ‘open’ ends of the wire. Sounds complicated, looks simple.
Here’s the rub;
1. Take two metal wires of different metallic composition.
2. Strip both ends of each wire.
3. Form an electrical junction by twisting one end of each wire
4. Measure the electrical potential between the wires on the end
opposite the junction.
You have just created a TC.
Thermocouples are simple. The theory behind them is not.
How does a Thermocouple work?
In 1921 a scientist named Thomas Seebeck discovered that when two
metals of dissimilar composition are joined together electrically in a
closed loop and heated at one end, a continuous electrical current is
created in that loop. The current flows for as long as there is a
temperature gradient from one part of the loop to another.
If the loop is broken and the electrical potential between the loose
ends is measured (as above), the electrical potential is a function of
the temperature of the remaining junction. This is known as the
thermoelectric effect or the Seebeck effect.
The physical cause of the Seebeck effect is beyond the scope of this
discussion. It involves thermoelectric charge carrier diffusion
asymmetries and all sorts of other stuff. It is enough to say that we
can exploit the Seebeck effect to reliably convert temperature to an
electrical potential that can be measured and recorded electronically.
(We may cover this in more detail at a later time depending on your
* See the EXCELLENT discussion about thermocouples in the Omega
publication ‘The Temperature Handbook’. (www.omega.com)
Why use a Thermocouple?
1.) Most obviously: to measure temperature. More specifically, to
‘transduce’ (is that a verb?) temperature into an electrical signal
that can be measured by an electronic instrument. Thus, a
thermocouple is a temperature transducer.
2.) Because your experiment might entail the destruction (either
intentional or unintentional) of your test article. Thermocouples are
cheap and (usually) don’t contain any materials that are harmful to
health or the environment. The same could not be said for a mercury
thermometer, for example.
3.) Your temperature ranges are very high or very low. While
Thermocouples are sensitive to temperatures, (otherwise, why would we
use them to measure temperature?) they are not sensitive to temperature
in their ability to be useful. Whereas an integrated circuit (IC) type
temperature transducer might have a limited operating range of
temperature, Thermocouples can be much more widely used. I have used
thermocouples directly in an open methane flame! Try that with an IC!
4.) Thermocouples are simple to use and extremely flexible. Being
inexpensive, TC’s are basically expendable. They require little or
no setup time and are reliable under a wide range of circumstances.
They are light, cheap, flexible, small, and easy to attach to Units
Under Test (UUT’s).
That’s up to you. Thermocouples are probably the most widely used
method for converting temperature into an electrical signal. They are
primarily useful whenever cost or ease of use is a factor. They are
cheap and easy to use if the correct equipment is utilized. That is
NOT to say, however, that thermocouples do not have limitations. In
many cases, there are better choices for temperature transducers.
We’ll talk about that at a later date as well.
Limitations of Thermocouples
1. Thermocouples are relatively INACCURATE. It is generally not
possible to resolve temperature differences less than about 1 degree
Celsius with a thermocouple.
2. The time constant of a thermocouple response is a function of the
size of the electrical junction where the two dissimilar metals meet.
So, while it is possible to create a thermocouple by twisting two
different wires together, great care is usually taken to minimize the
‘bead’ size so that reasonable time resolution can be obtained.
Rough Rule of Thumb: Thermocouples respond on the order of magnitude
of 1 – 10 Hz. It is usually pointless to acquire data from a
thermocouple at anything greater than 100 Hz. Remember, heat flow into
and out of a thermocouple bead is a function of the conduction,
convection and radiation from that bead. I have seen cases where
radiation effects were so high that we were over 100 degrees C from our
theoretical temperature. (The TC was wrong, not our theory..)
3. Interpolation of temperatures relies on high-order non-linear
polynomial curve fits which make it computationally expensive to
calculate temperature from the measured electrical potential. As an
example, a Type T thermocouple has a 14th order curve fit!!!
4. When a Thermocouple is attached to a data acquisition device,
additional thermocouple junctions are formed where the thermocouple
wires attach to the instrumentation. These junctions introduce
inaccuracies into the measurements and MUST be compensated for. This
requires sophisticated signal conditioning equipment with isothermal
terminal blocks and yadda yadda blah blah blah….. (We’ll talk more
about this next week). (Yes, I said “blah blah blah”!)
5. The electrical potential created by TC’s even at extreme
temperatures is typically very small, on the order of millivolts.
Thermocouple signals, therefore, are not very robust. Sophisticated
signal conditioning of the raw signal is usually required, and
re-transmit equipment is REQUIRED if the measurement has to be
conducted very far away from the signal conditioning equipment. A good
rule of thumb is that you must keep your TC cable to under 100 feet or
so. And don’t forget shielding, twisting, and proper termination of
the shield! (Another topic somewhere down the road.)
6. A thermocouple can only measure a temperature DIFFERENCE between the
junction and the point where the TC is connected to the
instrumentation. Therefore, the temperature of the isothermal terminal
block on the instrumentation must also be known. (Again, more next
Next week our discussion will follow an FAQ format, with the following
questions (and possibly others) answered.
Q: Do I need to calibrate my thermocouple?
Q: How do I connect a thermocouple to my measurement system?
Q: How do I convert the voltage to temperature?
Q: My measurements are off by several degrees, outside of the
published accuracy of a TC, why?
Q: There are like 100 different types of thermocouples out there,
which one should I use?
Q: Can I ‘share’ the output of my thermocouple with two different
data acquisition devices?
Q: How do I attach my Thermocouple to my UUT?
Q: How do I properly compensate for the different junctions created
by my instrumentation connections?
See you then!!!