Thermocouples (TCs)

The simplest sensors that only seem simple

ThermocouplesThermocouples are among the easiest temperature sensors to use and obtain and are widely used in science and industry. They are based on the Seebeck Effect or Thermoelectric Effect that occurs in electrical conductors that experience a temperature gradient along their length.

They are “simple”, rugged, need no batteries, measure over very wide temperature ranges and more. They have their quirks, too, like everything else.

The solution to thermocouple uses and problems lies in the details of a given application.

Have a read through, click on the key topics below to get more detail about the thermocouple (TC), how it works, color codes, recommended use limits and thermocouple standards.

The applications are where the seasoned TC user and the novice part company and the one with know-how usually achieves the desired result. If all else fails to impress you to be thorough and careful with thermocouples, consider these few facts:

1. Thermocouples measure their own temperature. You must infer the temperature of the object of interest by being certain there is no heat flow between them when you take the measurement.

That’s easier than it sounds in some cases.

2. Thermocouples can err in reading their own temperature, especially after being used for a while, or if the insulation between the wires loses its resistance due to moisture or thermal conditions, or there are chemical, nuclear radiation or mechanical effects in the immediate surroundings.

3. Beware of electrical hazards using thermocouples. They are electrical conductors and even refractory oxide sheathed models at high temperatures, should they contact a source of electricity, could kill you!

They can’t, it is physically impossible to have a temperature gradient at a point. Also, the Electric Field Strength (i.e. volts/meter) at such an impossible condition would be infinite, sufficient to tear the materials apart.

So, if you want to understand TCs better, start with the very basics; learn about the Seebeck Effect and how thermocouples really work!

Also, if you use thermocouples, you need to have some way to interpret their small and non-linear output voltages.

There’s lots of ways to do it, the simplest being a measurement of the output voltage and looking up the value in a table of millivolts dc versus temperature AND correcting for the cooler junction not being at 0 °C, or 32 °F, according to your units inclination.

The high end is to hook the thermocouples up to a modern readout display or a DAQ module plugged into a PC and read away!

However, it is not a fool proof business and there are many subtle things about thermocouples and their uses that have made well-intentioned engineers, who thought they understood them, look like fools.

The hooking up and reading out are most often the easy parts of a measurement.

The selection, installation details and the conditions of use play a big role in obtaining a measurement that is accurate and reliable.

It’s like many measurement subjects, the devil is in the details and “simple” thermocouples have a lot of details!

Their context of use is perhaps the biggest, especially where relatively high temperature-above a few hundred degrees (on anyone’s scale) are the object of measurement.

Above a few thousand degrees, there are a whole range of additional problems.

Thanks for visiting and if you plan to use them, be careful, don’t get fooled by their apparent simplicity. Learn the details, start with the basics (What measurement precision do I need?-Is it realistic for using thermocouples, etc.etc?)

In other words always apply a systematic, measurement engineering approach.

2 thoughts on “Thermocouples (TCs)”

  1. Ray —
    Both Measuring and Reference Junctions do the Sensing/Measuring”.
    They just do not GENERATE the voltage.

    Also, where is the link “to an excellent article”


    1. Thanks, Ray.

      You are correct of course,the Seebeck voltages are generated along the thermocouple. That distinction is lost on many folks who believe firmly that the actual net voltages are generated at the physical location of the two junctions. That would require a pair of voltage source of infinite emf, as we have said. That is nonsense.

      Also thanks for picking out the bad link to Bob Moffat’s paper in Electronics Cooling magazine, It was changed to: and the link on Page 3 on thermocouples has been fixed. That happen a lot these days as various references update their websites.


Leave a Reply

Your email address will not be published. Required fields are marked *