Global TMW:

Rowe's and Columns

Blog

Rowe's and Columns

« A cell phone for Halloween?... | Main | Measurement proverbs... »

Link This | Email this | Blog This | Comments (2)

Martin Rowe

Thermocouples raise reader temperatures

October 24, 2008 In the October print issue, my article "Thermocouples often described incorrectly" has raised a few eyebrows. In the article, I address the often misunderstood idea that a voltage develops across a thermocouples junction when a voltage actually develops across a wire when subjected to a thermal gradient. Several readers sent e-mails both prasing and criticizing the article. For example:

It was good to see your short piece on thermocouples in TMWorld. I have encountered more than one EE college professor who did not understand how a T/C works. It would be useful to tell your readers just where between the wire ends the emf is actually generated – in regions along the wire that are experiencing a thermal gradient.

This can become crucial to understand when interpreting the calibration of a ‘damaged’ T/C (where the damage may be mechanical, chemical, over-temperature operation, etc.). If in the cal lab the calibration thermal gradient is imposed on an undamaged length of the T/C conductors, then the resulting calibration looks good and misleads the calibration user. When a damaged T/C is calibrated by imposing the calibration thermal gradient on the ‘damaged’ region of the wire, then the results of the damage show up in the calibration.

Others questioned the fact that I didn't discuss cold-junction compensation, which you need because the junctions of thermocouple wires and terminal block form additional thermocouples and instrumentation must compensate for the error voltages that can occur. Because of the nature of the Test Digest page, I didn't have the space to cover that important topic.

Thermocouple information is available through many online sources. To learn more, see the thermocouple section in Omega Engineering's technical reference section. There, you can follow the "Using Thermocouples" link for a detailed description of thermocouple measurement principles. Other links discuss thermocouple properties and construction.

Posted by Martin Rowe on October 24, 2008 | Comments (2)

Industries: Bench and Modular Instrumentation, Machine Vision and Inspection

November 19, 2008
In response to: Thermocouples raise reader temperatures
mjohnm commented:

I'm afraid there are several factual errors in your article on thermocouples. Emfs ARE produced at the junction between dissimilar metals - THIS is the Seebeck effect and the emf is the Peltier coefficient (which depends on the temperature of the junction). You correctly state that emf is also developed along the length of a wire of a single metal - this is the Thomson effect.

(The Peltier coefficient should not be confused with the Peltier effect which is the inverse Seebeck effect - if a current is passed through the junction between dissimilar metals, heat is either absorbed or released depending on the direction of the current. This is the principal behind Peltier coolers.)

The simplest thermocouple circuit involves two junctions between dissimilar metals. An emf is generated in the circuit if it is broken in one metal and a meter attached. The total emf generated by that circuit is the difference between the Peltier coefficients at the two junctions plus the difference in the Thomson emfs in the two metals. If one junction is held at a constant temperature the emf has an approximately parabolic dependence on the temperature of the other junction.

I wish to write a short article for you (between 1/2 and 1 page) properly describing thermoelectricity and listing some precautions on how to make valid measurements (e.g. the need to avoid the "neutral point"). I can be reached at mjohnm@fracture.rutgers.edu if you would like such an article.