From Wikipedia, the free encyclopedia
Jump to: navigation, search
          This article is of interest to the following WikiProjects:
WikiProject Physics (Rated C-class, High-importance)
WikiProject icon This article is within the scope of WikiProject Physics, a collaborative effort to improve the coverage of Physics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 High  This article has been rated as High-importance on the project's importance scale.
WikiProject Electronics (Rated B-class, Mid-importance)
WikiProject icon This article is part of WikiProject Electronics, an attempt to provide a standard approach to writing articles about electronics on Wikipedia. If you would like to participate, you can choose to edit the article attached to this page, or visit the project page, where you can join the project and see a list of open tasks. Leave messages at the project talk page
B-Class article B  This article has been rated as B-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
WikiProject Technology (Rated B-class)
WikiProject icon This article is within the scope of WikiProject Technology, a collaborative effort to improve the coverage of technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
B-Class article B  This article has been rated as B-Class on the project's quality scale.
Checklist icon

Voltage created by Temperature Gradient vs Voltage created at junction[edit]

I'm deleting the following:

A common myth regarding thermocouples is that junctions must be made cleanly without involving a third metal, to avoid unwanted added emfs.[5] This may result from another common misunderstanding that the voltage is generated at the junction.[6] In fact, the junctions should in principle have uniform internal temperature, therefore no voltage is generated at the junction. The voltage is generated in the thermal gradient, along the wire.

The current theory as taught in the Physics Department of every accredited University in the United States is that it occurs at the junction of dissimilar metals.

A straightforward treatment on the subject can be found here:

I mention the Galvanic Anode, which has been used to protect iron/steel ships for centuries. Place a massive Zinc Block on an iron hull, and an automatic voltage is created by the junction. The iron becomes slightly negative, while the zinc becomes slightly positive. The negative iron repels negative ions that would otherwise corrode it (Cl-, OH-, etc . . . ), the positive zinc on the other hand attracts these, and so corrodes much faster than it would normally. This is why it is sometimes called a sacrificial zinc anode. Note there is no thermal gradient in this system. — Preceding unsigned comment added by Csdidier (talkcontribs) 02:01, 21 July 2015 (UTC)

The link you provide says nothing about thermoelectricity. It is a discussion of band diagrams and built in voltages which are unrelated to power-driving (voltmeter) voltages.
I'm sorry that the treatment was too abstract for you to see how the link is directly related to thermoelectricity. Although it does not state the term explicitly, it is completely related to thermoelectricity, in fact it explains succinctly the driving mechanism behind it. More importantly, it is the scientific explanation of the phenomena, which the reference contained in the section I deleted claims to be a myth. Here is the reference: [1]
Now in this reference, Martin Rowe makes the following claim:
many descriptions out there get it wrong and that misinformation keeps circulating. I once took that misinformation as correct, but no more. Here's the misleading statement, paraphrased:
"A thermocouple is made of two dissimilar metals joined to form a junction. A voltage occurs across that junction that changes with temperature."
There is no voltage generated at the junction where the two metals meet.
Martin Rowe is stating quite clearly that Dr. Michael Fortner from the University of Illinois is wrong, or rather that he is providing misleading information.
Furthermore, if we are to believe Martin Rowe, then the first line of paragraph 2 of this very Wikipedia article, is also misleading information:
Any junction of dissimilar metals will produce an electric potential related to temperature.
Wikipedia Page article with a reference that states that a preceding statement is false? A Wikipedia page needs to be somewhat consistent.
In addition, if we are to believe Martin Rowe, then this entire Wikipedia Page needs to be deleted:
Because Martin Rowe states that "There is no voltage generated at the junction where the two metals meet."
Here is another reference, which states the exact same theory that Dr. Michael Fortner does, the only difference is that Dr. Robert M. Metzger explicitly states the connection to thermoelectricity for those who unable to make that connection themselves:


Above excerpt from:

The Physical Chemists' Toolbox - Robert M. Metzger

An unfortunate aspect of solid state physics terminology is that the word "voltage" gets used in too many different ways. Electrostatic potentials, electron band energies, and Fermi levels can all be stated as voltages but it is only the Fermi level—the thermodynamic quantity—for which differences can supply real work.

Solid state Physicists know what they are talking about, when they use the term "voltage". It is the same "voltage" that Chemists use, Electrical Engineers use, etc . . . Electrostatic Potential is voltage. If you close the circuit, it won't be static anymore. I have heard of Capacitors (electrostatic potential) so large that they can power a radio for up to a half hour. Solid state physics terminology does not refer to electron band energies and Fermi Levels as voltages. Since these are energy levels, they use units of energy, such as the Joule, or the electron-Volt. I think Michael Fortners diagram used electron-volts. Are you confusing electron-volts with Volts? An electron-volt is a unit of energy. A Volt is a measure of Voltaic pressure.
Here is an actual discussion of thermocouples from a university: (I am sure there are many more, that is just the first hit I found when I googled for *thermocouple temperature gradient*). Observe slide 5 where it is stated in words and equations that the emf occurs in temperature gradients.
There's no author name on the link you provided, so I can't tell if it is authored by a professional, or a crack-pot.
Regarding the electrochemistry of iron and zinc, that is just another type of emf unrelated to thermoelectricity. It also relies on the presence of an electrolyte, since aside from the small current, that iron and zinc masses are in equilibrium (0 volts difference) as they are shorted to each other; the disequilibrium occurs in the electrolyte and at the metal surfaces. Incidentally, electrochemical behaviour is also not related to built in voltages in the semiconductor sense; I don't think it's possible to calculate the 0.32 V iron-zinc cell open circuit voltage on the basis of band diagrams.
I think you are talking about a Galvanic Cell. I was talking about a Galvanic Anode, which is completely different. A Galvanic Cell requires electrolytes to generate the EMF, whereas a Galvanic Anode is an example of an EMF generated from Zinc in contact with Iron (no water needed, no electrolytes needed, etc . . . ) If I didn't know any better, I would think that you were employing a gimmick called the Strawman Argument. I wasn't talking about Galvanic Cells, I was talking about Metal Metal junctions. Why do you want to change the subject? Do you think Galvanic Cells and electrolytes are related to Thermocouples? I don't think electrolytes are involved in Thermocouples. Csdidier (talk) 02:24, 22 July 2015 (UTC)
(By the way, it's usually best if you add Talk page discussions at the bottom, for better chronological ordering.)
--Nanite (talk) 06:29, 21 July 2015 (UTC)
Please do not insert your replies in the middle of others' comments, this is considered poor talk page etiquette as it makes it difficult for others to follow the flow of the conversation.
Back to the topic at hand, ...
If you are interested in learning about the underlying physics of thermocouples see the page Seebeck effect and it should be clear from the equations that the emf only arises where there is a temperature gradient. If you prefer an explicit statement from a named professor see slide 6 of this PPT from this guy. Let me know if you decide he is a crackpot and I can find you another.
If you are not interested in learning more, all I can tell you is that you have no idea what you are talking about.
Stick to the article content. This is just being petty. As well, your referenced article is not enlightening. It looks amateurish and could not be used as a credible reference in this article.
Kckid (talk) 11:27, 13 November 2015 (UTC)
Yes, I know exactly what you believe implicitly obvious from the PDF you linked, and I also know it is simply a common misconception (I used to have the same when I was an undergraduate, long before I was a thermoelectric materials researcher and had to get it right to do my job properly). If you really believe that a contact potential difference "explains succinctly the driving mechanism" behind thermoelectricity, I invite you to provide or derive an equation that gives the correctly thermocouple voltage in terms of contact potentials. Here's a hint: no such equation can possibly exist; the thermocouple voltage depends on Seebeck coefficient which is a bulk property whereas the contact potential is a surface property. There is indeed an electrostatic potential step at the interface of two metals, but this potential cannot sustain a current (and so it can't possibly be responsible for the power output from thermocouples). The articles on Fermi level, electrochemical potential, voltage, and yes even on Volta potential are all very clear on this point, that contact potentials do not produce a steady reading on your voltmeter.
Here is a nice pedagogical article I found which explains in far more detail how this contact potential misconception appears in thermoelectricity discussions. (Don't worry, you aren't alone, as they say even some scientists have gotten confused about Seebeck effects.)
Cheers, --Nanite (talk) 18:35, 22 July 2015 (UTC)
After reviewing these presentations from Mechanical Engineering folks, I might not call them crackpots but, their work is both visually awful and mathematically weak. If one buys into some of these meager explanations, the manufacture of thermopiles would be a total waste of money. One could simply coil up wire and put it in a source of heat and generate electricity, we know this is not true. Regarding, "A note on the electrochemical nature of the thermoelectric power," I don't see how this adds to the argument that this article should be written in terms of gradients along conductors. Please refer to Figure 1 of that article, it regards junctions, no less. The main purpose of that paper, is discussion of equations vis-a-vis the thermoelectric effect in semiconductors using diffusion equations. And as is stated, therein, the article contains original equations. Hardly, what pedagogical means. And it has nothing to do with a thermal gradient along a wire generating electricity as you appear to believe. Arguing the Seebeck effect in metals and the fact that voltages appear at thermocouple junctions is neither here nor there. Without the junction no thermal power is generated and the usefulness of thermocouples is negated.
Kckid (talk) 11:27, 13 November 2015 (UTC)
Look, the Seebeck effect is very simple in an open circuit thermocouple: \nabla V = -S \nabla T. S is a material bulk property and not a junction property. From this one equation everything, and I mean, everything about thermocouple circuits can be derived very easily.
There are indeed some introductory sources that say the thermocouple voltage is generated at the junction. This is wrong, but it produces the right results for the most basic thermocouple system and so the error is not immediately apparent.
If you read any serious (non-introductory) source about the physics behind thermocouples or Seebeck effect they never say that the voltage is generated at the junction and often explicitly note the misunderstanding. I do not want to go around finding new sources for you guys all day long and I don't have much access to paid articles, but here are some more sources I could find:
  • ASTM Manual on the Use of Thermocouples in Temperature Measurement: The Seebeck electromotive force (emf) is the internal electrical potential difference or electromotive force that is viewed externally as a voltage between the terminals of a thermocouple. This Seebeck source emf actually occurs in any electrically conducting material that is not at uniform temperature even if it is not connected in a circuit.' The Seebeck emf occurs within the legs of a thermocouple. It does not occur at the junctions of the thermocouple as is often asserted nor does the Seebeck emf occur as a result of joining dissimilar materials as is often implied. Nevertheless, for practical reasons (Section 2.1.3) it is always the net voltage between paired dissimilar materials that is used in thermocouple thermometry.
  • Kerlin, Thomas W.; Johnson, Mitchell "Practical Thermocouple Thermometry", very start of chapter 2: Voltage is not produced at the junction of the thermocouple wires. •Voltage is produced along the portions of the thermocouple wires that experience temperature differences.
  • CRC Handbook of Thermoelectrics, Chapter 1 Intro (by Daniel Pollock). . The Seebeck effect does not arise as a result of the junction of the dissimilar materials, nor is it directly affected by the Thomson or the Peltier effects;... These responses are in contrast to that of the relative Seebeck effect, which exists as long as the temperature gradient is maintained, regardless of whether current flows or not.
Too frequently the RSE [Relative Seebeck EMF] has been incorrectly described in the literature as being a consequence of the external contact potential, or Volta effect, between dissimilar materials. The external contact potential is not a thermoelectric effect. An external contact potential is induced when two different materials are brought sufficientlyclose to each other, but are not in physical contact, so that electron transfer between them results in a common Fermi energy level in each. This mechanism is independent of temperature and vanishes virtually instantaneously (~ 1e-15 s) when the materials make physical contact. The external contact potential has no relationship whatsoever to any thermoelectric phenomenon.
Regarding whether a circuit without junctions is useful or not, that is besides the point. The basic misunderstanding (that an equilibrium material junction at uniform temperature can somehow produce a voltage) is the root of many misconceptions about thermocouple circuits.
Is that less amateurish enough for you? --Nanite (talk) 22:09, 13 November 2015 (UTC)

Thermocouple vs. Thermistor[edit]

The explanation of cold junction compensation says you need a secondary thermistor to correct errors in thermocouple reading. This begs the question, "why not just use a themistor?". I assume the answer has to do with dynamic range (and maybe precision, although the article says thermocouples are difficult to get better than 1 degree celsius resolution...). But it would be nice if there were some explanation in the article. --User:Chinasaur

The primary reason is "dynamic range" (a good thermistor usually has a narrow temperature range), with other reasons being self-heating (unlike a thermocouple, you have to apply current to it) and the very non-linear resistance vs. temperature curve. Hmmm, the real way to answer this is with the various electronic temperature measurement pages having the pros and cons of each style of device. I'll put that on my (long) todo list. -- Kaszeta 20:04, 5 Oct 2004 (UTC)
How would you program the Voltage-Tempature relationship? 02:51, 6 September 2006 (UTC)t59
From my experience and experiments a thermocouple has a much greater accuracy than +/- 1K. Oystsot (talk) —Preceding undated comment added 17:22, 7 August 2009 (UTC).

Deleted 'Dawson' Theory[edit]

deleted the following:

"Recently discovered by MA Stephen Dawson, thermocouples may be used in conjunction to calculate circuit Q factors as well as measure Impedance. (XL, XC): The formulas: 2π x F x L OR 1 / (2π x F x C) may be used to calculate such things. Stephen Dawson's discovery states that these can be calculated by a thermocouple device, to gain accurate results an orifice plate may be used within a DP Cell. "

As written, it conveys no useful information, and makes confusing assertions without references, and sounds moreover like self-advertising. There may be a relation between thermocouples and electrical circuit Q, but none that I can find in any textbook, and none that I have encountered in a decade of working with thermocouples.

Sounds good. I don't even understand what they're implying, now that I think of it. Maybe using thermocouples inside an oscillator or something? - Omegatron 18:09, Apr 4, 2005 (UTC)

The above is very much nonsense. The formulas listed are just the basic first year formulas for capacitive and inductive impedance. An orifice plate withing a DP cell? Okay for starters a DP cell is about the size of a quarter is a piezoelectric or capcitive sensor about 5mm thick and an orifice plate is a big, heavy, steel plate used in industrial pipe-works to set up differential pressures in flowing fluids to measure flow rates. -Jim Eld

Two junctions[edit]

Why are two juctions needed, as I currently understand it one juction will produce a EMF, so why are two juctions needed? mickpc

Without a second junction, how could the electric circuit ever be closed? Think about it: We have the voltage generated at the (hot) junction between, say, the copper and iron wires, but now those electrons have to get back home. Somewhere, there must be a second place where the copper and iron touch (well, unless you're simply measuring the electrostatic field that is generated).
Atlant 12:19, 2 August 2005 (UTC)

Ok, well technically you can simply use the perant metal as a wire so u can use one junction. But my question revolves around; if the actual junction causes the EMF or the perant metal does, reading the article again I am tending to think that the EMF is devolped across the perant metal. mickpc

"Think about it..." The universe does not yield to common sense or rumination. And this is about electrostatics, an emf generated at the junction. There is no parent metal, there are two metals.
Kckid (talk) 07:45, 13 November 2015 (UTC)
I think you need to draw yourself a diagram. Draw the iron wire. Draw the copper wire. Connect them at one end of your diagram to create the hot-side thermocouple. Now figure out how you connect the wires at the other side of your diagram to create a complete circuit.
Atlant 11:09, 3 August 2005 (UTC)

I asked my lecturer and he said that only one junction is needed and this is generated from the junction itself, now either he is wrong or the article is wrong (and I doubt he is entirly wrong) mickpc

Your lecturer is correct.
Kckid (talk) 07:45, 13 November 2015 (UTC)
Like I said, you figure out how to get the electrons back around the circuit without another metal-metal junction and you post it here.
Atlant 23:20, 21 August 2005 (UTC)

Quite simply, just use the parent metal as the lead wire (or use a lead that is the same composition as the perant metal). This is tyically done to extend the distance between the two junctions or to the output transducer. mickpc

I'm having trouble understanding why you don't get this. Draw the picture! Assuming you have an actual circuit (a "circle"), you are *ASSURED* of having at least two junctions (where A meets B and then where B meets back up with A), and if you have binding posts and other hardware in the circuit, you may have far more than just two junctions.
Atlant 23:28, 27 August 2005 (UTC)

Please dont get me wrong here, I am only trying to clarify (deepen) my understanding, so I am open to constructive critsisim. You can have a single junction for a thermocouple, just use two dissimlar metals (wires) and join them at the junction. This will produce an EMF. Why you typically use a reference junction I am not entirly sure on, I am thinking that it is most likely the linearisation and stabilty. mickpc

Great; you've got an EMF, but what will you *DO* with it? Unless you have an electrometer handy, there's no way to read that EMF that doesn't draw at least a little DC current from the thermocouple. And the moment you start to draw DC current, you have to close the circuit. And then you're back to the problem I've been trying to raise to your attention: a complete circuit must have at least two junctions.
Atlant 12:11, 30 August 2005 (UTC)

Point taken, well I should have thought more about the this before hand, I read that the voltages are produced at the junctions but the article Peltier-Seebeck effect it is said that it is due to the diffusion mickpc

Just wanted to point out (as my physics mentor told me long ago) that there are always at least 2 junctions in a thermocouple sense circuit. The other junction, is created when the thermocouple is connected to the EMF sense device (ie the thermocouple meter, or sensing circuitry). The volt meter probes will create the other junctions when one connects the iron-copper wires to read the TC junction voltage.

The Seebeck effect causes a voltage to form that is proportional to the difference between the temperatures at the two respective junctions. You need to complete the circuit. Your hot junction is the measurement point, the cold is your tranducer (right at the terminal, unless youve got a nice TC input card.) Jim Eld —Preceding unsigned comment added by (talk) 04:26, 18 April 2009 (UTC)

To clarify for mickpc or others, any modern thermocouple measurement runs the thermocouple leads to a DC pre-amplifier, to magnify the very small voltage about 100 times. Because after it's magnified, the thermocouple part of the circuit is pretty much done and the rest of the circuit doesn't have to worry about screwing up the measurement. Well, any microvolt DC preamp starts with two metal input leads of the same metal; you may assume it's copper. They actually do use the same metal; it's very difficult for the preamp chip maker to guarantee low temperature coefficient of input offset (also called drift, specified in uV/degC) without carefully avoiding different metals in the input leads, because those would act as unintentional thermocouples. So, in our thought-expt., preamp-neg-copper connects to thermocouple-neg-copper (same metal, so not a thermocouple junction) and thermocouple-pos-iron connects to preamp-pos-copper, which is the 2nd junction, and it better be at room temperature. Then out in the measuring tip, thermocouple-pos-iron is joined directly to thermocouple-neg-copper, and this is the 1st junction. That makes it a closed loop circuit, for which the loop-integral of thermal emf is well defined. True the preamp does have sort of an open-circuit inside, but really it's within one silicon chip where the voltage difference is being counted up carefully. More appropriate to think of it as a capacitor that measures it's charge, than an open circuit. jimswen (talk) 22:33, 23 August 2009 (UTC)

Enough legitimate experimental research has shown that a thermocouple junction generates an EMF. It's not a misconception or pie in the sky theoretics. Plenty of reference material has already been cited on this talk page. It is not necessary to muddy the water with discussions about "thought experiments" and contrived or particular measurement systems. To put a fine point on it, both are examples of original research and have no place in Wikipedia. If you care to argue more about measurement, please reference "Quantum Measurement" Braginsky, Vladimir B., Khalili, Farid Ya, Thorne, Kip S., first.
Kckid (talk) 07:45, 13 November 2015 (UTC)

Re: Compensating/Extension cables[edit]

I agree of course, but was just trying to keep it a bit simpler. However, I have re-added the bit about adding the compensating voltage to the thermocouple voltage to get accurate result. Dave 21:59, 24 March 2006 (UTC)

Other kinds of thermocouples?[edit]

In a book by Patrick Moore thermocouples is said to be used to measure temperatures at Venus before the space age. This makes no sense to me as thermocouples are used in situ. Can anyone explain how thermocouples work in this case? Gunnar Larsson 20:12, 25 March 2006 (UTC)

I'm not an astronomer but I suspect the instrument used to measure Venus's temperature was a bolometer which is a specialized form of radiation thermometer that converts infrared energy into a tiny temperature rise, which then warms a thermocouple. So, it is an indirect use of the thermocouple principle. ( And now Wikipedia has got to the point where if you just blindly link a technical term like bolometer, chances are very good you come up with a blue link not a red link!) --Wtshymanski 17:26, 26 March 2006 (UTC)
That sounds exactly right; I couldn't remember the term last night.
Atlant 21:19, 26 March 2006 (UTC)
Ah, this makes more sense to me. Thanks! :-) Gunnar Larsson 18:40, 27 March 2006 (UTC)

Would it be possible to also mention the properties of type D and G thermocouples? H rossouw (talk) 08:45, 5 March 2009 (UTC)


A bit more information about thermopiles would be appreciated as I was redirected here from "thermopile". My understanding is that a thermopile is a number of "stacked" thermocouples to generate useful amount of power, where as a thermocouple can only output enough power to be useful as a sensor. I have no references for this piece of information other than sales information from an unvented propane heater.

Great listing of thermocouple types. Thanks. John H.

I use thermopiles on a daily basis to measure the power of laser beams. I added a section about this. I assume that the function in the propane heater is similar to that in heating appliances: the thermocurrent is used to keep a valve open. Putting several thermocouples in series increases the voltage. Han-Kwang 23:05, 7 July 2006 (UTC)
Right, but lots of convection heaters also use the thermopile voltage (through the thermostat) to also operate the main gas valve; the article discusses this.
Atlant 23:25, 7 July 2006 (UTC)
Yes, I can imagine so if it has to regulate a thermostat, but you have modified the section about guard flames. Search the net for "guard flame thermocouple millivolt" and you'll see plenty of service instructions explaining that its voltage should be in the 10--30 mV range, which corresponds to 200-600 K temperature difference over a single thermocouple. You don't need much to power a solenoid that is just there to keep a spring-loaded valve in the open position. Han-Kwang 11:02, 8 July 2006 (UTC)

This article is terrible, and has many errors, both technical and grammatical. For example, the list of coefficients for a K-type starts with n=0, which means a temperature difference with no voltage difference. And heat and temperature are confused. I could go on and on. Someone should start from scratch instead of trying to edit this.

Hal Lewis —Preceding unsigned comment added by Hallewis (talkcontribs) 19:49, 2 March 2009 (UTC)

Hi, wikipedia noob here so I don't know the protocols. Thermopiles are often used in industry to either increase the °K to mV gain of the sensor when stacked in series, or they are stacked in parallel in order to build in redundancy. They are cheap so it's better to just install 10 in parallel then send one of your guys out to repair it 9 times. -Jim Eld

Incorrect info[edit]

I believe that the included coefficients for Type K thermocouple aren't correct. There is no way that the coefficients of larger powers should get exponentially bigger.

 —Preceding unsigned comment added by (talk) 19:18, 10 March 2009 (UTC) 

Just say me wether the area of contact of the two metals do depend on the Volt per degree Centigrate. if not so then parellel connection should not have any variation in its output, but here it does showes variation. Why? and Why? —Preceding unsigned comment added by (talk) 13:41, 30 March 2009 (UTC)

Different issue: The "(nic-iron)" after Type N doesn't sound correct to me. It's at least not as descriptive as the others. Type N uses Ni/Cr/Si vs Ni/?/Si, with the Si there mainly to make protective surface oxidation for high temperature use. I forget what the "?" trace metal is; it may not be very important to the EMF. Even if it is Fe, maybe the description should read ("nic-iron") to show it's referring to a type-name, not to a pair of specific metals. jimswen (talk) 22:50, 23 August 2009 (UTC)

Curie temperature of iron[edit]

The number appears to be wrong. Please see discussion here:

Guy Macon 09:59, 15 July 2010 (UTC)

Followup: User Wtshymanski corrected it here and on the curie temperature page. —Preceding unsigned comment added by Guymacon (talkcontribs) 15:54, 15 July 2010 (UTC)


I noticed the thermoelectric effect navbox template has been moved to the very bottom of the article. This template is meant to be used on the side (unlike some others, see for example the bottom of states of matter. Incidentally, the other links in the navbox should show it being used as intended. I will put it back to the side but I just wanted to check here beforehand in case anyone had any concerns. Thanks, David Hollman (Talk) 10:39, 31 August 2010 (UTC)

Like most navboxes, it's a giant distraction at the top of the article. Do we need it at all? Wouldn't a few in-text links and carefully-chosen entries in "See Also" do a lot more for explaining the topic, instead of a loose collection of facts jammed willy-nilly into a navbox? Please don't put it at the top of the article, the picture isn't even relevant to thermocouples. --Wtshymanski (talk) 12:59, 31 August 2010 (UTC)
I would concur with that in the majority of cases. Kbrose (talk) 03:37, 16 September 2010 (UTC)
From my own experience, when I am not very familiar with a topic a good navbox not only provides links but also structure, which I find very useful. Another advantage is that as the template evolves all the articles benefit, rather than having to keep lots of "see also" sections updated. (Actually, this page [1] explains the benefits pretty well. Also note the mentioned study). However I do appreciate the "clutter" concern... In some articles navboxes are placed in an "introduction" section instead of at the very top, which seems to work okay; this article doesn't have a real "introduction" but given the content of the navbox the "principle of operation" section might make sense. What do you think about putting it there?
In any case I will remove the navbox from the very bottom since that just looks broken. David Hollman (Talk) 15:29, 31 August 2010 (UTC)
I read the "navbox on every page" essay and I disagree with it. Putting "hundreds" of items in a navbox would be an example of Wikilaziness; instead of recommending to the reader the one or two links that most relevantly explain the subject matter, instead waste the reader's time with scores of items that have been free-asssociated with no obvious sequence or ranking by relevancy. An article is instantly de-orphaned if it has a link, either in-line with the text or else in a "see also". The essay admits that a "navbox" does the same function as a "see also" section, but without the benefit of an editor recommending items that are most particularly relevant to the matter at hand. Ultimately you'd just need one navbox saying "Other stuff" and put the whole of Wikipedia in it...after all, you might otherwise miss the connection between thermocouples and the Franco-Prussian War. (The advice to search for articles that "need" navboxes by an Easter-egg hunt using "random article" is precious...if you don't know what's relevant to an article, for goodness sake don't randomly put in a navbox!) --Wtshymanski (talk) 21:50, 31 August 2010 (UTC)
I'm sure most people would agree that these templates need to be well structured and not become random or poorly organized lists. However I don't have the impression that this particular navbox has these problems; although if you have any suggestions for its improvement I'm sure those would be welcome (Template talk:Thermoelectric effect - not the busiest talk page in Wikipedia). David Hollman (Talk) 20:36, 7 September 2010 (UTC)
Isn't that a little like asking for suggestions on improvements in the process for tanning baby skin for briefcases? Some things ought not to be improved. --Wtshymanski (talk) 20:47, 7 September 2010 (UTC)


My point was that unless there are specific issues with this particular navbox, I can't see any reason it should not be included on this page, particularly given that they are used throughout Wikipedia. While I appreciate your personal opinion that navboxes are not useful, I don't think that overrides the broad consensus which seems to be that they are. David Hollman (Talk) 21:11, 15 September 2010 (UTC)

It's been so long that I've forgotten what it looked like. Let's see, my specific issues and objections are:
  1. Object depicted isn't a thermocouple.
  2. Title of navbox is out of context until you read how a thermocouple works
  3. "Principles" section is redundant with principles explanation in the text.
  4. "Applications (general)" section isn't about applications, but randomly sticks in "Thermoelectric materials" and "thermoelectric cooling". And, it's not about thermocouples. Navboxes should be written by people who know the difference between "applications" and "materials".
  5. "Applications (Power generation)" just looks another of a bad and confusing set of headings - why are "applications" broken down this way? Aren't these more sensibly given in the article text?
  6. "Applications (Sensing)" again looks like a poorly-factored heading, and randomly lumps "thermocouple" and "thermopile" together.

Then there's the usual infobox nonsense letters at the bottom. "V D E" sounds like a German standards organization. It's big, it's internally confused, it's irrelevant, it's distracting, it's redundant...but it's a nice shade of blue. It's not an asset to the article. If there was such a thing as a good navbox, this wouldn't be it. "If a dozen Wikipedia editors do a foolish thing, it's still a foolish thing." I have removed similarly badly-constructed and ill-concieved navboxes elsewhere, with remarkably little comment. --Wtshymanski (talk) 21:52, 15 September 2010 (UTC)

It seems to me that this impression of usefulness of navboxes is highly subjective and only relies on the aspect that people like to create them. On Wikipedia people do all kinds of things to avoid writing a good article or creating good prose to explain subjects. Lists of facts, compilations of data, List articles, navboxes, etc. are a lot easier to create than writing something meaningful that others can understand and appreciate. Unfortunately navboxes usually don't explain much at all. I am not saying they aren't useful at all, but they should be extremely well designed in terms of link content and if possible they should fit into the See-Also sections or that general area, which is where WP puts that kind of functionality. The portal pages are for guided learning, not the articles. They should be reserved to subject matter content and have a link to their larger context in the See-Alse section. Kbrose (talk) 03:37, 16 September 2010 (UTC)
Examining the 'Benefits of navboxes' expose, it is pretty clear that the result of that study are bogus, statistically invalid, see my comments there. Kbrose (talk) 04:29, 16 September 2010 (UTC)
Hi, I made the navbox.
My original motivation was to make it clear what all the articles are related to the thermoelectric effect, because they're substantially overlapping, disorganized, and in some cases horribly inadequate. When the articles are all properly self-contained, organized, and interlinking, (someday haha), then the template will be become irrelevant except for the core articles on the thermoelectric effect. I've been gradually organizing and merging and deleting and interlinking and copy-and-pasting over the past couple years, but there's still a huge amount of work to do on this group of articles.
In this case (Thermocouple) we have a well-written and comprehensive article, so it's less important here. I'm not strongly opposed to deleting the template from this article. I am strongly opposed to deleting it in, say, thermoelectric effect (a core article relating to all the others), or thermopile and thermoelectric generator (which are crappy and not-self-contained), for example.
Ha ha, but I do know what the difference between applications and materials are. :-P If you look at the article thermoelectric materials, you will see that the topic is actually "materials for thermoelectric applications". Maybe I should rename it (again).
The point is that it's a list of articles in the heading of thermoelectric effect, and that someone interested in one may want to see all the other articles out there, since the text of the articles is still inadequate in many cases. The categories are less important. They could obviously be improved. "Applications (Sensing)" is especially bad, I agree with that. --Steve (talk) 07:27, 16 September 2010 (UTC)

(edit conflcit)

Wtshymanski, thank you for your constructive comments. I will add a note to the navbox talk page pointed here so perhaps they can be addressed.
Kbrose, I think you make some very good points, particularly the subjective impression created by navboxes (or anything else on WP for that matter). Obviously there may be many audiences for any given article, and it is probably impossible to please them all equally. Perhaps the jury is still out on the "best" way to organize information ("see also", project pages, navboxes, categories... etc.) but I think its positive for people to experiment, try things out, see what works, until a consensus is formed. I would think that the optimal way to organize things will vary among topic areas and this may have to be found by trial and error / iteration. David Hollman (Talk) 07:29, 16 September 2010 (UTC)
Have a look at Wikipedia:Disinfoboxes, too. --Wtshymanski (talk) 19:01, 16 September 2010 (UTC)

Some references?[edit]

This page seems to be very lacking in references, which is a shame as there is a lot of real-world information on it. Maybe the original contributors might wish to share their sources? (talk) 23:46, 27 October 2010 (UTC)

I agree. For example, user just corrected the maximum temperature for a Type J thermocouple from 700 C to 750 C. If there had been a reference to it is likely that the original error would not have occurred.

Here is a list of the Omega thermocouple reference tables.

Adding these as references and double-checking the Wikipedia page against them might be helpful. I am on a hot project right now, but if nobody else does it I will try to add the references and double check the figures n a week or two. = Thermocouple Type B (degrees C) = Thermocouple Type B (degrees F) = Thermocouple Type C (degrees C) = Thermocouple Type C (degrees F) = Thermocouple Type E (degrees C) = Thermocouple Type E (degrees F) = Thermocouple Type J (degrees C) = Thermocouple Type J (degrees F) = Thermocouple Type K (degrees C) = Thermocouple Type K (degrees F) = Thermocouple Type N (degrees C) = Thermocouple Type N (degrees F) = Thermocouple Type R (degrees C) = Thermocouple Type R (degrees F) = Thermocouple Type S (degrees C) = Thermocouple Type S (degrees F) = Thermocouple Type T (degrees C) = Thermocouple Type T (degrees F)

Guy Macon 14:46, 29 October 2010 (UTC)

What about the thermopower page?[edit]

When I hear the word thermocouple, I think about the device that is used to control temperature also Thermopower, where could a link to this page be put? — Preceding unsigned comment added by Scikris (talkcontribs) 00:23, 1 February 2012 (UTC)


Why the names or Types of thermocouples are random (K,E,S,B..), not serial(A,B,C,D..)? Is there any reason? Chinu giet08 (talk) 15:10, 10 March 2012 (UTC)

I don't know the history, but I think they were serial, and are listed now by their popularity (or other parameters, like alloying metal) rather than alpha order. Materialscientist (talk) 05:48, 11 March 2012 (UTC)

Insulation types[edit]

Note that copying tables like this [2] is not a copyright violation (such data are not copyrightable). I don't know whether this table fits in the article though. Cheers. Materialscientist (talk) 23:08, 5 December 2012 (UTC)

Are you sure? The page where it was copied from certainly says "© 2011 - 2012 Pelican Wire Company, Inc. All rights reserved." I would be surprised if someone spent a lot of time gathering all that information, and just because they published it in tabular form, there was no copyright protection for their company. Such information certainly needs wikifying though: what is the point of that column headed 'PART #' here? I'm sure there's something in WP:MoS about presenting info in ALL CAPS too. (Copied here from my talk page so that other editors here may comment too) --Nigelj (talk) 23:19, 5 December 2012 (UTC)

note on lead image[edit]


It's worth noting that the lead image at the present is a bit misleading, as the thermocouple is not being used at all to measure room temperature in that case. The reason I say that is that the multimeter is measuring basically zero volts out of the thermocouple, but it is performing cold junction compensation (using a second internal thermometer, based on thermistor or something like that). Thus the thermocouple has absolutely nothing to do with the temperature being displayed on the multimeter: I could short the inputs together using a simple copper wire and see the same reading.

It would be nice to have an image showing a multimeter measuring the temperature of something else which has a different temperature. --Nanite (talk) 22:21, 10 December 2013 (UTC)

The thermocouple is being used to measure room temperature. Assuming that the room temperature is indeed 19° Celsius, the type K thermocouple in use will provide an output of 0.758 mV of e.m.f. The junctions between the thermocouple wires/connector and the multimeter terminals will each provide an e.m.f. though as we do not know what the metallic composition of the terminals is, we cannot speculate as to the magnitude of these e.m.f.s. The cold junction compensation is designed to calibrate out these later two e.m.f.s leaving the meter to indicate the temperature represented by the e.m.f. of the thermocouple itself, in this case 19° Celsius. (talk) 14:59, 14 March 2014 (UTC)


Type-K Polynomial Coefficients 0–500 °C, 0.05 °C accuracy; for Tref = 0 °C[2]
n a_n
0 0 °C
1 25.08355 °C/mV
2 7.860106×10−2 °C/mV2
3 −2.503131×10−1 °C/mV3
4 8.315270×10−2 °C/mV4
5 −1.228034×10−2 °C/mV5
6 9.804036×10−4 °C/mV6
7 −4.413030×10−5 °C/mV7
8 1.057734×10−6 °C/mV8
9 −1.052755×10−8 °C/mV9

I removed the following table since it takes much space and adds very little to the article, especially since it only describes type K. And if we would have this kind of information for all thermocouples, the article would be full of wikitables. Anonimski (talk) 19:50, 4 January 2014 (UTC)


Thermoelectric voltage temperature difference[edit]

Example: I have a thermocouple with the hot junction at the temperature of 575 °C and cold junction at the temperature 550 °C. Let's say I measured the thermoelectric voltage U=0.1 mV. Now, I will take the same thermocouple with the hot junction on the temperature 50 °C and the cold junction on the temperature of 25 °C. The temperature difference in both cases is 25 °C. Will I measure the same voltage U=0.1 mV or not??? Is the thermoelectric voltage only a function of the temperature difference or because of the non-linearity of Seebeck coefficient I will get some other U?? Thanks. Tomáš. — Preceding unsigned comment added by (talk) 12:11, 25 September 2014 (UTC)

Temperature differential? Huh?[edit]

In this edit, the following text was added.

"where there is a temperature differential experienced by the different conductors"

This doesn't make sense. The voltage of a thermocouple is created at the junction of the two metals. The temperature at that junction is the same for both metals. There is no temperature differential. Heating or cooling the rest of the thermocouple wire has no effect on the voltage being output by the thermocouple junction otherwise all thermocouples would need to be housed in a temperature controlled environment to make them work. That is not the case. Either the above statement is wrong or I don't understand what the editor was trying to say. (talk) 17:28, 12 June 2015 (UTC)

You're right, both conductors experience the same temperature differential and their temperatures are identical at the junction. So the edit text could be misinterpreted. (By the way, the voltage doesn't arise at the junction, rather, it arises in the regions of temperature gradients within each of the conductors.) --Nanite (talk) 18:50, 12 June 2015 (UTC)
The voltage "arises" at the junction, not in the regions of temperature gradients within each of the conductors. The Seebeck effect in the metals causes an electron migration along the conductors but, by different amounts in the two different conductors. Heating or cooling the wires along their length indeed causes no measurable change in the thermo-electric voltage observed at the sensing end as points out. The junction must be near the temperature of interest not just anywhere along the wires.
Kckid (talk) 11:42, 13 November 2015 (UTC)