Talk:Electromotive force/Archive 3

Dubious definition

The "Formal definition" says "The EMF of a source (electromagnetic, chemical, thermal or otherwise) may be defined as the work done by an external agent, per unit charge, with sign reversed, in bringing a test charge once around a circuit that contains the source and no other source." What is this supposed to mean? In a circuit with a battery, the result would always be zero, unless there's a magnetically induced emf as well. Doesn't the emf of the source need to be defined in terms of its terminals, not a closed path? Dicklyon (talk) 19:03, 21 June 2009 (UTC)

The "seat of EMF" concept as presented is also dubious. All the sources I find apply this concept to the question of where exactly in an electrochemical cell the EMF comes from. They don't say that a battery is a seat of emf, but that it has one, somewhere, in it. Dicklyon (talk) 20:01, 21 June 2009 (UTC)

I taken the dubious bits out now. Please someone review Electromotive_force#Formal_definition_of_electromotive_force for consistency of definition, terminology, and sources. Dicklyon (talk) 20:24, 21 June 2009 (UTC)

I just realized that "Griffiths" was among the old-style ref listings, and that it's searchable on Amazon. The cited page for the loop integral (p.293) or actually the page before, introduces a distinction between different electric forces, such that the electrostatic part cancels out, but there's another part that doesn't, attributable to a source of emf. I don't understand this. Does anyone, who can explain? Dicklyon (talk) 23:09, 21 June 2009 (UTC)

The definition there helps: "the work done, per unit charge, by the source", and says that in some books emf is defined that way. It makes sense because it makes clear that "the source" needs to be identified, and that the emf is a property of the source, not of the closed loop, and that the loop integral doesn't count the work expending in pushing the current through the rest of the loop, for example. In the statement we had, "The EMF of a source (electromagnetic, chemical, thermal or otherwise) may be defined as the work done by an external agent, per unit charge, with sign reversed, in bringing a test charge once around a circuit that contains the source and no other source," the connection between "a source" and "an external agent" was not at all clear, but I see now what it was getting at. But as Griffiths points out, you don't really need to integrate around the loop, just across the source itself, if the source is not a loop (since $f_{s}$ is zero outside the source, it says). It also says "there's some subtlety involved", so it needs more study before we see if we can re-incorporate some of this into the article. Dicklyon (talk) 23:18, 21 June 2009 (UTC)

There could be a distinction in definitions between publications that treat EMF in terms of fundamental phenomenon, such as the net effect of electric fields, and publications that treat EMF in terms of simplified lumped-element circuit analysis. Those with formal university-level education in electricity know that lumped-element circuits are only an approximation, but this is seldom explicitly acknowledged in publications (and in some areas of study, never causes practical problems). --Jc3s5h (talk) 15:39, 22 June 2009 (UTC)

My understanding of "integration around a closed loop" is that the conservative part of the E-field makes no contribution to a closed loop, so all that is left is the EMF contribution. The whole point of EMF is to express how "external agencies" (for example, non-conservative fields and chemical sources) result in charge separation that, in turn, causes electrical potential difference. Its about converting energy in various forms to electrical form. Brews ohare (talk) 16:28, 22 June 2009 (UTC)

But doesn't that integral give zero even when there's a battery in the loop? Most of the sources I looked at had integrals from A to B; the ones with closed loops were all about magnetic effects (transformer emf) unless I missed something. Dicklyon (talk) 21:14, 22 June 2009 (UTC)

Self discharge

The sentence: "However, one might note that even in the open-circuit condition, where no current is drawn, self-discharge occurs due to various secondary reactions." has been deleted from the article.

This sentence with its Wiki link to the topic should be restored for these reasons:

(i) It is a topic that shows up in many discussions of EMF in batteries.

(ii) It is a strength of Wikipedia that a reader can find links to related material; the utility of Wikipedia is thus diminished by elimination of such links. Brews ohare (talk) 16:40, 22 June 2009 (UTC)

I can't really comment without seeing the context of the statement. However, in a general article about EMF, a statement that only applies in particular situations should not be made, lest readers think it is an inherent property of EMF. --Jc3s5h (talk) 16:49, 22 June 2009 (UTC)

The context describes the main chemical reactions occurring in a battery that cause charge separation, and so this sentence serves mainly to caution the reader that other reactions are present as well, and have observable everyday effects. Brews ohare (talk) 16:58, 22 June 2009 (UTC)

The connect to the topic appeared to be tenuous, esp. lacking a source that would make that connection. Putting it back with reference to source discussing self-discharge in the context of battery emf seems like a good idea. Dicklyon (talk) 21:16, 22 June 2009 (UTC)

Formal definition of electromotive force

I find the replacement of the formal definition originally present:

"The EMF of a source (electromagnetic, chemical, thermal or otherwise) may be defined as the work done by an external agent, per unit charge, with sign reversed, in bringing a test charge once around a circuit that contains the source and no other source."

is not equivalent to the present definition:

"The emf along a path between two points A and B is the integral of the electric field aligned with the path."

The problem with the present definition is that the conservative portion of the electric field contributes to this definition, and is not part of the EMF. That is why Griffiths and others use a closed path.

The closed-path definition used later in the same section contradicts the leading formal definition. Brews ohare (talk) 16:50, 22 June 2009 (UTC)

That's exactly what I'm confused about, too. It's not clear to me how the the later one contradicts the first one, but it seems likely that at least in some cases they are incompatible, like for a resistive loop with magnetic induction in it. Definitions in the literature are not mutually consistent. That's why I added the quote in the lead, so that the reader will at least know that there are several conceptions that go under the name emf. If we can include several sourced definitions, even if not mutually consistent, and say how they relate to each other or to various points of view, that might be a good way to increase the value of this article. Or if there's a dominant or standard definition that we can base the article on, and explain variations relative to that, that might be OK, too. So far, I haven't found a description that I'm able to interpret sensibly for both the chemical and transformer types of emf. Dicklyon (talk) 21:23, 22 June 2009 (UTC)

Transformer emf

Transformer emf is also an agency that can cause charge separation, as evidenced in Faraday's law of induction that includes both motional and transformer emf. In transformer emf, for instance, a coil in open circuit configuration subject to a ∂B / ∂t sees a nonconservative E-field (one with a curl) that drives electrons to one end, and leaves a positive charge at the other end of the coil. This is the charge separation that results in a conservative E-field that drives current if the coil is attached to a resistor.

Also, if one integrates around a closed curve to find the work on a test charge, the conservative E-field produces a zero contribution because it contributes the same single-valued electrical potential at the start and at the end of the closed path, while the non-conservative E-field will contribute the negative of the transformer emf. Brews ohare (talk) 20:12, 22 June 2009 (UTC)

Those explanations make sense individually. But in a loop that's not open circuited, there's no charge separation, right? Say a loop of resistors. What does it mean to have an open-circuit loop that you integrate around anyway? And how does the loop integral work with a battery in a circuit with a resistor? Isn't the loop integral zero? Or how do you determine what to integrate in that case? I understand the concept that if you're doing the loop integral you want to separate the energy input mechanisms from the energy output or dissipation mechanisms, but what definition makes that work? Dicklyon (talk) 21:29, 22 June 2009 (UTC)

Let's try a motional emf. A segment of wire moves in a fixed B-field. The Lorentz force produces an emf, and charge separation occurs. If we imagine the build-up of this charge as the motion begins, it steadily increases until the conservative electric field from the charge at the ends of the wire is large enough that the eE force on the electrons inside the wire balances the Lorentz ev × B force. At that point we have the open-circuit voltage at the ends of the wire. If now a resistor is attached across the ends of the wire, a current flows. That current is attempting to restore the charge neutrality condition in the wire. As charge leaves one end of the wire to travel through the resistor it has to be replaced, or the wire ends return to charge neutrality. The motion of the wire tends to maintain the charge difference, but due to the "internal resistance" of the source, the actual voltage between the ends of the wire drops a bit. How much it drops is clearly a balance between how much current the emf can supply and how much the resistor takes away. So, for example, if the moving conductor is a cheese grater, a different amount of current is supplied by the motion of the cheese grater than if it were a solid copper plate. Perhaps the literature on generators or motors has such a discussion somewhere? Brews ohare (talk) 23:29, 22 June 2009 (UTC)

Let's try a battery: initially a chemical reaction occurs driven by the lower energy of the charge separated ionic products. The charge separation creates a E-field, and this field grows until it is no longer energetically favorable for the reaction to continue because the work done against theE-field is equal to the energy reduction due to the reaction.Then the reaction is arrested, except for self-discharge effects. Then we have the open-circuit voltage at the terminals. If a resistor is placed across the terminals, a current flows trying to re-establish charge neutrality on each terminal. The steady-state voltage across the terminals occurs at a value that makes the current supplied by conversion of the reactants equal the current through the resistor. Brews ohare (talk) 23:40, 22 June 2009 (UTC)

Yes, that's all clear, but doesn't answer my question. If we take emf to be a two-terminal path integral, it's OK. But in the loop integral, how do you ever get other than zero for a circuit with a battery? Some authors evidently try to separate fields into parts that associated with emf and parts that are not, but I don't see you doing that. So if the emf in the loop is zero, like KVL, that's OK, but then we still need the two-terminal path integral to say what the emf of the battery is, no? My point is that the loop integral can't be the whole story -- and in many sources, it's not, as they rely primarily on the A to B integral. Dicklyon (talk) 00:03, 23 June 2009 (UTC)

I'd say that within circuit theory you always get zero. The emf doesn't show up but the voltage generated by the emf does. That is how Kirchhoff's laws work: the sum of the IR drops = sum of the voltage from voltage sources. What you are asking, it seems to me, is how do you calculate what the voltage provided by a source of emf is. For a battery that seems to be an exercise at bottom in electrochemical reactions and thermodynamics. For a loop moving in an electromagnetic field, it can be done by using all the Maxwell equations and not just a macroscopic subset like Kirchhoff's laws. Is that the point? Brews ohare (talk) 00:23, 23 June 2009 (UTC)

Yes, sort of; but how to calculate is probably easy, once we have an adequate definition that covers those cases clearly. Dicklyon (talk) 06:52, 23 June 2009 (UTC)

That version of KVL seems like a tautological way to avoid defining emf or sources; I see the KVL article has a patch for transformer emf, but it doesn't say anything about batteries. Of course, that could be because of my revert] of an unfamiliar and unsourced version of the law that sounded like nonsense at the time, when I didn't know people had all these conceptions of emf different from voltage. Dicklyon (talk) 07:02, 23 June 2009 (UTC)

some sources

Sources provided by Brews ohare; comments by others:

• charge separation and emf
• charge separation and emf
• charge separation and emf
• moving rod in magnetic field – good stuff on magnetic emf, yet the indirect definition on p.795 seems to exclude magnetic induction. p.898 introduces "induced emf" without definition. Dicklyon (talk) 16:47, 23 June 2009 (UTC)
• emf as energy conversion – same book as above, the p.795 that I commented on. Dicklyon (talk) 16:55, 23 June 2009 (UTC)
• emf is not voltage – this page acknowledges imprecise use of the term, applies a definition that excludes the magnetic or tranformer emf, and states that confusion does not result; I don't see the word "voltage" on that page. Dicklyon (talk) 16:45, 23 June 2009 (UTC)
• emf as non-electrical energy
• emf as non-electrical energy
• emf and line integral
• emf and contact potential
• emf and contact potential
• emf and contact potential

Brews ohare (talk) 04:22, 23 June 2009 (UTC)

Would you say that these represent a variety of viewpoints on emf? Or a consistent viewpoint? If the former, we should attempt to identify a few to report, and if the latter, we should figure out a good definition that represents both the battery and the closed-loop induction cases. I've looked at a lot of sources already, but am at a bit of a loss to interpret it all. Dicklyon (talk) 06:51, 23 June 2009 (UTC)

Ross quotation

"To some authors it is synonymous with 'voltage.' To others it means the open-circuit voltage of a battery. To a third group of authors it means the open-circuit voltage of any two-terminal device. This use is met most often in connection with Thevenin's theorem in circuit theory. To a fourth group it means the work accounted for by agencies other than differences of the (not measurable) Galvani potentials. Such authors equate the current–resistance product of a circuit branch to the sum of voltage plus e.m.f. A fifth group extends this use to field theory. The authors of this group equate the product of current density and resistivity to the sum of electric-field strength plus an e.m.f. gradient. A sixth group applies the term to electromagnetic induction. These authors define e.m.f. as the spatial line integral of the electric-field strength taken over a complete loop. To them the term 'counter e.m.f.' means something. We therefore think it advisable to avoid the term e.m.f. altogether."

The author's conclusion here obviously is not going to happen: emf is a very strongly established term and is not going to go away. Besides, it is a useful concept. I will argue below that this author is making too big a deal about controversy, and that there is no real division in interpretations of emf. He lists 6 usages:

1. emf = voltage; this is just sloppy usage that crops up because the potential difference generated by a source of emf happens to have the same numerical value as the emf. Voltage refers to a potential difference, and is due to the conservative portion of an E-field. On the other hand, emf is an expression of conversion of energy from some other form (e.g. chemical bonds) to electrical form as a potential difference caused by separated charges. See #some_sources. This confusion in terms was noted in an earlier version of the Wiki article, but this caution was removed as conjectural.
2. emf=open-circuit voltage of a battery; general usage does not restrict emf to apply only to a battery, and even within this restriction, no-one restricts it to the open-circuit voltage. Also, it isn't a voltage.
3. emf = open-circuit voltage of any two-terminal device; not a common case, and emf is not a voltage
4. emf=an extra term in Kirchhoff's voltage law: never heard of such a thing. Kirchhoff's law is an expression of the conservative E-field component.
5. emf as emf gradient: this terminology shows up in the theory of semiconductor devices like diodes, for example. In this context it amounts to the gradient of the quasi-Fermi level. The quasi-Fermi level notion is an attempt to incorporate the non-electrical forces driving a current due to variations in material composition. For example, the built-in voltage in a pn-junction is said to be due to emf differences, the same idea as the charge-separation notion in the customary definition of emf.
6. emf = electromagnetic induction: this is a reference to Faraday's law of induction, which is not a different definition, just a different source of emf

I'd say this author is (i) making a mountain out of a mole hill (ii) has confused sources of emf with emf itself (iii) confuses voltage with emf and (iv) actually has omitted the most common and accepted definition of emf, namely the one that originally appeared in the article:

Electromotive force, or more commonly, an EMF, is a term used to characterize electrical devices that supply electrical energy to a circuit, such as voltaic cells, thermoelectric devices, solar cells, electrical generators and transformers.[1] The EMF of a device is the energy per unit charge provided by that device to the circuit.[2] Thus, for a given device, if an electric charge Q passes through that device, and gains an energy W, the net EMF for that device is the energy gained per unit charge, or W/Q. EMF has SI units of volts, or joules per coulomb.

I would recommend deletion of the misleading Sydney Ross quotation and reversion to the original definition or perhaps some more directly worded version from #some sources. The Sydney Ross quote is only a note to an historical essay by a non-scientist, is not supported by citations or discussion, but only the author's say-so, and is inaccurate. If the supporting citations for the original definition are deemed insufficient, more are found on this talk page at #some_sources. Brews ohare (talk) 13:26, 23 June 2009 (UTC)

Brews, the work you've undone was all based on sources. You can't just remove sourced stuff and replace it with a different interpretation without telling us the source. The discussion above was left hanging, without answers to the questions or sources for the answers. You're asserting a particular POV, not even well defined yet as far as I can tell, and calling sources that dissagree confused. You add orignally added the statement that "Occasionally, EMF is confused with the electrical voltage that it generates" without a clue as to why you said so. The quote from Ross was what I found when I looked for what was behind this statement. Ross rings true, because when I search sources I find a variety of interpretations and definitions, not a uniform POV based on an accepted definition. If I've got this wrong, show we which sources you consider to be definitive, and we can at least mention the others as alternative POVs, and then when we're clear on what the POVs are we can work on figuring out how prominent each is. You've embarked on a string of controversial edits. Slow down and work on getting a consensus, we don't have to reset back to a sourced state. Dicklyon (talk) 15:00, 23 June 2009 (UTC)

In particular, the opening definition, "the external work that must be expended to produce an electric potential difference" doesn't apply to the resistive loop where there are no potential differences, does it? And "The electric potential difference is created by the external agency by separating positive and negative charges, thereby generating an electric field" is also inapplicable in the case of a loop with equal potential and charge density all the way around it. This edit "delete confusing material" removes the move common formula and definition that I found in sources, and didn't say why it's confusing. So let's discuss please. Dicklyon (talk) 15:05, 23 June 2009 (UTC)

Dick: I've removed the Ross quote for reasons given: please indicate what specifically you object to in this list of reasons. I also eliminated the suggestion that the integral of E between two points in an open loop is emf. That is simply wrong (only the non-conservative E contributes) and contradicts other things in the article. I'll discuss it further with you if you like, although I believe the discussion already given demonstrates that fact conclusively and it is sourced.

The definition is not mine, it came from the source. Can you flesh out the "resistive loop" example for me? - I don't get it. Likewise the "loop with equal potential and charge density all the way around it" looks like an exception, but I cannot imagine what it's about. Brews ohare (talk) 15:18, 23 June 2009 (UTC)

It seems that you either do not accept the definition in terms of an agency causing charge separation, or think it is one of many. You have suggested at some stage that Faraday's law of induction is an exception to this definition, and I have explained that it is not. Do you accept that there is no exception here? I am highly skeptical that you can actually come up with any sourced example system that does not fit the definition, although you may be able to find some verbal expression of opinion that is simply the conjecture of a misguided soul like Ross. Brews ohare (talk) 15:30, 23 June 2009 (UTC)

Under some definitions, the integral of E from A to B is not emf; in others it is. Ross is the only source I've found that has surveyed the usage of emf in primary and secondary sources and did an analysis of its various meanings. Why would you dismiss this analysis out of hand? The reasons you give are that you disagree with some of the usages he found; yet he found them, he says, and why would you doubt that? I find them, too (maybe not all 6), which is why it's so hard to write a definition that works for everything. I understand where you're coming from, the emf is just that voltage that represents externally applied energy, but that's not the only definition that's common in sources, and isn't even a definition that has been made clear anywhere, as far as I can find. And you keep saying "emf is not a voltage"; what does that mean, for a quantity measured is volts? Dicklyon (talk) 15:41, 23 June 2009 (UTC)

Hi Dick: I have given my reasons, and asked you to point out specifically where you disagree. Please do that. Please provide me with links to alternative definitions that support Ross's confusion. In doing so, please supply example systems not mere comment, and please avoid the 19th century, which is what Ross is reporting on. (As you know, the notion of emf used by Maxwell, e.g., is not the modern view.)

Your view that Ross "has surveyed the usage of emf in primary and secondary sources and did an analysis of its various meanings" doesn't seem to apply to the cited work by Ross, where no such discussion occurs in his note on emf, and he says explicitly that he has deliberately avoided the use of the term emf throughout his article.

The view that "emf is not a voltage" can be found in #some sources, and is consistent with the sourced definition. As you know any energy divided by e becomes volts. There is no reason to think that means anything measured in volts is a voltage (that is, an electrical potential difference).

Again, please answer my questions about examples that are exceptions to the sourced definition I've cited. Brews ohare (talk) 15:54, 23 June 2009 (UTC)

Here is a source written for translators, specifically describing different uses of the term emf, some of which are voltages. It specifically states that the usage in electrochemistry is a different one from the use in electromagnetics. I'm not saying it's impossible to unify these, but this scholar, at least, didn't see it that way. I was not familiar with the distinction you asserted that, that "voltage" always means "electrical potential difference", but that may be so; if so, it seems to conflict with the definition you provided in the lead, "external work that must be expended to produce an electric potential difference" which appears to define emf as a voltage by that definition. Dicklyon (talk) 16:38, 23 June 2009 (UTC)

Well, Dick you now have an ally in Crispmuncher. Therefore, rather than battle against these odds, I will retire from this effort with these comments to your last communication:

I think you should interpret his intervention as basically a procedural objection, not a support for my content position. If you can work with us via normal procedures, your help will be welcome. But if you won't, then retiring from this article is the right move. As he mentioned, he's had trouble with your procedure before; so have I, and I'm not going to sit idle while you do to more articles what you did to speed of light and others. Dicklyon (talk) 17:24, 23 June 2009 (UTC)

1. It is not only "possible to unify" the E&M definition taken from Faraday's law of induction, this unification has been outlined specifically in the article (at least in some of its forms) and in discussion with you. Your continuing failure to address this point despite requests to do so is peculiar.

Can you say what source supports that unification clearly? I've been doing my best address exactly this point, but you have been failing to follow up my questions, probably because things are just moving too fast. Dicklyon (talk) 17:24, 23 June 2009 (UTC)

2. There is no doubt that some mechanisms (e.g. a battery or an electrical generator) can creates a charge separation that also creates an electrical potential difference. That does not mean the agency creating the separation is a voltage; it means the voltage is the result of the charge separation. That appears to me to be straightforward logic. However, if you doubt it, you can break down and consult the references in #some sources. Brews ohare (talk) 16:57, 23 June 2009 (UTC)

It's unclear to me what this means: agency is not a voltage. I have not introduced the notion of an agency here; what source does that? I think I get your point that the emf represents the work done on charges in a circuit, from external agencies; what I don't see is a good source that explains that in a way that answers my questions about both batteries and magnetics, without tautology, and works in general; that's why I keep asking. What I do see is lots of sources with other interpretations, including tertiary sources explaining that there exist other interpretations. Let's try to represent sources fairly and clearly, so others will be able to verify that what we write is true, or at least backed up by reliable sources, rather than idiosyncratic interpretation. Dicklyon (talk) 17:24, 23 June 2009 (UTC)

It is not for him to justify his actions. He is not the one removing references in favour of unsourced statements, or removing details of disputes and disagreements in favour on one particular arbitrary view. You might want to consider your general editing style: I first encountered you recently on speed of light and now I see that you are doing much the same thing all over the physics articles. Think carefully before making any changes. Don't read something, make some changes and then discuss whether you have understood it correctly. Get the groundwork done beforehand or else even legitimate improvements to an aricle are likely to be reverted. I'll remind you you have already received a warnign for disputive diting - take that on board.

In regards the substance of this article you seem to find abhorent the equation of voltage and emf. Although I know that many people do this, personally I don't disagree with you - it is akin to stating that length and metres are the same thing. However in your recent edits (which I reverted) you seem to equate work and force. That is not really up for debate as they are clearly different. If you understand this then in debating all and sundry on multiple articles at once you inevitably lose precision. Another reason to slow down in favour of more considered edits rather than incoprating whatever it is you have just read on Google with no selectivity whatsoever. CrispMuncher (talk) 16:44, 23 June 2009 (UTC)

To conclude that work and force is equated in what you deleted is just sloppy reading. Please show specifically how you come to that conclusion, which is not implied by any logic that I can think of. Bear in mind that all this material is sourced by citations to textbooks, so be careful. Brews ohare (talk) 16:56, 23 June 2009 (UTC)

Reversion without comment to erroneous conception of emf

Crispmuncher: The material you restored includes statements that are at variance with the literature as found in the section #some sources. How about defending your views on the Talk page instead of intervening in an uniformed manner on the article page? Please provide any exceptions you know of to the very conventional and sourced definition I provided, viz:

In physics, electromotive force, emf (seldom capitalized), or electromotance is the external work that must be expended to produce an electric potential difference. The electric potential difference is created by the external agency by separating positive and negative charges, thereby generating an electric field.

Brews ohare (talk) 16:33, 23 June 2009 (UTC)

BTW you sloppily also failed to proof-read your changes which now include a hodge-podge of undefined notation for the line integral. Brews ohare (talk) 16:41, 23 June 2009 (UTC)

Did it ever occur to you that I was in the process of writing some comments here before you instantly reverted and cried that I had not replied? A detailed rebuttal takes time. This is why I eventually gave up at speed of light. Indeed I have now had to enter all this text three' times as a result of your rapid-fire editing. CrispMuncher (talk) 16:44, 23 June 2009 (UTC)

C, thanks your your help. By the way, when you get an edit conflict, you can usually just copy your text from the lower edit window into the right place in the upper window and save; no need to retype. When you have several comments interspersed it's more complicated, but you can at least copy and save what you've typed, like I just had to do... Dicklyon (talk) 17:02, 23 June 2009 (UTC)

I made no reversion whatsoever. The article remains erroneous and in contradiction with sources at #some sources. The article retains the sloppy editing errors introduced by hasty reversion by Crispmuncher and his/her failure to proof read his/her changes. Brews ohare (talk) 17:40, 23 June 2009 (UTC)

The definitions in the sources are vague and contradictory. For example, one definition says "the external work that must be expended to produce an electric potential difference." So if the emf source is a secondary battery, is it the "external work" the reduction in stored chemical energy? Is it an artificial concept where one models the battery as a Thévenin equivalent circuit and the external work includes both the energy dissipated in the circuit-under-test, and in R_th? Or perhaps it is the pro-rata share of the energy that was used to charge the secondary battery. Some of the sources describe the emf from a transformer winding, but does not seem to include energy dissipated in the resistance of the winding.

I suspect when all is said and done, emf will turn out to be a vague concept for which there is no single rigorous definition. --Jc3s5h (talk) 16:55, 23 June 2009 (UTC)

That seems to be what many of the sources say; I suspect there's also a crisp concept here, corresponding to what Brews has in mind, if we can find an adequate definition and sourcing; there's no reason we can't represent both of these situations, and/or more. Dicklyon (talk) 17:02, 23 June 2009 (UTC)

I do not agree that the sources are vague and contradictory. It seems pretty clear to me that if you separate two opposite charges, work must be done. If that work is done by motion in a magnetic field, or by chemical rearrangement of some ions in a molecule, or by charge transfer across a pn-diode, that agency is an emf that results in conversion of some energy stored one way into energy stored in the separation of charges. There is nothing vague about this. Brews ohare (talk) 17:07, 23 June 2009 (UTC)

So now emf is an agency? What about the resistive loop in magnetic field? There's no charge separation, no variation in charge density, no variation in potential. What definition of emf makes it work for this case? Conceptually it's clear that the emf is equal to the voltage that would drive that amount of current in the loop; but what source provides a definition that leads to the right answer here? Dicklyon (talk) 17:14, 23 June 2009 (UTC)

So an agency creates an emf. Do we need a lawyer here? Brews ohare (talk) 21:05, 23 June 2009 (UTC)

EMF in closed resistive loop

I'd be happy to engage you here, but I'm unsure of the set-up. Am I to move a closed wire loop in a fixed B-field, for example? That is an example of Faraday's law: an emf is generated according to his law. The current that flows as a result of the emf depends of course upon the resistance of the loop.

The example you take can be viewed as a limiting case, I suppose. We can break the loop, obtain a charge separation, find the open circuit voltage. Then we reduce the resistance of the break from infinity to a finite value. The current that flows depends upon the resistance of the break. Of course, this current also flows in the loop, and so generates a voltage drop (a case of what is called internal resistance). That drop reduces the voltage the loop places across the break. The steady-state voltage across the break is the one that results in the same voltage across the break as the open-circuit voltage less the internal IR drop in the loop itself.

That example seems not to cause any difficulties, do you agree? Brews ohare (talk) 17:29, 23 June 2009 (UTC)

By a resistive loop I meant a physical "wire" with distributed resistance, such that the potential is everywhere equal. I can see how one could make a lumped-circuit approximation, breaking the loop into point resistances and extended conductors, such that the voltages induced across the wires would be dropped across the resistors; this would allow separating emfs from voltage drops, measuring each with voltmeters, etc. But the question is this: what sourced definition of emf applies to this (distributed, not lumped) situation? Dicklyon (talk) 17:35, 23 June 2009 (UTC)

I believe the example discussed above fits the description of a "physical wire with distributed resistance". The loop in the example can have any resistance distribution along its length one might like, and the "break" can be tiny. In the limit that the R of the break approaches the R one would like, a closed loop of arbitrary resistance variation results. A lumped circuit approximation is not necessary, although I think such an approximation still would establish the point: the EMF is that of Faraday's law, which applies to any closed loop regardless of its resistance. The EMF of Faraday's law is therefore the same as that of the loop with a tiny break of infinite resistance. In other words, in open circuit, the EMF of Faraday's law is opposed exactly by the open circuit voltage due to charge separation. (That is why no current flows in the open-circuit case.) In closed circuit, the EMF of Faraday's law is reduced by the internal voltage drop due to internal resistance of the loop. Brews ohare (talk) 17:48, 23 June 2009 (UTC)

Quotes about emf

Some ideas at variance with the introduction to the present article: they all say emf causes a voltage difference, and is not itself a voltage difference.

• Cook "Although the definition of ε is very similar to the integral giving a static potential difference ΔV about a closed path and both quantities are measured in volts, ε and ΔV are physically very different; ε is in general not equal to zero and hence arises from nonconservative force fields while ΔV us necessarily always zero and corresponds to a conservative force field." This source oddly is cited as support in the introduction when it actually contradicts the introduction by saying emf is not voltage.

I had actually added that source on the opening sentence in support of the alternative term electromotance, and hadn't really looked at what else it said yet. Now that I look at it, it is the closest I've seen to a workable complete definition to support your preferred definition, as it includes specifically the Faraday induction term and some terms (not very well defined) related to electrochemical and thermal processes. I think we can use this; not that it's the only definition, but it's as close as I've seen to the definition you would want, yes? Dicklyon (talk) 21:39, 23 June 2009 (UTC)

Take a shot at it , Dick. It might work. Some later discussion also needs revision to fit this change. Brews ohare (talk) 22:25, 23 June 2009 (UTC)

• MacDonald "the 'battery-like' EMF must have pushed positive charges in the horizon toward the equator and negative charges toward the pole, until the charge separation created sufficient electric field of its own to counteract the EMF ..."
• Lerner "An electric potential difference is produced by a system that does external work on electric charge. ... The external work per unit charge that must be expended to produce an increase ΔV in the electric potential of the charge is called the electromotive force.
• Halpern & Erlbach "To maintain the steady flow of charges therefore requires an external source of energy that in effect takes charges leaving one end and brings them back to the other... The external source, in effect, maintains a net positive-negative charge separation between the two ends of the conductor, which of course is what causes the potential difference to be maintained. ... The energy per unit charge supplied by the external source in maintaining the voltage is called the EMF ("electromotive force"), and it is the EMF that replenishes the electrical energy lost as the charges flow within the conductor."
• Bhatnagar "In order to maintain continuous flow of current, the positive charge which arrives at the negative terminal must be pushed ... to the positive terminal by some non-electrical forces acting within the seat of emf."
• MeenaKumari "Sources of emf include electric generators, batteries, fuel cells etc., which all convert non-electrical energy into electrical energy." Brews ohare (talk) 18:43, 23 June 2009 (UTC)

I contend that these sources clearly override the unsupported opinion of an historian (discussing the 19th century concepts of Volta) following his personal and undocumented review of whatever (uncited) literature on emf he may have run across. He says so far as he can tell the term is so muddy nobody should use it, which might be taken as fact, or as simply an admission that he never did get the concept. Certainly many modern authors use the concept and have a shared understanding of its meaning. Brews ohare (talk) 19:01, 23 June 2009 (UTC)

Actually, I don't think it can be over-ridden. I can find sources to support at least several of the interpretations he cites; just because there are a lot that do it one way, doesn't mean there aren't other ways. But I'm willing to strike a balance. I don't think we can say "anyone who calls an emf a voltage is simply confused" or anything to that effect; instead, let's state what the alternative interpretations exist, and which main interpretation the article will focus on. It doesn't have to be hard. Dicklyon (talk) 01:23, 24 June 2009 (UTC)

Some others, that at least illustrate that some authors think emf is a voltage: Dicklyon (talk) 02:16, 24 June 2009 (UTC)

• Croft Electromotive force, abbreviated e.m.f., and sometimes called voltage, electric pressure, or difference of potential, is used to designate the "push" that moves or tends to move electrons... (if it's "sometimes called voltage", is he saying it's a voltage? is he saying it's a difference of potential, or is he just saying it's sometimes called that? this book also has lots of illustrations of hydraulic analogy, which someone was looking for for the Ohm's law article.)

• Karmel et al. section ELECTROMOTIVE FORCE (EMF) OR VOLTAGE really does treat the terms as equivalent

• Hann the term orignally stood for electro-motive force and implies the voltage generated by... (definitely he's saying it's a voltage; but not that all voltages are emfs) and a different meaning of emf occurs in the field of Electro-Chemistry (but he doesn't really explain, so may actually be confused in this case) and despite its rather nebulous definition and apparent double-meaning, English-speaking technologists are reluctant to discard the term and emf in favor of simply voltage, ... (gives silly reason).

• Dorf induced voltage or emf (this phrase with which he labels a formulaic result can hardly be consistent with saying that emf is not a voltage)

• Edmonds has a clear definition for the magnetically induced emf: The line integral of the electric field around a given path is called the electromotive force or emf. He later notes it is similar to a voltage change but doesn't say it is one. And as soon as he mentions battery he stops talking about emf, maybe because he realizes that his definition doesn't work for a battery-driven circuit.

Hi Dick: Dorf does not really support the view that voltage=emf; he uses a closed path and says induced voltage = emf, by which he means the voltage from Faraday's law of induction. Likewise, Hann refers to a "near" synonym and says it implies " the voltage generated by a particular chemical or magnetic arrangement." Croft (1917) refers to emf apparently as either voltage or emf from Faraday's law, depending upon the example discussed. Hence, he doesn't really come down on either side. Likewise, Karmel uses a line integral between two points and says "in general the voltage depends upon the path chosen" and later makes the inconsistent statement "In later chapters we will discover the different sources of emf other than batteries which provide potential difference between their terminals as a result of internal chemical reaction." and adds "Eq. 3,21 is 'general enough' (my quotes) in conventional electric circuits." The fact is, you cannot get a battery emf from a line integral through the battery because taking a charge through a battery involves a problem in quantum statistical mechanics to find how the energy of the entire ionic-electronic system changes with position of the charge, and calling that a line integral is sweeping a lot under the rug.

As for Edmonds, I'd say I agree with him that Faraday's law produces emf, and I'd say we have sources that say this case is not different from a battery: an external agency creates a charge separation and the work done in doing so is the emf.

So what have we got here? We have two possible supports for the emf=voltage idea, one rather ambivalent about it (and rather dated) and one (I'd say) far from clear. Neither is a straightforward endorsement of the emf = voltage concept. Both might be explained as simply a bit sloppy. In the Wiki article I'd say that they have to be portrayed as a marginal and ambiguous view. Brews ohare (talk) 02:45, 24 June 2009 (UTC)

I think we can report it as "another view" without necessarily judging it as marginal or ambiguous; we can also balance with sources that specifically say emf is not a voltage. Dicklyon (talk) 03:33, 24 June 2009 (UTC)

Well, I suppose that it can be brought up, but this idea has no merit from either a calculation or intuitive viewpoint, as the idea of voltage handles the conservative part and the "separation of charge" emf handles the rest, leaving no role for the combined form except to increase the level of confusion and ambiguity. Brews ohare (talk) 05:33, 24 June 2009 (UTC)

Rojansky source

In the intro, Rojansky is cited as a supporting source for saying emf is the same as voltage. This source almost supports the article. It does equate emf to integral of E-field between two points, but I'd say that if one feels impelled to mention this view, it should be flagged as a minority opinion (possibly the opinion of one). I say that for these reasons: (i) This view is not compatible with Kirchhoff's law, except in special cases (namely the cases where the emf is zero). (ii) This view is not compatible with the "separation of charge" definition, which is the dominant view. (iii) Rojansky explicitly limits this discussion on p. 187 to the case where there actually is zero emf according to Faraday's law - he suggests a "redefinition" is necessary when a Faraday's law emf is present. (iv) Other sources of emf (besides Faraday's law) don't come up for discussion - since Faraday's law also is excluded by Rojansky himself, his definition of emf actually applies only where other authors would say there was no emf at all. Brews ohare (talk) 20:23, 23 June 2009 (UTC)

Actually what he says is that emfs "are also called voltages" which is not quite as strong as what you said he said. His open-circuit approach certainly does lead to a potential difference equal to the emf, right? But he refers to the electrostatic case as a special case and also introduces the loop integral; I presume the E in that includes magnetic effects, as opposed to being merely the potential gradient? Not clear. Anyway, I find an awful lot of such difficult-to-interpret sources. I'm not really to decide that some of them are just daft. I think it's more like there's no clear definition, so everyone makes up one that suits them, and they vary. Dicklyon (talk) 02:25, 24 June 2009 (UTC)

No, his open circuit approach mixes up the conservative and non-conservative contributions, except his caveats actually exclude any case where real emf's occur. He also disqualifies himself by saying his definition is not general and has to be modified to take Faraday's law into account. He also has not discussed batteries per se.

You may be anxious to adopt the "it's all confusing" stand point to avoid adopting a sharply focused definition that you are not yet confident in. That confidence can be increased by thinking about it more. Or perhaps we have to find an irrefutable source so you don't have to think about it.

The unsourced assertion of confusion ("Occasionally, EMF is confused with the electrical voltage that it generates...") was inserted into the article by you; I took it out. If confusion exists (which it might), then our job is to represent the points of view involved, whether they confuse or not. I'm certainly not looking to adopt an "it's all confusing" stance, just trying to understand the landscape well enough to report it accurately. I'd also like to report a sharply focused definition that works, but so far I'm not sure I've seen it. Your "formal definition" with the loop integral of the force per charge is pretty good if you can define what part of the force per charge it refers to; it seems to require more definitions to motivate a way to separate those forces due to external agencies from those not. How do you do that, e.g. how do you split up the fields or forces in a battery? I find the definitions rather lacking still. Your "vector field F represents the force per unit charge on a charge carrier" didn't even try, and left a definition that gives zero without magnetic effects, and therefore doesn't work for batteries, right? Dicklyon (talk) 03:26, 24 June 2009 (UTC)

If you insist upon presenting the view emf=voltage you are going to have to say that no-one really says it outright, but a small minority of sources appear to imply it is a possibility in some cases, and certainly it is not the generally accepted view. The use of a line integral over a line segment is to be avoided: a closed path is the way to go, and it is best restricted to the Faraday's law case because the line integral cannot be calculated through a battery. Brews ohare (talk) 02:49, 24 June 2009 (UTC)

My view is not emf=voltage. That's just one of the things in the literature that needs to be represented in the article. My personal views don't enter into this (though of course my understanding and interpetation of the literature does, but I'm pretty open minded there, as I feel like there's good stuff to be learned here, historically and usage wise).

I did not intend to say it was your view, but that it was a view you wanted to see represented. I shudder to hear there are other issues as well, but maybe this one can be settled first? Brews ohare (talk) 05:29, 24 June 2009 (UTC)

And why can't a line integral be calculated through a battery? Do you mean as opposed to a closed loop through a battery? Or what? Dicklyon (talk) 03:26, 24 June 2009 (UTC)

Well, I'm inclined to say something in words like "an external agency creates a charge separation and the work done in doing so is the emf", restrict the line integral discussion to the Faraday's law case where it is simple to evaluate because the Lorentz force is simple, and use thermodynamics for the battery. I don't know how to calculate the portion of a line integral of E·dl that sits inside the battery. Some macroscopic simplification like thermodynamics that can handle energy in multiple forms is more tractable, and seems to be the way the battery is most commonly handled. See the article on this. Brews ohare (talk) 05:29, 24 June 2009 (UTC)

The line integral may not be useful as a definition to calculate to, but conceptually it's no problem, I think. The problem comes when you have a circuit with current flowing, and the forces that are part of the emf may be hard to distinguish (via definition) from forces due to resistance (or electrostatics and Ohm's law). The definitions suggest that some part of the force, or field, is supplied by external agencies; but in a loop with e-field integrating to zero, how does one really define or determine what parts of the e-field are from external agencies and what parts aren't? Is just that you have to agree to agree that you know what part is external, part of emf, and what is not? I remain a bit confused about how to clarify, in a way that works for all kinds of emf. Dicklyon (talk) 06:22, 24 June 2009 (UTC)

Needing a break

Brews, take over for a while, now that you have heard my point. Try to keep one main interpretation in focus without excluding the others. And don't do so much at once that nobody can intervene and comment as it goes. Please. Dicklyon (talk) 06:17, 24 June 2009 (UTC)

Revision details

The introductory section contained a lot of repetition and some self-contradictory elements. The introduction now contains the most prevalent view of emf, and other views are summarized in the section Terminology.

The discussion of units is consolidated under the section Notation and units of measurement, which existed before, but was not the only place where this was discussed.

The line integral formulation has been extended using the formula of Cook in the section Formal definitions of emf. The usage of a line integral across a line segment has been deleted, as it contradicts the Cook formulation and is (as far as I know) advocated only by Rojansky, who has at best a minority view. (I'd say a useless view).

The quotation by Ross has been trimmed. As discussed on this Talk page, he rather exaggerates the importance of variations in usage.

Several sources have been added. Brews ohare (talk) 16:57, 24 June 2009 (UTC)

So you punted on the opening definition, making it an open-circuit definition instead of the loop integral definition? Too hard to find a unifying definition? Dicklyon (talk) 03:47, 25 June 2009 (UTC)

I regard the "charge separation" definition as the guts of the matter, and the closed loop integration as a follow-up theorem or corollary. It is presented under Formal definition, which may not be the best title. Brews ohare (talk) 09:42, 25 June 2009 (UTC)

Why the italics on emf? Dicklyon (talk) 06:42, 25 June 2009 (UTC)

There was a mixture of emf and emf so I made them all the same. I guess I felt emf emphasized that it was not a word, but an acronym. Brews ohare (talk) 09:42, 25 June 2009 (UTC)

Why the partial of the Sydney Ross quote, omitting his analysis of the different meanings in the literature, yet keeping his statement of one generalized meaning? Sort of defeats the purpose of quoting a historian. It seems a fair characterization of the variations that we encounter among the sources. Dicklyon (talk) 06:54, 25 June 2009 (UTC)

Ross quotation: reprise

I've made comments earlier about this quote. To summarize, the quote is based upon an aside by Ross to justify his avoidance of the term electromotive force, and is not a scholarly assessment of the history of the term. As such, I don't regard it as an historian's assessment, but as an off-the-cuff summary. You will notice that he does not cite any literature to support his statements, nor does he provide any discussion to indicate he has weighed the merits or importance of the different usages. In fact, I think it is a pot pourri he dredged up, and many of the usages are not significantly employed, and some are equivalent although he has listed them as distinct. The Wiki article does describe the main alternative usages in a balanced manner, and treats the Faraday's law and electrochemical emfs as examples of the same charge-separation definition, rather than following Ross and treating them as different. Quoting the full Ross remarks would then involve the article in historical debate, which I don't think is warranted as Ross does not support his views.

I added a number of sources that make different use of emf than the opining definition. Mostly they are from circuit theory, and are focused on solving circuit problems rather than carefully examining how emf and voltage are related.

I'd be happy to omit Ross altogether and use Graneau, the other source. Brews ohare (talk) 09:42, 25 June 2009 (UTC)

If you find someone to back up your "I don't like it" argument, I'll consider it. Until then, it's in – it's one of the few analysis of usages that we have, and appears at least to be a rather careful analysis of a wide range of literature. Dicklyon (talk) 04:54, 26 June 2009 (UTC)

Solar cells

The solar cell bit had some confusions, including in the lead, so I simplified the statement in the lead and worked on the description of how it works. It's not as different as it was saying. Dicklyon (talk) 08:03, 25 June 2009 (UTC)

I'm no expert on solar cells, and the solar cell article isn't much help. I'm somewhat concerned over whether this example opens a Pandora's box of emf sources that have no connection to the opening definition (or any of the others). The current is driven by diffusion, and the photo emf seems to be incidental. What do you think? Brews ohare (talk) 09:42, 25 June 2009 (UTC)

I am a bit of an expert on photodiodes, usually in reverse bias as a detector, but have passing familiarity with their usage as solar cells as well. There's some diffusion involved in places, like when the photon absorption occurs outside the depletion region, but I think that to say that's what drives the current is quite wrong; the source you had cited for that didn't appear to support it, so I took it out. Dicklyon (talk) 16:07, 25 June 2009 (UTC)

Hi Dick: I wonder if you noticed Eq. 6.17, which contains the diffusion lengths as determining the open circuit voltage of the photocell? That would seem to support the view that diffusion drives the current in this device. See also Short circuit current which appears also to determine the current in general. Brews ohare (talk) 16:27, 25 June 2009 (UTC)

Yes, I saw that. Long diffusion length increases the probability that the carriers will survive to separate, rather than recombine; but he certainly doesn't say that diffusion is what drives it. It's involved in part of the current flow, but is not what makes the emf. Dicklyon (talk) 16:47, 25 June 2009 (UTC)

In other forms of emf like batteries and generators, an open circuit condition is obtained when the developed emf creates a back voltage that arrests the continuation of charge separation. In the solar cell, light is continuously applied, and there is a photo emf. However, it seems that unlike the other cases, in open circuit conditions there is not an arresting of the generation mechanism (for example, light emission does not increase to counter the light absorption). Can you provide some discussion of this point and contrast it with the other cases? Brews ohare (talk) 19:30, 25 June 2009 (UTC)

It can be viewed as a voltage-dependent arrest of the action, perhaps, but more specifically it's a recombination via forward current, like a self-discharge path. Think of it as a current source in parallel with a diode. See solar cell#Equivalent circuit of a solar cell. Dicklyon (talk) 21:49, 25 June 2009 (UTC)

Brews, your change to the "unlike in batteries" statement is an improvement, as it's not as false as what it had said about the terminal voltage, but I'm not convinced that the effect you're trying to describe is really so unlike in batteries. Do you have a source for this contrast? Dicklyon (talk) 16:40, 26 June 2009 (UTC)

Hi DIck: BTW, have you noticed that photo emf appears to refer to the voltage introduced, rather than the work done in creating the charge separation. I'd guess the work/charge done in the charge separation is the photo-voltage because the rest of the photon energy just is dissipated?

Rojansky definition is his own, and is at variance with the remainder of the WIki article on emf

This is a very poor source, and presents a minority view of one that emf is given by a line integral over a segment of path.

${\mathcal {E}}=\int _{A}^{B}{\boldsymbol {E\cdot }}d{\boldsymbol {\ell }}\ ,$

All other sources use a closed path to eliminate the contribution of the conservative E-field. Rojansky's formula based upon a segment keeps the conservative E-field contribution. That may be viewed as simply an error, or as adoption by Rojansky of a definition that is neither potential difference (as done by some texts) nor emf (as defined by work done in charge separation), but the sum of the two. He directly acknowledges this fact in noting that his emf in general is path dependent. Nobody else does that.

The Rojansky equation above contradicts the immediately following equation in the WIki article:

${\mathcal {E}}=\oint _{C}{\boldsymbol {E\cdot }}d{\boldsymbol {\ell }}\ ,$

because the closed loop integral has zero contribution from the potential difference, which is zero for a closed loop. Brews ohare (talk) 10:12, 25 June 2009 (UTC)

I have replaced the misleading view of Rojansky with the correct but much more restricted use of a line integral inside a source of emf due to Griffiths. Brews ohare (talk) 14:53, 25 June 2009 (UTC)

If you define emf as in the lead, as an open-circuit voltage, then this electrostatic integral always gives the right answer, doesn't it? It's the emf associated with a two-terminal device, if evaluated at the terminals. I can find sources beside Rojansky if you like; it doesn't appear to me to be an outlier. Dicklyon (talk) 15:53, 25 June 2009 (UTC)

Dick: The lead-in versions I have used do not define emf as an open-circuit voltage, but as the work done per unit charge in creating such a voltage. The emphasis is upon non-electrical forces, and only then upon the resulting voltage difference due to charge separation. I have carefully extracted the relevant portion of Griffiths to replace the nonsense of Rojansky. Brews ohare (talk) 16:01, 25 June 2009 (UTC)

emf is not a voltage

The revision of 06:38, 25 June replaced:

A voltage difference is considered to be the result of an emf, in most usages of the term. The emf is typically considered to be the work done per unit charge in pushing charge through a battery against the battery's voltage difference, for example. However, it is common in circuit theory, for example, to refer to the voltage created by the emf itself as the emf.

with:

Not every voltage difference is considered to be an emf, in most usages of the term. The emf is typically considered to be the voltage created by a source, or the work done per unit charge in pushing charge through a battery against the battery's voltage difference, for example.

with the comment "try this minor adjustment instead".

In fact this is not a minor adjustment at all, but a complete change of meaning that contradicts the opening definition of the article. In fact the replacement text is itself internally inconsistent, defining emf as a voltage difference, and next saying it is the work done against the voltage difference. The definition adopted in the intro is that emf is the work done per unit charge in creating the voltage difference. The key point in emf is the conversion of energy from other forms to electrical form Brews ohare (talk) 10:24, 25 June 2009 (UTC)

A consequence of Brews ohare's argument is that any electrical device that (1) plugs into a household wall outlet and (2) does not contain, and is not connected to, any transformer, photo cell, or receiving antenna, should be analyzed without reference to the concept of emf. --Jc3s5h (talk) 14:26, 25 June 2009 (UTC)

I'd suggest that to do that analysis without emf, one uses the voltage difference between the prongs of the plug and Kirchhoff's laws (say). If one wishes to inquire as to the origin of the voltage difference across the prongs, then one must go into the details of the source of potential difference: is it a battery, generator or whatever. That is where emf comes up. Brews ohare (talk) 14:58, 25 June 2009 (UTC)

Yes, that is just how to do it. I wonder, though, whether the literature does that, or whether the literature describes the hypothetical voltage source used for analysis as an emf. I checked two circuit analysis books on my shelf, and no variation of the term "emf" is listed in either index. --Jc3s5h (talk) 15:20, 25 June 2009 (UTC)

It appears that circuits texts often use emf as interchangeable with voltage difference due to a battery or generator. That usage is pointed out in the Terminology section (or it used to be before Dicklyon's changes, I'm not sure now). Brews ohare (talk) 15:25, 25 June 2009 (UTC)

The statement referenced the Physics text of Halliday and Resnick, but I didn't see any support for it there, so I took it out. The interpretation of emf as a voltage is pretty widespread, and we should acknowledge it fairly. If you want to clarify, you can find a source that clarifies the relationship and talk about that, too. Dicklyon (talk) 15:55, 25 June 2009 (UTC)

Dick: Fig 28-4 in halliday and resnick indicates the emf Є as the voltage difference caused by the emf. You do not appear to have noticed that two or three other circuit references are cited in the Terminology section to support the notion that emf is used interchangeably with "voltage difference" in this field. Brews ohare (talk) 16:04, 25 June 2009 (UTC)

The chapter is called "Circuits" but the field of the writer is physics; it doesn't mention "circuit theory"; how does this say anything about what field this usage is common in? Dicklyon (talk) 16:34, 25 June 2009 (UTC)

Agreed, I haven't established it is a common usage, although this text is in its 6th edition and is a major player. However, there are many other sources cited in Terminology that make the point that halliday and resnick are not alone. Brews ohare (talk) 16:43, 25 June 2009 (UTC)

Of course they're not alone; this usage is widespread; my objection is to your saying that it a minority usage associated with the circuit theory field. Dicklyon (talk) 16:49, 25 June 2009 (UTC)

Do you have sources outside circuit theory (other than Rojansky, who uses his own definition of emf that is not defined by the endpoints A and B, but is path dependent)? Brews ohare (talk) 16:54, 25 June 2009 (UTC)

Collaboration style

Brews, your style of rapid revision, with dozens of consecutive edits that move in a direction for which there is no concensus and in many cases clear opposition, is making this process very difficult. I don't know what to do. Ideas? Dicklyon (talk) 16:30, 25 June 2009 (UTC)

I don't agree about lack of consensus, exactly. I think the major issue is whether you accept the Griffiths replacement for Rojansky at Electromotive_force#Formal_definitions_of_electromotive_force. Brews ohare (talk) 16:45, 25 June 2009 (UTC)

You've done 33 more consecutive edits to the article today since I complained about your style. I still think WP:CONSENSUS is important. Dicklyon (talk) 21:40, 25 June 2009 (UTC)

Hi Dick: Most of these edits are pretty uninteresting stuff. I think the biggy is the one above: the Griffiths material. It's sourced, it's accurate and it agrees with what else has been said. It does not agree with Rojansky's path-dependent emf definition. Brews ohare (talk) 22:27, 25 June 2009 (UTC)

Hi again Dick: Apparently you have no comment on the Griffiths substitution for Rojansky?? Brews ohare (talk) 06:06, 26 June 2009 (UTC)

Ross quotation again

Attempts to shore up this quote using google to find citations have been unsuccessful. All actually used variants in the use of emf are well documented in the Wiki article, showing Ross to have made some erroneous distinctions, and to have grossly exaggerated the murkiness of the concept. Putting the Ross material into the article only confuses things, and makes the reader wonder why the quote is there when it contradicts the article and no explanation is given.

There are basically two usages of emf. One is the lead-in: work per unit charge in creating an electrostatic back voltage. The other equates emf to the electrostatic voltage itself, rather than its creation. The second usage, which ignores the whole idea of how the back voltage comes about, occurs primarily in circuit analysis, where the origins of the back-voltage are not at question, but only the electrical consequences.

Because the second usage is a synonym with electric potential difference, it is worthy of only a subsidiary discussion in the Terminology section. The first usage comes up in Faraday's law of induction, in batteries, in semiconductor devices, and anywhere where the mechanism of generation of back voltage is of interest.

Ross's plethora of other distinctions and usages are undocumented, undocumentable, and marginal. The quote should be dumped. Brews ohare (talk) 14:41, 26 June 2009 (UTC)

differ in sign?

Brews has cited Griffiths to support his interpretation that different definitions differ in the sign of emf. This is not in Griffiths, and seems exceedingly unlikely. More likely, emf is defined as positive for a device adding work to a circuit, but the conventions on what direction to integrate differ. Can we find a source that actually comments on sign conventions, like for the direction of the loop or path integrals, so that we can clear this up? Dicklyon (talk) 16:57, 26 June 2009 (UTC)

Dick: You are on the wrong track here: notice the minus sign in Griffiths two-terminal equation, which does not appear in the closed-loop equation. Also, notice the term "back voltage" in the quotation: the adjective "back" means the voltage is in the opposite direction to the emf. That also is why Griffiths says there is zero net motance inside the source of emf: the emf is countered by the electrostatic field; that is, the electrical voltage is opposite in sign to the emf itself. This is why zero current flows in the open circuit condition. Brews ohare (talk) 17:10, 26 June 2009 (UTC)

I've taken this out of a few places again. The cited Griffiths page doesn't comment on the relative polarities of emf and voltage. What is the convention? When you have a 1.5 V battery connected to a resistor, putting 1.5 V across the resistor and across the battery, measured from the + side to the minus side, that corresponds to a + emf of the battery, as it's adding work to the circuit, right? This mean you run your test charge around in the same direction as the actual current to do the integral? And if you measure the voltage across the battery by looking around the loop in that direction it will be -1.5 V across the battery; is that what you mean by the opposite sign? But when emf is used to mean voltage, that's not the direction that the measure voltage in. The interpretation that the signs differ is your way of saying that you don't measure the voltage in the conventional way perhaps? The voltage that corresponds to the emf is the voltage that the battery supplies to the circuit. At least in all the stuff I ever saw. But point out where Griffiths explains things and talks about "back" and I'll try to get your point. Dicklyon (talk) 06:22, 28 June 2009 (UTC)

By opposite in sign I meant that inside the battery (as an example source) the electric field due to the charge separation opposes the forces driving the separation. In open circuit condition, then, the net work done in taking a charge from A to B is zero, because what is gained from the separating forces is lost in climbing the potential hill. This is what Griffiths says on p. 293: "the net force on the charges is zero". I take that to mean emf + o.c. voltage = 0. Brews ohare (talk) 14:35, 28 June 2009 (UTC)

Emf in diodes, imrefs, and thermodynamics

Dick: The whole idea of emf in diodes bothers me. It is apparent that charge separation occurs in diodes and the potential drop across the diode is due to this charge separation. During the establishment of this drop, a current flows, driven by the energy difference of electrons in the two materials. That all sounds like an emf at work.

However, there is no continuing supply of energy conversion once the built-in potential is established, and putting a load across the junction will not draw a current. That is not like the emf in a battery. Therefore, the definition of emf is faulty, because it does not distinguish such cases of transient establishment of charge separation from the ability to provide steady-state maintenance of current.

The p.d. in a diode does not cause a current and does not enter Kirchhoff's law. The failure of the diode p.d. to enter Kirchhoff's laws seems to put the lie to the 98% of all sources that equate voltage to electrical potential difference. Some sources Quimby Neamen attribute this failure to the need to account for electrode contact potentials as well as the junction built-in potential. If that is done, all these potentials will add to zero, leaving no contribution to Kirchhoff's law. Although true, I'd say its a cop out, and the real thing to look at is the Fermi level, which is flat.

In the photodiode the potential drop across the diode (which pushes no current) is reduced by the photo-voltage. The current that flows in the load does flow in accord with the photovoltage, like the battery case, but you can't calculate the emf the way the definition suggests as some kind of line integral. The emf appears to be given by a different type of formula, one connecting conversion of light energy to electrical energy.

Thermodynamics may be the way to go, or maybe quasi-Fermi levels. In any event, it is energy conversion, whether static or transient, that is the key, and emf denotes an energy imbalance that pushes charge. What do you think the article should say about all this? Brews ohare (talk) 18:28, 26 June 2009 (UTC)

I think we should say not much besides what we find in sources. I think of it this way: ignore the shunt R in the model, and assume you know what the series R is, and treat the diode as part of the elementary "cell"; then at zero current the terminal voltage is the emf, even in the shorted case in the dark. In the dark, then it appears that the emf is nonzero for the open-circuit cell and zero for the short-circuit cell, with no work being done in either case. In the light, the emf similarly adjusts itself to the load voltage, since the cell can't provide more than 1 electron per converted photon. From the terminal voltage and current in any condition, the emf is easy to calculate. Predicting the terminal I–V curve from light level and cell details is more complex, but there are models. From these models, you could work out how emf varies along the I–V curve. Basically, the extracted energy per photon is varying as the bias changes, being greatest for the largest forward bias, though the conversion efficiency goes down there. It's complicated, but the emf is still the work done per charge delivered. The mode I'm more familiar with is the back-biased detector, in which the absorption of photons causes stored potential energy to be dissipated, so the emf is not the concept I've worked with. Dicklyon (talk) 21:27, 26 June 2009 (UTC)

I've forgotten exactly how to analyze the band-bending and built-in potential relative to Fermi level and terminal voltage. I think the key is that when light is provided, you get carriers in the other band, far from the Fermi level. Maybe this changes it to a quasi Fermi level (does that mean non-equilibrium?). Anyway, I don't think I buy that explanation about the metal–Si junctions providing their own potential differences; isn't that what ohmic junctions are about? Doesn't the diode terminal voltage float to an actual nonzero difference in the absense of any load? I thought it did, but now I'm not sure. Dicklyon (talk) 21:33, 26 June 2009 (UTC)

Caveat on Griffiths formula

I believe that the photo diode can be included in the opening definition and in Griffiths line integral provided the E-field included is only that due to the mechanism causing charge separation (the light). That way the built-in electric field caused by thermodynamic equilibrium between dissimilar solids is ignored, and you get the correct photo-voltage if you know the correct charge distribution induced by the light. Brews ohare (talk) 13:05, 27 June 2009 (UTC)

Recent edits

(the path is taken from the negative terminal to the positive terminal to yield a positive emf, indicated work done on the electrons moving in the circuit when current flows in that direction in the source, or from positive to negative through the external part of a closed circuit; measured in this direction, the electric field is negative and the voltage difference, negative terminal relative to positive terminal, is negative).

Dick: By integrating the total field over a segment external to the source of emf, a contribution from both the conservative and the non-conservative E-field may be picked up (e.g. in cases when a magnetic field is present). Hence, the emf calculated on an open path with a segment external to the source does not produce the correct value. That is why Griffiths spends so much time on this equation. Please read Griffiths' discussion and revert this edit. Brews ohare (talk) 13:35, 28 June 2009 (UTC)

I deleted a portion of this edit that conflicts with the following discussion. I'm happy to leave it like that. Brews ohare (talk) 14:47, 28 June 2009 (UTC)

A transformer coupling two circuits may be considered a source of emf for one of the circuits, just as if it were caused by an electrical generator; this example illustrates the somewhat arbitrary nature of the decision of what is the circuit and what is an external agency.

This example does nothing of the kind. The division of the circuit is not arbitrary. It is an example of why "transformer" emf is called that, and in either of the two transformer-coupled circuits, this is a straightforward application of the definition of a source of emf. Please revert. Brews ohare (talk) 13:35, 28 June 2009 (UTC)

I modified this sentence to leave out the "arbitrary decision" stuff. Brews ohare (talk) 14:49, 28 June 2009 (UTC)

How does one decide what part of a circuit to call "external" when there are mutual inductances? What transformer windings are part of the circuit, and which are external? It all seems pretty arbitrary. How do you decide which voltages are due to emf and which are just interactions of elements of your circuit? Dicklyon (talk) 20:33, 28 June 2009 (UTC)

If an electrical voltage is not connected to an external resistor, then an electric current will not flow through that resistor (Ohm's Law). If an emf is applied across the resistor, between the terminals of the source there must exist a true electric field that produces a voltage difference that exactly matches the IR drop in the resistor.

Why was this deleted? Do you disagree with it, or do you think it is redundant? Brews ohare (talk) 13:35, 28 June 2009 (UTC)

It just doesn't make any sense. Maybe "If an electrical voltage is not connected to an external resistor" was intended to mean "If no electrical voltage is applied across a resistor" or something like that. And what does it mean to apply an emf, especially given your insistence that an emf is not a voltage. And what is a "true electric field" in contrast to? I just couldn't find any sensible interpretation or intent, and the section seemed more coherent without it.

Deleted reference to "Basic Electricity". This reference was included as an example of the use of the term emf as a voltage in circuit theory and was deleted as "not referring to emf": Here is a quote from this source discussion Kirchhoff's voltage law on p. 76:

"Kirchhoff's Voltag Law can be written as an equation as shown below:

$E_{a}+E_{b}+....+E_{n}=0$

where E_a, ... etc are the voltage drops and emfs around any closed circuit loop."

The "voltage drops" in the figure are the IR drops and the emf's are the battery voltages, showing quite clearly the usage that was to be illustrated. This reference should be restored. Brews ohare (talk) 13:35, 28 June 2009 (UTC)

My search of the cited page must have missed that. Sorry if so. Dicklyon (talk) 20:33, 28 June 2009 (UTC)