Talk:Transformer

Earlier talk is archived at :

• Talk:transformer/Archive1
• Talk:transformer/Archive2
• Talk:transformer/Archive3

Resultant flux in core

(hdg for nav)

If both lots of flux are not included in the diagram of a transformer, you tend to get into all sorts of trouble in explaining whats happening- which is the situation we have now. Both lots of flux should be shown in the diag. Pri flux generated by pri winding and sec flux generated by sec wdg. Then they both cancel out and everybody's happy!--Light current 22:13, 28 January 2006 (UTC)

Your unique perspective challenges us all. Everyone else seems to have a different understanding of how a transformer works. Only one flux, two opposing MMFs. --Wtshymanski 22:16, 28 January 2006 (UTC)

Maybe people believe what they read, but not what has been written? OK. Come up with the alternative beliveable, thought out, argument that makes proper sense to everyone. My interpretation is the standard one if only you can read between the lines of the math! (which are not really very close together)--Light current 22:19, 28 January 2006 (UTC)

How a transformer works

First you apply an ac voltage to the primary. This gives rise to a pri current which tend to cause a flux in the core. If the sec is loaded, a current will flow in it creating an opposing flux. This leaves only the magnetising flux in the core . Very simple!.

Paul, Now you will see that in order to get flux in the core in the first place, you need to apply a primary voltage. THis is what starts the whole process off. So it it ludicrous to say that the primary flux CAUSES the primary voltage. Just plain Madness!!!--Light current 00:11, 29 January 2006 (UTC)

If you could get perfect magnetic coupling between windings, you wouldnt need a core at all!--Light current 02:02, 29 January 2006 (UTC)

Go complain to Michael Faraday. The standard way to write this is that voltage is proportional to the change in flux. It makes no sense to do it the other way around in wikipedia. Also, I think you are confusing the input voltage with the voltage across the primary. The voltage across the primary is not at all what starts the process. It is a response to a change in flux (change in current). At the start, there is zero voltage across the primary. You attach an input voltage. This causes a small amount of current. The winding reacts to this current with a back EMF. There is no EMF until there is current. Pfalstad 03:55, 29 January 2006 (UTC)

The primary flux causes the primary EMF. I'll give you that. But not the applied voltage becuase thats whats applied. Nothing happens until you connect that source of voltage! The problem lies in the way the article is trying to explain the operation as one of 'cause and effect'. As I said before, if the cct was treated as two coupled coils like it is in most texts, the the equations are simply solved and no problem of causality arises. But maybe this is too complicated for the page?--Light current 04:01, 29 January 2006 (UTC)

The second equation is labeled "Similarly, the voltage induced across the secondary winding is". That is clearly talking about the EMF, which is caused by the flux. Would you feel better if we said "the voltage induced across the primary winding is proportional to the rate of change of flux"? That makes it clear that we are talking about the EMF not the input voltage. And if you want you can say "this must be equal to the input voltage Vf". The text is a bit unclear about which is the applied voltage and which is the EMF. (They're equal, so it's easy to get them confused.) The applied voltage is labeled $V_{f}$ in the diagram, but lowercase v is used in the text to mean either the emf or the applied voltage. Pfalstad 04:13, 29 January 2006 (UTC)

Well thats certainly better Paul. Why not try that?:-)--Light current 04:16, 29 January 2006 (UTC)

Mathematical equality is not assignment and has no sequence implied. --Wtshymanski 19:00, 30 January 2006 (UTC)

Yes but we're trying to explain how a transformer works in a 'sequential' manner. I say that that's the wrong way to do it. So what you have sown, you are now reaping!--Light current 22:16, 30 January 2006 (UTC)

The truth is that it can be the current or the voltage that induces the flux in the winding(s), to prove this you can look at the extreme cases, assuming an ideal lossless transformer: A. Short circuit at the secondary. if an ideal current source was applied to the primary no voltage would be iduced at either end, but there would be current at both. B. Open circuit, if a voltage source were applied no current would flow in either, but we would see proportional voltages at each end. Of course, the problem with this explaination is that it does not account for the transients that are present in a real transformer, which you need to take into account to really explain what happens sequentially in a transformer, but that is really too deep for wikipedia I think, becuase it is different for different scenarios. All that i really wanted to say is that the way it is right now, in the opening is fine i think, but the section on coupling by mutual induction is symantically wrong. Magnetomotive force is IpNp, it has nothing to do with Vp (although it's a common confusion, because of the fact that magnetomotive force is analogous to electromotive force). Anwyay, i'm not going to make this change right now. But it's something to consider. Nozog 09:58, 25 April 2006 (UTC)

Telephone history

Can anyone find any history of audio transformers, i.e., telephone circuits, etc? This side is missing in the present article. --Wtshymanski 19:00, 30 January 2006 (UTC)

Thers quite a lot on uses of tranformers (inc hybrid transformers) in telephone systems in the book

Understanding Telephone Electronics, by Fike& Friend. A Radio Shack publication of 1983 but developed (written by) by Texas Instruments Learning Center.--Light current 22:50, 30 January 2006 (UTC)

Tickler windings

Sorry I didnt see any text relating to this diag when I deleted it. However, isnt a tickler winding just a loosely coupled loop? Why use this name that I have not heard before?--Light current 15:03, 7 March 2006 (UTC)

It's called a "tickler" because in an Armstrong or Hartley oscillator, it provides the feedback (the "tickle") that keeps the main LC tank circuit oscillating. ...

OK its an old fashioned term in use before I was born! But is it still current terminology?--Light current 15:13, 7 March 2006 (UTC)

...That's an interesting question! I don't see many true RF circuits these days so I'm no expert, but it's been a dog's age since I've seen a true Armstrong or Hartley oscillator, so even if the term still is correct, I'll bet it's obsolescent.

Atlant 16:26, 7 March 2006 (UTC)

It's refered to alot in crystal sets. J. D. Redding 17:27, 7 March 2006 (UTC)

I thought crystal sets had only one coil. Thats what the art says.--Light current 04:03, 27 March 2006 (UTC)

A tickler winding was a loosely coupled feedback winding that was used in tube type regenerative receivers. By providing an adjustabe amount of positive feedback into back the front-end detector stage, these receivers operated right on the edge of oscillation in order to achieve high overall signal gain. When the detector received a CW signal the circuit would break into oscillation. The receiver oscillator would then combine with the incoming CW signal, and the resulting lower beat frequency could be heard as an audio signal. Originally developed by Armstrong in 1913. It is, indeed, an archaic term used only by us archaic electronics/radio hobbyists. Bert 05:21, 27 March 2006 (UTC)

So do we include it or what?--Light current 15:57, 27 March 2006 (UTC)

I'd recommend that it either be removed, or at least reworded. I am not familiar with modern RF designs that use regnerative feedback via a tickler winding. I'll reword it to at least be more accurate. Bert 03:11, 28 March 2006 (UTC)

The term 'feedback winding' I have heard and I think its still current. Mind you, tickler seems to be current as well![1]--Light current 03:15, 28 March 2006 (UTC)

Well, looks like the term may still be in use after all - guess we should leave it in... :^) Bert 04:18, 28 March 2006 (UTC)

Agreed (reluctantly)!--Light current 08:45, 28 March 2006 (UTC)

Front page

Transformer will appear on the front page of Wikipedia on Friday 28th April. If you can find ways to polish the article in time for that date, it would be welcome! --BillC 17:48, 30 March 2006 (UTC)

Where was the featured article proposed. I have seen no info on this and it doesnt appear in the featured article list. So whats going on?--Light current 17:58, 30 March 2006 (UTC)

Fluxes in the core

Do we have time to correct the erroneous description of the fluxes in the core?--Light current 17:50, 30 March 2006 (UTC)

And while we're at it, lets get that baffling mechanical analogy out, too. This article is no-where near featured status. --Wtshymanski 18:30, 30 March 2006 (UTC)

No we need to get the core flux sorted first if this diag is to be on main page!!--Light current 22:12, 30 March 2006 (UTC)

The picture is OK. Flux looks fine to me - it's even got the current arrows going the right way around. I know you don't believe in flux, but your opinion is decidedly in the minority. --Wtshymanski 22:37, 30 March 2006 (UTC)

My view is the normally accepted view given in the textbooks of two opposing fluxes largely cancelling out and leaving only the magnetising flux. The diagram is misleading in that it implies no flux generated by secondary winding and a large flux in the core. This is nonsense!--Light current 22:44, 30 March 2006 (UTC)

Partially agree. However, it's not nonsense if the transformer has no secondary load - the core flux will indeed be a maximum under this condition. Bert 01:33, 31 March 2006 (UTC)

Not sure if its a maximum with no sec load. I thought it would just be the mag flux. Isn't this how we measure magnetising current? (by the o/c test). And its the magnetising current that causes the magnetising flux. Since we are attempting to show the transformer's normal operation (ie with a sec load), I think the diag needs modifying. Sorry--Light current 02:00, 31 March 2006 (UTC)

I don't see how the diagram implies anything about the magnitude of the flux or which winding generates it. I don't see how it is possible for the diagram to show or imply much about those things. How would you explain what is a "large" flux. I would think that the magnitude can't very well be characterized except to state how close the iron is to saturation. --C J Cowie 00:50, 31 March 2006 (UTC)

The diagram may not imply anything about the size of the flux to you and me, but to a novice reader it may give the impression that there is no restraint on the size of the flux- whereas there in fact almost complete cancellation of the primary flux by the secondary flux. --Light current 02:00, 31 March 2006 (UTC)

I don't think that the diagram should attempt to show more or less than it shows. Additional information must be conveyed by the text of the article and perhaps other diagrams. Perhaps the caption should point out that the directions of the currents and flux are for an instant in time. Perhaps the flux should be termed "magnetizing flux" or "net flux" or just "flux" rather than "resultant flux."

I think two opposing fluxes should be shown! If this is not possible then the flux should be renamed magnetising flux.(or net flux). Yes I agree the term resultant is a bit ambiguous- ie what is it the result of?--Light current 02:37, 31 March 2006 (UTC)

Have you seen opposing fluxes shown or discussed in texts? The discussion that I have been reading talks about opposing magnetomotive forces resulting in no change in flux due to the presence of a secondary current. I think that showing opposing fluxes may cause confusion rather than clarify the situation, but perhaps the concept of opposing mmf is too abstract. --C J Cowie 02:57, 31 March 2006 (UTC)

Yes I have. For instance its in Higher Electrical Engineering by Shepherd, Moreton and Spence. publ Pitman Paperbacks, 2nd ed 1970. ISBN 0 273 40063 0 (page 261. chapter on transformers.)

It actually talks of four fluxes in the core with two coils on it:

1. Phi m1, a component due to the current in winding 1 which links all the turns of both windings
2. Phi m2 a component due to the current in winding 2 which links all the turns of both windings
3. Phi l1, a comonent due to the current in winding 1 which links all the turns if winding 1 but none of those in winding 2 ( pri leakage flux)
4. Phi l2 , a component due to the current in winding 2 which links only the turns of winding 2 (sec leakage flux)

The case we consider is when both windings are carrying current. In this case, two opposing fluxes are shown in the diagram.--Light current 03:11, 31 March 2006 (UTC)

I have found another reference that uses essentially the same argument:

Electrical Engineering Fundamentals by J P Neal dept of Elec Eng, University of illinois. publ McGraw Hill 1960 Library of Congress No 59-13210. In sect 7-9 on mutual inductance , p301, a transformer core is shown with two windings. I quote from the text (p302):

On referring to the figure, the reference directions of the mutual fluxes, phi 12 and phi 21 are in opposition along the common flux path--Light current 03:28, 31 March 2006 (UTC)

In Electric Machinery by Fitzgerald, Kingsley & Umans (see Electric motor, they have an ideal transformer diagram like the one we are discussing and state: "Note that for the reference directions shown … the mmfs of i1 and i2 are in opposite directions and therefore compensate. The net mmf acting on the core therefore is zero, in accordance with the assumption that the exciting current of an ideal transformer is zero." They also have a diagram showing the leakage fluxes, but show only the "resultant mutual flux" linking all turns of both windings. In Circuits Devices and Systems by Ralph J. Smith, he states "To maintain the magnetic flux at the required value, the mmf N2I2 produced by the secondary current must be offset by an equal and opposite mmf, N1I1 produced by additional primary current I1'." --C J Cowie 03:42, 31 March 2006 (UTC)

You are exactly right CJ! But the resultant mutual flux must be zero!! THe condition of equal MMfs must be the same as that of equal fluxes must it not?--Light current 22:15, 31 March 2006 (UTC)

Yes. The two approaches to the explanation lead to the same conclusion. I am still thinking about the advantages and disadvantages of one approach vs. the other. --C J Cowie 23:01, 31 March 2006 (UTC)

Featured article or picture?

Sorry if I was misleading. Transformer isn't a Featured Article, needing to go through peer review and then FA candidacy. However, Image:Transformer3d col3.svg will be a featured picture of the day on the front page then, and will contain a small intro section taking the reader straight to the Transformer article. It will, therefore be subject to more eyes than normal. The plan for next month's front page Featured Pictures is at Wikipedia:Picture_of_the_day/April_2006.-BillC 19:11, 30 March 2006 (UTC)

Now that you've got this picture, could you do the "turn of wire turning in a magnetic field" that the generator article needs desparately? I've had Inkscape for months but no time to practice using it (and not enough skill to use it for a decent picture without practice).--22:37, 30 March 2006 (UTC)

Flawed analogy

A gearbox can transmit torque at zero speed. Depending on the flavor of the analogy used, this is like a transformer transmitting power at DC or transmitting current with no voltage, or some such darn implausible thing. Maybe a transformer is more like a hydraulic torque converter, though if you can understand how that works you'll probably be able to understand the transformer physics equally well. --Wtshymanski 21:11, 31 March 2006 (UTC)

Bill, I recommend you beg, borrow or steal a copy of 'Dynamical Analogies' by Harry F.Olsen (Publ D.van Nostrand) 2nd ed 1958. Lib of congress card no:58-14102. When you have found a copy, look at p58 and tell me why you disgaree with this learned gentleman!--Light current 21:19, 31 March 2006 (UTC)

Please read the text again. I think your analysis of the analogy is wrong:

In this comparison, current is equivalent to shaft speed, voltage to shaft torque. In a gearbox, mechanical power (speed multiplied by torque) is constant (neglecting losses) and is equivalent to electrical power (voltage multiplied by current) which is also constant.

A transformer can transmit voltage (torque) with no sec current (zero speed)! No problemo!!!8-)) --Light current 23:22, 31 March 2006 (UTC)

Branchlist

Regardless of any merits of its inclusion, the new branchlist box is disrupting the top of the article. If you really want navboxes of this nature, they're generally better placed as horizontal boxes at the bottom of the article. See for example any of the Swiss canton navboxes, or those relating to a pope. --BillC 20:10, 1 April 2006 (UTC)

Im afraid I cant agree here Bill. The branch list must me at the article top to aid speedy navigation back and forth. After all, when you dial up a new page, it displays from the top down.--Light current 20:12, 1 April 2006 (UTC)

But the "branchlist" is in fact a menace to navigation - it clutters up the top of the article which is supposed to grab the user and give them enough information to decide if they need to read the rest or not. LINKS are for navigation. Many of the branchlists have been stunningly irrevant to the pages they appear on - maybe someone things tubes and MOSFETs are alike, but I doubt J. Random Encyclopedia User wants to be diverted off-topic at the very start of the article. --Wtshymanski 21:16, 1 April 2006 (UTC)

Ah well, its going to be a trade off isnt it. I dont regard it as a diversion, more as a handy way to waltz aroun the topic. Those who know where they are going will ignore it anyway! The lesser of two evils lets say! Can you make +ve suggestion?--Light current 21:20, 1 April 2006 (UTC)

Well, it has adversely affected the formatting at the top of the article. The article no longer has a lead image, the actual picture of a transformer is now split straddled across two sections, and some of the links in the branchlist have little or no relevance to transformers. You have asked for a positive suggestion; how about moving the lead picture back to where it was and reconsidering the format and placement of the branchlist? --BillC 21:29, 1 April 2006 (UTC)

Yes I will consult with Lindosland reagrding the placement of the branchlist. Thank you for your comments.--Light current 21:31, 1 April 2006 (UTC)

Ive moved the lede img to the left . Is that any good?--Light current 21:37, 1 April 2006 (UTC)

To be honest, no. --BillC 21:42, 1 April 2006 (UTC)

Well where would you suggest putting it . (Be polite please!)--Light current 21:44, 1 April 2006 (UTC)

I wish we had a better picture - what an ugly little pole-pig. Didn't we used to have a better picture? --Wtshymanski 21:59, 1 April 2006 (UTC)

There's this in the Commons. The image description translates as 'three-phase medium voltage transformer cross-section'. --BillC 02:50, 8 April 2006 (UTC)

I would suggest placing a horizontal box at the bottom, or not at all. I still don't see what the rationale behind this branchlist is; there are a number of items that are not relevant to transformers, and there are some significant omissions. Where, for example, is generator or electric power transmission? What does the branchlist provide that categories do not? --BillC 22:19, 1 April 2006 (UTC)

Well BillC, the idea is supposed to make it easier for newbies and others to find thier way around their topic of interest. For instance, if I was a newbie, and got to a page that wasnt quite what I was looking for, I wouldnt want to have to scroll all the way down to the bottom and hack my way thro' a complicated category list. I want to hit the 'back button' (or equivalent) or choose another subject from the conveniently palced menu at the top of the page.--Light current 22:39, 1 April 2006 (UTC)

About the Universal EMF equation

What is the unit of the cross section? I find the term "The value 4.44 collects a number of constants required by the system of units" pretty inaccurate, as it gives the feeling that this equation comes from nowhere. Does someone have more information about this? CyrilB 22:13, 1 April 2006 (UTC)

4.44 = 2Pi/Sqrt2 --C J Cowie 22:20, 1 April 2006 (UTC)

As C J Cowie said. Though, it's not collecting constants required by system of units, simply from the mathematical definition of a sinusoidal waveform, so that last part could go. The units of cross section are necessarily the same as those of the flux density B, which is measured in units of flux per unit area. --BillC 22:24, 1 April 2006 (UTC)

If we defined both E and B as RMS values, then we could get rid of the 1/(root 2). This would also remove the requirement that the flux waveform be sinusoidal. Is there any reason not to do this? --Heron 12:46, 15 April 2006 (UTC)

I believe that the universal EMF equation was derived from Faraday's law for use with AC power equipment intended for use on sinusoidal power supplies and operating with sinusoidal flux waveforms. The peak value of flux comes from the instantanious values in Faraday's law. I think that the peak value remains of interest in that a design constraint is that the maximum flux density is normally limited to prevent saturating the iron. I think that your revision has clarified the article, but the universal EMF equation is usually presented with the 4.44 constant because it makes a convenient design formula.

Thanks. As long as there's a reason, then it's fine. --Heron 16:08, 15 April 2006 (UTC)

Format of lead paragraph

LC, it must be clear that inserting the branchlist, if nothing else, is really spoiling the format of the start of the article. I have inserted a breakafterimages template to try to improve things a little. If you move text from Introduction into the lead section, then the space there will be filled and it should look better. In fact, there is no real need for an 'Introduction' at all. Why not move all of that into the lead? Per WP:LEAD, an article of this length should expect 2-3 medium-length paragraphs in its lead. This was tried before, but you removed it. --BillC 21:47, 3 April 2006 (UTC)

Done it! Any better?--Light current 22:16, 3 April 2006 (UTC)

I'll leave you to it. --BillC 23:49, 3 April 2006 (UTC)

What? Is that not what you suggested? Or Ihave I misunderstood you?--Light current 00:01, 4 April 2006 (UTC)

Electrostatic coupling?

What is electrostatic coupling? that should be explained (its mentioned on this page. Fresheneesz 10:07, 14 April 2006 (UTC)

Electrostatic coupling is a defect in practical transformers. Because there is always some capacitance between the various windings, there is some amount of capacitive coupling between the windings. This is referred to (somewhat archaically, I guess) as electrostatic coupling. Obviously, it becomes a bigger factor at high frequencies (where the impedance posed by a given size of capacitor is reduced). Even for line transformers, it may be a factor as it can lead to the coupling of high-frequency noise from primary to secondary.

Where it's deemed to be a significant factor, transformers often contain a grounded shield (e.g. a non-closed sheet of copper) separating the windings.

(Feel free to extrapolate what I've just said into the article.)

Atlant 12:12, 14 April 2006 (UTC)

Yes, 'capacitive' is better than 'electrostatic', and the shield is briefly mentioned in the article. I changed the word in the article. --Heron 12:24, 14 April 2006 (UTC)

Thanks!

Atlant 12:56, 14 April 2006 (UTC)

Improvements have helped

I have been asked to come back and comment on the improvements made to this page since I last commented. And I must say that the improvements have helped a lot, well done to all those who were involved in making the changes. --Badharlick 06:43, 21 April 2006 (UTC)

Phase relationship

Question moved from article to talk page:

What is the relationship between the AC voltage waveform on the primary and the AC voltage waveform on the secondary? Are the two waves synchronized in time and frequency? Are they in phase? — Preceding unsigned comment added by 207.255.22.129 (talk • contribs)

The flux in the core is 90° in phase behind the primary voltage. The voltage induced on the secondary is 90° ahead of the flux, so if the primary and secondary windings are in the same sense, the primary and secondary voltages are in phase with each other (i.e. synchronised). If the secondary is wound in the opposite sense, there is a 180° phase change between primary and secondary. The two waveforms have the same frequency, as a transformer alone cannot change this. (Magnetic saturation can introduce non-linearities into the equation, which has the effect of introducing component frequencies, but this is an effect associated with a non-ideal transformer.) (BTW, questions should be posted to talk pages, not to articles, thanks.) --BillC 16:59, 28 April 2006 (UTC)

rewrite

I have renamed "High frequency operation" to "Operation at different frequency". 400 Hz is not "HF". I do not know if this a good solution. I have added additional data to this section.

• Analogies are useful, this article needs something different. alex 17:19, 28 April 2006 (UTC), 9 may 2005

Read Dynamical Analogies by Harry F. Olson ;-)--Light current 17:49, 30 April 2006 (UTC)

The explanation reads over-complicated. "Transformers spin voltage analog to a mechanical gear box gear ratio" fully does it. User:Akidd_dublin 9 may 2006

I have now edited this out. It is preserved as an example for over-complicated explanation, which any already exists at an other place. I am collecting examples of such language usage here: User:Akidd_dublin/cleanup/language_usage Akidd dublin•tl•ctr-l 15:10, 10 May 2006 (UTC)

I have reinstated the gearbox analogy (which is a good one) until someone can come up with something better.--Light current 17:30, 10 May 2006 (UTC)

That's fine, however it is already explained within the gear box article. If a schoolteacher talks that way, i do not like it, because it is over-complicated. Akidd dublin•tl•ctr-l 12:39, 12 May 2006 (UTC)

Sorry this is an analogy for transformer operation, not an explanation of gear boxes. 8-| --Light current 13:59, 12 May 2006 (UTC)

possible split up of the article

• transformers
  • physical principe
  • scope of application

• transformers_(iron_core) (new article)
  • power grid transformers (low power)
  • signal transformers
  • power transmission transformers

• transformers_(ferrite) (new article)
  • power supply transformers
  • flyback transformers
  • filters (reference)

• transformers_(special) (new article)
  • isolation transformer
  • voltage transformer

I wouldn't divide it up that way. I would take the construction portion and move it into a separate article; and I would take most of the items under "Transformer designs" and move them to separate articles. Pfalstad 14:19, 30 April 2006 (UTC)

Does it really need splitting when its just got stable :-? --Light current 18:00, 30 April 2006 (UTC)

Probably it will need splitting one day. These types of transformers are searched by different people. The pictures could be more/better quality. Probably the main page can take the principe of operation, and a gallery of two/three good pictures (buy a new transformer for that purpose). User:Akidd_dublin 9 may 2006

I agree on dividing the article. It is too long.--Lenilucho 03:31, 10 May 2006 (UTC)

General principles of transformers

type A transformers (power freq, say)

type B transformers (audio, say)

etc ....--Light current 00:27, 17 May 2006 (UTC)

"Generally" is too vague

How about a citation for that paragraph that it is generally "safe" (whatever that means) to operate a transformer above its rated frequency? It's too vague and I don't know what specific instances it refers to if not for power transformers. Is it safe to operate a wall-wart on a 400 Hz aircraft supply? Too many qualifications required for this article's intention. --Wtshymanski 17:33, 16 May 2006 (UTC)

I agree its vague. Its my experience! It actually refers to saturation of the core due to insufficient primary inductance. Saturation will not occur at higher frequencies if the applied voltage is the same. Do you not agree with this?

Give me time to find some refS! What is the articles intention? 8-?--Light current 21:31, 16 May 2006 (UTC)

Well Ive changed this now. But Im not really sure if it says anything new anymore. What do you think Bill?--Light current 02:12, 17 May 2006 (UTC)

If no flux, then how does energy get from one side to the other?

If there's no flux in the core, how does energy get from the primary to the secondary? I have once again deleted the idiosyncratic assertion that an ideal transformer has no flux. --Wtshymanski 17:40, 17 May 2006 (UTC)

No resultant flux!(apart from mag flux) Primary flux is cancelled by secondary flux, Its a feedback mechanism and its obvious to anyone. Pri flux cannot increase more than infinitesimally before it is cancelled by the sec flux. This again is obvious to any engineer! 8-)--Light current 23:48, 17 May 2006 (UTC)

An ideal transformer, loaded with a pure resistance on its secondary, will look like a pure resistance at its primary terminals. ie no 'back emf' is detectable anywhere!! How can this be? 8-|--Light current 00:31, 18 May 2006 (UTC)

Good lord. Youth is wasted on the young. If I had your energy I'd be rich. If there's no flux, then *how* does energy get from primary to secondary? The article SAYS its the time rate of change of flux that produces the voltage in the winding (primary or secondary) - if there's no flux, its time rate of change is zero and there's no voltage! If the resultant flux in the core is zero, then we don't need the primary winding at all, and hey presto, we get energy from nowhere. I'd revert but you've got many valid edits in the last 1023 or so since I last saw the article yesterday. Saving every line as you change it helps get the edit count up, eh? --Wtshymanski 17:48, 18 May 2006 (UTC)

I dont undertand your first statement. Unfortunately Im not rich. Does transfer of energy require flux? or just the threat of flux or mutual cancellation of flux?

Let me try to explain it another way: Assume there is massive flux caused in the core by the action of the primary winding. However, there is equally massive flux in the opposite direction generated by the secondary current. These large rates of change of flux generate voltages. One is the applied primary voltage, the other is the secondary voltage.

Do you agree that pimary AT = sec AT or not?(neglecting mag current).If so you can immediatly see there can be no resultant flux in the core. We need a primary winding to create an alternating magnetic field to cancel the flux generated by the secondary winding. Energy transfer is by the mechanism of mutual flux cancellation. Physically what does this mean? I dont know. What does flux mean?

I will as you a converse question: Why do some transformers not need a core (trans line transformer et al?)

Some times the thoughts come to me one by one. I save the page when I have entered a valid thought and then see that the page still makes sense overall. I do not have any interest in my edit count as such. I notice tho' that this seems to be an increasingly favorite form of attack from editors with low (usually very low) edit counts and some vandals.--Light current 18:18, 18 May 2006 (UTC)

I have two references to the fact that the standard transformer analysis trests the transformer as having mutually opposing fluxes that sum to zero. They are listed in the article. You yourself (WTS) have stated that the MMFs cancel- so how can there be flux with no resultant MMF?--Light current 03:18, 25 May 2006 (UTC)

Others' comments

I don't have a lot to say about the question of flux/no-flux; I think the answer depends on how you define things and where you measure. I will suggest, however, that you forget the idea of "the core" versus "the air". The core is just a means to concentrate the magnetic flux and thereby increase the inductance, but short of core saturation and to a first approximation, there's no difference, in principle, between the way an iron core transformer and an air core transformer operates.

Atlant 18:25, 18 May 2006 (UTC)

There is a magnetic flux in the core, and there's nothing subjective about it. There's a clear explanation on allaboutcircuits.com. I'll summarise it thus. The voltage across the primary creates a flux in the core. This flux depends only on the primary voltage, regardless of what the secondary is doing. Why? Because E_p=dΦ/dt, and E_p has constant amplitude; therefore, Φ, which is common to both windings, must have constant amplitude. If a load is connected to the secondary so that a current flows in it, then there is an invisible tug-of-war between the two windings. The secondary tries to induce an infinitesimal bit of extra flux, which is immediately and exactly counterbalanced by an infinitesimal increase in flux in the opposite direction from the primary, so the net flux stays constant. --Heron 19:32, 18 May 2006 (UTC)

Well Atlant, I tend to agree with some of your above statement, but wherever you measure in the core, you will still find only the magnetising flux regardless of load current. I was trying to point out the need for flux cancellation in a loaded transformer. Highly permeable cores do, of course, concentrate the flux and increase pri inductance as well as increase coupling between windings.--Light current 22:58, 18 May 2006 (UTC)

Yes Heron, I'm talking about when there is a secondary loading, in which case sec. current flows and the fluxes are cancelled. This is the normal operation of a transformer. Otherwise I agree with the last part of your post except that its the pri current (H =IN, B=uH, phi=BA) that creates flux in the core. Also, I submit the the net flux you quote is merely the magnetising flux - nothing more.--Light current 22:58, 18 May 2006 (UTC)

WTS to LC

Nom d'un chien. OK, nevermind a transformer. Take an ideal core out of your stock of physics thoguht-experiment supplies, wind some resistance-free wire around it, and impose a perfect sinusoidal voltage on the coil. The voltage across the the coil is proportional to the time rate of change of flux linked by the coil, right? It's a perfectly standard physics-experiment type inductor, correct? OK, now without telling your experimental apparatus anything, sneakily wind a second coil - my question is, how does the apparatus "know" that now it's an ideal transformer and so suddenly has no flux in it? --Wtshymanski 02:44, 19 May 2006 (UTC)

Soon as sec current starts to flow, the core flux is completely cancelled by the sec MMF. Thats how the apparatus 'knows' that it has been transformed (into a transformer!) Simple! 8-))--Light current 13:02, 19 May 2006 (UTC)

The rate of change of flux in the core is equal to the ratio of the voltage across the primary and the number of turns on the primary which is also to the ratio of the voltage across and the secondary and the number of turns on the secondary Thus, if we let the number of turns in the primary and secondary go to infinity (while, of course, keeping the turns ratio the same), the rate of change of the core flux goes to zero for any primary or secondary voltage. Assuming that at some time the core flux equaled zero, the core flux will thus remain zero. Further, this will yield a transformer that works at DC. While this condition is not usually stated as a property of an ideal transformer, it seems logical to me that a truly ideal transformer should work at an arbitrarily low frequency, right? Alfred Centauri 04:15, 19 May 2006 (UTC)

Yes Alfred, I agree with you. Another way of putting it, though, is to say that instead of letting the number of turns approach infinity, we could let the permeability of the core approach infinity. This would form what I understand to be an ideal transformer (assuming zero R leads, no leakage ind etc) and indeed there is no reason why it should not work at an arbitrarily low frequency. AS you say 'the (net) core flux will remain zero' but we still have energy transfer! BTW the term DC is a difficult concept (cf arbitrarily low freq) -maybe to be discussed elesewhere 8-))--Light current 12:58, 19 May 2006 (UTC)

Yes, Light current, it is true that that "net flux" I mentioned was the magnetizing flux. However, I am unhappy with your statement that "in an ideal transfomer they cancel so that there is no overall resultant flux in the core". Shouldn't you say "they cancel, leaving only the magnetizing flux"? We all seem to agree that the magnetizing flux exists regardless of whether there is a secondary winding or not, and regardless of whether any current is flowing in the secondary. As for the original question of "how does the energy get from one side to the other?", the answer seems to be that the magnetic field consists of two variables: MMF (F, in amperes) and flux (Φ, in webers) (see this paper from TI; PDF). Just as electrical power is the product of both voltage and current, magnetic energy is the product of both MMF and flux. (Check it dimensionally: MMF x flux = A x Wb = A x V.s = energy.) So, if you have no magnetizing flux, then you can't have energy transfer, but the flux by itself is not the energy. --Heron 09:46, 19 May 2006 (UTC)

Heron, depends how you are defining ideal. In an ideal transformer, I thought the magnetising inductance was infinite (inf perm core matl). In this case, there would no mag current and no mag flux required.

So we should strictly say 'in a practical transformer the pri and sec fluxes cancel leaving only mag flux'. The energy must be the product of the MMF and the 'potential' flux. Trouble is, we dont know the potential flux until we have sec current flowing 8-). So again, it seems as if actual flux is not the primary agent of energy transfer. Maybe its the 'potential flux'. Energy transferred is not IMO the product of the MMF and the mag flux. If it were, then the ideal trans could pass no energy.

Think about o/c voltage sources. What energy do they provide?--Light current 12:49, 19 May 2006 (UTC)

I didn't realise that 'ideal' implied 'no magnetising flux', so I think I agree with you now. Perhaps the article needs to be explicit about what 'ideal' means. --Heron 18:26, 19 May 2006 (UTC)

Purpose of a transformer core

What is the real purpose of a transformer core since it seems there is no flux in it. 9-)--Light current 13:58, 19 May 2006 (UTC)

To increase the efficency of the mutual coupling between the two (or more) coils. And as I think you guys finally argued, in transformer with a loaded secondary coil, there is instantaneous flux in the core proportional to the time-derivative of the voltage.

Atlant 14:10, 19 May 2006 (UTC)

If E=4.44fnab, and b is zero, and f,n, and a are finite -- then E must be zero. Got to have (changing) flux to make volts. --Wtshymanski 02:16, 25 May 2006 (UTC)

So if you could get perfect coupling, as with some transmission line transformers, you wouldnt need a core (indeed some transmission line transformers dont have one). No flux, no core-- this is getting weird. So what really is the essence of a transformer?--Light current 14:19, 19 May 2006 (UTC)

To enable complete cancellation of the magnetic field generated by the primary and thereby extract energy from the primary circuit, a core must be used to concentrate the field produced by the primary. and sec current must exist so that it can cancel this field. When the magnetic fileds are completely cancelled, all the input power is passed to the output. The core therefore acts merely as a conduit/coupling device to enable magnetic field cancellation.--Light current 14:56, 19 May 2006 (UTC)

However, the magnetic fields never completely cancel, even in a perfect transformer. The primary winding has self-inductance. When the primary is driven by a sinusoidal voltage source, there is always an associated "magnetizing" current and magnetizing flux, irrespective of core material. For a non-lossy core and perfect wire, the magnetizing current will lag the applied voltage by 90 degrees so no real power is expended (i.e., a perfect inductor). Although it is true that incremental changes in primary and secondary fluxes cancel as the secondary begins passing current, this cancellation has NO impact on the magnitude of the underlying magnetizing flux. Note also that, for most ferromagnetic core transformers, the magnetizing flux is NOT small. In most designs, the magnetizing flux comes close to saturating the core. Bert 18:11, 19 May 2006 (UTC)

Bert, I agree there is always a mag current in a real transformer. Since this current is lagging the pri voltage (current) by pi/2, it is wattless (apart from heating up the windings a bit). THe cancellation I have been talking about does NOT include the mag flux. I thought I was pretty clear on that.

It is true that the value of mag current does not depend upon load current or power transfer but only on primary applied voltage. The magnetising current/flux is purely a consequence of the non infinite inductance of the pri winding and takes no part in the actual energy transfer between pri and sec. It seems to me that some editors think that the mag flux plays some essential part in this operation of a transformer and maybe they think it varies with load current. I can see you are not one of them. 8-) So we can say that there is no component of flux in the core that is in phase with the applied voltage/current in a real trans with resistanceless windings and k=1. 8-)--Light current 18:50, 19 May 2006 (UTC)

BTW the reason that mag flux comes close to saturating the core in most trans is that the designer is trying to save on wire for the pri wdg -- ie its an economical (or is it poor) design!--Light current 18:55, 19 May 2006 (UTC)

Tapered microstrip

Has anyone noticed that a piece of tapered microstrip (or other type of transmission line) acts as a transfomer? They're usually used to match different impedances of course, but in doing so, they naturally change the voltage and current at the o/p compared with the input.

Where is the core? Where is the flux? Where are the windings? What IS a transformer? Curiouser and curiouser. 9-))--Light current 14:28, 19 May 2006 (UTC)

I think the answer to this conundrum is to consider the transmission line as a series of mutually-coupled inductors. That is, the transmission line acts as a series of distributed primaries and secondaries, each pair performing aportion of the step-up or step-down function that the transmission line as-a-whole manages to achieve.

Atlant 15:55, 19 May 2006 (UTC)

Well of course , the explanation you would get from a microwave engineer is that the em energy flow is constant thro the line, the impedance of the line changes, so hey presto... volts and amps must change to satisfy ohm law! The actual physical process I find a bit mind boggling at the moment. (Ive always wanted to use the word 'boggling' on WP -now I have 8))--Light current 19:23, 19 May 2006 (UTC)

Diagram

I think the diagram needs to be altered to show two opposing fluxes: one due to pri AT nd other due to sec AT. Any comments?--Light current 13:52, 19 May 2006 (UTC)

No, I disagree. It would make it horribly messy. You've argued for this before. --BillC 16:55, 19 May 2006 (UTC)

Yeah but its actually wrong as it stands! All it needs is for that large arrow to be split into two smaller ones pointing toward each other. 8-|--Light current 17:15, 19 May 2006 (UTC)

The new reference should be added of course, and well done to Heron for researching it. I note that the Edwards and Saha paper shows a transformer diagram that is identical (bar the reversed secondary winding sense and concomitant reversed secondary current) to the transformer diagram here. Is our diagram still wrong? --BillC 17:43, 22 May 2006 (UTC)

Interestingly, the Edwards/saha paper seems to show a non zero net flux in the core, but that could be just because of the reversed sense of the secondary winding compared to our diag. As you know, I personally think our diag is wrong about the net flux, but Im happy to go with what the research paper concludes if everyone else is! 8-)--Light current 23:36, 22 May 2006 (UTC)

Blame it all on...

As part of our ongoing discussion, Light current asks:

What IS a transformer? Curioser and curioser. 9-))

Blame it all on the magnetic monopole. Seriously, while we understand all of this EM stuff at a sufficient level to design and use technology based on electromagnetism, when you really get down to thinking about it, magnetism itself is still magic, a complete mystery to physicists. So if we start asking the really deep questions, the state of the art simply isn't advanced enough to provide answers. ;-)

Atlant 15:48, 19 May 2006 (UTC)

I hate not knowing how transformers work, so I've been searching the web for several days to get some more information on this question. Nearly every website repeats the same formula about MMFs summing to zero, and none of them explains how the energy gets from one side to the other. Then I Googled up these two posts - [2] and [3] - on sci.physics.research from 1999. They mention an entity that we haven't considered yet: the leakage flux. According to those posts, this is what transfers the energy. The magnetizing flux is a red herring. Atlant, I take your point about the ultimate inscrutability of magnetism, but perhaps it's not magic after all! --Heron 20:38, 21 May 2006 (UTC)

If the MMFs sum to zero, does not this necessarily mean that the flux sums to zero (not counting mag flux)? I think it does. But there seems to be some doubt expressed by other editors. 8-? --Light current 21:33, 22 May 2006 (UTC)

Energy transfer

What Heron? No energy transfer without leakage flux/inductance?? So no energy transfer in an ideal transformer (which has perfect coupling)? 8-o

OTOH, looking at it from a field theory viewpoint, E X H must come into it somewhere! Hmmm. I shall need to ponder. 8-| --Light current 00:01, 22 May 2006 (UTC)

Try Power flow in transformers via the Poynting Vector by J.Edwards and T.K Saha (PDF). The only way that an ideal transformer can operate without leakage flux is if the primary and secondary windings perfectly enclose the core and occupy exactly the same position in space. The closest you can get to that is a toroidal transformer with interwound primary and secondary. --Heron 11:16, 22 May 2006 (UTC)

Yes Heron- ref to the other web posts, some thought last night, a very brief look at Edwards and Saka paper and my own beliefs in energy transfer mechanisms show that this may be a very fruitful path to persue. Well done and 10/10 for your research efforts. 8-)).This paper should of course go in the references. The conclusions apparently confirm what we all knew: the best sort of transformer is a toroidal core wound with biflar wire! Funnily enough-- this then tends to approach the form of a transmission line transformer--Light current 12:17, 22 May 2006 (UTC)

For what it's worth, the same kind of analysis can be performed on a simple electric circuit consisting of a voltage source and a resistor. The current through the conductors set up H fields that add in the region between the conductors and subtract in the region outside. Likewise, the electric fields (due to the surface charge densities) of the conductors add in the region between the conductors and subtract in the region outside. In the region between the conductors, E X H is in the direction from the voltage source to the resistor. Once we see this, the (now) obvious step of looking at the magnetic dual of this electric circuit gives us the energy transfer mechanism of the transformer right away. Fascinating! Alfred Centauri 13:22, 22 May 2006 (UTC)

Ah yes- but is it a transmission line? 9-))--Light current 13:31, 22 May 2006 (UTC)

Everything is a transmission line ;-) ; it's just that sometimes the frequency is low enough that you can ignore transmission-line effects.

Atlant 13:36, 22 May 2006 (UTC)

Blue touch paper lit! Now I retire immediately! 8-))--Light current 13:38, 22 May 2006 (UTC)

Yeah, LC knows my stand on this. Any physically extended electrical device will display TL properties. Further, if a device is constructed similarly to a TL (like many capacitors are), they will behave precisely like TLs constructed in exactly the same way. I know this is a trivial point but I think it's worth making again. Alfred Centauri 13:43, 22 May 2006 (UTC)

No I think its a very important point and can aid in the understanding of how things actually work as it seems to be doing in this and other cases. After all, if you can explain all electrical devices in terms of energy flow, dont you think that is a great step forward? The great problem I feel is knowing when to use the circuit approach, and when to use field theory. They must meet somewhere in the middle 8-|--Light current 14:32, 22 May 2006 (UTC)

Purpose of core - second thoughts

It would seem from the exposition of Edwards and Saha, that a core is necessary in a standard type transformer to 'conduct' the H field, generated by the pri winding, to the sec winding. In very closely coupled systems, like bifilar wound TL transformers, maybe the H field doesnt have so far to go and therefore a core is not necessary. But the question still is: what is the resultant flux in the core?--Light current 15:10, 22 May 2006 (UTC)

Really? I'm still under the impression that H is medium independent - isn't it B that the core guides? Put another way, the H field is generated by the current in windings. Outside the core, the associated B field is small and equal to mu-naught*H whilst in the core, the associated B field is large and equal to H times the permeability of the core, right?

Also, I think that one of the main affects of the core is to increase the self-inductance of the primary and secondary thus allowing the transformer to be used at much lower frequencies. Gotta go teach and class and think about this some more. Alfred Centauri 18:31, 22 May 2006 (UTC)

'Conduct' was a really bad word for me to use here! Any other word is better!--Light current 22:57, 22 May 2006 (UTC)

Temporarily returning the subject of electrical circuits, Alfred, our friends Edwards and Saha just happen to have produced a paper on that too. See Establishment of Current in Conductors [4]. My poor old engineering brain is still struggling to free itself from the 'electrical power is electrons flowing through wires' stuff that it was taught, and this paper is a great help. (I'm not an agent for Edwards and Saha, by the way; I just like their style.) --Heron 19:28, 22 May 2006 (UTC)

I thought it said H field (mmf) is guided by the core but Im going to read it again.8-|--Light current 21:20, 22 May 2006 (UTC)

Quote from article conc: Core reflects H field and directs power flow

The H field max in the single ph trans shown in the paper occurs at the centre of the gap between the pri and sec windings. As we have noted before, the H field tapers off as you progress toward the core and is very low inside the core and even lower outside the transformer. In this sense therfore, the H field is trapped in the gap and is essentially constant from pri winding to sec. In this sense also, it could be said that the core also guides the H field (as well as the B field inside of course) --Light current 21:36, 22 May 2006 (UTC)

OK, I see what LC is getting at. You're not saying the core conducts the H-field like a wire conducts a current but are instead saying that the geometry of the core concentrates the H-field within the gap between the arms of the core. Am I reading you correctly? Alfred Centauri 22:00, 22 May 2006 (UTC)

Yes this appears to be what the paper is saying. I should not have said 'conduct'.It sorta makes sense to me 8-)--Light current 22:14, 22 May 2006 (UTC)

3 ph from 2 ph

Dear Sirs,

Is it possible to power a three phase condensor unit, atop a building supplied with two phase power?

Sincerely, Sergei