I don't think I have to repeat what I already wrote in the last thirty or so pages. Re-read them if you want.
You touted this 'two capacitor' example as the (yet another) definitive circuit that would jave brought 'team Lewin' to the ground.
It turned out not to be at all different from the original ring with two resistors.
In steady state resistances are exchanged with reactances. The induced field drives charges on the plates of the capacitors and instead of a conduction current we end up with a displacement current.
The setup is the same as that I used for the ring with the two resistors: a ring with two resistors around a toroidal core. There are no surprises. Everything is as it should be: a drop of 308 mV on the 4.7uF cap, negligible - basically zero - drops in the copper section, and a drop of 66 mV in the 22 uF cap.
This time I didn't attach the scope's probes, but I would have seen the same 180 degrees phase inversion.
If you want to waste time, you can compute the reactances, and the current.
I am asking you again: why did you insist this example was something special? It was a dud.
I am asking you again: why did you insist this example was something special? It was a dud.
Will a ring made exclusively out of capacitors show an induced emf? No wires, only capacitors in series replacing the ring, as well as the two original resistors on each side. What would be the voltage VAD in this case? Do you agree that now that there are not 'Lewin zero volt' short circuits anywhere?
I am asking you again: why did you insist this example was something special? It was a dud.
Will a ring made exclusively out of capacitors show an induced emf? No wires, only capacitors in series replacing the ring, as well as the two original resistors on each side. What would be the voltage VAD in this case? Do you agree that now that there are not 'Lewin zero volt' short circuits anywhere?
And here we go again.
After the nth "killer question" is answered, not a comment.
And then "killer question" n+1 takes its place.
Answer that and the cycle repeats.
What if we made a ring of alternating capacitors and thermistors?
And a ring with thermocouples orbiting a neutron star?
What about a ring made of anti-matter?
And what if we changed the gravitational constant of the universe?
You are just wasting people's time.
When Lewin argues that the voltage between the terminals of the opposing resistors in the ring is zero or that the induced voltage in the voltmeter probes is also zero, he is incorrect, and that is the source of this controversy. The guy messed up and he won't admit it. He even went so far as of trying to change the definition of KVL. Too bad for Lewin that Maxwell himself wrote the definition of KVL in his 1873 book, which contradicts what he says.
Let the components of the current at any point be u, v, w.
[...]
By Ohm's Law the current is proportional to the electromotive intensity. Hence X, Y and Z must be linear functions of u, v, w. We may therefore assume as the equations of Resistance,
X = R1u + Q3v + P2w,
Y = P3u + R2v + Q1w,
Z = Q2u + P1v + R3w.
So, the resultant EMF inside a wire no matter its multiple origins must be its current multiplied by its resistance. So much for modelling a wire as a resistance in series with a battery. This agrees perfectly with what is being said all the time: the wires and resistors in Lewin's circuit only have a voltage that is their respective resistance times the current intensity and nothing else.
Also you did not answer my 2nd question: how one can practically demonstrate path-dependency using transformer with let's say, 100 turns of secondary.When you measure from the outside you can look at the measure in two ways:
the voltmeter shows the actual voltage in the gap between taps
the voltmeter shows the voltage along the portion of filament between the taps (which is the ohmic drop that is nearly zero volts) plus (or minus depedending on orientation) the linked emf (emf of one turn times integer number of turns linked).
This has been one heck of a ride.
Really the first thing I need to get off my chest is, I really appreciate the work by bsfeechannel and Sredni, and later thinkfat. Thank you for not giving up, and I also need to apologize. I don't remember what I exactly posted on the previous round (is it already two years or what?) but it was about bsfeechannel's writing style and completely uncalled for. Seeing how much pure stupidity you have to deal with, I admire the coolheadness you can still maintain.
Thank you for revisiting this again, because the fact that time has passed helps to see it in new light. It's also amusing to see how many have "converted". For the rest, quality is indeed going down.
Additionally to thinkfat doing full 180, my real eye openers were actually helpful comments by jesuscf and Jesse.
These guys repeatedly called all the math used gibberish, confusing, etc. etc., and it really hit home because I agree with them. Math as shown by Sredni is pretty much gibberish to me, as well. But there is one big difference; while hate admitting being wrong as well, I hate intellectual dishonesty even more. I simply can't go there.
Which gets us into the matter of credentials as discussed repeatedly. Personally, still in high school I really shone in math. Those integrals were a breeze. Coming to university however, the first few courses were still with full scores, but then my scores started dropping. It seems, while integrals as taught in high school were easy, now when multiple integral symbols with those small extra symbols like circles started to appear, number of different alphabets and writing styles to signify different entities blew through the roof, it become harder. I was able to pass exams still but had fundamental issues digesting it.
It didn't help that my curriculum was kind of special mix, originally designed to mass-produce design engineers for still growing Nokia (which later colossally failed), but I'm not sure if that mix worked that well. Only one course in very fundamental electromagnetism (dealing with these subjects) so while we had all the related math, it was disconnected. I never learned the Maxwell equations well. Then we mostly had signal processing, some computer science, some electronics, etc.
Finally, I failed the exam in vector analysis. This was the first time me literally failing math. In the end, I admitted the fact that I'm just not a math wizard, and I don't need to be. Different people have different skills. I know enough math to know what I need to know, so that I can ask for help from the right people if the task requires it. This also makes me appreciate those who can deal with complex math. Their skills and understanding in electromagnetics have enabled all the nice modern things we have, so that more simple-minded engineers like myself can just buy a lumped module which deals with all the voodoo internally.
And now we get back to the eye opening moments. The bullshit generated by J & J in this thread could work for some readers, say, a carpenter or plumber with no university math background.
But for me, it doesn't work. Sredni's math might look gibberish-y enough for me so that I understand how J & J feel about it, but on the other hand, I still remember enough of the Vector Analysis and Electromagnetic fields and waves that I can see chances are very high Sredni knows exactly what he's talking about. Even if that is unsure, it's certainly sure that J & J have absolutely no idea about the math involved whatsoever, even if they are capable of copy-pasting some terms and putting them together to form sentences cargo cult way - think about "lumpable paths" & co.
So, finaly I asked myself what is the real dispute here, behind the endless loop of fog screens.
It is the question:
Is Lewin's original circuit lumpable?
Because non-lumpability is hard to prove (like nonexistence of God), requiring mathmematical concepts not everybody seemingly understand or agree about, the burden of proof has to be turned around: prove that KVL works.
And I think the only sustainable, intellectual honest way of debating about this would be this very process:
* Show an equivalent circuit -> show that the real circuit behaves like that equivalent circuit
* Now disputing this is easy; just show how the given equivalent circuit does not work. Single datapoint suffices.
* Now the author is required to come up with a new equivalent circuit, until it works - or doesn't.
And this is what I have not seen. I'm serious. I have not seen a circuit diagram, showing an equivalent circuit, and Kirchoff based calculations that get the result matching with the experiment.
Instead, I have seen videos of special 3D arrangements made for the task.
So I think I know where the issue lies.
It's the fact that a 2D circuit drawing with nodes, KVL applied, is not sufficient alone. To get it working, extra "hidden" information is added; not necessarily even hidden, not everyone is dishonest with this, but nevertheless it seems the exact 3D construction is the crucial part. This cannot be conveyed with simple circuit mesh diagram, otherwise it would have been done already.
So the model is not a equivalent circuit that can be drawn as a diagram; no, the model is a photograph or usually even a video showing a careful 3D construction. Only within this model, the KVL seems to hold.
As soon as you change this model - while the 2D circuit diagram keeps the same! - the model breaks up. This is then called "bad probing".
But Kirchoff laws and simple 2D circuit diagrams with lumped components were never supposed to cover such complex cases. Sure, you can force this approach, by adding new layers of information (like, add a photograph describing the exact layout required, to the point of showing correct ways of probing), but then the question is, is this way of modeling beneficial? Or maybe going to the lower physical level with wider generalizations, using known, true and tested principles - for example, Maxwell equations on paper, or modern EM field solvers/simulators - would make more sense?
In any case, it's quite a mystery how this discussion actually started, and though I originally contributed this to Lewin's "flashy" way, I now think it's more because of how Mehdi represented it. After all, many have seen the Lewin's lecture, and it did mean nothing until Mehdi "set up the stage" so to speak, introducing the concept of "being wrong", which is actually quite ridiculous if you think about it deeper.
Lewin's lecture isn't that special; it's completely normal to see lecturers show examples "how not to do something". It's quite a stretch to think this shows that "the lecturer doesn't know how to do it". No shit Sherlock, that's the whole point. I could understand this from someone who has never attended university lectures, but it's surprising to see from Mehdi.
Really, the eye opener should be the fact that "correct probing" requires carefully thought out geometric constructions. This is not what I mean when I, as a practical engineer, talk about correct probing. For me, correct probing means avoiding extra loop caused by the scope's ground clip, instead using the small springy thing to connect to the terminals of the small output capacitor, for example, directly. But this is only possible if the thing to be measured is a small physical point (like a 0805 capacitor). In other words: a lumped circuit, or a close approximation of it! If you are measuring between wide area of circuit, under influence of external field, then obviously the presence of leads cannot be avoided in any way. They become parts of the circuit. The definition that "correct probing" is the one that happens to give consistent results per equivalent circuit + KVL is of course backwards. We should use capable enough models so they can model the real world, not the other way around (constraint the real world until it matches with the simple model), because latter limits out capability of building useful circuits. We can do better.
If you don't know how to use KVL, it doesn't mean that KVL doesn't work!
If you don't know how to use KVL, it doesn't mean that KVL doesn't work!Look at it this way: Lewin has shown an experiment where clearly the measured outcome doesn't match with KVL. This is undisputed. You guys are spending great effort in building significantly different experiments (to the point they look something completely else even to a plumber's or carpenter's eye) that shows that in these experiments, KVL holds. Yet when suggested to add small modifications to the experiments, you won't do it.
Except that the measured outcome of Lewin's experiment DOES match KVL calculations! Just draw the correct equivalent circuit, solve, and presto: the KVL solution matches experimental measurements!!! Maybe I am pointing out the obvious, but if you don't solve the correct circuit you'll get an incorrect answer.
2) This circuit has been solved correctly in electromagnetics textbooks by experts in the field.
A consequence of Faraday’s law of induction is that Kirchhoff’s loop rule (which states that [integral]E · ds = 0 around a closed path) is no longer valid in situations where there is a changing magnetic field. Faraday has taken us beyond the comfortable realm of conservative electric fields. The voltage difference between two points now depends on the path between them. Problem 7.4 provides an instructive example of this fact.
3) Both Jesse and I have done experiments that demonstrate that KVL and measurements results match perfectly. Just need to be carefull when measuring.
4) Many people have debunked Lewin's claims, including Mehdi, RDS accademy, and Mabilde. Neither Jesse nor I are the first ones.
QuoteA consequence of Faraday’s law of induction is that Kirchhoff’s loop rule (which states that [integral]E · ds = 0 around a closed path) is no longer valid in situations where there is a changing magnetic field. Faraday has taken us beyond the comfortable realm of conservative electric fields. The voltage difference between two points now depends on the path between them. Problem 7.4 provides an instructive example of this fact.
Purcell & Morin, Electricity and Magnetism, 3rd Edition, Chapter 7.6
Check example 6.6 from Electromagnetics by Notaros. That is how Lewin's circuit should have been solved. Pay attention to Figure 6.10b where the equivalent circuit to be solved is correctly presented.
EDIT: " [integral]E · ds = 0 around a closed path" is not the general definition of KVL. KVL must include all EMFs, included magnetically induced EMFs. Says who you may ask? Maxwell himself!
Check example 6.6 from Electromagnetics by Notaros. That is how Lewin's circuit should have been solved. Pay attention to Figure 6.10b where the equivalent circuit to be solved is correctly presented.
EDIT: " [integral]E · ds = 0 around a closed path" is not the general definition of KVL. KVL must include all EMFs, included magnetically induced EMFs. Says who you may ask? Maxwell himself!
Soooo... the EMF is located on top of the resistors, it seems. Half just above R1, and half just above R2. How many centimeters, exactly? The text does not say. Can you locate with a bit more accuracy? No?
Or maybe...
Maybe that's the "equivalent circuit" that allows you to "solve the problem from the circuit theory point of view" and that is one of the introductory textbooks that do not explain clearly to their easily distracted audience what they intend for V. Oh, wait, but it does explain what V is! Page 269, eq. 6.18
Eq = - grad V
(Eq is what I call Ecoul) and V is... the electric scalar potential. Only half of the potentials required to describe the total electric field. And the text also says so explicitly on page 277, formula 6.43
E(t) = - dA/dt - grad V
"We see that both potentials are needed for E..."
(the same expression I used to express Etot = Eind + Ecoul, even if recently I decided to call the scalar electric potential phi, instead of V - exactly to avoid this kind of confusion you are having)
So...
where is exactly the EMF, again?
(Lewin problem is solved as an exercise on Purcell, Morin: Berkeley Physics vol 2, Electricity and Magnetism 3rd edition)
(Yes, I have checked with the kitchen. They do have deja-vus)
Also you did not answer my 2nd question: how one can practically demonstrate path-dependency using transformer with let's say, 100 turns of secondary.When you measure from the outside you can look at the measure in two ways:
the voltmeter shows the actual voltage in the gap between taps
the voltmeter shows the voltage along the portion of filament between the taps (which is the ohmic drop that is nearly zero volts) plus (or minus depedending on orientation) the linked emf (emf of one turn times integer number of turns linked).
Sorry, I do not see how this demonstrates path-dependency. Let's say - I measure 10VAC on 100-turn secondary of my transformer. Please tell how to set-up 2nd AC voltmeter that would demonstrate "path-dependency", measure 0VAC while connected to same terminals as 1st voltmeter measuring 10VAC? Do you agree that 100 turns of 2nd voltmeter test leads on transformer needed?
Lewin's team, please come down to the Earth and answer this simple question - about how to demonstrate path-dependency on real transformer with 100-turn secondary using two AC voltmeters one showing nominal AC volts, another 0V. Many will be pleased to discover something they do not know yet. [disclaimer] I am usually not that persistent, but I feel like Killer Question (© Sredni) has been asked.
Yes, the current question is always the killer question.
Until it gets answered. They it becomes the forgotten question.
And killer question n+1 makes it appearance.
Lewin's team, please come down to the Earth and answer this simple question - about how to demonstrate path-dependency on real transformer with 100-turn secondary using two AC voltmeters one showing nominal AC volts, another 0V. Many will be pleased to discover something they do not know yet. [disclaimer] I am usually not that persistent, but I feel like Killer Question (© Sredni) has been asked.
It didn't occur to you that university stuff is not supposed to be down to Earth. It's not supposed to be dumbed down and simplified. This is about the theoretical basis, which has to be understood when designing RF layouts, or when designing EM field solver simulators. Someone have to do those, even if you only buy a transformer or pre-certified RF module and use them within the datasheet conditions.
And indeed, designing an actual RF communication circuit would be a perfect example where Kirchoff laws do not apply. This is obvious to anyone seeing an antenna, a wire that goes nowhere, yet radiates energy out. Still, I'm sure if you try enough, you'll be able to somehow make use of Kirchoff laws by modeling the transmitter and receiver pair as a schematic. After all, you can use infinite number of circuit elements. But is it the right way?
If I understood your question right, one of the many possible answers to your question would be, by piercing through the possible packaging and looping the probe wire 100 turns around the core. This is completely unfruitful. No one uses Maxwell equations in everyday engineering with everyday off-the-shelf transformer, the whole idea of that component is that it performs as a lumped circuit, a black box.
This was never anyone's intention. You have been only fighting against made-up strawman for years. Come on! Let the academics do the academic discussion their way; the track record is excellent. We "practical engineers" should learn something about it, even if we don't go fully there. Forget the ego for a while.
Thing is, if you had some basic understanding of the matter, you'd know that your question is irrelevant to the matter.
Thing is, if you had some basic understanding of the matter, you'd know that your question is irrelevant to the matter.
Really? - Most pathetic excuse of no explanation ever seen. If you are not able to explain your "science" to your grandmother - you do not know it well enough.
Thing is, if you had some basic understanding of the matter, you'd know that your question is irrelevant to the matter.
Really? - Most pathetic excuse of no explanation ever seen. If you are not able to explain your "science" to your grandmother - you do not know it well enough.
I'm not spoonfeeding you. The answers you're looking for are all in this thread. There's no value in repeating them over and over.