Does Kirchhoff's Law Hold? Disagreeing with a Master

[ogden]

In case you insist that there is just one wire and just one current meaning no path for return current - then I say that you don't even have circuit, thus Kirchoff's Circuit Law do not apply. There is huge difference between "do not apply" and "do not hold".

Nope. Kirchhoff says clearly that the wires can be connected in an entirely arbitrary way. So there is no requirement for them to form a circuit or to provide a path for a return current.

Exactly. Kirchhoff says "wires", thus more than single wire you desperately insist on. When you have single wire - you do not have conditions to apply KCL. That's why I looked for return current which I BTW found in the "spatially distributed capacitor", yet you managed to shift goalposts and change your mantra into talk about dotted area, not whole circuit. It is like showing one body of mass to happily conclude that Newton's law of universal gravitation do not work. It's not even unscientific. It's utterly stupid.

[bsfeechannel]

Yes the source of the error is people like Dr. Lewin applying it directly to a circuit without consideration if its applicable in those CONDITIONS.

Nope. Lewin was absolutely impeccable in the "application" of the laws. He used Kirchhoff to show that it works when you do not have varying magnetic fields.

He substituted the varying magnetic field for the battery and then showed how wrong it is to apply Kirchhoff to a circuit with varying magnetic fields. You will measure different voltages, because they are now path-dependent.

And left the problem for his students to solve. The solution is Maxwell and only Maxwell.

Fix those conditions by properly modeling the thing as a circuit mesh,

We've already proved in this thread that Lewin's circuit can't be modeled by lumped components.

if not then forget about Kirchhoffs circuit laws and stick to Maxwell.

The problem with your reasoning is that you think that if things are connected like a circuit, then Kirchhoff MUST hold. You think that if you find a single circuit where it can't be employed to explain its behavior then the whole theory is cactus.

This is BULLSHIT.

There are circuits for which Kirchhoff holds and others for which it doesn't. By now you should have known the difference. And the difference is that if the circuit is drenched with a varying magnetic field, KVL is out. If not, then you can merrily use your KVL and your precious LTSPICE to model it.

Lewin's circuit, if you have not paid the due attention to it yet, is just a sophisticated version of Faraday's original demonstration of induction. Lewin is just reenacting the same experiment with two fancy oscilloscopes and two resistors of relative high value. But the experiment is essentially the same.

Saying Lewin is wrong is the same as denying Faraday's law.

Okay yes i was using more the software engineering definition of abstraction layers rather than the mathematics definition of abstraction. With the mathematics definition it is indeed the other way around.

Abstraction in software is commonly thought of from the point of view of the user, but it conforms with the general definition of abstraction in which the "abstracted" software is in fact more inclusive than the special case.

And i fully agree, all of the levels of explaining electricity work fine as long as you use them within those limitations. Hence why i never really had any problems applying KVL to Dr. Lewins experimental circuit from his lecture. Any paradox between KVL and Maxwell with that circuit is simply down to using them wrong.

We've already showed that your "modelling" just don't apply to Lewin's circuit. It introduces gross errors. If I connect the lead of MY voltmeter (not yours with 250mV in series with the probes) from one resistor to the other where a wire should be in your "model" I will measure 0.5V, whereas in Lewin's circuit this voltage is zero, and everyone measured exactly zero volts.

There can't be any paradox between KVL and Maxwell. Lewin changed the condition for the validity of KVL and invited his students to explain the new behavior of that circuit. Those that think that Kirchhoff applies to just about any circuit are trying to find each one a different explanation. While those who understand Faraday's law could explain what is going on immediately.

I'm pretty sure QED had nothing to do with the development of the first transistors.

Then the first working BJT device popped up by ]accident from a bunch of people trying to build a better solid state RF mixer.

Wow! Shockley, Bardeen and Bratain won the Nobel Prize by accident! Lucky bastards!

The usual transistors can still be explained rather well with Maxwells fields pushing charged particles around.

Isn't this what essentially QED is: electrodynamics applied to subatomic particles like the electron?

Mostly just a case of controlling where the charge carriers are rather than making use of any quantum mechanical effect.

Strange. I had the impression that electrons received and emitted photons when they change their quantic levels of energy in the electronic band structure.

I'm pretty sure very few forum members here work in a semiconductor fab, let alone one that works with such fine feature capabilities or work on building quantum computers. Hence why very very few engineers would have a good reason to dig deeper than Maxwell, heck for 95% of cases even Kirchhoff is close enough(As long as you know about the other 5%).

I also had the impression that every electronics engineer around the world learns at least some rudimentary concepts of QED as part of their regular course which is required to understand solid state electronics. But I may be wrong.

[bsfeechannel]

Exactly. Kirchhoff says "wires", thus more than single wire you desperately insist on. When you have single wire - you do not have conditions to apply KCL.

Alright. I created an ogden version of KCL fail for you. I hope it now can make its way into your everything-must be-a-circuit head.

That's why I looked for return current which I BTW found in the "spatially distributed capacitor",

Cool! Can you connect an ammeter and measure this current flowing through a... wire?

[ogden]

Alright. I created an ogden version of KCL fail for you.

Brilliant. Exactly what I was looking for. Could you sign it and put high resolution file on google drive? - So I can show what stupid ingenuity looks like.

[Berni]

if not then forget about Kirchhoffs circuit laws and stick to Maxwell.

The problem with your reasoning is that you think that if things are connected like a circuit, then Kirchhoff MUST hold. You think that if you find a single circuit where it can't be employed to explain its behavior then the whole theory is cactus.
This is BULLSHIT.
There are circuits for which Kirchhoff holds and others for which it doesn't. By now you should have known the difference. And the difference is that if the circuit is drenched with a varying magnetic field, KVL is out. If not, then you can merrily use your KVL and your precious LTSPICE to model it.
Lewin's circuit, if you have not paid the due attention to it yet, is just a sophisticated version of Faraday's original demonstration of induction. Lewin is just reenacting the same experiment with two fancy oscilloscopes and two resistors of relative high value. But the experiment is essentially the same.
Saying Lewin is wrong is the same as denying Faraday's law.

No i said that KVL holds in every circuit mesh model. Not every physical circuit, this is an important distinction so don't carelessly throw them in the same bag.

I'm not saying Lewin is wrong. Kirchoffs circuit laws indeed don't work in such a use case. All i did was show a method of using KVL in a way that does work in that circuit, useful in cases where you would want to apply other circuit analysis tools.

And i fully agree, all of the levels of explaining electricity work fine as long as you use them within those limitations. Hence why i never really had any problems applying KVL to Dr. Lewins experimental circuit from his lecture. Any paradox between KVL and Maxwell with that circuit is simply down to using them wrong.

We've already showed that your "modelling" just don't apply to Lewin's circuit. It introduces gross errors. If I connect the lead of MY voltmeter (not yours with 250mV in series with the probes) from one resistor to the other where a wire should be in your "model" I will measure 0.5V, whereas in Lewin's circuit this voltage is zero, and everyone measured exactly zero volts.
There can't be any paradox between KVL and Maxwell. Lewin changed the condition for the validity of KVL and invited his students to explain the new behavior of that circuit. Those that think that Kirchhoff applies to just about any circuit are trying to find each one a different explanation. While those who understand Faraday's law could explain what is going on immediately.

I did say there is no paradox in my own post so why are you arguing for the same thing.

Oh and if you do have a voltmeter that can integrate the work needed to move an electron along any chosen path id love to see a write up on its operation. Would give you a pretty good chance at a Nobel prize even.

If you are so good at it how about showing me how to correctly analyze this simple circuit:
https://www.eevblog.com/forum/chat/does-kirchhoffs-law-hold-disagreeing-with-a-master/msg2189216/#msg2189216


Feel free to use any method you want, as long as it shows what the circuit will do once given power.

...
Wow! Shockley, Bardeen and Bratain won the Nobel Prize by accident! Lucky bastards!
...
Isn't this what essentially QED is: electrodynamics applied to subatomic particles like the electron?
...
Strange. I had the impression that electrons received and emitted photons when they change their quantic levels of energy in the electronic band structure.
...
I also had the impression that every electronics engineer around the world learns at least some rudimentary concepts of QED as part of their regular course which is required to understand solid state electronics. But I may be wrong.

The inventors of the transistor certainly did amazing work in the field of semiconductors, but just saying that making a transistor was not what they ware trying to do when they made one. Its not the only major discovery that had a little bit of luck in it.

I was certainly ever shown any QED in lectures about semiconductors, tho to be honest those ware pretty dull lectures so i mostly just memorized enough stuff to pass the test, rather than show much interest. Its all mostly just electrostatic fields with a bit of electron physics thrown in. Nice to know about, but not terribly useful to know in deep detail. Much like teaching software engineers some assembler, not really practical for the majority of cases but good to know the basics.

Those electrons do certainly interact trough photons according to QED, but the overall behavior is more sensible to explain with Maxwell since it still works just fine in there.

[ogden]

Latest news

[sectokia]

In light of this debate firing up again on Mehdi's forums/reddit etc:

I just thought of an experiment that absolutely settles this once and for all:

-Remove the volt meters. No probes or extra loops are needed.
-Insert amp meters inline with the resistors.
-Put the resistors in known quantity of water such that by measuring the change in temperature, using the specific heat of water, you derive the power given off by the resistor.
-Having the resistors power and the current reading: You now can calculate the voltage across the resistor.

Notice that in such an experiment there is only ever 1 loop. So, there cannot be 'induction' in the meter probes, nor is there any 'secondary loops', nor can there be any 'bad probing'.

If you do this experiment: You end up with the exact same voltage as you do when you put the meters on each side.

So the presence of the meters does not matter!

This is proof the voltages summed around the resistor loop do not equal zero.

KVL DOES NOT HOLD!

Medih = wrong, Lewin = right.

[Zucca]

Now an ultimate idea/experiment to settle the discussion forever?

Not after 42 pages of technical discussion. Very unlikely.

 

[AVGresponding]

In light of this debate firing up again on Mehdi's forums/reddit etc:

I just thought of an experiment that absolutely settles this once and for all:

-Remove the volt meters. No probes or extra loops are needed.
-Insert amp meters inline with the resistors.
-Put the resistors in known quantity of water such that by measuring the change in temperature, using the specific heat of water, you derive the power given off by the resistor.
-Having the resistors power and the current reading: You now can calculate the voltage across the resistor.

Notice that in such an experiment there is only ever 1 loop. So, there cannot be 'induction' in the meter probes, nor is there any 'secondary loops', nor can there be any 'bad probing'.

If you do this experiment: You end up with the exact same voltage as you do when you put the meters on each side.

So the presence of the meters does not matter!

This is proof the voltages summed around the resistor loop do not equal zero.

KVL DOES NOT HOLD!

Medih = wrong, Lewin = right.

You'd have to thermally isolate the water extremely well, or heat leaking away would badly skew the result. The volume, pressure, and purity would also have to be very closely controlled.

[Berni]

How did this thread get picked up from the dead.

That suggested experimental setup is overly complicated for no reason. You can get thermal RF power meters off the shelf that already have a calibrated resistor and temperature sensor inside.

Results would be the same as with voltmeters anyway. There is just 1 loop, hence 1 current flowing around it, so same current in both resistors, then the rest is ohms law.

There is no paradox here, just two groups of people with a different view how the experimental setup is interpreted as a electrical circuit. One of the interpretations looking more correct from a physics point of view, other interpretation looking more correct from an electronics engineering point of view. Both interpretations are correct when put in the appropriate context.

[sectokia]

Medhi wasn't just claiming that KVL still holds, he made numerous other claims including that the voltage at the meters wasn't the voltage over the resistors, but instead was the voltage over the resistors, plus/minus the voltage 'induced' in the meter probes.

[EEVblog]

In light of this debate firing up again on Mehdi's forums/reddit etc:

Mehdi has a forum?

[m k]

First it goes so that circuit must be closed.
Later that changes so that thinking concentrates to a single spot.

EM of EMI is energy and 'I' is not unidirectional.

If I say that electricity can't be inside a copper I definitely must define things more closely, since even that copper atom is still copper, it's not how metal is usually understood.
An atom must have a fundamental inside and outside, where inside energy is atom's energy and outside energy is what we can call electricity.
From every electron that energy is also ball symmetrical, so at the beginning no quanta has a definite side.

So all electrons of all atoms of all materials have ball symmetrical energy quanta interactions, for electricity that includes a return path and everything between comes and goes.
What we can sense is statistical difference of all that.

So it doesn't mean that energy in cable flows in insulation, excluding local nuances.
It means that without an insulation the energy can more possibly go somewhere else.
Cable is of course a waveguide, but what part is meant to what function is not necessary a straight forward thing.

When we measure current, it's the same despite the side of a load.
So how is its energy, it isn't, it's just part of the vector.

[ballsystemlord]

There are 42 pages of text here, not to mention hours of videos linked.

Is there a set of posts/video combos that will accurately summarize the discussions herein?

I ask not because I doubt Kirchhoff, or anyone else, but because I thought this would be an interesting learning experience, and I'd rather not spend weeks trying to read/watch up on the matter. And some of the linked pics no longer exist. 404 not found....

Thanks!

[EEVblog]

There are 42 pages of text here, not to mention hours of videos linked. Is there a set of posts/video combos that will accurately summarize the discussions herein?

Basically, physicists and engineers think of things differently and explain things differently, and that's OK.
And Lewin's practial demo leaves a lot to be desired in terms of being a robust example.

[SiliconWizard]

There were a lot of side-discussions, but what it boils down to is definitions - for which Lewin is right, and as I remember, he has one written course that was more useful to get his point than several hours of videos and his attempts at demos.

One related point is that EEs usually think that a lumped-element model can model reality in a useful enough way, while being analyzable with the classic circuit analysis tools, which are easier to manipulate than EM equations, and much faster to simulate.

[m k]

One easy starting point is glass.

All our energy is the same, only originating material and frequency changes.

Seeing less through thicker glass is clear and easily accepted, it's just not transparent completely.
But situation change when viewing angle changes enough.
At some point happens a visible diffraction.

So a material has a very distinctive property that is actually very close to not being part of it at all, its surface.
That is also why and where frequency start counting heavily.

[HuronKing]

In light of this debate firing up again on Mehdi's forums/reddit etc:

Mehdi has a forum?

There is a fan Reddit group about him.

[EPAIII]

Someone asked for a summary. And you mention physicists and engineers. My degree is in physics and I worked for over 45 years as an engineer in electronics, so both in one.

The first time I saw Lewin's demonstration I just laughed. How a PhD and professor in physics could make such an elementary mistake in metering technique was totally beyond me. I spotted what was actually happening after just a few minutes. What has amazed me since then is how so many people can waste so much time on this. I even wasted some of my time trying to correspond with Lewin. Instead of discussing this, he just shut me off. He did not want to hear it and probably had realized he was wrong and could not defend his prior claims in an honest discussion.

My summary: Kirchhoff was a genius. Lewin is/was an idiot. And there you have it.

There are 42 pages of text here, not to mention hours of videos linked. Is there a set of posts/video combos that will accurately summarize the discussions herein?

Basically, physicists and engineers think of things differently and explain things differently, and that's OK.
And Lewin's practial demo leaves a lot to be desired in terms of being a robust example.

[RoGeorge]

summarize the discussions

1. Kirchhoff's second law (sum of voltages in a loop is zero) holds for conservative fields.
2. Kirchhoff's second law does not hold for non-conservative fields.

Both true.  Second one is more of a corner case, and in engineering it is dealt with differently.  Thinking it in terms of non-conservative fields is rather a mathematician's approach (a non-conservative field is a calculus thing).  Anyway, Kirchhoff never said it would cover such corner case.  His laws were developed in a different context.  What we learn now as Kirchhoff's laws is a better abstraction, though a little bit of the original context was lost with the new formalism.

Again, nothing wrong with that.  The professor is technically correct (he has a better formal understanding, it took me a lot of extra reading to get his viewpoint - at first I thought he's either wrong or joking - he was neither), but that doesn't change physics, and doesn't change engineering either.

What the professor is saying was always like that, just that not everybody realized it can be interpreted that way.  Engineers are also correct in their approach.  Kirchhoff's laws are watertight (his laws were about conservative fields).  The experiment is also correct.  Measurements are correct, too.  The voltage "paradox" is really there, and it is no paradox, that's how it is expected to be.

Everybody is right, and everybody is talking about the same thing, except that there is more than one way to skin a cat.  And that's why the 42 pages.