Kirchhoff's Loop Rule Is For The Birds

[John Heath]

This statement was made by Dr Walter Lewin. There are great videos by Mr Lewin on fundamentals of Coulomb force and magnetism. He knows his onions. His argument against Kirchhoff's Loop Rule can be seen here.



Mr Lewin is a great physics teacher but on this one issue I am on the fence. A  Dirk Van Meirvenne felt the same way to the point of making a you tube video counter argument. It was well thought out covering every small detail of the Walter experiment. What followed was an example of the power of the internet. None other than Mr Lewin himself joins in the comments section of the video arguing his original position. It does not get better than this. The Dirk Van Meirvenne video.



 

[HalFET]

If this is the experiment I think it is (can't watch video on bus), then he didn't take into account that his wires and voltmeter are non-ideal. Kirchoff only works if all significant variables are accounted for. And honestly there are bigger f-ups in quantum physics when it comes to electricity.

[orin]

I thought we just went through this one a couple of months ago. 

There are no quantum effects here, just basic first year college electromagnetism.

Yes, the circuit theory we know and love assumes that there are no changing magnetic fields* - which patently isn't the case when you have a bloody great big solenoid going through the middle of your circuit and you put a step voltage across the solenoid.

There's really nothing to see here.  Unwanted changing magnetic fields induce unwanted currents/voltages.  That's why we minimize loop area in our circuits and use twisted pairs... something you can't do when the above mentioned solenoid is in the middle.

*we get away with the magnetic fields in inductors as they usually aren't within another loop of the circuit... though it is a problem with RF filters that are physically close together - you'll see coils and toroids mounted perpendicular to each other to minimize the effect.

[John Heath]

I thought we just went through this one a couple of months ago. 

There are no quantum effects here, just basic first year college electromagnetism.

Yes, the circuit theory we know and love assumes that there are no changing magnetic fields* - which patently isn't the case when you have a bloody great big solenoid going through the middle of your circuit and you put a step voltage across the solenoid.

There's really nothing to see here.  Unwanted changing magnetic fields induce unwanted currents/voltages.  That's why we minimize loop area in our circuits and use twisted pairs... something you can't do when the above mentioned solenoid is in the middle.

*we get away with the magnetic fields in inductors as they usually aren't within another loop of the circuit... though it is a problem with RF filters that are physically close together - you'll see coils and toroids mounted perpendicular to each other to minimize the effect.

I do not see quantum mechanics in here either. It is just classical issues of a changing magnetic field. I think it boils down to not using a proper toroidal transformer and not accounting for scope probe ground lead being effected by the changing magnetic field. 

[John Heath]

I'm not sure if you have been constantly trolling this forum with your quantum understanding of the world or the creed for it, but my advise is that no EEs make money by understanding how down to quark level things work. We make money by understanding the macro world it behaves in.

I don't care what is the mechanism down to the quantum level, all I care is how it behaves in a macro level -- I measure current, I don't count charge. There are physics forums out there and such questions should be posted there.

If you want to use a new theory to prove the old, commonly accepted theory is wrong, design an experiment and prove it, or use absolutely WATERTIGHT mathematical method to prove it.

And even if you can prove it is wrong, if you can't design an experiment to show it, it's likely that it won't affect us, at least not with current technology.

There is no quantum mechanic here. Just Kirchhoff's Loop Rule in an expanding magnetic field. I am not a troll. I just have a keen interest in the fundamentals of electronics. We are getting off on the wrong foot here. Best to nip this one in the bud before it festers. When I asked you how you made a home brew spectromass meter I was not being sarcastic. I do want to know. And if I said something that has turned you off I apologize.

[John Heath]

If this is the experiment I think it is (can't watch video on bus), then he didn't take into account that his wires and voltmeter are non-ideal. Kirchoff only works if all significant variables are accounted for. And honestly there are bigger f-ups in quantum physics when it comes to electricity.

On a bus ? A cell phone no doubt while being handed off from RF tower to RF tower while the bus moves. We live in good times. It is the video you are thinking of. Hope you have time to watch it when you are home as the Dirk Van video is an interesting and detailed repeat of the Walter experiment.

[John Heath]

When I asked you how you made a home brew spectromass meter I was not being sarcastic. I do want to know. And if I said something that has turned you off I apologize.

I was totally not mad on that. I didn't reply because I didn't want to hijack the thread, and I didn't want to bump the other thread while there were already a good discussion on MS technology, rather than IMS.
What I don't like is the attitude, particularly the questioning attitude towards what most people consider correct.
I did exactly the same before, but over time, I realized that 99.999% theories in the books are correct, and suspicion only makes me paranoid, while reality proves that classic theories are almost always correct, at least at a macro scale.
I have no interest in being another Einstein, and I will keep my feet in the engineering field for good. Science? Not for me, and I believe most people here believe in practical engineering more than fundamental science somewhere up in the sky.

Agree. It is human nature that everyone likes a good camp fire story. When QM says entanglement it is "oooo star trek" so that becomes the popular belief. The voice of reason with a simple classical explanation , Bohmian mechanics , is as boring a fish wrapped in news paper. By the by there is a theard going on now on mass spectro meters. I am not nagging I'm just saying.

[Zero999]

What I don't like is the attitude, particularly the questioning attitude towards what most people consider correct.
I did exactly the same before, but over time, I realized that 99.999% theories in the books are correct, and suspicion only makes me paranoid, while reality proves that classic theories are almost always correct, at least at a macro scale.

I have no problem with that attitude. I think it's a good thing, as long as the person listens and is willing to lean. I can appreciate it being annoying, especially when some questions seem silly, but it's still a healthy attitude to have.

[T3sl4co1l]

Kirchoff works at any pointlike node.

If a circuit is slow relative to its size, then currents will not be phase shifted due to gross delay, and current in = current out of a wire.

If the circuit is fast relative to its size, so that phase shifts are significant, you must maintain the pointlike condition.  Instead of a current flowing in and out of a wire, you have a transmission line pair, launching at point A (where the wire and GND return currents match) and terminating at point B (where the same is again true).

Transmission line analysis takes you several decades higher in frequency, until that model breaks down again -- when the cross section of a transmission line is no longer pointlike, but supports higher (TEM/TE/TM) modes, it behaves as a waveguide instead.  Then you have no choice but to treat every point in space as a pointlike Kirchoff node, in other words, use the continuous-space, 3D field equations.

Transmission lines are the second most general circuit analysis tool.  RLC equivalents are 3rd, but the most commonly taught (to the extent that most people actually believe resistance, capacitance and inductance are elemental).  The absolute most general (and physically realistic) method is simply E&M analysis, these days usually in a FEM tool, where the spacial fields and AC steady state or transient time evolution are produced.

Tim

[rfeecs]

As Orin said, we went through this in another thread a few months ago.  He pointed out a paper that describes exactly the experiment that Dr. Lewin performed.  The paper describes it very clearly and in detail and it all makes sense:
http://www.phy.pmf.unizg.hr/~npoljak/files/clanci/guias.pdf

A lot of confusion is created by Dr. Lewin's view of the definition of Kirchoff's law, which is described here:
http://web.mit.edu/8.02/www/Spring02/lectures/lecsup4-1.pdf

He also refers to this lecture supplement:
http://web.mit.edu/8.02/www/Spring02/lectures/lecsup3-15.pdf

His view is that Kirchoff's law does not apply to any circuit containing a changing magnetic field.  In this case, Faraday's law applies:


He always strictly separates the two sides of this equation, and when he sums up voltages along a path, he only counts the E dot dl part as voltage, and then this equals the other side caused by the magnetic field.

So his view is that Kirchoff's law does not apply to circuits with inductors.  Most electrical engineers where not taught this way.  But Dr. Lewin insists this is the only correct way to look at it, and any other way is DEAD WRONG.

[John Heath]

There are different ways of looking at it. From my shoes the test was incomplete. The current to voltage is displaced in time during change. For this reason thousands of measurements have to be made for the increase and decrease of the magnetic field. Only when everything settles down do you average out all measurements and find that it does equal equals zero. To take one measurement when the inductor current is displaced in time is not a complete test.

[mtdoc]

I've watched some of these videos as well as many of Dr. Lewin's MIT electromagnetism course lectures.

As a former college teacher, I alternate between thinking that he is really serious about his claim that almost all college physics teachers don't understand this and have it wrong or that the gleam in his eye (and his obviously purposeful court jester dress) means that he is really just trying to be provocative and get students to really *think* - Forcing them to grapple with the underlying principles at play rather than just accepting the KCL/KVL dogma. 

I know by forcing me to think about what is going on I've gained a deeper understanding (along with some confusion ).  That's what a true teacher always strives for.

On the other hand, he may just be getting a little disinhibited with age...

[RoGeorge]

I've just watched those videos, and there is no mention there about quantum mechanics. I don't understand why in the first few replies of this thread the QM is repeatedly mentioned.

The videos are about basic electromagnetism, and there are no "contradictions", so the reply video has nothing to "solve", IMO.

Regarding the experiment, if we take into account the magnetic circuit, too, not only the electric circuit, then there is nothing to write home about. Also, if we want to be pedantic about it, then each of the 2 voltmeters are measuring the induced voltage in their own loop, loop made by the measuring probes.

Maybe I am missing something very important, because I don't see what's the point.
- Kirchhoff's laws are for DC circuits, not for AC. IMHO, to mention Kirchhoff is a little offtopic for a transient regime.
- In a constant field, the energy to move electric charges depends only of the start point and the end point. The path between start and end point does not matter (i.e. the shape and position of the probes).
- In a variable field, the energy to move electric charges depends not only of the start and the end point, but also depends of the path we move from start to end (i.e. depends of the position and shape of the probes, because the wire inside the probes determines the path for our charges here). The importance of the charge's path is pretty obvious if we keep in mind that an integral is just a sum of many pieces. If you start altering the pieces (i.e. varying the field) while we are summing the pieces, of course we will get various results for various charge's paths.

We are talking here about energy instead of voltage because, by definition, voltage is the energy to move a unit of charge, and is measured in Joules per Coulomb, AKA Volts.

To sum it up:
1. No AC, no problems with the probes. (AKA the path of the electrical charges does NOT matter, AKA the measured voltage does not depend of the path, AKA the shape described by the probes of your voltmeter doesn't matter)
2. If you have AC, be careful how you lay down your probes. Induction will mess up the readings. (AKA the path of the electrical charges DOES matter, AKA be careful how you lay and shield your probes, or else you'll measure something else than intended)
3. Kirchhoff is for DC, not for AC, so leave Kirchhoff's laws out of this, they are not applicable.

My personal interpretation is that the Prof. Walter Lewin agrees with 3., but he is just teasing us in order to make his students more aware about 1. and 2., unless I am completely missing the point, of course.
Regarding the shirt, he was always wearing funny shirts during the filmed classes. Anyway, a big thank you, Prof. Walter Lewin, for all the physics classes, and for all the passion put in teaching.

[nctnico]

I watched about a quarter of the video and then got bored. What is (obviously) missing in the first diagram is an AC source which represents the induced voltage. Add that and all laws of physics are met. If I had a teacher like that I'd want my money back. I want to be told how things work and not decipher half arsed riddles which presume the laws of physics are wrong.

[helius]

It isn't surprising that the automatic response from engineers to this example is so negative: using the laws of electromagnetism would require calculus.

[John Heath]

There was another fine video made on the Lewin's view of Kirchhoff's Law by Cyriel Mabilde. By this time I had already formed an opinion and put it the comment section. What I said is just an opinion. I want to know what others think of this counter argument video to the Lewin's position. Also it would not be a surprise if Mr Walter Lewin himself leaves a comment as he is the real deal.

[RoGeorge]

As Orin said, we went through this in another thread a few months ago.  He pointed out a paper that describes exactly the experiment that Dr. Lewin performed.  The paper describes it very clearly and in detail and it all makes sense:
http://www.phy.pmf.unizg.hr/~npoljak/files/clanci/guias.pdf

A lot of confusion is created by Dr. Lewin's view of the definition of Kirchoff's law, which is described here:
http://web.mit.edu/8.02/www/Spring02/lectures/lecsup4-1.pdf

He also refers to this lecture supplement:
http://web.mit.edu/8.02/www/Spring02/lectures/lecsup3-15.pdf

His view is that Kirchoff's law does not apply to any circuit containing a changing magnetic field.  In this case, Faraday's law applies:


He always strictly separates the two sides of this equation, and when he sums up voltages along a path, he only counts the E dot dl part as voltage, and then this equals the other side caused by the magnetic field.

So his view is that Kirchoff's law does not apply to circuits with inductors.  Most electrical engineers where not taught this way.  But Dr. Lewin insists this is the only correct way to look at it, and any other way is DEAD WRONG.

Thank you for mentioning those 3 PDFs. 

Now, it's all clear, and again, it does not raise any contradictions with what I was taught. I remember here (in the former communist country Republica Socialista Romania) we were always using Kirchhoff to solve only DC circuits, without any variable magnetic fields. Kirchhoff was introduced in general school at about 14 years old (in the Physics manual for the 8th grade, http://manualul.info/Fizica_VIII_1987/). Since the 8'th grade to university we were always using Kirchhoff for DC only, where time (or any other kind of variation) was not involved. I don't recall stretching Kirchhoff into using it for alternating current, variable fields or any other kind of transient phenomena.

While the Prof. Lewin's physics is correct, I don't understand why he is trying to disprove a claim that was never made, that claim being "Kirchhoff can be used with transient phenomena".

My wild speculation here will be that the professor noticed that some students were trying to apply Kirchhoff in conditions where those laws were not applicable, so the professor needed to underline that the Kirchhoff's law are applicable only in certain conditions.

Unfortunately, he picked up a title, "Kirchhoff's Loop Rule Is For The Birds", which made everybody goes into denial. If he would have titled it "Kirchhoff's is for DC only, with no variable fields" or "The voltage depends of the charge's path in non conservative fields", then there will be nothing to debate, I guess. Clickbait titles are a waste of time, if not straight dangerous.

  Thumbs down for the statement "Kirchhoff's Loop Rule Is For The Birds". That is not a scientific statement, it's a flame. 

[orin]

There was another fine video made on the Lewin's view of Kirchhoff's Law by Cyriel Mabilde. By this time I had already formed an opinion and put it the comment section. What I said is just an opinion. I want to know what others think of this counter argument video to the Lewin's position. Also it would not be a surprise if Mr Walter Lewin himself leaves a comment as he is the real deal.

This demo is not the same as Lewin's.

In Lewin's demo, the magnetic field is negligible outside the solenoid where the scope probes are attached (it's a long solenoid and it is stated that they measured it in one of Lewin's demos).  It is not the case with this coil.

Look at the scope probe ground lead at 5:02 in the video and the loop it makes.  We've all seen the magnetic field generated by a solenoid or magnet and how it curves away from the end (see https://en.wikipedia.org/wiki/Solenoid for a nice colorful picture).  Even though his probe and ground lead are in a vertical plane, there is still changing flux contained within the loop.  If he removed his board and just made a loop of his probe/ground lead in the same position, he would still get an emf indicated by the scope, so he is not just measuring the emf induced in his horizontal loop.

This video is more of a demonstration of what happens if your scope ground lead is too long!

There was some doubt raised about the total emf induced in Lewin's experiment.  But all he needed to know was the geometry of the solenoid and dI/dt.  He can easily measure dI/dt.  The flux at the center¹ of the solenoid is u_0*N*I*pi*r^2/l where N is the number of turns, r is the radius of the solenoid, l is the length of the solenoid and u_0 is the permeability of free space.  As long as the loop in the experiment is around the center of the solenoid, he can calculate the induced emf in the loop.

¹It's half that at the ends... another difference in this video.

[rstofer]

It isn't surprising that the automatic response from engineers to this example is so negative: using the laws of electromagnetism would require calculus.

But the results don't make sense in my physical universe.  Can I really get a sign and magnitude change by simply moving my probes?  Or if I use two meters?  What if the distance between A1/A2 and D1/D2 approached zero?  Am I really going to see in practice what the physics shows?  I doubt it!

Is there a video that shows exactly what Dr. Lewin described (in English)?
Without obvious measurement issues?

The video linked in Reply 20 shows that the effect described by Dr. Lewin could be the result of a incorrect measurement procedure.  In the video at around 22:00, the author talks about reassuring our grandkids that Dr. Lewin's Super Demo may have a problem.  How did he know my grandson is embarking on an EE program?

I think I lump this in with the elegant proof that 2 = 1:

https://www.math.toronto.edu/mathnet/falseProofs/first1eq2.html

One day I was walking my German Shepherd Dog as we approached a Christmas tree laying at the curb for disposal.  A gust of wind came up and the tree started blowing toward us, rolling along the curb.  Never in my dog's life experience had he been attacked by a tree.  While he would ordinarily defend against any threat, this was beyond his understanding.  He was used to peeing on trees, not fighting them off.

Same with me and Dr. Lewin's video.  I don't know what to make of it so I will choose to move along.

[rfeecs]

It isn't surprising that the automatic response from engineers to this example is so negative: using the laws of electromagnetism would require calculus.

But the results don't make sense in my physical universe.  Can I really get a sign and magnitude change by simply moving my probes?  Or if I use two meters?  What if the distance between A1/A2 and D1/D2 approached zero?  Am I really going to see in practice what the physics shows?  I doubt it!

Is there a video that shows exactly what Dr. Lewin described (in English)?

The distance between A1/A2 and D1/D2 don't matter.

This paper is the exact experiment Dr. Lewin does:
http://www.phy.pmf.unizg.hr/~npoljak/files/clanci/guias.pdf

This video is not in English, but is seems to be demonstrating the same concept, and Dr. Lewin commented favorably on it.  It looks more like the configuration of his experiment, with a longer solenoid in the middle:



You can think of it as a transformer.  The wires going to the multimeter form part of a secondary loop.  When you move the wires to the other side you reverse the direction of the loop and change which resistor is included in the loop.

Electrical engineers would account for the "voltage" induced across the wires and it should all add up and make sense.

[John Heath]

So the tree was peeing on the dog not the dog peeing on the tree. I can see how this would confuse him.

I like this video better as it shows the Walter Lewin point in a clearer way. He is putting his test leads right on the loop then twisting the leads to null out external magnetic fields. This way he can not measure a magnetically induced voltage in the loop as both the test lead wire and loop wire cancel each other out. This leaves only the voltage drop across the resistor being measured.

I can think of a practical purpose in this. If someone wanted to measure the RF AC being picked up by an antenna they would ground one test lead then reach up with the other lead to measure the RF AC at the top of the antenna. They would read 0 volts as the antenna and test lead are in the same magnetic RF field therefore canceling each other out. It would be confusing. However if that same someone was paying close attention to Walter Lewin they would not be confused by a 0 RF voltage reading at the top of the antenna. 

[rstofer]

You can think of it as a transformer.  The wires going to the multimeter form part of a secondary loop.  When you move the wires to the other side you reverse the direction of the loop and change which resistor is included in the loop.

Electrical engineers would account for the "voltage" induced across the wires and it should all add up and make sense.

I do think of it as a transformer.  In fact, I was immediately drawn to the manner in which my clamp-on AC ammeter coil works.  Current through the wire generates a current in a winding which flows through and deflects the meter and with a little scaling tells me how many amps are flowing.  Simple and intuitive!  I have been using an Amprobe for more than 50 years.  Got it down pretty good.

Metering type current transformers also come to mind.  A simple winding surrounding a conductor and we have a secondary current that has some relationship to the primary current.  Mine were usually on the order of 1000:5 or higher.

And, yes, all the voltages gains and drops around a closed path sum up to 0V - just like Kirchhoff said.
But you have to account for all of the loops...

I'll just concede to being an engineer and leaving physics to someone else.

[T3sl4co1l]

It goes to show, you cannot separate the measurement from the system.  It is physically necessary that a voltage measurement be performed between two points separated by space.  And that space necessarily can have nonzero magnetic flux, and therefore induction.

To think otherwise is only to fool yourself.

Drawing the incomplete circuit, with one loop enclosing flux, and arbitrary probing, is manifestly wrong -- the circuit can only be correct if all loops and induced voltages are included.  Then the flattened schematic representation is valid and usable.

That's why, as I said before, the only valid application of Kirchoff is at a pointlike node.  If the fields are negligible, then all points in the circuit are pointlike.  Such is the case at DC, and low AC frequencies.  At higher frequencies, one must restrict the region to smaller and smaller volumes or cross-sections, to maintain the pointlike condition.  One-dimensional transmission line analysis takes advantage of this.  At very high frequencies, where the fields interact with themselves (i.e., standing and propagating waves in waveguides, cavities, free space, etc.), it is only meaningful to use the infinitesimal form, which is then integrated over the volume in question, yielding the wave equations for the system.

Riddle: in the above picture or video, what voltage (in ideal, imaginary, absolute terms) is induced across the meter leads, in what locations?  Is it zero, or some fraction, positive or negative, of the total induction in the large loop?

It's a simple yet fun example.  Students are taught with overly simple cartoons -- like a region of flux extending out of the page, that's magically zero just outside of the region.  Yet even with an infinite solenoid (not a very clearly truncated one as pictured), the flux does not go to zero instantly outside the enclosed area of the solenoid; it does so gradually, and in a complicated manner (due to near-fields from the finite wires the solenoid is inevitably made from).   For a short solenoid, the flux is not at all axial, but crowning out like a fountain, so that there is a large radial component to the flux direction.  Consider this against the question I've just posed.

Tim

[orin]

So the tree was peeing on the dog not the dog peeing on the tree. I can see how this would confuse him.

I like this video better as it shows the Walter Lewin point in a clearer way. He is putting his test leads right on the loop then twisting the leads to null out external magnetic fields. This way he can not measure a magnetically induced voltage in the loop as both the test lead wire and loop wire cancel each other out. This leaves only the voltage drop across the resistor being measured.

What a wonderful demo!  Even though we don't see the phase difference between measuring from either side, you can see clearly that as you move the measuring loop, you get different results.  (Lewin's comment notes that he moved to his demo, rather than using a transformer with a single turn secondary, to show the phase difference.)

Now, to put the final nail in the coffin, you'd want to measure just across one of the wires connecting the resistors.  With the twisted pair to the meter and laying the leads as close to the wire connecting the the resistors as possible, what would you read on the voltmeter?

Answer: Darn close to zero, depending on how well you minimized the loop area between the measuring leads and the wire connecting the resistors.  (Doesn't matter where the changing magnetic field is coming from, if the loop area is zero, the flux is zero for all time, hence no induced emf!)

I must admit, that I didn't pay this demo that much attention when I watched Dr. Lewin's 8.02 lectures, but the attempts to prove him wrong made me revisit it.  I have seen nothing that disproves Dr. Lewin's point.

[Kalvin]

Would the demo be less baffling if Dr. Lewis had stated that he just showed a single-turn transformer with and an oscilloscope measuring the induced voltage across a loop including the measurement leads, and that Kirchhoff's loop rule may not apply to circuits with transformers, couped inductances and transmission lines?