Sick of ridiculous KVL infighting

[instrumental]

40 pages and a few months on, some nerds don't seem to know when to quit arguing. The video Mehdi made about a "principled disagreement" with Lewin's lectures is rough as well, and now this has spiralled into numerous YouTube videos trying and failing to explain EM theory. I'm surprised to see so much discord in a community of professional electronics engineers.

I'm quite tired of seeing this, so I am hoping to provide a fairly definitive explanation of:

• Why KVL doesn't always work (e.g. why we don't use it for RF circuits or optics),
• Why KVL can be used to analyse low-frequency circuits (power, audio, etc), and finally,
• Where the gap is.

So, here we go.

Voltages measure electrostatic potential; the electric field is the gradient (slope) of the voltage, and the voltage is the potential function of the electric field.

For a potential function to be defined on the field, the field has to be conservative. For a field to be conservative, it has to be path-independent (you can calculate voltage between two points by taking any path through the field; the voltage will only depend on the endpoints). For a field to be path-independent, it has to be curl-free.

Imagine a toy boat circling a drain in the bathtub. We can divide the flow field into divergence-only (flow towards the drain) and curl-only (swirling around the drain) flows. Ignoring the divergence for now, imagine the toy boat is initially stationary. The flow will take the boat with it, and it will begin to pick up speed as work is done by the flow on the boat. Once the boat has made a full revolution of the drain, it is in the same initial place, but it is now in motion; it has acquired energy. This should illustrate to you that a circulating field is not a conservative field; we can't define potential energy between two points because different paths will do different quantities of work.

Now, we might ask, what is the curl of the electric field? This is given by Faraday's law of induction -- it's proportional to the derivative of the magnetic field with respect to time. If we have a changing magnetic field, we have a non-conservative field.

When do we have a changing magnetic field? Well, magnetic fields are made by currents; changing electric fields also create magnetic fields, which gives rise to EM wave propagation. So, any time we have a changing field or current (e.g. AC at any frequency), we can no longer define voltage.

There are two ways out of this. Physicists often work with 4-potentials in EM theory. A vector potential can be defined for the divergence-free field; a scalar potential (voltage) can be defined for the curl-free field, and the two together produce the full field. This magnetic vector potential is significant: the Aharonov-Bohm effect illustrates this nicely. Unfortunately, this doesn't mean we get KVL.

More commonly, approximations can be made which allow the re-introduction of voltages. If dB/dt is sufficiently small that it can be ignored, we can approximate the field as conservative. This is manifestly the case for power lines (50-60Hz) and holds (as a rule of thumb) roughly until the scale of the circuit under consideration is close to one-tenth the wavelength of the signal (e.g. a 1MHz signal has a wavelength in the tens of meters; you don't worry about MHz routing on a small PCB, but you run into RF effects with larger-scale antennae. By 1GHz, you're dealing with centimeters of wavelength and you're squarely in RF territory).

TL;DR field theory explains phenomena from DC, to RF, up to the highest-energy gamma rays, with visible light in between. You wouldn't apply KVL to RF or optics, would you?

[PlainName]

I'm quite tired of seeing this, so I am hoping to provide a fairly definitive explanation of:

Looks to me like you think your explanation is the authoritative explanation and, in order that no-one gets sidetracked by lesser explanations, you've started a new thread just for this. Why couldn't you just put it at the end of the current long thread? Imagine if every participant there started a new thread just to highlight their own authoritative explanation - the entire forum would consist of root posts about this stuff.

If you really want to make it count, delete this thread and just argue the merits in the proper place. Or just sigh and walk off knowing you are right and everyone else can wallow in their mistakes without further bothering you.

[Siwastaja]

you've started a new thread just for this. Why couldn't you just put it at the end of the current long thread?

While I generally agree that creating multiple threads about the same matter is a bad idea - also against forum rules -, in this special case that particular thread moves so fast, driven by what basically is a semiautomated spambot - which is also against forum rules - copypasting same responses and same dummy "questions" weeks and months straight. Any sensible comment gets lost in the noise in a matter of a day, which is probably exactly the reason for such smoke screen.

In reality, the fact is there is no large dispute. The few that run it make it look like large. Their capacity of even generating unique messages is exceeded long ago, which is why jesuscf just basically copypastes.

Quite frankly, it's nothing but playing games and throwing the toys and sand around, by the few remaining (is there more than 2 anymore?) people who can't admit their mistake or shortcomings, like many, including me, could (for me, it sure did take time).

This is why I would like to give a pass to this opening post which, despite being outside of the sandstorm thread, tries to actually be helpful and explain the issue.

Electromagnetism is non-trivial, and for "practical engineers", it can be damn difficult to admit our math is lacking. But I also can't stand intellectual dishonesty.

But really, the game is soon over. The quality of the "opposition" has gone down significantly, from the serious attempts of building usable equivalent circuits in SPICE, that model the experiment, to just rigging physical experiments with careful layout (which never was part of the experiment; this was the eye-opener for me; I can see the difference between circuit layout and good probing), shouting, and thinking that adding more exclamation marks makes KVL hold better.

And why I call it intellectual dishonesty? The answer is simple: because the people who built these Youtube experiments, who constructed the layouts that give "expected" multimeter number on photographs or videos, themselves witnessed the path dependency during tweaking of the layout, to get the error within the expected <2% or so. Think about it.

It sure seems more like flat-earth stuff every day.

[PlainName]

Any sensible comment gets lost in the noise in a matter of a day, which is probably exactly the reason for such smoke screen.

What will prevent this thread also suffering the same fate once it's been discovered?

If someone is disrupting a discussion through foul means then surely the answer is to ask a mod to deal with them. I don't see that giving them more threads to disrupt achieves much. Of course, perhaps the mods will be too busy, fail to see your point or otherwise let things run, but you don't know unless you ask.

But, on the whole, I am not averse per se with skimming the crust off into another smaller and newer thread. It was just the "Listen, my explanation is the right one, so I'm a special case" which poked me.

[nctnico]

you've started a new thread just for this. Why couldn't you just put it at the end of the current long thread?

While I generally agree that creating multiple threads about the same matter is a bad idea - also against forum rules -, in this special case that particular thread moves so fast, driven by what basically is a semiautomated spambot - which is also against forum rules - copypasting same responses and same dummy "questions" weeks and months straight. Any sensible comment gets lost in the noise in a matter of a day, which is probably exactly the reason for such smoke screen.

In reality, the fact is there is no large dispute. The few that run it make it look like large. Their capacity of even generating unique messages is exceeded long ago, which is why jesuscf just basically copypastes.

Quite frankly, it's nothing but playing games and throwing the toys and sand around, by the few remaining (is there more than 2 anymore?) people who can't admit their mistake or shortcomings, like many, including me, could (for me, it sure did take time).

This is why I would like to give a pass to this opening post which, despite being outside of the sandstorm thread, tries to actually be helpful and explain the issue.

Electromagnetism is non-trivial, and for "practical engineers", it can be damn difficult to admit our math is lacking.

Agreed. IMHO the whole point is to understand when a simplification works and when not. Showing electrical phenomena using simple multimeters also raises my eyebrows; I've seen that go wrong too many times.

When skimming through the other thread an interesting question popped into my mind: where does the electricity go that is being fed into an antenna (*). 

* Yes, this is formulated in a very poor way on purpose.

[Siwastaja]

What will prevent this thread also suffering the same fate once it's been discovered?

Nothing, and it will likely happen, after which moderators likely lock up this thread as a duplicate. Just saying I understand why the OP felt like opening a new thread.

If someone is disrupting a discussion through foul means then surely the answer is to ask a mod to deal with them.

Let's face it, that's not realistically going to happen. The discussion is on-topic, actually core of the forum, so this is not a valid excuse to intervene. All that is left is to carefully analyse what's going on, whether the means are foul, and do an informed decision. Moderation on this forum has never worked this way for a simple reason: lack of resources. Moderation is pretty much hit-and-run strategy, and I don't blame Dave & Simon, they have hands full and need to make quick decisions on simple matters (like totally offtopic politics).

So yeah, I pretty much agree with your points.

[Nominal Animal]

40 pages and a few months on, some nerds don't seem to know when to quit arguing.

Ah yes: because they disagree with you, they are nerds.

And you are professional, eh?

No, I don't think so.  I do believe asshat and a coward is a much more appropriate term for the likes of you.

I bet there is also a reason why you created a second account to post this thread –– which, by the way, is against the forum rules.  When your own argument is too weak to gain traction, what better to create an astroturfing account and call the others names and try to make them look ridiculous?

Oh fuck I hate these social games.

[bdunham7]

In reality, the fact is there is no large dispute.

Once I figured that out I tried--rather clumsily I suppose--to point out what I saw as the primary issue leading to the disagreement and later the misdirection of attention as to which 'path' was important--it took me a bit longer than I'd like to admit to catch onto this--that sort of made it look like a much more noteworthy issue than it really is.  But the main contestants weren't interested in any of that, just proving that they were superior and engineers are stupid.  And then along came some additional contestants to confirm that. 

The OP in this thread has a reasonable, short explanation for those that don't want to read 39 pages of name calling and drivel.  However, I would replace

So, any time we have a changing field or current (e.g. AC at any frequency), we can no longer define voltage.

with "there are multiple possible definitions of voltage" or something like that.  Proving that we can't define voltage by measuring something with voltmeters seems a bit silly.

[bdunham7]

No, I don't think so.  I do believe asshat and a coward is a much more appropriate term for the likes of you.

I bet there is also a reason why you created a second account to post this thread –– which, by the way, is against the forum rules.  When your own argument is too weak to gain traction, what better to create an astroturfing account and call the others names and try to make them look ridiculous?

Did I miss something?   

[SiliconWizard]

Proving that we can't define voltage by measuring something with voltmeters seems a bit silly.

Should be obvious... but that's very well put. =)

[Nominal Animal]

Did I miss something?   

Perhaps?  The OP is using a newly created account to call others nerds just because their own opinion is not gaining traction.

(That opinion being "this subject should be clear to everyone, so stop discussing it, it scares me".)

It is a social tactic that happens to push one of my buttons.  I absolutely hate it.

(No, I'm not drunk/high/angry, just very sensitive to these kinds of social-based attacks when someones opinion is not gaining traction.  Others do not care, but I do: I've had enough of them.)

[bdunham7]

Perhaps?  The OP is using a newly created account to call others nerds just because their own opinion is not gaining traction.

I don't know whether the OP is a sockpuppet or a longtime lurker that finally wanted to say something.  I'll presume the latter unless the mods bounce him.

Ordinarily I might agree with you but the thread in question is so vile and polluted I don't think it is even possible to have a sane conversation within it.  I tried to point out a few things that I thought would lead to some agreement, but I regret even wasting the time.

[Nominal Animal]

Ordinarily I might agree with you but the thread in question is so vile and polluted I don't think it is even possible to have a sane conversation within it.  I tried to point out a few things that I thought would lead to some agreement, but I regret even wasting the time.

I avoided that thread for the same reason.  However, starting a new thread with an insult is not a valid way to start a conversation, is it?

(I understand that to some, it is not an insult.  But to me, the entire pattern [of starting a new thread with an insult, claiming that no-discussion is the "professional approach", with their own opinion as "the obviously correct answer, case closed"] is a very sore button.)

I really, really hate such detestable attempts at social manipulation.

To me, the entire discussion is a bit funky, because as nctnico said,

IMHO the whole point is to understand when a simplification works and when not.

and this observation should also be extended to the model used to describe a situation.

One thing I absolutely love about physics simulations is that no matter what you do, the first step when you get some results is to analyze whether it makes any sense.  A part of that is to guesstimate the various factors (say, to within a few orders of magnitude), whether the model includes everything that should be included, and so on.  Only a small part of that is estimating the approximations used; it is the appropriateness of the model used that is the key.

I do find it interesting that the skill of shifting ones mind across model complexity levels is relatively rare.  In programming, I've seen flamewars between top-down (starting with an overall plan, and finishing with actual code) and bottom-up (starting with actual code, then connecting them together, to build a larger more complex whole) approaches, that completely baffles me: I proceed with the hardest problems first, until I can build a reliable model of the end result, happily skipping between complexity levels as needed.

I can see exactly why the same would happen with those who are used to treating everything as a field, and those who work with circuits.

Perhaps it is too hard for most humans to encompass both at the same time –– it is to me, that's why I have had to learn to skip and switch as needed ––, but assuming ones favourite model suffices everywhere is ... well, insufficient/wrong/silly.  In physics simulations, one would trip on it immediately, and fail.

Molecular dynamics is an excellent example.  If we take only the outermost interacting electrons in atoms, and model the rest of the atoms (both nucleus and the rest of the electrons) as a single point charge, and only consider the rest states of each atom, we can model all chemical bonds to a very high degree.  (The charges themselves are modeled as quantum mechanical waves, see e.g. Hartree-Fock method; this is why these simulations are called "quantum mechanical" or "ab initio", starting from the simplest possible interaction model.)

However, the math is so onerous, that even the largest supercomputers have issues with more than some thousands of electrons.  Simplify the interaction model, for example via Embedded Atom Model for metals (albeit you need multi-band EAM for some metals like ferrochrome), and you can model millions to billions of atoms, and get essentially the same results.

Which one is correct?  Well, neither, because both are approximations.  For some systems, both are precise enough to yield useful information, and are used every day in materials research (even in now-mundane things like thin film tech, ion implantation, and so on).
(And yes, there are lots of QM/ab initio simulators using VASP or Dalton, and scoffing on those who use classical potential models or force fields (naming varies between physics, chemistry, and biology, even though they all do more or less similar simulations).  It, too, is horribly silly.  And very often leads to someone, usually an established professor, making an argument from authority, which is even more disgraceful.)

[Simon]

In reality, the fact is there is no large dispute.

Once I figured that out I tried--rather clumsily I suppose--to point out what I saw as the primary issue leading to the disagreement and later the misdirection of attention as to which 'path' was important--it took me a bit longer than I'd like to admit to catch onto this--that sort of made it look like a much more noteworthy issue than it really is.  But the main contestants weren't interested in any of that, just proving that they were superior and engineers are stupid.  And then along came some additional contestants to confirm that. 

The OP in this thread has a reasonable, short explanation for those that don't want to read 39 pages of name calling and drivel.  However, I would replace

So, any time we have a changing field or current (e.g. AC at any frequency), we can no longer define voltage.

with "there are multiple possible definitions of voltage" or something like that.  Proving that we can't define voltage by measuring something with voltmeters seems a bit silly.

The problem is that thanks to certain not as prolific as the masses think and misguided public figures ventures, engineering is now "cool" and the bar for interest has dropped down very low and suddenly every idiot that usually would be more interested in posing down the local pub to attract girls thinks he is an expert because he saw a video made by someone as dumb as them that explained it for them.

[Siwastaja]

I don't think it's a sock puppet account, the writing style does not seem to match with anyone. OTOH, I'm not an expert in analysing human communication, so could be wrong. It could be a sockpuppet with careful change in writing style. Maybe an actual expert can chime in, if we have one.

The fact that the poster has 1 post is not surprising. Because the matter is discussed outside of EEVBlog forum, it brings people in when they notice it's being discussed here. Has happened before.

In any case, it's up to Nominal Animal to prove his claim that this is a sockpuppet account. Do you actually have some information we others don't? Or are you just participating in these name calling social games you so much hate? Calling one a sockpuppet is no different from any other name calling that's going on. And don't get me wrong, I totally see the point, I hate social games too, but you must admit they sometimes just suck you in - and it doesn't make you feel good about yourself.

Regarding social manipulation, this is what I hate above all, and having had about three years to think about this, I'm becoming more and more certain that social manipulation is exactly what Mehdi originally did, and did it really really well. This is difficult to say for me because I have been quite entertained with Mehdi's videos, admired his down to earth style of teaching basic electronics and related physics, and always thought he's a very decent human being, and the last person to play such social games.

On the other hand, worst social manipulators tend to be among those who appear very positive, appear to hear everybody out and have a balanced discussion, appear easily approachable. This isn't a surprise, because if you are nasty and call others with names, who's going to believe you and fall into your social trap? No one. Successful social manipulation needs to be hidden. What we see in the current thread of 40+ pages, isn't social manipulation; it's just gang membership, namecalling, generally behaving like little kids, with some physics education thrown in. However, we can assume that social manipulation is what started it, not a coincidence. This is a fair assumption because the Lewin's experiment has been widely known for decades; nothing changed. What suddenly made it controversial? So who manipulated people into this? There are not too many options. Maybe Lewin, by posting thousands of hours of physics teaching videos from 1990's - early 2000's, hiding a manipulation attempt in one, to have it triggered two decades later? Is this realistic? So I think, if successful, actual social manipulation is part of this, there are no other viable perpetrators than Mehdi. He has to be at the root, and if not, no one is, it is then "just happening".

I also think another massive contributing factor was Lewin's initial response to Mehdi. This reponse was, IIRC, quite nasty and totally unhelpful, and something a good teacher should never let slip. It sure helped create the impression that "Mehdi has to be right", and strengthen the Mehdi followers, initially me included (until I finally got over it, and started thinking about the experiments and physics itself. Again, the eye opening moment was when I admired the amount of effort put in to the exact layout of the experiments (think Mabilde), and then bang, I realized that circuit layout is different from "good probing"; and "good probing" alone didn't provide the expected results. And that layout wasn't described in the equivalent circuit diagrams).

In other words, Mehdi's behavior, at least on surface, was better than Lewin's. Which is also exactly why, if I have to place bets on who's the social manipulator, I will vote Mehdi. And now, this is getting a bit far-fetched, but if Mehdi's a real social manipulator, then he has also known beforehand Lewin's personality (which is all over his videos), and knew that: "buahaha, I will win if I keep my head cool, make my opponent angry by pulling the right knobs, so that he slips out insults; then I only have to ask a few professors to assess my claims, and get a report I can quote out of context."

This is still not to say I'm 100% sure Mehdi is a social manipulator. I'm still 50-60% about this, and I will never be sure about this. Maybe things just went the way they did, with no bad intentions. But think about it!

[HuronKing]

In other words, Mehdi's behavior, at least on surface, was better than Lewin's. Which is also exactly why, if I have to place bets on who's the social manipulator, I will vote Mehdi. And now, this is getting a bit far-fetched, but if Mehdi's a real social manipulator, then he has also known beforehand Lewin's personality (which is all over his videos), and knew that: "buahaha, I will win if I keep my head cool, make my opponent angry by pulling the right knobs, so that he slips out insults; then I only have to ask a few professors to assess my claims, and get a report I can quote out of context."

This is still not to say I'm 100% sure Mehdi is a social manipulator. I'm still 50-60% about this, and I will never be sure about this. Maybe things just went the way they did, with no bad intentions. But think about it!

What's irritating is that Mehdi didn't even wait for Dr. Belcher, whom Lewin suggested he go talk to, to answer his questions before releasing his initial video accusing Lewin of 'bad probing.' I could see why Lewin would be dismissive of Mehdi at face value, and then become outright contemptuous of Mehdi claiming an MSEE while refusing to see something that a 2nd year physics student learns (that line integrals are path dependent in non-conservative fields, the end). Did Mehdi even watch any of Lewin's other lectures in the course? Don't know!
Then add onto it Mehdi's legions of literally millions of subscribers bombarding Lewin with hate-mail... and taking a victory lap and dragging Feynman along for the ride... it's really too much.

And this is more than just a stupid internet drama to me. I am an EE educator. I teach power and motor lab courses and I ALWAYS have a handful of students who have seen Mehdi's videos on KVL and need to have the pseudoscience shaken out of their heads and so I include a vector calculus review and thorough discussion  of Romer's paper as my first lecture.
 

[Simon]

Unfortunately the electroboom channel speaks to the lower levels with it's constant epileptic fit provoking gags. This is why it's not a good idea to get more people into the field by appealing to their stupidity. You need to make sure the capable are given the chance no matter their socio economic standing. Then you will collect more. The people that electroboom's videos appeal to should maybe abstain from procreating!

[tszaboo]

Guess what, KVL also doesn't hold, when a wire is hot, and thermionic emission happens.
It doesn't mean it is useless. And who cares. Let grey haired physics professor rant on their own lectures. When I had a prof like this at university, I just stood up and took the class with another one, that was an engineer.

[Nominal Animal]

In any case, it's up to Nominal Animal to prove his claim that this is a sockpuppet account. Do you actually have some information we others don't? Or are you just participating in these name calling social games you so much hate? Calling one a sockpuppet is no different from any other name calling that's going on.

Dammit, you're absolutely right.

I guess certain words in the title ("ridiculous") and initial post ("nerd") just hit me in a sore spot.  I find the original topic interesting (because there is no simple answer, since both approaches involve modeling the situation in a specific way, and it just isn't possible to say that one is superior to the other –– they are simply appropriate in different contexts), and maybe the combination of perceived insult and social manipulation tactics pushed me over the edge... but I do acknowledge that I have not contributed positively to the topic, and this is purely my own mistake and error, and I do apologize.

(I do see the pot-kettle-black on my part here, too: I essentially made the same error as the OP of this thread.  I have no proof OP had any kind of manipulative intent.  Apologies for that, too. )

[HuronKing]

Unfortunately the electroboom channel speaks to the lower levels with it's constant epileptic fit provoking gags. This is why it's not a good idea to get more people into the field by appealing to their stupidity. You need to make sure the capable are given the chance no matter their socio economic standing. Then you will collect more. The people that electroboom's videos appeal to should maybe abstain from procreating!

Indeed it does, but his narrative is seductive, even to budding engineers.

"Lewin is an egg-headed physicist who doesn't do anything practical unlike us manly-men practically practical engineers!"

Meanwhile, I'm there thinking,
"Uhh... none of us could even take an Applied EM or a Motors course without first passing the MIT 8.02 equivalent..."

And don't get me wrong. I used to wear my Full-Bridge Rectifier T-shirt because his shtick was funny and he wrapped up some nice rudimentary lessons in slapstick comedy.

But nowadays I use his channel as a case study for my students about engineers becoming susceptible to pseudoscientific notions.

I didn't even understand exactly what Lewin meant when he called Mehdi a Flat-Earther, but now I do.
For most navigational purposes, we consider the Earth to be flat. It's a good approximation. No one takes the curvature of the planet into consideration when driving to work. 99% of the time - we are Flat-Earthers in practice, even if we pay lip-service to believing in the sphericity of the Earth, none of us need to know that in our daily lives.

But that 1% of the time, when we step onto an airplane to fly across the planet, NOW the curvature of the Earth matters. So, we need spherical Earth science to navigate properly.

KVL is our Flat-Earth approximation. Faraday's Law is the spherical Earth.
KVL is just a special case of Faraday's Law and only holds in specific conditions and assumptions. Just like our spherical Earth can be considered to be locally flat under the right conditions.

But when Mehdi claims KVL ALWAYS HOLDS... he's saying the Earth is always flat, all the time. And so now I share Lewin's outrage at Mehdi telling millions of viewers that the Earth is always flat... err... that KVL always holds.

[Simon]

If I remember rightly it was lewen that tried to show the theory does not hold up by going beyond it's scope. As much as I respect the man as an inspiring lecturer in this case he delved into something outside of his sphere and did not make clear why it was happening.

[Siwastaja]

A really good teacher needs to keep their head cool in case of objection, and continue the quest of making people learn; spend all the time necessary to help student find the source of their error. Lewin colossally failed in that, possibly because of personality issues, possibly because he did not consider Mehdi as his students. But there is a third possibility, that he understood from day one that Mehdi's being intellectually dishonest with no intention of learning, and "skip" the whole game as unfruitful waste of time. Such is risky behavior, but so is sexually harassing one's students. Lewin's definitely is not a model character. When you say it like this directly, everybody will agree that it's irrelevant when it comes to the physical discussion. Yet, it's very usable tool to create suitable impressions. Lewin's not trustable. Mehdi's a very nice guy, who just makes a rational argument...

Guess what, KVL also doesn't hold, when a wire is hot, and thermionic emission happens.
It doesn't mean it is useless. And who cares. Let grey haired physics professor rant on their own lectures. When I had a prof like this at university, I just stood up and took the class with another one, that was an engineer.

This comment is fascinating, I think it kind of proves my point exactly.

Imagine Mehdi's video never existing in a first place.

I'm rephrasing tszaboo's comment in my own words, and I hope I don't significantly change the meaning, as if this comment appeared after attending Lewin's lecture with the demonstration shown:
"Oh, Lewin demonstrated a case where KVL does not hold, I understood it, and it's correct. There are other cases where KVL fails, too, like thermionic emission. So what? No one ever called KVL useless, but I still want to make it super clear, that it is not useless. I don't like the character of Lewin and due to that, will prefer other lecturers."

See how ridiculous this is in this changed Mehdi-free context? Lectures are full of demonstrations. I have never heard anyone say "so what, there are others as well".

But we see many comments like tszaboo's, and why is that? Because a straw man exists, a straw man of "KVL being useless", which is just ridiculous, no one ever claimed that. This strawman also includes the idea that Lewin somehow claimed that KVL is unusable for most everyday engineering tasks, or something like this.

This strawman is especially appealing to those who never attended university, and as a result, not understand basic university-level mathematical notations and more importantly, admit the importance of basic terminology. Which is fine, but Lewin's lecture videos are in university context. Mehdi's reply isn't.

But this strawman clearly is also appealing to some who did attend university, like tszaboo, or myself.

But where did this strawman originate? I'll give tszaboo benefit of doubt, his intentions are at least neutral.

Hence the "Mehdi's a social manipulator theory". Believe it or not, it's not important, but important is to think what's happening.

[instrumental]

Hey folks!

Any sensible comment gets lost in the noise in a matter of a day, which is probably exactly the reason for such smoke screen.

What will prevent this thread also suffering the same fate once it's been discovered?

If someone is disrupting a discussion through foul means then surely the answer is to ask a mod to deal with them. I don't see that giving them more threads to disrupt achieves much. Of course, perhaps the mods will be too busy, fail to see your point or otherwise let things run, but you don't know unless you ask.

But, on the whole, I am not averse per se with skimming the crust off into another smaller and newer thread. It was just the "Listen, my explanation is the right one, so I'm a special case" which poked me.

Firstly -- no, I'm not a sock-puppet, I'm a longtime lurker who happened on this thread yesterday while skiving off documentation. I'd already been somewhat aggravated at the YouTube scene over this dispute, and made an account to post about this (and also to complain about the SCPI interface on the RS-KEL103, but that's another matter). Nice to meet you all :-).

Secondly -- yes, it's a bit presumptive of me to come and post an "authoritative" explanation instead of posting in the other thread, where it would be more on-topic. Firstly, I was hoping more for a meta-discussion about the argument, and wanted to bracket off the problem itself. Secondly, I'd like to understand what exactly about this problem is so tricky to understand; if my explanation of the problem is unsatisfactory, perhaps you have a more intuitive explanation.

We've gone this way now, and I'm happy to be having this discussion. Please excuse me! I lost the plot a bit, writing the screed, then (in a slightly hungover state) forgetting to return to the raison d'etre of the thread in the first place -- asking why this disagreement is occurring, and what can be done to put an end to it.

If I remember rightly it was lewen that tried to show the theory does not hold up by going beyond it's scope. As much as I respect the man as an inspiring lecturer in this case he delved into something outside of his sphere and did not make clear why it was happening.

Lewin's demonstration is not the clearest. I'm still scratching my head over it! Most of his demos are excellent; this is trying to demonstrate a counterintuitive phenomenon with counterintuitive, roundabout testing. I think this also leads to so-called"experimentalists" changing the free parameters of the demo and disagreeing about the interpretations of the results -- a sign of a bad experiment!

Perhaps to sketch an alternative, I'd opt to attach the ground clip to the tip of an oscilloscope probe and demonstrate that I can still sense "voltages" when time-varying magnetic flux is running through the loop -- it's a handy trick to know for qualitative EMC, and simple enough that the odd interpretations around Lewin's demo ("what if we move the resistors around the loop") are evaded. No free parameters this way, hard to get it wrong :-).

A really good teacher needs to keep their head cool in case of objection, and continue the quest of making people learn; spend all the time necessary to help student find the source of their error. Lewin colossally failed in that, possibly because of personality issues, possibly because he did not consider Mehdi as his students. But there is a third possibility, that he understood from day one that Mehdi's being intellectually dishonest with no intention of learning, and "skip" the whole game as unfruitful waste of time.

This is a great point. From Lewin's perspective, this is a settled matter; EM theory has hung around for over a century, and he'd been working in the field for over half a century at the time. He runs into the issue of communication with a student who doesn't quite get it. It's made more difficult by the fact that Mehdi himself is positioned as an educator, who frames "disagreeing with a master" as regular in the ordinary course of science.

So, any time we have a changing field or current (e.g. AC at any frequency), we can no longer define voltage.

with "there are multiple possible definitions of voltage" or something like that.  Proving that we can't define voltage by measuring something with voltmeters seems a bit silly.

I appreciate this! I meant this more in the context of the definition of the voltage as the potential function of the electric field (in which case the maths breaks) -- this should definitely be clarified. Thanks!

Many thanks for the discussion, all!

[RoGeorge]

Might be Prof. Lewin's account.   

Can't tell about others, but I'm very well and happy about my nerdyness.  To me, being called a nerd is a compliment, thank you! 

About Prof. Lewin, I think he is correct, no doubt for me now.  Makes sense in every aspect I am aware of, thought I'm only an amateur physicist.

At first, my "mind conditioning" unintended induced by years of tinkering with electronic circuits made me think Mr. Lewin was joking, especially since one of his papers about the subject was dated 1st of April. 

But then, I've put aside the EE approach and tried to look only in terms of physics laws and physics and definitions (Voltage is the work needed to drag a charge from A to B), and it was clear like day and night that the professor is correct.  Also, what he was pointing out and his mind blowing demonstration doesn't brake Electrical Engineering in any way.

EE deals with the same results by "blaming" the induction in the voltmeter's cables.

Overall, I'm grateful he pointed out to non-conservative fields with a click bait title and a mind blowing experiment.  Otherwise I would have missed that aspect of non-conservative fields entirely.  Though, the biggest lesson to me was how strong brainwashing can be (here a case of unintended brainwash).

The strongest cage to break is the mind conditioning that slowly grows to imprison us.

[Simon]

My memory is that Lewin demonstrated that the law did not hold, but gave no further explanation as though he had proved someone wrong. As a lesson that is a failure.

[PlainName]

My memory is that Lewin demonstrated that the law did not hold, but gave no further explanation as though he had proved someone wrong. As a lesson that is a failure.

Disclaimer: I know nothing about all this (in fact, this thread is the first time I've tripped over it). If you're teaching someone, surely you don't show them the answers to every problem straight off? You pose a problem and they figure out how to solve it. Teaching is simply giving them the means to find a solution, not giving them the answers. So posing an issue (something doesn't work as expected) it seems to me to be fine to then let them figure out wtf is going on. Later, either no-one figures it and you have to explain, or it is figured and you smooth over the rough edges.

[Berni]

This stuff is still going?

The reason why people can't come to an agreement with each other is that both sides are correct from a given point of view. The physicists and electronics engineers look at it in different ways, that both work within the given context. One of the explanations deals with electric/magnetic fields in 3D space, the other deals with voltages and currents inside a circuit diagram. Both come to the same result.

Just stick to the explanation that works best for you stop trying to convince others that your way is the 'only and only correct way'. Both explanations are idealized cases that ignore certain other natural effects(and rightly so, since why include stuff that have negligible effect on the result).

It's the same as arguing about light being treated as particles as being wrong. Mr. Maxwell there clearly shows how light is also just a electromagnetic wave. Yet you wouldn't use waves to calculate at what angle a laser beam will bounce off a mirror. Yet then again Maxwells equations don't describe the packet nature of light, so are they also wrong? No point in being overly pedantic for no reason. As long as an equation works within the context it was made, just make sure to use it within the correct context.

[PlainName]

No point in being overly pedantic for no reason

But... it's the internet.

[Berni]

No point in being overly pedantic for no reason

But... it's the internet.

Indeed...

[instrumental]

This stuff is still going?

The reason why people can't come to an agreement with each other is that both sides are correct from a given point of view. The physicists and electronics engineers look at it in different ways, that both work within the given context. One of the explanations deals with electric/magnetic fields in 3D space, the other deals with voltages and currents inside a circuit diagram. Both come to the same result.

Just stick to the explanation that works best for you stop trying to convince others that your way is the 'only and only correct way'. Both explanations are idealized cases that ignore certain other natural effects(and rightly so, since why include stuff that have negligible effect on the result).

It's the same as arguing about light being treated as particles as being wrong. Mr. Maxwell there clearly shows how light is also just a electromagnetic wave. Yet you wouldn't use waves to calculate at what angle a laser beam will bounce off a mirror. Yet then again Maxwells equations don't describe the packet nature of light, so are they also wrong? No point in being overly pedantic for no reason. As long as an equation works within the context it was made, just make sure to use it within the correct context.

I'm not sure I agree with this. Both sides are correct up to a point; then, one answer strongly dominates the other. It's a matter of knowing when certain approximations can be made and when those approximations no longer hold.

Newtonian gravity was sufficient to calculate orbital trajectories to land us on the moon. No need to invoke general relativity to tackle that challenge. But we would consider GR a better theory than Newtonian gravity, no? It's a matter of understanding when and where the framework breaks down; which one is more correct and which one provides a simpler, more user-friendly method to calculate.

It's the same here; Maxwell's result demonstrated that light is a wave. Then, the discovery of the photoelectric effect yielded issues and inconsistencies which manifested in the development of an entirely new framework (quantum mechanics) which successfully yields an explanation and tools to calculate -- and gives us wave-particle duality. However! -- I work on infrared instruments, and nobody uses quantum optics here; regular old Zemax is good enough, thank you very much. Why bring in something so complicated when regular ray/wave optics will do?

KVL is a step back; an even further simplification which assumes a conservative electric field. This assumption breaks in some cases (as mentioned earlier), but is a useful approximation for most circuits we see in our day-to-day lives. KVL is a useful tool if it provides useful results; we should understand where it breaks down so we aren't caught pissing into the wind on a £100k deliverable.

And that's the nutshell of my point -- sure, both are "valid," but only up to a point, and we need to have a good understanding of where that point is.

[Simon]

Both sides are correct up to a point; then, one answer strongly dominates the other. It's a matter of knowing when certain approximations can be made and when those approximations no longer hold.

Bingo, so why are we still discussing it? in one sentence you have explained the actualy solution, both are right, both may be wrong, it's a matter of point of (technical and context) view.

[instrumental]

Both sides are correct up to a point; then, one answer strongly dominates the other. It's a matter of knowing when certain approximations can be made and when those approximations no longer hold.

Bingo, so why are we still discussing it? in one sentence you have explained the actualy solution, both are right, both may be wrong, it's a matter of point of (technical and context) view.

Well, if we agree about the solution after a few sentences, why are YouTube channels with millions of subscribers feuding about it? Partly it's just a kvetch about the way this topic is discussed. Partly it's speculating about how best to explain the problem.

But this problem is important. My third-year instrumentation lecturer always liked to remind us that it was fields all the way down, and even at DC. He was a physicist at heart, and it showed; he was a wizard to us engineers. (Perhaps I'm biased this way -- spent a few years building AMR magnetometers, and miss the field, if you pardon the pun.)

Carrying that mentality forward, I've seen things go wrong in industry because people ignore this -- if you forget about fields, you forget that you shouldn't route over splits in ground planes (or, really, split ground planes without rationale -- but that's a whole different religious argument). I worked on a receiver board for a LiDAR on an upcoming lunar mission and the receiver is borked for this reason (SPI bus routed over a GND split near the signal path) -- needs substantial, costly, time-consuming redesign and respin, complicated by the fact that the designer doesn't understand the issue and won't concede that it's a fields problem (despite every fingerprint being there, from qualitative EMC to being able to measure the SCLK signal rising edge on the TIA output). Forget about fields and you'll never be able to solve many of the noise issues you're dealing with -- and don't hold any hope for understanding why your boards have failed EMC.

So, I put it to you -- are we educating new engineers right? Are we motivating the issues at hand well, and are we giving concise and intuitive explanations? If we're triggering holy wars then the explanations at hand are not good enough, and we need to do better. As to what those might look like? Well, I've no idea what's intuitive to newcomers of the field; I've been at this too long. What helps? How do we avoid more of the kind of 39-page inane fighting on this forum, and between YouTube electronics education superstars?

[Berni]

I'm not sure I agree with this. Both sides are correct up to a point; then, one answer strongly dominates the other. It's a matter of knowing when certain approximations can be made and when those approximations no longer hold.

What answer strongly dominates is again dependent on the context.

Story 1:
Someone is designing some sort of magnetic apparatus that involves pulsing a large electromagnet with some wires running around it, so perhaps the question is what is going to happen to those wires. In this case you model up your apparatus and run it all trough an EM field solver, find what sort of fields are going to be in there, what they do to that wire you care about....etc. Define your path along what you want to measure the voltage. We got the result we need and there we go. Maxwell all the way here. No use what so ever for any of Kirchhoffs stuff.

Story 2:
Someone is upgrading some sort of existing magnetic apparatus with a large pulsing electromagnet and has issues with wires that for some reason have to run near it. They don't have the CAD models of the thing since the thing was built by a company that shut down that branch of business 10 years ago, so doing an EM simulation would involve a good bit of caliper work to reverse engineer it and the costumer needs this thing working by the end of the week. So perhaps instead haul out that trusty old boatanchor network analyzer, hook it up to the electromagnet and the wire(keeping in mind that probing will be included in the measurement), measure the S parameters of it across the frequency range of interest. Fire up there favorite spice flavor and plonk it in as a transformer with those parameters. Now they can experiment with what circuitry they need to add around it to fix the problem they are having with the electromagnet interference. No Maxwell involved in any of this, it is all Kirchhoff paired with some AC circuit theory and numerical integration methods.

No point in playing a fan boy of one or the other.
All that matters is that you use the right tool for the job and at the same time understand the limitations of the tool.

People often form opinions on there own needs and preferences, but just because something doesn't fit your needs or liking does not make it wrong, especially when that thing is actually the right tool to use for someone else, giving them excellent results. For this reason i find both explanations equally correct. But feel free to continue waging the Lewin vs KVL war on the forums, looks like there will always be someone up for discussion to try and convince towards one or the other side.

EDIT: Btw i think Lewins KVL failiure demonstration is a great one. Really provokes some deep thought into what voltage actually is. This is exactly what the students need. Its just that the explanation of why KVL breaks here is a bit lackluster(He claims that KVL is simply wrong and does not work).

[bdunham7]

Both sides are correct up to a point; then, one answer strongly dominates the other. It's a matter of knowing when certain approximations can be made and when those approximations no longer hold.

I don't think that's true in the Lewin/KVL 'debate'.  It's not like classical physics vs modern physics or anything like that.  There's a stark disagreement over whether the voltage between two points in a given apparatus is a uniquely defined single number or can be different numbers depending on 'the path'.  The debate devolves from there.

[SiliconWizard]

And, Lewin's demonstration itself - oh I know, let's not get into that all over again - was flawed IMHO. Let's put another coin in the machine.
I think this guy makes a few good points, and exposes them in a calm and structured way:

[bdunham7]

And, Lewin's demonstration itself - oh I know, let's not get into that all over again - was flawed IMHO. Let's put another coin in the machine.

The demonstration was great.  His explanation contained clever elements of misdirection.  I think everyone willing to consider reasonable explanations has already figured it all out and left the building--at least in that thread.  As for what remains, I've seen contested divorces that were less contentious.  The whole thread is now even sillier than the can't-go-faster-than-tailwind guy, and that's saying a lot.

[Siwastaja]

My memory is that Lewin demonstrated that the law did not hold, but gave no further explanation as though he had proved someone wrong. As a lesson that is a failure.

If you actually watch it, there's nothing special about it: typical demonstration, with explanation, but it's math heavy; you need to have solid understanding in the math.

It's also series of lectures, not hit-and-run. Students are also assumed to follow a textbook.

Remove small part of it from context, without understanding surface and line integrals and curls and whatnot, and the result is what it is.

[snarkysparky]

Both sides are correct up to a point; then, one answer strongly dominates the other. It's a matter of knowing when certain approximations can be made and when those approximations no longer hold.

I don't think that's true in the Lewin/KVL 'debate'.  It's not like classical physics vs modern physics or anything like that.  There's a stark disagreement over whether the voltage between two points in a given apparatus is a uniquely defined single number or can be different numbers depending on 'the path'.  The debate devolves from there.

Faradays integral law says the line integral of electric field around any closed path is equal the the time rate of change of flux enclosed by the path.
What the path is is not specified. 

So I vote for electrical potential in non time changing fields to be path independant.

[bdunham7]

Faradays integral law says the line integral of electric field around any closed path is equal the the time rate of change of flux enclosed by the path.
What the path is is not specified. 

Yes, but Faraday's Law does not itself directly represent a physical phenomenon.  It is simply a provable result of integral calculus that all paths that completely enclose a specific amount of changing flux will have a particular amount of total EMF around that path.

So I vote for electrical potential in non time changing fields to be path independant.

I'm not sure I follow the reasoning. 

[snarkysparky]

reasoning is that IF faradays integral law is correct the result is path independent.   Old man Faraday didn't say what path is to be taken.

EQ 4 

https://hep.uchicago.edu/cdf/frisch/p142/Purcell_Chapter2.pdf

[Naej]

In case someone wants to learn the difference between voltage and potential difference, and why:
https://physics.princeton.edu//~mcdonald/examples/lewin.pdf
https://physics.princeton.edu//~mcdonald/examples/volt.pdf
https://physics.princeton.edu//~mcdonald/examples/voltage.pdf

Have fun.

[emece67]

.

[RoGeorge]

That integral form is not the definition of voltage.  That's a formula, an equality where V = V/m * m, so V = V after m simplifies away.  It defines V in terms of V, not nice.


I'll take as correct the definition it was given in the primary school, and was used since the early notions of electricity, in terms of energy (work) and charge.  If we imagine we have to drag (as in move, or transport) a charge between to points in space, from A to B, then we define:

Voltage is the work required to move an unit of electric charge between two points A and B.

\[V_{AB} = W/q\]
Where:
- \$W\$ is the mechanical work, the energy involved in moving the unit of charge, energy measured in Joules
- \$q\$ is the electric charge, measured in Coulombs.

[bdunham7]

As long as some say that the «voltage» between points A and B is ΦB−ΦA (being Φ the electric scalar potential) and some others say that it is −∫BAE⃗ ⋅dl⃗ , there cannot be any agreement

A very succinct statement of what I tried to say much earlier.  Unfortunately, the 'KVLers' apparently didn't understand it either.  However, a few issues with your conclusions...

• the second is easier to measure (just ensure there's no varying B field inside the probe wires), but it is path dependent

Actually in these cases there is a varying B-field inside the probe wires, just not between the probe wires and the ring.  What the voltmeter ends up seeing is the EMF around the whole outer loop minus whatever voltage drop there is due to resistance in that outer loop.  The fact that this value is the same as what you would expect across the inner resistor is just math.

Apparently IEC thinks that the "voltage" is that of the 2nd definition, and normal voltmeters, including the 121GW (incidentally, what about a firmware update allowing the user to select the desired voltage definition to use?), do measure according to such definition, so I will continue (it was what they taught me when young) using such definition of voltage. Thus, I will see systems where 2 voltmeters connected to the same points throw different measures, shit happens.

Actually, in any reasonable setup where the body of the meter is not placed in ridiculous places, a voltmeter like the 121GW will display a value very close to the ΦB−ΦA 'scalar potential' or 'KVL volts' (my tonge-in-cheek term for it from the other discussion) that is present directly across it's input jacks.  If you do place the meter in some place with a significant curled E-field going through and around the meter, you will likely simply get unpredictably erroneous readings due to the complex internal circuitry of the meter.  The 'path dependence' is all about the test leads, which are just wires that form another circuit.  This might seem like nitpicking, but I think it is not.  Special firmware will obviously not allow you to magically read KVL volts, but it might be possible to come up with a probe set that does.[/list]

[bdunham7]

Voltage is the work required to move an unit of electric charge between two points A and B.

That is one definition and for many things, likely the most sensible.  However, it is not the only reasonable or possible definition.

[Psi]

40 pages and a few months on, some nerds don't seem to know when to quit arguing. The video Mehdi made about a "principled disagreement" with Lewin's lectures is rough as well, and now this has spiralled into numerous YouTube videos trying and failing to explain EM theory. I'm surprised to see so much discord in a community of professional electronics engineers.

I'm quite tired of seeing this, so I am hoping to provide a fairly definitive explanation of:
.....

[emece67]

.

[emece67]

.

[thinkfat]

Note: laying the probes in a way that they are always orthogonal to the induced E field is an easy way (supposed you know how the hell is the induced E field!) to measure \(\Phi_B-\Phi_A\), but not the only choice. What you need to ensure is that the line integral of the induced E field along the probes is 0.

Wouldn't the latter be true also for any closed path that doesn't contain any time-varying magnetic flux?

[RoGeorge]

Voltage is the work required to move an unit of electric charge between two points A and B.

That is one definition and for many things, likely the most sensible.  However, it is not the only reasonable or possible definition.

1. No and 2. Yes

1.  If that is a definition for many other things, then give some examples.  A definition must uniquely identify the thing we want to define.  Maybe you spotted in that definition a mistake nobody seen before.  Highly unlikely but possible, so give an example to understand what you meant there.



2.  Physics laws are all linked together, like a big network, or a mesh of relations and interactions, a rats nest, a spaghetti code, call it how you like, e big entanglement of relations.  At some point somebody observers something new, and define that new discovery as being such and so, in terms of some other common knowledge and definitions.

But that new thing discovered is already entangled with other laws and relations, and formulas of how to deduce one from another are clarified and so on.  Usually the first definition stays, sometimes it is polished and changed, but there is no absolute definitions in physics.

Everything is defined in relation with some other things.  We all agree to consider some 7 units as fundamental, and define everything else in terms of those 7 picks:  https://en.wikipedia.org/wiki/SI_base_unit  but that is only a convention, like a random pick if you want to say it like that.  Those 7 units are so mostly because of historical reasons, can be as well other set of units.  Some green Martians can have a completely different set of definitions for the same Universe, and that wouldn't change Physics at all.

Because all physics laws we know so far are linked together, we can pick whatever starting point as a definition.

So, a definition can be relative to any other known thing, but it must let no doubt about the new thing it defines.  Also, it must be in relation with something else.  It is not allowed to define "a volt = a volt + a horse after you removed the horse".

[SandyCox]

The point is that circuit theory is a (very useful) abstraction. The notions of electric and magnetic fields are not part of this paradigm. There is no Faraday’s law or Ampere’s law within the paradigm of circuit theory.
In circuit theory, components are represented by symbols and the “wires” form a graph, describing the way in which the components are connected. Components are represented by their terminal properties, in the form of mathematical equations, i.e.
v = Ri,
i = C dv/dt,
v = L di/dt,
etc.
While these equations are derived from the laws of electromagnetism, the actual physics of how the components work do not form part of the paradigm. Semiconductors are also represented by symbols and their terminal properties. Circuit theory tells us absolutely nothing about the underlying semiconductor physics.
The loops we draw in circuit theory and the actual physical loops are not the same. The physical loops have parasitic resistance, inductance and capacitance. Magnetic fields can couple into the physical loops. We can model these effects, to a certain extent, by creating symbols and equations that capture the underlying physics. For example, the equivalent circuit of a transformer.
Within the paradigm of circuit theory, Kirchhoff’s laws are perfectly valid.
Circuit theory remains an extremely valuable abstraction. Most of us simply would not be able to design circuits if we had to take the full Maxwell’s equations, and semiconductor physics, into account at every step of the design process.
Dr Lewin, with all respect, you are wrong!
Apparently, you are unaware of the fact that circuit theory is an abstraction. You are mixing concepts from two different paradigms.

[thinkfat]

And back to square one...

[snarkysparky]

Aren't we talking about a completely static situation.  No field quantities are changing with time?  That was the original question?

If so isn't the line integral of E * dl    path independent?

[SandyCox]

Not static. Note the di/dt and the dv/dt in the equations.

There is no concept of electric field within the paradigm of circuit analysis. It is only introduced in the next level of understanding. Also no concept of distance or path along which to integrate. Just symbols described by their terminal properties and links (virtual wires) connecting them.

[snarkysparky]

The original problem is a battery powering a light bulb.  So yes the fields have no time varying component after the transient.   

And after the transient the power flows entirely in the wires.   

Veritasium is WRONG in his statement.

[SandyCox]

I fully agree, but this thread is about KVL.

[RoGeorge]

Aren't we talking about a completely static situation.

No, there is nothing static in Lewin's paradox.

He said:  I'm gonna show you how two voltmeters connected in parallel can indicate two different voltages.

And everybody was, "yeah right, maybe if one is defective, you are trolling us, right?".  And then Lewin said something like:  "you know what, even more, I'm gonna make one show +0.9V, and the other show -0.1V, both in parallel, how about that, do you believe me?".  And everybody was:  "chill down professor, you must be drunk, or something".

Then the professor cobbles up a circuit with 2 voltmeters in parallel, and "bang!" he released a full blown EMP (Electro-Magnetic Pulse) in the middle of his circuit.  And one voltmeter goes +0.9V, the other goes -0.1V, just like he predicted, yet with both voltmeters connected in parallel.

And the professor is: "told, you!", and everybody else was: 
Then after a second, "wait a minute, you weren't suppose to do that!" (as in, we all assumed a static situation, or else said a conservative field).

Well, "assumption is the mother of all fuck-ups", I never said it was static, and my point is one of the most glorified of yours EE rules doesn't hold in some very particular situations, like I just shown you with my clever circuit and this EMP, or else said instead EMP, a non-conservative field for the most pedantic, sais the professor.

Well, the professor never actually said all those words, that was just my artistic rendering of how the whole debate started, technically.

-----------------------

From here on, even more human feelings and emotions are thrown into the game, electricians and engineers saying "my rules you point out as limited/sometimes wrong are working just fine for me, and I have my ways to deal with that particular EMP you showed, we use induced voltage in the probing wires instead", and the professor said "but that really is because your rule doesn't hold for non-conservative fields", and so on.

From here, each side gets more and more stubborn into its own interpretation, and bang!, 40 pages of fights on EEVblog only, and countless other debates elsewhere! 

 



See for yourself at minute 50:50 (and those experimental results are not coming out of nowhere, it was all explained how something like that is possible, just watch the lecture(s) preceding the minute 50:50 demo):

8.02x - Lect 16 - Electromagnetic Induction, Faraday's Law, Lenz Law, SUPER DEMO
Lectures by Walter Lewin. They will make you ♥ Physics.

[Simon]

Carrying that mentality forward, I've seen things go wrong in industry because people ignore this -- if you forget about fields, you forget that you shouldn't route over splits in ground planes (or, really, split ground planes without rationale -- but that's a whole different religious argument). I worked on a receiver board for a LiDAR on an upcoming lunar mission and the receiver is borked for this reason (SPI bus routed over a GND split near the signal path) -- needs substantial, costly, time-consuming redesign and respin, complicated by the fact that the designer doesn't understand the issue and won't concede that it's a fields problem (despite every fingerprint being there, from qualitative EMC to being able to measure the SCLK signal rising edge on the TIA output). Forget about fields and you'll never be able to solve many of the noise issues you're dealing with -- and don't hold any hope for understanding why your boards have failed EMC.

So, I put it to you -- are we educating new engineers right? Are we motivating the issues at hand well, and are we giving concise and intuitive explanations? If we're triggering holy wars then the explanations at hand are not good enough, and we need to do better. As to what those might look like? Well, I've no idea what's intuitive to newcomers of the field; I've been at this too long. What helps? How do we avoid more of the kind of 39-page inane fighting on this forum, and between YouTube electronics education superstars?

You don't need a degree in physics to know that you should not route a trace from one ground plane to another, this is because you don't need to know down to the atomic level why it's a bad idea, at a basic level you just need to know not to do it, a slightly better engineer will know the verbal reasoning but won't need to worry about the math. It is the physicists job to tell you the math detail but for the purposes of designing a PCB that's irrelevant, because as a good engineer you know not to do it.

i have no idea about fields, but I put items including a SMPS through military EMC testing successfully, no physics knowledge required, I'm just a good engineer. If you asked to to prove down to the math why my method worked I will tell you to consult a physicist. I can tell you broadly what is going on but I won't be calculating the rf emissions for you. No one sits down and makes calculations to that level when they design for EMC, they learn what works and what does not and hopefully get mentored by anther experienced engineer. Theoretical evaluation of an EMC scenario is far more complicated than designing with well informed instinct based on intuition and experience. That is the difference between engineers and physicists.

So as he is the ultimate physicist why did he do a demo to disprove a theory but not explain that the theory was not meant to hold at this level and then explain what was going on. He just seemed to disprove someone with no explanation himself.

[PlainName]

at a basic level you just need to know not to do it

I think this is what distinguishes, say, a script kiddy from a programmer. Anyone can follow rules, but knowing why you follow them allows you to apply them properly and appropriately (and, sometimes, not to). As they say, "Rules are for the obedience of fools and the guidance of wise men."

[Simon]

at a basic level you just need to know not to do it

I think this is what distinguishes, say, a script kiddy from a programmer. Anyone can follow rules, but knowing why you follow them allows you to apply them properly and appropriately (and, sometimes, not to). As they say, "Rules are for the obedience of fools and the guidance of wise men."

Knowing why is different to doing the mathematical proof about it. If I already understand enough to know that it is not a thing I should do. Do I need to to the theoretical math and prove it ? no I don't, because I know enough to know broadly that the outcome is a fail.

[SiliconWizard]

at a basic level you just need to know not to do it

I think this is what distinguishes, say, a script kiddy from a programmer. Anyone can follow rules, but knowing why you follow them allows you to apply them properly and appropriately (and, sometimes, not to). As they say, "Rules are for the obedience of fools and the guidance of wise men."

Yes. But you'd also be a fool thinking that just because you apply more complex rules, then you're not just following rules, and that you actually understand something. =)

[PlainName]

That's a nice scarecrow (aka strawman) with the ever more complex rules

All I pointed out was that it's better if you know why a rule is there. You don't need to go down rabbit  holes or have more complex rules, just know why that rule is the rule. For instance, you put a diode across a relay coil as a matter of course. It's a rule. But if you didn't know why it's a rule then you don't understand what your circuit is doing. Sure, it will survive better than if you didn't know the rule was there, but you're just tickboxing. However, knowing why you put the diode there doesn't imply anything about quantum mechanics or the 40 lower layers of sub-rules that might be invocable.

In the specific case of the ground planes, why do you not route over gap? If you don't know and just follow the rules, the probability is that when you need to do it (perhaps you have no choice) you won't be able to choose the least worst way of doing it. Perhaps it doesn't actually matter in this design and having separate planes is more important. Who knows? (Clearly, not the person who unknowingly rule-follows.)

[SiliconWizard]

That's a nice scarecrow (aka strawman) with the ever more complex rules
All I pointed out was that it's better if you know why a rule is there. You don't need to go down rabbit  holes or have more complex rules, just know why that rule is the rule. (...)

That was just a pinch of humility, which never hurts, and which some seem to be severely lacking here (not talking about you).

But of course, the more you know... the more you know. In a way. =) (Jokes aside, because beyond some point, excessive "knowledge" can actually end up counterproductive.)

Yes, applying rules without understanding them is never a good thing. You should at least understand the basic principles underlying them - which will help applying them wisely.

You don't necessarily need to resort to Maxwell equations  - or the Schrödinger equation  - to design and analyze electronic circuits. The moment you think you do, either you *really* do, or you are probably going to teach physics rather than design things.

Of course knowing why you use some rules while routing PCBs *is* a necessity. But knowing the principles doesn't mean you need to dig ultra deep. When was the last time you used Maxwell equations to route a PCB?

And precisely, I think a good engineer (and admittedly not all are "good") must be good at applying physics. Engineering is applied science. And believe it or not - I'm pretty sure a few will fiercely disagree - good engineers are often better at applying physics than many physicists (at least, those that are theoretical physicists). Experimental physicists are a different matter. I have a deep respect for them (I have for theoretical physicists too, don't get me wrong, it's just that we are talking about applied science here!) Many theoretical physicists are not good at experimental physics. Which is why it's not unusual that the ones having devised sophisticated theories and the ones that have been able to observe them through experiments are different people/teams.

So anyway, just a few random thoughts. My point is, certainly, if you don't understand the underlying principles, you're going to have a tough time, but conversely, just because you perfectly master the underlying principles to the minute details doesn't mean you'll be good at applying them in all circumstances.

[nctnico]

In short: engineers apply science; turn applicable theory into something that works in the real world.

[Nominal Animal]

And precisely, I think a good engineer (and admittedly not all are "good") must be good at applying physics. Engineering is applied science. And believe it or not - I'm pretty sure a few will fiercely disagree - good engineers are often better at applying physics than many physicists (at least, those that are theoretical physicists). Experimental physicists are a different matter. I have a deep respect for them (I have for theoretical physicists too, don't get me wrong, it's just that we are talking about applied science here!) Many theoretical physicists are not good at experimental physics. Which is why it's not unusual that the ones having devised sophisticated theories and the ones that have been able to observe them through experiments are different people/teams.

I agree.  Similar difference exists between mathematicians and physicists, too.  Things like galling have a rather funky causation chain, and are obvious in hindsight, but very, very difficult to predict from the theory alone.  Yet, just about anyone dealing with metal-to-metal sliding contacts knows about galling, and how lubrication helps avoid it.  (Even then, cold welding (in vacuum) was a bit of a surprise when it was discovered in the 1940s.)

Kirchhoff's circuit laws precede Maxwell's equations for classical electromagnetism.  Neither is exactly correct: they are both limited models.  As far as we currently understand, for a full description of electromagnetism, we need to turn to quantum electrodynamics.

QED is also currently used for the most precise simulations of chemistry (including molecular dynamics and materials physics), by only modeling the outermost interacting electrons for each atom.  Even there, because of its calculation-intensive nature, we are limited to cases with at most some tens of thousands of electrons, plus the volume must repeat in every direction (periodic boundary conditions).

Now, mathematically, you do get from QED to Maxwell's equations at the limit of taking the reduced Planck constant to zero, \$\hbar \to 0\$.  This is deceptively simple, because it really does not just mean that we ignore the quantum nature of the universe –– even though there are a lot of highly respected theoretical physicists that will laugh at that and say that of course it does –– because that difference is what produces some of the unexpected effects: just like galling I mentioned before is unexpected to those considering atomically perfect metallic crystalline surfaces sliding against each other, but obvious and easy to explain in hindsight, when you already know it does happen.

(Do remember, that physicists still cannot exactly agree if and why hot water freezes faster than cold water in the exact same conditions.  This is easy to experimentally verify.  I do believe the reason has been found through simulations, and is essentially that heating water will affect the O-H bond length decreasing the heat capacity of the molecule, but when shedding the heat, the bond length shrinks slower than the other degrees of freedom that comprise the heat of the molecule (various vibration modes).   This is why I don't see anything "strange" in arguing whether KVL or Maxwell is correct in some specific situation, because to me, it is normal argument about which imperfect model is better applied since the "correct" one, QED, is too complicated to apply here.  But remember galling: math alone does not say what effects there are if you simply approximate a single constant a bit.  A lot of things go wonky in geometry if you assume e.g. \$\pi = 3.14\$ i.e. rational, for example.)

Full disclosure:  I don't like QED.  I am not a mathematician, and getting my feeble brain to work with QED and the approximations required to work with it (like the Hartree-Fock method) so overwhelms it that I then have no touch with the physical systems I'm trying to work with.  Lagrangian and Hamiltonian mechanics I can just about grok, and I can work with quantum physics in general (since it has a nice definition of observables that keeps me in touch with the actual physical systems I can measure and, uh, observe).  So, I won't be offering much wrt. KVL-vs-Maxwell.  I'm best suited to making the computational tools for others to work with, really.

[RoGeorge]

e.g. \$\pi = 3.14\$

Can be between 3 and 4, according to mathematicians: 

When Pi is Not 3.14 | Infinite Series | PBS Digital Studios


though Pi is 3.0 for construction engineers and that's nothing but a scratch, because cosmologists often assume Pi is 1.0! 

[PlainName]

Her hands need tying down or the video cropping or something. Very hard to watch with those beating time.

[RoGeorge]

She was a very good host for that channel, that good that after she left the channel in order to have more time to finish her PhD, they tried a few replacements but none were as charismatic and skilled as she was, and after a few more months of trying, PBS decided do end the "PBS Infinite" channel.

"PBS Infinite Series" was a very good channel, always with outstanding content and making complicated concepts easy to follow, brief but showing just enough to make one curious and wanting to learn more.

[RoGeorge]

you do get from QED to Maxwell's equations at the limit of taking the reduced Planck constant to zero, \$\hbar \to 0\$.

He, he, nice!  I didn't know that before, so I've thought "how's so" while peeling potatoes, trying to cherry-pick for an explanation.  Well, \$\hbar \to 0\$ would mean making quantization smaller and smaller, until instead of quantities coming in certain chunks, it all become a "smooth" continuous function, just like we used to have before QED.  Q.E.D. 

But the idea of energy coming in chunks was introduced in order to patch the wrong prediction of the so called ultraviolet catastrophe.  With smooth energy exchange, the calculated spectrum of a black body radiation was very different from the measured spectrum.  And somehow, adding a new rule about energy exchanges happening only in certain chunks of energy, the quantization, fixed that wrong prediction.

Now, I would expect to get some failed prediction in terms of QED, too, when \$\hbar \to 0\$ (as it is in Maxwell), similar with the failed predictions from the spectrum of the black body radiation when the quantization aspect was not considered, but I don't know any examples where Maxwell fails to predict correctly.

What would be such an example?

[bsfeechannel]

You don't need a degree in physics to know that you should not route a trace from one ground plane to another, this is because you don't need to know down to the atomic level why it's a bad idea, at a basic level you just need to know not to do it, a slightly better engineer will know the verbal reasoning but won't need to worry about the math. It is the physicists job to tell you the math detail but for the purposes of designing a PCB that's irrelevant, because as a good engineer you know not to do it.

i have no idea about fields, but I put items including a SMPS through military EMC testing successfully, no physics knowledge required, I'm just a good engineer. If you asked to to prove down to the math why my method worked I will tell you to consult a physicist. I can tell you broadly what is going on but I won't be calculating the rf emissions for you. No one sits down and makes calculations to that level when they design for EMC, they learn what works and what does not and hopefully get mentored by anther experienced engineer. Theoretical evaluation of an EMC scenario is far more complicated than designing with well informed instinct based on intuition and experience. That is the difference between engineers and physicists.

So as he is the ultimate physicist why did he do a demo to disprove a theory but not explain that the theory was not meant to hold at this level and then explain what was going on. He just seemed to disprove someone with no explanation himself.

"All generalizations are dangerous. Even this one." People attribute to Alexandre Dumas. We can't generalize and say that knowledge of physics is dispensable for the "good" engineer. Electronics engineering is a vast field. Some areas can get away with basic knowledge of circuits. Other areas are at the edge of the technological advancement and require really knowledgeable people.

So, the difference between an engineer and a physicist not always can be translated into the knowledge of physics. And in many cases you can have a bit of overlap between the two camps. Just to cite two classical examples, Michael Faraday, a physicist, while discovering the phenomenon of magnetic induction invented the very first transformer, which is now an ubiquitous engineering device. Moritz Jacobi, an engineer (and also a physicist) proved the theorem of maximum power transfer.

Lewin never wanted to disprove Kirchhoff's laws. What he wanted to show is that those laws are not applicable to all circuits and in that case the more general theory is Maxwell's equations. He left the explanation of his demo as a homework to his students. The "ridiculous" KVL threads we have is us responding to the challenge, and discussing why some have flunked it miserably. I have learned an awful lot with it, especially with Sredni's and Huronking' posts and with several others'.

[Nominal Animal]

Now, I would expect to get some failed prediction in terms of QED, too, when \$\hbar \to 0\$ (as it is in Maxwell), similar with the failed predictions from the spectrum of the black body radiation when the quantization aspect was not considered, but I don't know any examples where Maxwell fails to predict correctly.

What would be such an example?

The Wikipedia article lists a few.

The photoelectric effect is probably the easiest case to verify and understand, where Maxwell's equations fail to predict the phenomena.  Maxwell's laws suggest that a low-frequency (long-wavelength) light beam at high intensity would accumulate enough kinetic energy in the outermost electrons of the atoms, exciting them until they're kicked out from the atoms, and thus releasing photoelectrons, but that does not happen.  The photons need to exceed a threshold energy (frequency, wavelength) for it to happen.

A typical silicon photovoltaic cell has a minimum bandgap of about 1.12 eV.  Photons having this energy have wavelength of about 1100 nm, which is near infrared, just outside human visible spectrum.  Maxwell's equations don't predict any kind of threshold effect, and photons with longer wavelengths (and thus lower energies) should work just fine, just provide less energy.  However, in reality, photons with lower energy (higher wavelengths) won't produce a photoelectric effect.

This is easier to see with semiconductors with bandgaps in the visible spectrum, for example gallium phosphide (as used in e.g. old-style green LEDs), which has a bandgap of 2.26 eV, corresponding to photon wavelength of about 550 nm.  Use a red laser (over 600 nm) to exite gallium phosphide, and nothing happens.  Use a blue laser (under 500 nm), and it exhibits the photoelectric effect.

(A laser, such as a cheap laser pointer, is just an easy source of monochromatic light, which is the point here: the photons all have the same wavelength.  In a laser, they also have the same phase, but that is not important here.  This experiment works with non-coherent monochromatic light of suitable wavelengths.)

[Lewin] left the explanation of his demo as a homework to his students.

Funnily enough, I see this as the biggest difference between mathematicians or theoretical physicists, and experimental physicists or others who apply mathematics as a tool.

Mathematicians show the formulae and their derivation, and leave the interpretation and application to the student.
Experimental physicists and those who teach applied mathematics describe a family of problems, their description in mathematical terms, and the applied tools that can be used to find the solutions –– exactly the part that is left to the students by mathematicians.

To some people, the latter is the natural, better approach.  To some people, the former is the natural, better approach.  A lot of people can work things out either way.  Very, very few people can teach both ways effectively.  And this creates a big part of the dichotomy, since very few people can bridge the two effectively.  It also explains why some believe string theory is physics, while others consider it only mathematics thus far.