Sick of ridiculous KVL infighting

[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".