"Veritasium" (YT) - "The Big Misconception About Electricity" ?

[TimFox]

Fluid mechanics is not my field, but I believe that the pressure against the pipe walls is different when there is fluid flow, compared to that with no fluid flow.
Certainly, if the pipe is noticeably elastic (thin rubber), one expects the walls to expand somewhat when there is fluid flow.

[cdev]

All this makes me think about, and silently thank Maxwell and Poynting for, so long ago, asking the questions. How amazing this all is.

[IanB]

Fluid mechanics is not my field, but I believe that the pressure against the pipe walls is different when there is fluid flow, compared to that with no fluid flow.
Certainly, if the pipe is noticeably elastic (thin rubber), one expects the walls to expand somewhat when there is fluid flow.

Yes, but it is slightly more complicated than that. Firstly, when there is no flow, the pressure inside the pipe is a free variable (an initial condition), and it can be set to anything. Secondly, if we start with that initial condition and no flow, and then commence the flow of fluid, the pressure will actually decrease due to conservation of energy (the sum of pressure energy and kinetic energy is conserved). When the fluid starts moving it gains kinetic energy, which causes the pressure to fall to conserve the total energy within any element of fluid.

You can observe this with a rubber garden hose having a spray nozzle. If you shut off the spray, the hose will bulge due to the backed up water pressure. If you open the nozzle the hose will relax. Whether the nozzle is open or closed, the water inside the hose has exactly the same energy--the energy it had when it came out of the faucet--the only difference is whether it is flowing or not.

[dom0]

I'll admit upfront that I have only skimmed the thread.

I'll start with a simplification, which makes the experiment easier to conduct: Just remove one of the "legs". Now we have a pretty simple setup: A twin-lead transmission line, shorted at the end, between a pulse generator and its load. Anyone arguing for the "can only light up after one light-second or more without violating causality" (<- which seems to be a popular stance outside EE forums) position should be OK with this, as there is still a light-second worth of wire between the bulb and the source.

How does that behave? If load impedance ~ transmission line impedance, then we'd expect half the output voltage near-instantly across the load, as the impedance across the transmission line is..well the impedance of the transmission line right up until the wavefront bounces off the short at the end and has returned (2x delay time). Then the impedance looking into the transmission line will jump to ~zero and full output voltage appears across the load.

With a coax this is really nice to demonstrate:



No delay:



Some ripples and other reflections going on, because Zsource=Zload=50 Ohm, but ze cable = 75 Ohm.

With a transmission line things get more wild.

A twin-lead transmission line with 1 meter spacing and a reasonable conductor thickness has an impedance of around 1 kOhm. So clearly you're not going to light up a 12 V car lamp with that.

I constructed a simple ~10 meter long ladder line with ~50 mm spacing (~600 Ohms impedance) and used the same 50 Ohm source and load as before:



Reflections galore. Some voltage across the load instantly as before, but I wouldn't count that as "lighting up".



With a 600 Ohm load, we're only seeing about ~35 % of the output voltage across the load. I'm somewhat sure that this is because my setup is grounded and a ladder line suspended ~70 cm above ground has a fairly significant impedance to ground. With a high-impedance load I'm getting close to 50 % output, suggesting to me that the common-mode / ground impedance is around 500-600 Ohm.

So clearly you need to do this in space.

[eti]

I'll admit upfront that I have only skimmed the thread.

I'll start with a simplification, which makes the experiment easier to conduct: Just remove one of the "legs". Now we have a pretty simple setup: A twin-lead transmission line, shorted at the end, between a pulse generator and its load.

How does that behave? If load impedance ~ transmission line impedance, then we'd expect half the output voltage near-instantly across the load, as the impedance across the transmission line is..well the impedance of the transmission line right up until the wavefront bounces off the short at the end and has returned (2x delay time). Then the impedance looking into the transmission line will jump to ~zero and full output voltage appears across the load.

With a coax this is really nice to demonstrate:



No delay:



Some ripples and other reflections going on, because Zsource=Zload=50 Ohm, but ze cable = 75 Ohm.

With a transmission line things get more wild.

A twin-lead transmission line with 1 meter spacing and a reasonable conductor thickness has an impedance of around 1 kOhm. So clearly you're not going to light up a 12 V car lamp with that.

I constructed a simple ~10 meter long ladder line with ~50 mm spacing (~600 Ohms impedance) and used the same 50 Ohm source and load as before:



Reflections galore. Some voltage across the load instantly as before, but I wouldn't count that as "lighting up".



With a 600 Ohm load, we're only seeing about ~35 % of the output voltage across the load. I'm somewhat sure that this is because my setup is grounded and a ladder line suspended ~70 cm above ground has a fairly significant impedance to ground. With a high-impedance load I'm getting close to 50 % output, suggesting to me that the common-mode / ground impedance is around 500-600 Ohm.

So clearly you need to do this in space.

Nice CLEAR 'scope photos - these exceed the standard I expect (my standards are very high) 

[Per Hansson]

@Per Hansson: Sorry if I misunderstood you, but I'm not sure you completely got the point made in the video.

I'm also not sure I get this part: "power is already flowing in the wires up to the point of the switch"
If the circuit is open, then no power is flowing. Or is there?

That's how I find the "water flow" analogy misleading when applied to electricity. Unlike water behind a closed faucet, electric power is not pushing against the open switch. There is no power flowing until the switch is closed. But, sorry again if I misunderstood what you said.

Oh I most certainly misunderstood it, the question asked by the video is just a silly game with no actual real world use.
The point I tried to make, but worded badly is that I think the reason for the "faster than anticipated" turn-on of the light is due to a silly thing:
Remove the switch from the circuit and instead use the cables themselves to turn on the light, thus acting as the switch.
What I propose is that the way the switch is positioned it is already at potential, even if there is no path for the electrons to flow.
It's a bit like the Schrodinger's cat: try to touch the two opposing wires where the switch is to measure any power? Well now you will see

[IanB]

Further pontification on the subject of "flow" and what does it actually mean?

According to standard electrical theory, a conductor carrying a steady DC current has a magnetic field around it. But the same conductor carrying no current has no such field.

However, what does electrical current, or flow of charge, mean in this situation? The wire with a current, and the wire without a current, are identical in internal structure and charge distribution. Furthermore, every electron is identical to every other electron, therefore if electrons had been displaced, how could you even tell?

It seems that the distinguishing manifestation of current in a conductor is the magnetic field surrounding it. Why, therefore, say that the current induces a magnetic field? Why not say that the magnetic field actually is the current, since this is the observable, detectable manifestation? It seems more real than magic pixies inside the wire.

[End of philosophical musings]

[bdunham7]

For the pipe system, the power transfer is calculated at the boundaries. We can calculate the work done by the surroundings to push 1 kg of water into one end of the pipe, and we can calculate the work done on the surroundings when 1 kg of water leaves the other end of the pipe. So magically, energy has been transferred from one boundary to the other. But it might be a different kg of water that came out compared to the one that went in. So the energy was not carried along the pipe by that kg of water. Perhaps we can have a model that says the energy transfer was by a combination of pressure gradients and mass fluxes, and that satisfies us.

But philosophically, if the water is entirely uniform, the state of the pipe before and after is identical. If nothing has changed in the pipe, what part did the pipe play in the exercise? Apparently the pipe is not material to the result of the calculations. Only the boundary conditions matter.

Not 'magically', there is a process occurring within the pipe whether you observe it or not.  Something--a pump--did work on a kg of water by pushing it a distance into the pipe against a force (pressure).  If you like to think in 1 kg increments, then that kg also pushes on the next (or previous) kg also doing about the same amount of work, and so on, until the penultimate kg does the work on the last kg by pushing it out of the pipe against some sort of load, whether that is an orifice or a turbine or whatever.  If this is a continuous unvarying process, then the forces against the pipe are the same the entire time and the pipe doesn't see any of the work except for a small amount due to turbulence and friction--even though the pipe provides the reactionary force that restrains the water from moving in any direction except along the pipe.

So steady DC current...continuous displacement of charges....constant magnetic field around wire.....static electric field from wire depending on local charge density which is invariant....you can fill in the blanks.

[bdunham7]

Why not say that the magnetic field actually is the current, since this is the observable, detectable manifestation?

The same reason that we don't say that the heat produced by the current in any non-superconductor is the current.  It is an effect with a causal relationship, but it doesn't adequately describe 'current'. 
But it is true there is an extremely close relationship between magnetic fields and current--you don't get one without the other. 

[EEVblog]

Now we ask "how is the energy being conveyed?" There is no electric field in the (ideal) wires, any flow of charge through the wires has no generation/loss of electrical energy by definition of the wires being ideal wires in a DC system. In fact, in our theoretical system, the perfectly matched source and load dipoles effectively (but not completely apart from within the wire) cancel each other out with only substantial fields in a parallel region between the dipoles. Again, where is the energy "coming and going"? the creation and loss of electrical energy (from and into other forms) happens only "inside" the source and load.

*snip*

Thus we see the "true" purpose of the wire is just charge conduit to maintain an certain equilibrium where there is a flow of electrons through both the source and the load but no net change in the total charge of the system, charge is cycled in a circular manner between the source and load, a "circuit" if you will

The actual "energy" is coupled via the electric field.

Yup, inside the wire.

I'll say it explicity in case it's not clear what I'm getting at. The Veritasium video is all about the power/energy flowing outside the wire.
At DC that doesn't happen, it's inside the wire. Whatever physics mechanism you want to use the describe that doesn't matter, because whilst there is an external magnetic field around the wire at DC, it's not moving. And therefore by definition the mechanism of power transfer must hence be inside the wire.

This is why I think (and probably others) that this is getting a bit philosophical, akin to the Lewin vs Electroboom. How you you define where energy is "flowing"?

Correct. In fact I'd go as far to say the argument, at steady state DC, is absolutely pointless to the practical field of engineering.
This is why engineering has developed it's own laws, tools, and models to solve these problems.

So what's more important, the wire or the electric field?

The wire.
Good luck trying to deliver 1000W through your 30 guage wire and its magnetic field.

[EEVblog]

One question for you Dave (because I'm going to go out on a limb and guess that your next objection is me conflating "power flowing in a particular place" with "magnitude of Poynting vector", i.e., just accepting Veritasium's video at face value):

Veritasium (and many physicists) propose the Poynting vector, S = E x H as the best way to answer the question "how much electromagnetic power is flowing in this particular point in space". If you don't like this, what do *you* propose as an alternative? IMHO P = IV isn't really up to the task of answering "how much power is flowing in this particular point in space", because I and V aren't fields in space (or at least, V isn't well defined until you choose a ground point).

Or, do you agree that the Poynting vector is the right formula to use, but disagreeing with Veritasium about what that field actually looks like in the case of a simple steady state battery+lamp circuit?

Having looked at this again, I do not disagree.
What I'm saying is that it's of no practical value to an engineer to analyse it like this.
Just like his video is of no practical value to an engineer. Engineers have developed tool methods for practically implementing the flow of energy and signal transmission for this very reason.

[EEVblog]

Have to say. These are some fantastic explanations and explorations being posted by people Much better than my own attempts.
Edit: Dave, you could always try the explorational/learn along type video if this is more than you'd want to present authoritatively.

I cannot present any physics authoritatively. I'm not a physicist, and I suck at physics. I'm just a lowly practical engineer who knows this is all a game of spherical cows.
At DC it's just stupid and pointless to argue any of this.
If the physicists want to design practical stuff using their model of Poynting vectors at DC, go right ahead, I'll sit on the sidelines laughing my arse off.

[EEVblog]

I'll say it explicity in case it's not clear what I'm getting at. The Veritasium video is all about the power/energy flowing outside the wire.
At DC that doesn't happen, it's inside the wire. Whatever physics mechanism you want to use the describe that doesn't matter, because whilst there is an external magnetic field around the wire at DC, it's not moving. And therefore by definition the mechanism of power transfer must hence be inside the wire.

About the moving/not moving question relating to the DC magnetic field.

There is a simple analogy for this. Consider a motor/generator arrangement, with a cylindrical shaft connecting them. Apparently, the rotating shaft is conveying power from the motor to the generator.

But now, suppose the shaft is entirely uniform in its internal structure and external appearance. The surface is absolutely smooth and free of blemishes. In this state, any rotational position of the shaft is identical to any other position. If you close your eyes and someone rotates the shaft, you cannot tell afterwards if, or by how much the shaft has rotated, because it is completely and 100% uniform in every way.

This leads to an interesting situation. Nothing is changing in any measurable way, and yet, somehow, power is being transferred from the motor to the generator without any observable movement in the system. If the only thing you are allowed to observe is the shaft, you cannot tell if, or how much, power is being transferred at any instant. The flow of energy cannot be seen or measured while it is in transit. It can only be detected at the origin and destination, by its effects.

Similarly, with the DC circuit arrangement, all the fields and potentials, electrical and magnetic, are unchanging. The transfer of power does not require observable movement.

That is just physics hand waving. How is the power transfered outside the wire in the magnetic field under DC conditions?
Remember, the entire premise of the video is that the energy/power flows outside the wire. We all know how to explain this for high frequency, and it's useful at a practical level. But how do you explain it at DC that is of any use practically.

[EEVblog]

What I'd like to see in a video from Dave:

1. Understanding what is really going on with the bulb turning on "instantly" according to the distance between the bulb and switch: (it just the capacitance).
2. Relaxing the ridiculous assumptions that make the above result seem significant (infinite impedance in the bulb, zero impedance battery, and zero resistance wire)
3. Explaining how I^2 R losses are extremely important in practice, and they DO depend on the volume of copper, and so it is still correct from an intuitive point of view to say that the copper carries current.

All that is planned to be in the video.
Step 1: Show how this is a very basic transmission like problem.
Step 2: Explain how there is nothing new here for an engineer, we know about drift velocity, we know about power being transfered in the electromagnetic fields, it's all fundamental stuff.
Step 3: Explain how the undersea cable thing was an early development of transmission line theory.
Step 4: Explain how engineers have developed models, tools, and methods to get practical real world results istead of having to use Poynting vectors et.al.
Step 5: Explain the DC conumdrum and how there is essentially zero practial in continuing the extend Poynting vector theory into the DC realm. It's just a game of spherical cows. And how this is just another silly physicist vs engineer debate like Walter Lewin and how KVL is wrong.

[bdunham7]

You can observe this with a rubber garden hose having a spray nozzle. If you shut off the spray, the hose will bulge due to the backed up water pressure. If you open the nozzle the hose will relax. Whether the nozzle is open or closed, the water inside the hose has exactly the same energy--the energy it had when it came out of the faucet--the only difference is whether it is flowing or not.

This ignores the flow-dependent pressure drop in the faucet and plumbing....

[BrianHG]

What I'd like to see in a video from Dave:

1. Understanding what is really going on with the bulb turning on "instantly" according to the distance between the bulb and switch: (it just the capacitance).
2. Relaxing the ridiculous assumptions that make the above result seem significant (infinite impedance in the bulb, zero impedance battery, and zero resistance wire)
3. Explaining how I^2 R losses are extremely important in practice, and they DO depend on the volume of copper, and so it is still correct from an intuitive point of view to say that the copper carries current.

All that is planned to be in the video.
Step 1: Show how this is a very basic transmission like problem.
Step 2: Explain how there is nothing new here for an engineer, we know about drift velocity, we know about power being transfered in the electromagnetic fields, it's all fundamental stuff.
Step 3: Explain how the undersea cable thing was an early development of transmission line theory.
Step 4: Explain how engineers have developed models, tools, and methods to get practical real world results istead of having to use Poynting vectors et.al.
Step 5: Explain the DC conumdrum and how there is essentially zero practial in continuing the extend Poynting vector theory into the DC realm. It's just a game of spherical cows. And how this is just another silly physicist vs engineer debate like Walter Lewin and how KVL is wrong.

Sounds like a >45min video.  Will this be the first you attempt a partial script?

[rs20]

What I'm saying is that it's of no practical value to an engineer to analyse it like this.
Just like his video is of no practical value to an engineer. Engineers have developed tool methods for practically implementing the flow of energy and signal transmission for this very reason.

I fully agree with everything you say here. If you have a steady state DC circuit, solving field equations in 3D space and integrating Poynting vectors over a plane slicing the universe into two halves is ridiculous, impractical, inefficient, etc etc.

But, just don't forget that if you go to all that ridiculous effort, it gives the correct answer, despite all the apparent power flux being localized outside the wires. So labelling this method of analysis as "totally impractical" is 100% fine, labelling it as flat-out "wrong" would be... rather difficult to defend? (But reading between the lines I suspect you've already reached this conclusion.)

[EEVblog]

Further pontification on the subject of "flow" and what does it actually mean?
According to standard electrical theory, a conductor carrying a steady DC current has a magnetic field around it. But the same conductor carrying no current has no such field.
However, what does electrical current, or flow of charge, mean in this situation? The wire with a current, and the wire without a current, are identical in internal structure and charge distribution. Furthermore, every electron is identical to every other electron, therefore if electrons had been displaced, how could you even tell?
It seems that the distinguishing manifestation of current in a conductor is the magnetic field surrounding it. Why, therefore, say that the current induces a magnetic field? Why not say that the magnetic field actually is the current, since this is the observable, detectable manifestation? It seems more real than magic pixies inside the wire.

Because engineers have ot desing practical stuff. Going around thinking that the magnetic field is the current adds zero value. In fact it's just detrimental.
Skin effect is zero at DC, so therefore there is no value to be gained in thinking about the magnetic field other than a by-product of the curent flow.

[EEVblog]

What I'm saying is that it's of no practical value to an engineer to analyse it like this.
Just like his video is of no practical value to an engineer. Engineers have developed tool methods for practically implementing the flow of energy and signal transmission for this very reason.

I fully agree with everything you say here. If you have a steady state DC circuit, solving field equations in 3D space and integrating Poynting vectors over a plane slicing the universe into two halves is ridiculous, impractical, inefficient, etc etc.

But, just don't forget that if you go to all that ridiculous effort, it gives the correct answer, despite all the apparent power flux being localized outside the wires. So labelling this method of analysis as "totally impractical" is 100% fine, labelling it as flat-out "wrong" would be... rather difficult to defend? (But reading between the lines I suspect you've already reached this conclusion.)

Yes. Like I said before, it's not wrong, I'm sure the physics works out, there is just no point to it outside of a physics lecture. And this is why physics courses do not produce engineers.

[EEVblog]

What I'd like to see in a video from Dave:

1. Understanding what is really going on with the bulb turning on "instantly" according to the distance between the bulb and switch: (it just the capacitance).
2. Relaxing the ridiculous assumptions that make the above result seem significant (infinite impedance in the bulb, zero impedance battery, and zero resistance wire)
3. Explaining how I^2 R losses are extremely important in practice, and they DO depend on the volume of copper, and so it is still correct from an intuitive point of view to say that the copper carries current.

All that is planned to be in the video.
Step 1: Show how this is a very basic transmission like problem.
Step 2: Explain how there is nothing new here for an engineer, we know about drift velocity, we know about power being transfered in the electromagnetic fields, it's all fundamental stuff.
Step 3: Explain how the undersea cable thing was an early development of transmission line theory.
Step 4: Explain how engineers have developed models, tools, and methods to get practical real world results istead of having to use Poynting vectors et.al.
Step 5: Explain the DC conumdrum and how there is essentially zero practial in continuing the extend Poynting vector theory into the DC realm. It's just a game of spherical cows. And how this is just another silly physicist vs engineer debate like Walter Lewin and how KVL is wrong.

Sounds like a >45min video.  Will this be the first you attempt a partial script?

It might be actually. Maybe not script word for word, but at least have my points in order and stick strictly to them. It needs to be a 10min video I think.

[bdunham7]

Remember, the entire premise of the video is that the energy/power flows outside the wire. We all know how to explain this for high frequency, and it's useful at a practical level. But how do you explain it at DC that is of any use practically.

If you're going to do a smackdown video on this, you might want to also make the point that to the extent that there are EM fields outside the wire, those are generally seen as parasitic losses when it comes to power transmission, not a mode of energy transfer.  At DC there is no EM field propagation, only local fields due to the steadily moving charges.  As the frequency increases, there is more EM radiation that takes energy out of the system, which is why we transition to transmission lines such as twin-lead, coax and finally waveguides--those are to keep those EM fields contained and the energy going in the right direction.  IOW, the reason we have coax is because at 1GHz, the 'fields outside the wire' will radiate all the energy, not direct it down the wire.

[tszaboo]

I think his theoretical physics checks out, that's why things like RF waveguides work. What doesn't check out, is the animations and the pictures, they are way out of proportion, and some of the arrows, going from the battery to the bulb are nonsensical.
And it all this "power flowing" discussion depends on the definition of power in an electrical system. You can define it as P=UI or you can  define it with deltaE and deltaB. Both are right, one is unusual.
However, he totally drops the ball with that 1 lightsec long battery cicuit. The question is, what is the step response of a very long twin lead transmission line. We can measure the inductance of a 10m long cable, calculate the mutual capacitance (attofarad or something like that) and it is possible to build a lumped element model of the cable. And The model doesn't even need to go to 1 lightsec, because hey, scopes are fast. So let's take 1ms, or 300 Km. The inductance is probably quite large, resistance as well.

[IanB]

This ignores the flow-dependent pressure drop in the faucet and plumbing....

Not ignores, neglects. The result would be the same even with hypothetical pipes and hoses that have no flow resistance. Just the same as the assumption of perfect wires with no electrical resistance in circuits.

[sandalcandal]

What I'd like to see in a video from Dave:

1. Understanding what is really going on with the bulb turning on "instantly" according to the distance between the bulb and switch: (it just the capacitance).
2. Relaxing the ridiculous assumptions that make the above result seem significant (infinite impedance in the bulb, zero impedance battery, and zero resistance wire)
3. Explaining how I^2 R losses are extremely important in practice, and they DO depend on the volume of copper, and so it is still correct from an intuitive point of view to say that the copper carries current.

All that is planned to be in the video.
Step 1: Show how this is a very basic transmission like problem.
Step 2: Explain how there is nothing new here for an engineer, we know about drift velocity, we know about power being transfered in the electromagnetic fields, it's all fundamental stuff.
Step 3: Explain how the undersea cable thing was an early development of transmission line theory.
Step 4: Explain how engineers have developed models, tools, and methods to get practical real world results istead of having to use Poynting vectors et.al.
Step 5: Explain the DC conumdrum and how there is essentially zero practial in continuing the extend Poynting vector theory into the DC realm. It's just a game of spherical cows. And how this is just another silly physicist vs engineer debate like Walter Lewin and how KVL is wrong.

Sounds like a >45min video.  Will this be the first you attempt a partial script?

It might be actually. Maybe not script word for word, but at least have my points in order and stick strictly to them. It needs to be a 10min video I think.

Giving "good" coverage of all that within 10 mins is going need a really tight script. Will be impressive if you manage to squeeze all that in only 10 min (or even 30min), particularly considering your usual presentation style. The original Veritasium video is nearly 15min and we can see how much that "skips over".

[EEVblog]

Giving "good" coverage of all that within 10 mins is going need a really tight script. Will be impressive if you manage to squeeze all that in only 10 min (or even 30min), particularly considering your usual presentation style. The original Veritasium video is nearly 15min and we can see how much that "skips over".

It would have to be pretty much just verbalising the points I've made here, there would be little if any proof or detailed explaination.
The only thing I'd really be explaining is the fact that practical engineeirng differs from theoretical physics.