2nd analysis, I think I was originally wrong! - Veritasium 'The Big Misconception About Electricity'

I made a video response to Veritasium's "The Big Misconception About Electricity" video, and while most people agreed with my response and Dave from EEVBlog had a very similar analysis, after thinking about it some more I think both Dave, me, and pretty much everyone was wrong and Veritasium was right all along!
A few commenters on the original video correctly pointed out that I should probably try real C and L values. I originally brushed it off thinking these values would be so small that it would be inconsequential, and while I know of transmission lines I assumed the characteristic impedance would be in the megaohms.
I think a lot of people familiar with ECE also thought the same. Even the EEVBlog video mentioned the current would be really "nuff all", but this is not true! When plugging in real world values, we actually do get a significant amount of power, enough to light a LED.
Not only that, this new analysis shows the light can actually be continuously on, and that the instantaneous "blip" on-time from both my original video and EEVBlog's video is probably also wrong and an artifact of how we did our analysis.
Also this analysis shows that only under a very special case will you have a 1s full brightness response. Most cases either the light is on immediately, or ramps up way after 1s. Basically this core fact that almost everyone agreed on was at least true is also kind of wrong!
I originally believed that Veritasium was correct and made my video to show Derek was correct, but only on a technicality because I thought he needed a magic light bulb that lights under even a tiny current. However now I think he is correct all along and it's even realistically feasible to boot.

Addendum: With a perfectly matched load of 1.5k ohms, here the steady state power is 96mW, and our transient response is half that at 48mW (50%). If we use a 300 ohm load, we only get 2.7% power transfer. With a 6k load, you'll get lower power but a higher % of power at 10V (75% in the first 1s).

Timestamps
0:00 Background
1:20 Original Model and Explanation
3:35 Quick Transmission Line Characteristic Impedance Review
6:30 Simple Transmission Line Simulation
8:28 Simple Transmission Line Simulation - Unmatched Loads
10:33 Simulation of Veritasium's Question
12:16 Using a Realistic Light Bulb
13:34 The "Blip" Mistake
15:31 The "1s" Mistake
17:45 Using a Light Bulb of Higher Resistance
19:12 Conclusions
20:49 Bonus: simulation with cut ends
23:18 TLDR Summaries

[timbeaton]

Interesting to see that another channel has actually done this as a real experiment, although only using about a kilometre of wire, but got confirmation of the initial transient, ramping up to the full voltage after a short delay.

[clems]

First folks, the batt is DC. And the wiring is Not a TEM transmission line..
.

[victortitov]

Finally! A thorough analysis that doesn't have glaring errors! (at least from my level of knowledge). Thanks. BTW, this stairstep rise in current is provided by veritasium himself, but it's only a glimpse, and full explanation is in expert's responses Ve himself linked below his video.

[Observer]

Z Y,
All of this reasoning is PERFECTLY correct!!
THIS IS THE FIRST VIDEO I have seen out of four or five that has the analysis almost exactly correct.!!!
As a *VERY* experienced RF engineer, I can say that this is the first time I've seen anybody doing anywhere near correct analysis. Well, it's actually the second, technically, because the video by Ben Watson, who did an HFSS simulation shows the initial delay before the small current flows in the bulb.
However, he didn't really acknowledge it very well in his video, but it shows in his time domain simulation around time 9:30.
.
Once you understand that the transmission line characteristic impedance looks like a resistor it becomes very simple.
.
When you close the switch you have the two characteristic impedance resistors in series with the battery and bulb except they are joined by 1 meter wires.
I was really getting worried, but I finally saw that around time 20:30, you finally mentioned the one meter distance between the battery and the bulb.
.
It will indeed, take the time for the fields to traverse that 1 m and create that small initial current through the bulb.
.
You are correct that you should not model the transmission line with the capacitor first. That leads you to the wrong conclusion of that spike.
.
All of your simulations about the reflections are also correct.

Now, it does take 1 meter's worth of time travel for those fields to couple between the wires and create the characteristic impedance that looks like the resistor.
Ben Watson used different distances, but it does show an initial time delay equivalent to the 1 m distance in the veritasium model and it shows the low current after that time delay. Then, it shows the higher current later on.
.
However, all, and I repeat, all of your analysis of the reflections with the various values of resistor are COMPLETELY correct.
The only thing that appears incorrect here is the fact that you shows that that small current flows immediately after the switch closes, but then you talk about the one meter distance later on.
So this video has it correct even though you take way too much time for me, obviously to explain it for all the people who don't know transmission lines.
This is really a quite trivial problem for an RF Engineer.
However, the way Veritasium laid it out, he made it extremely complex and I'm close to agreeing with Dave-EV blog, that he did it on purpose.
.
Also, the video by Ben Watson using the HFSS simulation clearly does show the equivalent of that 1 meter delay in his time domain plot out around 9:30 video time.
Regards
P.S.
I didn't listen to the rest of the video where you open circuit the two ends and figure you probably got that correct.

[skepticalextraterrestrial]

1m/c is 3.34 nanoseconds, and you are simulating in 5 microsecond time steps. Moreover, you are using a transmission line model but the current at the lamp at t=1m/c depends on the current at the source at t=0+. And the current at the source at t=0+ is not a current flowing across an infinitely long line, but a moving point charge. This is a dramatically different problem. The magnetic field of a point charge falls as the square of the distance. Even at 1m/c seconds, you have a charge flowing only along 1m of wire (and not the full current yet at that, thanks to inductance). The magnetic field produced by this 1m of current is negligible. If it weren't negligible and produced anything like you are calculating, then the wires in the walls of your house would interfere with everything in the room. The trick here is that the line is LONG, and that adds up to a LONG electromagnetic field. BUT that LONG electromagnetic field is not established in 1m/c seconds. Finally, Derek's video misleads by giving the impression that this is the mechanism by which your house circuits are powered and that "energy does not flow in wires." This is not an accurate description.

[bobwhite]

Appreciate the effort - although using sensible values for C and L (not just ratio) would have shown a much less significant spike at the outset (which would have made that red herring of a an initial spike less of a distraction in first place). In any case, without Veritasium giving the specification of the bulb, it is impossible to conclude whether being lit is a plausible outcome. So even though you are showing several milliamps, the power to a matched bulb (1.5K impedance) would be less than 100 milliwatts. Definitely enough to light discrete LEDs, but that is not what the bulb in video is. It is showing as pretty bright in a full sun background. Even a very efficient LED bulb would be around 5W to do that. So, 1/50 the power arrives at first, and eventually, at steady state turns into the full power. The CLEAR inference from the video was that since power is transmitted in the fields, we can expect something on the order of full power after that first few ns. Yes, he mumbles something about the reflections needing time to settle, but getting almost two full orders of magnitude less power than full power at the magic t = 1m/c mark is bogus for claiming the bulb is lit. He was trying to make a point, but used a vastly underspecified and nebulous problem statement to make it. It really takes away from his otherwise decent treatment of the topic.

[argcargv]

Another thing to consider is that a led is a nonlinear circuit element. It will look like a high impedance until the led starts to conduct. In that way it would sort of self match to the transmission line impedance so it would be guaranteed to light.

[motronix-gr]

Steady state analysis is very different from transient analysis. I don't understand why you mix and match equations of steady state with the step function...For example Impedance can be used only in steady state analysis, but you use it when the input is the step function...there is no definition of transient impedance...that i know of ... at least... unless you say that Impedance=dv/di ... very confusing for me... (Impedance is valid only for sinusoidal voltage and current ONLY in the steady state... correct me if i am wrong...)

[rgmoore]

What if the battery and the light are 1m apart but the wire path isn't. For instance, if the wire path takes some enormous loop as far apart from each other as possible.

[Green_House]

Not really the main point of the original video.
Veritasium assertion was that energy doesn't travel in the wires but it does travel in fields around the wires.
We all know that when the switch is activated, there will be a transient step voltage in the circuit and various capacitive and inductive effects come in to play, but what's happening after say 1 minute once all transients are over. Does the energy travel in the wires or fields? And remember, we are talking DC here!

[ZongyiYang]

Addendum: I forgot to mention just how much power is transmitted at steady state compared to the initial transient power. Our best case scenario is the perfectly matched load of 1.5k ohms, here the steady state power is 96mW, and our transient response is half that at 48mW (50%) (Edit: also "best case" as the most power transferred, you can actually get higher % of power with higher resistance but lower overall power). If we use a 300 ohm load, we only get 2.7% power transfer. If we move to something like a 12V 100W light bulb, we'd only get an initial transient power of 0.09mW compared to the 100W at steady state (almost 0%). So whether or not you consider the energy to be transferred through the air is significant compared to the wire is dependent on what load is chosen.

With a 6k load, you'll get lower power but a higher % of power at 10V (75% in the first 1s). That is significant! Whether or not you think using a small lower power light bulb vs a 1 or 100W light bulb is inline with the spirit of the problem is up to you!

[ButchNews]

Oh, and another thing. He was WRONG about the trans cable iron cladding. It IMPROVED the signal. Think co-axial cables. He didn't understand the cable issue which, even today, has about a 60 milli-second delay.

[ShiehJimmy]

Thank you for running the simulation for us. I didn’t even think the resistor and the reflection would affect the results!

[smostars]

Veritasium's application of the Poynting Vector is the problem. Feynman talks about this misunderstanding in his Lectures volume II book.

[ggcmod]

The main point of all the controversy is that it must be defined when "the light bulb is on", because yes, the simulations indicate that there is an energy blip, but the point is that in the Veritasium video at the end Derek says that the bulb will turn on immediately after one second.

[X-boomer]

I had a go at this last week because like you I thought the line separation was way too high and current would be negligible. But I used a 1mm wide conductor which gave me an impedance of 1.0kΩ, and a standard 12V 0.25W bulb resistance of 48Ω

I think the conclusion I would have to take away from all this is that the devil is in the detail and all veritasiums hand waving about ideal this and that is fundamentally wrong. If you choose impossible values for a problem that is dependent on a complex interaction of factors you might not even have thought of, you're going to get an impossible answer.

[urnoob]

question was totally deceptive by not putting the meter unit
i had no idea what 1/c meant before thinking about it smh

[rsmt]

14:30 I think this is why models usually have the initial and final elements be L/2 (or C/2). Splits the difference between having too much influence from inductance and capacitance.

[marcfruchtman]

I really like that you spent the time to run a simulation. A pretty amazing effort !
If the original question is unmodified, then the technical correctness is unchanged, and the question as originally posed is an EM Wave problem, not a transmission line problem. Though, I do think it was "intended" by Veritasium to be a transmission line problem.

I really admire the attempt to validate the power output of the transmission line theory to light a "non-magic bulb", because, ultimately, we don't use magic bulbs in real life. This sim appears to show fairly reasonable amounts of current reaching the bulb, which is somewhat unexpected. However, the voltage still takes quite some time to ramp up. So there would still be some delay to overcome the forward voltage of the LED.

However, I have to view the simulation with a lot of skepticism as it relates to "real world" results. Is there a way to simulate this with real world wire lengths and wire that has actual resistance and get similar results. For example, if you change this to two 10 meter loops of ~22 gauge copper wire (insulated) 1 meter apart, will you get reasonable correlation?

Two 10 meter lengths can be reasonably validated physically with oscilloscope. The actual capacitance and Inductance values of the wire can be measured. (Though i don't have equipment sensitive enough to do it).

I would really love to see something that "physically" correlates.

Thanks for spending the time to make this video and the simulation. Note: This doesn't change how I feel toward the Veritasium video. I still find it very misleading. And I disagree that this is how electricity is "really" delivered to my house in order to power my appliances which seemed to be a fairly major message delivered by the video.

[stephenjackson]

There is some energy that goes from the switch to the bulb in 1m/c seconds. But it is not the amount of energy that can travel in the wires. It's more like the bulb is experiencing a small magnetic force when the switch is closed (but not exactly a magnet, just sort of like that in intensity...) The real power to fully brighten the bulb to intensity must still traverse the entire wire. If the full electrical power traversed the gap, then you'd be shocked just as badly putting your finger between the bulb and the switch as you are if you grab the live bare wires when electricity is flowing. That decidedly does not happen.