I love how this video demonstrates that proper experimental design relates to just careful thinking and understanding of assumptions and first principles, and does not have anything to do with how fancy the equipment is. The whole experiment used dead bug circuits and a super cheap scope, but was crafted to be able to gain them necessary insights, and it does so with sufficient resolution to prove the points.
For improved accuracy with differential probes, you should have connected the grounds between them! The second thing, and I think you did do this, but non the less: It is very important to use the x10 setting on the probe for two reasons: to increase the input impedance of the probes (the thing you tried to do with capacitors), and to enable you to trim the probe's compensating capacitors (little screws on the side) to eliminate ringing. Trimming is done with a probe attached to the test generator contacts just beneath the display in the middle of your scope. The problem you saw with the probe disrupting the signal patterns is from the ground connection of the trigger probe attached to the switching circuit. You really should have added some proper analogue switchings to the circuit, like SN74AC66, and also separate batteries for the oscillator and the driving the line, to mitigate ground coupling through the scope.
While i agree with you that this is an excellent video, it does go into the weeds on some of the EE that your average joe would fine more confusing than the generalizations on the main channel.
This was really cool, please make more EE content. 90% of ECE content on youtube doesnt go past first 2 weeks of Electronics 1 so its really cool to see some applied signals content.
2:49 My EE background says for smoothest power you want a combination of different capacitors. I would use a 10uF electrolytic, a 100nF MLCC, and a 1nF ceramic in parallel. Each capacitor is most effective in a certain frequency range due to parasitic resistance and inductance. By combining different types and values you can cover the whole frequency range of interest.
@@AlphaPhoenix2 The idea that you need multiple different values or types of decoupling capacitors in the way that stefansynths suggests is a common misconception in electrical engineering. Generally the way you would want to use decoupling capacitors is to place as large a capacitance value with as low a parasitic inductance and (sort of) as high of an esr as you can get away with as close to the power pins of the device as you can in a way that minimizes loop inductance. Generally the lowest parasitic inductance capacitors you can get are those in the smallest packages you can get, so SMD ceramic capacitors. Nowadays these are available with relatively high voltage and high capacitance ratings, so there is generally no need to opt for electrolytic capacitors if you want a 10uF capacitor that can handle 9V. A drawback of certain ceramic capacitor types is that, depending on the dielectric, their capacitance can drop significantly when higher voltages are applied. Another drawback of ceramic capacitors is that they can have too low of an esr. There are controlled esr ceramic capacitors, but they are rarer and more expensive. If needed, you could put a resistor in series with a ceramic capacitor, though that obviously increases area usage, bill of materials and loop inductance. I find this video gives the best explanation about decoupling capacitors Watch code: WdlN8bHw-w0
In fact (and if people are still reading, 5 months on), above a certain frequency (governed by the parasitic parameters, due to the physical construction), capacitors behave as inductors (their complete opposite). They will still store charge as any capacitor, but the voltage vs. current waveform follows that of an inductor. That's why we use multiple capacitors of different values (actually orders of magnitude) and construction types (usually electrolytic and ceramic) and physical sizes to cover the full frequency range.
20:45 "And it's just that easy" is quite the statement! That is a LOT of painstaking work with many mistakes that can be made. Very impressive research and so cool that a reasonable digital oscilloscope can capture this.
The twisted pair is a transmission line. Your rapidly switched pulse is a mix of RF (wide band) and DC. We need to be careful and not equate the DC bulk Resistance of the twisted pair to DC flow, with the Impedance of the twisted pair to an RF wave front. Different critters. The pulse that is impressed and propagated at the velocity factor of the transmission line when the switch fires each time, is primarily a radio wave that rides up on the skin depth of the wire (frequency dependent) and its propagation along the wire is resisted by a combination of reactance and surface resistance, that we call impedance. Along with that is the DC current that is more controlled by resistance and propagates through the bulk of the metal at a different velocity. This can be shown by a long balanced line made of two, very thin wall, hollow copper pipes. Measure the DC resistance of the line, then measure the impedance and velocity factor of the line at some arbitrary, but stable, radio frequency. Now, replace the hollow pipe with solid round bar of the same diameter and repeat the measurements. The DC resistance change will be significant while the velocity factor of the balanced line to the pulse will remain close to the original.
All 3 vudeos wefe just amazing!! Seriously! I would make two minor changes though. For the Schmidt trigger, add a couple.of 0.1uF or even 10nF capacitors.ij parallel to the large ones you have. As close to the vcc/gnd pins of the chip as possible. Rhe big capacitors help but their internal resistance and inductance means that they are "slow". So for the very fast transients you want a fast 0.1uF capacitor bormally known as a "decoupling capacitor". This will give you much better rise times Swcond improvement would be to use a differential probe inatead of the math function. Thia way you sont have to worry about the disconnected ground leads and the inductance that those leads (pf the red and black probes) create. Remwbwr that all ground ports of the scope are common ao you have a "back" path or at least aome potential interference from the ground leads hanging around. A differential probe does the math in analog way nect to the measuremen point and ahould give you wven better noise immunity (although the traces did look very goos) Finally, given you hace a "repeatable" trace.. if you apace put the pulaea enough (say at 10khz) you can use the averaging funxtion of the scope to get better dynamic range /.noise performance by accumulating hundreds of runs of the same repetaed event. Some scopres (sampling scopes9 can evwn increas the time resolution by adding a slight jitter to the trigger and getvsamples at slightly different times but due to the repetitive bature of the signal, tgey can build a higher resolution (time wise) "Image" Finally ypu can use the trigger input od the svope instead of one of the channels, directly from the Schmidt trigger output (maybe uaing one of the channelse separately). That way you can sanple using channels 1 and 3 and thus run the scope at dohble the rate. For example of the scipe is 1gigasample/s that usually applies to channels 1 and 3 by themselves. But if you capture all 4 channels.or 1 and 2 (or 3 and 4) then the rate goes down to 500 megasamples per channel
When I watched the main video I thought, wow, he must've had a shitload of probes to take all these measurements at once. It never occurred to me that you just took a shitload of measurements with three probes. Thank you for your hard work!
Itis scary how much work and thought went into this. Crazy inspiring work. Would love to see how you animated the electron motion according to the observed experimental voltages.
Brian, this is probably the most interesting series of videos I’ve ever seen. No hyperbole here. So many levels of fundamental understanding about electricity, impedance, metrology, being shown and explained in an intuitive way. One question I was left with: how does current depict itself in the top graph of voltage? Is it the increased area under the curve? The slope of the wave front? I know charge has to flow to change the voltage within the line, but I’m having trouble connecting it to the measured values. Thanks.
the "fake electrons" in the diagram, and the associated "electron drift speed" plot were based on how much current would need to flow in order to see the observed changes in potential along the wire. The first bounce is easy: If you assume the capacitance of the cable is the same everywhere, then every additional volt is a proportional change to the electron density (linear density) in the wire. Once you pick an arbitrary (ridiculously massive) capacitance, you can invert the "electrons per length of wire" plot to get a "distance to the next electron" plot at any location along the wire. If you know the electron at the far end of the line isn't moving (for relativistic reasons, don't need to assume anything about electricity), you can use this to iterative draw each "electron dot" spaced back from the end of the wire the correct amount. The really cool thing is that as you play the animation forward, this construction (based completely on a fudged electron-density-to-voltage characteristic, produces motion, because you NEED electrons to shift down the wire slightly to produce these voltage waves. Subsequent bounces are harder, because then you don't have any stationary electrons to use as a foundation, and you need to grab an arbitrary (actually carefully algorithmically chosen but whatever) and say "this particular fake electron isn't going to change in speed this iteration" and you base everything off of that. I realized after the fact that I could have based everything off the electron at the supply end and then measured the ACTUAL current using a small resistor (which I did at one point), but I was actually much happier to see that you can get good realistic motion based exclusively on assumptions about what current is necessary to produce a certain voltage distribution. My FIRST attempt was actually a full simulation - I dumped a bunch of electrons in the wire and let them be accelerated by the field, but unfortunately the scope traces at different distances were SLIGHTLY offset from each other, and that meant that while a human could identify the waves easily, the algorithmic electrons were bunching up as they got attracted to all the slightly-elevated trace locations and it was a nightmare.
Thanks. Took me a bit to work through, but I think I got your approach: Represent voltage i.e. relative charge density as the direct inverse of distance between the charges. Back-calculate the position of each charge relative to the stationary reference. Add the incremental movement of the reference to all charges. I puzzled through my question again, and this is my long-winded logic on current in the voltage graph: Consider arbitrary positive charges, zero net charge initially, and a perfect step input wave. When the “voltage wave” travels through the length of wire, the net charge in the wire increases. This increase is proportional to the voltage(charge density) and the length of wire(capacitance) under/behind the wave. The speed of the wave determines the amount of charge moving through the starting point per unit time. Charge movement is current. So the voltage integrated over the length of wire is total charge, and the change of this integral over time is current. If we move the virtual “current sensor” to somewhere along the wire, we can just integrate starting at that point. Of course, this breaks apart as soon as there’s a current coming in from the end.
AlphaPhoenix: "We take thousands of measurements so you don't have to." Mind blowing project. Amazing attention to detail. And a result that changed my understanding of electricity. I'm subscribing... and continued success!
To switch the Hex schmitt trigger inverters quickly you need to supply current to the IC over a large frequency band (Fourier analysis), which means you have to have instantaneous current flow at high frequencies to minimize the signal rise and fall times, meaning smaller value supply capacitors (higher self resonance) as you get closer to the IC power pins while minimizing the device lead inductance (lead length). But, you will quickly be limited by the IC package internal lead inductance. You will have several capacitors in parallel of different values and types all working together to meet the high frequency IC switching current demand. The aluminum electrolytic capacitors in the video have much too small a self resonance frequency to supply the needed current to switch the IC quickly, and the bulk capacitance is largely negated by high inductance of the long lead lengths.
These three videos have been the most fantastic learning experience I have had in a very long time. I have for years always wondered about this exact same thing and now I feel like a window has been put in a place where there was once a wall. Thank you so much for your work!!!
This project is a masterpiece. Well done. Thanks for showing the setup as well, this answered all my questions, along with all the questions I didn't ask but should have
Regarding using the electrolytic caps to bypass the schmitt trigger's power... the common practice is to add a ceramic cap in parallel with the electrolytics. The ceramics are much faster to respond to any changes. As for the differential probing of the twisted pairs... the probes already have a much higher impedance than the twisted pair. Blocking the DC with the series caps doesn't seem like it would improve things. Differential probing is required, of course, because you want the voltage difference between the red and black wires, and that black wire is definitely not at the scope's ground potential. I really enjoyed the part where you explained the need to get data at all of the points and animate it, and then said "it's just that easy". 🙂
good to know! the vast majority of my electronics experience is with very slow things so electrolytics have been good. maybe those weren't doing anything at all here lol
@@AlphaPhoenix2 Doing this sort of high frequency electronics work has a variety of pitfalls. You might want to browse a book titled High Speed Digital Design by Howard W. Johnson, ISBN 0-13-395724-1. There are techniques to improve the scope probe performance too. This video has a quick explanation of why you can't use the scope probe ground lead for high frequency measurements.... ua-cam.com/video/henwiLzSieA/v-deo.html
If you want perfect equal current on pluss and minus cable, maybe just use a small high speed transformer as your supply. Current should be equal on both sides in my mind 😅
Thank you for this detail, brilliant set of videos. (The description's link to your main channel has a funny character at the end, doesn't work without editing target)
Very nice explanation about how you made the measurements and processed the data. To be honest, I thought you were pulling it via USB directly from scope, but saving the CSV is a very good and simple approach. I never thought about CSV saving because my scope is damn slow there.
After seeing probably that exact video about the schmitt trigger TDR a few months ago, I've built one aswell and used it to meassure the reflection in a few meters of coax. Allthough I was mostly trying to meassure the speed of light in said coax, aswell as learning about transmission lines in general. These 3 videos you made really halped me understand the concept of transmission lines a lot better, especially where that magical impedance value comes from and why reflections occur. I'll also definitley attempt some of the experiments you demonstrated with twisted pair transmission lines, especially using multiple probes to cancel out the crosstalk. Really love your videos and looking forward to more of them :) I also just so happen to have the exact same oscilloscope you do, it's really a lot of oscilloscope for a humane price
Could you put an instrumentation amp (powered by a coin cell, perhaps) at your measurement head, so as to have a lower-capacitance input? Also, with an instrumentation amp, you can drive the shields of the short cables actively to the mean of the input signals so as to minimize the effective capacitance. Center-tap the gain-setting resistor and run it through a unity-gain buffer. Failing that, they do make x100 scope probes so that you can have even lower input capacitance. In your pulser, you probably don't need those big electrolytics. The battery is effectively a huge electrolytic capacitor. You need a small amount of capacitance with very low inductance right by the chip. With the dead-bug setup you have, just solder a 100 nF ceramic between pins 7 and 14 of that 74AC14. I suspect you'd do well making your pulser around a line driver like the AD8132. Impedance matching when you launch the pulse does help! I saw you with a bunch of termination resistors. How close did the measured characteristic impedance of the line come to your calculations? Now I'm sitting here with a head full of imaginary Smith charts and wondering about whether I could design a current transformer to measure things less invasively. (But I haven't designed any magnetics in years, and I'd be really, really rusty at this point.)
Ok, that was awesome. Love the detailed explanation. If you could post the processing scripts and raw data for people to check their workflow, that would be extra (CC/OpenSource would be great, but could make it a Patreon perk).
Just wow! I can only imagine how much work it took you to make all of these. For electrical engineers the results are pretty easy to understand - the long wire is inductance and capacitance "distributed" along the wire length.
Great work on that visualization! I think your setup would greatly benefit from using coax instead of twisted pair, and from using BNC T-adapters as monitoring points (even though you've stated your motivation for a custom-made lines at the end of the video) - see my comment under your main video for more detail. By the way a great learning source about transmission line transients is a book by G.A. Mesyats called "Pulsed power", more specifically, chapter 2.
Could you do this same sort of thing with an antenna? Could you use that to see how and where it's picking up a signal? Would a received signal cancel itself out if it's wave is bounced back at cosign of the received wave? That why some antenna's like old TV antenna's have a branching fern leaf pattern to provide some provide some constructive interference at the main branch? Lastly does it move faster as it's an external force or the same because of how electrons move in the wire? That being the most like Veritasium example and probably the more interesting one to look at in relation to it. I'm sure it's just oscillating or resonating not actually moving through it like this though and that would be an interesting topic as well.
The yellow trace is not due to crosstalk. The chemical battery is not -5V and an infinite ground. It is a redox reaction, pushing electrons on one side and pulling the on the other. Half the potential is on each end. So you will see symmetric behavior in the positive and negative wires. Both of them are getting a wave of current due to the chemical action at that electrode.
You, Sir, did an incredible amount of work creating these videos. The thought, the setup, the programming, and so much more - boggles both the mind and the body. Extraordinarily impressive!
Terrific trilogy. IMO only two aspects are missing: 1) it need to be said that there is a voltage drop in your battery corresponding to the load 2) why does the voltage fixate at least temporarily after the first wave propagated? It is not obvious. On the one hand I can say that it stays same because each new section of low resistance wire looks exactly the same regardless of how many is left but that's not enough.
17:00 also you could just place the scope probe close to the wire without making any contact and it would be very similar in nature but less repeatable
I tried this (just clamping the insulation) but my probes had too much internal loading (maybe because I didn't tie the grounds together, as someone commented), but I was also worried about different amounts of wire squish altering the capacitor between the probe and the wire
that trigger circuit.. I love the idea, and you nailed this whole video (taught me some stuff; thanks!) BUT I know it's faster to just solder it together.. I get that traces (w/o a proper matched "ground" line) just get more complex as freq goes up. Still.. Dude.. there has to be a better way.. looking at it as is makes me cringe.. Maybe there's a pigment/insulating spray glue that's totally inert, easy to razor-off (like to measure etc), and simple to use..??? IMHO, you gotta just make a proper one.. the idea of a proper trigger/clk for your scope feels right what with all the research you do.. no? Oh.. and thanks a bunch for this series!! It's really helping me find the tiniest of holes in my foundation! very cool
Actually, I don't want to do this myself. Even though I have almost everything I would need. I am big fan of backseat science, and you left me wholly satisfied doing so much diligence I would have wanted to put in myself.
To anyone who wants to recreate this experiment, it will produce very poor results if you try to use a single-ended scope probe to measure a differential signal on a twisted pair by connecting the probe's ground lead (black aligator clip) to the second wire as a signal reference. That works when you're measuring single-ended signals on a circuit board with a common ground plane, but in a balanced transmission line, you need to make a differential measurement (like he is doing) using two 'scope probes (one on each side) and use the oscilloscope's "A-B" math function. Using two probes on each wire of a balanced pair allows both sides to be measured like a signal (because both sides *are* signals), and each side gets the same 'scope probe loading (if both probes are matched) and it doesn't unbalance the transmission line (creating different circuit conditions that alter the signal). Note, both scope probes need to be the same cable length and same probe type and should also be properly compensated for high frequency peaking with a square wave calibrator (the little screw adjustment on the probe box squares up the edges). If you try to connect scope probe ground clips as a signal reference, it will change the topology of what you're measuring because it alters the signal return paths of differential transmission lines by connecting them to the oscilloscope ground. It also connects them to AC power safety ground, if your oscilloscope is powered from AC. Using a battery powered 'scope can eliminate the AC ground connection, but if more than one probe is used, the scope probe ground references still get electrically tied together at the 'scope chassis. So, he uses two 10:1 probes to make each measurement, and this isolates the oscilloscope ground from the circuit under test, and it doesn't alter the transmission line's signal returns, nor does it unbalance the transmission line. Note, when doing differential measurements, you should still connect the scope probe ground clips to something that is common mode to your circuit. Ideally, they should be tied closely together and connected to the ground of your signal source. He is letting the oscilloscope tie the grounds together and using the ground reference of the trigger channel to tie it to the signal source ground. In short, using differential measurements are necessary on balanced pairs of wires.
I am amazed that you are willing to put in this much work. You must be loving it and you must be doing it because of that. As others have mentioned the trick to not introduce any big change by the testing equipment is to use very high impedance probes. Unfortunately many of your viewers of the actual test video do not realize that at those short time scales as you are using there is really nothing that can be called DC. They look at it from a DC point of view instead of from an AC point of view which I believe makes it more difficult for them to understand it. I have for a long time known that your results are the fact of what happens (and picked c) but I have never seen it so thank you. I have often seen the results of it in data communication and antenna cable performance.
I find it really difficult to communicate that short pulses are AC. in one video i went into detail (on a very similar circuit) and showed an FFT and everything and I had irate comments telling me that "pulsed DC" was different than "AC". whatcha gonna do?
@@AlphaPhoenix2 I appreciate your frustration. I must admit that I myself have changed my view on it over the last few years and that is after an electronic engineering education that I had about 45 years ago where I did learn that every pulse can be calculated as made up out of sine waves. At first glance one would of cause say that DC is when the current is just going one way and AC is when it is going forward and backward (alternating). That is a fine if we are talking about what most people see and experience most times. It is only when one get's into working or thinking about what happens on a time scale of less than a second (especially milli, micro and nano seconds) that it really matters. Personally I have worked with that with computers etc. since I finished my engineering education. Now after I have retired I have started to analyse things a little better as I also have some interests in physics and then I realize that the little bastard electrons are what is moving charges around and they or their influence can not move faster than the speed of light which compared to frequencies isn't that fast after all. To explain that a pulse is AC the one thing I can think of right now is to explain how such a pulse can be made up of higher frequency harmonics of the pulse and that must be the case in real life too as if all higher frequencies are filtered out a pulsed DC will only rise and fall as a sine wave of the pulse frequency. It also demands to explain that all circuits consist of inductors, resistors and capacitors. The functioning at super cooled circuits where the resistance may be zero I am not quite sure of now. I am still learning at age 77. A place where a pulsating wave is used to make a higher frequency - a harmonics - is in many high frequency radio transmitters. They often start up generating a lower frequency that is easier to control , then cut the sine wave's top and bottom to resemble a square wave and with filtration picking out the desired frequency. That was used in a TAXI radio I modified to use as a Ham radio transmitter many years ago. An AC pulse can of cause also be said to be a DC with an overlaying AC but the initial start when a DC is connected to a circuit can in my opinion always be seen as part of an AC wave. Unfortunately there are people that are so sure of them self's and what they have learned that they have to be unpleasant in their remarks. I know that although I have learned and experienced a lot then I really don't know very much for sure and if people can disprove what I am saying then I am willing to change my mind if it makes more sense to me after their explanation. To be simply told that I am wrong without an explanation doesn't cut it.
Thanks for the explanation of why you twisted your own wire - I'd also leapt to "can't you just use cat-5 or something", but needing the linear resistance make a lot of sense.
Great video ! I'm really glad , this video Explains and shows that you are aware of the many things And issues that 'may' influence the readings etc It Takes away the miriad of questions I had as to the accuracy and reliability of the setup used . So we can now view the Readings in perspective relative to all the other factors involved 🧐 of course there are still a couple of things I wonder about , But that is always where the next experiment comes from ! Experiment to clarify , demystify and fully understand that what we see and already innately understand by gut feeling
So Capacitors are shock absorbers? What's the delay of that? Reminds me of sound waves in baffles. Eventually leakage occurs and how you control that his how you transmit the energy. Entropy is such a good model for these reasons. More holistic. Something I find cool is using microscopic baffles with phonons to direct thermal energy where it is more conductive to escape and not reverberate in its container.
if you made a cube of up/down twisted pair, you could probably make your maze and measure and animate THAT. tho I would hate to have to do the data collection for that without automation.
Thinking about how twisted pairs naturally reduce EMI (the reason they are used in telephone and Ethernet cables). It would be interesting to see this same setup but not twisting the pair, just straight wires touching each other. How much worse would the EMI be? Would probably require some home built contraption to put a wrap around the pair to maintain even contact between them. The thinnest available speaker wire maybe? Thermocouple wire is often made this way, but it's pricey.
Рік тому
Oh you are using LDO, I was very confused about the voltage scale in the previous 2 videos.
I believe the reflection characteristics are going to be different depending on the frequency ur switching, but ur vid shows the general concept. Could be cool to show an impedance mismatch and it's the reflections in this style of vid
It would be really cool if you could somehow extend this to the flow of electrons in the battery itself. I would love to see a visual of the forces and flow enacted upon the electrons inside the battery.
since you AC coupled your probes at the tips how did you get steady state measurements? I thought everything was detailed and explained well and done great but I was a little confused on that part. A series cap on the tip of the probes should only allow for changes in the circuit to be measured and not the actual DC voltage right? I think you do an excellent job with these series of videos about circuits.
Just a question... Would setting the scope to AC coupling instead of DC coupling have the same (or similar) effect as what you did with the capacitors in series with the probe? Or would that make it hard to keep the readings stable over multiple readings at different positions on the wire?
Love the content it's so cool! An editing note: your narration audio is pretty quiet. Right click video, stats for nerds, volume normalized: -14db. That should read 0 for proper loudness.
Super interesting, thank you for going to all of this effort watching those plots slosh around is incredible! One thought though, you're using singled ended probes, yes you've taken two of them and used a math function to make them look differential but ultimately each measurement is single ended. The effect of that is that each has some input impedance to it's reference, a reference they share. In most cases this is like 10MR or so. The problem I wonder about is that because of the above you are effectively connecting the start of the cable run to the point you're measuring together through the scope. With around about 15MR, thas isn't much of course. But by a similar theory of the electrons "predicting" currents until proven wrong later by hitting the load, surely you are getting currents flowing through the scope, bypassing the cables you intend to measure? I'm sure the effect would be minor regardless. But one way to move closer to solving this would be an active differential probe. Better yet, two, one for each measurement point. I think ideally you'd have a scope with isolated channels, but you'd still want a differential probe to get the purest measurements
Would you be interested in a PCB with test points and possibly the pulse circuit built onto it? I could send it your way and have it done in a few days. the trace spacing can be as close as .15ish mm. could save you some space! I imagine the coupling would be worse though!
So for a circuit, there has to be a positive and a negative wire, right (heads up, might use positive/red/negative/black incorrectly, sorry...trying to use 'Positive' to mean electrons flowing INTO the wire, and negative to mean electrons flowing back to the battery): I'm a bit confused with the control wire to control cross-talk...I imagined the three 'looms' (just my nickname for them) each being one prong of the Y configuration. If this were the case, then both the positive and negative wires would have to be on each loom wrapped together, like a highway wraps two directions together within one string of highway....each loom being one 'string' of 'highway' containing both directions on the Y arrangement (thus the three looms, one for the beginning and a junction seperating that for both a looping string of 'highway' loom and a last string of highway that doesn't return/isn't connected). Is the 'control' wire just your negative wire? Or did I completely misunderstand the arrangement of the 'looms' and the Y prong configuration? Basically...would 20 probes on the black wire measure equal changes as 20 additional and separate but equally spaced probes on the red wire at the instant the battery is connected? Aren't we doing something differently, measuring them together via a single probe in your experiment? Who's to say electrons even move in both of the wires when the initial charge dives into the wire, couldn't that just be one of the wires while the other doesn't experience anything (unless the red and black are soldered together, in which case it does eventually make it around)? @AlphaPhoenix @AlphaPhoenixChannel Like, specific points on both the red and black wires seem like they are eventually graphed combined as a single point...why not separately? Did you have data that showed for certain that, for each length along the wire, changes happen simultaneously and equally to the positive and negative wires? For instance, I didn't understand how you can say/show (18:22 on original video) the red wire (negative?) wire ending up with fewer electrons than it started while also showing a graph that only seems to be demonstrating what's happening on the positive/black wire, when I can't really tell how you separated those measurements from "different sides of the highway at the same point" if at all. It might be an assumption, but I'm not sure how we could say (on a disconnected loop) the red/negative wire ends up with less electrons if we were only measuring points on the black wire (the wire we are 'pushing' electrons into). I'm wondering if when you say 'electrons slam into the end of the wire', I think we have to assume you mean they slam into the end of the black/positive wire only. But I'm more interested in the data...did you have data show what happened on the red wire when this experiment at 18:22 on original video was performed (as in, what we would assume to be the opposite of slamming into the end of a wire). If electrons do leave the red/negative wire, does that mean by pumping electricity into object A, we can use that to inherently pull out electrons of object B even without a circuit being created? I'm just concerned there might've been an assumption (or data was only shown for the black wire/half of the pathway rather than the entire 'loop' of wire made up of both black and red directions). I think no electrons moving on the red/negative wire while the 'signal' makes it's way around the loop via the black/positive wire might be a valid way to view this, but the data would obviously prove that right or wrong by showing either negative voltage or no change at all when the battery is initially connected to this disconnected circuit. I'd also be interested, if you had the data, whether the changes in voltage on both wires within the disconnected circuit were exactly proportional to the charge sent down the black/positive wire, or only proportional (meaning, did we discover whether the battery 'pull' is equal to it's 'push' of electrons, or even how that force might change as it traverses the wire as less and less electrons belong within the negative/red wire in total like eventually trying to suck air from an airless vacuum vs sucking air from a pressurized or equalized balloon as would be the initial case) Thanks if you see this, my apologies if it's a misunderstanding on my behalf!! :) Music major here hahaha, won't be offended :)
Good question - I did briefly measure this pumping by placing two small resistors near the switch so I could measure the current in both wires, and the “distant” pumping effect was present. I talk about it a little bit in the other follow up video with the experiment details!
You can see this on an old analog monitor if you put a tea on the coax and continue single down coax without a load resistor or a nother monitor you will see the single reflection giving you a double image which will change depending on the length of the coax.
Could you redo this using a multiplexer to avoid manually change the probes? You could first prototype a manually change of probes but pressing a button to change the input of multiplexer. And evolve to an automatic one. Thanks for the very instructive video.
@@AlphaPhoenix2 Ok you're right, as the calculation are relative to the initial signal time, if the length and the constitution of all cables probes are the same it should be ok 😉 (what did you think of my experiment proposal ? 🙂)
((just in case : the experiment proposal in the comment section of the video "Can you replace all of the mobile electrons in a wire? (Back of the Envelope episode 2)" 😇 ))
Finally a UA-camr that actually goes into detail with experimental setup and allows us to conduct our own independent research ❤
He's legitimately brilliant.
This is the kind of thing I wish I'd done as a lab in py208!
Most detail-oriented person the planet. Incredible!
I think every experimental research paper should have a video like this, it would make replication a lot easier
I love how this video demonstrates that proper experimental design relates to just careful thinking and understanding of assumptions and first principles, and does not have anything to do with how fancy the equipment is. The whole experiment used dead bug circuits and a super cheap scope, but was crafted to be able to gain them necessary insights, and it does so with sufficient resolution to prove the points.
THREE AlphaPhoenix videos in one day?! Christmas came early!
unfortunately i will not have a christmas video ready in two weeks lol
For improved accuracy with differential probes, you should have connected the grounds between them!
The second thing, and I think you did do this, but non the less:
It is very important to use the x10 setting on the probe for two reasons: to increase the input impedance of the probes (the thing you tried to do with capacitors), and to enable you to trim the probe's compensating capacitors (little screws on the side) to eliminate ringing.
Trimming is done with a probe attached to the test generator contacts just beneath the display in the middle of your scope.
The problem you saw with the probe disrupting the signal patterns is from the ground connection of the trigger probe attached to the switching circuit.
You really should have added some proper analogue switchings to the circuit, like SN74AC66, and also separate batteries for the oscillator and the driving the line, to mitigate ground coupling through the scope.
I agree the 3 videos should be a mini series or a really long video on your main channel. Stuff this good shouldn’t be resigned to the second channel
While i agree with you that this is an excellent video, it does go into the weeds on some of the EE that your average joe would fine more confusing than the generalizations on the main channel.
This was really cool, please make more EE content. 90% of ECE content on youtube doesnt go past first 2 weeks of Electronics 1 so its really cool to see some applied signals content.
2:49 My EE background says for smoothest power you want a combination of different capacitors. I would use a 10uF electrolytic, a 100nF MLCC, and a 1nF ceramic in parallel.
Each capacitor is most effective in a certain frequency range due to parasitic resistance and inductance. By combining different types and values you can cover the whole frequency range of interest.
I'd never considered that capacitors could have frequency varying impedances. thanks!
@@AlphaPhoenix2 The idea that you need multiple different values or types of decoupling capacitors in the way that stefansynths suggests is a common misconception in electrical engineering. Generally the way you would want to use decoupling capacitors is to place as large a capacitance value with as low a parasitic inductance and (sort of) as high of an esr as you can get away with as close to the power pins of the device as you can in a way that minimizes loop inductance. Generally the lowest parasitic inductance capacitors you can get are those in the smallest packages you can get, so SMD ceramic capacitors. Nowadays these are available with relatively high voltage and high capacitance ratings, so there is generally no need to opt for electrolytic capacitors if you want a 10uF capacitor that can handle 9V. A drawback of certain ceramic capacitor types is that, depending on the dielectric, their capacitance can drop significantly when higher voltages are applied. Another drawback of ceramic capacitors is that they can have too low of an esr. There are controlled esr ceramic capacitors, but they are rarer and more expensive. If needed, you could put a resistor in series with a ceramic capacitor, though that obviously increases area usage, bill of materials and loop inductance. I find this video gives the best explanation about decoupling capacitors Watch code: WdlN8bHw-w0
In fact (and if people are still reading, 5 months on), above a certain frequency (governed by the parasitic parameters, due to the physical construction), capacitors behave as inductors (their complete opposite). They will still store charge as any capacitor, but the voltage vs. current waveform follows that of an inductor. That's why we use multiple capacitors of different values (actually orders of magnitude) and construction types (usually electrolytic and ceramic) and physical sizes to cover the full frequency range.
20:45 "And it's just that easy" is quite the statement! That is a LOT of painstaking work with many mistakes that can be made. Very impressive research and so cool that a reasonable digital oscilloscope can capture this.
The twisted pair is a transmission line. Your rapidly switched pulse is a mix of RF (wide band) and DC. We need to be careful and not equate the DC bulk Resistance of the twisted pair to DC flow, with the Impedance of the twisted pair to an RF wave front. Different critters. The pulse that is impressed and propagated at the velocity factor of the transmission line when the switch fires each time, is primarily a radio wave that rides up on the skin depth of the wire (frequency dependent) and its propagation along the wire is resisted by a combination of reactance and surface resistance, that we call impedance. Along with that is the DC current that is more controlled by resistance and propagates through the bulk of the metal at a different velocity.
This can be shown by a long balanced line made of two, very thin wall, hollow copper pipes. Measure the DC resistance of the line, then measure the impedance and velocity factor of the line at some arbitrary, but stable, radio frequency.
Now, replace the hollow pipe with solid round bar of the same diameter and repeat the measurements. The DC resistance change will be significant while the velocity factor of the balanced line to the pulse will remain close to the original.
All 3 vudeos wefe just amazing!! Seriously!
I would make two minor changes though. For the Schmidt trigger, add a couple.of 0.1uF or even 10nF capacitors.ij parallel to the large ones you have. As close to the vcc/gnd pins of the chip as possible. Rhe big capacitors help but their internal resistance and inductance means that they are "slow". So for the very fast transients you want a fast 0.1uF capacitor bormally known as a "decoupling capacitor". This will give you much better rise times
Swcond improvement would be to use a differential probe inatead of the math function. Thia way you sont have to worry about the disconnected ground leads and the inductance that those leads (pf the red and black probes) create. Remwbwr that all ground ports of the scope are common ao you have a "back" path or at least aome potential interference from the ground leads hanging around.
A differential probe does the math in analog way nect to the measuremen point and ahould give you wven better noise immunity (although the traces did look very goos)
Finally, given you hace a "repeatable" trace.. if you apace put the pulaea enough (say at 10khz) you can use the averaging funxtion of the scope to get better dynamic range /.noise performance by accumulating hundreds of runs of the same repetaed event.
Some scopres (sampling scopes9 can evwn increas the time resolution by adding a slight jitter to the trigger and getvsamples at slightly different times but due to the repetitive bature of the signal, tgey can build a higher resolution (time wise) "Image"
Finally ypu can use the trigger input od the svope instead of one of the channels, directly from the Schmidt trigger output (maybe uaing one of the channelse separately). That way you can sanple using channels 1 and 3 and thus run the scope at dohble the rate. For example of the scipe is 1gigasample/s that usually applies to channels 1 and 3 by themselves. But if you capture all 4 channels.or 1 and 2 (or 3 and 4) then the rate goes down to 500 megasamples per channel
When I watched the main video I thought, wow, he must've had a shitload of probes to take all these measurements at once.
It never occurred to me that you just took a shitload of measurements with three probes.
Thank you for your hard work!
Itis scary how much work and thought went into this. Crazy inspiring work.
Would love to see how you animated the electron motion according to the observed experimental voltages.
Ah, I see you are an aficionado of 3D "just solder it on anywhere" circuit design. Applause, applause. You win 4 internets.
Brian, this is probably the most interesting series of videos I’ve ever seen. No hyperbole here. So many levels of fundamental understanding about electricity, impedance, metrology, being shown and explained in an intuitive way.
One question I was left with: how does current depict itself in the top graph of voltage? Is it the increased area under the curve? The slope of the wave front? I know charge has to flow to change the voltage within the line, but I’m having trouble connecting it to the measured values.
Thanks.
the "fake electrons" in the diagram, and the associated "electron drift speed" plot were based on how much current would need to flow in order to see the observed changes in potential along the wire.
The first bounce is easy: If you assume the capacitance of the cable is the same everywhere, then every additional volt is a proportional change to the electron density (linear density) in the wire. Once you pick an arbitrary (ridiculously massive) capacitance, you can invert the "electrons per length of wire" plot to get a "distance to the next electron" plot at any location along the wire. If you know the electron at the far end of the line isn't moving (for relativistic reasons, don't need to assume anything about electricity), you can use this to iterative draw each "electron dot" spaced back from the end of the wire the correct amount. The really cool thing is that as you play the animation forward, this construction (based completely on a fudged electron-density-to-voltage characteristic, produces motion, because you NEED electrons to shift down the wire slightly to produce these voltage waves.
Subsequent bounces are harder, because then you don't have any stationary electrons to use as a foundation, and you need to grab an arbitrary (actually carefully algorithmically chosen but whatever) and say "this particular fake electron isn't going to change in speed this iteration" and you base everything off of that.
I realized after the fact that I could have based everything off the electron at the supply end and then measured the ACTUAL current using a small resistor (which I did at one point), but I was actually much happier to see that you can get good realistic motion based exclusively on assumptions about what current is necessary to produce a certain voltage distribution.
My FIRST attempt was actually a full simulation - I dumped a bunch of electrons in the wire and let them be accelerated by the field, but unfortunately the scope traces at different distances were SLIGHTLY offset from each other, and that meant that while a human could identify the waves easily, the algorithmic electrons were bunching up as they got attracted to all the slightly-elevated trace locations and it was a nightmare.
Thanks. Took me a bit to work through, but I think I got your approach:
Represent voltage i.e. relative charge density as the direct inverse of distance between the charges.
Back-calculate the position of each charge relative to the stationary reference.
Add the incremental movement of the reference to all charges.
I puzzled through my question again, and this is my long-winded logic on current in the voltage graph:
Consider arbitrary positive charges, zero net charge initially, and a perfect step input wave.
When the “voltage wave” travels through the length of wire, the net charge in the wire increases.
This increase is proportional to the voltage(charge density) and the length of wire(capacitance) under/behind the wave.
The speed of the wave determines the amount of charge moving through the starting point per unit time.
Charge movement is current.
So the voltage integrated over the length of wire is total charge, and the change of this integral over time is current.
If we move the virtual “current sensor” to somewhere along the wire, we can just integrate starting at that point.
Of course, this breaks apart as soon as there’s a current coming in from the end.
You deserve a Nobel price just for this video, really !
Love your posting of the extra stuff and outcuts here! Thank you for all the hard work!
AlphaPhoenix: "We take thousands of measurements so you don't have to." Mind blowing project. Amazing attention to detail. And a result that changed my understanding of electricity. I'm subscribing... and continued success!
To switch the Hex schmitt trigger inverters quickly you need to supply current to the IC over a large frequency band (Fourier analysis), which means you have to have instantaneous current flow at high frequencies to minimize the signal rise and fall times, meaning smaller value supply capacitors (higher self resonance) as you get closer to the IC power pins while minimizing the device lead inductance (lead length). But, you will quickly be limited by the IC package internal lead inductance. You will have several capacitors in parallel of different values and types all working together to meet the high frequency IC switching current demand. The aluminum electrolytic capacitors in the video have much too small a self resonance frequency to supply the needed current to switch the IC quickly, and the bulk capacitance is largely negated by high inductance of the long lead lengths.
Your youtube channel(s) helped me understand more about electricity and it's properties more than any other content on this platform. Thank you! ❤
These three videos have been the most fantastic learning experience I have had in a very long time. I have for years always wondered about this exact same thing and now I feel like a window has been put in a place where there was once a wall. Thank you so much for your work!!!
This is a great work of engineering AND teaching. This is rare, not only in the electricity field. It deserved some sort of academic recognition
This project is a masterpiece. Well done. Thanks for showing the setup as well, this answered all my questions, along with all the questions I didn't ask but should have
Regarding using the electrolytic caps to bypass the schmitt trigger's power... the common practice is to add a ceramic cap in parallel with the electrolytics. The ceramics are much faster to respond to any changes. As for the differential probing of the twisted pairs... the probes already have a much higher impedance than the twisted pair. Blocking the DC with the series caps doesn't seem like it would improve things. Differential probing is required, of course, because you want the voltage difference between the red and black wires, and that black wire is definitely not at the scope's ground potential. I really enjoyed the part where you explained the need to get data at all of the points and animate it, and then said "it's just that easy". 🙂
good to know! the vast majority of my electronics experience is with very slow things so electrolytics have been good. maybe those weren't doing anything at all here lol
@@AlphaPhoenix2 Doing this sort of high frequency electronics work has a variety of pitfalls. You might want to browse a book titled High Speed Digital Design by Howard W. Johnson, ISBN 0-13-395724-1. There are techniques to improve the scope probe performance too. This video has a quick explanation of why you can't use the scope probe ground lead for high frequency measurements.... ua-cam.com/video/henwiLzSieA/v-deo.html
Yessss! I was wondering exactly what this video answes. You're the best, dude.
If you want perfect equal current on pluss and minus cable, maybe just use a small high speed transformer as your supply.
Current should be equal on both sides in my mind 😅
Thank you for this detail, brilliant set of videos.
(The description's link to your main channel has a funny character at the end, doesn't work without editing target)
This channel is about to take off big time. Brilliant video!
Also I did similar experiments when I got my first DSO.
My mind was blown away...
Very nice explanation about how you made the measurements and processed the data.
To be honest, I thought you were pulling it via USB directly from scope, but saving the CSV is a very good and simple approach. I never thought about CSV saving because my scope is damn slow there.
can't get enough of this series. And the work you put in to make those measurements... sheesh!!
Thank you for this video! I was just about to ask you multiple questions that this video answered. Especially the excel/matlab how-tos. Subbed!
I have the exact same model scope so this video was especially cool to see
After seeing probably that exact video about the schmitt trigger TDR a few months ago, I've built one aswell and used it to meassure the reflection in a few meters of coax. Allthough I was mostly trying to meassure the speed of light in said coax, aswell as learning about transmission lines in general. These 3 videos you made really halped me understand the concept of transmission lines a lot better, especially where that magical impedance value comes from and why reflections occur. I'll also definitley attempt some of the experiments you demonstrated with twisted pair transmission lines, especially using multiple probes to cancel out the crosstalk. Really love your videos and looking forward to more of them :)
I also just so happen to have the exact same oscilloscope you do, it's really a lot of oscilloscope for a humane price
Could you please link that TDR video?
ua-cam.com/video/9cP6w2odGUc/v-deo.html
"Focus you fu...!" Ups, that's from another channel. Absolutely love everything that you do. Respect.
Could you put an instrumentation amp (powered by a coin cell, perhaps) at your measurement head, so as to have a lower-capacitance input? Also, with an instrumentation amp, you can drive the shields of the short cables actively to the mean of the input signals so as to minimize the effective capacitance. Center-tap the gain-setting resistor and run it through a unity-gain buffer.
Failing that, they do make x100 scope probes so that you can have even lower input capacitance.
In your pulser, you probably don't need those big electrolytics. The battery is effectively a huge electrolytic capacitor. You need a small amount of capacitance with very low inductance right by the chip. With the dead-bug setup you have, just solder a 100 nF ceramic between pins 7 and 14 of that 74AC14.
I suspect you'd do well making your pulser around a line driver like the AD8132. Impedance matching when you launch the pulse does help!
I saw you with a bunch of termination resistors. How close did the measured characteristic impedance of the line come to your calculations?
Now I'm sitting here with a head full of imaginary Smith charts and wondering about whether I could design a current transformer to measure things less invasively. (But I haven't designed any magnetics in years, and I'd be really, really rusty at this point.)
Great video! Love the insight into the process.
This is an outstanding set of videos. Thank you so much for this! 😊
Ok, that was awesome. Love the detailed explanation. If you could post the processing scripts and raw data for people to check their workflow, that would be extra (CC/OpenSource would be great, but could make it a Patreon perk).
Just wow! I can only imagine how much work it took you to make all of these. For electrical engineers the results are pretty easy to understand - the long wire is inductance and capacitance "distributed" along the wire length.
That set of videos should be shown at schools, and mandatory at trade schools
Верное замечание👆👍👍👍
Great work on that visualization! I think your setup would greatly benefit from using coax instead of twisted pair, and from using BNC T-adapters as monitoring points (even though you've stated your motivation for a custom-made lines at the end of the video) - see my comment under your main video for more detail. By the way a great learning source about transmission line transients is a book by G.A. Mesyats called "Pulsed power", more specifically, chapter 2.
From the first few sentences, my thought was: "A true scientist!"
I will try all in my power to repeat this experiment.
THIS is the content i love the most
Fantastic work! Thanks so much for sharing the work behind the scenes and the thought processes that went into this fascinating work.
Finally, an example where 3D graphs are perfectly sensible
one of the VERY few. somebody should tell Nature editors
Could you do this same sort of thing with an antenna? Could you use that to see how and where it's picking up a signal? Would a received signal cancel itself out if it's wave is bounced back at cosign of the received wave? That why some antenna's like old TV antenna's have a branching fern leaf pattern to provide some provide some constructive interference at the main branch?
Lastly does it move faster as it's an external force or the same because of how electrons move in the wire? That being the most like Veritasium example and probably the more interesting one to look at in relation to it. I'm sure it's just oscillating or resonating not actually moving through it like this though and that would be an interesting topic as well.
That answered my concerns with how the experiment was set up. Awesome! Thank you for doing such an amazing job at scientific experimentation.
That was way cool, thanks for walking us through the deatild setup!
amazing videos, I always thought about showing this kind of experiment in a transmission line lesson and you just did it fantastically
The yellow trace is not due to crosstalk.
The chemical battery is not -5V and an infinite ground. It is a redox reaction, pushing electrons on one side and pulling the on the other. Half the potential is on each end. So you will see symmetric behavior in the positive and negative wires. Both of them are getting a wave of current due to the chemical action at that electrode.
Sincerely, extraordinary experiment presentation and its results and unique visualizations. Congratulations!
You, Sir, did an incredible amount of work creating these videos. The thought, the setup, the programming, and so much more - boggles both the mind and the body. Extraordinarily impressive!
Great video, excellent experiment, but... nice watch dude!
Terrific trilogy.
IMO only two aspects are missing:
1) it need to be said that there is a voltage drop in your battery corresponding to the load
2) why does the voltage fixate at least temporarily after the first wave propagated? It is not obvious. On the one hand I can say that it stays same because each new section of low resistance wire looks exactly the same regardless of how many is left but that's not enough.
Fantastic work!
17:00 also you could just place the scope probe close to the wire without making any contact and it would be very similar in nature but less repeatable
I tried this (just clamping the insulation) but my probes had too much internal loading (maybe because I didn't tie the grounds together, as someone commented), but I was also worried about different amounts of wire squish altering the capacitor between the probe and the wire
Loved the animations great job
I love your experiments ! Great video ! By the way, why not using coaxial cable ?
that trigger circuit.. I love the idea, and you nailed this whole video (taught me some stuff; thanks!)
BUT
I know it's faster to just solder it together.. I get that traces (w/o a proper matched "ground" line) just get more complex as freq goes up.
Still.. Dude.. there has to be a better way.. looking at it as is makes me cringe..
Maybe there's a pigment/insulating spray glue that's totally inert, easy to razor-off (like to measure etc), and simple to use..???
IMHO, you gotta just make a proper one.. the idea of a proper trigger/clk for your scope feels right what with all the research you do.. no?
Oh.. and thanks a bunch for this series!! It's really helping me find the tiniest of holes in my foundation! very cool
This was surprisingly riveting!
Very impressive! Good job!
Really nice explanation of your materials and methods. Love this video! It made me appreciate impedance more.
Actually, I don't want to do this myself. Even though I have almost everything I would need.
I am big fan of backseat science, and you left me wholly satisfied doing so much diligence I would have wanted to put in myself.
This is amazing! Thank you for your effort! ♥️
To anyone who wants to recreate this experiment, it will produce very poor results if you try to use a single-ended scope probe to measure a differential signal on a twisted pair by connecting the probe's ground lead (black aligator clip) to the second wire as a signal reference. That works when you're measuring single-ended signals on a circuit board with a common ground plane, but in a balanced transmission line, you need to make a differential measurement (like he is doing) using two 'scope probes (one on each side) and use the oscilloscope's "A-B" math function.
Using two probes on each wire of a balanced pair allows both sides to be measured like a signal (because both sides *are* signals), and each side gets the same 'scope probe loading (if both probes are matched) and it doesn't unbalance the transmission line (creating different circuit conditions that alter the signal). Note, both scope probes need to be the same cable length and same probe type and should also be properly compensated for high frequency peaking with a square wave calibrator (the little screw adjustment on the probe box squares up the edges).
If you try to connect scope probe ground clips as a signal reference, it will change the topology of what you're measuring because it alters the signal return paths of differential transmission lines by connecting them to the oscilloscope ground. It also connects them to AC power safety ground, if your oscilloscope is powered from AC.
Using a battery powered 'scope can eliminate the AC ground connection, but if more than one probe is used, the scope probe ground references still get electrically tied together at the 'scope chassis.
So, he uses two 10:1 probes to make each measurement, and this isolates the oscilloscope ground from the circuit under test, and it doesn't alter the transmission line's signal returns, nor does it unbalance the transmission line.
Note, when doing differential measurements, you should still connect the scope probe ground clips to something that is common mode to your circuit. Ideally, they should be tied closely together and connected to the ground of your signal source. He is letting the oscilloscope tie the grounds together and using the ground reference of the trigger channel to tie it to the signal source ground.
In short, using differential measurements are necessary on balanced pairs of wires.
I am amazed that you are willing to put in this much work. You must be loving it and you must be doing it because of that.
As others have mentioned the trick to not introduce any big change by the testing equipment is to use very high impedance probes.
Unfortunately many of your viewers of the actual test video do not realize that at those short time scales as you are using there is really nothing that can be called DC. They look at it from a DC point of view instead of from an AC point of view which I believe makes it more difficult for them to understand it. I have for a long time known that your results are the fact of what happens (and picked c) but I have never seen it so thank you. I have often seen the results of it in data communication and antenna cable performance.
I find it really difficult to communicate that short pulses are AC. in one video i went into detail (on a very similar circuit) and showed an FFT and everything and I had irate comments telling me that "pulsed DC" was different than "AC". whatcha gonna do?
@@AlphaPhoenix2 I appreciate your frustration. I must admit that I myself have changed my view on it over the last few years and that is after an electronic engineering education that I had about 45 years ago where I did learn that every pulse can be calculated as made up out of sine waves.
At first glance one would of cause say that DC is when the current is just going one way and AC is when it is going forward and backward (alternating). That is a fine if we are talking about what most people see and experience most times. It is only when one get's into working or thinking about what happens on a time scale of less than a second (especially milli, micro and nano seconds) that it really matters. Personally I have worked with that with computers etc. since I finished my engineering education.
Now after I have retired I have started to analyse things a little better as I also have some interests in physics and then I realize that the little bastard electrons are what is moving charges around and they or their influence can not move faster than the speed of light which compared to frequencies isn't that fast after all.
To explain that a pulse is AC the one thing I can think of right now is to explain how such a pulse can be made up of higher frequency harmonics of the pulse and that must be the case in real life too as if all higher frequencies are filtered out a pulsed DC will only rise and fall as a sine wave of the pulse frequency. It also demands to explain that all circuits consist of inductors, resistors and capacitors. The functioning at super cooled circuits where the resistance may be zero I am not quite sure of now. I am still learning at age 77.
A place where a pulsating wave is used to make a higher frequency - a harmonics - is in many high frequency radio transmitters. They often start up generating a lower frequency that is easier to control , then cut the sine wave's top and bottom to resemble a square wave and with filtration picking out the desired frequency. That was used in a TAXI radio I modified to use as a Ham radio transmitter many years ago.
An AC pulse can of cause also be said to be a DC with an overlaying AC but the initial start when a DC is connected to a circuit can in my opinion always be seen as part of an AC wave.
Unfortunately there are people that are so sure of them self's and what they have learned that they have to be unpleasant in their remarks. I know that although I have learned and experienced a lot then I really don't know very much for sure and if people can disprove what I am saying then I am willing to change my mind if it makes more sense to me after their explanation. To be simply told that I am wrong without an explanation doesn't cut it.
Man, standing ovation! 👏👏👏
Thanks for the explanation of why you twisted your own wire - I'd also leapt to "can't you just use cat-5 or something", but needing the linear resistance make a lot of sense.
It would have been interesting if you had discussed what defines the impedance of the wire. It seems not to be only the ohmic resistance.
This is absolutely mind melting 🤯
Great video !
I'm really glad , this video Explains and shows that you are aware of the many things
And issues that 'may' influence the readings etc
It Takes away the miriad of questions I had as to the accuracy and reliability of the setup used .
So we can now view the Readings in perspective relative to all the other factors involved 🧐
of course there are still a couple of things
I wonder about ,
But that is always where the next experiment comes from !
Experiment to clarify , demystify and fully understand that what we see and already innately understand by gut feeling
Wow, wow, wow. Pure genius!
So Capacitors are shock absorbers? What's the delay of that? Reminds me of sound waves in baffles. Eventually leakage occurs and how you control that his how you transmit the energy. Entropy is such a good model for these reasons. More holistic.
Something I find cool is using microscopic baffles with phonons to direct thermal energy where it is more conductive to escape and not reverberate in its container.
Incredible work thank you
This is fantastic - I wonder what will the dynamics be for a electric motor where back EMF plays a role (AC circuit)
I love your videos man!
Are you coming to the FLL world finale in Houston in May?
if you made a cube of up/down twisted pair, you could probably make your maze and measure and animate THAT. tho I would hate to have to do the data collection for that without automation.
Why not put this on your main channel? It's just as interesting as the main videos!
Thinking about how twisted pairs naturally reduce EMI (the reason they are used in telephone and Ethernet cables). It would be interesting to see this same setup but not twisting the pair, just straight wires touching each other. How much worse would the EMI be? Would probably require some home built contraption to put a wrap around the pair to maintain even contact between them. The thinnest available speaker wire maybe? Thermocouple wire is often made this way, but it's pricey.
Oh you are using LDO, I was very confused about the voltage scale in the previous 2 videos.
I believe the reflection characteristics are going to be different depending on the frequency ur switching, but ur vid shows the general concept. Could be cool to show an impedance mismatch and it's the reflections in this style of vid
It would be really cool if you could somehow extend this to the flow of electrons in the battery itself. I would love to see a visual of the forces and flow enacted upon the electrons inside the battery.
since you AC coupled your probes at the tips how did you get steady state measurements? I thought everything was detailed and explained well and done great but I was a little confused on that part. A series cap on the tip of the probes should only allow for changes in the circuit to be measured and not the actual DC voltage right?
I think you do an excellent job with these series of videos about circuits.
Just a question... Would setting the scope to AC coupling instead of DC coupling have the same (or similar) effect as what you did with the capacitors in series with the probe?
Or would that make it hard to keep the readings stable over multiple readings at different positions on the wire?
So then it's possible to have any combination of impedance and resistance?
Love the content it's so cool! An editing note: your narration audio is pretty quiet. Right click video, stats for nerds, volume normalized: -14db. That should read 0 for proper loudness.
I am glad that you have done these experiments to show aspects of the nature of electricity that even just 10 years ago were mostly theoretical.
I learned this very stuff in school 25 years ago and it was very old then. This isn't new, but he did a very good job of showing it.
Super interesting, thank you for going to all of this effort watching those plots slosh around is incredible!
One thought though, you're using singled ended probes, yes you've taken two of them and used a math function to make them look differential but ultimately each measurement is single ended. The effect of that is that each has some input impedance to it's reference, a reference they share. In most cases this is like 10MR or so. The problem I wonder about is that because of the above you are effectively connecting the start of the cable run to the point you're measuring together through the scope. With around about 15MR, thas isn't much of course. But by a similar theory of the electrons "predicting" currents until proven wrong later by hitting the load, surely you are getting currents flowing through the scope, bypassing the cables you intend to measure? I'm sure the effect would be minor regardless. But one way to move closer to solving this would be an active differential probe. Better yet, two, one for each measurement point.
I think ideally you'd have a scope with isolated channels, but you'd still want a differential probe to get the purest measurements
Instead of capacitors, could you use a hall effect sensor to measure without interfering?
What a damn good video man!
Would you be interested in a PCB with test points and possibly the pulse circuit built onto it? I could send it your way and have it done in a few days. the trace spacing can be as close as .15ish mm. could save you some space! I imagine the coupling would be worse though!
So for a circuit, there has to be a positive and a negative wire, right (heads up, might use positive/red/negative/black incorrectly, sorry...trying to use 'Positive' to mean electrons flowing INTO the wire, and negative to mean electrons flowing back to the battery): I'm a bit confused with the control wire to control cross-talk...I imagined the three 'looms' (just my nickname for them) each being one prong of the Y configuration. If this were the case, then both the positive and negative wires would have to be on each loom wrapped together, like a highway wraps two directions together within one string of highway....each loom being one 'string' of 'highway' containing both directions on the Y arrangement (thus the three looms, one for the beginning and a junction seperating that for both a looping string of 'highway' loom and a last string of highway that doesn't return/isn't connected). Is the 'control' wire just your negative wire? Or did I completely misunderstand the arrangement of the 'looms' and the Y prong configuration?
Basically...would 20 probes on the black wire measure equal changes as 20 additional and separate but equally spaced probes on the red wire at the instant the battery is connected? Aren't we doing something differently, measuring them together via a single probe in your experiment? Who's to say electrons even move in both of the wires when the initial charge dives into the wire, couldn't that just be one of the wires while the other doesn't experience anything (unless the red and black are soldered together, in which case it does eventually make it around)?
@AlphaPhoenix @AlphaPhoenixChannel Like, specific points on both the red and black wires seem like they are eventually graphed combined as a single point...why not separately? Did you have data that showed for certain that, for each length along the wire, changes happen simultaneously and equally to the positive and negative wires? For instance, I didn't understand how you can say/show (18:22 on original video) the red wire (negative?) wire ending up with fewer electrons than it started while also showing a graph that only seems to be demonstrating what's happening on the positive/black wire, when I can't really tell how you separated those measurements from "different sides of the highway at the same point" if at all. It might be an assumption, but I'm not sure how we could say (on a disconnected loop) the red/negative wire ends up with less electrons if we were only measuring points on the black wire (the wire we are 'pushing' electrons into).
I'm wondering if when you say 'electrons slam into the end of the wire', I think we have to assume you mean they slam into the end of the black/positive wire only. But I'm more interested in the data...did you have data show what happened on the red wire when this experiment at 18:22 on original video was performed (as in, what we would assume to be the opposite of slamming into the end of a wire). If electrons do leave the red/negative wire, does that mean by pumping electricity into object A, we can use that to inherently pull out electrons of object B even without a circuit being created? I'm just concerned there might've been an assumption (or data was only shown for the black wire/half of the pathway rather than the entire 'loop' of wire made up of both black and red directions). I think no electrons moving on the red/negative wire while the 'signal' makes it's way around the loop via the black/positive wire might be a valid way to view this, but the data would obviously prove that right or wrong by showing either negative voltage or no change at all when the battery is initially connected to this disconnected circuit.
I'd also be interested, if you had the data, whether the changes in voltage on both wires within the disconnected circuit were exactly proportional to the charge sent down the black/positive wire, or only proportional (meaning, did we discover whether the battery 'pull' is equal to it's 'push' of electrons, or even how that force might change as it traverses the wire as less and less electrons belong within the negative/red wire in total like eventually trying to suck air from an airless vacuum vs sucking air from a pressurized or equalized balloon as would be the initial case)
Thanks if you see this, my apologies if it's a misunderstanding on my behalf!! :) Music major here hahaha, won't be offended :)
Good question - I did briefly measure this pumping by placing two small resistors near the switch so I could measure the current in both wires, and the “distant” pumping effect was present. I talk about it a little bit in the other follow up video with the experiment details!
Amazing series of video, what software did you use for the animations?
Looks like MATLAB
yup@@ebrewste
You can see this on an old analog monitor if you put a tea on the coax and continue single down coax without a load resistor or a nother monitor you will see the single reflection giving you a double image which will change depending on the length of the coax.
Could you redo this using a multiplexer to avoid manually change the probes? You could first prototype a manually change of probes but pressing a button to change the input of multiplexer. And evolve to an automatic one.
Thanks for the very instructive video.
This setup reminded me of how you can emulate slow motion camera by using a stroboscope.
Is the time the signal goes into the oscilloscope from the probe taken in account ? (Subtracted to get the real time it got in the wire ?)
dont know how long, should be the same for all
@@AlphaPhoenix2 Ok you're right, as the calculation are relative to the initial signal time, if the length and the constitution of all cables probes are the same it should be ok 😉
(what did you think of my experiment proposal ? 🙂)
((just in case : the experiment proposal in the comment section of the video "Can you replace all of the mobile electrons in a wire? (Back of the Envelope episode 2)" 😇 ))
it would be very interesting to test this with an ac power supply and a tuned wire or LC circuit