As an RF Engineer, I live by this principle. It was really cool to see you approach it from the DC perspective and I am very impressed by your painstaking oscilloscope measurements!!! I have only ever seen plots like this in simulations, never with true measurements.
You can also view terminating resistor as substitute for the rest of infinite wire. This way it's easier to understand why there is no reflected wave in this case
@@AlphaPhoenix2 The Design of CMOS Radio-Frequency Integrated Circuits has a very amusing chapter on transmission lines containing the gem: > We've just seen that an infinite ladder network of infinitesimally small inductors and capacitors has a purely real input impedance over an infinite bandwidth. Although structures that are infinitely long are somewhat inconvenient to realize, we can always terminate a finite length of line in its characteristic impedance. Energy, being relatively easy to fool, cannot distinguish between real transmission line and a resistor equal to the characteristic impedance.
I learned the same thing, but 20 years ago... Anyway what buggs me about the original verstasium post was the original claim that EE didnt learn this. And.. well we bety much do. Impedance matchning is like the core of high frequency circit design. Really with out it most modern radio systems would not work at all, like a wifi circut
@@Jefferson-ly5qe yea,, i don´t know exactly where the cut of is, we did basically all the labs in 2.5 and 5.1Ghz and then when i worked with it it was basically minimum 800Mhz ... Ironically no i work with transport system and the highest frequency we use is 400Hz (note, Hz, not kHz or MHz), typically 8000V and around 1000A. The people that did the system in the 1970 basically just made a gigantic coaxial cable and bent it to form. This turned out to be incredibly expensive. Of cause we talk about 100 of km of cable, so that is millions and millions. Some of the AC we could simply eliminate but some of it need to be around, So we basically try to balanced stamped metal sheets. Resistance of the old system was also a bit to high. Makes it a bit more complicated by the load of the vehicle change the induction. Still its not really that complicated compare to like a 4G muti frequency base station that i worked with prior. its really just a different kind of complication because everything is huge.
this is not deep inside. It is an explanation made by a man that simply doen’t have enough technical skills (Ohm’s law , resistance definition for example)
The fact that more or less consumer grade scopes can actually capture this is utterly brilliant. And then all the graph visualisations, especially the one with the 8 or so different resistance levels and see how impedance matching magically works... very very cool! And fairly important with everything with antenna's, especially high powered ones that will reflect significantly if not correctly matched, blowing up amplifiers in the process.
All hail the Standing Wave Ratio meter & the 50 Ohm dummy load. I never expected to see hard data showing these waves of electrons as clearly as the sea reflecting waves from a harbour wall. This has to be one of the most revelatory data sets I've ever come across. Dat here in appreciative, slight stunned silence as things I've used in RF & audio all fall into place!
Coming from the days where we would have to allow the scope to warm up for an hour before use, seeing the phrase "consumer grade scopes" blows my mind. But yeah, I guess that's where we are at. :)
@@trevorus I have one as well. I bought it for a project that I was working on a couple of years ago. Supposedly it can be modded to do light duty as a spectrum analyzer.
I don't often write comments, but WOW this was awesome. This, combined with the last video somehow got me to see an intense beauty in this "low-level", more advanced type of electrical properties. Thank you so much for this
It was really interesting to see the dynamical demonstrations of pulses propagating and hitting the different loads! I didn't expect the real world waveforms to be so close to the ones you see in textbooks. Bravo to you!
As someone working in testing motors for EVs, this video and the main channel one were GOLD! The Portuguese PhD who tries to educate me at work will love this, and hopefully i can understand more about EMC!
Let me tell you that this just made me make sense of the DMX line impedance loads used to avoid "reflections" in the line.... I knew about this ripples or waves of "voltage" but I couldn't figure it out why a load of an "apparent impedance" could work to "destroy" them.... This is just amazing.... Thank you.
If you use the electron motion visualization again in future, adding colour based on speed may make it a easier to see the changes in velocity, so blue is slow, red is fast, and rainbow or gradient between. Also, this video should be required watching for every EE student, it explains and shows impendance 10x better than anything I got in school.
As a ham radio operator and long time mechanic and experimenter you have just explained something I could not see in my mind for 50 years. Now that I have a visual, it is unbelievably simple. I was like you about the coax cable. I never could see how a 15 foot piece of 50 ohm could be the same as 200 feet. That really bugged the piss out of me. Thanks for the video.
To have an idea of where the 150 ohm characteristic impedance of the cable is coming from, you can think of the wire having some inductivity and that there is some capacitance between the two twisted wires in the cable. If you would compute the series inductance and parallel capacitance, the resulting impedance would be 150 ohm. Of course, when modeling this, you would split the cable into an infinite number of sections and each section would have its inductance and capacitance resulting in the characteristic impedance of 150 ohms. You can thing of the wave propagation as the transfer of energy between those inductors and capacitors in the infinitesimal segments, each having a 150 ohm characteristic impedance. This also helps to think what happens when the wave leaves the last segment of the wire and hits a short or open.
To understand that, I would have to know what inductivity and capacitance was but yes, it kind of makes sense. I'm still not sure how you can measure both impedence and resistance in Ohms. If they have the same units, shouldn't they be the same quantity? I guess this is like how "power consumption" is measured in "Watt-hours" and "energy" is measured in "Joules"?
@@aspzx You would measure them in Henry (Ohm * sec) for the inductance and Farad ( C / V ) ...I hope I still remember correctly, for the capacitance. (where C is Coulomb and V is volt)
@@aspzx So both resistance and impedance are the same ‘thing’, except that resistance in DC is a special case. The formulas for impedance of capacitors and inductors are complex (as in have a real and imaginary part) that depend on frequency, and at DC what happens is that the impedance of an inductor goes to zero and the impedance of the capacitor goes to infinity (which is the same as an open circuit), so once it reaches that steady DC state you can basically assume that both the inductance and the capacitance that is inherent in the wires doesn’t exist anymore, and all you see left over is the resistance. So impedance is the same thing, but the dynamic nature of the signal (in the case in the video it’s a transient, the voltage changing in a ‘step’ from zero up to the voltage of the battery when it’s switched on) is bringing out extra elements that you just can’t see at DC.
@DrenImeraj And the consequence of this is that the "width" of the traveling step-pulse wave front imposes a limit on the practical bandwidth of the wire?
The wire also has Reactance ie its "resistance" to the flow of AC current as in AC the magnetic field is rising and falling all the time affecting the flow of the electrons.
When I was a teen, I had a book about electronics by one G.M Scroggie. He described "transmission lines" in terms of a series of inductors in series with the two sides of the line, with capacitors across the two lines between each inductor. Basically a whole line of series/parallel tuned circuits which has an impedance related to the impedance of the capacitors and inductors. I didn't understand why, AT ALL. Now, thanks to THIS VIDEO, over 40 years later, I understand enough to SEE why you show it as tuned circuits. Not only that, but I know why the inductors are in series with the line, and the capacitors between the lines. I also see how the impedance works. I now understand why a CB Radio enthusiast needs a "Standing Wave Ratio" meter to ensure reflections are minimised at the aerial. This one series of videos has BLOWN MY MIND and I love it.
At D.C., yes, it reaches an equilibrium quickly but the point of impedance matching is at high frequencies where peaks and dips of frequency response will occur if impedances aren't matched. I find it very interesting and enlightening seeing these high speed captures you are doing in understanding these phenomena. It makes it so much easier to see what is happening. It is distributed capacitance and inductance which creates the 150 Ohm (or whatever impedance, it depends on the cable configuration / geometry) and the distributed resistance which determines the Q (usually negligible), but the basis of why this happens is shown in the transient responses you are demonstrating. This is significant and will be appreciated by teachers in time. Please keep up the good work!
Now it makes sense, why I had to terminate a Modbus/RS485 wire pair bus with a 120 Ohm resistor, exactly to prevent signal reflection on the line. Thank you so much for all your videos, always looking forward to them!! 😊
One on each end I hope. :) BTW, replacing the 120 Ω resistor with two 60 Ω resistors in series and connecting a "Properly sized" capacitor from the center point of the two resistors to ground will give you about a 3db noise reduction. See Jan Axelson's "Serial Port Complete" page 123 starting with "Terminations for Short Lines" for details.
Yes it's starting to make some sense ! VERY imformative videos !! I'm battling another BUS system at Work, a Carlo Cavazzi Dupline system. (a type of large field automation system) The signal is a Square wave with about 7 volt RMS and at 132 mS there are 128 waves, with different pulse width. The transmission line is branched in several branches of different length, in total probably over 50 branches and combined, 2000+ meters of wire. The system is generally used without any Termination at open ends, but we have lots of communication problems, which I attribute to open end cable reflections. I wish someone at Carlo Gavazzi knew 5% of what others here know, so they could answer my questions 😐
@@daniel635biturbo Now that sounds like a fun one! :) While adding impedance matching termination resistors should help out, it appears that your network is already below voltage. Adding termination resistors will increase the load on your network. Okay, so the system runs at 1khz, and the choice of cable is up to you. Still, line terminations should help, but it would be dependent on what cable you used. Sounds like your voltage is quite a bit low (7 volts vs 8.2). Possibly a node is dragging the network down. Since your system seems to be overloaded already, I would hold off on line terminations. You should have a "Stiff" voltage source before adding any more load. At the data rates that the system runs at, a DC meter should suffice to (only) check voltage. But a scope would be pretty much mandatory to see what is really going on. And I see what you are saying with the 128 "waves", that corresponds to the 16 x 8 matrix of device addresses with a sync pulse at the start. So the entire I/O "matrix" updates at 1hz? Looks like someone figured out how to automate Morse code. What will they think of next? :) My first recommendation, the easiest to install, and the one likely to have the best bang for the buck, is to chop the line in half and put a repeater in the middle. That will probably fix you up without having to do anything else. And it's not like you are delivering broadband speeds to the other end of the network. :) Then you could measure each individual network and see if one is now at the magical 8.2 volts while the other is still at 7 volts. If this is the case, you probably have a module (node) dragging the buss down. And a scope will show if a node is at a different DC bias (the pulse will jump up or down for that node). So is this for building automation, mining, metering, or?... And finally PLCS.net is your friend, there are a handful of guys there that have used it. www.plctalk.net/qanda/showthread.php?t=11592&highlight=carlo+gavazzi+dupline
@@MrWaalkman Big thanks for your response, you probably now know more than me about how it works. The RMS voltage is dependent of the duty cycle, so when I look at it in the ocilloscope I get probably 8,2 volt peak, have not really checked. The system dates back to the 1990 in our factory, and are installed with non twisted 1,5mm2 homogenous copper leads, without any shielding. Back in the 1980s Electromatic developed this in Denmark, but now the system is sold and manufactured in Italy. And they seem to have lost A LOT of knowledge, my Swedish support can't really support my questions. Earlier in the 1980s the installation recommendations were "free" do as you like, but now they ask why we don't have twisted pairs, or shielded cables.... Now, I can't really put my finger on what the problem really is, but I can see cable reflections on the waveform, and It's different depending where I measure in the system. The problem presents itself with "Ghost signals" so the controller sets some inputs as TRUE for one communication cycle. And that is a real problem handling in the "PLC" code, If I can't trust the signals, and the filtering methods are very clunky to say the least. (0,5 seconds+one cycle) The controller should be able to handle 450mA load, and we are only at 30mA on the system that give me trouble. After lots of research I found out that they sell "Termination units" I believe that it is a resistor and a capacitor and possibly a diode. But have not cracked one open yet, this module is also called Impedance module, wonder why 😊 They recommend installing one (DT02) at 1200m from the controller in one cable end, but not several, as it "decreases possible transmission distances" The termination unit don't seem to add any significant load, and the wave looks better, with one installed. What I also find is that most "Ghost signals" occur during daytime, when the factory is in production, so the cable reflections are not solely the problem. But perhaps the cable reflections make the system more sensitive to other signal noise, that occurs during daytime production. Anyhow.... The Plctalk site seems down now, but I will take a look later, good tip !
This doesn't just apply to long wire transmission lines. Changes in impedance on PCB traces or lines connected to them creates reflections and that's generally where noise comes from. Robert Feranec has a bunch of videos talking to a signal integrity expert, Eric Bogatin, and he basically explains everything you did in these videos but in the context of PCB design.
Rick Hartley did a similar thing on Altium's channel. Video name "How to Achieve Proper Grounding - Rick Hartley" He ranked EMI issues above signal integrity. He explains it better than i do.
@@kazuviking I try and overclock SPI on an atsame51 chip so I can push frames to a ~115KB SPI display as fast as possible. I think the limit is supposed to be 18MHz according to the datasheet. If I add 33 ohm series resistors on the cheapest china 2 layer boards I can manage to get 24MHz before things get corrupted. That's just the first value I tested, on I think a 12 mil trace. It would be nice to go faster, and maybe even pass EMI testing even though I'll never sell anything, but I have no idea how I'd even begin to figure the characteristics of the out of spec peripheral. I'm going to have to check out this videos, thanks dudes!
PCB traces are still modeled as transmission lines they just have a different geometry. PCB tend to be a strip, microstrip, and coplanar *TEM transmission lines* Coaxial line Two-wire line Parallel plate line Strip line Microstrip line Coplanar waveguide *high-order transmission lines* Rectangular waveguide optical fiber
Thank you! I am 53, have been interested in antennas and electronics my whole life, and never really understood cable impedance properly until now. One of the best explainer videos ive ever seen!
The data collection and subsequent visualization / animation here was incredible! This is remarkable to see how the measurements really demonstrated exactly what you were describing. Bravo for what I am sure was a significant undertaking in capturing all of this data!
Never seen this much honesty in anything related to electricity. My heartfelt thanks and prayers to universe that every EE in the world follows you. I hate just assuming things because somebody says so with oomph and authority. You deserve 8 million subscribers not just 8K. Just keep it basic and honest. This content is like watching Faraday himself at work. So innocent and humble.
I love the oscilloscope/line plots! I've always understood transmission impedance as the infinite ladder of inductors and capacitors, but never seen it measured/plotted like this -- very helpful!
Thanks! I “learned” about impedance back in 1976. But I never truly grokked it until now. Your incredible patience taking all those behind the scenes oscilloscope readings really paid off.
This is very visual. Thanks for doing this incredible tedious and crazy task of measuring! You can simulate transmission lines without dirt effects in ltspice or qspice and compare the results.
Absolutely amazing experiment and visualization! Thanks for putting all the effort into this. The only thing I find confusing or misleading is that somehow the power supply needs to "know" how much current it is "supposed to send" down the wire. The amount of current, when the switch closes, is immediately determined by the potential difference between the power supply and and the transmission line divided by the impedance (resistance and reactance) of the transmission line, all things you explain very nicely. From my perspective, there's no need to bring in agency that somehow the power supply needs to expect, communicate or know what's at the end of the wire. There's no causality to be broken. Keep up the good work! I love your videos!
"The amount of current, when the switch closes, is immediately determined by the potential difference between the power supply and and the transmission line divided by the impedance (resistance and reactance) of the transmission line, all things you explain very nicely." The electricity has to "learn" where it is going and how much is work is there when it arrives. It is a *process*. That is what makes these videos so brilliant.
Same. I mean, I understood the math, but not the why. Actually seeing the reflections? Worth so much more than all of the textual explanations of SWR I've ever seen.
the explanations didn't ever sit with me, but after I had the initial animations made so i could imagine what the waves looked like, I was plugging random resistors into the cable to make measurements and one of them didn't ripple. i just kind of stopped for a minute and went "ooohhhhhhhhhhh"
I work with PCB layout a lot, this series reminds me of a presentation from Rick Hartley on how to achieve proper grounding with circuits on and off circuit boards. That one was packed with a bunch of seemingly counterintuitive goodies as well. Fact is, we have been too dependent on the DC laws and the model we built in our head about the AC domain was some sort of extension from DC, and it turned out to be overly intuitive and incorrect. Rick described traces on the PCB as waveguides for electromagnetic fields to travel along, and it's not the electrons moving inside them that carries the engery, but rather the electromagnetic field around the conductors. When people get it seriously wrong, they end up facing a ton of crosstalks on a circuit board and scratch their heads wondering how the pulses "leaked out" of the traces. No, they were never inside the traces.
What a fantastic video. I've been doing electronics engineering for two decades and never had the intuition that this video just gave me. I love how empirical it is!
What I love about this is that it's basically just standard wave mechanics. Like, the graph of the voltages along the wire seemed fairly intuitive to me because it looks pretty well exactly like the wave mechanics I learned in high school physics.
Great video! I've been dealing with electronics for almost 20 years now, and this made me realize the following about impedance matching: It's basically like pulsing a wave of water at a water park, but it immediately goes through a gate. You then want as much of the pulse to make it through another gate at the end. The way you pulse the water influences the size of the gates that you want to not waste any wave.
Be it AC or DC, impedance matching will give you the maximum power transfer. When you have impedance matched, it essentially is removing the reflection by making it appear to not be entering a new medium. Any differences in impedance causes a reflection.
And to add to the previous comment .. analogies exist for optics, acoustics, and any other realm of physics that involves the transfer of energy from one medium to another. This is why some optical lenses have "anti-reflection" coatings, why some speaker designs are shaped in certain ways, and so on. You want the energy to see, in its intended direction of travel, as gradual a change in it the characteristics of its propagation medium as possible. Otherwise, boing! energy goes bouncing backwards.
The role of impedance matching in minimizing the reflections of the electrical waves is exactly what you would expect from the role impedance matching has in minimizing reflection of acoustic waves. Think of the load as a boundary between two media through which the electrical waves travel. If the resistance of the load is higher than "expected" it will cause free-end reflection (you can see that the reflections from all loads greater than 150 ohms are not inverted from the original pulse). If the resistance of the load is lower than "expected" it will cause fixed-end reflection (the reflections from loads lower than 150 ohms are inverted from the original pulse). If the resistance is just right, it's like a transition between boundaries of identical densities, no reflection will occur. Really cool to see this!
This video managed to teach me weeks of undergrad EE material in a matter of minutes. Hands down the best explanation of impedance (and electric flow in general).
Thanks man, you cleared it up. I'd been stuck with this contradiction (actual resistance of coax vs rated impedance) for about 2 weeks straight, since I started working with an SDR. You made my day!
As an electronic engineer I have learned all this but in a dry theoretical way. You make it visible. Thanks for your enormous effort. You must love it. I have used the reflection principle to find damage on data cables and cable studs to filter out unwanted frequencies. An open ended cable stud of 1/4 wave length of cause represent a short when connected parallel (at the right frequency and some of it's harmonics only of cause). I am subscribed to your video's and like watching them as although I know a lot of it I do not know it all and I still learn from both you and some of your viewers comments.
Oh man finally this is an intuitive&accurate explanation to what is impedance. The clumped bit explains what is the mysterious “reflection”. So the water model is precise, if you only consider gravity. Gravity itself is the impedance in water model.(surface tension and maybe all other fundamental fields contributes a bit but not too important?) Also the effect of EM field is just propagate too fast to become intuitive. Using an accurate water analogy is good for analysis. Also the slope of potential difference between two time points(how long?) is the value of impedance?
At 5:20, impedance matching fully made sense after all these years of trying to understand it. This and the other video have been monumental to my understanding of electricity. Thank you for all this!
Wow! Thanks for this intuitive demonstration. I have been dealing with those 50 ohm coaxial cable since 1990. I knew how important the terminator is for the ethernet system but never figured out why 50 ohm. Now I know that 50 ohm terminator minimizes the signals bouncing back.
14:12 I think that the pulse travelling down each fork in the wire is about 67% of the pulse in the 1st wire. That's because the 1st wire sees a termination of 75 Ω (the parallel of the 2 lines in the fork), and therefore 1/3 of the pulse will be reflected back (as shown). That 1/3 is given by the reflection coefficient Γ = (Zʟ-Zo)/(Zʟ+Zo) where Zo is the line impedance Zo = 150 Ω and Zʟ is the load terminating that line which in this case is Zʟ = 75 Ω.
I hope Brian reads your comment. I came up with the same ratio: at the fork boundary, the transmitted voltage pulses are +2/3 of the incoming pulse and reflected pulse is −1/3 only. As for power, transmitted powers are both 4/9 of initial power and reflected power is 1/9, hence a 100 % total.
Haven't seen the main video yet, but in just the first two minutes you've helped me understand RF impedance better than an entire semester of antenna design.
This applies perfectly well to RF, it's just that the system never settles, because the driving voltage is continuously changing. Then you get waves continuously propagating in both directions to create standing waves.
This is just freakin' amazing. I really, truly want to thank you for all this work. What a great service to those interested in learning about impedance.
This is amazing. I did an electrical engineering school 20 years ago, with between 4 and 8 hours a week of electromagnetic waves or transmission lines, and I never got as good an insight as to what was actually happening as after watching less than an hour of your videos. Great job!
From my experience, engineering school gets too wrapped up in teaching equations and methods with not enough focus on understanding the concept that drives the method or equation. Some professors are incredibly bad about this and others understand it totally
This topic always fascinates me. And I was struggling trying to understand impedance matching until I got into ham radio and started reading ARRL Antenna Book. In the transmission lines theory section, it explains Z matching for AC circuits and it's importance. Basically, a transmission line represents myriads of capacitors and inductors in series that make the line reactive and gives its characteristic impedance. When the load Z at the output of the line exactly matches the characteristic Z the line "thinks" it's just a continuation of itself.. No reflections occur, resulting in full transfer of power with no losses (except the material's resistance). But yeah, your "recognizance'' pulse is super cool. Another, mind blowing thing for me is that if, for example, a line had no resistance and was 100% lossless, the current would still directly depend on the voltage in accordance with the Ohm's Law, just as though resistance was there!... I don't have an engineering background so it really seems cool to me . Surely, for engineers it's not news at all)))
The fact I watched this and thought it "obvious" afterwards (despite never being taught it before), *really* goes to show how well explained this was. Incredible.
This was fascinating to watch, especially with the addition of all the o-scope plots. Seeing how the waves propogate through the circuit, reflect, and eventually find an equilibrium was so cool. Thanks for making this!
Thank you. This video provided answers I have been looking for for about 10 years. When I was young I was really passionate about music. That got me into physics and I eventually became an engineer (though not an electrical engineer). After watching your videos I feel like I have finally understood some of the questions that younger me had about speakers, amplifiers, and electric instruments. I have really enjoyed the work you have done on electricity here. The other content is great too. You are my favorite TH-camr and this is my first comment.
I've always found impedance matching somewhat intuitive, but honestly I'm seriously impressed with your experimental setup and the resulting graphs. Awesome job!! Thanks for all the effort and the great video!
Radio: where transients are all you have. Seriously, though, this is a great explanation for both line impedance and an actual visualization of SWR. The only thing missing is a Smith chart. In studying for my Extra license, there's so many sources of impedance mismatching that the cause of impedance in the transmission line is rather glossed over. There's an emphasis on avoiding reflections (because each bounce causes more loss to heating the wire due to internal resistances) but not much of a dive into the 'why'. I was writing a comment asking "Where did the 150 ohms come from?" and then you actually answered it.
This is a brilliant and painstaking way to de-mystify what electrons are doing when you turn on DC, and gives insight into why standing waves exist when you have an RF source. I'm going to use this pair of videos with new hams, Thanks so much!!
This brings me back to doing impedance phasor diagrams, and using simple trig to find the Z of and RLC circuit. Years ago...can't remember anything other than it was fun!
I have been trying to understand impedance matching to make a pcb wifi antenna for an esp32 chip and this explains exactly why the antenna needs to be impedance matched. I have watched a couple of hour long pcb design seminars trying to wrap my head around it and you made it make sense within 5 minutes.
I have always known since I first saw the leaves in a Leyden Jar, that electron density is always a factor in figuring out the details of how circuits work. I'm glad you have publicized this aspect of 'live' circuits.
Incidentally, the resistive wire kind of shows why we use series termination in digital lines that aren’t highly impedance controlled - having a signal switching on and off, we put a small series resistor, often near the source (transmit side) to absorb some of the reflection that comes from impedance mismatch or discontinuities when the signal is being switched. Having the whole line being resistive isn’t desirable because it’d absorb (I.e. waste) a lot of power, and you’d have to drive a lot more current into it, but a small series termination resistor helps absorb reflections and ringing without much loss.
Source termination is actually fascinating if you match it to the line impedance - at the transmitting end it forms a divider with your microstrip and thus only half the voltage goes down the wire, but at the receiving end with its near-infinite impedance, the voltage doubles due to the reflection and the signal is preserved! Then the reflection goes back up and gets eaten by the source termination resistor, although you can send multiple pulses and the waves will just go through each other.
In a digital line, you have all the odd harmonics making up the square wave, therefore you want the characteristics impedance of the line to be non-dispersive ( having the same phase velocity for all frequencies) to avoid having a "slew" in your square wave pulse (which would be caused by different present harmonics having different phase velocities).
Excellent. Truely, these two videos are invaluable and I want to show them to a lot of undergraduate and graduate students. Congratulations on explaining the concepts so well!
one extra thing to add to this very fine explanation and visualization of transmission lines - the two wires are not only communicating electrostatically, but also magnetically. the two wires in close proximity have magnetic linkage, like a transformer, and so as voltage gradient starts electron flow, that electron flow has a magnetic implication that also generates voltage gradient in the other wire. The two electron flows attempt to be equal and opposite, to minimize the amount of energy that has to be stored in the magnetic field, with the result that magnetic energy is almost only present in the space between the wires.
Heavy duty stuff - I’m going to have watch this again and again. When physicists look at light wave transmission down an optical fibre we worry about a backreflection from the end of the fibre. It is caused by a refractive index difference between glass and air and we can ‘match it out’ with a drop of oil (no longer a good idea with todays fibre connector designs.) we picture the light wave bouncing off the glass to air interface at the end of the line. In our heads it looks like the graphs you have displayed here. Ultimately the physics is the same as it is an electric field driving a charge distribution but I’m going to have to think about it all over again. Bravo on your brilliant work.
Shweeet!!!!! More about my favorite part of the main video!! Thanks for the link to this one! Yeah man, this is so cool! Learning more about electrical shenanigans makes me understand more and more why it's such a good analog to sound
stellar content. This beats any explanation I've seen before for line impedance and impedance matching that doesn't require a very solid understanding of maxwell's laws.
Wow, just wow. What a great way to convey a bunch of really tough concepts. I took a computerized instrumentation design course for my physics undergrad about 35 years ago at Cornell. We had an analog oscilloscope, and we're doing our data capture using an Apple 2 computer. It's amazing how far we've come in that time. There's no way we could have done anything like this with that equipment. Like you, I've been generally familiar with the concept of line and circuit impedence ever since those days. For instance, I've always known that speaker wire is typically 8 ohms and knew that was important to maintain the quality of the sound. Still, I always felt I was missing the big picture. I also know that within networks, having proper line terminators is important and was even more so back in the day when we used bnc cables. These reflections are the reason. Imagine if these pulses were bouncing back from every connection or disconnected line. So now you need to move on to what this means in A/C circuits. It becomes even more important in that context where you don't ever have a steady state, and you are always dealing with some component of the line impedance. I know I these circuits you can easily get RLC coupling between the impedance, the inductance, and the capacitance of the wires and the rest of the circuit often creating a virtual high or low pass filter that degrades the signal. This could even be a good opportunity to do a cross-over with one of the more math heavy channels (3 blue 1 brown, maybe? ). You could handle all the practical demonstration side of it like you've done here. Grant could then talk about signal filtering, complex analysis, and / or how differential equations enable us to characterize these circuits. Those are really tough topics, but a demonstration like this would really help make them tangible. I'd also like to see what happens if you two collaborated on the data visualization. You both do an amazing job but have different styles based on the kind of content you do. Thanks!
THANK YOU. I asked the exact same question at 8:53 in university 10 years ago and got a very vague awnser. Am am so happy I finally have a intuitive explanation.
This discussion goes from 0 to 100 for me. I know you provided all the necessary information and explanation but I'm still having a hard time equating this to SWR and impedance matching. BTW, you're a very smart person, thanks for sharing your knowledge and insights.
Really helpful way to observe something which is very hard to understand. Thanks for your hard work in putting this together and thank you for calling out the anthropomorphisms used too often to describe why something happens in the physical world.
please for the love of knowledge, science and life, DO NOT STOP MAKING VIDEOS. I greatly enjoyed taking my time over 1 hour of pausing and re-iterating every sentence and fragment of this wonderful explanation until I was absolultly certain I understand every notch and intrigue you presented its a revitilizing experiance in todays lack-of-attention-span-world to actually focus clearly on something, even if its just for an hour, and having an absolute blast diving into a fascinating subject - presented so clearly thank you.
Yes. Huge help. I am an engineer ( mechanical ). However I work with DC circuits and RF coaxial all my life for fun and hobbies and work… I thought I understood Impedance and SWR, standing wave ratio. However this was an epiphany… With RF you need to get best impedance match to get most RF energy out transferred to antenna or RF signal from antenna to receiver. Although RF is not DC it is subject to the same transient phenomenon. Seeing your videos gives me deeper insight. Thank You.
Outstanding work to capture and present this concept. Since you are more practical and motivated than me - perhaps a line constructed of a chain of series inductors and parallel capacitors to represent the resistive "characteristic" impedance with the 'basic' elements that make up a cable (inductive conductors and capacitive coupling between conductors). Also helps explain why different cables have different impedances. Beautiful video - thanks again. Will show some others at the workplace to help reinforce the concept.
Change the spacing of the wires on the board and change the characteristic impedance and rerun the test to show that the initial pulse really would follow that new impedance! This is a great demo of the benefits of transmission line matching.
Good explanation, and great visualisations! Though I wouldn’t say that the initial current is “too much” or “too little” because I don’t think it’s useful to think of a transient in terms of the circuit’s future steady state. It’s exactly what it needs to be based on its current state (the impedance of the line that the transient has traveled down so far), and it obviously it doesn’t ‘want’ or ‘expect’ to see a load that is the same as the characteristic impedance, it has no information about what’s coming (and it goes without saying that it has no ability to desire anything!). Usually the *designer* wants impedance to be matched (in an audio, RF or high-speed digital circuit especially), but the electrons are just going to do what they do.
I think that the anthropomorphisation helps with intuition. Everyone obviously understands that the wavefront isn't actually a living being. And I do think that it makes sense to think of the wave as "expecting" something. If it hits anything other than the characteristic impedance along the line, it ceases to be a wave and becomes two waves (forward and reflected). Thus, the wave only exists as long as the impedance doesn't change and can be thought of as "expecting" a specific impedance. The concept of a wave is already a very human abstraction to begin with, so using that kind of language doesn't really change much.
WOW, now i understand why Profi-Bus needs a special cable, terminating resistors, why branching is not allowed / does not work and it is called mirroring on a hole new level :D ... Thank you so much!
As an RF Engineer, I live by this principle. It was really cool to see you approach it from the DC perspective and I am very impressed by your painstaking oscilloscope measurements!!! I have only ever seen plots like this in simulations, never with true measurements.
You can also view terminating resistor as substitute for the rest of infinite wire. This way it's easier to understand why there is no reflected wave in this case
fantastic way to phrase it! it looks like the wire goes on forever
@@AlphaPhoenix2 The Design of CMOS Radio-Frequency Integrated Circuits has a very amusing chapter on transmission lines containing the gem:
> We've just seen that an infinite ladder network of infinitesimally small inductors and capacitors has a purely real input impedance over an infinite bandwidth. Although structures that are infinitely long are somewhat inconvenient to realize, we can always terminate a finite length of line in its characteristic impedance. Energy, being relatively easy to fool, cannot distinguish between real transmission line and a resistor equal to the characteristic impedance.
@@kilovoltamp Sounds like an interesting read. I'll have to check it out thank you
@@kilovoltamp That quote sounds like a mash up of Turbo Encabulator and The Missile Knows Where It Is
@@kilovoltampthe line "energy, being relatively easy to fool" sounds like someone is going to get a good zap ⚡
25 years ago I learned all this at university. But the depth of insight you've added to that basic knowledge is a revelation.
Incredible work here
I learned the same thing, but 20 years ago...
Anyway what buggs me about the original verstasium post was the original claim that EE didnt learn this.
And.. well we bety much do. Impedance matchning is like the core of high frequency circit design. Really with out it most modern radio systems would not work at all, like a wifi circut
@@matsv201yeah, once you get into the Mhz you start to realise that everything is a capacitor and an inductor
@@Jefferson-ly5qe yea,, i don´t know exactly where the cut of is, we did basically all the labs in 2.5 and 5.1Ghz and then when i worked with it it was basically minimum 800Mhz ...
Ironically no i work with transport system and the highest frequency we use is 400Hz (note, Hz, not kHz or MHz), typically 8000V and around 1000A.
The people that did the system in the 1970 basically just made a gigantic coaxial cable and bent it to form. This turned out to be incredibly expensive. Of cause we talk about 100 of km of cable, so that is millions and millions.
Some of the AC we could simply eliminate but some of it need to be around, So we basically try to balanced stamped metal sheets. Resistance of the old system was also a bit to high.
Makes it a bit more complicated by the load of the vehicle change the induction.
Still its not really that complicated compare to like a 4G muti frequency base station that i worked with prior. its really just a different kind of complication because everything is huge.
this is not deep inside. It is an explanation made by a man that simply doen’t have enough technical skills (Ohm’s law , resistance definition for example)
The fact that more or less consumer grade scopes can actually capture this is utterly brilliant. And then all the graph visualisations, especially the one with the 8 or so different resistance levels and see how impedance matching magically works... very very cool! And fairly important with everything with antenna's, especially high powered ones that will reflect significantly if not correctly matched, blowing up amplifiers in the process.
All hail the Standing Wave Ratio meter & the 50 Ohm dummy load.
I never expected to see hard data showing these waves of electrons as clearly as the sea reflecting waves from a harbour wall.
This has to be one of the most revelatory data sets I've ever come across.
Dat here in appreciative, slight stunned silence as things I've used in RF & audio all fall into place!
Coming from the days where we would have to allow the scope to warm up for an hour before use, seeing the phrase "consumer grade scopes" blows my mind. But yeah, I guess that's where we are at. :)
I picked up a NanoVNA, and it's been really helpful tuning antennas. Generator and scope in one device!
@@trevorus I have one as well. I bought it for a project that I was working on a couple of years ago. Supposedly it can be modded to do light duty as a spectrum analyzer.
@@MrWaalkman
I don't often write comments, but WOW this was awesome. This, combined with the last video somehow got me to see an intense beauty in this "low-level", more advanced type of electrical properties. Thank you so much for this
You said exactly what I was going to say. Incredible work!!
It was really interesting to see the dynamical demonstrations of pulses propagating and hitting the different loads! I didn't expect the real world waveforms to be so close to the ones you see in textbooks. Bravo to you!
Ow, you are here too) xD
Ахах) да, привет
As someone who has learnt this at uni, i can say these videos have been the best explanation i have ever seen
As someone working in testing motors for EVs, this video and the main channel one were GOLD! The Portuguese PhD who tries to educate me at work will love this, and hopefully i can understand more about EMC!
Let me tell you that this just made me make sense of the DMX line impedance loads used to avoid "reflections" in the line.... I knew about this ripples or waves of "voltage" but I couldn't figure it out why a load of an "apparent impedance" could work to "destroy" them.... This is just amazing.... Thank you.
If you use the electron motion visualization again in future, adding colour based on speed may make it a easier to see the changes in velocity, so blue is slow, red is fast, and rainbow or gradient between.
Also, this video should be required watching for every EE student, it explains and shows impendance 10x better than anything I got in school.
Totally
As a ham radio operator and long time mechanic and experimenter you have just explained something I could not see in my mind for 50 years. Now that I have a visual, it is unbelievably simple. I was like you about the coax cable. I never could see how a 15 foot piece of 50 ohm could be the same as 200 feet. That really bugged the piss out of me. Thanks for the video.
To have an idea of where the 150 ohm characteristic impedance of the cable is coming from, you can think of the wire having some inductivity and that there is some capacitance between the two twisted wires in the cable. If you would compute the series inductance and parallel capacitance, the resulting impedance would be 150 ohm. Of course, when modeling this, you would split the cable into an infinite number of sections and each section would have its inductance and capacitance resulting in the characteristic impedance of 150 ohms. You can thing of the wave propagation as the transfer of energy between those inductors and capacitors in the infinitesimal segments, each having a 150 ohm characteristic impedance. This also helps to think what happens when the wave leaves the last segment of the wire and hits a short or open.
To understand that, I would have to know what inductivity and capacitance was but yes, it kind of makes sense. I'm still not sure how you can measure both impedence and resistance in Ohms. If they have the same units, shouldn't they be the same quantity? I guess this is like how "power consumption" is measured in "Watt-hours" and "energy" is measured in "Joules"?
@@aspzx You would measure them in Henry (Ohm * sec) for the inductance and Farad ( C / V ) ...I hope I still remember correctly, for the capacitance. (where C is Coulomb and V is volt)
@@aspzx So both resistance and impedance are the same ‘thing’, except that resistance in DC is a special case. The formulas for impedance of capacitors and inductors are complex (as in have a real and imaginary part) that depend on frequency, and at DC what happens is that the impedance of an inductor goes to zero and the impedance of the capacitor goes to infinity (which is the same as an open circuit), so once it reaches that steady DC state you can basically assume that both the inductance and the capacitance that is inherent in the wires doesn’t exist anymore, and all you see left over is the resistance. So impedance is the same thing, but the dynamic nature of the signal (in the case in the video it’s a transient, the voltage changing in a ‘step’ from zero up to the voltage of the battery when it’s switched on) is bringing out extra elements that you just can’t see at DC.
@DrenImeraj And the consequence of this is that the "width" of the traveling step-pulse wave front imposes a limit on the practical bandwidth of the wire?
The wire also has Reactance ie its "resistance" to the flow of AC current as in AC the magnetic field is rising and falling all the time affecting the flow of the electrons.
When I was a teen, I had a book about electronics by one G.M Scroggie. He described "transmission lines" in terms of a series of inductors in series with the two sides of the line, with capacitors across the two lines between each inductor. Basically a whole line of series/parallel tuned circuits which has an impedance related to the impedance of the capacitors and inductors. I didn't understand why, AT ALL. Now, thanks to THIS VIDEO, over 40 years later, I understand enough to SEE why you show it as tuned circuits. Not only that, but I know why the inductors are in series with the line, and the capacitors between the lines. I also see how the impedance works. I now understand why a CB Radio enthusiast needs a "Standing Wave Ratio" meter to ensure reflections are minimised at the aerial. This one series of videos has BLOWN MY MIND and I love it.
The link to the main channel in the video description leads to an error page, because of an ')' at the end of the link.
It is because of the three dots.
@@jeevanraj5305the three dots are a visual thing done by TH-cam lol
Lol
At D.C., yes, it reaches an equilibrium quickly but the point of impedance matching is at high frequencies where peaks and dips of frequency response will occur if impedances aren't matched. I find it very interesting and enlightening seeing these high speed captures you are doing in understanding these phenomena. It makes it so much easier to see what is happening. It is distributed capacitance and inductance which creates the 150 Ohm (or whatever impedance, it depends on the cable configuration / geometry) and the distributed resistance which determines the Q (usually negligible), but the basis of why this happens is shown in the transient responses you are demonstrating. This is significant and will be appreciated by teachers in time. Please keep up the good work!
Your 17 minute explanation has made more sense than an entire semester of transmission lines! Amazing! My circuit has been completed😅
I’ve been trying to understand this for 20 years and this video finally made it click. Thank you so much!
Now it makes sense, why I had to terminate a Modbus/RS485 wire pair bus with a 120 Ohm resistor, exactly to prevent signal reflection on the line. Thank you so much for all your videos, always looking forward to them!! 😊
One on each end I hope. :)
BTW, replacing the 120 Ω resistor with two 60 Ω resistors in series and connecting a "Properly sized" capacitor from the center point of the two resistors to ground will give you about a 3db noise reduction. See Jan Axelson's "Serial Port Complete" page 123 starting with "Terminations for Short Lines" for details.
Yes it's starting to make some sense ! VERY imformative videos !!
I'm battling another BUS system at Work, a Carlo Cavazzi Dupline system. (a type of large field automation system)
The signal is a Square wave with about 7 volt RMS and at 132 mS there are 128 waves, with different pulse width.
The transmission line is branched in several branches of different length, in total probably over 50 branches and combined, 2000+ meters of wire.
The system is generally used without any Termination at open ends, but we have lots of communication problems, which I attribute to open end cable reflections.
I wish someone at Carlo Gavazzi knew 5% of what others here know, so they could answer my questions 😐
@@daniel635biturbo Now that sounds like a fun one! :)
While adding impedance matching termination resistors should help out, it appears that your network is already below voltage. Adding termination resistors will increase the load on your network.
Okay, so the system runs at 1khz, and the choice of cable is up to you. Still, line terminations should help, but it would be dependent on what cable you used. Sounds like your voltage is quite a bit low (7 volts vs 8.2). Possibly a node is dragging the network down.
Since your system seems to be overloaded already, I would hold off on line terminations. You should have a "Stiff" voltage source before adding any more load.
At the data rates that the system runs at, a DC meter should suffice to (only) check voltage. But a scope would be pretty much mandatory to see what is really going on.
And I see what you are saying with the 128 "waves", that corresponds to the 16 x 8 matrix of device addresses with a sync pulse at the start. So the entire I/O "matrix" updates at 1hz?
Looks like someone figured out how to automate Morse code. What will they think of next? :)
My first recommendation, the easiest to install, and the one likely to have the best bang for the buck, is to chop the line in half and put a repeater in the middle. That will probably fix you up without having to do anything else. And it's not like you are delivering broadband speeds to the other end of the network. :)
Then you could measure each individual network and see if one is now at the magical 8.2 volts while the other is still at 7 volts. If this is the case, you probably have a module (node) dragging the buss down. And a scope will show if a node is at a different DC bias (the pulse will jump up or down for that node).
So is this for building automation, mining, metering, or?...
And finally PLCS.net is your friend, there are a handful of guys there that have used it. www.plctalk.net/qanda/showthread.php?t=11592&highlight=carlo+gavazzi+dupline
@@MrWaalkman Big thanks for your response, you probably now know more than me about how it works.
The RMS voltage is dependent of the duty cycle, so when I look at it in the ocilloscope I get probably 8,2 volt peak, have not really checked.
The system dates back to the 1990 in our factory, and are installed with non twisted 1,5mm2 homogenous copper leads, without any shielding.
Back in the 1980s Electromatic developed this in Denmark, but now the system is sold and manufactured in Italy.
And they seem to have lost A LOT of knowledge, my Swedish support can't really support my questions.
Earlier in the 1980s the installation recommendations were "free" do as you like, but now they ask why we don't have twisted pairs, or shielded cables....
Now, I can't really put my finger on what the problem really is, but I can see cable reflections on the waveform, and It's different depending where I measure in the system.
The problem presents itself with "Ghost signals" so the controller sets some inputs as TRUE for one communication cycle.
And that is a real problem handling in the "PLC" code, If I can't trust the signals, and the filtering methods are very clunky to say the least. (0,5 seconds+one cycle)
The controller should be able to handle 450mA load, and we are only at 30mA on the system that give me trouble.
After lots of research I found out that they sell "Termination units" I believe that it is a resistor and a capacitor and possibly a diode.
But have not cracked one open yet, this module is also called Impedance module, wonder why 😊
They recommend installing one (DT02) at 1200m from the controller in one cable end, but not several, as it "decreases possible transmission distances"
The termination unit don't seem to add any significant load, and the wave looks better, with one installed.
What I also find is that most "Ghost signals" occur during daytime, when the factory is in production, so the cable reflections are not solely the problem.
But perhaps the cable reflections make the system more sensitive to other signal noise, that occurs during daytime production.
Anyhow.... The Plctalk site seems down now, but I will take a look later, good tip !
This doesn't just apply to long wire transmission lines. Changes in impedance on PCB traces or lines connected to them creates reflections and that's generally where noise comes from. Robert Feranec has a bunch of videos talking to a signal integrity expert, Eric Bogatin, and he basically explains everything you did in these videos but in the context of PCB design.
Rick Hartley did a similar thing on Altium's channel. Video name "How to Achieve Proper Grounding - Rick Hartley" He ranked EMI issues above signal integrity. He explains it better than i do.
@@kazuviking I try and overclock SPI on an atsame51 chip so I can push frames to a ~115KB SPI display as fast as possible. I think the limit is supposed to be 18MHz according to the datasheet. If I add 33 ohm series resistors on the cheapest china 2 layer boards I can manage to get 24MHz before things get corrupted. That's just the first value I tested, on I think a 12 mil trace. It would be nice to go faster, and maybe even pass EMI testing even though I'll never sell anything, but I have no idea how I'd even begin to figure the characteristics of the out of spec peripheral.
I'm going to have to check out this videos, thanks dudes!
PCB traces are still modeled as transmission lines they just have a different geometry. PCB tend to be a strip, microstrip, and coplanar
*TEM transmission lines*
Coaxial line
Two-wire line
Parallel plate line
Strip line
Microstrip line
Coplanar waveguide
*high-order transmission lines*
Rectangular waveguide
optical fiber
Thank you! I am 53, have been interested in antennas and electronics my whole life, and never really understood cable impedance properly until now. One of the best explainer videos ive ever seen!
4 years of engineering bachelor, and now i understand why the characteristic impedance is seen briefly by the supply..... Thanks!
Great physical explanation of impedance. Best I have seen...
The data collection and subsequent visualization / animation here was incredible! This is remarkable to see how the measurements really demonstrated exactly what you were describing. Bravo for what I am sure was a significant undertaking in capturing all of this data!
Never seen this much honesty in anything related to electricity. My heartfelt thanks and prayers to universe that every EE in the world follows you. I hate just assuming things because somebody says so with oomph and authority. You deserve 8 million subscribers not just 8K. Just keep it basic and honest. This content is like watching Faraday himself at work. So innocent and humble.
This might be the best explanation of characteristic impedance on the internet. So intuitive, thank you.
Im just getting into hobby RF. perfect video to help me understand impedance matching and tuning.
the animation with the several different line impedances is fantastic. i wish this video existed way back when i was in engineering school!
I love the oscilloscope/line plots! I've always understood transmission impedance as the infinite ladder of inductors and capacitors, but never seen it measured/plotted like this -- very helpful!
Thanks! I “learned” about impedance back in 1976. But I never truly grokked it until now. Your incredible patience taking all those behind the scenes oscilloscope readings really paid off.
This is very visual. Thanks for doing this incredible tedious and crazy task of measuring!
You can simulate transmission lines without dirt effects in ltspice or qspice and compare the results.
Absolutely amazing experiment and visualization! Thanks for putting all the effort into this. The only thing I find confusing or misleading is that somehow the power supply needs to "know" how much current it is "supposed to send" down the wire. The amount of current, when the switch closes, is immediately determined by the potential difference between the power supply and and the transmission line divided by the impedance (resistance and reactance) of the transmission line, all things you explain very nicely. From my perspective, there's no need to bring in agency that somehow the power supply needs to expect, communicate or know what's at the end of the wire. There's no causality to be broken. Keep up the good work! I love your videos!
"The amount of current, when the switch closes, is immediately determined by the potential difference between the power supply and and the transmission line divided by the impedance (resistance and reactance) of the transmission line, all things you explain very nicely." The electricity has to "learn" where it is going and how much is work is there when it arrives. It is a *process*. That is what makes these videos so brilliant.
I’m an Extra Ham radio operator. This was really helpful in understanding coax. That has driven me crazy for a long time.
Same. I mean, I understood the math, but not the why. Actually seeing the reflections? Worth so much more than all of the textual explanations of SWR I've ever seen.
the explanations didn't ever sit with me, but after I had the initial animations made so i could imagine what the waves looked like, I was plugging random resistors into the cable to make measurements and one of them didn't ripple. i just kind of stopped for a minute and went "ooohhhhhhhhhhh"
@@AlphaPhoenix2 These are the best moments! 😁🤘
@@AlphaPhoenix2why only the specific one stopped the reflections?
I work with PCB layout a lot, this series reminds me of a presentation from Rick Hartley on how to achieve proper grounding with circuits on and off circuit boards. That one was packed with a bunch of seemingly counterintuitive goodies as well. Fact is, we have been too dependent on the DC laws and the model we built in our head about the AC domain was some sort of extension from DC, and it turned out to be overly intuitive and incorrect. Rick described traces on the PCB as waveguides for electromagnetic fields to travel along, and it's not the electrons moving inside them that carries the engery, but rather the electromagnetic field around the conductors. When people get it seriously wrong, they end up facing a ton of crosstalks on a circuit board and scratch their heads wondering how the pulses "leaked out" of the traces. No, they were never inside the traces.
What a fantastic video. I've been doing electronics engineering for two decades and never had the intuition that this video just gave me. I love how empirical it is!
What I love about this is that it's basically just standard wave mechanics. Like, the graph of the voltages along the wire seemed fairly intuitive to me because it looks pretty well exactly like the wave mechanics I learned in high school physics.
This is all stuff that I've studied in uni but no professor explained it this clearly! Great job!!
Great video! I've been dealing with electronics for almost 20 years now, and this made me realize the following about impedance matching:
It's basically like pulsing a wave of water at a water park, but it immediately goes through a gate. You then want as much of the pulse to make it through another gate at the end. The way you pulse the water influences the size of the gates that you want to not waste any wave.
Be it AC or DC, impedance matching will give you the maximum power transfer. When you have impedance matched, it essentially is removing the reflection by making it appear to not be entering a new medium. Any differences in impedance causes a reflection.
And to add to the previous comment .. analogies exist for optics, acoustics, and any other realm of physics that involves the transfer of energy from one medium to another. This is why some optical lenses have "anti-reflection" coatings, why some speaker designs are shaped in certain ways, and so on. You want the energy to see, in its intended direction of travel, as gradual a change in it the characteristics of its propagation medium as possible. Otherwise, boing! energy goes bouncing backwards.
Masterpiece.
I cannot imagine the amount of time and measures you spent on this project.
Thank you!
I have waited all my life to intuitively understand this! Thank you, thank you.
The role of impedance matching in minimizing the reflections of the electrical waves is exactly what you would expect from the role impedance matching has in minimizing reflection of acoustic waves. Think of the load as a boundary between two media through which the electrical waves travel. If the resistance of the load is higher than "expected" it will cause free-end reflection (you can see that the reflections from all loads greater than 150 ohms are not inverted from the original pulse). If the resistance of the load is lower than "expected" it will cause fixed-end reflection (the reflections from loads lower than 150 ohms are inverted from the original pulse). If the resistance is just right, it's like a transition between boundaries of identical densities, no reflection will occur.
Really cool to see this!
This video managed to teach me weeks of undergrad EE material in a matter of minutes. Hands down the best explanation of impedance (and electric flow in general).
Thanks man, you cleared it up. I'd been stuck with this contradiction (actual resistance of coax vs rated impedance) for about 2 weeks straight, since I started working with an SDR. You made my day!
The animated graphs from this video series are really well done. Really helps cement everything you are talking about.
As an electronic engineer I have learned all this but in a dry theoretical way. You make it visible. Thanks for your enormous effort. You must love it.
I have used the reflection principle to find damage on data cables and cable studs to filter out unwanted frequencies. An open ended cable stud of 1/4 wave length of cause represent a short when connected parallel (at the right frequency and some of it's harmonics only of cause).
I am subscribed to your video's and like watching them as although I know a lot of it I do not know it all and I still learn from both you and some of your viewers comments.
I love these well made videos, that clearly explain something complex/technical, and invoke an "oh my gosh, so that's why" epiphany.
Oh man finally this is an intuitive&accurate explanation to what is impedance. The clumped bit explains what is the mysterious “reflection”.
So the water model is precise, if you only consider gravity. Gravity itself is the impedance in water model.(surface tension and maybe all other fundamental fields contributes a bit but not too important?)
Also the effect of EM field is just propagate too fast to become intuitive. Using an accurate water analogy is good for analysis.
Also the slope of potential difference between two time points(how long?) is the value of impedance?
The measurements and visualizations are incredible in these videos! Gives me a real sense of understanding how it all actually works.
At 5:20, impedance matching fully made sense after all these years of trying to understand it. This and the other video have been monumental to my understanding of electricity. Thank you for all this!
Wow! Thanks for this intuitive demonstration. I have been dealing with those 50 ohm coaxial cable since 1990. I knew how important the terminator is for the ethernet system but never figured out why 50 ohm. Now I know that 50 ohm terminator minimizes the signals bouncing back.
14:12 I think that the pulse travelling down each fork in the wire is about 67% of the pulse in the 1st wire. That's because the 1st wire sees a termination of 75 Ω (the parallel of the 2 lines in the fork), and therefore 1/3 of the pulse will be reflected back (as shown).
That 1/3 is given by the reflection coefficient Γ = (Zʟ-Zo)/(Zʟ+Zo) where Zo is the line impedance Zo = 150 Ω and Zʟ is the load terminating that line which in this case is Zʟ = 75 Ω.
I hope Brian reads your comment. I came up with the same ratio: at the fork boundary, the transmitted voltage pulses are +2/3 of the incoming pulse and reflected pulse is −1/3 only.
As for power, transmitted powers are both 4/9 of initial power and reflected power is 1/9, hence a 100 % total.
I have learned more in these series of video than in my entire post secondary education. This visualization changes everything!
Haven't seen the main video yet, but in just the first two minutes you've helped me understand RF impedance better than an entire semester of antenna design.
This applies perfectly well to RF, it's just that the system never settles, because the driving voltage is continuously changing. Then you get waves continuously propagating in both directions to create standing waves.
This is just freakin' amazing. I really, truly want to thank you for all this work. What a great service to those interested in learning about impedance.
This is amazing. I did an electrical engineering school 20 years ago, with between 4 and 8 hours a week of electromagnetic waves or transmission lines, and I never got as good an insight as to what was actually happening as after watching less than an hour of your videos. Great job!
From my experience, engineering school gets too wrapped up in teaching equations and methods with not enough focus on understanding the concept that drives the method or equation. Some professors are incredibly bad about this and others understand it totally
This topic always fascinates me. And I was struggling trying to understand impedance matching until I got into ham radio and started reading ARRL Antenna Book. In the transmission lines theory section, it explains Z matching for AC circuits and it's importance. Basically, a transmission line represents myriads of capacitors and inductors in series that make the line reactive and gives its characteristic impedance. When the load Z at the output of the line exactly matches the characteristic Z the line "thinks" it's just a continuation of itself.. No reflections occur, resulting in full transfer of power with no losses (except the material's resistance).
But yeah, your "recognizance'' pulse is super cool. Another, mind blowing thing for me is that if, for example, a line had no resistance and was 100% lossless, the current would still directly depend on the voltage in accordance with the Ohm's Law, just as though resistance was there!... I don't have an engineering background so it really seems cool to me . Surely, for engineers it's not news at all)))
The fact I watched this and thought it "obvious" afterwards (despite never being taught it before), *really* goes to show how well explained this was. Incredible.
As a ham radio operator, this is helping me more than anything else I've came across.
I thought the original video was cool, but that is way beyond that - real gem of practical knowledge.
This was fascinating to watch, especially with the addition of all the o-scope plots. Seeing how the waves propogate through the circuit, reflect, and eventually find an equilibrium was so cool. Thanks for making this!
Thank you. This video provided answers I have been looking for for about 10 years. When I was young I was really passionate about music. That got me into physics and I eventually became an engineer (though not an electrical engineer). After watching your videos I feel like I have finally understood some of the questions that younger me had about speakers, amplifiers, and electric instruments.
I have really enjoyed the work you have done on electricity here. The other content is great too. You are my favorite TH-camr and this is my first comment.
I've always found impedance matching somewhat intuitive, but honestly I'm seriously impressed with your experimental setup and the resulting graphs. Awesome job!! Thanks for all the effort and the great video!
Radio: where transients are all you have.
Seriously, though, this is a great explanation for both line impedance and an actual visualization of SWR. The only thing missing is a Smith chart. In studying for my Extra license, there's so many sources of impedance mismatching that the cause of impedance in the transmission line is rather glossed over. There's an emphasis on avoiding reflections (because each bounce causes more loss to heating the wire due to internal resistances) but not much of a dive into the 'why'. I was writing a comment asking "Where did the 150 ohms come from?" and then you actually answered it.
This is a brilliant and painstaking way to de-mystify what electrons are doing when you turn on DC, and gives insight into why standing waves exist when you have an RF source. I'm going to use this pair of videos with new hams, Thanks so much!!
This brings me back to doing impedance phasor diagrams, and using simple trig to find the Z of and RLC circuit. Years ago...can't remember anything other than it was fun!
"very many oscilloscope traces" Indeed. Thank you for sharing your hard work.
This series is great, and the idea of combining multiple oscilloscope traces is fantastic. I hope you will continue it with some AC meassurements.
I have been trying to understand impedance matching to make a pcb wifi antenna for an esp32 chip and this explains exactly why the antenna needs to be impedance matched. I have watched a couple of hour long pcb design seminars trying to wrap my head around it and you made it make sense within 5 minutes.
I have always known since I first saw the leaves in a Leyden Jar, that electron density is always a factor in figuring out the details of how circuits work. I'm glad you have publicized this aspect of 'live' circuits.
Incidentally, the resistive wire kind of shows why we use series termination in digital lines that aren’t highly impedance controlled - having a signal switching on and off, we put a small series resistor, often near the source (transmit side) to absorb some of the reflection that comes from impedance mismatch or discontinuities when the signal is being switched. Having the whole line being resistive isn’t desirable because it’d absorb (I.e. waste) a lot of power, and you’d have to drive a lot more current into it, but a small series termination resistor helps absorb reflections and ringing without much loss.
It's just like the damper in a shock absorber. If you only have a damper it burns up, and if you only use the springs it bounces too much.
Source termination is actually fascinating if you match it to the line impedance - at the transmitting end it forms a divider with your microstrip and thus only half the voltage goes down the wire, but at the receiving end with its near-infinite impedance, the voltage doubles due to the reflection and the signal is preserved!
Then the reflection goes back up and gets eaten by the source termination resistor, although you can send multiple pulses and the waves will just go through each other.
In a digital line, you have all the odd harmonics making up the square wave, therefore you want the characteristics impedance of the line to be non-dispersive ( having the same phase velocity for all frequencies) to avoid having a "slew" in your square wave pulse (which would be caused by different present harmonics having different phase velocities).
Excellent. Truely, these two videos are invaluable and I want to show them to a lot of undergraduate and graduate students. Congratulations on explaining the concepts so well!
Wow! This is impressive! I only wish professors would take the time to teach like you.
one extra thing to add to this very fine explanation and visualization of transmission lines - the two wires are not only communicating electrostatically, but also magnetically. the two wires in close proximity have magnetic linkage, like a transformer, and so as voltage gradient starts electron flow, that electron flow has a magnetic implication that also generates voltage gradient in the other wire. The two electron flows attempt to be equal and opposite, to minimize the amount of energy that has to be stored in the magnetic field, with the result that magnetic energy is almost only present in the space between the wires.
Heavy duty stuff - I’m going to have watch this again and again. When physicists look at light wave transmission down an optical fibre we worry about a backreflection from the end of the fibre. It is caused by a refractive index difference between glass and air and we can ‘match it out’ with a drop of oil (no longer a good idea with todays fibre connector designs.) we picture the light wave bouncing off the glass to air interface at the end of the line. In our heads it looks like the graphs you have displayed here. Ultimately the physics is the same as it is an electric field driving a charge distribution but I’m going to have to think about it all over again. Bravo on your brilliant work.
Excellent work. This is one of the best videos I've ever seen.
I admire the ingenuity and patience needed to setup this experiment.
Shweeet!!!!! More about my favorite part of the main video!! Thanks for the link to this one!
Yeah man, this is so cool! Learning more about electrical shenanigans makes me understand more and more why it's such a good analog to sound
I'm SO glad you made this video because my first thought upon watching the first one was that it makes impedance matching make perfect sense!
stellar content. This beats any explanation I've seen before for line impedance and impedance matching that doesn't require a very solid understanding of maxwell's laws.
Wow, just wow. What a great way to convey a bunch of really tough concepts. I took a computerized instrumentation design course for my physics undergrad about 35 years ago at Cornell. We had an analog oscilloscope, and we're doing our data capture using an Apple 2 computer. It's amazing how far we've come in that time. There's no way we could have done anything like this with that equipment.
Like you, I've been generally familiar with the concept of line and circuit impedence ever since those days. For instance, I've always known that speaker wire is typically 8 ohms and knew that was important to maintain the quality of the sound. Still, I always felt I was missing the big picture. I also know that within networks, having proper line terminators is important and was even more so back in the day when we used bnc cables. These reflections are the reason. Imagine if these pulses were bouncing back from every connection or disconnected line.
So now you need to move on to what this means in A/C circuits. It becomes even more important in that context where you don't ever have a steady state, and you are always dealing with some component of the line impedance. I know I these circuits you can easily get RLC coupling between the impedance, the inductance, and the capacitance of the wires and the rest of the circuit often creating a virtual high or low pass filter that degrades the signal.
This could even be a good opportunity to do a cross-over with one of the more math heavy channels (3 blue 1 brown, maybe? ). You could handle all the practical demonstration side of it like you've done here. Grant could then talk about signal filtering, complex analysis, and / or how differential equations enable us to characterize these circuits. Those are really tough topics, but a demonstration like this would really help make them tangible. I'd also like to see what happens if you two collaborated on the data visualization. You both do an amazing job but have different styles based on the kind of content you do.
Thanks!
THANK YOU. I asked the exact same question at 8:53 in university 10 years ago and got a very vague awnser. Am am so happy I finally have a intuitive explanation.
I finally understand impedance matchin. I'm a electronics technician an your video filled this gap in my knowledge. Thanks.
This was super informative and very beautifully illustrated! I can only imagine all the work that went into it, thanks for a great video :)
Beyond impressive. This is main channel video quality, even if it's a follow-up video. This one is almost more interesting than the first!
Gosh this and the other video on measuring waves of electricity was absolutely brilliant. Thanks for opening my eyes.
This discussion goes from 0 to 100 for me. I know you provided all the necessary information and explanation but I'm still having a hard time equating this to SWR and impedance matching. BTW, you're a very smart person, thanks for sharing your knowledge and insights.
Woah! This was always a hand-wavey topic for me that I never put in the effort to fully understand. What an epic visualization!
Really helpful way to observe something which is very hard to understand. Thanks for your hard work in putting this together and thank you for calling out the anthropomorphisms used too often to describe why something happens in the physical world.
please for the love of knowledge, science and life, DO NOT STOP MAKING VIDEOS.
I greatly enjoyed taking my time over 1 hour of pausing and re-iterating every sentence and fragment of this wonderful explanation until I was absolultly certain I understand every notch and intrigue you presented
its a revitilizing experiance in todays lack-of-attention-span-world to actually focus clearly on something, even if its just for an hour, and having an absolute blast diving into a fascinating subject - presented so clearly
thank you.
Awesome video and explanation. This helps me understand why impedance matching is so critical for audio
Yes. Huge help. I am an engineer ( mechanical ). However I work with DC circuits and RF coaxial all my life for fun and hobbies and work… I thought I understood Impedance and SWR, standing wave ratio. However this was an epiphany… With RF you need to get best impedance match to get most RF energy out transferred to antenna or RF signal from antenna to receiver. Although RF is not DC it is subject to the same transient phenomenon. Seeing your videos gives me deeper insight. Thank You.
The animation hitting the different loads is top notch!
Best explanation of electrical impedance I've ever heard/seen. Your videos are brilliant. Well done. Thx
This is one of the best pieces I've ever seen on the subject as far as being understandable.
As a ham radio operator this was really cool to watch. Thanks!
Outstanding work to capture and present this concept. Since you are more practical and motivated than me - perhaps a line constructed of a chain of series inductors and parallel capacitors to represent the resistive "characteristic" impedance with the 'basic' elements that make up a cable (inductive conductors and capacitive coupling between conductors). Also helps explain why different cables have different impedances. Beautiful video - thanks again. Will show some others at the workplace to help reinforce the concept.
I studies transmission lines long ago but this video really brought it home. Nice job!
Change the spacing of the wires on the board and change the characteristic impedance and rerun the test to show that the initial pulse really would follow that new impedance!
This is a great demo of the benefits of transmission line matching.
Good explanation, and great visualisations! Though I wouldn’t say that the initial current is “too much” or “too little” because I don’t think it’s useful to think of a transient in terms of the circuit’s future steady state. It’s exactly what it needs to be based on its current state (the impedance of the line that the transient has traveled down so far), and it obviously it doesn’t ‘want’ or ‘expect’ to see a load that is the same as the characteristic impedance, it has no information about what’s coming (and it goes without saying that it has no ability to desire anything!). Usually the *designer* wants impedance to be matched (in an audio, RF or high-speed digital circuit especially), but the electrons are just going to do what they do.
How do you know electricity is not a living, conscious being? That's just a speculation or yours
I think that the anthropomorphisation helps with intuition. Everyone obviously understands that the wavefront isn't actually a living being.
And I do think that it makes sense to think of the wave as "expecting" something. If it hits anything other than the characteristic impedance along the line, it ceases to be a wave and becomes two waves (forward and reflected). Thus, the wave only exists as long as the impedance doesn't change and can be thought of as "expecting" a specific impedance.
The concept of a wave is already a very human abstraction to begin with, so using that kind of language doesn't really change much.
WOW, now i understand why Profi-Bus needs a special cable, terminating resistors, why branching is not allowed / does not work and it is called mirroring on a hole new level :D ... Thank you so much!
Thank you for making this follow up. It's how I wanted the first video to finish! Awesome.
As an electronics developer I find this a very good way to show how this works. :)