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 ⚡

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)

You said exactly what I was going to say. Incredible work!!

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

It is because of the three dots.

​@@jeevanraj5305the three dots are a visual thing done by UKposts lol

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 !

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

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"?

@@aspuzling 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)

@@aspuzling 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.

Totally

Ow, you are here too) xD

Ахах) да, привет

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.

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).

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.

"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.