Verry interesting video :) i never saw someone before who dosent use a 75 ohm signal generator on a 75 ohm load. In the past i taught it would be necessary to avoid reflection but today i know its not. I like that you verry clearly showed that only the termination determined the reflection.
Using a matched impedance at both ends is always a good thing, but my generator doesn't really have proper output impedance setting... On the other hand, you can add a series termination at the beginning of the line (between generator and cable) and by doing that the signal will bounce only 1 time at the far end of the cable, but the reflection will be absorbed at the beginning of the cable. If the end of the cable is high Z, the signal at the beginning will be signal/2 (75R-termination+75Z of the cable divided) and the signal at the end will be (signal/2)*2 (because of the impedance mistmatch) so the signal passes unaffected
@@FesZElectronics good point but i think most people try to get the most power of the generator like they do in audio amplifier impedance matching. i have two questions: 1. have you ever made a video about miller effect in amplifiers and in general with impedance chaning maybe on a complex numers level? and 2: have you ever worked with microcontrolers or do you want to make videos about some applications with them?
To be honest, 1 - no an 2 - I did work with uC but I don't really plan any videos with that at the moment. I mainly want to focus on analog applications. Would there be any subject in particular you would be interested in though?
@@FesZElectronics no actually not. Analog stuff ist anyway more interesting and there are not many educational videos on TH-cam with high quality like this. If i could choose a video it would be tutorial about antennas and how they radiate and pick up electromagnetic waves, because i never had an intuitive feeling for this behavior.
your graph with the red line in the middle and Z1 and Z2 reminded me of Descartes law of refraction that I hated back in school, since both light and electricity are both EM waves, I think they are somehow related but in a subtle way, but I may be just imagining things, anyways great video Fesz.
Hello, I am doing an experiement. I need to transmit a signal very quickly from one connection to another. First the connection will be open. Then closed. Then open again. Kind of like a switch. If I was going to speed up the transfer of the signal during the closed time, should I increase the frequency? Sorry I'm kind of new to this. I would be thankful for any help.
Great video, this really solidified the concept of transmission lines for me. Learnt about impedance matching for low frequency circuits back in analog electronics class, so it’s interesting to see how impedance matching is an elegant solution to multiple problems. I wonder, at the limit of a continuous change in impedance across a transmission line (like a coax cable with constantly increasing r1/r2), would there be any reflection issues at all? I also wonder how much of an issue a BNC connection in the middle of a transmission line makes, the only BNCs I’ve seen are made for RG58, so I assume there are slightly different ones made for RG59? I’d also be interested in seeing some sort of “ltspice toolkit” full of spice models and examples and such that you use. Then again, I’ve been meaning to swap to KiCAD’s internal spice simulator.
Well a coax cable should have uniform impedance throughout its length, it should not change. On the other hand, because of its series resistance (the wiring is not perfect) the cable has losses. In a proper coax datasheet you have the cables attenuation for a standard length (100m) and for various frequencies. But its not a reflection issue, its just losses caused by the dielectric and the wire. Regarding the connection between lines, it all depends on the signal passing trough it and the transmission length of the BNC connector. If the added bit has a Td comparable to the signal then it will cause problems, if not,I think it can be ignored. There are some LTspice examples you might find interesting in the installation file under LTspice>Examples>Educational. I covered this in an old video th-cam.com/video/MwfeMjcM4zY/w-d-xo.html
Light in vacuo travels 1 foot in 1 nanosecond (2sf). This is about the only current use for the foot as it gives Americans a feel for the speed of light.
lol @5:01 Well you what always gets me? I am fine if people just want to use their own system and units, but please if you make a table anyway, just convert it to units everyone understands. So why not using both in the same table as well as the calculations? It will take you a couple of minutes more writing the article, yet it is so much more readable. Excellent video again! Btw, I would bump up your resolution in LTSpice, you will get a lot more points that way to see the ringing better.
Hello P_Mouse! It's not a resolution problem. The signal should look all boxy. My measurements looked more sine wavy because of the signals used (slow rise-fall times) and the BW limitation of the measurement equipment. I'm happy you enjoyed the video!
Well, to be fair, I simulated an ideal line. The real line has some extra series resistance that causes the line to attenuate signals based on their frequency - my practical line has from the datasheet an 0.9dB attenuation/100ft at 10MHz but at 6GHz it reaches 21.1dB/100ft of attenuation. By the way, in the simulation command line I had a maximum time step set to 0.01n for the last simulation (the one with 2 pieces of transmission line) and the signal simulated had a transition time of 1n so there should be quite enough points.
You can take advantage of the coax as an inductor, and just add extra plates as capacitors. This sort of technique is useful in very high frequency applications >1GHz. Similar technique is called a "Distributed element filter" - you will see this sort of layout in very high frequency circuits.
If the IC works in the GHz range, you should. I've seen in some high frequency IC's (I think it was the HMC994) that the silicon is no longer encased in epoxy, but rather the space is left void (or filled with air); I guess this was done to keep a close control of the dielectric constant around the IC and the bond wires.
Well, when it comes to high speed circuitry, usually the proper simulation model to be used is an IBIS model. This should contain information about how each I/O behaves over a wide frequency range when signals are passing trough it. So on the one side you can use some basic simulation models to make sure the circuit works (models that take into account the basic internal functions) and then finish the design by simulating the interfaces between the components using the IBIS models to ensure that signals get transmitted from one point to another without to much distortion.
far better than my Dr. Ing. prof. X who has 30 years of experience. keep doing it. thanks a lot.
I love this channel, your metode for explain concepts is very good. Please don't stop uploading your videos.
most underrated channel ever. awesome video
Great clear explaination. Thank you.
Yet another great video! Even my cat, Tigrut, looked very interested of your explanations.
Kitty :D
great video, we really want to see another videos on the topic
thank you
Verry interesting video :) i never saw someone before who dosent use a 75 ohm signal generator on a 75 ohm load. In the past i taught it would be necessary to avoid reflection but today i know its not. I like that you verry clearly showed that only the termination determined the reflection.
Using a matched impedance at both ends is always a good thing, but my generator doesn't really have proper output impedance setting...
On the other hand, you can add a series termination at the beginning of the line (between generator and cable) and by doing that the signal will bounce only 1 time at the far end of the cable, but the reflection will be absorbed at the beginning of the cable. If the end of the cable is high Z, the signal at the beginning will be signal/2 (75R-termination+75Z of the cable divided) and the signal at the end will be (signal/2)*2 (because of the impedance mistmatch) so the signal passes unaffected
@@FesZElectronics good point but i think most people try to get the most power of the generator like they do in audio amplifier impedance matching. i have two questions: 1. have you ever made a video about miller effect in amplifiers and in general with impedance chaning maybe on a complex numers level? and 2: have you ever worked with microcontrolers or do you want to make videos about some applications with them?
To be honest, 1 - no an 2 - I did work with uC but I don't really plan any videos with that at the moment. I mainly want to focus on analog applications. Would there be any subject in particular you would be interested in though?
@@FesZElectronics no actually not. Analog stuff ist anyway more interesting and there are not many educational videos on TH-cam with high quality like this. If i could choose a video it would be tutorial about antennas and how they radiate and pick up electromagnetic waves, because i never had an intuitive feeling for this behavior.
Keep the good job. Thanks a lot.
As usual ,very informative video . Keep on !!!
Great content. Waiting for further content on transmission lines.
your graph with the red line in the middle and Z1 and Z2 reminded me of Descartes law of refraction that I hated back in school, since both light and electricity are both EM waves, I think they are somehow related but in a subtle way, but I may be just imagining things, anyways great video Fesz.
You are brilliant!
Hello, I am doing an experiement. I need to transmit a signal very quickly from one connection to another. First the connection will be open. Then closed. Then open again. Kind of like a switch. If I was going to speed up the transfer of the signal during the closed time, should I increase the frequency? Sorry I'm kind of new to this. I would be thankful for any help.
Thanks for good videos
can you make a reaction or experiment as a response for veritasium with the 1s Transmission thought experiment
Great video, this really solidified the concept of transmission lines for me. Learnt about impedance matching for low frequency circuits back in analog electronics class, so it’s interesting to see how impedance matching is an elegant solution to multiple problems. I wonder, at the limit of a continuous change in impedance across a transmission line (like a coax cable with constantly increasing r1/r2), would there be any reflection issues at all? I also wonder how much of an issue a BNC connection in the middle of a transmission line makes, the only BNCs I’ve seen are made for RG58, so I assume there are slightly different ones made for RG59?
I’d also be interested in seeing some sort of “ltspice toolkit” full of spice models and examples and such that you use. Then again, I’ve been meaning to swap to KiCAD’s internal spice simulator.
Well a coax cable should have uniform impedance throughout its length, it should not change. On the other hand, because of its series resistance (the wiring is not perfect) the cable has losses. In a proper coax datasheet you have the cables attenuation for a standard length (100m) and for various frequencies. But its not a reflection issue, its just losses caused by the dielectric and the wire.
Regarding the connection between lines, it all depends on the signal passing trough it and the transmission length of the BNC connector. If the added bit has a Td comparable to the signal then it will cause problems, if not,I think it can be ignored.
There are some LTspice examples you might find interesting in the installation file under LTspice>Examples>Educational. I covered this in an old video th-cam.com/video/MwfeMjcM4zY/w-d-xo.html
@@FesZElectronics can you explain how to simulate a long twisted pair or coax cable with loses also?
Clearly explained
Light in vacuo travels 1 foot in 1 nanosecond (2sf). This is about the only current use for the foot as it gives Americans a feel for the speed of light.
Thanks a lot! 👍👍👍
lol @5:01
Well you what always gets me?
I am fine if people just want to use their own system and units, but please if you make a table anyway, just convert it to units everyone understands.
So why not using both in the same table as well as the calculations?
It will take you a couple of minutes more writing the article, yet it is so much more readable.
Excellent video again!
Btw, I would bump up your resolution in LTSpice, you will get a lot more points that way to see the ringing better.
Hello P_Mouse! It's not a resolution problem. The signal should look all boxy. My measurements looked more sine wavy because of the signals used (slow rise-fall times) and the BW limitation of the measurement equipment.
I'm happy you enjoyed the video!
@@FesZElectronics are you sure? I actually don't think so. This is just a form of ringing caused by oscillations.
Well, to be fair, I simulated an ideal line. The real line has some extra series resistance that causes the line to attenuate signals based on their frequency - my practical line has from the datasheet an 0.9dB attenuation/100ft at 10MHz but at 6GHz it reaches 21.1dB/100ft of attenuation.
By the way, in the simulation command line I had a maximum time step set to 0.01n for the last simulation (the one with 2 pieces of transmission line) and the signal simulated had a transition time of 1n so there should be quite enough points.
Very interesting.
How can you use a transmission line combined with a capacitor to achieve a resonant LC tank?
You can take advantage of the coax as an inductor, and just add extra plates as capacitors. This sort of technique is useful in very high frequency applications >1GHz. Similar technique is called a "Distributed element filter" - you will see this sort of layout in very high frequency circuits.
Hi, I am new to simulations, what type of simulation is this in SPICE?
Hello! I use LTspice in all my videos as software, and in this video I used transient type simulations - its the basic time domain simulation.
@@FesZElectronics thank you
Would you treat the bondwire as transmission line?
If the IC works in the GHz range, you should. I've seen in some high frequency IC's (I think it was the HMC994) that the silicon is no longer encased in epoxy, but rather the space is left void (or filled with air); I guess this was done to keep a close control of the dielectric constant around the IC and the bond wires.
@@FesZElectronics i think this can be important also, but it looks not very widely supported by the design tools..
Well, when it comes to high speed circuitry, usually the proper simulation model to be used is an IBIS model. This should contain information about how each I/O behaves over a wide frequency range when signals are passing trough it. So on the one side you can use some basic simulation models to make sure the circuit works (models that take into account the basic internal functions) and then finish the design by simulating the interfaces between the components using the IBIS models to ensure that signals get transmitted from one point to another without to much distortion.