I am very happy that I found your channel! Excellent quality explanation and demonstration clarifying confusing and sometimes “mysterious” RF topics. Thank you very much for sharing your knowledge.
As an electro hobbyist i got a high value knowledge about RF topic here on Your channel what was looked as a mysterious woodoo magic for me.😊 Thank you . Blessing to you.
Nice to know that this also translates to twisted pair topology. I've been doing a lot of TDR twisted pair measurements at work lately, and was having a hard time finding information specific to twisted pairs vs coax.
I know the TDR method works on twisted pair. I haven't tried the VNA methods on twisted pair, so I am not 100% sure on that one. It would be an interesting experiment, though.
@@eie_for_you I'm using a LibreVNA at work to test twisted quad, actually - roughly 2.7km in our test setup. You might want to check LibreVNA out. It's not as convenient as a tiny/nano/pico vna, but it's considerably more versatile, and all the schematics/software/firmware is open source.
@@MadeInMichigan Hmmmm...haven't heard of LibreVNA before. Sounds interesting. The nanoVNA with the VNA Saver app claims to do TDR, but I've never used it that way. I might just have to experiment with it. Thanks!
There is one more way...Find the frequency of the 1/4 wave length of the line with the far end open circuit. Multiply that frequency by .5 and 1.5. Measure the R and J at those two frequencies. Calculate the two impedances using the R and J. Average those two impedances. This is essentially the same procedure that the FCC specifies to measure the sample lines of AM directional broadcast stations.
Very nice explanations! From my experience I have had measure really long twisted pair cable around 1km. In this case resistance is also matters. What would be the way to measure resistance using VNA, same as inductance?
Thank you! 🙂 Measuring the impedance of twisted pair cable was my first introduction to doing all of this. Communication cable for EEG amplifiers. Resistance ... I am assuming series resistance of the wire itself. You could short the far end and measure this with your an appropriate DVM, but I recommend using the methods I describe in this video: ruclips.net/video/M46EtjltOug/видео.html Regarding the inductance, as far as I know, you are not going to be able to get at the inductance all by its itself. What you read with the VNA is a composite impedance which is composed of the resistance, capacitance and the inductance.
Great video! I have a bunch of unknown cable (conducts are 2x 16awg, 5x 24awg) which I intend to use for some signals over the 5 smaller cables along with power over the larger ones. I am hoping to transmit an RS422/RS485 equivalent differential signal......Is the TDR method valid if was to pick two of these smaller gauge wires as a "pair" to know how to properly terminate the line at both ends? I have access to a DSO and sig-gen, so this method would be best :) Thanks!
Yes, actually! A company I worked for bought a 1,000 ft roll of multiconductor, twisted pair, custom cable that I had to test for impedance. The TDR method is the method that I used to do that.🙂
Great video learnt heaps, thanks. Question if I may. Regarding the third TDR method: your generator is set to 80K square wave 50% duty cycle (at 12:00) followed by a High Pass RC Filter with a cut off frequency of 451Mhz (7.5pF & 47ohms). I note the oscilloscope trace was a nice rounded single pulse with little high frequency components taking about 3 divisions (10nS/Div) or equivalent to about 33MHz. I'm guessing you very much reduced the duty cycle off screen. What is the purpose of the Resistor and Capacity (HPF) network? Thanks again
Thanks! The purpose of the RC network is to give me a highly differentiated signal so we only see the transitions in the square wave. We do this because we want VERY narrow pulses to shoot down the coax. This will produce very narrow positive and negative pulses. The repetition rate is such that we will never see the negative side of this. We trigger on the positive-going incident pulse and then watch for the reflected pulse some time later. The period of the square wave has to be such that the reflected pulse returns before the negative incident pulse is around and yet short enough to allow a fast response time. Hope this helps. 🙂
@@gregwmanning That is one way of thinking about it, but it is not actually the technically correct one. Because this whole apparatus is operating in the time domain, we have to consider the problem in the time domain, not the frequency domain. The leading edge of the square wave acts like a positive step function (a voltage going from 0 volts to some positive voltage in 0 seconds ... though we know it actually has a finite rise-time). Initially the voltage across the resistor and capacitor is 0. When the leading edge of the square wave comes along the capacitor acts like a short because it fully discharged and the current through it is maximum. This is the current that passes through the resistor. High current, fixed resistance...high voltage on the resistor. BUT the capacitor charges with the current passing through it and its voltage rises. This causes the current to decrease through the resistor and the voltage across the resistor decreases. Eventually, the capacitor voltage is equal to the applied positive voltage of the step function. The voltage across the resistor drops to zero. The amount of time it takes from the initial very positive peak to the drop to zero volts is determined by the values of R and C. ??????? So, why are the pulses so very rounded instead of the pulse I have described? We have two distinct reasons. First, it is because the actual rise-time of the square wave is anything BUT a true step function. Thus, the leading edge of the pulse gets mushed and rounded. Second, the overall frequency response of the whole measurement system with the coax and all makes it impossible to actually "see" the high frequencies necessary to observe anything that sharp. I hope this helps or have I stirred up the silt and muddied that water worse?
@@gregwmanning Here is a quick spreadsheet (in the ZIP file) for you to play with. drive.google.com/file/d/1m-MuAroXaS3AXYo7ej7EaXAEX0yiIiFJ/view?usp=sharing
@@eie_for_you You are very kind, with such a detailed and enlightening explanation. I have downloaded your spreadsheet and will digest it now. Many thanks Ralph.
thanks for your Excellent explanation I want to know if I want to calculate its Characteristics Impedance In 10 Mhz Frequency what changes in this test circuit like R(47Ohms) and C(7.5pF)
The frequency that I use for the TDR method is chosen so the falling edge of the square wave is far enough away from the rising edge so the rising edge is effectively acting as a step function. And yet, the repetition rate is high enough to make for a nice solid trace even on an analog scope. It isn't the test frequency because we are operating in the time domain, not the frequency domain for this test. The purpose of the RC network is to create a narrow pulse to send down the transmission line. In my opinion, there is no real practical reason to perform this test at 10 MHz. But, if you do ... I simply chose values for R and C so that I got a nice, highly differentiated pulse as viewed on the scope (no calculations, just picking values out of the air). I would change the capacitor if I changed anything first. Remember ... easy to forget ... we are operating in the time domain with the TDR method. It is so very easy to have one foot in the time domain and the other in the frequency domain and get mixed up on this.
Nothing formal like a TDR instrument. I used an oscilloscope, a square wave generator, a resistor and a small value capacitor. So, yeah, I had my own DIY TDR which is sufficient for the job.
Well, yes, that is one use. It can also be used with various antenna types, for impedance matching, for closed circuit TV video, and a lot of other things. This particlar roll of coax was used for an antenna feedline, believe it or not.
I am very happy that I found your channel! Excellent quality explanation and demonstration clarifying confusing and sometimes “mysterious” RF topics. Thank you very much for sharing your knowledge.
Yes! There is a LOT of "magic" and mystery in RF! You are very welcome! I am glad to have been a blessing to you. :-)
As an electro hobbyist i got a high value knowledge about RF topic here on Your channel what was looked as a mysterious woodoo magic for me.😊 Thank you . Blessing to you.
Thanks! Yes ... RF is a LOT of mysterious voo-doo! I am glad that my channel has helped to clear the fog a bit. :-)
very good information,thank you for your time
You are welcome! 🙂
Like your presentation
Thanks, man! 🙂
Nice to know that this also translates to twisted pair topology. I've been doing a lot of TDR twisted pair measurements at work lately, and was having a hard time finding information specific to twisted pairs vs coax.
I know the TDR method works on twisted pair. I haven't tried the VNA methods on twisted pair, so I am not 100% sure on that one. It would be an interesting experiment, though.
@@eie_for_you I'm using a LibreVNA at work to test twisted quad, actually - roughly 2.7km in our test setup.
You might want to check LibreVNA out. It's not as convenient as a tiny/nano/pico vna, but it's considerably more versatile, and all the schematics/software/firmware is open source.
@@MadeInMichigan Hmmmm...haven't heard of LibreVNA before. Sounds interesting. The nanoVNA with the VNA Saver app claims to do TDR, but I've never used it that way. I might just have to experiment with it. Thanks!
Nice explanation
Thanks! 🙂
There is one more way...Find the frequency of the 1/4 wave length of the line with the far end open circuit. Multiply that frequency by .5 and 1.5. Measure the R and J at those two frequencies. Calculate the two impedances using the R and J. Average those two impedances. This is essentially the same procedure that the FCC specifies to measure the sample lines of AM directional broadcast stations.
Hmmmm...I hadn't heard of that one before. Sounds like a fun thing to try. 🙂
Very nice explanations! From my experience I have had measure really long twisted pair cable around 1km. In this case resistance is also matters. What would be the way to measure resistance using VNA, same as inductance?
Thank you! 🙂 Measuring the impedance of twisted pair cable was my first introduction to doing all of this. Communication cable for EEG amplifiers.
Resistance ... I am assuming series resistance of the wire itself. You could short the far end and measure this with your an appropriate DVM, but I recommend using the methods I describe in this video:
ruclips.net/video/M46EtjltOug/видео.html
Regarding the inductance, as far as I know, you are not going to be able to get at the inductance all by its itself. What you read with the VNA is a composite impedance which is composed of the resistance, capacitance and the inductance.
Are you sure you're not Ralph Cable?
Great video!
I have a bunch of unknown cable (conducts are 2x 16awg, 5x 24awg) which I intend to use for some signals over the 5 smaller cables along with power over the larger ones. I am hoping to transmit an RS422/RS485 equivalent differential signal......Is the TDR method valid if was to pick two of these smaller gauge wires as a "pair" to know how to properly terminate the line at both ends? I have access to a DSO and sig-gen, so this method would be best :)
Thanks!
Yes, actually! A company I worked for bought a 1,000 ft roll of multiconductor, twisted pair, custom cable that I had to test for impedance. The TDR method is the method that I used to do that.🙂
@@eie_for_you fantastic! Thank you! Looks like I've got some homework to do then 🤣
@@mjunek It's fun to discover stuff. When you see it working, it is very satisfying! 🙂
Great video learnt heaps, thanks. Question if I may. Regarding the third TDR method: your generator is set to 80K square wave 50% duty cycle (at 12:00) followed by a High Pass RC Filter with a cut off frequency of 451Mhz (7.5pF & 47ohms). I note the oscilloscope trace was a nice rounded single pulse with little high frequency components taking about 3 divisions (10nS/Div) or equivalent to about 33MHz. I'm guessing you very much reduced the duty cycle off screen. What is the purpose of the Resistor and Capacity (HPF) network?
Thanks again
Thanks! The purpose of the RC network is to give me a highly differentiated signal so we only see the transitions in the square wave. We do this because we want VERY narrow pulses to shoot down the coax. This will produce very narrow positive and negative pulses.
The repetition rate is such that we will never see the negative side of this. We trigger on the positive-going incident pulse and then watch for the reflected pulse some time later. The period of the square wave has to be such that the reflected pulse returns before the negative incident pulse is around and yet short enough to allow a fast response time.
Hope this helps. 🙂
@@eie_for_you So RC network are only passing higher harmonics of the 80K square wave?
@@gregwmanning That is one way of thinking about it, but it is not actually the technically correct one. Because this whole apparatus is operating in the time domain, we have to consider the problem in the time domain, not the frequency domain.
The leading edge of the square wave acts like a positive step function (a voltage going from 0 volts to some positive voltage in 0 seconds ... though we know it actually has a finite rise-time).
Initially the voltage across the resistor and capacitor is 0.
When the leading edge of the square wave comes along the capacitor acts like a short because it fully discharged and the current through it is maximum. This is the current that passes through the resistor. High current, fixed resistance...high voltage on the resistor. BUT the capacitor charges with the current passing through it and its voltage rises. This causes the current to decrease through the resistor and the voltage across the resistor decreases. Eventually, the capacitor voltage is equal to the applied positive voltage of the step function. The voltage across the resistor drops to zero.
The amount of time it takes from the initial very positive peak to the drop to zero volts is determined by the values of R and C.
???????
So, why are the pulses so very rounded instead of the pulse I have described?
We have two distinct reasons.
First, it is because the actual rise-time of the square wave is anything BUT a true step function. Thus, the leading edge of the pulse gets mushed and rounded.
Second, the overall frequency response of the whole measurement system with the coax and all makes it impossible to actually "see" the high frequencies necessary to observe anything that sharp.
I hope this helps or have I stirred up the silt and muddied that water worse?
@@gregwmanning Here is a quick spreadsheet (in the ZIP file) for you to play with.
drive.google.com/file/d/1m-MuAroXaS3AXYo7ej7EaXAEX0yiIiFJ/view?usp=sharing
@@eie_for_you You are very kind, with such a detailed and enlightening explanation. I have downloaded your spreadsheet and will digest it now. Many thanks Ralph.
thanks for your Excellent explanation I want to know if I want to calculate its Characteristics Impedance In 10 Mhz Frequency what changes in this test circuit like R(47Ohms) and C(7.5pF)
The frequency that I use for the TDR method is chosen so the falling edge of the square wave is far enough away from the rising edge so the rising edge is effectively acting as a step function. And yet, the repetition rate is high enough to make for a nice solid trace even on an analog scope. It isn't the test frequency because we are operating in the time domain, not the frequency domain for this test. The purpose of the RC network is to create a narrow pulse to send down the transmission line. In my opinion, there is no real practical reason to perform this test at 10 MHz. But, if you do ... I simply chose values for R and C so that I got a nice, highly differentiated pulse as viewed on the scope (no calculations, just picking values out of the air). I would change the capacitor if I changed anything first. Remember ... easy to forget ... we are operating in the time domain with the TDR method. It is so very easy to have one foot in the time domain and the other in the frequency domain and get mixed up on this.
You had a TDR setup? I had a used GDO and bits of wire. 🙄
Nothing formal like a TDR instrument. I used an oscilloscope, a square wave generator, a resistor and a small value capacitor. So, yeah, I had my own DIY TDR which is sufficient for the job.
I guess it's TV coax!
Well, yes, that is one use. It can also be used with various antenna types, for impedance matching, for closed circuit TV video, and a lot of other things. This particlar roll of coax was used for an antenna feedline, believe it or not.
I am hearing you with care. ua3dpf