Thank you so much for the kind words. Amateur radio is what got me started in electronics so many years ago (after a brief introduction with a 100 mW CB Sears base station I got for Christmas as a pre-teen). I don't operate a lot, but I learned so much from getting licensed and studying books published by ARRL. It's nice to hear this is helping to give back some to the community. 73's
Greetings from middle-aged men in China (although it might not be very popular)! I became fascinated with electronics two years ago, I believe its not too late. Thank you very much, and admire your effort in education after retirement!
Thanks for the greeting ! Glad you enjoyed the videos. Yes - its never too late. I'm learning biology and bio-chemistry in retirement - now that I have more time. Best wishes from the US!
Stray Capacitance, Stray Inductance, ghosts in the machine. The little VNA's have become a real treasure for amateur electronic experimentation, and obliviously an asset even to the more experienced electrical engineer.
My friend, this is amazing. I'am learning english and I love electronics. In your videos, I find both of those to be high quality. Be sure I'am using my NanoVNA today. Many thanks. :-)
Brilliant video and explanation of the black art of RF design. Thanks. Just FYI, the nanoVNA-F V2 goes up to 3GHz and can use 301 plotting points. The v0.5.0 firmware doesn't include thru calibration though this isn't really a disadvantage.
saw another RF youtubers video where he recommends component measurement should be considered at point where Reactance equals source impedence(50 ohms for NanoVNA), at 90 degree points. Makes sense as it will form a proper voltage divider at that point.
That's reasonable, but the actual low-frequency value of inductance (or capacitance), is actually slightly off at those points on the Smith Chart. For an inductor, anywhere along the arc on the left side of the chart should give about the same value. For an inductor I usually go to a point about halfway between the 90 degree point and the far left where the impedance is zero. (except don't take it too close to the 0 Ohm point, since accuracy will fall off there of course). For a capacitor, it's the opposite. One needs to take the measurement on the arc between the bottom 90 degree point and the far right of the chart where Z goes to infinity as frequency tends to zero. So I think the 90 degree rule is pretty good.
Thanks! There's a few additional videos and playlists available. See the channel's associated website at ecefiles.org (or directly on RUclips at www.youtube.com/@MegawattKS ) Thanks for leaving the kind comment.
Great overview video. Careful though when measuring large or small impedances with an S11 measurement, the VNA gets a lot less accurate then. The "testing" version of nanoVNAsaver has some nice new features for measuring impedances using an S21 measurement. That would help improve accuracy of the measurements quite a bit and you may find that your high resistance test will give you a somewhat different result (though still having quite some effect from parasitics)
Good point - especially for the high R values well above 50 Ohms. I haven't tried the S21 method, but did write a paper a couple years back about measuring Q. We found that with HP8753 units at least, the measurements can be quite decent if the measured impedance is close to the cal-points (0 and infinity), and worst around +/- j50. But I don't know if the Nano uses the same calibration equations...
Hi... I really enjoyed this very educational video regarding NanoVNA measurements of RLC components. I'm learning how to use the NanoVNA to measure LC circuits and stand-alone inductors. Your analysis piqued my curiosity about how to interpret the measurements. My question concerns the measurement of the self-resonant point of an inductor. When you showed the model schematic, we saw R(big) to be in parallel with L, which I took as the equivalent parallel resistance exhibited at resonance by the parallel LC circuit formed by the parallel combination of the L and parasitic capacitance. So near the self-resonant point, the apparent value of R(big) increases, as shown in the demo. i think you mentioned that R is the series resistance. The value of R(big) was around 2.6K near the resonant frequency, which seems rather high for a series resistance for the coil. I think the VNA is measuring the resistance it sees, which in this case should be the equivalent parallel resistance R(big). If this interpretation is correct, this is great way to measure Q, which I calculated to be about 7 based on the measurements. Thanks again, looking forward to future videos on using the NanoVNA.
Hi Scott, You have hit on some very important points! The extraction of model element values and the analysis of Q is indeed tricky near self-resonance, and even in the technical literature there are some inaccurate approaches and innacurate measurements reported. I like your focus on the self-resonant frequecy (SRF) and what happens there and agree that what you are seeing at self-resonance is the parallel loss resistance Rbig - not the series R. That's where the literature can be misleading, and you have correctly realized the complexity. Here goes for an attempt at summarizing (I'm neglecting the 15nH/in interconnect L in the model here): At low frequency (well below the SRF), one can just do Q=X/R where X and R are the imaginary and real parts of the impedance reported by the VNA. In this regime, we can delete R(big) from the model and just use the series R. But starting somewhere around 1/2 of the SRF, this simple calculation breaks down and the model needs to include the parallel R(big). Indeed this then gives Q=0 at the SRF since the measured value is purely resistive as you said (the SRF is where the VNA curve changes from inductive to capacitive on the right side of the chart). But there is still a resonance, so clearly there is non-zero Q ! So in this regime (between frequency of about 0.5SRF and SRF) a different approach is needed. I'll follow-up with this thought in the next reply to keep this one reasonable length... -Bill
Here is what I have done to extract more meaningful Q values between frequencies of about 0.5*SRF and the SRF: 1) delete the series R from the model so that it simplifies to a parallel RLC circuit. Take a measurement around 0.2 SRF to find the reactance and compute the L value from that. 2) Use this low-frequency L value together with the SRF to compute the C value for the model. The next step is where the math requires complex number crunching, so forgive me if I don't explain it well: 3) Take measurements of Z=R+jX at the frequencies of interest between 0.5 and 1.0 times SRF. Convert this impedance Z to an admittance Y (Y=1/Z) and then subtract off the admittance of the capacitance found in step 2. This gives an admittance Y' of just the R(big) in parallel with L. The real-part of Y' is then equal to 1/Rbig. 4) if Q is desired at this frequency, use the parallel-mode Q formula Qp=Rbig/X_L where X_L is the reactance of the inductor in the model at this frequency 🙂
@@MegawattKS Bill - my interest in all this stems from a recent experience I had measuring inductance of large air-wound coils having moderate inductances. In one case, I had a loading coil for an antenna of about 150 uH consisting of about approximately 100 turns of enameled wire on a PVC form. i made this quite a while ago, but recently I used a NanoVNA to sweep from about 50 kHz to 10MHz to measure the inductance. I expected to see a normal Smith chart plot for an inductor in this frequency range, but this coil has a self-resonance at less than 2 MHz, well below the intended operating frequency of about 3.7 MHz. I know it is most likely due to parasitic interwinding capacitance that is rather large, because the 100 windings are right next to each other, and wound on PVC that has a high dielectric constant of about 3. However, I was not sure if this is an artifact or a real effect. I made two such coils to be used as loading coils to resonate a 40 meter 1/2 wave dipole in the 80 m band. I guess my point is that theoretically, this coil is made to the desired inductance geometrically, but its frequency response is another issue and not considered in the design criteria. Would this inductor work as a loading coil then if it turns into a capacitor at the frequency of interest? If it does work as an inductor at the operating frequency, then does the parasitic capacitance get swamped out by the antenna capacitance? I don't think so, because I calculated that the parasitic interwinding capacitance should be about 70-80 pf, which is on the order of the antenna capacitance for a 40 m dipole operated on the 80 m band, which the coil is to null out (e.g., say about 10 pf antenna capacitance), unless the capacitances are in series, in which case the add in parallel. The smaller one dominates, as you pointed out in the video. If you have any insight to this dilemma, I would love to hear your feedback. Thanks again for your great video series on all the subjects you cover on your channel. Scott ☺
@@scottgilbert7927 Hi Scott. I'll take a stab at this. I have experience with the theory, but no direct experience designing a loaded dipole. My first thought is that 150 uH sounds very large for the goal of loading a 1/2 wave 40 m dipole to work at 3.7 MHz. But where are the loading coils going? At the feedpoint or somewhere near the middle of each element? Maybe 150uH is correct for the latter case, but it's j3500 Ohms at 3.7 MHz, which is hard to realize. Assuming that's the goal, then I think the interwinding C is the culprit as you said. And the VNA is telling us its a capacitor, which it is. It won't work as an inductance load. Maybe try winding the coils with wire with some thicker insulation to cut down on C (though the coil length might grow of course), or use a larger diameter PVC form to decrease the number of turns needed? The other option I can think of is loading each side of the dipole at the feedpoint. I think the antenna you describe (40m dipole used at 80m) looks something like Z = 73/4 Ohms - jX with X maybe a couple hundred Ohms (maybe check with EZNEC simulator?). In that case, you only need L = X/(2*pi*f) = 8.6 uH. Of course you'll then need to build something like an L-match network to convert 73/4=18 Ohms to 50 or 75 as needed for the transmission line. Hope some of this helps. Great problem and application of the NanoVNA !
Interesting. I was always been told that modern high stability resistors have the resistive material wound around the body like a coil. They are trimmed by laser to match tolerance etc. So it's the resistive element is inductive. Old carbon resistors dont have the problem that modern resistors do. Wire wound are highly inductive. Non inductive resistors can be purchased. Modern resistors dont suffer with carbon migration when used on DC voltage. But this is what I have been taught over 30 years at Tech College but things may be different these days. Cheers from old George
Hi George. Now that you mention it, I think I heard that too. Not sure how many resistor formulations use that technique. But it's important to remember that even if they don't, there's tons of inductance (e.g. 20 nH per inch) for just the resistor itself (it has non-zero length), plus the length of any interconnects between it and a capacitor, etc. As you said, wire-wound resistors can also add substantial inductance - being basically coils - in addition to their own length and interconnects on a PCB/etc. Cheers.
The first VNA was not by HP, but by Rohde & Schwarz at the beginning of '50s: the ZDD. It is also a well known and sought after instrument between collectors.
Thanks! Can you provide a link to information on it? I can't seem to find it with a Google search on "Rohde & Schwarz ZDD" or related things. I did find an IEEE paper in ARFTG that talks about history, but it only says the 1964-released HP8410 was the first widely used one. It mentions some homodyne units from Hazeltine from the 50s, and Wiltron and Rantec in '63. ieeexplore.ieee.org/document/4119938
I'm very interested in using a VNA for the building and adjustment of toroids used in HF low pass filters. At the moment I'm working on a diplexer type filter design. Ideally, I need an instrument that will provide 1nH resolution. Do you believe the NanoVNA is suitable for that task? Perhaps you know of a better suited - ideally inexpensive instrument?
Technically the answer is yes - if you're below about 100 Ohms reactance at 14 MHz for example (100 Ohms is about 1000 nH there and the display shows 3 significant digits in the marker readout). BUT - this is resolution - not accuracy. Accuracy of a VNA is not nearly that good and depends heavily on whether you're measuring small to moderate impedances, or large ones closer to the right side of the chart, where things get really 'off'. As for alternatives, at VHF and above, a VNA is about the best one can do, without creating special test fixtures that apply resonance with a known capacitor value. At HF, there are somewhat higher accuracy methods without resonance use, but they're expensive - and frankly seldom worth messing with, IMO.
I am curious if you had tried the Inductors(uH range) on small toroids and observed them on VNA.....i unfortunately started with a inductor on FT37-43 while being a new user on NanoVNA. Took me days before realizing how bad the inductor on small toroid could be with load of stray capacitance, it was almost a diagonal line on the Smith chart instead of being a circle. With air core inductors(nH range) its not as bad to get it right, it forms a decent circle.
I just looked up the -43 core and it appears to be Nickel-Zinc, and claims to have a high Q (at lower frequencies - e.g. below 30 MHz). Try resetting the Stimulus frequency range to 0 to 30 MHz (or even lower frequencies) and then recalibrate it and try the measurement again. I think what you may be seeing with the diagonal line is the NanoVNA taking only a couple datapoints in the low frequency range and just connecting these points with a straight line. (if it's sweeping 0 to 1000 MHz and only takes 100 points, then the frequency step is 10 MHz.) I've seen it do this, and had to drop the frequency range and recalibrate to get it to take enough points to make an arc as it should from 0 to 30 MHz.
@@MegawattKS I had it sweeping 1-30Mhz...what i meant by diagonal line(exaggerated to make point) is that the Smith curve was kind of going all over the uH range instead of making anything that resembles a circle. I had other known value inductors(wound on bobbins) tested to double check if something wrong with calibration, they were working fine.
Really appreciate your present at but I have a question I can’t seem to get answered. In this video, you showed how (with an inductor) when you raised frequency the inductive-reactance increased-so good. In a couple of videos W2AEW & Gregg Messenger, they showed (when connected to an antenna) that as the frequency increased, the inductive-reactance decreased by rotating downward from the upper hemisphere past the Real-axis into the capacitive lower hemisphere. An antenna will always look more-and-more inductive as frequency is increased (i.e. antenna too long); I would think the vector should rotate upward as frequency increases as it did for your inductor. Tnx & 73…
Hi. Their videos sound correct. It turns out that dipoles (or monopoles or most other antennas) actually do go through periodic resonances. At each resonance, it transitions from capacitive to inductive, or inductive to capacitive. Actually inductors will do this too if measured at high enough frequencies (above what's called their self-resonant frequency). Here is a webpage that shows a graph of real and imaginary parts of an antenna's impedance versus frequency (reactive, imaginary part in blue). They didn't label it, but the horizontal axis is frequency in MHz for the 30m monopole analyzed. But dipoles/etc will do the same thing. Hope that helps. 73 www.siranah.de/html/sail018r.htm
@@MegawattKS well thank you for taking the time to both comment and providing the link. I did get a response from Alan saying that the measurement-point is at the coax, not the antenna unless it is at exact 1/2 wavelengths. I must be honest, I’m still trying to get my head around this still (since it always seems to go in the same direction) but will try some experiments to hopefully gain greater insight so, we might get with you again. Thanks Again & 73…
At 14:36, you were measuring a 9100 ohm resistor with the nanoVNA, I was wondering if you were using the S11 method? I think that 9100 ohms is at the edge of the accuracy of the instrument, and switching (although, not easy with this setup) to the series S21 and converting to S11z... Copper Mtn website has some interesting info on the VNA accuracy for various Z ranges.,.. I have indeed re-learned quite a bit from your videos, I've enjoyed taking some of the formula and coding them into Python as calculators... Thanks for you time sharing this info... Oh, I see Rjordon was all over this....
Good points. Yes, I was just using the S11 method. I'll have to look at the Copper Mtn site and see if they found the same results we did relative to improved accuracy near the short and open points on the chart after calibrating for S11 measurement using SOL references. Thanks for the comments!
i've seen alot of videos on this, i'm getting one. great video. why would you spend 400 dollars for an antenna analyzer, when you can spend 100 and use the nanoVNA and can also use it to test filters, amps, etc. 73s
@@joeb3300 Excellent. I used to pull a bench unit with a VGA output, stick it on a cart, and hook it to the classroom projector - but that would be easier for sure :-)
A+ info, pace, and presentation quality. Immediately upped my game, thanks! Where did you get the test boards? Can you provide a link to the Radio Class handout? I'd find them handy when teaching other hams. Edit: I now read that others have asked the same question going back a year or more. I understand the overhead of creating a site for distro. If these came from a class, can you point us to the folks who put on the class? I wondered if you rolled your own test boards. I also use U.FL connectors for their ease of handling and quick connect/disconnect. When handled carefully they last a long time and give good results.
Good to hear. Thanks for letting me know it helped! The boards that the radio was built using are ones I designed and fabbed through ExpressPCB (I like their simple free layout software). I agree on the U.fl connectors. We used those in the microwaves course and they are fine for repeat connections :-) The course itself that inspired these videos is one that I and others did at K-State. But we're all retired now, so to keep it alive, I did these vids. It's still on the books, but not actively being taught at the moment. Here's a related K-State web page. It's got some stuff that might help (e.g. the last bullet in the first section takes you to the class handout sheets). ece.k-state.edu/about/people/faculty/kuhn/
@@andros2112 Thanks ! I think I've got it back. They changed the URLs of faculty/staff last month by putting an 'about' into the URL path. Hopefully this page will stay intact. I'm currently lobbying for the administration to not delete emeritus faculty pages (or at least this one. Hopefully the department at least will keep the content, if not this exact URL...
The "RF Demo Kit" shown in the video was purchased from Amazon in the US. But it is basically sold by DeepElec. I'm not sure what stores you have access to, but you can also find it and other items on the DeepElec site on AliExpress here: deepelec.aliexpress.com/store/1101363542 And details on it, including some excellent close-up photos, are on the Deepelec site here: deepelec.com/rf-demo-kit/
No. There are a couple reasons. It's meant to look at "stimulus-response", for things like measuring gain (or 'return-loss'). Without the stimulus part of that, it cannot frequency/phase-lock, so it will not read the signal. The other critical reason is that it could be damaged by anything more than about a milliwatt of power. (actually, the damage level may be 10 dBm, but that's still 0.01 Watt max). I have noted on Amazon that there are a number of small low-cost power meter modules out there now for similar price range (e.g. 50 dollars) with a simple LCD display. But again, they're meant for very low powers, well under a watt.
@@MegawattKS than you for your reply. I want to measure max power output of radio transmitter, which has max power output 20dbm in data sheet. in data sheet this via has max 10db range , by adding attenuator I cannot measure ? if not can you also provide a link to pwoermeter please. thx
@@tomipiriyev The NanoVNA still can't do it, since it needs to look at something which is 'coherent' with the signal it puts out. (it could do an amplifier since it could provide the input, but for a transmitter, the signal is coming from somewhere else). What frequency are you wanting to do this at? If it's below 900 MHz, then you might want to look at getting one of the really cool "TinySA" spectrum analyzers (and use an attenuator as you said). But if its WiFi like 2.4 or 5.x GHz, you may do better with something like this (I'm not recommeneding this particular product, but it gives an idea maybe): www.amazon.com/Taidacent-Calibration-Measurement-50-0dBm-Attenuation/dp/B07JMKQTDW/ref=sr_1_3?crid=27JYNJNFYW5W8&keywords=rf+power+meter . I don't know how accurate it is though...
@@tomipiriyev Yes - the TinySA is different than the NanoVNA. The TinySA is a spectrum analyzer (showing power levels of signals vs frequency, or power density due to modulation vs frequency). The NanoVNA is a network analyzer designed to measure things like gain (ratio of size of output signal to size of input signal) vs frequency. The TinySA is closer to what I think you're asking about, but unfortunately it does not go above about 0.9 GHz, so it won't work. For 2.4 GHz, the cheapest and best route to just getting the power level is probably something like the power-meter board in the link (together with a power supply and 20 or 30 dB attenuator). And it won't have the same issues of reading inaccurately due to wideband modulation that a spectrum analyzer would.
Just a video editing glitch. I do a lot of "slash-and-burn" editing (throwing out parts of what is filmed to tighten the story). I tried to pause it to see after you pointed it out - and as far as I can tell, there seems to be a frame or two with one of the "spinny arrows" that Pinnacle Studio sometimes generates for unknown reasons. They seem to get generated during the video export process. I upgraded recently to a newer version, so hopefully there will be less of these types of glitches :-)
S-parameters please: measure components (passives) with nano vna Smith Chart(r +/-jw), put into circuit (simple), predict response with S-parameter algebra. 100Mhz would be perfect for this. Then: talk about signal graphs,....
Good ideas. That's currently deeper into math than I've been doing in these - but I'll try to think about this for down the road. I agree that this would work best at 100 MHz. Too often S-parameters at microwave frequency vary too much with the exact test fixture used in the reference design/measurement (e.g. PCB thickness and grounds used for actives). But as you said - 100 MHz would make things come out close to theory :-)
Sorry you're having difficulty viewing it. I checked and it says visibility is public. It also has 1700 views, so maybe this is some current RUclips error ?
![Lethal Company Music Video 🎵 "Follow The Orders" [VERSION A]](http://i.ytimg.com/vi/QeooZB4Z7RA/mqdefault.jpg)
The clearest teaching style I've come across for the VNA. The calm delivery and pace is ideal for learners.
Thanks!
OUTSTANDING instruction! Best video I"ve seen this year. Should be required viewing for all amateur operators. 73
Thank you so much for the kind words. Amateur radio is what got me started in electronics so many years ago (after a brief introduction with a 100 mW CB Sears base station I got for Christmas as a pre-teen). I don't operate a lot, but I learned so much from getting licensed and studying books published by ARRL. It's nice to hear this is helping to give back some to the community. 73's
A+ tutorial quality. Highly recommended for everybody who is into understanding high frequency circuits.
I like the important details this guy goes into. A much needed explanation for RF design.
Greetings from middle-aged men in China (although it might not be very popular)! I became fascinated with electronics two years ago, I believe its not too late. Thank you very much, and admire your effort in education after retirement!
Thanks for the greeting ! Glad you enjoyed the videos. Yes - its never too late. I'm learning biology and bio-chemistry in retirement - now that I have more time. Best wishes from the US!
Great video! You're an absolute champ explaining these in simple terms.
Thanks !
Stray Capacitance, Stray Inductance, ghosts in the machine. The little VNA's have become a real treasure for amateur electronic experimentation, and obliviously an asset even to the more experienced electrical engineer.
Well said !
This is an amaZing tool and discussion of parasitic as well as TV or high speed digital signals.
Great Video! I appreciate you taking the time to pass on your knowledge. I will have to watch this one again.
Glad it was helpful!
Amazing video, and very complete. Thank you for taking the time to create this!
So far I believe you do the best job explaining the use of the nanovna and tinysa. PLEASE keep it up!
Thanks, will do!
@ Harold Lacadie. Totally agree with your comment.
nano vna keeps on giving. cheers for the series
Agreed. Its an amazing instrument, with so many applications!
My friend, this is amazing. I'am learning english and I love electronics. In your videos, I find both of those to be high quality. Be sure I'am using my NanoVNA today. Many thanks. :-)
Thank you for the very nice comment. Enjoy the NanoVNA. It is an amazing instrument !
Excellent explanation of the basics. Thumbs Up!
Thanks. Glad it was helpful !
Brilliant video and explanation of the black art of RF design. Thanks. Just FYI, the nanoVNA-F V2 goes up to 3GHz and can use 301 plotting points. The v0.5.0 firmware doesn't include thru calibration though this isn't really a disadvantage.
You're welcome. And thanks for the info on the V2 !
Thanks for making this, it made a lot of sense and I learnt a lot. 👍
Glad it was helpful!
saw another RF youtubers video where he recommends component measurement should be considered at point where Reactance equals source impedence(50 ohms for NanoVNA), at 90 degree points. Makes sense as it will form a proper voltage divider at that point.
That's reasonable, but the actual low-frequency value of inductance (or capacitance), is actually slightly off at those points on the Smith Chart. For an inductor, anywhere along the arc on the left side of the chart should give about the same value. For an inductor I usually go to a point about halfway between the 90 degree point and the far left where the impedance is zero. (except don't take it too close to the 0 Ohm point, since accuracy will fall off there of course). For a capacitor, it's the opposite. One needs to take the measurement on the arc between the bottom 90 degree point and the far right of the chart where Z goes to infinity as frequency tends to zero. So I think the 90 degree rule is pretty good.
great vid, keep up the great work
Thanks! There's a few additional videos and playlists available. See the channel's associated website at ecefiles.org (or directly on RUclips at www.youtube.com/@MegawattKS ) Thanks for leaving the kind comment.
Great lessons, thx!
Hello my friend, my name y Jeorge from Medellín Colombia.
Many thanks for your video, it is great.
Congratulations, I'll listen to you later.
Hi Jeorge. You are very welcome. Thanks for leaving the comment.
Awesome video, dude. Thank you.
Thank you for this tutorial, Its very clear and very well explained. 73'.
Great overview video. Careful though when measuring large or small impedances with an S11 measurement, the VNA gets a lot less accurate then. The "testing" version of nanoVNAsaver has some nice new features for measuring impedances using an S21 measurement. That would help improve accuracy of the measurements quite a bit and you may find that your high resistance test will give you a somewhat different result (though still having quite some effect from parasitics)
Good point - especially for the high R values well above 50 Ohms. I haven't tried the S21 method, but did write a paper a couple years back about measuring Q. We found that with HP8753 units at least, the measurements can be quite decent if the measured impedance is close to the cal-points (0 and infinity), and worst around +/- j50. But I don't know if the Nano uses the same calibration equations...
Hi... I really enjoyed this very educational video regarding NanoVNA measurements of RLC components. I'm learning how to use the NanoVNA to measure LC circuits and stand-alone inductors. Your analysis piqued my curiosity about how to interpret the measurements. My question concerns the measurement of the self-resonant point of an inductor. When you showed the model schematic, we saw R(big) to be in parallel with L, which I took as the equivalent parallel resistance exhibited at resonance by the parallel LC circuit formed by the parallel combination of the L and parasitic capacitance. So near the self-resonant point, the apparent value of R(big) increases, as shown in the demo. i think you mentioned that R is the series resistance. The value of R(big) was around 2.6K near the resonant frequency, which seems rather high for a series resistance for the coil. I think the VNA is measuring the resistance it sees, which in this case should be the equivalent parallel resistance R(big). If this interpretation is correct, this is great way to measure Q, which I calculated to be about 7 based on the measurements. Thanks again, looking forward to future videos on using the NanoVNA.
Hi Scott, You have hit on some very important points! The extraction of model element values and the analysis of Q is indeed tricky near self-resonance, and even in the technical literature there are some inaccurate approaches and innacurate measurements reported. I like your focus on the self-resonant frequecy (SRF) and what happens there and agree that what you are seeing at self-resonance is the parallel loss resistance Rbig - not the series R. That's where the literature can be misleading, and you have correctly realized the complexity. Here goes for an attempt at summarizing (I'm neglecting the 15nH/in interconnect L in the model here): At low frequency (well below the SRF), one can just do Q=X/R where X and R are the imaginary and real parts of the impedance reported by the VNA. In this regime, we can delete R(big) from the model and just use the series R. But starting somewhere around 1/2 of the SRF, this simple calculation breaks down and the model needs to include the parallel R(big). Indeed this then gives Q=0 at the SRF since the measured value is purely resistive as you said (the SRF is where the VNA curve changes from inductive to capacitive on the right side of the chart). But there is still a resonance, so clearly there is non-zero Q ! So in this regime (between frequency of about 0.5SRF and SRF) a different approach is needed. I'll follow-up with this thought in the next reply to keep this one reasonable length... -Bill
Here is what I have done to extract more meaningful Q values between frequencies of about 0.5*SRF and the SRF: 1) delete the series R from the model so that it simplifies to a parallel RLC circuit. Take a measurement around 0.2 SRF to find the reactance and compute the L value from that. 2) Use this low-frequency L value together with the SRF to compute the C value for the model. The next step is where the math requires complex number crunching, so forgive me if I don't explain it well: 3) Take measurements of Z=R+jX at the frequencies of interest between 0.5 and 1.0 times SRF. Convert this impedance Z to an admittance Y (Y=1/Z) and then subtract off the admittance of the capacitance found in step 2. This gives an admittance Y' of just the R(big) in parallel with L. The real-part of Y' is then equal to 1/Rbig. 4) if Q is desired at this frequency, use the parallel-mode Q formula Qp=Rbig/X_L where X_L is the reactance of the inductor in the model at this frequency 🙂
@@MegawattKS Thank you Bill. This is very instructive. I read your next reply, and will comment there.
@@MegawattKS Bill - my interest in all this stems from a recent experience I had measuring inductance of large air-wound coils having moderate inductances. In one case, I had a loading coil for an antenna of about 150 uH consisting of about approximately 100 turns of enameled wire on a PVC form. i made this quite a while ago, but recently I used a NanoVNA to sweep from about 50 kHz to 10MHz to measure the inductance. I expected to see a normal Smith chart plot for an inductor in this frequency range, but this coil has a self-resonance at less than 2 MHz, well below the intended operating frequency of about 3.7 MHz. I know it is most likely due to parasitic interwinding capacitance that is rather large, because the 100 windings are right next to each other, and wound on PVC that has a high dielectric constant of about 3. However, I was not sure if this is an artifact or a real effect. I made two such coils to be used as loading coils to resonate a 40 meter 1/2 wave dipole in the 80 m band. I guess my point is that theoretically, this coil is made to the desired inductance geometrically, but its frequency response is another issue and not considered in the design criteria. Would this inductor work as a loading coil then if it turns into a capacitor at the frequency of interest? If it does work as an inductor at the operating frequency, then does the parasitic capacitance get swamped out by the antenna capacitance? I don't think so, because I calculated that the parasitic interwinding capacitance should be about 70-80 pf, which is on the order of the antenna capacitance for a 40 m dipole operated on the 80 m band, which the coil is to null out (e.g., say about 10 pf antenna capacitance), unless the capacitances are in series, in which case the add in parallel. The smaller one dominates, as you pointed out in the video. If you have any insight to this dilemma, I would love to hear your feedback. Thanks again for your great video series on all the subjects you cover on your channel. Scott
☺
@@scottgilbert7927 Hi Scott. I'll take a stab at this. I have experience with the theory, but no direct experience designing a loaded dipole. My first thought is that 150 uH sounds very large for the goal of loading a 1/2 wave 40 m dipole to work at 3.7 MHz. But where are the loading coils going? At the feedpoint or somewhere near the middle of each element? Maybe 150uH is correct for the latter case, but it's j3500 Ohms at 3.7 MHz, which is hard to realize. Assuming that's the goal, then I think the interwinding C is the culprit as you said. And the VNA is telling us its a capacitor, which it is. It won't work as an inductance load. Maybe try winding the coils with wire with some thicker insulation to cut down on C (though the coil length might grow of course), or use a larger diameter PVC form to decrease the number of turns needed? The other option I can think of is loading each side of the dipole at the feedpoint. I think the antenna you describe (40m dipole used at 80m) looks something like Z = 73/4 Ohms - jX with X maybe a couple hundred Ohms (maybe check with EZNEC simulator?). In that case, you only need L = X/(2*pi*f) = 8.6 uH. Of course you'll then need to build something like an L-match network to convert 73/4=18 Ohms to 50 or 75 as needed for the transmission line. Hope some of this helps. Great problem and application of the NanoVNA !
Interesting. I was always been told that modern high stability resistors have the resistive material wound around the body like a coil. They are trimmed by laser to match tolerance etc. So it's the resistive element is inductive. Old carbon resistors dont have the problem that modern resistors do. Wire wound are highly inductive. Non inductive resistors can be purchased. Modern resistors dont suffer with carbon migration when used on DC voltage. But this is what I have been taught over 30 years at Tech College but things may be different these days. Cheers from old George
Hi George. Now that you mention it, I think I heard that too. Not sure how many resistor formulations use that technique. But it's important to remember that even if they don't, there's tons of inductance (e.g. 20 nH per inch) for just the resistor itself (it has non-zero length), plus the length of any interconnects between it and a capacitor, etc. As you said, wire-wound resistors can also add substantial inductance - being basically coils - in addition to their own length and interconnects on a PCB/etc. Cheers.
Thanks for the amazing lecture!
You're very welcome!
Excelent video!!!
Thanks !
Interesting. The RF world is very strange. Thank you
You're welcome. It really is! One of my favorite sayings is "Amplifiers oscillate and oscillators don't" :-) Often due to these gotchas, I imagine...
Fantastic video!
Glad you liked it!
Excelente señor!!!!
DO you have a website where one can look at your vast collection of handouts and cool collection of radio schematics?
Sorry - not yet. But I should consider it. Thanks for the idea.
Excellent video 👌
Thank you !
The first VNA was not by HP, but by Rohde & Schwarz at the beginning of '50s: the ZDD. It is also a well known and sought after instrument between collectors.
Thanks! Can you provide a link to information on it? I can't seem to find it with a Google search on "Rohde & Schwarz ZDD" or related things. I did find an IEEE paper in ARFTG that talks about history, but it only says the 1964-released HP8410 was the first widely used one. It mentions some homodyne units from Hazeltine from the 50s, and Wiltron and Rantec in '63. ieeexplore.ieee.org/document/4119938
This is good stuff. Thank you for sharing :-)
Excellent! Thank you!
Glad you enjoyed it :-)
I'm very interested in using a VNA for the building and adjustment of toroids used in HF low pass filters. At the moment I'm working on a diplexer type filter design. Ideally, I need an instrument that will provide 1nH resolution. Do you believe the NanoVNA is suitable for that task? Perhaps you know of a better suited - ideally inexpensive instrument?
Technically the answer is yes - if you're below about 100 Ohms reactance at 14 MHz for example (100 Ohms is about 1000 nH there and the display shows 3 significant digits in the marker readout). BUT - this is resolution - not accuracy. Accuracy of a VNA is not nearly that good and depends heavily on whether you're measuring small to moderate impedances, or large ones closer to the right side of the chart, where things get really 'off'. As for alternatives, at VHF and above, a VNA is about the best one can do, without creating special test fixtures that apply resonance with a known capacitor value. At HF, there are somewhat higher accuracy methods without resonance use, but they're expensive - and frankly seldom worth messing with, IMO.
Bravo! Vy gud video.
Thank you.
Thanks!
Thanks a lot for posting these very nice VNA videos!
Thanks for the feedback. Glad you like them!
I am curious if you had tried the Inductors(uH range) on small toroids and observed them on VNA.....i unfortunately started with a inductor on FT37-43 while being a new user on NanoVNA. Took me days before realizing how bad the inductor on small toroid could be with load of stray capacitance, it was almost a diagonal line on the Smith chart instead of being a circle.
With air core inductors(nH range) its not as bad to get it right, it forms a decent circle.
I just looked up the -43 core and it appears to be Nickel-Zinc, and claims to have a high Q (at lower frequencies - e.g. below 30 MHz). Try resetting the Stimulus frequency range to 0 to 30 MHz (or even lower frequencies) and then recalibrate it and try the measurement again. I think what you may be seeing with the diagonal line is the NanoVNA taking only a couple datapoints in the low frequency range and just connecting these points with a straight line. (if it's sweeping 0 to 1000 MHz and only takes 100 points, then the frequency step is 10 MHz.) I've seen it do this, and had to drop the frequency range and recalibrate to get it to take enough points to make an arc as it should from 0 to 30 MHz.
@@MegawattKS I had it sweeping 1-30Mhz...what i meant by diagonal line(exaggerated to make point) is that the Smith curve was kind of going all over the uH range instead of making anything that resembles a circle. I had other known value inductors(wound on bobbins) tested to double check if something wrong with calibration, they were working fine.
thanks
Really appreciate your present at but I have a question I can’t seem to get answered.
In this video, you showed how (with an inductor) when you raised frequency the inductive-reactance increased-so good.
In a couple of videos W2AEW & Gregg Messenger, they showed (when connected to an antenna) that as the frequency increased, the inductive-reactance decreased by rotating downward from the upper hemisphere past the Real-axis into the capacitive lower hemisphere. An antenna will always look more-and-more inductive as frequency is increased (i.e. antenna too long); I would think the vector should rotate upward as frequency increases as it did for your inductor. Tnx & 73…
Hi. Their videos sound correct. It turns out that dipoles (or monopoles or most other antennas) actually do go through periodic resonances. At each resonance, it transitions from capacitive to inductive, or inductive to capacitive. Actually inductors will do this too if measured at high enough frequencies (above what's called their self-resonant frequency). Here is a webpage that shows a graph of real and imaginary parts of an antenna's impedance versus frequency (reactive, imaginary part in blue). They didn't label it, but the horizontal axis is frequency in MHz for the 30m monopole analyzed. But dipoles/etc will do the same thing. Hope that helps. 73 www.siranah.de/html/sail018r.htm
@@MegawattKS well thank you for taking the time to both comment and providing the link. I did get a response from Alan saying that the measurement-point is at the coax, not the antenna unless it is at exact 1/2 wavelengths.
I must be honest, I’m still trying to get my head around this still (since it always seems to go in the same direction) but will try some experiments to hopefully gain greater insight so, we might get with you again.
Thanks Again & 73…
At 14:36, you were measuring a 9100 ohm resistor with the nanoVNA, I was wondering if you were using the S11 method? I think that 9100 ohms is at the edge of the accuracy of the instrument, and switching (although, not easy with this setup) to the series S21 and converting to S11z... Copper Mtn website has some interesting info on the VNA accuracy for various Z ranges.,.. I have indeed re-learned quite a bit from your videos, I've enjoyed taking some of the formula and coding them into Python as calculators... Thanks for you time sharing this info... Oh, I see Rjordon was all over this....
Good points. Yes, I was just using the S11 method. I'll have to look at the Copper Mtn site and see if they found the same results we did relative to improved accuracy near the short and open points on the chart after calibrating for S11 measurement using SOL references. Thanks for the comments!
i've seen alot of videos on this, i'm getting one. great video. why would you spend 400 dollars for an antenna analyzer, when you can spend 100 and use the nanoVNA and can also use it to test filters, amps, etc. 73s
Agreed ! (Assuming one has good enough eyesight ;-)
@@MegawattKS I attach my nanoVNA to the PC and have a wonderful 24" (diagonal) display. Not great for field work, but wonderful for classes.
@@joeb3300 Excellent. I used to pull a bench unit with a VGA output, stick it on a cart, and hook it to the classroom projector - but that would be easier for sure :-)
A+ info, pace, and presentation quality. Immediately upped my game, thanks! Where did you get the test boards? Can you provide a link to the Radio Class handout? I'd find them handy when teaching other hams. Edit: I now read that others have asked the same question going back a year or more. I understand the overhead of creating a site for distro. If these came from a class, can you point us to the folks who put on the class? I wondered if you rolled your own test boards. I also use U.FL connectors for their ease of handling and quick connect/disconnect. When handled carefully they last a long time and give good results.
Good to hear. Thanks for letting me know it helped! The boards that the radio was built using are ones I designed and fabbed through ExpressPCB (I like their simple free layout software). I agree on the U.fl connectors. We used those in the microwaves course and they are fine for repeat connections :-) The course itself that inspired these videos is one that I and others did at K-State. But we're all retired now, so to keep it alive, I did these vids. It's still on the books, but not actively being taught at the moment. Here's a related K-State web page. It's got some stuff that might help (e.g. the last bullet in the first section takes you to the class handout sheets). ece.k-state.edu/about/people/faculty/kuhn/
@@MegawattKS FYI the link seems to be down. Excellent presentation BTW. You have a new subscriber!
@@andros2112 Thanks ! I think I've got it back. They changed the URLs of faculty/staff last month by putting an 'about' into the URL path. Hopefully this page will stay intact. I'm currently lobbying for the administration to not delete emeritus faculty pages (or at least this one. Hopefully the department at least will keep the content, if not this exact URL...
Hi thanks for your video, one question, the ground plane must be connected to gnd terminal or it's isolated?
Yes. Should be connected to ground.
@@MegawattKS Thank you
WHERE FROM CAN ONE ORDER THESE CONVERSIONS AND ELECTRONICS FOR NANO VNA ? KINDLY THOMAS KALLMYR UDDEVALLA SWEDEN
The "RF Demo Kit" shown in the video was purchased from Amazon in the US. But it is basically sold by DeepElec. I'm not sure what stores you have access to, but you can also find it and other items on the DeepElec site on AliExpress here: deepelec.aliexpress.com/store/1101363542 And details on it, including some excellent close-up photos, are on the Deepelec site here: deepelec.com/rf-demo-kit/
Information on the custom PC boards shown in the video is also available here: ecefiles.org/rf-circuit-prototyping/
can this be used ad rf pwoermeter, to measure the max power output of rf module?
No. There are a couple reasons. It's meant to look at "stimulus-response", for things like measuring gain (or 'return-loss'). Without the stimulus part of that, it cannot frequency/phase-lock, so it will not read the signal. The other critical reason is that it could be damaged by anything more than about a milliwatt of power. (actually, the damage level may be 10 dBm, but that's still 0.01 Watt max). I have noted on Amazon that there are a number of small low-cost power meter modules out there now for similar price range (e.g. 50 dollars) with a simple LCD display. But again, they're meant for very low powers, well under a watt.
@@MegawattKS than you for your reply. I want to measure max power output of radio transmitter, which has max power output 20dbm in data sheet. in data sheet this via has max 10db range , by adding attenuator I cannot measure ? if not can you also provide a link to pwoermeter please.
thx
@@tomipiriyev The NanoVNA still can't do it, since it needs to look at something which is 'coherent' with the signal it puts out. (it could do an amplifier since it could provide the input, but for a transmitter, the signal is coming from somewhere else). What frequency are you wanting to do this at? If it's below 900 MHz, then you might want to look at getting one of the really cool "TinySA" spectrum analyzers (and use an attenuator as you said). But if its WiFi like 2.4 or 5.x GHz, you may do better with something like this (I'm not recommeneding this particular product, but it gives an idea maybe): www.amazon.com/Taidacent-Calibration-Measurement-50-0dBm-Attenuation/dp/B07JMKQTDW/ref=sr_1_3?crid=27JYNJNFYW5W8&keywords=rf+power+meter . I don't know how accurate it is though...
@@MegawattKS tiny SA is different than nanoVNA? I mean working principle, I am a little confused.
I will measure 2.4Ghz .
@@tomipiriyev Yes - the TinySA is different than the NanoVNA. The TinySA is a spectrum analyzer (showing power levels of signals vs frequency, or power density due to modulation vs frequency). The NanoVNA is a network analyzer designed to measure things like gain (ratio of size of output signal to size of input signal) vs frequency. The TinySA is closer to what I think you're asking about, but unfortunately it does not go above about 0.9 GHz, so it won't work. For 2.4 GHz, the cheapest and best route to just getting the power level is probably something like the power-meter board in the link (together with a power supply and 20 or 30 dB attenuator). And it won't have the same issues of reading inaccurately due to wideband modulation that a spectrum analyzer would.
5:25 what was the flashing frame?
Just a video editing glitch. I do a lot of "slash-and-burn" editing (throwing out parts of what is filmed to tighten the story). I tried to pause it to see after you pointed it out - and as far as I can tell, there seems to be a frame or two with one of the "spinny arrows" that Pinnacle Studio sometimes generates for unknown reasons. They seem to get generated during the video export process. I upgraded recently to a newer version, so hopefully there will be less of these types of glitches :-)
@@MegawattKS Shotcut is an opensource program for video editing and transcoding. Could be helpful, too :)
S-parameters please:
measure components (passives) with nano vna Smith Chart(r +/-jw),
put into circuit (simple),
predict response with S-parameter algebra.
100Mhz would be perfect for this.
Then: talk about signal graphs,....
Good ideas. That's currently deeper into math than I've been doing in these - but I'll try to think about this for down the road. I agree that this would work best at 100 MHz. Too often S-parameters at microwave frequency vary too much with the exact test fixture used in the reference design/measurement (e.g. PCB thickness and grounds used for actives). But as you said - 100 MHz would make things come out close to theory :-)
Thank You for a very cogent Lesson!
You're very welcome. Thanks for leaving this feedback. It is very much appreciated.
E=IR
60 and 130$? It's $16 on ebay WITH shipping!
Why do you bother posting these when they are declared PRIVATE?
Sorry you're having difficulty viewing it. I checked and it says visibility is public. It also has 1700 views, so maybe this is some current RUclips error ?
If it was private, it wouldnt have showed up on my feed like an easter egg
YOU DON`T HAVE THE TIME TO READ AND ANSWER THESE QUESTION DO YOU.....KINDLY THOMAS KALLMYR UDDEVALLA SWEDEN.
Please see answers in the other comments below. 🙂
E=IR