NanoVNA and TinySA for Radio Design
Vložit
- čas přidán 8. 07. 2024
- Using the NanoVNA and TinySA to illustrate how radio / wireless devices work. This video concentrates on showing the front-end filtering and amplification in a superhet FM broadcast band receiver design. It also overviews some key instruments that have become reasonably affordable in recent years, allowing a home RF lab to be assembled for a few hundred dollars.
- Věda a technologie
I've returned to this series after a break and I'm really glad that is is as good as ever and is teaching me a lot. You don't fall into the trap that a lot of experts do when they teach. That is assume the student has more knowledge than they actually have. When you have been doing something for many years, it's easy to assume that one's knowledge is common knowledge. You don't fall into that trap, thank-you.
Thanks ! I'm definitely glad to hear that it's the right balance. I totally agree with the need to stay with the audience, and try to constantly think about what the listener knows and might not know at any given time.
Just wanted to say many thanks for the serial. I am very much enjoying watching it, as I just starting developing interests for radio technique and setting up my workbench.. 👌
You're very welcome. And thanks for leaving the comment. It is very nice to hear that its helpful.
Thank You, I enjoyed your video.
Excellent video, very practical and objective ! Thanks for sharing! Looking forward to the next video!
Thanks ! Glad it was helpful. Next one will probably be on bandpass filter design, and then maybe one on matching networks. I've been slow lately. Seems like a month went by between the one on RLC component parasitics and this one. Hopefully I can get the next one out sooner :-)
thanks for the video! I'm a ''maker''. So it's sometimes hard to find, the correct quick tech-info! YUP! technology products are definitely becoming cheaper, and a lot easier to access. this really gives me the ability, to make my product a lot quicker! I'm always delighted to see Engineers, bring something to the CZcams table. especially something that can help me advance skill-set's in life!
Thanks for the feedback ! I'm glad you found it helpful. I need to get working on the next one which hopefully will be helpful as well. The plan is to discuss how to design bandpass filters for particular frequencies and bandwidths after assessing the "Q" of the components :-)
thank you i like it
Forgive me! I am just seeing this vid for the first time. I am a career RF/Antenna/Microwave Design Engineer. I really do appreciate some of tne new stuff coming out (I have seen some of this stuff for the past 3 - 5 years). It is exciting to see VNA's, for example, which are VERY affordable.
They are, albeit, somewhat limited in both frequency & capability. If you are designing just a bit beyond WiFi/Bluetooth, these things are great. If you are trying to design higher freq stuff into the S band & above, or several different types of receivers & RF filters, then a more complex system may be required. Although, I have purchased 2 of the NANO VNA's and set them up using MATLAB & Python to get the full S-parameter (S11, S12, S21, S22) measurement for filters & some other mixer type responses. It isn't easy but can be done. Higher freq stuff (over 6GHz) will require a better VNA, however.
Overall, it is very encouraging to see some younger engineers out there using their gifts to bring some of us not so entrepreneurial-types useful and very economical tools!
Totally agree. I do miss not having built-in attenuators to characterize active circuits, and the number of frequency points is a little limited in a single sweep, but overall the NanoVNA is amazing. I've been tempted to get the 6.3 GHz one, but haven't yet. On a related note, have you seen the "TinySA Ultra"? Erik and team extended the TinySA spectrum analyzer to go to 6 GHz now (though the spurious-free'ish region is only through 800 MHz I think). And it has signal generation capabilities too, and output level control down to about -100 dBm. I have two of the original TinySAs. One for spectrum analysis and one for signal generation :-) Amazing stuff !
@@MegawattKS I have looked at the TinySA Ultra, but I already have an HP8592B SpecAn. Still, it might be interesting to get one and compare results.
I bought the HP years ago from a salvage company who was liquidating a lot of Motorola stock of test instruments. It goes up to 26GHz, so I am good for the freq band of interest. The sad part is I was also trying to purchase an Agilent 5071A VNA from the same company, but the bidding got WAY out of my budget. I think someone purchased the unit for around $18k. They got a good deal for that price!
@@brianweber1973 Cool. We had an 8592L (22GHz lower-cost version) at school that we used (and still use) in the microwaves-and-antennas course. It's a solid performer! The nice thing about the TinySA units (if one doesn't need to go into the GHz spectrum) is how portable they are. Even at $60 for the base model, they're useful for around-the-house propagation measurements and RFI investigations, checking cable-TV signal levels, generating RF signals, and such :-)
Excellent Video. Thank You! I have a question. I've seen notes on some systems where they talk about the 'system bandwidth' what are they referring to? the pre-select filter BW or the IF filter BW?
Thanks! The answer to the bandwidth question kinda depends on context. There are at least 3 different cases I can think of: 1) the pre-select as you mentioned, which determines vulnerability to interferers, 2) the IF filter you mentioned - which is classically equal to the spectral width of the signal so it minimizes noise without hurting the signal information, and 3) the final information bandwidth - which is dependent on the modulation type, and data-rate if it's a digital system. My guess is that they're referring to the final IF bandwidth which matches the modulation's spectral width. But you'd have to look at it in context of the particular system and what's being discussed. Hope that helps. 73's
how about putting links to the stuff your list in your video description..... thanks for this video.
You're welcome. I'll try to make a list here in this reply so they can be Googled to see options and details (don't want to favor a particular site or vendor/distributor - more fun comparing the offerings and seeing new stuff when searched :-). Here goes - starting with the simple ones: "NanoVNA" (lots of different versions - mine is the 4" screen NanoVNA-F), "TinySA" (some different types, but mainly different places to buy), "RF Demo Kit for NanoVNA-F - by deepelec" (The green board with all the cool Smith Charts and components), "NanoVNA Testboard Kit" (the little blue board the filter was built on), "Interstellar Electronics SMD Professional Assortment Kit" (the surface mount parts box), "Inductor Sample Book" (goes by various names/vendors - but that should pull up useful hits), " SMA RF DC-Block DC to 6 GHz 50 Ohm", "U.FL to SMA Male Coaxial RG178 Low Loss Cable", "Nooelec SMA Attenuator Kit - Bundle of 6pc 2W 50 Ohm SMA in-Line Attenuators", "6.5ft Low-Loss Coaxial Extension Cable RG58 (50 Ohm) SMA Male to SMA Female", "Digital Multimeter, ACEGMET", "ADS5012h Handheld Digital Portable Oscilloscope", "50PCS IPEX U.FL SMD SMT Solder for PCB Mount Socket Jack Female RF Coaxial Connector". Hope that's helpful.
You mention a course you used to teach - any of that material online anywhere?
Thanks for the video! I haven’t picked up a tiny sa yet (been making do with the oscilloscopes fft and a rtl-sdr dongle), but have been eyeing now that I’m getting back into homebrew.
The only on-line version of that course material so far is on this "MegawattKS" CZcams "channel". I've selected a handful of favorite lecture/demos and tried to adapt them to a general audience, with minimal assumptions on prior math background/etc. Instead, they are focused around the NanoVNA. But much of the course content is sorta there. I always resisted doing "distance-learning" versions of the classes due to their lab nature. But now that the NanoVNA and TinySA are out there, it's getting more practical for sure :-)
I'm definitely impressed with the TinySA and would recommend it. I got mine through Amazon and I think I paid more than the $55, but supposedly you can get them at that price from other sellers. The performance of the unit is much better than I thought it would be. And for the 300 MHz band, the internal architecture is excellent (multi-pole LPF followed by upconversion to 433 to put a tight filter on it - and then downconvert, digitize, and DSP filter). Pretty similar to the big-boy architectures, which probably accounts for its reasonably spur-free performance ...
@@MegawattKS Thanks!
Very interesting video. Can you tell me the name of the very small connectors that appear in the video at 5.00? They are shown in a strip .
They are U.FL connectors - developed by Hirose. I Googled U.FL and there's actually a Wikipedia page on them if your curious for more info - including some other names for them :-)
@@MegawattKS They are also called IPX and IPEX.
Random question: A typical signal level input to a receiver might be 50 micro-volts. If this is so, then how does a crystal detector work when the diode has a .3 volt forward drop? It seems like the voltage is about 1/6000 of what is required to make the diode conduct.
Good question. A few things that I can think of combine to make this plausible. The main thing is that semiconductors have a sub-threshold exponential behavior - which is still non-linear. The forward drop depends on amount of current (or visa-versa). In addition, there is some voltage step-up (maybe 10x or 20x) from the tapped tuning coil that does an impedance transformation from 50'ish Ohms to a couple kOhms that the earpiece presents at the output. Finally, the earpiece is right in the ear and the ear has about 100 dB of dynamic range. So very little RF power is needed to get enough audio power to hear it. 50uV at 50 Ohms works out to -73 dBm (if I punched the calculator correctly), which is maybe 70'ish dB below max volume (1mW directly into the ear) - so it's within the ear's dynamic range 🙂
One other thing. Even thought the points above make it plausible, it's still a push at 50 uV. Are we sure crystal sets can work at 50 uV? They're typically used for long-wave signals with very long wire antennas that can pull in a lot of voltage (from V = E * Length). Disclaimer - I've never actually used a (good) crystal set.
I searched for data on a Germanium 1N34 diode and found this discussion. There's some nice graphs that show the early cutin of some detectors electronics.stackexchange.com/questions/223813/the-classic-1n34a-ge-diode-has-a-vf-of-1v-how-could-a-diode-with-such-a-high
@@MegawattKS Thanks .... that was helpful!
Excellent video! Thank you for sharing your knowledge on RF.
Please, I would like to ask a question: if I needed to measure the frequency response and insertion loss of a filter that is not intended to work with 50 ohm source and load (for example, a Medium Wave AM pre-select filter, with a ferrite coil, which also works as a very high impedance magnetic antenna), would it be possible to do such measurement with the NanoVNA, as it's designed to work with 50 ohm sources and loads? If the MW filter were connected to the NanoVNA directly, the bandwidth would be much larger than the expected one in this case, because this MW AM filter should match a very high impedance rather than 50 ohms. I have tried to use an impedance transformation at the input and output of the LC resonant circuit by using two secondary windings on the coil (one for the input and one for the output), so that the LC resonant circuit sees a high impedance on both ends and matches to 50 ohms of the input and output of the NanoVNA. It's weird that If I connect only the S11 port to the circuit through one winding, the S11 reads out 50 ohms, but, if I connect the S21 port too to the other winding, then the S11 in this case reads out 25 ohms instead of 50, so it ruins my measurements.
Thank you!
Interesting. It sounds like it's sorta working - but just a bit lossy? Is the bandwidth roughly what you expect? I.e. narrow, but about a factor of 2 wider than it should be? If so, then I'm guessing there's just 3dB of loss somewhere. You also mentioned the coil is a loop antenna. Have you accounted for the expected radiation "losses" from that?
The few times I've actually tried to test a filter that wasn't 50 Ohms, I used either transformers (which I think is what you're describing essentially), or used high-value resistors. For example, we measured 330 Ohm ceramic FM IF filters using some simple 4:1 impedance transformers (didn't have 6:1). On a research project we did it a different way. We just put series resistors in at the end of the coax cables of the value we needed the high-Z (non 50) filter ports to be. Of course the S21 measurement has to be adjusted, and/or one needs to do an S21 cal with those in-place. There's a fair amount of dynamic range in the VNA (even the NanoVNA), to handle this. Finally, one might try an active probe - but that would only work on the port 2 side. Port 1 would still need a transformer or resistor...
Thank you for your answer and advice. I think I figured out what's going on: my set-up is a home made ferrite core coil with 80 turns to be used as a magnetic antenna to pick up MW broadcast stations, together with a variable capacitor to make up an LC filter to select the desired frequency. The Q factor of the coil is 30 at 1 MHz (not too high because I'm not using Litz wire and I'm probably getting losses due to the skin effect). So, I needed to characterize the LC filter to measure how it's frequency response to the MW stations. I imagine we could model the magnetic antenna as a voltage source with a series impedance of Rp, which is about 144 K (maybe this assumption is wrong, I don't know). So, if I connected a load of 144 K, it would be in parallel with Rp, which gives 72K... And I think this is what I didn't take into account before... Because, now, the source resistance would be 144 K, but the equivalent load resistance would be 72 K (just half)... And I think that's why the NanoVNA is reading out 25 ohms in S11(through the matching transformer) instead of 50 when both ports of the NanoVNA are connected. The BW I get is 92 kHZ, with -2.62 dB at the center frequency, and I guess they're due to the insertion loss of the filter.
So, in this case, I think the measurements now make sense. What I don't know if this is the right way of measuring an LC filter that also works as a magnetic antenna... In this case, would Rp appear as the parallel loss resistance of L or as the source impedance of the equivalent voltage source of the antenna?
Thank you and congratulations for such great videos on RF
@@user-pi1ih6wn3j I think that all sounds good, except you said series impedance of Rp. That should be "parallel impedance", yes? Honestly if you're satisfied with the selectivity as-is, I think the impedance mismatch is not a big problem. If I understand the setup right, then I think the answer is that Rp will be the parallel (Norton) resistance of the received signal source used in analyzing the circuit. But converting Norton to Thevenin, Rp is also Rsrc for a source model comprised of a voltage in series with resistance. Does that make sense?
@@MegawattKS Thank you for your answer. Yes, I think it makes sense now. I've managed to plug it to the BNC input of a receiver and the signal now in Medium Wave is strong!
I'll continue watching the video series to design the other parts of the AM radio.
Thank you.
So you could use these items to replace a radio service monitor
Possibly. But they don't have the certified performance or the user-interfaces those products have. (Full disclosure: I have never used a service monitor - so I'm just guessing).
Where did you teach radio design? Is it possible to see the classes on the Internet?
I was at Kansas State University (aka K-State). Yes. While the classes themselves are not on-line in the traditional way, I did abstract the key material from the course and put it on CZcams a few years ago after I retired. Search CZcams for "Radio Design 101" or go to the playlist here: czcams.com/play/PL9Ox3wpnB0kqekAyz6blg4YdvoEMoJNJY.html . There is also an associated website which has links to that and other related things, including educational material on antennas and propagation. It even has the full lecture notes I used when teaching the radio design course (the lecture notes are a little rough - but the videos are OK ) Here's the website: ecefiles.org/ It's not fancy, but the content is there 🙂 Hope its helpful.
Hi, is there some significant difference between 1206, 0805 and 0402 smd components in case of frequency response? I mean, a LPF wit 0805 will be the same as 0402 with the same components values? Thanks
It depends a little on frequency of operation. But the PC board trace inductive parasitics are probably more important than the component size within that range. Keeping a tight layout and using ground vias to the ground-plane properly (immediately next to the pad that is supposed to be grounded) is critical, especially above VHF. The one exception to this answer applies to SMD inductors. Small inductors (e.g. 0402) can have much higher series resistance than larger ones (0805, 1206, or larger air-wound). This leads to high signal losses and wider bandwidths that expected. That's why the circuits shown at 11:49 in the video used air-wound inductors. A high quality factor was needed for the bandpass filtering. That isn't usually available in the smaller SMD types because they use spiral-wound flat inductor designs with small, thin traces inside the chip. But you can buy air-wound inductors in SMD 1206'ish size that can have Q up to 50 or more. Hope that helps.
@@MegawattKS thank you very much for this answer. I'm developing a circuit with an ESP32(2.45GHz)using chip antenna with 0805 components because it is easier to solder. I already ordered a NanoVNA 2 from Bangood to try to make the pi matching circuit a little better. I calculated the pi matching circuit with an online calculator. I know that the Esp32 has an output impedance of 35ohm+0j and the NanoVNA is 50ohm. I will try to solder a coaxial cable in the pin of esp32 to try measure the impedance of the matching network. But, how to know it will work if both the esp32 and NanoVNA has different output impedance? I will solder the coaxial cable without the esp32 in the board. Thank you very much.
@@anlpereira 0805 components should definitely work, but again, make sure that the interconnect traces on the PCB are short/compact in the circuit itself ('transmission lines' leading to/from the antenna can be longer since they maintain 50 Ohm impedance if they are the proper width). See this video for a deep-dive into this topic: czcams.com/video/R0mRTigYzco/video.html&lc= I have not used the ESP32 myself, but based on some schematics I find on the web and your comment that the output Z is 35+j0, it should work fine even without a match. If you compute the "return-loss" or SWR for a 50 Ohm load on a 35 Ohm source (or the other way around), its actually a very good match. I.e. it should be better than -10 dB S11 even without a match. On a related topic, have you considered adding a bandpass filter between the chip and antenna? I see some people have done that, and depending on the level of interference you expect to see, it might be useful. But maybe that can be added later if testing reveals it might be needed. In any case, please pay close attention to VNA calibration and coax/cable connections to make sure the parasitics are well controlled to get good measurements. Enjoy the NanoVNA !
@@MegawattKS yes, I place three components in a Pi configuration. I saw in the antenna datasheet that I have to place a 3nH series inductor with two 0F shunt capacitors, but I think these values are considering a 50Ohm source impedance. As soon as I have the circuit and NanoVNA and make some tests I will share the results. Thank you very much again for answers.
Could you recommend a book on RF design?
There are several, depending on specific interests and what one's background is (e.g. how mathematical one wants to be). The one I recommended in the communication circuits class I taught was: "RF Circuit Design 2nd Edition by Christopher Bowick ". It has the advantage that it covers filter design in similar ways to what is done in this video and it is a good complement to the Radio Design 101 video series on this channel. And it goes into some areas in more depth - like series/parallel realizations of impedances underlying matching to complex loads. Amazon has a "Look Inside" for this where you can see the table of contents as well as read a good portion of the first chapter or two to see if this fits you're interests and you like the style. There are other texts of course that cover systems, microwaves, digital modulations, RFIC design, etc.
@@MegawattKS Thanks. I have the first edition of Chris Bowick's book. I found it excellent. The level was exactly right for me. I didn't know there was a new edition. I'll take a look.
Mike (in Germany)
I AM looking for your amP diy to increase My helium lorawan Antenna! Could You máke a vídeo about thks
Sorry - I didn't understand the question. Can you rephrase it? E.g. "increase My helium lorawan Antenna". I've been slow to get the next video out. It will be on filter design - but I've been considering making it the first in a series on design topics...
If I've deciphered your question correctly you have a helical antenna that you use for LoraWAN and you want to know if this amplifier helps you? LoraWAN runs at 868MHz if I remember correctly, that would be quite a bit beyond what the circuit here is tuned for
Do you have a paypal account? I like to send you some $$. I benefited a lot from your videos. You seem to enjoy being an educator.
There is so much demand to learn digital communications
I am getting into it as a hobbyist
I am a retired now.
Cheers from Palestine
Sorry - no. I just create these in hopes of contributing to the common good - after retiring from university teaching :-)