Hi All,
A couple of weeks ago I was looking for a way to make precision measurements of inductances and in this context posted a question about this in an inductance meter thread here on diyaudio.
One of the replies came from 1audio who suggested a nanoVNA Vector Network Analyzer which may cost as little as ~30 Euros (or maybe even less):
New 2.8 inch LCD Display NanoVNA VNA HF VHF UHF UV Vector Network Analyzer Antenna Analyzer + Battery| | - AliExpress
These nanoVNAs are available in different models ranging from (as far as I can see) 10 kHz up to 3 GHz bandwidth and somewhat different specs.
A couple of videos on how to use them can be found here:
#359 How to properly use a NanoVNA V2 Vector Network Analyzer & Smith Chart (Tutorial) - YouTube
nanoVNA - Measuring Inductors and Capacitors (Vers. 3) - YouTube
Apparently they can be used to make precision measurements of inductors and capacitances - and another diyaudio member mentioned that they might also be used for amplifier loop gain measurements - ... but I am wondering what else they may be used for? May they e.g. be used for precision phase noise measurements, or .... ?
Before maybe venturing into buying one it could be interesting to hear what the really practical uses may be for such a device ...
Thanks for any insights on this you may have
Cheers,
Jesper
A couple of weeks ago I was looking for a way to make precision measurements of inductances and in this context posted a question about this in an inductance meter thread here on diyaudio.
One of the replies came from 1audio who suggested a nanoVNA Vector Network Analyzer which may cost as little as ~30 Euros (or maybe even less):
New 2.8 inch LCD Display NanoVNA VNA HF VHF UHF UV Vector Network Analyzer Antenna Analyzer + Battery| | - AliExpress
These nanoVNAs are available in different models ranging from (as far as I can see) 10 kHz up to 3 GHz bandwidth and somewhat different specs.
A couple of videos on how to use them can be found here:
#359 How to properly use a NanoVNA V2 Vector Network Analyzer & Smith Chart (Tutorial) - YouTube
nanoVNA - Measuring Inductors and Capacitors (Vers. 3) - YouTube
Apparently they can be used to make precision measurements of inductors and capacitances - and another diyaudio member mentioned that they might also be used for amplifier loop gain measurements - ... but I am wondering what else they may be used for? May they e.g. be used for precision phase noise measurements, or .... ?
Before maybe venturing into buying one it could be interesting to hear what the really practical uses may be for such a device ...
Thanks for any insights on this you may have
Cheers,
Jesper
Last edited:
I have considered buying one of these as well, but honestly I was overwhelmed with the variety of offerings on AliExpress and didn't bother properly comparing what feels like at least a dozen options. The EEVblog forum or some RF forum is probably a better place for this topic, maybe there even is an overview or collection of these things somewhere. I haven't dug deep enough probably.
you can do a lot with a VNA. First keep in sight the input and output level limits and the useable frequency range. The one linked goes from 50 KHz to 300 MHz. That should be fine for most audio related applications.
With a VNA you can check frequency response, phase shift, create bode plots, measure losses in cables, crosstalk and shielding (limited by the systems dynamic range), measure resistors and reactive components, crystals, tuned circuits etc.
I would get one of these to explore but I already have this: Product Details | Keysight and just spent $80 on a pair of 75 Ohm to 50 Ohm minimum loss matching pads to make the 75 Ohm unit more useful. Once those arrive I'll try to show what can be done with a VNA.
What I think will be my next real tool will be a spectrum analyzer. Its clear there are EMI and RFI issues galore in digital audio stuff. The spectrum analyzer I have (Tek 7S12) is too big and clumsy to cart around to look for emi stuff. There are some really cheap analyzers on eBay but I'm not confident they will do what I want yet.
With a VNA you can check frequency response, phase shift, create bode plots, measure losses in cables, crosstalk and shielding (limited by the systems dynamic range), measure resistors and reactive components, crystals, tuned circuits etc.
I would get one of these to explore but I already have this: Product Details | Keysight and just spent $80 on a pair of 75 Ohm to 50 Ohm minimum loss matching pads to make the 75 Ohm unit more useful. Once those arrive I'll try to show what can be done with a VNA.
What I think will be my next real tool will be a spectrum analyzer. Its clear there are EMI and RFI issues galore in digital audio stuff. The spectrum analyzer I have (Tek 7S12) is too big and clumsy to cart around to look for emi stuff. There are some really cheap analyzers on eBay but I'm not confident they will do what I want yet.
Hardware Versions | NanoVNA V2
As I suspected there appears to exist somewhat of an "original" as well as more or less authorised copies or modified implementations. Right now it seems to be somewhat of a challenge to get your hands on the more recent hardware versions of the "original". But on AliExpress every second seller claims to be the official brand store or whatever, so always check at least twice before buying...
As I suspected there appears to exist somewhat of an "original" as well as more or less authorised copies or modified implementations. Right now it seems to be somewhat of a challenge to get your hands on the more recent hardware versions of the "original". But on AliExpress every second seller claims to be the official brand store or whatever, so always check at least twice before buying...
Last edited:
It appears that the official latest version NanoVNA V2 Plus4, 4 inch 4GHz Vector Network Analyzer is not available through proper channels: NanoVNA V2 Plus4 4 inch 4GHz Vector Network Analyzer|Oscilloscope Parts & Accessories| - AliExpress
Not available here either: 4" NanoVNA V2 Plus4 from HCXQS group on Tindie
Not available here either: 4" NanoVNA V2 Plus4 from HCXQS group on Tindie
Last edited:
I bought one of these for home use:
4" NanoVNA V2 Plus4 from HCXQS group on Tindie
The larger display and the metal cabinetry appealed to me compared to other versions. The shielding gives better dynamic range and gets rid of a lot of little glitches from radiated stuff.
The nanoVNA actually is surprisingly close in performance to the Agilent 8753 units at work. Way smaller, though. And gobs less expensive. No fan noise. You could buy several NanoVNAs each year just for what the yearly 8753 calibration costs.
The thing is, measuring to the last tenth of a dB (as you might believe you can do with an expensive VNA) with regard to transmission and return loss is sort of a fool's errand anyway. Even the 8753's drift that much over a few hours. Besides - what measurement are you going to attempt that require that level of precision? What would it matter?
I think you need to keep in mind what these instruments will do in general and get the one that will do what you want for your application. Generally, if you use SMA connectors a lot and terms like "S22" come up often in your casual conversation, then the VNA may be for you. If you want to see what the gain and phase margin is for that new power amp you just built, then a Bode type system as found in a lot of new oscilloscopes or Analog Discovery units might be better. They are somewhat interchangeable from the 10 km view, but one will be more convenient for your use than the other.
Keep in mind that VNAs are for closed system measurements. Apply a stimulus and look at the response. You might think that they can do measurements of other sources, but that really isn't fair to say. What you want for that is a spectrum analyzer. Think of a spectrum analyzer as a sensitive oscilloscope that displays in the frequency domain. A visual radio receiver, if you like. Note that it will measure down to microvolts.
You can spend lots of money on a spectrum analyzer,. But, in the same spirit of buying a NanoVNA, I purchased one of these:
TINYSA TINYSA
Shockingly great, especially for the price. It allows me to sniff about (with the right probes) to see what's radiating on a pc board, right down to the component level. Think your amplifier may be oscillating at 52 MHz? Want to be sure? Want to find out just how much high frequency crap is spewing from your audio file server into your preamp? This will show it.
No, it's not as good as a Keysight X-series analyzer, but the X-series analyzers are comparable in price to Audio Precision test systems. More or less. Too expensive for me. And too large. And too loud. A tinySA is plenty good for my needs.
Both of these gadgets are hard to get right now, because they use chips! Chips are in short supply these days.
Both of these gadgets will connect to your computer, so that you can have a larger display and more functionality. The software is free and there's a few choices.
Both of these gadgets get routine firmware upgrades that you can download and install for free. The upgrades to date have fixed the occasion bug and added a bunch of new features.
I'd check out the home pages for the developers of both of these. There's lots of really good information there.
Important: There's too many clones of both products out there in the world. Most really don't work nearly as well and are prone to failure. It's really worth it to spend the extra couple bucks to get the real deals.
One last thing... Measuring phase noise is challenging. Generally speaking, that means expensive if you want any sort of sensitivity and accuracy. Those X-series spectrum analyzers can make the measurements if you purchase the right options. But, even there you are limited with regard to close in phase noise. I personally think that you're better looking at a DAC's output in the audio spectrum instead.
(Wait! I lied! There's one other thing - here's a more general purpose gadget that I alluded to before. It might be more suited to your task: USB Oscilloscope and Logic Analyzer - Digilent Analog Discovery 2)
4" NanoVNA V2 Plus4 from HCXQS group on Tindie
The larger display and the metal cabinetry appealed to me compared to other versions. The shielding gives better dynamic range and gets rid of a lot of little glitches from radiated stuff.
The nanoVNA actually is surprisingly close in performance to the Agilent 8753 units at work. Way smaller, though. And gobs less expensive. No fan noise. You could buy several NanoVNAs each year just for what the yearly 8753 calibration costs.
The thing is, measuring to the last tenth of a dB (as you might believe you can do with an expensive VNA) with regard to transmission and return loss is sort of a fool's errand anyway. Even the 8753's drift that much over a few hours. Besides - what measurement are you going to attempt that require that level of precision? What would it matter?
I think you need to keep in mind what these instruments will do in general and get the one that will do what you want for your application. Generally, if you use SMA connectors a lot and terms like "S22" come up often in your casual conversation, then the VNA may be for you. If you want to see what the gain and phase margin is for that new power amp you just built, then a Bode type system as found in a lot of new oscilloscopes or Analog Discovery units might be better. They are somewhat interchangeable from the 10 km view, but one will be more convenient for your use than the other.
Keep in mind that VNAs are for closed system measurements. Apply a stimulus and look at the response. You might think that they can do measurements of other sources, but that really isn't fair to say. What you want for that is a spectrum analyzer. Think of a spectrum analyzer as a sensitive oscilloscope that displays in the frequency domain. A visual radio receiver, if you like. Note that it will measure down to microvolts.
You can spend lots of money on a spectrum analyzer,. But, in the same spirit of buying a NanoVNA, I purchased one of these:
TINYSA TINYSA
Shockingly great, especially for the price. It allows me to sniff about (with the right probes) to see what's radiating on a pc board, right down to the component level. Think your amplifier may be oscillating at 52 MHz? Want to be sure? Want to find out just how much high frequency crap is spewing from your audio file server into your preamp? This will show it.
No, it's not as good as a Keysight X-series analyzer, but the X-series analyzers are comparable in price to Audio Precision test systems. More or less. Too expensive for me. And too large. And too loud. A tinySA is plenty good for my needs.
Both of these gadgets are hard to get right now, because they use chips! Chips are in short supply these days.
Both of these gadgets will connect to your computer, so that you can have a larger display and more functionality. The software is free and there's a few choices.
Both of these gadgets get routine firmware upgrades that you can download and install for free. The upgrades to date have fixed the occasion bug and added a bunch of new features.
I'd check out the home pages for the developers of both of these. There's lots of really good information there.
Important: There's too many clones of both products out there in the world. Most really don't work nearly as well and are prone to failure. It's really worth it to spend the extra couple bucks to get the real deals.
One last thing... Measuring phase noise is challenging. Generally speaking, that means expensive if you want any sort of sensitivity and accuracy. Those X-series spectrum analyzers can make the measurements if you purchase the right options. But, even there you are limited with regard to close in phase noise. I personally think that you're better looking at a DAC's output in the audio spectrum instead.
(Wait! I lied! There's one other thing - here's a more general purpose gadget that I alluded to before. It might be more suited to your task: USB Oscilloscope and Logic Analyzer - Digilent Analog Discovery 2)
Bear in mind RF test kit is not designed for audio frequencies - the Nano VNA I have is rated from 50kHz upwards, which might make some measurements of components rather suspect. And of course everything is in the 50 ohm regime, and the further away from this value (larger or smaller), the less accurate readings will become.
Fantastically good value device for crude characterization of a VHF filter or amp or antenna, not so relevant for measuring an audio component!
Fantastically good value device for crude characterization of a VHF filter or amp or antenna, not so relevant for measuring an audio component!
192k soundcard = VNA from 20 Hz to about 80kHz, with tons of dynamic range and SFDR.
You can also make an impedance meter with a soundcard, among other things.
HF VNA is useless for high impedances, but high impedances don't exist at HF
How else would you know the impedance of your power supply, the phase margin of your feedback, or if your filters actually get rid of the noise?
2 clocks with slightly different frequency -> mixer -> soundcard -> python script -> phase noise
This downmixes the phase noise to low frequency so there is a lot more of it to measure than when you divide the clock. But if you want to divide it:
44.1k WCLK -> 192k soundcard (but you need a low phase noise soundcard)
Last edited:
A 50 Ohm VNA is a poor LCR meter if R, Xc or Xl is high. You will not get accurate results. HP sold a 1GHz 1 port VNA based Z meter a long time ago, HP4191. It was not accurate.
Newer 1GHz designs are Voltage-Current based. Inject constant current and measure amplitude and phase shift across component.
Newer 1GHz designs are Voltage-Current based. Inject constant current and measure amplitude and phase shift across component.
This all started with a question about measuring a small (1 uH) coil. The reactance at 1 KHz is zilch. it occurred to me that a NanoVNA at 10 MHz could do a decent job of quantifying small caps and inductors, for not much money. A vintage Tek 130 would cost more be less flexible and almost as old as us vintage audio people. More conventional LC meters are even more expensive.
An analog discovery can do more in some respects, less in others and it does cost more.
It all depends on what you are trying to do.
An analog discovery can do more in some respects, less in others and it does cost more.
It all depends on what you are trying to do.
For audio work there are plenty of cheap LRC meters that will generally suffice. As was stated you can also use a crappy PC, a few resistors and op-amps to measure Z or L or C. SW is free as well.
A VNA meant for RF use does a very crappy job of measure L/C if the Xc/Xl is say above 150 Ohms. To try and measure inductors or caps meant for audio use on an RF VNA is not recommended. In general you will be near or above the components self resonance.
A VNA meant for RF use does a very crappy job of measure L/C if the Xc/Xl is say above 150 Ohms. To try and measure inductors or caps meant for audio use on an RF VNA is not recommended. In general you will be near or above the components self resonance.
I don't think this is useful for audio. For an LCR Meter, the DER DE-5000 is great value for the money. It's particularly helpful for screening capacitor ESR when deciding to swap out PSU caps. It has kelvin probes and you can do LCR measurements at 100/120Hz, 1kHz, 10kHz and 100kHz. For ham radio work, this is also actually totally sufficient. But really the DE-5000 is meant for you to hand-wind small coils, test unknown parts, screen parts.
The VNA is revolutionary for teaching electronics though. If you want to expose people to principles at RF, this $50 widget is truly amazing. I can see this being used to illustrate concepts by visualization to high school students that normally wouldnt see this till college.
The VNA is revolutionary for teaching electronics though. If you want to expose people to principles at RF, this $50 widget is truly amazing. I can see this being used to illustrate concepts by visualization to high school students that normally wouldnt see this till college.
Actually for RF work, an LCR meter like the DER DE-5000 that doesn't measure at the actual operating frequency is not as useful as you may think. For RF filters, you need to know the parasitics of a component so they can be used in your modeling and confirmed with testing. An RF inductor measured at 100KHz will measure quite differently at 100MHz.
Unfortunately a VNA won't measure high Z components accurately. You need an old school Q meter or a modern LCR meter.
Unfortunately a VNA won't measure high Z components accurately. You need an old school Q meter or a modern LCR meter.
Hi all,
& thanks for all your interesting feedbacks
When reading them I am sort of thinking that it looks as if posting about this topic is like entering a Pandora's box - so many options and different approaches to "similar" functionality.
Now, since this is a topic (RF) that is not my main strength I am inclined to just see where the thread goes and pick up the various inputs you post here. However, just a couple of clarifications about the use I have for measuring coils:
- I, hopefully, soon will be making a couple of the University of York low phase noise oscillators at 22.579 MHz & 24.576 MHz frequencies. To this end I would like to confirm the inductance of the coils used here so as to make reasonably sure that they are the value they should be.
If I understand the feedbacks you have given correctly this should be possible with a VNA operating at 22.xx & 24.xx frequencies and then reading the inductance measured by the VNA. Would that be correct?
@peufeu:
Sounds like a very interesting approach to phase noise measurements ... Do you reckon that this setup would perform close to the performance of the so-called Timepod phase noise measuring device?
http://www.miles.io/timepod/ss_1pg.pdf
@Kn0ppers:
About those smaller VNAs I have seen them ranging from 10 kHz to the 4 GHz mentioned here so quite a span of frequencies ....
@CG & 1audio : You both mention a spectrum analyzer & CG also the Analog Discovery 2.
I have actually been considering the Analog Discovery 2 not least because it is also 14 bits resolution so it will show more detail than my current oscilloscope (200 MHz picoscope, unfortunately "tricky triggering" functionality). However, as far as I can see the Discovery's impedance measurement relies on the built-in signal generator which only goes to 12 MHz so it is a limited bandwidth.
But I am thinking that maybe a Discovery & spectrum analyzer could be a feasible combination. The first covering the lower frequencies and the other the higher frequencies ...
But maybe there are even better solutions - e.g. better accessible spectrum analyzers out there?
Thanks again for your feedbacks & have a good weekend,
Jesper
& thanks for all your interesting feedbacks
When reading them I am sort of thinking that it looks as if posting about this topic is like entering a Pandora's box - so many options and different approaches to "similar" functionality.
Now, since this is a topic (RF) that is not my main strength I am inclined to just see where the thread goes and pick up the various inputs you post here. However, just a couple of clarifications about the use I have for measuring coils:
- I, hopefully, soon will be making a couple of the University of York low phase noise oscillators at 22.579 MHz & 24.576 MHz frequencies. To this end I would like to confirm the inductance of the coils used here so as to make reasonably sure that they are the value they should be.
If I understand the feedbacks you have given correctly this should be possible with a VNA operating at 22.xx & 24.xx frequencies and then reading the inductance measured by the VNA. Would that be correct?
@peufeu:
Sounds like a very interesting approach to phase noise measurements ... Do you reckon that this setup would perform close to the performance of the so-called Timepod phase noise measuring device?
http://www.miles.io/timepod/ss_1pg.pdf
@Kn0ppers:
About those smaller VNAs I have seen them ranging from 10 kHz to the 4 GHz mentioned here so quite a span of frequencies ....
@CG & 1audio : You both mention a spectrum analyzer & CG also the Analog Discovery 2.
I have actually been considering the Analog Discovery 2 not least because it is also 14 bits resolution so it will show more detail than my current oscilloscope (200 MHz picoscope, unfortunately "tricky triggering" functionality). However, as far as I can see the Discovery's impedance measurement relies on the built-in signal generator which only goes to 12 MHz so it is a limited bandwidth.
But I am thinking that maybe a Discovery & spectrum analyzer could be a feasible combination. The first covering the lower frequencies and the other the higher frequencies ...
But maybe there are even better solutions - e.g. better accessible spectrum analyzers out there?
Thanks again for your feedbacks & have a good weekend,
Jesper
Unless the inductor/capacitor "plots" away from the outer ring of a Smith chart you will be unable to measure them with any reasonable accuracy. Your better off using manufacturers data.
Unless the VNA has a very good bridge in it and you can use a multi-term calibration routine with very good standards your component data will be suspect.
In real life you cannot get precision inductors for RF work. You can get variable inductors that work up to maybe 200MHz. You can get fairly good tolerance capacitors but in real life you will never get access to them unless you are buying very high volumes.
Unless the VNA has a very good bridge in it and you can use a multi-term calibration routine with very good standards your component data will be suspect.
In real life you cannot get precision inductors for RF work. You can get variable inductors that work up to maybe 200MHz. You can get fairly good tolerance capacitors but in real life you will never get access to them unless you are buying very high volumes.
> Now, since this is a topic (RF) that is not my main strength
I got a N2PK VNA, but I don't do any RF, it was just for audio
(although one could argue that 50MHz clocks are not really "low frequency")
> I, hopefully, soon will be making a couple of the University of York low phase noise oscillators at 22.579 MHz & 24.576 MHz frequencies. To this end I would like to confirm the inductance of the coils used here so as to make reasonably sure that they are the value they should be.
Busaboy is right, VNA can't measure anything accurately above say 100 ohms. It's very good for low impedance, for example if you want to know the ESL of a ceramic cap mounted on your board, then it's great. But for a 100µH inductor, it's useless. If you use a coil at high frequency though, you should also worry about its self-resonant frequency (SRF). VNA won't help with that unless the coil is tiny enough that its impedance peak is below what it can measure accurately.
So you need an instrument that measures impedance, not just inductance. I thought it was for a speaker crossover or something. For this application you can measure just the inductance, so I said you can also do it with a soundcard. But if you want to know the SRF to make sure the coil is still a coil at the frequency you're using it at, then, no soundcard.
> sounds like a very interesting approach to phase noise measurements ... Do you reckon that this setup would perform close to the performance of the so-called Timepod phase noise measuring device?
lol, my setup cost 10 cents (one NOR gate plus some junk), so, probably not.
I did not bother to calibrate it. All I was interested is which one of these cheap canned oscillators performed the best, so I bought pairs, mixed the outputs, and picked the one with the nicest FFT. To adjust frequency on the oscillators I just used some more or less discharged AA batteries. I also listened to them in the DAC and the one with the lowest jitter also sounds the best. It's Vectron VCC1, unfortunately unobtainium now.
I've used the same trick to measure how much jitter WM8805 adds to the signal. I wired a WM8805 as transmitter, pumped I2S into it, sent its SPDIF output to another WM8805 as receiver, mixed the BCLK from the original I2S and the BCLK out of the receiver with a NOR gate, then lowpass. Since they're the same frequency, the HF goes away with the mixing and lowpass, and you get the phase noise/jitter directly with the soundcard. It turned out the internal XO of WM8805 is garbage, also very sensitive to layout, but with an external oscillator, it's absolutely awesome.
I got a N2PK VNA, but I don't do any RF, it was just for audio
(although one could argue that 50MHz clocks are not really "low frequency")
> I, hopefully, soon will be making a couple of the University of York low phase noise oscillators at 22.579 MHz & 24.576 MHz frequencies. To this end I would like to confirm the inductance of the coils used here so as to make reasonably sure that they are the value they should be.
Busaboy is right, VNA can't measure anything accurately above say 100 ohms. It's very good for low impedance, for example if you want to know the ESL of a ceramic cap mounted on your board, then it's great. But for a 100µH inductor, it's useless. If you use a coil at high frequency though, you should also worry about its self-resonant frequency (SRF). VNA won't help with that unless the coil is tiny enough that its impedance peak is below what it can measure accurately.
So you need an instrument that measures impedance, not just inductance. I thought it was for a speaker crossover or something. For this application you can measure just the inductance, so I said you can also do it with a soundcard. But if you want to know the SRF to make sure the coil is still a coil at the frequency you're using it at, then, no soundcard.
> sounds like a very interesting approach to phase noise measurements ... Do you reckon that this setup would perform close to the performance of the so-called Timepod phase noise measuring device?
lol, my setup cost 10 cents (one NOR gate plus some junk), so, probably not.
I did not bother to calibrate it. All I was interested is which one of these cheap canned oscillators performed the best, so I bought pairs, mixed the outputs, and picked the one with the nicest FFT. To adjust frequency on the oscillators I just used some more or less discharged AA batteries. I also listened to them in the DAC and the one with the lowest jitter also sounds the best. It's Vectron VCC1, unfortunately unobtainium now.
I've used the same trick to measure how much jitter WM8805 adds to the signal. I wired a WM8805 as transmitter, pumped I2S into it, sent its SPDIF output to another WM8805 as receiver, mixed the BCLK from the original I2S and the BCLK out of the receiver with a NOR gate, then lowpass. Since they're the same frequency, the HF goes away with the mixing and lowpass, and you get the phase noise/jitter directly with the soundcard. It turned out the internal XO of WM8805 is garbage, also very sensitive to layout, but with an external oscillator, it's absolutely awesome.
Last edited:
There are Murata inductor kits, they are excellent. The kits and the inductors.
And you get them at low quantities from Digikey and others.
And my DG8SAQ VNWA has a quite good bridge made from SMD parts and it
also does 12 term error correction. And below a few 100 MHz it does not deliver
results that are really different from the R&S ZVB 8 next to it.
It has a very active user group VNWA@groups.io with a lot of people who know
what they are talking about. Much interesting stuff for any VNA user.
Ok, the ZVB 8 has a better dynamic range, can do power sweeps where "power"
is justified and goes much higher. But the price tag was a different league.
Gerhard
Attachments
Last edited:
Weapons grade nonsense. 100 Ohm is still in the middle of the Smith chart.
No way. Just think where you can get spectrum plots that go down to -175 dBc.
Perfect scenario for injection locking.