NanoVNA Custom Software

[Jacon]

Hi battlecoder,

You should subscribe to this board ASAP:

https://groups.io/g/nanovna-users

It's really the main place of info, help & devel works for this marvelous unit 

[ted572]

The NanoVNA menu can be like working in the dark searching for where to go next.  That is until you get used to it, and appreciate it’s features and very easy operation.  So for a kick start and a reminder, I made this PDF Table of the Menu Options in the NanoVNA with Firmware (edy555) 0.3.1 (the latest). Please let me know if you find any errors or missing info.

[Jacon]

...  So for a kick start and a reminder, I made this PDF Table of the Menu Options in the NanoVNA with Firmware (edy555) 0.3.1 (the latest). Please let me know if you find any errors or missing info.

This, IMHO, is the most actual, accurate & convenient Menu info page:
https://oristopo.github.io/nVhelp/html/Menu.htm

Oristo prepared quite comprehensive docu page also:
https://oristopo.github.io/nVhelp/html/nVhelp.htm#U0

[bradley1evblog]

Hello Joe, I'm new to the nanovna but delighted with the product. I'm trying to find your Nano Custom Software (for a pc) as shown in your video .... is it available for download anywhere?

[joeqsmith]

Hello Joe, I'm new to the nanovna but delighted with the product. I'm trying to find your Nano Custom Software (for a pc) as shown in your video .... is it available for download anywhere?

Bradley,  while I have not released my software there are a few open source projects being worked on that may work well for you.  I can't say I have personally tried any of them but I suspect they are much closer to what you would want from a software interface than what I put together for doing my own evaluation of the Nano.   

As I have stated a few times, mine is really more of an engineering research tool than anything.  It supports some hardware to run specific tests that the average user is not going to have access to.   It's also become a lot more complex over time and I have yet to put any sort of manual together for it. 

I was going to make one last video showing some of the latest features and maybe provide some additional ideas to those creating the open source code but I lost interest in the project.   If you find that the open source projects are lacking in some way, I suggest you politely ask the authors and see if they will support you. 

Enjoy your Nano.   It is a pretty nice device, especially at the $50 price.   

[pa3hfu]

I installed VNASaver on my iMac. Installation was succesful. But when I connect my unit by USB and click connect to NANOVNA I get this failure:

MacMini-I:.config mathieuverhorst$ cd nanovnasaver
MacMini-I:nanovnasaver mathieuverhorst$ nanovnasaver
NanoVNASaver 0.2.0
Copyright (C) 2019 Rune B. Broberg
This program comes with ABSOLUTELY NO WARRANTY
This program is licensed under the GNU General Public License version 3
See https://github.com/mihtjel/nanovna-saver for further details
Settings: /Users/mathieuverhorst/.config/NanoVNASaver/NanoVNASaver.ini
2019-11-28 11:40:42,308 - NanoVNASaver.NanoVNASaver - ERROR - Tried to open  and failed: [Errno 2] could not open port : [Errno 2] No such file or directory: ''

What can be wrong?
If I cannot connect to my Mac, I even cannot update my VNA-f in the future...
Thanks in advance again!!

[radiolistener]

pa3hfu, I'm not sure if NanoVNA-F is compatible with NanoVNA-SAVER, but your error looks like you have issue with port name. It may be related with missing driver or bad USB connection.

[pa3hfu]

Thanks RadioListener,  It was a bad USB connection. Another cable ==> succes!! 

[ted572]

RF Demo Kit - NanoVNA Test Networks Data
The following PDF provides information on the RF Demo Kit manufactured by SYSJOINT.CON, EEPELEC.COM, etc. This document describes the required VNA settings (of course not just the NanoVNA) to duplicate the response cures depicted on the unit's PCB for each RF circuit.
Edit: Without this type of information the RF Demon Kit is of very little value, if any for a novice.  Although with it you can verify that your VNA (any type/model) is working properly, and the RF Demo Kit then can also assist with learning how to use a VNA for measuring low value LC (i.e. uH, pF) components, use of a Smith Chart, etc, in addition to simply Antenna SWR.
Note: I’m not affiliated with any eBay seller of the RF Demo Kit, and in fact I held off posting this for a month or so because I didn’t want them to use this info to sell their product. But I caved and decided to share it with other users that may be interested. Although this is not rocket science, and anyone with a VNA could have done this very easily.

[joeqsmith]

Just an update.

I've had a few people ask me about how the Nano compares with other VNAs.  I will ask for specifics but it seems to be a more general question in general.  I think the last person asked about reliability and I responded that I had the Nano for a few months but my old 8754A has spanned five decades now and still running.  (with the original capacitors BTW lol) 

I see where there continues to be an effort to produce a 3GHz pocket VNA that doesn't use harmonics.  If they come up with a product, I plan to have a look.   

Playing with the Nano may end up costing me a chunk of change as it's rekindled my interest in replacing my old VNA.  I am reading up on the more common instruments available today.       

I was thinking that I may do some sort of video that compares these four systems.   Obviously reliability can't be a metric but if you have something you would like to see, feel free to chime in. 

A few things off the top of my head are noise, cost, features, noise, stability, accuracy, noise, dynamic range, ease of interfacing it with a PC, noise and sweep rates.   I may do something with noise while I am at it.   

[ogden]

Main difference between two: 8754A is instrument with certain and defined accuracy, NanoVNA is "low cost tool". Latter is more like pocket DMM which BTW is sold w/o proper enclosure, another - benchtop multimeter with (possibly valid) calibration sticker on it. Both have their use.

[joeqsmith]

While it wouldn't take much to improve on my old HP,  I would like to find something that will support lower frequencies (<1MHz) and faster data rates to the PC.  I dare say without the Labview interface, the old HP would be worthless as I no longer have a grease pen to write on the screen.    With Windows 10, I now need to run it under a virtual XP box and I can't see investing the time to do this.   I want to get something new enough to at least support Ethernet.    Thinking something along the lines of the old Agilent PNAs like this one:

https://www.ebay.com/itm/Agilent-PNA-E8358A-300kHz-9GHz-2-Port-4-Receiver-RF-Vector-Network-Analyzer-015/223568031067?hash=item340db16d5b:g:rw4AAOSwK0ldFTxx

I've been reading the manuals and like the system but they do seem to have some high asking prices for them.   

[DH1AKF_Wolfgang]

Please look here:
https://groups.io/g/nanovna-users/wiki/shellcommands

[joeqsmith]

Please look here:
https://groups.io/g/nanovna-users/wiki/shellcommands

If there is something specific you would like me to have a look at, just upload it here or to Google docs.   

I have to be registered to gain access to this area.  I was a member for a very short time.  The majority of posts were about how to print a case for it.    There were very few technical discussions and so I left. 

That ebay listing advertises "Fully Functional,  aligned with test data", but we can clearly see one of the source attenuators is damaged.   


https://www.ebay.com/itm/Agilent-PNA-E8358A-300kHz-9GHz-2-Port-4-Reciever-RF-Vector-Network-Analyzer/273440234503?_trkparms=aid%3D1110001%26algo%3DSPLICE.SIM%26ao%3D2%26asc%3D20160323102634%26meid%3Df1cbcf5bc5c74606920468d8c5276bd9%26pid%3D100623%26rk%3D2%26rkt%3D6%26sd%3D223568031067%26itm%3D273440234503%26pmt%3D0%26noa%3D1%26pg%3D2047675&_trksid=p2047675.c100623.m-1

I was thinking, pay a little more and you could get a working one.     We can see "Seller refurbished: An item that has been restored to working order by the eBay seller or a third party not approved by the manufacturer. This means the item has been inspected, cleaned, and repaired to full working order and is in excellent condition."   Again, slide down to the diagnostics page and we can see the unit fails the diagnostics. 

Most listings don't offer any details so it's a pretty high risk for such an old system that is no longer serviced and parts may be difficult to locate. 

I haven't looked at what they are actually selling for yet but  for a 20 years old, damaged unit,  the prices see way off.    Maybe time for some very low offers.   

[Roger Need]

I have to be registered to gain access to this area.  I was a member for a very short time.  The majority of posts were about how to print a case for it.    There were very few technical discussions and so I left. 

Joe,

The groups.io nanaovna users group has changed a lot since you first saw it. High SNR now... 

There a now a lot of people with extensive VNA experience posting good technical stuff at a high level.  Some are VNA calibration commercial vendors and one is developing a new low cost VNA's that go up to 3.5 GHz for $100 USD (just went into production) . Hugen who made the nanovna so popular by creating an assembled product posts from time to time as well as several authors who have created forked firmware and software for the device.  There is also an extensive wiki on all facets of software, hardware and documentation plus many informative user manuals. 

Roger

[joeqsmith]

I have to be registered to gain access to this area.  I was a member for a very short time.  The majority of posts were about how to print a case for it.    There were very few technical discussions and so I left. 

Joe,

The groups.io nanaovna users group has changed a lot since you first saw it. High SNR now... 

There a now a lot of people with extensive VNA experience posting good technical stuff at a high level.  Some are VNA calibration commercial vendors and one is developing a new low cost VNA's that go up to 3.5 GHz for $100 USD (just went into production) . Hugen who made the nanovna so popular by creating an assembled product posts from time to time as well as several authors who have created forked firmware and software for the device.  There is also an extensive wiki on all facets of software, hardware and documentation plus many informative user manuals. 

Roger

Good to hear that the group has evolved.   

In my first post, I had mentioned the plan was to compare the 3GHz unit with a few others once it becomes available.   As I mention, their Wiki appears locked for member's only.   So if there is anything specific people would want me to look at, they would need to make it publicly available.   

https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg2887094/#msg2887094

There was a member here that was doing a fair amount of work on the firmware.  I haven't looked at their progress in a few months but the later versions I was testing were a lot more stable that anything else I had seen at the time.

[RCinFLA]

You are basing this on ADC bits.  ADC bits is not the limiting factor on dynamic range.  It is the synthesizer, followed by harmonic mixing that is limiting dynamic range.

Also the lower maximum sampling rate of the TLV320 limits the frequency of the I.F. which in turn results in more sideband noise from the synthesizer which limits dynamic range.

[joeqsmith]

You are basing this on ADC bits.  ADC bits is not the limiting factor on dynamic range.  It is the synthesizer, followed by harmonic mixing that is limiting dynamic range.

Also the lower maximum sampling rate of the TLV320 limits the frequency of the I.F. which in turn results in more sideband noise from the synthesizer which limits dynamic range.

I would have no way of knowing what this is in regards to.   Who is "you" and what was based on ADC bits? 

[OwO]

Yes, so V2 uses the built in ADC on the GD32. The input noise floor is still far above the ADC's own noise floor so it was "free" cost savings. The bottleneck on S21 noise floor is the PSRR of the receive mixer and the fact that LF noise can't easily be filtered (and there is no power budget available for a linear regulator instead of a switching one).

The bottleneck on S11 trace noise is synthesizer phase noise, both the close-in noise (slow wobbling of the center frequency/phase) and the noise at the image frequency (f_RF + f_IF * 2). The close-in noise can be cancelled by the reference receiver, but the image noise can't. At 12kHz IF the two noise contributions are about the same, which means the noise improvement due to the reference receiver is only 3dB, which is why we went with a switched reference channel in V2.

[profdc9]

May I make a suggestion?  Synchronize the data acquisition to the IF cycles, rather than free acquiring the IF because the microcontroller and the frequency synthesizer are on different clocks.  This also helps with phase noise.  The VNA I designed acquired four samples per cycle, so that one can do simple addition and subtract to get in-phase and quadrature components.  This is not presently done in the NanoVNA.  This will help reduce the sensitivity to phase noise and allow much longer integrations (smaller IF bandwidth) to achieve a lower noise floor if needed.

Also, I don't think the NanoVNA as built presently stays within the linear region of the mixer.  It overdrives the inputs of the SA612 severely.  The datasheet indicates the proper level of drive.  The RF pins should be much less than kT/q or 26 mV, usually 10 mV is the maximum desired drive.  The oscillator pins should be no more than 200 mV.  I stick with 100 mV to stay in the linear region.

Gilbert cells are good for RF receivers but their dynamic range is not that great because they only accept small signals.  The dynamic range in practice is determined by the noise voltage on the low-end (a few microvolts) up to the limit where the intermod products start to distort the mixing signal (about 10 mV).  EU1KY did a good job in his design because he maximized the signal available to the SA612 given these constraints.   The tests being done on the NanoVNA on open, short, load don't test this at all.  Try measuring 10000 ohms S11 in the HF or 1000 ohms at UHF and see what happens.  Don't make these mistakes again with a new design.

Also, it wouldn't hurt to have a proper fixture so that you can do these tests properly.  I designed one for my VNA.  Perhaps you could use that or make one of your own.  I very much tried to wring every bit of accuracy out of the VNA design even if it meant the data acquisition would be slow.

Anyways if any of this interests you, take a look at

http://github.com/profdc9/VNA

[ogden]

microcontroller and the frequency synthesizer are on different clocks

No they are not. In NanoVNA and NanoVNA.v2 everything related to signal path (mixers & ADC) is clocked from single source, si5351.

[joeqsmith]

https://www.rtl-sdr.com/a-nanovna-2-0-is-in-the-works-same-price-frequency-range-up-to-3-5-ghz-higher-dynamic-range/

In the above article, dated Oct 4th 2019, they mention initial sales in roughly 3 months.  I assume then that things are pretty much wrapped up with research/development and it is now in the production phase .   Is this correct, or is the project still in the design phase?

[profdc9]

That doesn't mean that the acquisitions are synchronized so that the same part of each IF cycle is sampled.   How many clocks of the crystal on the SI5351A correspond to the RF and LO signals and the MCLK driving the codec and the microcontroller?  They are all on separate fractional dividers.  Being clocked from the same crystal is a necessary, but not sufficient, condition for the IF acquisition to be synchronized.  It doesn't take much error in the phase to limit your dynamic range, and the microcontroller needs to know the phase fairly precisely to demodulate the IF signal, especially if there no other signal indicating the relative phase relationship between the microcontroller's clock and the IF.

Anyways, do it or don't do it, it doesn't matter to me.  The NanoVNA is a result of groupthink.  There's not much verification of performance against real loads.  All of the emphasis is on precision, not accuracy.

https://groups.io/g/nanovna-users/topic/nanovna_real_resistance/33508471

You might want to understand the sources of error in the present NanoVNA before moving on to the next one.  The SI5351A is used out of spec up to 300 MHz and the SA612 is heavily overdriven.  What do you expect?  But since it seems its mostly used for testing whether antennas are matched or not, it's probably good enough for most of its users who have never used a VNA before and don't know what to expect.  I can't even get a straight answer to how to calibration works on the NanoVNA.  Is there anyone besides edy555 that knows how that works?  I sat down and derived the equations myself and tried to figure out how the NanoVNA firmware implements calibration.

Anyways no matter how it turns out that NanoVNA v2 will be a success, so there's no point in criticizing it.

[OwO]

Yes I know what you are talking about. People on that list "verify" their calibration by remeasuring their short, open, and load. I suggested the coaxial stub test to check for linearity and so far only one person has done a proper test.
V2 is a separate effort and doesn't reuse anything from the original design. The IF signal's period is exactly 25 ADC clock cycles. The mixer is a high linearity one (over 20dBm IIP3) and is driven at least 20dB below P1dB. I posted some linearity and error magnitude calculations on the list where I concluded you must stay 20dB below IP3 and 10dB below P1dB to get reasonable EVM.

[profdc9]

Well it's good, if you're going to make a fresh start, it's better of course to not repeat mistakes.  I think some kind of fixture would be helpful, however, to go with the new VNA to verify its operation.  That way someone who receives it can verify it is working properly to spec.

I have used a zif socket, and if used carefully, can actually work up to about 500 MHz!  But that's not going to cut it for the V2, so think about a good set of standards and a fixture to go with it.  That will both help to create it and put to rest questions about its performance.  Even a simple PCB with a coplanar waveguide on it, something like this:

https://www.edn.com/pcb-fixtures-improve-component-measurements/

could be a big help.