NanoVNA Custom Software

[Alextsu]

I did finding their site where you can DL the manual for the new LiteVNA.  Maybe someone will benchmark it against the others.   
https://zeenko.tech/litevna?spm=a2g0o.detail.1000023.1.b5a97ff8tHYEOr

Yellow are the mapping circles.

Some preliminary info about nev SAA-2 from Hugen here.
https://groups.io/g/nanovna-beta-test/topic/new_v2_version_hardware/83829273
BPF compared against R&,S measurements

[joeqsmith]

Looks like its an invite only.  Too bad they didn't make it viewable.

[Alextsu]

DR for S21  BPF & f<3 GHz is better than 80 dB, for f>3GHz it's about 70 dB.
I shall ask for permission to make a repost here.

[joeqsmith]

Is that with any averaging?

***
Looking at the data I took with no averaging, the claim is it would out perform your VNA as well as the V2+4 I have above 3GHz.   They also claim it is faster than the V2Plus4 (>550 points/s ).   And, half the price.   

If they use the old ASCII interface,  my software may work but I have my doubts.  A while back I had a discussion with Dislord about the problems I was seeing with their firmware.   One thing they were pushing was to not use the command set I have been using as it was obsolete.  It may no long support these early commands.   Then there is the whole stability issues I see with the firmware.   

[DiSlord]

LiteVNA use V2 exchange protocol

For compare perfomance better use last LiteVNA hardware (i have only first test board)
I run and connect NanoVNA_r10 to LiteVNA all work, data exchange ok

[joeqsmith]

LiteVNA use V2 exchange protocol

For compare perfomance better use last LiteVNA hardware (i have only first test board)
I run and connect NanoVNA_r10 to LiteVNA all work, data exchange ok

I am curious if you get the same update rate with NanoVNA_r10 as with other software. 


Alex,

If I plot the GMS with GMAX, you will see the two programs match.   I do not currently plot this data but can certainly add it if you like. 

[DiSlord]

if set 10x AVG, SweepTime 3088
If set 2x AVG, SweepTime 675
For 1x avg show strange value 1 - 35 (for 1x avg sweep speed ~570 points/sec)

[joeqsmith]

if set 10x AVG, SweepTime 3088
If set 2x AVG, SweepTime 675
For 1x avg show strange value 1 - 35 (for 1x avg sweep speed ~570 points/sec)

I should have been more clear what I was asking.  I have no way to support averages so set the VNA to none.   

Connect and set the date points to 101 and the start/stop to 1MHz/300MHz.  Leave it in Smith chart. 

Goto the setup/Diagnostic tab and select the NanoVNA Raw Data tab.   CH1_Z should not be updating and the Freq(Hz) should be as shown.  If not, the range or number of points is not set right. 

What does the Response Time show?   This is the time from when I request data to when I actually receive it.   My original will bounce around from 600 - 800ms. 

You could lower the Refresh from 200ms to 50ms and throttle the trigger by increasing the PgmTimeout to 2000ms but it shouldn't matter.     

[DiSlord]

LiteVNA use V2 protocol (and use NanoVNA V2+4 software), as i see screenshot you show NanoVNA-H exchange protocol and software

V2 protocol not allow ask only one channel data, always send all

I connect to LiteVNA and made screensots

[joeqsmith]

Very nice.  You're using the install software.   You may want to try it with the current version.   

I never wrote a regression test for the new protocol as that V2+4 pretty much worked out of the box.   The problems I have seen with the V2P are it locks up at random.  This can be anywhere from a half hour to six hours.   When it locks, it requires a power cycle to recover.   It did turn out to be firmware, or at least I found a version that I have ran over 48 hours without a single glitch.   I tried one of their latest releases and the problem was back. 

[joeqsmith]

Autoscale still has a bit of a problem.   All three antennas from my demos.   

[Alextsu]

Alex,

If I plot the GMS with GMAX, you will see the two programs match.   I do not currently plot this data but can certainly add it if you like.

Thanks, leave it as You like)
....
I don't know whether it's possible or not to do in LabView (in *.exe files), but it might be good to provide the end user with an opportunity to scale the size of the app's main window at the screen.
While I am comfortable with the present settings, others might like to change them, depending on the screen resolution, etc.

[DiSlord]

I never wrote a regression test for the new protocol as that V2+4 pretty much worked out of the box.   The problems I have seen with the V2P are it locks up at random.  This can be anywhere from a half hour to six hours.   When it locks, it requires a power cycle to recover.   It did turn out to be firmware, or at least I found a version that I have ran over 48 hours without a single glitch.   I tried one of their latest releases and the problem was back. 

I update software to last, and run sweep on LiteVNA, at this moment it work more then 10 hour (201 points and 10 to  300MHz sweep)

As i remember you use my firmware for stable work (in it i made lot of fixes), LiteVNA also use part of this firmware (better say my V2Plus/Plus4 fw use LiteVNA code)

[joeqsmith]

Alex,

If I plot the GMS with GMAX, you will see the two programs match.   I do not currently plot this data but can certainly add it if you like.

Thanks, leave it as You like)
....
I don't know whether it's possible or not to do in LabView (in *.exe files), but it might be good to provide the end user with an opportunity to scale the size of the app's main window at the screen.
While I am comfortable with the present settings, others might like to change them, depending on the screen resolution, etc.

I'll add it to the next release to avoid future questions on why they differ.  Also, I will add an easy way to select log or linear scale. 

For scale see the following posts and those near them:

https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg3617154/#msg3617154

https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg3621182/#msg3621182

I have no plans to change this.

[joeqsmith]

I never wrote a regression test for the new protocol as that V2+4 pretty much worked out of the box.   The problems I have seen with the V2P are it locks up at random.  This can be anywhere from a half hour to six hours.   When it locks, it requires a power cycle to recover.   It did turn out to be firmware, or at least I found a version that I have ran over 48 hours without a single glitch.   I tried one of their latest releases and the problem was back. 

I update software to last, and run sweep on LiteVNA, at this moment it work more then 10 hour (201 points and 10 to  300MHz sweep)

As i remember you use my firmware for stable work (in it i made lot of fixes), LiteVNA also use part of this firmware (better say my V2Plus/Plus4 fw use LiteVNA code)

Thanks for checking my latest software.  I know you said your LiteVNA was an older revision but it would still be interesting have the data for it.  I wrote a simple procedure for the two measurements.  If you could upload the TS files, I would add them.   
https://nanorfe.com/forum/VNA-output-level.html

I don't think they ever sorted out why the V2P hangs.   The firmware I am using is not even listed on OWOs page, so I archived it with my software.   I have not changed the firmware in the V2+4.   That firmware is also not shown on their site and is a couple of days older than their stable version.   Guessing there was some bugs they fixed that do no effect me.   I only run them with the PC so most of the features added to the firmware are wasted on me. 

Yes, we had a discussion about using your firmware with the H4.   

Starting:
https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg3584301/#msg3584301

Last post:
https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg3586208/#msg3586208

I have not used the H4 since except to measure it's system dynamic range and noise floor.   That old communications interface was poor.  I am really glad you and Hugen adopted the newer interface.  It's robust, fast, reliable and very easy to implement.   

[Schorsch3]

Hello I am new here, congratulations for this great software.
I have a SAA-2N here and work with vna-qt SW.
Is it possible to use your SW also for the SAA-2N ?
PC with WIN7 Prof. is available.
Unfortunately I do not get along with the installation description yet.
Is there a step by step instruction to install your software ?
I have read that a "Main" SW is required. Where can I download this?
Many questions, please ask for info if possible.

Greetings

Translated with www.DeepL.com/Translator (free version)

[Alextsu]

I have a SAA-2N here and work with vna-qt SW.
Is it possible to use your SW also for the SAA-2N ?
PC with WIN7 Prof. is available.

Yes, it's possible, but is not recommended if You are not fluent with LabView-based apps.
It's a steep learning curve to start from.

Unfortunately I do not get along with the installation description yet.
Is there a step by step instruction to install your software ?
I have read that a "Main" SW is required. Where can I download this?
Translated with www.DeepL.com/Translator (free version)

1) Download & install the "Main" installer package from this link:
https://github.com/joeqsmith/NanoVNA_Software/releases/tag/0.10

2) Unzip, install.

3) For the latest version of SW, at Github, download the .zip file from this link:
https://github.com/joeqsmith/NanoVNA_Software/blob/NanoVNA_V2Plus/NanoVNA_V2Plus_2p08.zip

Unpack it to the path at C:/program files (x86)/... where the original SW (NanoVNA_V2Plus.exe) is located.
Replace the existing *.exe file with the current version file.

[DiSlord]

Thanks for checking my latest software.  I know you said your LiteVNA was an older revision but it would still be interesting have the data for it.
https://nanorfe.com/forum/VNA-output-level.html

My firmware for V2/V2Plus/V2plus4/LiteVNA support this registers (allow control output power and average settings)
    //    -- 40: average setting
    //    -- 41: si5351 power
    //    -- 42: adf4350 power
    registers[0x40] = current_props._avg;
    registers[0x41] = current_props._si5351_txPower;
    registers[0x42] = current_props._adf4350_txPower;

I wrote a simple procedure for the two measurements.  If you could upload the TS files, I would add them.   

How i can measure this data? Need measure open ports after calibration?

Here Hugen screenshots: 6dB attenuator on S11 port (S21 port open) and 60dB between ports, after ~3.8GHz possible see dynamic lost

[joeqsmith]

I wrote a simple procedure for the two measurements.  If you could upload the TS files, I would add them.   

How i can measure this data? Need measure open ports after calibration?

Here Hugen screenshots: 6dB attenuator on S11 port (S21 port open) and 60dB between ports, after ~3.8GHz possible see dynamic lost

To measure the S11 noise floor,  I used the standard cal kit that was provided with the V2.  The standards were connected directly to the VNA's ports (no cables or adapters were used).  Standards were properly torqued.   Three different bands were used, 1MHz - 1.5GHz, 1.5GHz-3GHz and 3GHz - 4.4GHz, calibrating for each band.

For the new VNA, add another band from 4.4GHz to 6GHz.  Cal the instrument for one of the bands using  the supplied SOL standards, then with the load installed, measure S11 and save it to a Touchstone file.   Do that for all four bands.

The specs for the System dynamic range for the V2P is 60dB < 3GHz and 70dB < 3GHz, without averaging.  I will assume System mean no amplification.   As before, the standard cal kit that was provided with the V2 was used.  The standards were connected directly to the VNA's ports.  For the thru, a single cable was attached from port 1 to 2.   I used one of the blue cables supplied with the V2P.  A;; connections were properly torqued.   Three different bands were used, 1MHz - 1.5GHz, 1.5GHz-3GHz and 3GHz - 4.4GHz, calibrating for each band.   Both ports were terminated to 50 ohms when making the measurement.

I used my software to make all of these measurements.  The ideal model was used in all cases.   My software supports the leakage term.   When measuring above 3GHz, I show it with and without the leakage term.

 
So, again cal the unit and theninstall a load on each port and measure S21.   Save that data to a Touchstone file.  Do this for all four bands.   Zip up all 8 Touchstone files and upload them here.   


My software I think uses the 0x42 address for the ADF4350 power.    You can find that setting under the Setup/Diagnostics tab,  Filters/Attenuators.  Once you enter a new number it should send it to the VNA.  You should be able to cal the unit, or just normalize with a cable shorting port 1&2 together.  Look at S21 after changing it.  You should be able to see the change.   Of course, you need to be in the range that it's using the ADF4350.     

[joeqsmith]

If you want to just run a single sweep like Hugen shows from 1MHz to 6GHz, that would be fine.   I was running in smaller bands but no reasons you couldn't do this.     

[Schorsch3]

Thanks for the quick information, the installation was successful.
Now I have to learn to work with it.

greetings Georg 

[DiSlord]

Made SOLT calibration and after single sweep, both ports open 10MHz to 6400GHz on 1001 points, save result to SD card

After add 50Om load on both ports (use different 50Om load (not calibration) on S11) and save again

[joeqsmith]

Thanks for taking the time to run these.  Both ports loaded is fine.  This will get us both and save's you time.  Would have been better to keep the cal standard on port 1 for the noise floor, but not a big deal.  I was just looking to get an idea where it was. 

Sounds like you ran it standalone (no PC).  For the higher frequencies (>3GHz) are you using the leakage term? 
Did you use any sort of smoothing on the single sweep?   

The dynamic range is impressive.

[DiSlord]

Yes, i run standalone (prevent noise from external power)

I have V2, and can see leakage between ports if put fingers between RF connetors (near buttons)
On LiteVNA i see small noise level drop after ~3GHz if add 50Om load on S11 measure port, also need warm device near ~5min

No any smooth used, only avg = 10x, this option increase measure data size for one point and give better quality, but slow measure time.

[joeqsmith]

Then actually 10 sweeps averaged, not one sweep smoothed.  This makes a lot more sense when looking at your data.  Your single sweep comment threw me off.   As I said, I took all my data with a single sweep, no average.   If there is a problem with the USB connection causing noise, I want to know about it.  I think what you have provided is good enough.  I wouldn't repeat the test unless you're bored.  We have the new hardware on order.   I'll let you know how it performs in a few months.   
 
If you copied the V2Plus4 firmware as a starting point, I expect the calibration accounts for the leakage term.  OWO had brought the need for it to my attention after I started porting my software over to it.