NanoVNA Custom Software

[N2CUA]

Yeah I never got that concept either ..
My sister is a bit like that .. but she's family so I just accept it for what it is and try to find ways to make her see ( gently ) that its just not the case .. Takes time ... but I think she is coming around .. a bit of tough love though .. 

I hope you change your mind at some point .. but thats your call of course..

Thanks for the kind words ..

You would think with all the available data on the USB bus coming from the VNA that there would be a more robust interface ..
But still ... for the price its great and I dont have to lug my Rigol spectrum analyzer around when helping friends with their antennas ..

Randy

[Psynosaur]

Thanks to the OP joeqsmith.

Your videos are very educational and well put together

Then thanks 5q5r for your nanovna - saver app, it quite a treat playing with it

Kind regards

Keep the information coming!

ZS1SCI

[joeqsmith]

5q5r

I tried your prebuilt EXE on Windows 10.  It connected to the Nano and I was able to get it to plot the data without any problems.   It looks like a very good start.   So far, I think you are the only one who has stepped up and shown something other than lip service.   With the number of people asking for open source, Python code, you are making a lot of people very happy.

[joeqsmith]

Yeah I never got that concept either ..
My sister is a bit like that .. but she's family so I just accept it for what it is and try to find ways to make her see ( gently ) that its just not the case .. Takes time ... but I think she is coming around .. a bit of tough love though .. 

I hope you change your mind at some point .. but thats your call of course..

Thanks for the kind words ..

You would think with all the available data on the USB bus coming from the VNA that there would be a more robust interface ..
But still ... for the price its great and I dont have to lug my Rigol spectrum analyzer around when helping friends with their antennas ..

Randy

I'm a bit surprised myself that the PC software was released in its current state.    You could have a look at 5q5r's code.  Obviously they are just getting started but it looks promising.

https://github.com/mihtjel/nanovna-saver

[hendorog]

I haven't been following this thread, but if anyone is interested:

Here are some examples of offline calibrations, comparative tests with my HP 8753 and another NanoVNA, and a TDR example.

https://github.com/hendorog/nanovna_test
This uses the free Jupyter Notebook which you need to install if you want to play on your machine with this code.
Apologies in advance as the comments are pretty sparse, most were done as part of conversations on groups.io.

Note, I am not grabbing the data automatically yet, this is just for offline analysis at the moment. I have been using the NanoVNA C# app to get the s1p files to date.

[radiolistener]

hendorog, what mixers are used in your nanovna version - SA602 or SA612?

Can you take S1P for some load different than 50 Ohm?
For example 500 Ohm (SWR=10.0) will be a good choice for testing nano vna precision.
You can create such dummy load with SMA connector and two 1k SMD resistors soldered on it.

Such dummy loads - 500 Ohm and 5 Ohm is a must have for VNA testing and calibration. They both should give SWR=10.

[hendorog]

hendorog, what mixers are used in your nanovna version - SA602 or SA612?

Can you take S1P for some load different than 50 Ohm?
For example 500 Ohm (SWR=10.0) will be a good choice for testing nano vna precision.
You can create such dummy load with SMA connector and two 1k SMD resistors soldered on it.

Such dummy loads - 500 Ohm and 5 Ohm is a must have for VNA testing and calibration. They both should give SWR=10.

The last graph on this page https://github.com/hendorog/nanovna_test/blob/master/NanoVNA%20test.ipynb shows nanovna and 8753 measurements of a reference attenuator provided with my cal compared with reference data provided by Dr Kirkby.
(Note the 8753 was intentionally setup to match the nanovna, without any cal kit information. When properly configured for the cal kit, the 8753 will match the green line closely)

This gives a good idea of the out of the box accuracy of the device using its own (uncharacterised) cal kit against an independently measured standard above 50MHz.

Will do some more tests for you, but if I don't get it done tonight it might be a while due to other commitments.

My two devices (white gecko and black hugen) both have 602 mixers. I have read that the two models of mixers are internally identical but no idea if that has been confirmed.

[hendorog]

Also, just to add there is a known bug in the released NanoVNA sharp application when using it to do calibrations.

The measurement of the other standards are discarded unless the Short is the first standard measured.

This is because the Short button click event in the NanoVNA sharp app does a call to the 'cal_reset' operation - which wipes any earlier measurements out.

So if you do a cal in Open Short Load order using the NanoVNA sharp application, there will be an error of about 0.3dB - not much, but significant enough to be important when doing comparisons as can be seen in the attached pic.

First pic attached showing the effect of doing OSL cal order.
Second pic attached which is the same as in the previous post, showing the workaround of doing SOL cal order.

[N2CUA]

Inspired by Joe's great YouTube videos on the NanoVNA, I went ahead and botched together a little Python app that does some of the same things (most importantly, multiple sweeps to increase the amount of data points.)

It's open source and on github, as "NanoVNA-Saver" (because I initially just made it to save s1p/s2p files). I daren't post a link in my first post on this forum though :-)

Thanks Joe for the inspiration, and I hope some of the rest of you may find it useful.

Would you be up to helping me set up my Python on windows 10? Its installed but having issues with installing PyQT5. Python is completely new to me ..  Not sure if its worth all this just to have a better interface ..

[radiolistener]

My two devices (white gecko and black hugen) both have 602 mixers. I have read that the two models of mixers are internally identical but no idea if that has been confirmed.

It looks that SA612A works much better. My NanoVNA have SA602A mixers. Here is my 500 Ohm terminator measurement:

Series RLC:


Parallel RLC:


Also see S1P file for SA602A version with 500R terminator on CH0 in attachment.


Here is also screenshot for 500R terminator from another user, who have NanoVNA with SA612A:


As you can see, SA612A looks much better and works almost up to 900 MHz with above 200 Ohm result. While SA602A is going below 200 Ohms at just 160 MHz. But I'm not sure. May be this is just quality issue with my NanoVNA (for example bad soldering or something like that).

It will be good if you can confirm that your SA602A version also has bad performance on Series RLC screen. So, if SA602A is the root of cause for that, then it probably can be fixed by replacement SA602A mixers with SA612A

[N2CUA]

Inspired by Joe's great YouTube videos on the NanoVNA, I went ahead and botched together a little Python app that does some of the same things (most importantly, multiple sweeps to increase the amount of data points.)

It's open source and on github, as "NanoVNA-Saver" (because I initially just made it to save s1p/s2p files). I daren't post a link in my first post on this forum though :-)

Thanks Joe for the inspiration, and I hope some of the rest of you may find it useful.

Would you be up to helping me set up my Python on windows 10? Its installed but having issues with installing PyQT5. Python is completely new to me ..  Not sure if its worth all this just to have a better interface ..

I got it working .. Just have to learn PYTHON now .. lol

[hendorog]

My two devices (white gecko and black hugen) both have 602 mixers. I have read that the two models of mixers are internally identical but no idea if that has been confirmed.

As you can see, SA612A looks much better and works almost up to 900 MHz with above 200 Ohm result. While SA602A is going below 200 Ohms at just 160 MHz. But I'm not sure. May be this is just quality issue with my NanoVNA (for example bad soldering or something like that).

It will be good if you can confirm that your SA602A version also has bad performance on Series RLC screen. So, if SA602A is the root of cause for that, then it probably can be fixed by replacement SA602A mixers with SA612A

Interesting - based on the hint I mentioned, I searched and found this story on the internet:
https://soldersmoke.blogspot.com/2009/06/na5n-on-ne602.html

I'll do that test, tonight if I have time, and I will also sweep the loads on my 8753 for comparison.

[radiolistener]

it seems that my NanoVNA with SA602A may be OK. Probably this measurement error caused by 1 cm SMA-M to SMA-M adapter. I tried to add two adapters on another VNA and got similar measurement error.

Just incredible, 1 cm SMA to SMA adapter leads to so high measurement error... 

[N2CUA]

it seems that my NanoVNA with SA602A may be OK. Probably this measurement error caused by 1 cm SMA-M to SMA-M adapter. I tried to add two adapters on another VNA and got similar measurement error.

Just incredible, 1 cm SMA to SMA adapter leads to so high measurement error... 

Can you try a calibration with all cables and connectors that will be used in place ?
Other than the needed calibration items of course

[Flynt]

A greeting to the whole group.
I'm a new user.

I tried to follow all the Treead, but I don't read English well, (I'm Italian).

I would like to ask you if anyone can kindly tell me where to recover this software "Custom Lab View for Nano VNA"

THANK YOU ANTICIPATEDLY WHO WILL MAKE ME AID ...

[_Wim_]

A greeting to the whole group.
I'm a new user.

I tried to follow all the Treead, but I don't read English well, (I'm Italian).

I would like to ask you if anyone can kindly tell me where to recover this software "Custom Lab View for Nano VNA"

THANK YOU ANTICIPATEDLY WHO WILL MAKE ME AID ...

Hello Flynt, for the moment this software is not available for download. It was developed by joeqsmith to show what is possible with the Nanovna, to make some educational youtube video's about the NanoVNA and to inspire others to make apps of their own (which has worked already, thanks to 5q5r)

[Flynt]

Thanks Wim
I've already written a private message to joeqsmith, to know how to get it ...

I'm not able to program...

[_Wim_]

I think this software is a good alternative to what joeqsmith has made:

https://github.com/mihtjel/nanovna-saver

[_Wim_]

The executable can be downloaded from here if you do not want to run it in phython (download the zip file)

https://github.com/mihtjel/nanovna-saver/releases

[radiolistener]

Can you try a calibration with all cables and connectors that will be used in place ?
Other than the needed calibration items of course

The problem here is that all calibration loads should be connected through the same length as 500 Ohm terminator.

Here is my calibration kit and 500 Ohm terminator:


As you can see, it's about 15 mm longer than calibration terminators. It leads to additional delay in wave propagation and as result it leads to measurement error.

You can check my previous S1P file for 500R terminator, see Group Delay S11, it's about 0.25 ns.

Now I tried to setup ELECTRICAL DELAY = 247 picoseconds (see SCALE menu in NanoVNA) to get software correction for that delay. I tried different values and 247 ps give me the best result. So it seems like these 15 mm leads to 247 ps delay. And now, with 247 ps delay correction, I get much better result

Series RLC:


Parallel RLC:


Group Delay S11:


S11:


VSWR:

[radiolistener]

Here is another finding. When I tried to get the same measurement (with electric delay 247 ps) for 500R terminator with no connection to PC (powered from accumulator), it shows 500 Ohm and VSWR=10 almost up to 900 MHz.

For example:
600 MHz: Rs = 550 Ohm, VSWR = 11.07
700 MHz: Rs = 501 Ohm, VSWR = 10.03
800 MHz: Rs = 478 Ohm, VSWR = 9.62
900 MHz: Rs = 462 Ohm, VSWR = 9.08

So, it seems that USB connection to PC affects measurements...

Tried more tests about that. When I connect it to powered off power bank, there is no influence, but after power on, there is a change in measurement above 500 MHz. It seems that external power supply leads to increase power supply voltage and it affects measurement.

I tried to calibrate it with external power supply, but it seems, that calibration cannot eliminate measurement error which appears with external power supply. So, the only way to get the best measurement is to disconnect device from external power supply. Very strange. Any idea how to fix it?

[rhb]

I checked the pinout of  the STM32F303CCT6 with 48 KB of RAM and 256 KB of flash today and it should be a drop-in replacement for the existing STM32F072CBT6.  The 3x increase in RAM and 2x increase in flash would allow a substantial increase in features such as TDR, storing multiple calibration ranges, etc.  The groups.io list is a firehose of activity.

I was going to do the chip swap myself, but one of the sellers, hugen, has offered to provide a unit with the '303 part.  I expect that will be very popular with people wanting to write FW with additional features.  So I'm hoping for an announcement that '303 based units are available for  an extra $5-10 US.  In quantity 2400 from Digikey, the price differential is $1.28 US.  $1.92 vs $3.20 for the MCU.  That makes it an attractive change for an OEM if it allows more features.

The 16 KB memory of the current versions is a significant limitation.  An important application for the additional memory is TDR from VNA data. There are already multiple implementations for producing TDR traces from VNA data being done on a PC, but no room for it on the device.

The existing FW should run on the larger part unaltered, so buying a unit with the STMF303 part  makes it easy for a buyer who is not interested in writing FW to do updates as things develop.  This is the RF version of the $20 LCR-transistor tester.  Hang on to your hats!

Have Fun!
Reg

[rhb]

Here is another finding. When I tried to get the same measurement (with electric delay 247 ps) for 500R terminator with no connection to PC (powered from accumulator), it shows 500 Ohm and VSWR=10 almost up to 900 MHz.

For example:
600 MHz: Rs = 550 Ohm, VSWR = 11.07
700 MHz: Rs = 501 Ohm, VSWR = 10.03
800 MHz: Rs = 478 Ohm, VSWR = 9.62
900 MHz: Rs = 462 Ohm, VSWR = 9.08

So, it seems that USB connection to PC affects measurements...

Tried more tests about that. When I connect it to powered off power bank, there is no influence, but after power on, there is a change in measurement above 500 MHz. It seems that external power supply leads to increase power supply voltage and it affects measurement.

I tried to calibrate it with external power supply, but it seems, that calibration cannot eliminate measurement error which appears with external power supply. So, the only way to get the best measurement is to disconnect device from external power supply. Very strange. Any idea how to fix it?

My first guess would be EMI coming in on the power cable.  Does the cable you're using have chokes?

[tautech]

The sweep width and the few data points available is also a significant limitation on accuracy. Joe showed this earlier where the interpolation with so few points had it drawing a series of straight lines on the Smith chart.
The user needs keep this data point limitation firmly in mind and use a sweep width that's appropriate and a heap of averaging if necessary.

[radiolistener]

My first guess would be EMI coming in on the power cable.  Does the cable you're using have chokes?

I also thought about it. But this is not the case. I can connect any length of cable and it doesn't affect measurement. Also I tried to connect cable to power bank with disabled output and it doesn't affect measurement. But when I press power on button on power bank, it leads to measurement error. So, the the issue appears when external voltage is applied and don't depends on the cable.