Hi All,

There were requests for an executable version of nanovna-saver that works with the v2 and I think I built one for Windows. This was created from Holger's development branch which you can see here on Github: https://github.com/zarath/nanovna-saver

Among other things, yesterday I patched a bug that stopped the vna2 from running faster the one the second sweep so now the first sweep runs at normal (first) speed and later sweeps are all about twice as fast. Nice...

The executable can be downloaded from my stash at: http://www.audua.com/Files/nanovna-saver_6_12.zip . Just unzip it and run it anywhere.

No guarantees but it worked once for me. It may warn you that the exe is dangerous because it's from the web. Tell it to ignore that.

-Mark

Hi Mark
Thank you for the exe. It works with the new Hugen S-A-A 2 v 2.2 . The was a couple of crashes during calibration and sweeping but in the end I found out the USB connection is not that stable so unplut and replug before starting the NanoVNA-Saver I went OK
Kind regards
Kurt

Mark, thank you very much for your release!

Do you think its OK if we add it on the group software section?

I am happy to hear from Kurt that also works with Hugen's 2.2!

Hi,
Sure, do whatever you'd like with it... I personally like it way better than vna-qt because of (1) able to show log charts and (2) better calibration with non-ideal stuff. I wish the markers worked better (and the serial comm) but I used the S1p export today and woohoo that was nice to use with excel.

-Mark

Hi Mark,

my Win10 vm is broken again....
can you give
https://github.com/zarath/nanovna-saver/suites/791026510/artifacts/8439281
a try?

73
Holger

Hi Mark,
how do you set the markers?
They seem to reset to the min sweep frequency at every scan.

Thanks!

Ioannis

Mark could you please check if you got my private message?

Cheers

Yeah, I spent a couple hours this morning trying to fix the markers and gave up. I'm very much a QTPython novice and couldn't find the linkage to repair. I'll probably look at them again because they are just so annoying. I have no idea if this is an artifact from developer branch or if they always have been broken.
-Mark

On 12.06.20 20:56, Ioannis E. Kassotakis wrote:
> Hi Mark,
> how do you set the markers?
> They seem to reset to the min sweep frequency at every scan.

This bug creept in between 0.2.2 and 0.2.3 development branch in
Rune's original repository..

I hope, I'll find the code change, that causes this problem.

73
Holger

Hello All,

On 12.06.20 12:42, MarkZ wrote:
> Hi All,
>
> There were requests for an executable version of nanovna-saver that
> works with the v2 and I think I built one for Windows. This was
> created from Holger's development branch which you can see here on
> Github: https://github.com/zarath/nanovna-saver

major work done. I've released a S-A-A-2 aka VNA V2 Edition
of NanoVNASaver on github.

https://github.com/zarath/nanovna-saver/releases/tag/v02.3.0-dg5dbh

This fixes also the annoying marker frequency reset bug after each
sweep.

Currently S-A-A-2 will be polled with 300 points / scan. Should
be choosable in next updates.

Nearly all Classes are in separate files now. This refactoring
can cause new errors of course, but I want to add further
analysis methods, like calculating optimal tune networks for
antennas for example. Therefore I've tried to get the code more
modular. As more parts of the code aren't GUI dependable it
would be easier to do automated test

Have fun,
73
Holger (DG5DBH)

Hi Holger,
My attempts have failed, the program can only run on a 64-bit machine

73, Gyula HA3HZ

Hi Gyula,

On 15.06.20 18:39, Gyula Molnar wrote:
> Hi Holger,
> My attempts have failed, the program can only run on a 64-bit machine

sorry I've no access to 32Bit develop environments. So I can only
suggest a source installation for that targets.
I've added a comment in the release text.

73
Holger

Hi Holger,

Thanks for the information.
The source code under Python also runs under Windows 32 bits anyway.
I can't test in the absence of S-A-A-2, but I can run the program.

73, Gyula HA3HZ

Hello Holger,

I have tried your nanoVNA-saver fork 0.2.3.0.
Unfortunately after saving a calibration, I can not apply it again. There is a message:
"All of short, open and load calibration stepsmust be
completed for calibration to be applied.", see screenshot.
So, once I have left the program I need to calibrate again.

Now I have tried version 0.2.3.1.
Unfortunately in a 4 segment scan the inner 2 segments are missing, see appended screenshot.

73, Rudi DL5FA

Hi Rudi,
I think this bug fix set (calibration and multi-segment) was in flux. My .exe has both fixes, i'm pretty sure, if you want to use that. Holger's latest code is more up to date but releases tend to lag coding. http://www.audua.com/Files/nanovna-saver_6_19.zip

Regards,
Mark

On Tue, Jun 23, 2020 at 09:38 PM, MarkZ wrote:
>
> I think this bug fix set (calibration and multi-segment) was in flux. My .exe
> has both fixes,
>
Hello Mark,
Thank you very much for the hint.
Unfortunately I am using MaxOS, and the newest release 0.2.3.1 for MacOS (Python source) from Holger does not contain those fixes.

73, Rudi DL5FA

On Tue, Jun 23, 2020 at 02:24 PM, <reuterr@web.de> wrote:

Unfortunately I am using MaxOS, and the newest release 0.2.3.1 for MacOS (Python source) from Holger does not contain those fixes.
==========================
Rudi,
Holger's newest Python source code release (from 2 days ago) is 0.2.3.4 at https://github.com/zarath/nanovna-saver/releases/tag/v0.2.3.4-dg5dbh . I just verified it addresses the multi-segment bug which Mark submitted an issue report for. I didn't check the calibration issue you posted about.

- Herb

Hello Holger,

I have tried your nanoVNA-saver fork 0.2.3.4.
Unfortunately after saving a calibration, I can not apply it again. There is a message:
"All of short, open and load calibration stepsmust be
completed for calibration to be applied.", see screenshot.
So, once I have left the program I need to calibrate again.

Because the nanoVNA V2 does send raw data only, a local calibration in nanoVNA-saver
which can be loaded from file is very helpful.

72, Rudi DL5FA

Hello Rudi,

On 24.06.20 06:43, reuterr@web.de wrote:
> I have tried your nanoVNA-saver fork 0.2.3.4.
> Unfortunately after saving a calibration, I can not apply it again.
There is a message:
> "All of short, open and load calibration stepsmust be
> completed for calibration to be applied.", see screenshot.
> So, once I have left the program I need to calibrate again.
you're right.

There is a glitch in the save calibration code which creates
illegal cal files when through and isolation are calibrated.

I'll tell you when I've fixed this issue.

73
Holger, DG5DBH

Hello Holger,

Thank you very much for your quick explanation.
Good success :-)

73, Rudi DL5FA

Hi Rudi,

On 24.06.20 15:53, reuterr@web.de wrote:
> Hello Holger,
>
> Thank you very much for your quick explanation.
> Good success :-)
>
> 73, Rudi DL5FA

The safe reload bug is fixed. But it is still just usable with the "old"
VNAs.
I've done some refactoring at the calibration code, to simplify the code
and to understand how it works. The new code still works fine with the
original NanoVNA and calibration is working with V2, but as soon as I
save and reload calibration data, everything runs wild with the V2.
So further investigation is needed.

I haven't created a new release yet, just source code on development branch.

73
Holger, DG5DBH

On Wed, Jun 24, 2020 at 10:24 PM, Holger Müller wrote:
>
> The safe reload bug is fixed. But it is still just usable with the "old"
> VNAs.
>
Hello Holger,

Thank you for the information.
For me it is not urgent to have it for the V2,
but it would be nice to have it in the future.
Sometimes I got a program crash under MacOS,
and then I need a new calibration, to proceed with measuring.

The official PC program vna-QT (Linux, MacOS, Windows) is
also good, but is some features missing from nanoVNA-saver.

73, Rudi DL5FA

Hi Holger
Great work, thank you.
I have been absent on this group (needs to read 96 messages 😊 ) for some time doing other related activities (TinySA) and just tested your work on W10 64Bit both with the S-A-A 2 and Hugen S-A-A V2_2 and so far super performance.
Kind regards
Kurt

Holger,
I've been complaining for a while that there was something wrong with the through calibration in the NanoVNA-Saver branch for V2. I just checked your repository and noticed that the v0.3.4 change log announced "Fixed bug in Through Calibration".

I upgraded to version 0.3.4 and Through Calibration now works as well in NanoVNA-Saver as it does on the V2 itself and the native NanoVNA-QT application.

Thanks for addressing this issue. I was beginning to wonder if I was the only person experiencing it. If there was an "upgrade V2 firmware" option in NanoVNA-Saver I wouldn't have much need for the NanoVNA-QT application.

Again, thanks - Rune would be proud of you.

- Herb

Hi Herb,

On 10.07.20 02:27, hwalker wrote:
> I upgraded to version 0.3.4 and Through Calibration now works as well
> in NanoVNA-Saver as it does on the V2 itself and the native
> NanoVNA-QT application.
>
v0.3.5 is on the way. There was some stability issues and problems
with multi segment sweeps.

Pre release is looking for testers :-)

https://github.com/NanoVNA-Saver/nanovna-saver/releases/tag/v0.3.5

73
Holger, DG5DBH

A quick question regarding nanovna-saver and calibration ;)

If one calibrate with 30 segments (more than 30k points), 50Kz-3Ghz and save/recall this calibration, do you think its enough to have full range calibrated?

This would help by having less calibration files/procedures.

Or, my question could be, what is your calibration process/flow?

All the best,

Luis CT2FZI

Hi Holger
Downloaded and installed.
Will test
I have a question:
Rune did not know how to implement a shunt C for the load in calibration, there is R. L and delay but not a C shunt.
I did long ago made him a report that there is a difference if including the shunt C as a negative L value:
Do you have a possibility to crack that problem ?
Kind regards
Kurt

Hi Kurt,

On 10.07.20 13:26, Kurt Poulsen wrote:
> Hi Holger
> Downloaded and installed.
> Will test
> I have a question: Rune did not know how to implement a shunt C for
> the load in
calibration, there is R. L and delay but not a C shunt.
> I did long ago made him a report that there is a difference if
including the shunt C as a negative L value:
> Do you have a possibility to crack that problem ?> Kind regards

sorry, I'm not so deep in the coefficient data of calibration
elements and haven't look much around in that part of code.

I even have no calibration elements with measured calibration data
certificates...

73
Holger, DG5DBH

Hi Holger
Fair enough. Hopefully Rune soon get well and find motivation and energy to tune in again. Se owe him a lot credit
Kind regards
Kurt

I tried to start the nanovna-saver.exe, after unpacking the archive with many files. I am getting the starting screen, it identifies the NanoVna v2 (in my case COM10), but after pushing connect it crashes. Am I doing something wrong? Are there any complied exe for the 0.34 and previous versions?
OK1VAW, Vojtech.

Hi Vojtech,

On 10.07.20 20:15, ok1vaw wrote:
> I tried to start the nanovna-saver.exe, after unpacking the archive
> with many files. I am getting the starting screen, it identifies the
> NanoVna v2 (in my case COM10), but after pushing connect it crashes.
> Am I doing something wrong? Are there any complied exe for the 0.34
> and previous versions?
Can be found as assets in the github releases.

But please file an issue at github. Preferable with an attached
debug log. This can be created by nanavna-saver.exe -D debug.txt

73
Holger, DG5DBH

Hello Holger,
I will fill that later, leaving soon for weekend. I have tried the version 0.3.4, there was not this issue, but about few minutes of run the Nanovna_saver freezes and the measurement stops, the Nanovna V2 freezes too with the display in the USB Mode, and I get python error "2020-07-10 22:58:09,400 - NanoVNASaver.NanoVNASaver - ERROR - Tried to open COM10 and failed: could not open port 'COM10': PermissionError(13, 'Přístup byl odepřen.', None, 5) ", the czech (" 'Přístup byl odepřen.'" means Acces was denied). I have keep to run in the same condtions the version 6_19 (on the saver it shows 0.2.3-dg5gbh), although it runs significantly slower, I have no issues with the stability running continously more than 2 hours without any interruption.
The computer is 6core Xeon X5660 and the processor load is negligible, so there should be some issue in communication on USB or between the saver and V2.
Another bug concerns the s21 gain graph on the attached screenshot -x axis auto division. There is a even number of vertical lines, so the center of frequency line is not at 2400MHz center, but away and the x axis divisions are strange and different from the s11 right.

just had the 0.3.5 version running for about two hours ... no problems
found

no com error

just works :-)

dg9bfc sigi

Am 10.07.2020 um 22:02 schrieb ok1vaw:

> Hello Holger,
> I will fill that later, leaving soon for weekend. I have tried the version 0.3.4, there was not this issue, but about few minutes of run the Nanovna_saver freezes and the measurement stops, the Nanovna V2 freezes too with the display in the USB Mode, and I get python error "2020-07-10 22:58:09,400 - NanoVNASaver.NanoVNASaver - ERROR - Tried to open COM10 and failed: could not open port 'COM10': PermissionError(13, 'Přístup byl odepřen.', None, 5) ", the czech (" 'Přístup byl odepřen.'" means Acces was denied). I have keep to run in the same condtions the version 6_19 (on the saver it shows 0.2.3-dg5gbh), although it runs significantly slower, I have no issues with the stability running continously more than 2 hours without any interruption.
> The computer is 6core Xeon X5660 and the processor load is negligible, so there should be some issue in communication on USB or between the saver and V2.
> Another bug concerns the s21 gain graph on the attached screenshot -x axis auto division. There is a even number of vertical lines, so the center of frequency line is not at 2400MHz center, but away and the x axis divisions are strange and different from the s11 right.
>
>
>

tested the pre release ... sees to work fine so far

dg9bfc sigi

Am 10.07.2020 um 10:30 schrieb Holger Müller:

> Hi Herb,
>
> On 10.07.20 02:27, hwalker wrote:
>> I upgraded to version 0.3.4 and Through Calibration now works as well
>> in NanoVNA-Saver as it does on the V2 itself and the native
>> NanoVNA-QT application.
>>
> v0.3.5 is on the way. There was some stability issues and problems
> with multi segment sweeps.
>
> Pre release is looking for testers :-)
>
> https://github.com/NanoVNA-Saver/nanovna-saver/releases/tag/v0.3.5
>
> 73
> Holger, DG5DBH
>
>
>

On Fri, Jul 10, 2020 at 03:30 AM, Holger Müller wrote:

...v0.3.5 is on the way. There was some stability issues and problems with multi segment sweeps.

Pre release is looking for testers :-)
==================================
Holger,
The v0.3.5 RC is working well on both my V2 and H-4 with Windows10, x64.

1. Support for DiSlord's 201pt firmware for the H4 in a dropdown menu (101 or 201) works well.
2. The same dropdown menu support for the V2 works well with even more points up to 1024.
3. Save and recall calibrations is working for both the H4 and V2.
4. Screen shots for the H4 is working well.
5. Through calibration for both the H4 and V2 is fixed.
6. I haven't encountered any errors with multi-segment sweeps, although with the V2 and H4 selectable sweep points I find less of a need to use multi-segment sweeping.
7. I haven't had an opportunity to test any of the other new features that were added to v0.3.5RC.

Good job and thanks for your contributions. The re-factoring in-line with PEP-8, use of "f" strings , and splitting out some of the classes will hopefully make future support of NanoVNA-Saver easier.

- Herb