Hi Guys,
Can anybody explain to me why the Windows binary builds of
NanoVNA-Saver, supporting V2 boards, are of the 64-bit type. After all, this
program doesn't need loads of memory to execute, greater than double
precision arithmetic or super operational speed because the hardware is the
limiting factor.

Most of the Guys at my local Ham-radio club, who make RF'y things where
VNA's are most useful, are like me and use cast-off computers in their
workshop come radio shacks! There is generally no reason for upgrading to
the latest 64-bit I99 multi-core processor machines with 64Gb of memory,
that we'll never use, and Windows bloat-ware Version thingamabob, to enjoy
what we do. Now, to my mind a 32-bit native app will do the job, will also
work on 64-bit operating systems with no measurable improvement. So, please,
can one, or other, of you clever SW persons manage to put out a 32-bit
version of this program which will run on my scrap-yard potable?

You may have gathered that I, and a lot of my fellow Hams, are not all that
interested in b****ring about with Linux and x-compilers to get an
application running on the older generation of Windoze computers.

A response would be welcome.

Regards,

Gary

Hello

The good thing about nanoVNA-saver is that it's author freely and publicly
shared it's work.

Then someone on their free time and knowledge added support to V2.

Now, if you can compile it to 32 bits you can download the code from GitHub
and do it.

Maybe someone already did it for 32 bits?

Cheers

Luis CT2FZI

A terça, 28/07/2020, 18:18, Gary Cobb via groups.io <g3tmg=
talktalk.net@groups.io> escreveu:

> Hi Guys,
> Can anybody explain to me why the Windows binary builds of
> NanoVNA-Saver, supporting V2 boards, are of the 64-bit type. After all,
> this
> program doesn't need loads of memory to execute, greater than double
> precision arithmetic or super operational speed because the hardware is the
> limiting factor.
>
> Most of the Guys at my local Ham-radio club, who make RF'y things where
> VNA's are most useful, are like me and use cast-off computers in their
> workshop come radio shacks! There is generally no reason for upgrading to
> the latest 64-bit I99 multi-core processor machines with 64Gb of memory,
> that we'll never use, and Windows bloat-ware Version thingamabob, to enjoy
> what we do. Now, to my mind a 32-bit native app will do the job, will also
> work on 64-bit operating systems with no measurable improvement. So,
> please,
> can one, or other, of you clever SW persons manage to put out a 32-bit
> version of this program which will run on my scrap-yard potable?
>
> You may have gathered that I, and a lot of my fellow Hams, are not all that
> interested in b****ring about with Linux and x-compilers to get an
> application running on the older generation of Windoze computers.
>
> A response would be welcome.
>
> Regards,
>
> Gary
>
>
>
>
>

Here is a link to how to build or run a 32-bit version.

https://groups.io/g/nanovna-users/topic/75487354#15382

Roger

On 28.07.20 19:21, CT2FZI wrote:
> Hello
>
> The good thing about nanoVNA-saver is that it's author freely and publicly
> shared it's work.
>
> Then someone on their free time and knowledge added support to V2.
>
> Now, if you can compile it to 32 bits you can download the code from GitHub
> and do it.
And even better, you don't need to compile it. You can install the 32Bit
Python Interpreter run the source installation. So you can possibly use
it on arm win10 on a raspberry (not tested).

Reason why there is no 32Bit binary can be found in
https://github.com/NanoVNA-Saver/nanovna-saver/issues/284

Perhaps one can convince Microsoft (the owner of github) to give
us 32Bit Win10 python action images. But as they plan to drop
32Bit support in general, I doubt.

73
Holger, DG5DBH

[Edited Message Follows]

If you want to create your own 32 bit version, it's not that hard... here's how.

I manually run a program called pyinstaller to create a single exe (which is almost what Holger does). It's simple and works fine for me. He doesn't bother bundling up the stuff into a single .exe which almost certainly runs faster but my system is plenty fast.

If you don't want a single bundle, leave off the --onefile

1) Install GIT - this is the source code control system used by github. Available from https://git-scm.com/download/win
2) Install python3. This is available from python.org (https://www.python.org/downloads/release/python-385/). Make sure you add it to the path for simplicity.
3) install pyinstaller:
pip3 install pyinstaller
4) Run a command prompt and
a) clone the github nanovna-saver repository via:
git clone https://github.com/NanoVNA-Saver/nanovna-saver
b) cd nanovna-saver
c) pip3 install . to install the requirements (don't forget that trailing '.')
d) run NanoVNASaver to check that everything works. This will execute python. Assuming it worked...
e) pyinstaller --onefile nanovna-saver.py will bundle everything into the 'dist' folder

Mark

-- so in the 15 minutes it took me to research this there were two answers. hope i didn't step on anyone's toes. M

If you want to create your own 32 bit version, it's not that hard... here's how.

I manually run a program called pyinstaller to create a single exe (which is almost what Holger does). It's simple and works fine for me. He doesn't bother bundling up the stuff into a single .exe which almost certainly runs faster but my system is plenty fast.

If you don't want a single bundle, leave off the --onefile

1) Install GIT - this is the source code control system used by github. Available from https://git-scm.com/download/win
2) Install python3. This is available from python.org (https://www.python.org/downloads/release/python-385/). Make sure you add it to the path for simplicity.
3) install pyinstaller:
pip3 install pyinstaller
4) Run a command prompt and
a) clone the github nanovna-saver repository via:
git clone https://github.com/NanoVNA-Saver/nanovna-saver
b) cd nanovna-saver
c) pip3 install . to install the requirements (don't forget that trailing '.')
d) run NanoVNASaver to check that everything works. This will execute python. Assuming it worked...
e) pyinstaller --onefile nanovna-saver.py will bundle everything into the 'dist' folder

Mark

On 28.07.20 20:48, MarkZ wrote:
> If you want to create your own 32 bit version, it's not that hard...
> here's how.
>
> I manually run a program called pyinstaller to create a single exe
> (which is almost what Holger does). It's simple and works fine for
> me. He doesn't bother bundling up the stuff into a single .exe which
> almost certainly runs faster but my system is plenty fast.
>
> If you don't want a single bundle, leave off the --onefile

Thanks for the info. I'll use onefile in future...

(I general don't bother using binaries when dealing with python)

73
Holger, DG5DBH

[Edited Message Follows]

Hi Mark Z,
Thanks for the help.
I tried and I was able to create one file.
The auto-py-to-exe module uses the same pyinstaller. *

73, Gyula HA3HZ

[Edited Message Follows]

Hi Mark Z,
Thanks for the help.
I tried and I was able to create one file.
The auto-py-to-exe module uses the same pyinstaller.

73, Gyula HA3HZ

Hi Mark Z,
Thanks for the help.
I tried and I was able to create one file.
This will make the file smaller than with the auto-py-to-exe module.

73, Gyula HA3HZ

Gyula, hi :)

Next time you use auto-py-to-exe choose "one file" instead of "One
Directory":

[image: image.png]

I was also able to compile using nanovna-saver icon ;)

[image: image.png]

My version of the nanovna-saver only has the .exe file, no other files or
directories.


Cheers!





On Tue, 28 Jul 2020 at 20:57, Gyula Molnar <gyula.ha3hz@gmail.com> wrote:

> Hi Mark Z,
> Thanks for the help.
> I tried and I was able to create one file.
> This will make the file smaller than with the auto-py-to-exe module.
>
> 73, Gyula HA3HZ
>
>
>
>

[Edited Message Follows]

That's right ... there's only one file, but it's bigger.
The two pictures show you make from this and you made this.

ps.: I just quietly note that I don’t need this because Python runs under Windows, it can be started right away.
I’m only talking to those who only have a Windows 32-bit machine because they can’t use the latest nanovna saver.
Since I didn't have to install it just to test it, I tried it.

That's right ... there's only one file, but it's bigger.
The two pictures show you make from this and you made this.

Today I tested the v0.3.7-pre07 version of the nanovna-saver on github with "auto-py-to-exe".
This module uses the pyinstaller command, so the end result must be found in the output folder.

I did this using Python 3.7.4 and Git previously installed on a Windows7-32 bit machine.
The final .exe file can be run without Python.

73, Gyula HA3HZ

Hello everyone,

I have to agree with Gary on this one. Please, please, please create a 32 bit version. I have been following all of you for about four months. I see great things being worked on and discussed, but, I only understand about 1/3 of it. For many years I was a “power user.” I used and created data bases, spreadsheets, used word processors and created newsletters. But, I did not create those programs. I like observing at this level but I’m not any good at it.

Recently, I asked for some help installing Saver on a 32 bit machine. I got several good responses especially from Gyula. (By the way, great website Gyula.) He gave me step by step instructions. Actually I made good progress on installing Python and Saver on my machines. However, I got to the end and it would not work. See the following:




The above was my second attempt. I also tried on my shack computer with the same results.

Found too many “Access is denied” messages. I’ve gotten over being frustrated. I have been through a very stiff learning curve, but I am stuck again. I would appreciate some help getting beyond this obstacle.

But, here is the point. I bought a nanoVNA to use it. I expected to get the device, download the software and put it to use. For years I have had a miniVNA that I did just that with. I have literally spent weeks trying to install the software. (I’m 73 years old and need a bigger screen) When Gary sent this email he was looking for someone to create a 32 bit version. I think he is like me and does not have the skills to work at that level. I tried and failed. Some of you responded that he should just download another program and create a 32 bit version. I barely have an idea about what that means much less doing it. For some of you it appears that creating a 32 bit version would not be that difficult. Please do. (For the nanoVNA-H and the nanoVNA-V2)

Thank you for your time,

Joe

Joe Joncas WA7MHB
Lincoln County Radio Amateur Club
541-272-3207
joe1@bmihome.org

> On Jul 28, 2020, at 10:16 AM, Gary Cobb via groups.io <g3tmg=talktalk.net@groups.io> wrote:
>
> Hi Guys,
> Can anybody explain to me why the Windows binary builds of
> NanoVNA-Saver, supporting V2 boards, are of the 64-bit type. After all, this
> program doesn't need loads of memory to execute, greater than double
> precision arithmetic or super operational speed because the hardware is the
> limiting factor.
>
> Most of the Guys at my local Ham-radio club, who make RF'y things where
> VNA's are most useful, are like me and use cast-off computers in their
> workshop come radio shacks! There is generally no reason for upgrading to
> the latest 64-bit I99 multi-core processor machines with 64Gb of memory,
> that we'll never use, and Windows bloat-ware Version thingamabob, to enjoy
> what we do. Now, to my mind a 32-bit native app will do the job, will also
> work on 64-bit operating systems with no measurable improvement. So, please,
> can one, or other, of you clever SW persons manage to put out a 32-bit
> version of this program which will run on my scrap-yard potable?
>
> You may have gathered that I, and a lot of my fellow Hams, are not all that
> interested in b****ring about with Linux and x-compilers to get an
> application running on the older generation of Windoze computers.
>
> A response would be welcome.
>
> Regards,
>
> Gary
>
>
>
>

On 01.08.20 00:05, Joe Joncas wrote:
> I have to agree with Gary on this one. Please, please, please create
> a 32 bit version. I have been following all of you for about four
> months. I see great things being worked on and discussed, but, I only
> understand about 1/3 of it. For many years I was a “power user.” I
> used and created data bases, spreadsheets, used word processors and
> created newsletters. But, I did not create those programs. I like
> observing at this level but I’m not any good at it.

https://github.com/NanoVNA-Saver/nanovna-saver/issues/284#issuecomment-667496243

73
Holger, DG5DBH

Hi Holger
Thank you for the link. I work fine apart from the first measurements are not correctly made
See attachments
Kind regards
Kurt

[Edited Message Follows]

Hi Joe,

I understand your frustration. My hope was that one of you 32-bit-required folks would step up and offer to build an .exe every now and then. There's no real urgency to get the latest and greatest code (usually) and so if someone like Gyula could do a 32-bit build then just post it (in the wiki?) that would work fine imho.

Having it in the normal build is clearly a better solution, but it takes time and care to build 32 bit on a 64 bit system (if it's even possible). It's much simpler to do (and ensure working) on a 32 bit computer, which none of 'us' have.

I'll commit to spending some time this week working on the 32 bit build/checking on Github.

Regards,
Mark

Hi Joe,

I understand your frustration. My hope was that one of you 32-bit-required folks would step up and offer to build an .exe every now and then. There's no real urgency to get the latest and greatest code (usually) and so if someone like Gyula could do a 32-bit build then just post it (in the wiki?) that would work fine imho.

Having it in the normal build is clearly a better solution, but it takes time and care to build 32 bit on a 64 bit system (if it's even possible). It's much simpler to do (and ensure working) on a 32 bit computer, which none of 'us' have.

I'll commit to spending some time this week investigating how to do this reliably for Holger.

Regards,
Mark

Hi Mark,

On 01.08.20 12:56, MarkZ wrote:
>
> I'll commit to spending some time this week working on the 32 bit
> build/checking on Github.

I've done already

see

https://github.com/NanoVNA-Saver/nanovna-saver/actions?query=workflow%3A%22Windows+Release%22

73
Holger, DG5DBH

Hi Holger
Could you please have a look of what is wrong when more than one segement used or if in "Manage" the number of points selected are above 101 points (the alternative way instead of using number of segments). The problem is that then the first measurement is way off.
Kind regards
Kurt

Hi Kurt,

On 01.08.20 12:52, Kurt Poulsen wrote:
> Hi Holger Thank you for the link. I work fine apart from the first
> measurements are not correctly made See attachments Kind regards

I don't get the point, what's exactly wrong with first measurments?
Do you mean the first sweep which is taken?

Can you give me some more precise information?

73
Holger, DG5DBH

Hi Holger
No, if you look closely on the sweep of the UT141A rigid line at the very first point at the start frequency it is placed well above the 0 dB level where it should start. Then one step from start frequency (measurement point 2) then it correctly running from this point onwards correctly on the 0dB line. That only happens when more that one segment or in the "Manage" you choose the customer points beyond 51 or 101. For measurement points at or lower than 101 the first measurement point is spot on 0dB.
Kind regards
Kurt

[Edited Message Follows]

Hi Joe,
OK, it will go, you just gave the start command in the wrong place.
After installing Python and the modules then exit "cmd" mode.
Download the "Source code.zip" of the latest version from the nanovna-saver release page. link: https://github.com/NanoVNA-Saver/nanovna-saver/releases
Unpack this to drive C: / as administrator.
Then use a file manager to access the C: /nanovna-saver-0.3.6 directory (or the name you downloaded). Click nanovna-saver.py to start it. A small window will appear, after which the program itself will log in. I think maybe you may have made a mistake here that you wanted to boot from the "cmd" window.
Please reply, if it still doesn't work, I'll try to help.

73, Gyula HA3HZ

ps.:
Note: I am happy to help anyone who asks.
To do this, if you ask a question and someone answers, at least all you have to do is answer for the helper.
You can also do this directly by clicking on "Reply" at the end of the message you want to write to and then write the message and find a "Private" word in the lower right corner. Clicking on it will not take you to the forum, but to that email address. From then on, you can talk to each other, you don’t have to wait days.
Of course, if you regularly view your messages (email)
If you just want to complain that it doesn’t work for me, I read it and don’t react to it.

[Edited Message Follows]

Hi Joe,
OK, it will go, you just gave the start command in the wrong place.
After installing Python and the modules then exit "cmd" mode.
Download the "Source code.zip" of the latest version from the nanovna-saver release page. link: https://github.com/NanoVNA-Saver/nanovna-saver/releases
Unpack this to drive C: / as administrator.
Then use a file manager to access the C: /nanovna-saver-0.3.6 directory (or the name you downloaded). Click nanovna-saver.py to start it. A small window will appear, after which the program itself will log in. I think maybe you may have made a mistake here that you wanted to boot from the "cmd" window.
Please reply, if it still doesn't work, I'll try to help.

73, Gyula HA3HZ

Hi Joe,
OK, it will go, you just gave the start command in the wrong place.
After installing Python and the modules then exit "cmd" mode.
Download the "source.zip" of the latest version from the nanovna-saver release page. link: https://github.com/NanoVNA-Saver/nanovna-saver/releases
Unpack this to drive C: / as administrator.
Then use a file manager to access the C: /nanovna-saver-0.3.6 directory (or the name you downloaded). Click nanovna-saver.py to start it. A small window will appear, after which the program itself will log in. I think maybe you may have made a mistake here that you wanted to boot from the "cmd" window.
Please reply, if it still doesn't work, I'll try to help.

73, Gyula HA3HZ

Hi Holger and Herb,
Apologies, I can download from the releases page but not from page
you pointed at. I'll wait until it becomes a release!

Cheers
Gary

Hi Kurt,

On 01.08.20 16:20, Kurt Poulsen wrote:
> No, if you look closely on the sweep of the UT141A rigid line at the
> very first point at the start frequency it is placed well above the 0
> dB level where it should start. Then one step from start frequency
> (measurement point 2) then it correctly running from this point
> onwards correctly on the 0dB line. That only happens when more that
> one segment or in the "Manage" you choose the customer points beyond
> 51 or 101. For measurement points at or lower than 101 the first
> measurement point is spot on 0dB.

Wow, that's a good hint. I've seen this first displaced first point
in each segment before, but I don't found out why.
I suspect a firmware problem... but I haven't looked with lower point
count till now. As far as I know and what I've seen from -H4 Charts
with 201 points, it happens only with the V2.
That's a good hint to look at ...

Thanks!

73
Holger, DG5DBH

Hi Holger
Right that is only with the S-A-A-2 and Hugen S-A-A-2_2
I attach measurement made by the SA_QT software which does not shown any funny behavior apart for very small deviation at the very low frequencies
Kind regards
Kurt

Hi Kurt,

On 01.08.20 22:00, Kurt Poulsen wrote:
> Hi Holger Right that is only with the S-A-A-2 and Hugen S-A-A-2_2 I
> attach measurement made by the SA_QT software which does not shown
> any funny behavior apart for very small deviation at the very low
> frequencies Kind regards Kurt

see my attached screenshot of an uncalibrated scan with VNA_QT
from 0.5M to 300M with approx. 40cm of RG58.

Also that deviation... no matter which point setting...

No problems with frequencies > 30M at all..

73
Holger, DG5DBH

The deviation is because of the coupler design, which relies on a pair
of baluns to extract a difference signal. Baluns are not effective below
a few MHz, which the firmware corrects for but some deviations remain.
After calibration the deviations should be gone.

Hi Gabriel,

On 03.08.20 10:43, Gabriel Tenma White wrote:
> The deviation is because of the coupler design, which relies on a pair
> of baluns to extract a difference signal. Baluns are not effective below
> a few MHz, which the firmware corrects for but some deviations remain.
> After calibration the deviations should be gone.

Yes, calibration works.

There were sometimes phase jumps in the very first sweep point on the
S21 value.

The picture is with 10 segments on NanoVNA-Saver.

It's always the first value, regardless of point count and so on,
do you have an idea, where this can result from?

73
Holger, DG5DBH

Looks like stale data are getting in. After setting sweep parameters the
software needs to clear the on-device FIFO and discard at least one more
point. I'll take a look at the code in the new nanovna-saver.

Hi Gabriel,
On 03.08.20 10:59, Holger Müller wrote:
>
> There were sometimes phase jumps in the very first sweep point on the
> S21 value.

sorry, not the phase, the magnitude of first meassured S21 Point
is always to high.

Can be seen on S21mag (or better on S21vswr) graph in vna_qt, too.

73,
Holger, DG5DBH

Hello again....
hi Gabriel,
On 03.08.20 11:11, Holger Müller wrote:
> Hi Gabriel,
> On 03.08.20 10:59, Holger Müller wrote:
>> There were sometimes phase jumps in the very first sweep point on the
>> S21 value.
> sorry, not the phase, the magnitude of first meassured S21 Point
> is always to high.

No problems with preliminary firmware 20200619

73
Holger, DG5DBH

On 03.08.20 14:00, Holger Müller wrote:
>> Hi Gabriel,
>> On 03.08.20 10:59, Holger Müller wrote:
>>> There were sometimes phase jumps in the very first sweep point on the
>>> S21 value.
>> sorry, not the phase, the magnitude of first meassured S21 Point
>> is always to high.
> No problems with preliminary firmware 20200619
I have to revert....

still happens, but not always ...

73
Holger, DG5DBH