Hi everyone;

Does some guys are motivated to convert/rewrite the NanoVNA-V2 schematics/board to KiCad ?
As actual format done with Geda is clearly not usable ...
Lot of improvement could be done if it will be converted to KiCad which is really more advanced than geda (which is stopped since lot of years) to do RF/microwave stuff.

Best Regards,
Benjamin

Where do you find the schematics?

—
Chris

> Am 12.05.2020 um 21:37 schrieb Benjamin Vernoux <bvernoux@gmail.com>:
>
> Hi everyone;
>
> Does some guys are motivated to convert/rewrite the NanoVNA-V2 schematics/board to KiCad ?
> As actual format done with Geda is clearly not usable ...
> Lot of improvement could be done if it will be converted to KiCad which is really more advanced than geda (which is stopped since lot of years) to do RF/microwave stuff.
>
> Best Regards,
> Benjamin
>
>
>


--
Chris
--
I am not responsible for typos... they always respond to the bad robot...

https://gitlab.com/bepissneks/S-A-A

On Tue, 12 May 2020 at 23:41, dekagon <ck@cksd.de> wrote:

> Where do you find the schematics?
>
>

It seems schematic/board is available here
https://github.com/nanovna/S-A-A
Latest version (done with gEDA pcb 4.2.0 available here http://pcb.geda-project.org) produced seem to be
https://github.com/nanovna/S-A-A/blob/master/v2_2.sch
https://github.com/nanovna/S-A-A/blob/master/v2_2.pcb
https://github.com/nanovna/S-A-A/blob/master/v2_2.gsch2pcb

Gabriel Tenma White: You are welcome to clarify what is latest schematic/board used for the production.

The original thread is here https://groups.io/g/nanovna-users/topic/73993300#13216

But it is not clear which version is the good one and if it corresponds the actual version sold...

Regards,
Benjamin

Thank you for the links :)

Chris

> Am 12.05.2020 um 23:51 schrieb Benjamin Vernoux <bvernoux@gmail.com>:
>
> It seems schematic/board is available here
> https://github.com/nanovna/S-A-A
> Latest version (done with gEDA pcb 4.2.0 available here http://pcb.geda-project.org) produced seem to be
> https://github.com/nanovna/S-A-A/blob/master/v2_2.sch
> https://github.com/nanovna/S-A-A/blob/master/v2_2.pcb
> https://github.com/nanovna/S-A-A/blob/master/v2_2.gsch2pcb
>
> Gabriel Tenma White: You are welcome to clarify what is latest schematic/board used for the production.
>
> The original thread is here https://groups.io/g/nanovna-users/topic/73993300#13216
>
> But it is not clear which version is the good one and if it corresponds the actual version sold...
>
> Regards,
> Benjamin
>
>
>


--
Chris
--
I am not responsible for typos... they always respond to the bad robot...

On Tue, May 12, 2020 at 12:37 PM, Benjamin Vernoux wrote:

Does some guys are motivated to convert/rewrite the NanoVNA-V2 schematics/board to KiCad ?
As actual format done with Geda is clearly not usable ...
====================================================
hugen designed a working S-A-A-2 pcb, converting from gEDA to easyEDA fairly quickly. He might consider collaborating on a 4" S-A-A-2 design using KiCad since most open source projects use it. easyEDA has worked well for him on his NanoVNA-H, NanoVNA-H4 and tinSA projects. I guess its whatever you are familiar with.

- Herb

On Tue, May 12, 2020 at 02:41 PM, dekagon wrote:

Where do you find the schematics?
============================================
I uploaded the version 2.2 schematic in PDF format to the "Files" section ( https://groups.io/g/NanoVNAV2/files/User%20Manuals,%20How2,%20Examples/Gabriel%20v2_2%20schematic.pdf ).

Courtesy of Gabriel Tena White

- Herb

Herb,
I only asking because of my own interest. I want to build one for my own with a greater display and a Desktop housing. So my interest is whether I could split the board into two or maybe three boards which are separated from each other.
An RF board and the cpu board with usb c and the power board. The Display board is still separated only with a connector used.
I am also be aware that this vna will cost much more as the good and smart design from Gabriel but my goal is not to create another mobile version. I do want to have finally a desktop variant with compatibility to all the software as VNAsaver and so on.
If I will be successfully the another goal will be the firmware compatibility with the exception that the display driver section must be eventually modular in the firmware to easy adapt another lcd formats and chips. The solution would be to build in the support for several driver chips with the suitable fonts and build the apropriate binary for the fitting vna model...
And important for me is the fact that I want to get Gabriels permission to use the schematics and her files for that.
If all will be successful and it could be a selling project I surely will pay for the permission to sell some. But this is a future thing.

Kind regards
Chris
--
I am not responsible for typos... they always respond to the bad robot...

Dear Chris

agree with the principle of a desktop version - and potentially one can make use of a standard STM32XXXX Discovery type CPU board and only develop a carrier/motherboard for the Display and RF side to save time.

I support this direction and can potentially assist, having done both STM32 code porting and KICAD PCB design in the last years

Deon

I order this display module
https://aliexpress.ru/item/4000946111830.html?spm=a2g0s.9042311.0.0.61a533ed2iPjHf

And then recive SAA-2 replace it.
Also update code for use it

Hello Deon,

thank you for your answer.
But first I have to wait for Gabriels permission...

Chris

---
'Everything should be made as simple as possible, but not simpler.' -- Albert Einstein

Am 13.05.2020 um 10:33 schrieb randmental:

> Dear Chris
>
> agree with the principle of a desktop version - and potentially one can make use of a standard STM32XXXX Discovery type CPU board and only develop a carrier/motherboard for the Display and RF side to save time.
>
> I support this direction and can potentially assist, having done both STM32 code porting and KICAD PCB design in the last years
>
> Deon
>

Dislord,

it seems that you ordered the same as mine...
But I used another shop:
https://www.aliexpress.com/item/4000234986822.html?spm=a2g0s.9042311.0.0.7f2a4c4dDlAHsm

It would be a great help if you adapt the code for using a 4" lcd.
Do you already know whether you make a separate code fork for the 4"
version or modularize the V2 code
for using 2.8, 3.2 and 4.0" displays?

Chris
---

'Everything should be made as simple as possible, but not simpler.' -- Albert Einstein

Am 13.05.2020 um 11:25 schrieb DiSlord:
> I order this display module
> https://aliexpress.ru/item/4000946111830.html?spm=a2g0s.9042311.0.0.61a533ed2iPjHf
>
> And then recive SAA-2 replace it.
> Also update code for use it
>
>
>



--
Chris
--
I am not responsible for typos... they always respond to the bad robot...

Display module 2.8 or 3.2 use some driver and resolution 320x240 (ili9341) nothing need update

Display module 4` use different driver ST7796S (some as H4 version)

Sensor panel driver on both modules some (XPT2046)

ST7796S and Ili9341 commands some
Difference only in init

Problem on output, for H version i revrite plot module code for ease change resolution in config, need port this code (Gabriel do it in https://github.com/nanovna/NanoVNA-V2-firmware/tree/ui-changes-dislord) and adaptate for bigger resolution

For test possible only replace module, and replace init sequence (you can see it in H and H4 code)

On Wed, May 13, 2020 at 02:33 AM, dekagon wrote:

... But first I have to wait for Gabriels permission...
===============================
Gabriel previously posted the following when someone asked about using her open source design:
"Yeah sure, just change up the layout a bit, space things out, try to make it look like it's not a verbatim copy ;)"

She also posted the following:
"The PCB+partial PCBA would cost around ~$50 for 2pcs, and only slightly more for 5pcs. That will give you a board with one side populated (all the passives other than the baluns and the FBMH1608 ferrite beads). However it is a 5x10cm design, and you can fit in another 5x10cm design without additional PCB cost. Make sure you select the 0.2mm prepreg stackup (JLC7628).
Usually when I do prototyping runs I will combine many projects onto one board. If they are only separated by a silk line (no v-cut or long slots) there is no extra charge).
The commit 39d648 is the "golden" production version."

- Herb

Hi everyone,

After check the best quality/price for a good screen 4 inch (800x480) with a new STM32 MCU
My proposal is
- Use Value line STM32H750 (H7@480MHz ultra cheap available for less than 5USD for 1 unit on https://lcsc.com/product-detail/ST-Microelectronics_STMicroelectronics-STM32H750VBT6_C404010.html and more than 10 times faster than the STM32F1...) +
- Use 4 inch TFT LCD screen touch screen module IPS full view Ultra HD 800X480 like this one https://www.aliexpress.com/item/4000376042215.html which is clearly better quality than other screen for very low price

The aim is to design and build something better/faster (especially with bigger screen) than what exist for low cost just under 100USD
Advantage of moving directly to IPS TFT 800x480 is later we can use other size with same code (as anyway a refactoring is mandatory for any screen/resolution bigger than what exist) and we can also use smaller resolution by HW rescaling (supported natively by STM32H750) with clever code it will need only to manage pictures in hi-resolution 800x480 which can work on different screen size ...

Best Regards,
Benjamin

My ordered 4“ lcds are on the way. I think I get them this week. But the V2 are still stucked anywhere...
So for testing I must wait till the v2‘s are here too.

Btw I like the rewriting of your plot module. It is a better way for adapting lcd modules.

--
Chris
--
I am not responsible for typos... they always respond to the bad robot...

Herb,
this sounds very good... :) So now we must convert the schematics to Kicad otherwise I must learn to use easyEDA ;)

--
Chris
--
I am not responsible for typos... they always respond to the bad robot...

@Benjamin Vernoux: For the sub-$100 price point I'd aim for better RF performance as well (higher sweep rates enabled by lower trace noise), which would most likely mean using an FPGA because much higher IF frequencies are needed, possibly 2.5MHz. That would require at least 10 to 20MHz sample rate ADCs. The Spartan 6 FPGA costs around $4, and there are very cheap ADCs at 20MHz sample rate. Can you find a STM32 with a 10Msps ADC for a good price?

In the short term there will probably be a V2 with a 4'' display, but that will likely still use the same F303 microcontroller, which should be fast enough given that both the -F and -H4 use an equivalent microcontroller.

The "desktop" version of the VNA sounds interesting, but I think we might as well also upgrade the RF design. So the hforsten VNA would be a better starting point and not the S-A-A-2. I **may** end up designing a "proper" VNA with only focus on the best performance and less regards for cost, but of course such a design would never be manufactured for sale and will be DIY only. I have too many other projects to work on right now but I will get to it eventually.

Hi Gabriel,

Speaking about better RF performance for low cost for a V3.
I would heavily suggest the amazing low cost (less than 10USD for 1unit) multi-core LPC4370 (FET100) (instead of an expensive FPGA+external ADC) that I'm using in my products (I have designed the HW, FW and host tools) AirSpy R0/R2 and AirSpy Mini, it is clearly unbeatable price with 12bits ADC max 80MSPS which can be easily driven in real-time over USB HS (480Mbit/s) with 20MSPS like done in AirSpy R2.

For more details on AirSpy Firmware which could be reused for a very good VNA
https://github.com/airspy/airspyone_firmware
https://github.com/airspy/airspyone_firmware/wiki/AirSpy-firmware-directory-structure-and-software-architecture

We could even think about a platform with LPC4370 (mainly for ADC+USB HS) + STM32H750 (for heavy real-time computation+display) or doing all just with LPC4370 as it contains 2 CortexM0 + 1 Cortex M4F @max 204MHz
Any feedback is welcome

Best Regards,
Benjamin

Gabriel,

I fully agree to your arguments.
One of my thoughts were to rework the RF board for a higher RF
performance. But I are not sure too whether the STM cpus are
the right choice for the neccessary sampling speeds. The change does not
have to go for higher noise ratios.
The usable performance have also getting lower to -100 db in the higher
frequency area.
So you are right with hforsten design. It goes up to 6 GHz and looks
also very interesting.
A full 2-port version as from him is also the goal.
For me personally the mobility is not the point. Therefore I got the
plan to make a desktop variant...
It would be nice if you get some time for help...

Chris

---
'Everything should be made as simple as possible, but not simpler.' -- Albert Einstein

Am 13.05.2020 um 13:04 schrieb Gabriel Tenma White:
> @Benjamin Vernoux: For the sub-$100 price point I'd aim for better RF performance as well (higher sweep rates enabled by lower trace noise), which would most likely mean using an FPGA because much higher IF frequencies are needed, possibly 2.5MHz. That would require at least 10 to 20MHz sample rate ADCs. The Spartan 6 FPGA costs around $4, and there are very cheap ADCs at 20MHz sample rate. Can you find a STM32 with a 10Msps ADC for a good price?
>
> In the short term there will probably be a V2 with a 4'' display, but that will likely still use the same F303 microcontroller, which should be fast enough given that both the -F and -H4 use an equivalent microcontroller.
>
> The "desktop" version of the VNA sounds interesting, but I think we might as well also upgrade the RF design. So the hforsten VNA would be a better starting point and not the S-A-A-2. I **may** end up designing a "proper" VNA with only focus on the best performance and less regards for cost, but of course such a design would never be manufactured for sale and will be DIY only. I have too many other projects to work on right now but I will get to it eventually.
>
>
>



--
Chris
--
I am not responsible for typos... they always respond to the bad robot...

Benjamin,

I would also prefer a non FPGA solution because we have to give a
complete new code structure a new life.
Compatibility to the main forks of the nanovna-v2 code would be better
but gives us not the neccessary flexibility.

Chris

---
'Everything should be made as simple as possible, but not simpler.' -- Albert Einstein

Am 13.05.2020 um 13:16 schrieb Benjamin Vernoux:
> Hi Gabriel,
>
> Speaking about better RF performance for low cost for a V3.
> I would heavily suggest the amazing low cost (less than 10USD for 1unit) multi-core LPC4370 (FET100) (instead of an expensive FPGA+external ADC) that I'm using in my products (I have designed the HW, FW and host tools) AirSpy R0/R2 and AirSpy Mini, it is clearly unbeatable price with 12bits ADC max 80MSPS which can be easily driven in real-time over USB HS (480Mbit/s) with 20MSPS like done in AirSpy R2.
>
> For more details on AirSpy Firmware which could be reused for a very good VNA
> https://github.com/airspy/airspyone_firmware
> https://github.com/airspy/airspyone_firmware/wiki/AirSpy-firmware-directory-structure-and-software-architecture
>
> We could even think about a platform with LPC4370 (mainly for ADC+USB HS) + STM32H750 (for heavy real-time computation+display) or doing all just with LPC4370 as it contains 2 CortexM0 + 1 Cortex M4F @max 204MHz
> Any feedback is welcome
>
> Best Regards,
> Benjamin
>
>
>



--
Chris
--
I am not responsible for typos... they always respond to the bad robot...

Since Gabriel is already doing a SDR similar to Pluto how big of an
overlap, in basic building blocks, is there between a SDR and a VNA?
HPSDR has a VNA functionality but I'm not sure how applicable that is to
this case.

Speaking about better RF performance for low cost for a V3.

> I would heavily suggest the amazing low cost (less than 10USD for 1unit)
> multi-core LPC4370 (FET100) (instead of an expensive FPGA+external ADC)
> that I'm using in my products (I have designed the HW, FW and host tools)
> AirSpy R0/R2 and AirSpy Mini, it is clearly unbeatable price with 12bits
> ADC max 80MSPS which can be easily driven in real-time over USB HS
> (480Mbit/s) with 20MSPS like done in AirSpy R2.
>
> For more details on AirSpy Firmware which could be reused for a very good
> VNA
> https://github.com/airspy/airspyone_firmware
>
> https://github.com/airspy/airspyone_firmware/wiki/AirSpy-firmware-directory-structure-and-software-architecture
>
> We could even think about a platform with LPC4370 (mainly for ADC+USB HS)
> + STM32H750 (for heavy real-time computation+display) or doing all just
> with LPC4370 as it contains 2 CortexM0 + 1 Cortex M4F @max 204MHz
> Any feedback is welcome
>
> Best Regards,
> Benjamin
>
>