开云体育

I have been working with Hugen for the past two months to port the STM32F303CCT6 and 4" LCD for new NanoVNA-H. The original author, edy555, also provided valuable suggestions.

Software porting is about 90% finished. Some display and jog switch related UI still needs to be fine tuned. Please check my porting at



Hugen keeps improving the hardware design. Here are some features that we have discussed.
- STM32F303CCT6, which has 128KB flash memory, 40KB SRAM, and 8KB CCM, running at 72MHz. Thanks to F3's FPU, the compiled code size can be even smaller than the original F072's.
- 4" LCD 320x480, driven by ST7796S through SPI interface. Even with twice the pixel numbers than ili9341, the sweep refresh rate remains the same as original one, becasue of F303's higher clock rate.
- UART pins, which can be used for Bluetooth and firmware update. Any volunteer for the UART porting and mobile app?
- Larger battery to support larger LCD.
- The option of plastic and metal cases.
- The option of SMA and N connectors.
- Miscellaneous circuit improvements
* Improved analog front-end. See
* New power management chip and circuit.
* Enter DFU by pressing jog switch at power on. Jumping into DFU by software doesn't work for F303 yet.
* Option of hardware reset codec chip by MCU. This will help fix the power on issue.
* Dimmable LCD back light.

Hugen hasn't finalized his design yet. Your voice will help him converge the feature list faster. I hope that Hugen will freeze and release his design soon so that we can have new gadgets to play with.

73
Ken, AA6KL

In his past posts hugen said he believed a larger display unit would not market well. I'm glad to see that he is now considering not only a unit with a larger display but better processor also. I think the sales of the NanoVNA-F shows that a larger display is definitely the way to go. Hope design changes can also be made to the front end to achieve a lower noise floor.

- Herb

I vote for the larger display. I have one of each and can say
the larger display is much better.

73, Dick, W1KSZ
________________________________
From: [email protected] <[email protected]> on behalf of hwalker <herbwalker2476@...>
Sent: Thursday, November 21, 2019 8:44 AM
To: [email protected] <[email protected]>
Subject: Re: [nanovna-users] F303 and 4" LCD for next generation NanoVNA #circuit #flash_size #improvement #enclosure #battery

In his past posts hugen said he believed a larger display unit would not market well. I'm glad to see that he is now considering not only a unit with a larger display but better processor also. I think the sales of the NanoVNA-F shows that a larger display is definitely the way to go. Hope design changes can also be made to the front end to achieve a lower noise floor.

- Herb

On Thu, 21 Nov 2019 at 15:12, Ken Liao, AA6KL <kuohsing@...> wrote:

I have been working with Hugen for the past two months to port the
STM32F303CCT6 and 4" LCD for new NanoVNA-H. The original author, edy555,
also provided valuable suggestions.

Good.

Software porting is about 90% finished. Some display and jog switch
related UI still needs to be fine tuned. Please check my porting at



Hugen keeps improving the hardware design. Here are some features that we
have discussed.
- STM32F303CCT6, which has 128KB flash memory, 40KB SRAM, and 8KB CCM,
running at 72MHz. Thanks to F3's FPU, the compiled code size can be even
smaller than the original F072's.
- 4" LCD 320x480, driven by ST7796S through SPI interface. Even with
twice the pixel numbers than ili9341, the sweep refresh rate remains the
same as original one, becasue of F303's higher clock rate.

Why not use a larger display? The nanoVNA-F uses 4.3", and I know it's only
0.3", but it is nearly 10% at this size.

Also, make it possible to change the display if it gets damaged, without
needing to desolder, as this is a weak point. Use a connector.

The option of N connectors sounds good. But female N's need will need
proper support for calibration kits, which current firmware lacks. The
centre pin of the female N sits more than 5 mm higher than the reference
plane, (5 mm nearer the front of the connector, so this should be
compensated for).

- UART pins, which can be used for Bluetooth and firmware update. Any

volunteer for the UART porting and mobile app?

- Larger battery to support larger LCD.
- The option of plastic and metal cases.
- The option of SMA and N connectors.

N connectors are most welcome.

- Miscellaneous circuit improvements

* Improved analog front-end. See

* New power management chip and circuit.
* Enter DFU by pressing jog switch at power on. Jumping into DFU by
software doesn't work for F303 yet.

* Option of hardware reset codec chip by MCU. This will help fix the

power on issue.
* Dimmable LCD back light.

It would be good if there could be headers to use in place of the toggle
swtich, and USB connectors, so people can put in their own case and use
decent quality switches, so things don't break.

Hugen hasn't finalized his design yet. Your voice will help him converge
the feature list faster. I hope that Hugen will freeze and release his
design soon so that we can have new gadgets to play with.

73
Ken, AA6KL

Dave, G8WRB

Dimmable LCD back light.

Being able to completely shut off LCD backlight (and refresh, but that is only firmware)
if/when driving from BLE to save power.

Thanks for the update, Ken!

Hopefully, you are able to make the jog switch more responsive by using interrupts instead of polling.

As for the power circuits, it is nice to be able to use a single 3.7v Li cell with an inverter but in order to get noise levels down (you want a 3GHz top end), using two cells and running 7.2v through a low noise analogue regulator to 5V and then to 3.3V will remove most power supply noise.

The down side of 2 cells is requiring >8V for charging the batteries (and price) although with really good filtering, the unit can still be powered from the USB port.

Just some thoughts...
Larry

On Fri, Nov 22, 2019 at 12:06 AM, Dr. David Kirkby, Kirkby Microwave Ltd wrote:

- STM32F303CCT6, which has 128KB flash memory, 40KB SRAM, and 8KB CCM,

Correct, STM32F303CCT6 has 256KB FLASH.

hugen

Thanks Hugen's correction. And special thanks to Oristo8 and Larry for maintaining the wiki pages for us to quickly look for useful information.

Regarding the display size, NanoVNA-F's 4.3" IPS LCD is a nice one. But it comes at the cost of price, large battery, and extra weight. Also, its 800x480 resolution requires parallel interface to reach sub-second refresh rate. That's why NanoVNA-F uses 100-pin STM32F103VET6. Another type of 4.3" LCD has 480x270 resolution with parallel interface only. ILI9488 only supports 24-bit RGB at SPI interface, and ILI9486 is reaching end of production. So it seems that this ST7796S based 4" LCD is the best choice in term of price, weight, performance, and availability. Hugen has installed a FFC connector at the prototype, so it should be easy to replace the LCD if desired.

The LCD back light is driven by a DAC pin and a NPN BJT. So it should easy to add firmware support to control back light later.

For the SMA or N-type socket, it's better to have IPX/IPEX/UFL sockets on PCB, and screw the SMA or N-type sockets at the case. This will help prevent the PCB get damaged from screw/unscrew connectors. Then, the female N-type center pin height issue mentioned by Dr. David Kirkby can be solved too.

The jog switch is a difficult one. Edy555 uses GPIO interrupt and software debounce already. I replaced job switch's pull-up resistor from 10K ohm to 1K ohm. It seems to respond better, although not perfect yet. If the pins are exposed, we could either connect to external buttons or rotary encoder.

Regarding the power supply, I would hope to have a way to bypass the voltage pump circuit and use external 5V power supply when quiet power is needed.

About the BLE support, besides the VCC, GND, RX, and TX, do we also need other pins like SELECT, STATE and EN?

Ken

On Fri, 22 Nov 2019 at 15:11, Ken Liao, AA6KL <kuohsing@...> wrote:

Regarding the display size, NanoVNA-F's 4.3" IPS LCD is a nice one. But
it comes at the cost of price, large battery, and extra weight. Also, its
800x480 resolution requires parallel interface to reach sub-second refresh
rate. That's why NanoVNA-F uses 100-pin STM32F103VET6. Another type of
4.3" LCD has 480x270 resolution with parallel interface only. ILI9488 only
supports 24-bit RGB at SPI interface, and ILI9486 is reaching end of
production. So it seems that this ST7796S based 4" LCD is the best choice
in term of price, weight, performance, and availability. Hugen has
installed a FFC connector at the prototype, so it should be easy to replace
the LCD if desired.

What is the resolution of the 4” display you intend using? Is it less than
the the NanoVNA-F? If so, I would not buy one.

The NanoVNA-F comes with a 5600 mA hr battery, which I believe will run the
unit for about 18 hours. It is not that expensive.

Screen size is the biggest problem for me with the NanoVNA, so I am willing
to pay extra for a better screen.

For the SMA or N-type socket, it's better to have IPX/IPEX/UFL sockets on

PCB, and screw the SMA or N-type sockets at the case. This will help
prevent the PCB get damaged from screw/unscrew connectors. Then, the
female N-type center pin height issue mentioned by Dr. David Kirkby can be
solved too.

No, the problem I mentioned can *not* be solved by that. The problem is
that the centre pin of a female N sticks about 5 mm in front of the
reference plane. This is in the standard for an N connector - it is 0.207”,
which is just over 5 mm. So if you leave a female N connector open, it will
not have a zero phase angle.

I just calibrated an 8753ES which has adapters to female N. I then removed
everything from the female N connector. I attach a Smith Chart view as
well as a phase plot. One can see that at 1.5 GHz, the phase is -20.162
degrees. So if you assume that an open female has a phase of 0 degrees,
there will be significant error. Probably around 12 degrees at 900 MHz.

Any male short standard would be nothing like 180 degrees. So the lack of a
properly implemented calibration routine would cause *very significant
errors* with the N connector.

The jog switch is a difficult one. Edy555 uses GPIO interrupt and
software debounce already. I replaced job switch's pull-up resistor from
10K ohm to 1K ohm. It seems to respond better, although not perfect yet.
If the pins are exposed, we could either connect to external buttons or
rotary encoder.

That sounds good.

Regarding the power supply, I would hope to have a way to bypass the
voltage pump circuit and use external 5V power supply when quiet power is
needed.

About the BLE support, besides the VCC, GND, RX, and TX, do we also need
other pins like SELECT, STATE and EN?

Ken

Dave

--

Dr. David Kirkby,
Kirkby Microwave Ltd,
drkirkby@...

Telephone 01621-680100./ +44 1621 680100

Registered in England & Wales, company number 08914892.
Registered office:
Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United
Kingdom

Thanks for the info

A few thoughts:
EN for the ble module could be useful to shut it down as much as possible to lower the noise, or should we cut off its VCC in this case a p-chan mosfet on the VCC line could do.

If 2 batteries, then a boost converter could be used for charging them from the 5v. Boost converter is a mosfet, diode, coil, can be driven by the stm32 and monitored as well by the stm32 using an adc input.

Regards
Jose

On Fri, 22 Nov 2019 at 15:32, David J Taylor via Groups.Io <gm8arv=
[email protected]> wrote:

For me, having the extra resolution of the 4.3-inch display would be a
significant benefit, and the extra size would not be unwelcome. More than
101 points per plot would be possible. The increased weight would be more
than offset by increased usage time. We already see some of these
advantages in the NanoVNA-F.

73,
David GM8ARV

I agree with David - weight is not an issue.

I think if Hugen wants to sell these in a reasonable quantity, I think the
screen will need to be at least as big as the NanoVNA-F, as I think many
people will buy the NanoVNA-F for a better screen.


Dave.
--
Dr. David Kirkby,
Kirkby Microwave Ltd,
drkirkby@...

Telephone 01621-680100./ +44 1621 680100

Registered in England & Wales, company number 08914892.
Registered office:
Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United
Kingdom

I agree with David. I bought a Nano VNA which is very good for the money but my eyes are not good enough for it to be of any use as a portable instrument so I immediately ordered a NanoVNA-F purely for the larger screen. The NanoVNA is now primarily used with the PC.

73

Conrad PA5Y

________________________________
From: [email protected] <[email protected]> on behalf of Dr. David Kirkby, Kirkby Microwave Ltd via Groups.Io <drkirkby@...>
Sent: 22 November 2019 17:30
To: [email protected] <[email protected]>
Subject: Re: [nanovna-users] F303 and 4" LCD for next generation NanoVNA #circuit #flash_size #improvement #enclosure #battery

On Fri, 22 Nov 2019 at 15:32, David J Taylor via Groups.Io <gm8arv=
[email protected]> wrote:

For me, having the extra resolution of the 4.3-inch display would be a
significant benefit, and the extra size would not be unwelcome. More than
101 points per plot would be possible. The increased weight would be more
than offset by increased usage time. We already see some of these
advantages in the NanoVNA-F.

73,
David GM8ARV

I agree with David - weight is not an issue.

I think if Hugen wants to sell these in a reasonable quantity, I think the
screen will need to be at least as big as the NanoVNA-F, as I think many
people will buy the NanoVNA-F for a better screen.


Dave.
--
Dr. David Kirkby,
Kirkby Microwave Ltd,
drkirkby@...

Telephone 01621-680100./ +44 1621 680100

Registered in England & Wales, company number 08914892.
Registered office:
Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United
Kingdom

Has anyone considered using TNC connectors? Not as heavy as N Types and, according to Amphenol, good from DC to 10 Gig.

Paul

I've made a new PCB for testing, it's working well now, but At the moment ADC1 can't read VBAT data correctly, and I can't get into DFU mode through the menu yet, I haven't found the cause of the problem yet, maybe to fix it after Christmas. Most of the circuits are the same as the current NanoVNA-H, so I don't think this is the next generation of NanoVNA, it's just the NanoVNA that replaced the MCU and has a larger screen, but that's probably the version that many people in this group need. If there are suggestions for changes that can be directly replied to this message, I will improve it.

The attachment is a screenshot.

hugen

hugen,
The larger display size is great and if the > 50 dB dynamic range to 1500 MHz holds up in the final release then I'd say you have a winner. Does the larger flash_size allow you to increase the measurement points to at least 201? Will firmware developed for the NanoVNA-H be compatible with the F303 based product? Any idea what the target price will be?

- Herb

At the moment ADC1 can't read VBAT data correctly, and I can't get into DFU
mode through the menu yet, I haven't found the cause of the problem yet

If MCU is changed from original nanoVNA, then I guess
also need change to NANOVNA_STM32_F072/board.h
for BOOT_FROM_SYTEM_MEMORY_MAGIC_ADDRESS

ADC1 address is buried in ChibiOS/os/common/ext/CMSIS/ST/*

The firmware for NanoVNA-H needs to be recompiled before it can be used. You need to modify the MCU and LCD drivers and some UI changes. You can view the code of AA6KL. AA6KL has completed most of the work. Measurements up to 1.5G are difficult to guarantee 50dB dynamics. The current test results seem to be able to obtain relatively accurate measurements in a dynamic range of 40dB.

hugen

I have made these changes, but I always get an error interrupt. It may be caused by some register values not being reset properly.

hugen

Hi Hugen,

I recently purchased my from ALI Express.

I am totally new to the VNA.

How can one identify if it is a clone or not.

Regards

Oscar