BTT Octopus Pro F446 build

[yoft]

Hi,
I've been attempting to a build for the BTT Octopus Pro F446
https://github.com/bigtreetech/BIGTREETECH-OCTOPUS-Pro

It has a 32kb bootloader and 12MHz crystal.

I've written the board map without any issues. However, I'm having trouble with the build configuration.
I've never worked on linker scripts before so I was wondering if you had any guidance with this?

Thanks!

[terjeio]

Using Eclipse to build?
Perhaps this lnker script can be used, it assumes a 64K bootloader and no EEPROM emulation in flash. It has a 32Kbit EEPROM that should be used for settings storage.

[yoft]

Hi,
I've tried a few variations on the build, the latest is here. I'm still unable to communicate after flashing. I tried both 64 and 32k bootloader sizes and added #define EEPROM_ENABLE 32 to the board map.

Here is the latest attempt:
https://github.com/yoft/STM32F4xx/tree/new-octo

Thank you aain for your help!

[terjeio]

Communication via USB can be tricky to get right since the clock supplied to the USB peripheral has to be exactly 48 MHz. The clock setup differs between F446 variants making it even more tricky.

Try changing this line to:
#elif defined(BOARD_FYSETC_S6) || defined(BOARD_BTT_OCTOPUS_PRO)

Regardless of this you can measure the voltage on a stepper enable pin: removing a stepper driver from its socket and measure in the socket is IMO the easiest/safest. It should be 3.3V (or 5V if the signal is buffered) if the code was sucessfully started by the bootloader. Do this to determine the size of the bootloader.

[yoft]

Hi @terjeio,
I've tried your suggestions with no success. Both 32k and 64k bootloaders had 5v on the enable output.

I've found some settings used for this board from another source here

I've copied them into my repo

I'm seeing 5v on driver0 Enable output on the new linker script committed to my repo too. Is it possible this is a feature of the bootloader?

I've also tired setting
/#define USB_SERIAL_CDC 0
and using a usb/serial convertor @ 115200 baud and have had no response to that either. I tried installing the Klipper firmware however I've never used this before. When using the UART variant, it was periodically sent a weird character over serial every few seconds. When using the USB variant, it showed up in device manager as an Unknown Device, which is more than I had before.

I'm very grateful for your help in this

[terjeio]

I've found some settings used for this board from another source here

Ok, it seems the bootloader is 32K from that:

FLASH (rx) : ORIGIN = 0x8008000, LENGTH = 512K - 32K

Have you tried with BTTs clock setup?

And if you have a scope you can check the UART TX line, is there traffic there on boot and is the bit rate 115200 baud (~8.7 microsecond/bit)?

[yoft]

Hi @terjeio,
I've made some changes and grabbed a linker script from the BTT repo.
Here Are the changes.
And here is where I got the settings from.

The driver enable output appears to be 5v no matter which linkerscript I use. I've tried the first two driver enable outputs.

Compiler preprocessor symbols:
USE_HAL_DRIVER
BOARD_BTT_OCTOPUS_PRO
STM32F446xx
HSE_VALUE=12000000
_VOLUMES=1
_USE_IOCTL=1
_USE_WRITE=1

If you have any further ideas, I'd love to hear them.
Thank you again for your help.

[terjeio]

Do not enable the EEPROM until you have communication working. And disable flash emulation as well by leaving only line 2663 in this block of code.
If disabling USB CDC is the UART still dead? Do you have access to an oscilloscope for checking the TX line? Or are there LEDs on the USB <> UART breakout? If so is the TX LED briefly flashing on startup?

To 100% verify that the code has started you could try to continously toggle a GPIO pin early in the the startup sequence, after outputs has been initialized - before this line. This to determine if the clock tree is incorrectly setup or if the booloader/loader script combo is wrong.

[yoft]

I've tried disabling EEPROM and flash, however I receive a few errors and the build fails.
/../Src/flash.c:44: undefined reference to '_EEPROM_Emul_Start'
I'm confused by this as FLASH_ENABLE = 0 so flash.h should not be included.

[terjeio]

Try with this:
flash.zip

[yoft]

Hi @terjeio,
Good news!
I went back to basics and learnt how to write a simple application that blinks an output.
Whilst doing this, and using other outputs to indicate where errors had occurred I believe I found the issues!
The main issue is that it threw an error when attempting to configure PLL, I suspect that the bootloader configures the system clock to use PLL. So I've told it to use HSI, configure PLL , then use PLL again.
As soon as this was fixed, I was able to connect to the board through USB using IoSender! Huge relief!

Now just to sort out the EEPROM connection issues.

[ErdbeermarmeladebrotmitHonig]

I just updated the linker script, maybe there was one faulty region, I removed it.

[ashokmanker]

Still not detecting on PC. May be ld script is for 429VGTX, but processor in Octopus_pro is 429ZGTX. Thats also seems correct accroding to link shared by @yoft https://github.com/bigtreetech/BIGTREETECH-OCTOPUS-V1.0/blob/master/Firmware%2FMarlin-bugfix-2.0.9.3.x%2Fbuildroot%2Fshare%2FPlatformIO%2Fvariants%2FMARLIN_F429ZG%2Fvariant.cpp

[ashokmanker]

Any progress in this direction?

[ErdbeermarmeladebrotmitHonig]

Not from my side, as I am not working on it. I can't debug anything as I don't own any board with a STM32F429 chip. You mentioned that the linker script is for STM32F429VGTX but the BTT Octopus has a STM32429ZGTX, but in my opinion this should not make that much of a difference. However, you could try the linker script from Marlin, maybe it works with that one. Keep in mind, that you have to add the "_FLASH_VectorTable".
If it is not working with this linker script you could do what @yoft mentioned: try a blink script.

[ashokmanker]

Amyone else helping working 429 board