Remora - Rpi Software Stepping Using External Microcontroller via SPI

I tested the code one week ago on a F407VET, F407VGT, F407ZGT with SDIO SDcard, flashed with the SKR2 bootloader.
It works very good out of the box, and i can't wait for the final release.
After examining the code it would be easy to port it for F1xx, and maybe F7xx and H7xx (400 Mhz)
i'm not familliar with Mbed, I'm using only STM32CubeIDE, so i can't help that much, but i'm working on my own version with an different linuxcnc driver (not the remora.c, i started from rpspi.c) for now without success (the goal is to increase the step generation)
 

cakeslob wrote:

Im powered on dc in with 12v, and the serial is in the same place it was before on the tft spot. before I tried to load hello world, marlin was very chatty on the serial , and now nothing so I am wondering what is wrong, like if I put it into the wrong bootmode while loading from the sd card or something . Im only trying to get hello world working on the skr2 as a baseline first so I can modify it for a different target stm32 chip/board to test out. But compiling and loading it to the sd card doesnt seem like enough

Hi, I note that you used the online compiler. I fear that you may not be able to setup the online compiler to allow for the bootloader and custom vector table. If Marlin does not flash as per normal then the bootloader may be corrupted / written over by a incorrectly configured bin file. If this is the case then you'll need to get an ST-Link to re-flash the stock bootloader.

I'd recommend to only use Mbed Studio or CLi to ensure that the target specific configuration is used.

gtt38 wrote:

I tested the code one week ago on a F407VET, F407VGT, F407ZGT with SDIO SDcard, flashed with the SKR2 bootloader.
It works very good out of the box, and i can't wait for the final release.
After examining the code it would be easy to port it for F1xx, and maybe F7xx and H7xx (400 Mhz)
i'm not familliar with Mbed, I'm using only STM32CubeIDE, so i can't help that much, but i'm working on my own version with an different linuxcnc driver (not the remora.c, i started from rpspi.c) for now without success (the goal is to increase the step generation)

 

Step generation is all on the MCU side, LinuxCNC only needs the Servo thread running.

scotta wrote:

 
 
Step generation is all on the MCU side, LinuxCNC only needs the Servo thread running.

 

I already know that,for me a FPGA is the best solution for that reason but the pins are fragile from my experience. I had to use some additionnal component (bus switches) to protect the FPGA. I had very good result on a XC7A35T with XADC support, but the pb is the cost. The STM32H743VIT @480Mhz could be the perfect solution.

gtt38 wrote:

....... the pins are fragile from my experience. I had to use some additionnal component (bus switches) to protect the FPGA.

I use several Mesa 7i92 on plasma cutters, the worst environment for electronics, never had one fail. And i do not use anything else on IO, directly to drives, relays with their own transistors, relays on inputs. They just need a bit of care.
Not helping, am i ?

Any FPGA or ARM SOC requires protection from ESD/EMI, noise and spikes on its IO's when used in an industrial environment.

STM32H743VIT data sheet
www.st.com/resource/en/datasheet/stm32h743vi.pdf

Budgetary pricing @10K pcs is ~$7ea USD IF you can actually find them in stock. This price isn't very far off of an older Xilinx FPGA used in the 7i92. A proper IO board will still have the same IO protection and connectors. But maybe you don't want or need all that protection that the 7i92 and similar offers. I'm waiting for the chip prices and availability to settle down before I commit to a board design using the STM32's. They do offer more flexibility and features than the LPC17xx. www.st.com/en/motor-drivers/powerstep01.html is another option using STM.