big issue in communication with GRBL at least with UART

[robomechs]

big issue in communication with GRBL at least with UART

[robomechs]

And what's about usb cdc?

[mstrens]

I have not yet test with USB.
Anyway, with USB I can't trace the message with my digital scope (it is just a cheep chinese item).
Furthermore, to communicate between my ESP32 stand alone micro computer and GRBL I can't use USB. UART is by far the easiest way to let ESP32 communicates with STM32

[robomechs]

Ok.

[misan]

@mstrens There is a native port of GRBL for ESP32 (not for 6 axes though) https://github.com/bdring/Grbl_Esp32

[mstrens]

I know the ESP32 project but I wanted a system that can run is stand alone with a display, a touch screen and a sd card. ESP32 does not have enough pins to manage all this. My ESP32 project seems OK (just some more tests to finalise) but the issues are now with GRBL on STM32. With some other people we already made some PCB for the Blue pill and for the ESP32.
I hope we will find the solution for this issue with GRBL on STM32

[mstrens]

I made some more analysis.
I changed the code in order to set a pin HIGH each time the program enter the interrupt that generates step and dir signals. The pin is then set LOW when it exits the interrupt.
When I run the test Gcode, I saw that I got pulse about 5 usec HIGH and 5 usec LOW.
It means that the signal was at about 100khz.
I expect that this does not leave lot of time to the CPU to manage other tasks (there is less than 5 usec between interrupt).
I looked further the set up and I was using the default values (5200 steps/mm).
This high value combined with a high speed (2000 mm/min), explains that GRBL had to generate signals at such a high frequency and so lose some incomming characters.
I reduced the step/mm to 120 (realistic for my machine) and now it seems OK.
I will further investigate to be sure it is really OK.