SKR 1.4 weird connection problem to PC (Repetier)

Hi,

I've changed the controller of my CNC from Ramps to SKR 1.4. I had several problems while compiling but I luckily I could solve them on my own. Unfortunately now I'm stuck. I use Repetier to controll the CNC from my PC and connected the USB cable as usual. When hitting connect, Repetier says connected but the controlls aren't really working. When I hit the X+10mm button, first nothing happens, but after a very long time the CNC suddenly does the move it's supposed to. This waiting time can be anything from 10s to what seems like forever. That's the same behavior for entering my own gcodes like M122 for checking if UART is working.
Repetier says "No start signal detected - forcing start" and every now and then it says "Communication timeout - reset send buffer block".
When using the LCD menu, everything is working perfectly. So I guess the problem has to come from the USB communication.

When I first started the board, the LCD said "Err: EEPROM version". I solved that by hitting Restore Defaults, then Store Settings and then Load Settings. Now the Error is permanently gone, but could this be a hint on what's wrong?

I already tested connecting to my Mac using Repetier: Same thing...

I hope anyone has an idea for what settings are wrong in my config or what other problem it could be. Attatched is a .zip containing my config.

Thanks Max

I played a bit with the settings and got the problem to be consistent. Just had to reduce the baudrate from 250000 to 115200. Now I can send exactly one command which gets executed immediatly but nothing after that. For sending a second command I have to reconnect Repetier Host to the Board, but no reboot of the SKR 1.4 required.

Is it normal that the 32bit Boards don't restart on connect? My Ramps/Mega2560 always did that...

Quote
mf2909
Is it normal that the 32bit Boards don't restart on connect? My Ramps/Mega2560 always did that...

I've observed the same issue.. did you find a solution for that (to make it allow reconnection?)

This may be related, but time stamps don't track... can't hurt to try

Back in April Marlin changed the start keyword to echo:start.

Repetier is to dumb to be able to see the start in this string.

After a long discussion between developers it was agreed that changing to echo:start was in error.
This was reverted in todays bugfix.

To apply this fix it to your version simply open Marlin/src/MarlinCore.cpp
find SERIAL_ECHO_MSG("start");
and replace it with SERIAL_ECHOLNPGM("start");

Edited 1 time(s). Last edit at 12/03/2020 06:55AM by Dust.