Multiple motion system

[gloomyandy]

@Alva I think the first step will be to wait for 3.6.0-beta.2 (due shortly I think) and test again with that. If it still fails then try and create a simple test case that recreates the problem and then we will need to try and reproduce the problem.

[Alva]

@gloomyandy Okay, I will wait for the release then. Thank you

[droftarts]

@Alva It is available now: https://github.com/Duet3D/RepRapFirmware/releases/tag/3.6.0-beta.2

Ian

[Alva]

@droftarts

• daemon(2).g
• mini(1).gcode
  Hello , i have updated the machine to 3.6.0.beta2 and tried a simple logic as seen in the mini-1.gcode which is attached. while it starts the print i manually set the flag to enable the daemon.g and the daemon.g is also attached. And the same error occurred again. Do you think there is something wrong with the way that i understood the second motion system? or am i doing something wrong? weird that my logic worked in 3.5. fw version.
  Please let me know. Thank you.

[Alva]

@Alva Looks like there was a mistake in the 3.6.0.beta2 release and David updated the new one today. I will try out that one.

[dc42]

@Alva here are a couple of clarifications:

• When you select a tool (which becomes the current tool for the current motion system), the axes that X and Y are mapped to for that tool become owned by the current motion system.
• It's fine to drive the two motion systems from different GCode streams.

[dc42]

@Alva I just noticed this in your mini(1).gcode file:

G4 P0 S0.5

You should use either P or S in G4, not both. Similarly in daemon.g. However, this doesn't explain the error.

[dc42]

@Alva my best guess is that the move G1 X0 Y0 U0 in mini(1).gcode hasn't completed when daemon.g starts commanding U axis motion. This shouldn't be possible, because if U hasn't finished moving then U should still be owned by MS0, so the attempt to move it in daemon.g should fail with the "axis already owned by another MS" message. I'll check the M400 code.

[Alva]

@dc42 I have edited the code and logic a bit. Found different cases:

; File Name: 		/sys/daemon.g
if(!exists(global.daemonStop))
	global daemonStop = false	; Set to 1 to stop the daemon.g, it can be use to upload a new daemon.g file

while( global.daemonStop == false )
    if(state.status== "processing")
        G91
        M596 P1
        G1 U5 F300      
        M596
    G4 P0
M99; Exit

test_2nd_MS.gcode
.
This happened when i activated the daemon before the move to X0 Y0. But if i activate the daemon after that then it works perfectly. Cannot relate whats going on.

set global.daemonStop = true ; stopped the daemon
M118 S{"starting to move"}
G90
G1 X0 Y0
M400 S1
G91
set global.daemonStop = false ; activated the daemon which only use U axis
G1 X100 Y100 F200
M400 S1
G90
M118 S{"move done"}
M99

[Alva]

@Alva

[Alva]

@Alva Looks like it is a timing problem or something.

[dc42]

@Alva thanks, I will try to replicate it.

[Alva]

@dc42 Thank you

[dc42]

@Alva I have reproduced this and I am looking into it.

[dc42]

@Alva I created https://github.com/Duet3D/RepRapFirmware/issues/1064 and I am testing a fix.

dc42 created this issue in Duet3D/RepRapFirmware

closed Use of multiple motion systems causes Code 3 halts in 3.6 beta1 and beta2 releases #1064

[dc42]

@Alva please try the new Duet 3 Mini firmware binary at https://www.dropbox.com/scl/fo/pasag1g18orahwvn49qp2/ABWgw6D-TyKWxxurTmGwiUE?rlkey=r9h3rjyw1l4wh7xbr8yho37gw&dl=0.

[Alva]

@dc42 It seems to be working. I have tried all the above conditions and works as intended. Can i expect the same fix for Duet3 6HC board as well, as i observed the same behavior in it as well. Thank you

[dc42]

@Alva I have just put a 6HC build at the same location.

[Alva]

@dc42 Thank you , I will check it out

[T3P3Tony]

@Alva please can you confirm if you have tested this.