_CLEANING_MOVE macro doesn't account for COPY or MIRROR mode

[devbranch-vadym]

What happened

I've faced an issue when the filament ran out on T1 during the COPY mode print.
I've inserted a new filament as usual, which triggered loading filament on T1.

The extrusion was successful; however, I then unexpectedly got move out-of-range errors, resulting in Klipper shutdown and failed print.

Looking at the filament load code, seems like the issue is with the _CLEANING_MOVE macro:

https://github.com/Rat-OS/RatOS-configurator/blob/development/configuration/macros/load_filament.cfg#L426

						{% set loading_position = printer["gcode_macro T%s" % toolhead].loading_position|float %}
						{% set parking_position = printer["gcode_macro T%s" % toolhead].parking_position|float %}
						{% set speed = printer["gcode_macro RatOS"].toolchange_travel_speed|float * 60 %}

						{% set cleaning_position = loading_position %}
						{% if loading_position == parking_position %}
							{% if loading_position > 0 %}
								{% set cleaning_position = loading_position - 30 %}
							{% else %}
								{% set cleaning_position = loading_position + 30 %}
							{% endif %}
						{% endif %}

						# nozzle cleaning 
						G1 X{parking_position} F{speed}
						G1 X{cleaning_position} F{speed}
						G1 X{parking_position} F{speed}
						G1 X{cleaning_position} F{speed}
						G1 X{parking_position} F{speed}

It seems like it doesn't account for COPY/MIRROR modes' coordinate shift.
Since the same operation worked fine on T0 previously and works fine with T1 in single mode, my guess is that the macro tried to move the T1 in absolute coordinates (which are only valid in single toolhead mode) instead of applying COPY/MIRROR mode offset/calculation (e.g., something like x -= bed_size /2).

What did you expect to happen

The filament should have loaded without error.

How to reproduce

Start a print in COPY mode, let filament run out on T1, then insert a new filament.

Additional information

No response

[devbranch-vadym]

Probably some logic could be borrowed here:

RatOS-configurator/configuration/macros/priming.cfg

Line 317 in bdffe02

{% if target_idex_mode == 'copy' or target_idex_mode == 'mirror' %}

[devbranch-vadym]

@HelgeKeck suggested the printer is not in copy mode while it's on pause, so my guess about the coordinates offset is incorrect.

https://discord.com/channels/582187371529764864/859890291591217162/1408104783604092948

Here are related parts of my config:

# RatOS.cfg
[gcode_macro T1]
...
variable_loading_position: 532.8                          # filament load x position
variable_parking_position: 557.8                          # parking x position

# printer.cfg
[dual_carriage]
position_max: 558.283
position_endstop: 558.283

[gcode_macro T1]
variable_parking_position: 556.283

Unfortunately, I don't have logs (logrotate already ate them), but I remember it was something about moving to 5xx on the X axis.

My loading_position is not equal to parking_position, so cleaning_position = loading_position, meaning the nozzle cleaning code should have moved the toolhead between 557.8 and 532.8. That's safe within 558.283 (position_max) limit.

I'll check if I can reproduce this again.

[devbranch-vadym]

ratos-variables.cfg:

[Variables]
idex_applied_offset = 0
idex_xcontrolpoint = 210.97707130405206
idex_xoffset = 0.29643320294232467
idex_ycontrolpoint = 529.5963760982671
idex_yoffset = -0.02374941536629649
idex_zcontrolpoint = -0.7168749999984574
idex_zoffset = -0.012500001482515621
idex_zoffsetcontrolpoint = -4.032500001461502
nozzle_expansion_applied_offset = 0.1977749255718564
nozzle_expansion_coefficient_multiplier = 1.9114459736472382
nozzle_expansion_coefficient_t0 = 0.09406249999336236
nozzle_expansion_coefficient_t1 = 0.09249999999490433
t0__spool_id = 180
t0_filament = ('unknown', 'unknown', 220)
t1__spool_id = 203
t1_filament = ('unknown', 'unknown', 220)

(Ignore the last 4 lines, that's for Spoolman)

[devbranch-vadym]

Copy and paste of my new guess from Discord to keep things up to date:

I didn't have a chance to reproduce the issue normally yet, but I think I know what might be happening.

I just did this quick experiment but forgot to switch to T1:

G28
_CLEANING_MOVE TOOLHEAD=1

I got the same move out of range error as when the issue happened mid-print.

So I checked the load_filament.cfg and didn't find any T1 or _SELECT_TOOL called if the printer is homed anywhere.

As you said previously, pausing disables the COPY mode temporarily, but I can't see the tool change anywhere. Now I remember T0 was active in Fluidd when the original issue happened to me.

My guess is that it probably tried to move T0 (which was active) into T1 cleaning position (which is obviously out of range).