I have a situation where lane2 rams the extruder gears, buffer expands, extruder gears squeal, buffer compresses to zero. None of the other lanes squeal when ramming the buffer. My config sets lane2 & lane3 dist_hub: 110.0 for both and their PTFE lines are the same length. I have afc_unload_bowden_length: 1340.0 and afc_bowden_length: 1305.0. I have been getting a bunch of print PAUSE where I need to load lane2 manually to resume the print. In observations, I see the matte PLA filament was grinded down locally where I assume the drive motor was squealing continuing to ram the buffer when the filament had already reached the toolhead extruder gears. So I loosened the tension thumbscrew 3/4 of a turn and there is enough grip on the filament to feed/pull but the squealing still happens only for lane 2. Should I reduce my afc_bowden_length by ~10mm so that it doesn't ram as hard - and if I do will the other lanes still reach the toolhead extruder? The other lanes stop immediately once the buffer expands and trips its switch. Like, is there some logic to keep feeding in increments of 10 mm up to ~50 mm or so until the buffer is expanded? I haven't caught it live yet so I'm not sure if the reason why the lane2 filament is unloaded from the espooler was because the job was paused and that is some default behavior.
#buffer ram variation
lapis temple
I feel like the esteps of the motor on lane2 need to be tuned. Right now I have microsteps:16 and rotation_distance: 4.65 for all the lanes.
lapis temple
Yeah, it has the logic to keep jogging. I will reduce 1305 to 1290 for some trial and error. https://www.armoredturtle.xyz/docs/afc-klipper-add-on/installation/buffer-ram-sensor.html#required-configuration
lapis temple
Nah, maybe it wasn't e-steps. When I run AFC_calibrate_lanes both lane1 & lane2 return an unload length of 1325mm and a load length of 1305mm
shrewd heron
So currently AFC will always move the filament the full distances regardless if the switch is triggered or not, we have been working on filament homing which will stop once a switch has been triggered.
But for now you are probably better off setting you bowden length to the smaller number as AFC will do 20 more short moves until your sensor is triggered.
lapis temple
The smaller number doesn't seem to fix the problem. The fixed length (10mm) short moves stop shy of the fully extended length of the buffer so now pretty much all lanes get some squealing as the ramming happens more often. I feel like the ramming buffer ought to stop when it leaves the compressed state, not when it triggers the extended state and then backs off one short move. Then I know I'm at least one short move distance inside the buffer zone instead of one squealed grinding of filament. Can I change which trigger toggles buffer ramming completion?
shrewd heron
Currently there is not a way to change which sensor it uses
you can always try the homing branch, its up as a side branch right now...
lapis temple
Hardware comes more easily to me. I think I will modify the buffer CAD so that it has 10 mm longer length for the filament to travel after the advanced trigger is toggled to be in better agreement with the short_moves: 10mm distance.