Travel between very close but non-connected regions get combed

[richfelker]

I first noticed this a year or two ago printing screw threads oriented in the XY plane: the top one or two layers would get severe stringing between them, and seemed to be missing retractions. Setting a low retraction_combing_max_distance fixed it, so the problem was clearly that combing was being applied despite the travel being between disconnected regions.

I've finally tracked the bug down to:

https://github.com/Ultimaker/CuraEngine/blob/58cae8395a97879f36e960d8f551089c2ec394a5/src/LayerPlan.cpp#L384-L389

where Comb::calc simply returns true (do combing) if the distance is less than max_distance_ignored.

I've since found that the same problem occurs in some articulated print-in-place models where it makes it harder to break them free initially and weakens small linkages by not getting the full amount of material bonded in the intended places.

As a workaround, the machine_nozzle_tip_outer_diameter setting can just be set to 0, but I believe this should really be fixed. Retracted travel without any stringing over distances shorter than the nozzle outer diameter is possible, and failure to do so is very surprising and compromised the integrity of prints.

[Ghostkeeper]

A similar issue was fixed here, a few days after you reported this: https://github.com/Ultimaker/CuraEngine/pull/1619

Does this issue still occur in master? How would we reproduce it?

[richfelker]

Reading those changes, it looks like they may mitigate the specific case I hit, but I think an issue is still present. Comb::calc unconditionally says "yes, combing is possible" when the distance is less than max_distance_ignored, and this is just wrong. As far as I can tell the commits don't change that because the logic for max_distance_ignored short-circuits the whole rest of the function's logic for when combing is allowed.

[richfelker]

At the moment I don't seem to have a test case to trigger the problem, but it may also be a matter of other local changes I've made. I'll try to follow up with an unmodified tree later and see if I can find cases still affected.

[rburema]

I'll try to follow up with an unmodified tree later and see if I can find cases still affected.

Sorry for the wait. Did you find anything?