Wrong extrusion spacing for top layers when using "combine infill"

[simonkuehling]

Hi,

i observed a slicing bug with latest v1.1.5-stable when using the "combine infill" option, where the top solid layers seem to have wrong extrusion spacing - as shown in the pictures. Config, STL and GCODE attached below.

Extrusion widths from the gcode:

; perimeters extrusion width = 0.50mm
; infill extrusion width = 1.98mm
; solid infill extrusion width = 0.85mm
; top infill extrusion width = 0.85mm
; support material extrusion width = 0.50mm

A correct infill layer from the middle of the object:

One of the top solid layers - extrusion spacing too wide:

And between the infill and top solid layers there are two of these sparse layers that also seem to be wrong:

CONFIG
STL
GCODE

[simonkuehling]

I just confirmed that those odd top layers also occur when slicing without any top solid layers and 100% rectilinear infill - as shown here:

[simonkuehling]

ok, another update: seems like slic3r is just calculating an unreasonably wide extrusion due to thin layers (0.1mm) with a big nozzle (0.5mm).

I just noticed that the gcode sniplet in my initial report above states:

; infill extrusion width = 1.98mm

...which is nearly 4x the nozzle diameter - a little too much in my opinion ;-)
Maybe we should introduce some kind of maximum ratio for nozzle diameter <--> extrusion width?

[alranel]

Well, I'd say there's nothing wrong... first off, don't trust 3D previews when they try to render the volume of extrusions :)

The top layers you see are wrongly rendered in the preview... you can see empty gaps between perimeters too.

Regarding width, it is capped to a sane max value for every part of the print except for internal sparse infill where the maximum is used. Note that maximum means full native flow for the nozzle.

[simonkuehling]

Yes, i totally agree that the renderings may easily be wrong...

I just tried to investigate what i have seen on the real print before - the layers i pointed at did in fact extrude with large gaps between extrusion lines which resulted in a faulty upper surface.

Anyway - is the extrusion width of 1.98mm in the intended range for a 0.5mm nozzle with 0.1mm layers?

With a common cone-shaped nozzle tip i see no chance to spread the molten plastic wide enough to get a closed layer finish.
I measured the extrusion width on these spots to be roughly 1mm.
We would need some kind of spatula-shaped nozzle tip - or a bigger flat face around the nozzle bore to distribute the plastic in such wide lines...

[lordofhyphens]

Could you check this on https://github.com/alexrj/Slic3r/files/336282/slic3r-most_prs.zip and provide screenshots from slic3r's own preview?

The extrusion width calculation probably needs tuning, as I agree that the max width possible is probably around 80-90% of the diameter of the flat of the nozzle (which is different with different nozzles). Probably isn't hard to add a limit and option.

[simonkuehling]

@lordofhyphens - unfortunately i do not have a windows machine on hand, only running linux here... do you have a commit-reference that i could check out and compile?

[lordofhyphens]

Branch is fix-combine-infill, see my fork (lordofhyphens/slic3r) or the open pull request. On my phone or I would give the exact link.