KISSlicer 1.5 Beta 2 - in development

[lonesock]

Hi, all.

In the latest build, I suppress any loops immediately next to crowning paths so we don't get too much plastic in those regions. It looks as if, in that letter 'A', there are tiny crowning paths that create a void, but then are optimized away since they are too short. This gives me something to look at, thanks. Out of curiosity, do you need to set the threshold to 1 for it to work well, or does a tiny value (like 0.01) work as well?

thanks,
Jonathan

[inventabuild]

I need to set the threshold to 1 for it to work well (as in above post). Crowning = 0 and Crowing = 0.1 (see below) give me the same results. I am printing a box top with inlaid text against the bed to give both the top and the text a smooth finish. When there are gaps in the printed text it doesn't look as nice.

Oversample Res = 0.1, Crowning = 0 - Path.JPG (62.35 KiB) Viewed 6060 times

Oversample Res = 0.1, Crowning = 0.1 - Path.JPG (42.81 KiB) Viewed 6060 times

[lonesock]

Could you please post that model 'A's STL file, and let me know what extrusion width you are using?

thanks,
Jonathan

[inventabuild]

Hi Jonathan,

Attached are the two stls of the box top. Please let me know if you need anything else.

Thank you.

Mitch

SNAPS Box Top Text - Small Version.zip

(30.52 KiB) Downloaded 287 times

SNAPS Box Top - Small Version.zip

(34.42 KiB) Downloaded 260 times

[inventabuild]

...and Extrusion Width = 0.4mm

[lonesock]

OK, there is a beta 2.22 up on dropbox.

Here are some things to consider regarding crowning:

• Try setting the crowning threshold ~= 2 * extrusion width
• The crown path width is auto-adaptive, from 0.5 * extrusion width to 1.5 * extrusion width
• If you get disjointed crown paths it is possible they are right near the 0.5 or 1.5 extrusion width...try adjusting the extrusion width a tiny bit (like +/- 0.02 mm)

Please let me know how this goes!

thanks,
Jonathan

P.S. On the box with text in it, you don't need to actually cut the letters into the box, you could just set the extruder priority in the Mesh-to-Extruder map window. That way you give the text extruder a higher priority and it will automatically override the box material in the overlapping regions. It will hopefully make the modelling easier. [8^)

[inventabuild]

Jonathan,

Thank you for the tips. Vetting all the letters in the alphabet (to anticipate any words I might need to print) I found that V1.5 Beta 2.20 does a much better job of filling in the letters with Crowning than V1.5 Beta 2.22.

Attached are the stl's of the top (will try not cutting the text next time) and below are snapshots of KISSlicer's paths (Beta 2.20 is 1st pic, Beta 2.22 is 2nd pic). Both slices are done w/ Extrusion Width = 0.4, Resolution 0.1 and Crowing 0.8. Note letters B, H, I, Z, really almost all of them have better fill-in-the-letters paths in Beta 2.20 vs 2.22.

Alphabet Box Top Text - Tahoma Cap Text - Small Version.zip

(72.49 KiB) Downloaded 273 times

Alphabet Box Top - Tahoma Cap Text - Small Version.zip

(85.03 KiB) Downloaded 256 times

[radus]

Hi!

I'm try slice object with 1 mm size. see kisstest.stl
with 2.22 version no crowning.
But old 2.20 was good.

OK, there is a beta 2.22 up on dropbox.
Please let me know how this goes!

[Dreide]

I'm try slice object with 1 mm size. see kisstest.stl
with 2.22 version no crowning.
But old 2.20 was good.

As lonesock said: "The crown path width is auto-adaptive, from 0.5 * extrusion width to 1.5 * extrusion width".
With 0.4 extrusion width (at 0.2 layer height), the width covered by the perimeter extrusion is 0.84mm already, so 0.16mm is missing to fill 1mm (note that the shape of the outer side of the perimeter is assumed to be round, not rectangular, that's where the 0.04mm is coming from). Since 0.16 is smaller than half the extrusion width, no crowning is done.
IMHO it is reasonable to have crowning limits, because too small or too wide crowning paths are not extruded cleanly anyway. In general, whatever changes are made to KS, you can always find models for which these changes give worse results, but the goal should be to improve things for the majority of models.
The ultimate solution for thin wall and/or crowning problems would be to allow the extrusion width of regular paths to be locally adaptive, but this is obviously very complicated to implement and would also be computationally demanding.

[radus]

I'm try slice object with 1 mm size. see kisstest.stl
with 2.22 version no crowning.
But old 2.20 was good.

As lonesock said: "The crown path width is auto-adaptive, from 0.5 * extrusion width to 1.5 * extrusion width".
With 0.4 extrusion width (at 0.2 layer height), the width covered by the perimeter extrusion is 0.84mm already, so 0.16mm is missing to fill 1mm (note that the shape of the outer side of the perimeter is assumed to be round, not rectangular, that's where the 0.04mm is coming from). Since 0.16 is smaller than half the extrusion width, no crowning is done.

Just need change formula to auto-adaptive, from 0.25 * extrusion width to 1.5 * extrusion width
And crowning lets be work better!