3D Honeycomb - switch direction at smallest bridge point, rather than every layer

[gringer]

This is an update to my 3D Honeycomb / Truncated Octahedron infill code to make the infill pattern a lot stronger, and more likely to hold at very low infill percentages. The only bridging happens at the square horizontal faces, and only once per direction change. I have also made pattern density tweaks to make the infill percent better match the extruded density.

While the pattern does exclude two hexagonal walls with each direction, the strength should be similar to a fully-filled truncated octahedron pattern because the diamond verticals are preserved, and much stronger than the current implementation.

[SoftFever]

This is an update to my 3D Honeycomb / Truncated Octahedron infill code to make the infill pattern a lot stronger, and more likely to hold at very low infill percentages. The only bridging happens at the square horizontal faces, and only once per direction change. I have also made pattern density tweaks to make the infill percent better match the extruded density.

While the pattern does exclude two hexagonal walls with each direction, the strength should be similar to a fully-filled truncated octahedron pattern because the diamond verticals are preserved, and much stronger than the current implementation.

Thank you @gringer
Do you mind to post some screenshots/recording/comparisons/etc...?
It will help us to understand the proposed pattern(and attract community member to try/test and provide feedback to you).

[gringer]

Sure, here are a couple of images that demonstrate the new changes:

Here are a couple of video recordings of this new approach:

https://twitter.com/gringene_bio/status/1386901717019594752 [transparent filament]

https://twitter.com/gringene_bio/status/879212511836360704 [purple filament, sound is buggy]

FWIW, this was fixed in Slic3r in 2018, but that was shortly after Slic3r was forked for PrusaSlicer.

Here's what the infill pattern looked like in Slic3r before this change:

[SoftFever]

@gringer
This is nice. Even though its geometry is that of a honeycomb, the way it prints and the benefits it provides are more similar to gyroid. Plus, it prints faster than gyroid!

This would be a great addition to the infill patterns. I'd prefer it to be added as a new infill pattern instead of replacing the current '3D honeycomb'; maybe call it 'Truncated Octahedron'?
Let me know what you think.

[Eldenroot]

Yes, please - do not change the current one but add a new infill pattern, I like it!

[igiannakas]

So if I get this right; this pattern doesnt overlap itself, same as gyroid, right? If so that's fantastic! A faster alternative is great :)!!

[gringer]

I'd prefer it to be added as a new infill pattern instead of replacing the current '3D honeycomb'; maybe call it 'Truncated Octahedron'?

This is a bug fix, not a feature request. It's the same pattern, just an update. We discussed the naming of this when the pattern was first introduced into Slic3r, and it was decided to call it "3D honeycomb" rather than "truncated octahedron (3D)". At that time there were no other 3D patterns:

slic3r/Slic3r#1646 (comment)

I agree that because there are now other 3D patterns, it would make more sense to call it Truncated Octahedron, but that would touch a few more files.

This fixes many bugs in the existing pattern. It should replace the existing 3D honeycomb pattern because the existing pattern is too weak. Both patterns have identical geometry (i.e. they're both truncated octahedrons), with direction switching for half the pattern time. The only difference is that this one switches direction less frequently (i.e. at the smallest bridge point rather than every layer).

For the existing pattern to work at all, it needed to have special bridge extrusion settings, which means that its speed is different from other infill, and its line width is harder to change.

There's no benefit in keeping the existing pattern.

See a mechanical comparison of the patterns here by @unicornswag:

https://forum.prusa3d.com/forum/prusaslicer/prusaslicers-3d-honeycomb-infill-is-sub-optimal/

[gringer]

So if I get this right; this pattern doesnt overlap itself, same as gyroid, right? If so that's fantastic! A faster alternative is great :)!!

Yes, the truncated octahedron pattern in the 3D Honeycomb infill is quite similar to gyroid. It's basically a straight version of it, i.e. what gyroid would look like if the sine waves were replaced with lines having 45º direction changes.

Unlike gyroid, it has no direction changes in mid air, and therefore is likely to have fewer stringing issues for very low infill percentages (e.g. 5%). Due to being straight lines rather than sine waves it also has a smaller GCode footprint, which can help a bit with printer reliability when printing from an attached computer.

[SoftFever]

I'd prefer it to be added as a new infill pattern instead of replacing the current '3D honeycomb'; maybe call it 'Truncated Octahedron'?

This is a bug fix, not a feature request. It's the same pattern, just an update. We discussed the naming of this when the pattern was first introduced into Slic3r, and it was decided to call it "3D honeycomb" rather than "truncated octahedron (3D)". At that time there were no other 3D patterns:

slic3r/Slic3r#1646 (comment)

I agree that because there are now other 3D patterns, it would make more sense to call it Truncated Octahedron, but that would touch a few more files.

This fixes many bugs in the existing pattern. It should replace the existing 3D honeycomb pattern because the existing pattern is too weak. Both patterns have identical geometry (i.e. they're both truncated octahedrons), with direction switching for half the pattern time. The only difference is that this one switches direction less frequently (i.e. at the smallest bridge point rather than every layer).

For the existing pattern to work at all, it needed to have special bridge extrusion settings, which means that its speed is different from other infill, and its line width is harder to change.

There's no benefit in keeping the existing pattern.

See a mechanical comparison of the patterns here by @unicornswag:

https://forum.prusa3d.com/forum/prusaslicer/prusaslicers-3d-honeycomb-infill-is-sub-optimal/

Wow, 10 years ago? You rock!
I'm not too concerned about the names, we don't need to restart the naming discussion.

Personally, I haven’t used the current Honeycomb 3D infill due to the bridging issues. I’m confident your new 3D honeycomb will become a favorite among many users.

Regarding the replacement, unless we receive feedback suggesting a valid use case for the old version, I’m all for replacing it with the new Honeycomb 3D as well.

[SoftFever]

Looks good

[SoftFever]

Merged.
Thanks @gringer

[SoftFever]

Quick testing results show very high quality, even at high printing speeds.

Filament: PETG
Infills printing speed: at 126 mm/s

[SoftFever]

And of course I have to try 200 mm/s infill speed as well :)

[igiannakas]

Nice one! New default infill this 🎉