Magnet holes

[greg19]

This closes #53. This closes #92.

I will need to test this in practice and probably adjust the numbers, but the main functionality is there.

Magnet relief

I think the contact points of hex shape could be broken with it (I haven't tested it in practice), but if we plan to remove the Hex in the future I'd just ignore this (or disallow Hex+relief).

Crush ribs

I have been experimenting a little, and came up with this design. The number of ribs and the waviness can be adjusted.

The Gridfinity Rebuilt wraps a sine wave around a circle, but if I understand correctly it really creates a 360 sided polygon that approximates it. I wanted to create something built from smaller amount of primitives, so these are circle arcs that connect smoothly.

Baseplate magnet chamfers for all hole shapes

I saw that round shape had a chamfer, but the hex didn't so I added the chamfer to hex and crush ribs. Had to modify the logic a little to make it look nice and leave enough contact surface, so the angle isn't perfectly 45° (it depends on the hole shape).

Do you think these chamfers would be good also for bins?

Misc

• Changed utils.curve_to_wire to utils.curve_to_face, as it was always used for creating faces.
• Changed Part.makeLoft([Part.makeCircle(...), Part.makeCircle(...)]) in few places to Part.makeCone(...)

[greg19]

I did some testing on crush ribs, and numbers that worked best for me were: $diameter=5.7$, $\beta=0.5 \cdot \pi / 2$, $n=12$. It isn't that easy to insert first magnet, but it allows ~6 reinsertions without losing the grip.

From my understanding, there is a tradeoff between how easy it is to initially insert the magnet and the ability to remove it and reinsert. As the grip comes from the material deforming, it loses some of its ability with each insertion. I preferred $n=12$ over $n=8$ because my understanding is that the more ribs, the less deformation is needed in each of them to hold the magnet with the same total force, increasing the ability to reinsert magnets.

I tested only for two parameters: force needed for first insertion, and number of reinserts after which you can pull out the magnet with a force of another magnet. An interesting problem is how the parameters affect long term grip force, but I have no idea how this could be assessed.

Maybe if someone doesn't plan to remove the magnets at any point, the parameters can be tweaked (I guess by increasing the waviness and maybe slightly diameter) to make the initial insertion easier while retaining a strong initial grip.

[Stu142]

Thanks for the results. This seems like a good place to start.

[greg19]

For a status update on this PR, I want to test the removable magnets for bins and crush ribs in bins (so far I tested only baseplates). Maybe tweak the numbers a little (depending on testing), but the main code structure shouldn't change much now.

[greg19]

@Stu142 I finally was able to test the magnet remove channels and they seem to work OK. Sorry for stalling this for so long, but this PR now looks complete to me.