I wanted to share my progress on attempting to design a puzzle called a Bump Megaminx or "Bumpminx" as some people say. This idea has popped up here and there but I believe I have finally designed a workable version. Below are two photos:
File comment: Bumpminx-view 1
bump megaminx cad.jpg [ 249.68 KiB | Viewed 983 times ]
File comment: Bumpminx-view 2
bump megaminx cad 2.jpg [ 243.65 KiB | Viewed 983 times ]
First of all I need to demonstrate some of the problems with designing such a puzzle and how I solved each. The first problem is that if any bumping effect is applied on a normal Megaminx, extra Pyraminx-Crystal like pieces appear as cuts intersect each other:
File comment: Thanks to Hidetoshi for this image.
bump_system.gif [ 10.36 KiB | Viewed 983 times ]
As you may be able to tell from the above photos the way I got around this issue was to use very specific conical cutting surfaces, hence the curvy appearance of the cuts. I'm not going to go into specifics of the geometry of the cutting surfaces, but basically what results is that the cuts never intersect each other, and therefore these edges don't appear.
The second problem is the mechanism. On a Megaminx the mechanism is pretty close to the surface and if bumped it goes "out of bounds" or put simply isn't completely inside the puzzle. To solve this I based the mechanism around the shallowest cut depth, which gave it enough room to fit inside. The internals are smaller than normal, but it's actually a very stable core. Here's a view of the inside:
File comment: An internal view of the mechanism and labels.
bump megaminx 3.jpg [ 239.81 KiB | Viewed 983 times ]
Notice that the mechanism fits neatly and very close to the surface on one side, but on the opposite it is buried deep underneath the surface, which is a result of the Bumping.
The third problem (and probably the biggest) is that due to the shallow cuts that come from a Megaminx, there normally isn't much room to bump the puzzle (only about a mm difference per axis) and so many of the pieces become so similar in dimensions that they are nearly indistinguishable during solving. In my design the curved cuts allowed me to start with a slightly deeper cut depth than what would normally be allowed on a Megaminx, but sadly this didn't help much. I was only able to bump the axis system 6mm in one direction, 4mm in another, and 2mm along the final direction. This results in the cut depths of each side differing by 2mm, which I don't think is enough to be visible during solving. What do you think?
In conclusion, the Bump Megaminx is possible, but pretty impractical and time-consuming to make (I shudder at the thought of having to individually tolerate and fillet 63 parts...
I want to hear your thoughts on this puzzle. Would you be interested in seeing a physical version? Do you think it would be possible to recognize such small differences between pieces? Would this be practical? I really want to see this puzzle in real life someday having come up with the concept independently over a year ago. Any comments are appreciated.