07/05/19

The idea with the initial layout is to create something that works, not something that's the exact size and shape of a regulation 'Cube. Form can be considered after it functions correctly.

Starting with the central brass widget, the preliminary core assembly is based on a variety of open-source images and patent information.

All of the parts (except toolbox items like bolts and bearings) had to be designed from scratch, which is adding a considerable amount of research time.

The problem I'm running into is finding either; good images of the pieces from standard angles or a method to determine the geometry of an adjacent piece's internal structure indirectly using pieces that have already been modeled.

07/07/19

Geometry itself might actually be the key to their shape. The face pieces form a sphere internally, so it stands to reason that the actual internal structure of the pieces needs to interlock with, and move around, that sphere.

I take back what I said about YouTube not being useful here. Definitely some good information in these videos when it comes to making the design as accurate and close to real life as possible. https://www.youtube.com/watch?time_continue=9&v=DpUG5c_KDFc and https://www.youtube.com/watch?v=bgcScY7CiMs and https://www.youtube.com/watch?v=XVmBCfgQYCg <----Now with more unnecessary EDM.

The initial assumption that the central widget is made of brass was incorrect. They probably could be made of brass, but it's far more common for it to be made of nylon. The other pieces are made of injection molded ABS plastic. http://www.madehow.com/Volume-7/Rubik-s-Cube.html

07/08/19

The layout for the edge pieces is taking shape and it's becoming apparent that there's good news and bad news.

The Good News: The design for the edge and corner pieces isn't going to be as complicated as I initially thought. As suspected, the basic idea behind the internal structure of the non-face pieces is merely to interlock with adjacent pieces and when necessary move around inside the sphere formed by the face pieces.

The Bad News: Mating this so it works looks like it's going to be challenging. More research into advanced mates is going to be necessary.

07/09/19

School of Hard Knocks MOTD: Come up with a consistent fillet plan before you start modeling next time. It's an annoying waste of time to have to fix two or three features (that can easily cascade into an avalanche of many more broken features) in a model because of a hasty fillet choice early in the design process. Also, consider which features might need to change at some point due to interfering geometry (and those that definitely will not) when adding fillets. Critical features should probably stay unfilleted until near the end of the assembly process, unless doing so would compromise other features/parts.

Modeling of the Edge Pieces is complete and they have been added to the assembly.

I'm still having issues with figuring out how to mate everything in such a way that it functions as a Rubik's Cube would in the real world. Research into how the path mate function works was looking like it might be useful here, but further investigation is making that seem less likely. While creating a circular (or any closed shape) path for pieces on a face to travel around would certainly be possible with a path mate, it doesn't solve the problem of getting it to work on non-coincident perpendicular planes across an effectively infinite (4.3x10^19) set of possible face combinations.

So, in a nutshell the question really becomes, how do you create a set of mates that functions in an interlocking way? Is there a way to program mates with a set of if/else, or equivalent, statements?

07/10/19

After two days of searching it doesn't look like what I want to do with the mates is actually possible within the framework of SolidWorks. Full-real-time-collision-detection in something this relatively complex would probably require cloud computing levels of processing power, so it's definitely not going to happen on my desktop. You could sort of do it with the SolidWorks Mates Controller, but the sheer number of possible face combinations, and the fact that you would need to input the motion of every cube across every one of those (4.3x10^19) face combinations puts it far outside the scope of what I wanted to accomplish here. It's one thing to go a little overboard on getting the details just right, it's quite another to start with the intention of designing a Rubik's Cube and accidentally end up with plans for the Death Star.

Revised Objective: Model what would be a fully functional Rubik's Cube if it were to be produced. Ensure that none of the pieces interfere with each when a face is rotated as it would be with a real 'Cube. Test all of that by fixing the faces and rotating them.

07/14/19

Animation note: The model can be centered for the animation by deleting the Orientation and Camera View key in the motion study, then adjusting the model to the correct screen position, and then replacing the Orientation and Camera View key with the new position.

07/17/19

School of Hard Knocks MOTD: Edges in a model are not edges in a rendering. Edges in a model are not edges in a rendering. Edges in a model are not edges in a rendering.

Models of Corner Pieces are complete and have been added to the assembly. This completes the Rubik's Cube, so the next step will be to create a motion study to ensure proper fit and clearance for the 'Cube's internal structure.