As others have said, it should be easy enough to create the second version in wood or metal. With identical sized diameter rods, you just cut away until you have the pieces the required shapes. You just need to check that it's structurally sound after all those cuts though. As for specifics, you could always just use paint or dyes, or go for different actual materials of metal or wood types. Metal tends to give weight and heft to a puzzle, which has it's own appeal, while wood can introduce a more organic and natural element to how a puzzle feels in the hand.
I like that the first move in this puzzle is a combination of pieces moving. I wonder if you could incorporate a twisting motion, like the French Fries puzzle.
I can bring you in contact with Evgeniy Gregoriev, who can build you one. The price will likely be in the EUR 100-200 range, or perhaps more. If that does not deter you, then please contact me at oskarvandeventer.nl/index_contact.htm
I'm just guessing here, but could the flaw be that the outer shell can also be solved into a perfect cube shape with a volume that also perfectly contains the "infill only" part?
Coextruding is so cool. CNC Kitchen has some interesting videos where he makes his own filaments in this way, including one that’s PLA on the outside and TPU on the inside. 2:49 As somebody who mods a lot of PC games, that looks kinda like a normal map.
There are some 1.3 million positions, some are "Garden of Eden" positions in that there is no way of getting to them but they could be used as starting points. (Have the 4 empty slots in pairs facing each other, each side of each gap have pieces that want to cross blue one way and some other colour the other).
This puzzle looks like it’s using straight-cut gears, which seems like it could be quite loud. Herringbone gears might help with keeping it quiet and smooth.
What pigments are you using to make it glow in the dark? I’m personally fond of strontium aluminate, which not only glows even more brightly (and for longer) than the more common zinc sulfide, but is easy to tweak to glow different colors. It can also be charged with heat!
I think it needs some magnets. A few powerful magnets in the right places might actually make it even harder to turn! (See z3cubing’s most magnetic cube.)
See e.g. studio.ru-vid.comPLwdtSpXcZNXnWGKjE7U1r5zIjbeH5J_-4/; oskarvandeventer.nl/; en.wikipedia.org/wiki/Oskar_van_Deventer; www.puzzlemaster.ca/browse/inventors/oskar/?p=all
Now, that's really nice. It also makes me realise that I hadn't appreciated how the original version worked. So, how much time do I allocate to investigating if it can get blocked? Arggghh - it's too late my brain is already working on it.
Could you please describe a bit more the what moves where? In programming, when you do the multi-threading and utilizing the locks, there is an effective approach on how to prevent deadlocks - to establish that the "smaller key gets locked first" for any transaction. It is a well established technique/knowledge, I'd like to try to apply same principles to prove or disprove your claim, but I really need a bit more information about what exactly moves where and why. From the overview in this video - it looks like your puzzle can be reduced into the graph of states, and then trying to apply the simple rules on which states can be changed into which other states.
Each of the points where 2 wheels touch has a magnet embedded on either side of it, and each piece is also a magnet. Due to magnetic forces any non-blue piece will automatically jump from blue > yellow > green > red > blue etc when given the opportunity, while the blue pieces always jump in the opposite direction. So for example if the point where the yellow and green disks touch had a blue piece on the green side and an empty slot on the yellow side then that blue piece would jump over to the yellow side, but if that piece had been any other color it would not jump.
It is much simpler than you may think. If you have six blue magnetic tokens in the yellow disk, and six yellow tokens in the blue disk, then everything is stuck. This is because the blue tokens can only exit to the blue disk, and the yellow tokens can only exit to the yellow disk. So by reducing the number of blue tokens to five, we assure that this state can never happen, and hence the puzzle can never get blocked. (Reminder: the red-yellow-green tokens have one magnetic polarity, and the blue ones opposite polarity.)
If all discs move clockwise, blocking is impossible because holes can always receive discs from the spinner counterclockwise of them. This is true regardless of the number of holes, provided there is at least one disc and at least one hole. If only one disc moves counterclockwise, we can have a configuration where a hole can turn to the left and not receive a disc, and can turn to the right and not receive a disc. However, we're not stuck because with only one disc moving against the current, we know we can rotate one of the neighboring spinners to find a compatible disc for the hole. This argument actually works as long as the number of counterclockwise discs is lower than the spinner capacity. IE, with five counterclockwise discs, the hole must be able to match up with a sixth disc that wants to move clockwise -- or else there's another hole, which itself can't be stuck. There are a bunch of cases to consider for a full proof, but this seems plenty convincing. Again, the number of holes doesn't seem to affect the argument. With at least six counterclockwise discs and six clockwise, getting stuck is certainly possible, because we can have two adjacent spinners full and no way to get holes onto them.
Yes, my first version (the laser cut model in the video) had two discs moving clockwise and two anti-clockwise, thus the 'black hole' problem Oskar describes. As two colors have three gates (green and blue in Oskar's final version), you also have a choice on which direction the center (green-blue) gate operates.
@@mikewoodconsulting Whoops I forgot about the center gate! Honestly it seems as if the center gate doesn't change the circumstances for when it gets stuck, six discs of each chirality still can get stuck. The center gate just cuts in half the number of stuck configurations I would guess.
I don't understand how the puzzle gets blocked from the flipped magnetism. Does it actually get mechanically stuck and become unable to move? Or does the puzzle end up in a state where it cannot be solved, like from making a move that cannot be reversed even with a lot of extra moves?
Here is a constructed example. Imagine that blue and yellow spinner contain discs that want to travel counterclockwise, and red and green, discs that want to travel clockwise. There is a potential traffic jam at the yellow/green interface. This becomes an actual traffic jam if all four holes are in the blue and red disc (an easy thing to accidentally do). The holes can't move because all discs now want to move away from them.
Yep, believe me, it gets completely stuck and you have to physically remove a disk to un-brick the puzzle! Every gate in my first model has two magnets, one on each side of the gate, of opposite polarities. So one pushes and one pulls the discs through the gate.
If I understand correctly then the puzzle becomes fully blocked if there's no pieces that want to jump into the currently empty slots. So if you assume a 50/50 mix of clockwise moving and counterclockwise moving pieces then you could create a dead-end state by having all the empty slots on one disc while all of the pieces in the 2 adjacent discs are ones which only want to move *away* from the currently empty slots. In that state you would have no way of moving any piece to a different disk.
Very nice puzzle. I think it will be very nice to use the 'concentric' infill pattern for the top and bottom surfaces. This will make them a lot more aesthetic.
I think this would be great in metal! Like aluminium or steel where the path's are laser/water/plasma cut. Metal is much stronger and doesn't deform while playing.
In the classical puzzle, each layer on each piece has an 120-degree angle. Starting from each angle, We can construct a regular hexagon. Each piece becomes three layers of hexagons. If we make many copies of such pieces, I wonder if can tile the plane in 3 levels. I guess the tiling process of the plane may need a synchronous motion of all pieces.
