Very interesting and informative... thanks for posting all that. I think I'm starting to see how you can encode a maze in there. If I'm correct I think you may just need a one part nut... Correct? By that I mean there wouldn't be a second ring which was free to turn inside the nut. I'm really curious to see more.
I envisioned a two-piece design, with the pin mounted on a ring concealed inside a larger collar (which could be "nut" shaped for thematic reasons).
For the purpose of demonstration, I sketched a simple hysteresis maze about 12x12 in size. The user controls up-down movements, the maze controls left-right movements. It would be self-resetting, like Oskar's mechanical mazes, so that the pin will return to the start any time the user moves to the bottom. This maze will be mapped onto the bolt-like cylinder.
(I'm treating this as a dry run, so I won't slow down to fix mistakes as I go; I'll just make note of them for future reference.)
So here's the maze I'll use for this test design...
File comment: A hastily-designed hysteresis maze, just for demo purposes.
IMG_7476.JPG [ 208.36 KiB | Viewed 1129 times ]
The next step is to map this maze onto a bolt-like cylinder. Unlike the Mk3
I want the features of the maze to be visible so I chose to adapt the modulated-peaks technique by making the peaks flat and wide to make room for very distinct maze elements.
My next adaptation is to increase it to 12 threads, so each peak represents one entire column in the sketched hysteresis maze. The grooves of the threads will act as guides for the nuts, and also mark the boundaries between the column positions in the maze.
(At this point I discovered my first mistake: my maze sketch didn't allow any space for wall thickness. That is to say the lines in the maze are just lines, not solid walls. For now I arbitrarily thickened the features where possible, and omitted any that would have overlapped others.)
The "bolt" would look something like this...
File comment: Bolt-like hysteresis maze
Skewed shaft.png [ 100.73 KiB | Viewed 1129 times ]
(My second mistake, the 'maze' features have the wrong angles so they won't guide the pin smoothly from one position to the next. There are step-like jumps in places that ought to be aligned.)
The "nut" isn't nut-shaped at this point. It has 12 helical teeth-like threads that ride in the 12 threads on the bolt, with sufficient space to ensure it won't collide with the raised features of the maze.
Inside the nut there's a separate ring with a raised pin that's guided left-right by the maze. It needs some sort of mechanism to make it stop only at the 12 positions corresponding to the 12 columns in the maze. On my physical "nut" model I had used magnets, but there I just put 12 detents on the ring (not shown in this view). The sides of the pin are skewed so all sides will meet the skewed features of the maze at the same tangent.
File comment: Although it's not nut-shaped, this is the "nut" with a floating ring whose pin traverses the maze.
Skewed ring and pin.png [ 56.02 KiB | Viewed 1129 times ]
When the collar ("nut") is installed on the maze, the user can move it up or down by screwing the whole collar left or right. The movements of the collar are synchronized with the threads, so the pin inside the collar will only move left or right relative to the 'maze' when it encounters features of the maze that push it sideways onto adjacent peaks.
File comment: Collar (nut) installed on cylinder (bolt)
Skewed shaft and collar.png [ 82.85 KiB | Viewed 1129 times ]
This particular model has several faults that would prevent it from fully working, but it illustrates the basic idea...warts and all!
For reference, I've attached a 3-dimensional PDF of the bolt-like maze cylinder.