Created by Ryan Beal in Feb 2021
Printed on Ender3 with Crealty slicer; Priline pla (wood color)

naming convention is
r - rings
g - gaps
rs - random seed used in SCAD rotations (for debugging)
b - ball diameter

so a name like 3r_2g_rs42_b3-b5
would indicate a puzzle with 3 rings, 2 gaps in each ring, with multiple ball sizes

Design Notes:
Create your own randomly generated puzzle with Open scad!
After printing insert small steel balls at center of puzzle.
Shake or spin the puzzle, to force the balls to the outer storage ring
They should not be able to escape the puzzle except at the center. (see lip notes)

Balls:
I'm using these ones: uxcell 5mm Bearing Balls 304 Stainless Steel G100 Precision Balls 50pcs https://www.amazon.com/uxcell-Bearing-Balls-Stainless-Precision/dp/B07YKQDYLL/ref=sr_1_3?crid=TB3QOQR6P2TL&dchild=1&keywords=5mm+ball+bearings&qid=1614578279&sprefix=5mm+ball+%2Caps%2C309&sr=8-3

Solve times:
3r-1g: 10-20 sec
3r-4g: 15-30 sec
5r-3g: 20-40 sec

OpenSCAD Notes:
OpenScad is a programmer's 3d rendering tool.
you can tweak the build plate by changing the numbers in this function
general_puzzle(rows= 5, gaps=3)
You could create any size puzzle you wish.
Furthermore Tweak the random_seed variable
but note that not all puzzles/seeds are valid, since the gap could be too close to eachother, especially as you increase number of gaps. For now just try a few numbers until you got a puzzle worth keeping.
This may be fixed in future revision; but this would be "procedural generation" instead of "random generation", and that upgrade feels like too much work tonight. Especially since puzzle enjoyment is not linearly tied to number of gaps.

IMG Notes:
notice how in both compare you can see one is 'rougher' than the other,
this is the difference in the face num $fn var in SCAD to determine how smooth to render the stl files

Few of the print in IMG here are not perfect as you see some lift on the central wall lip
I didn't see this problem with the Eryone Wood filament, and decreasing the heat also did help
In the future I could also Tweak the lip to be shaped better for printing with more gradual steps / ramp instead of a sudden transition.
For me since the number of prints is limited, so i will just rework the piece, remove the burrs and 1 or 2 hangs which sometimes happen with a razor

Slicer Notes:
Slicer choice:
For me sli3r was creating a lot of stringing because the wipe was not done on internal area before lift and

continue; even tho the options are set the sli3er output was not as clean.
I used Crealty slicer as it generated more appropriate g-code for these objects.

Wall lip:
Double check your gcode to ensure the slicer is doing someting sensible at the layer transition for the lip.
the sudden transition is not ideal, but seemed to work just fine with my printer/filament settings

Lip Notes (cont):
Puzzles with gap in the same angle for multiple row will create escape points as in my 3r_4g file
if they are even slightly offset there will be no ball escape; There are 2 good example to see this in the 5r-3g file