In this post I want to share some of the design iterations and steps that have been made over the past months. About some info about the "Why?" behind the design check out this post about the history of the dUltimate design.
The first mockup
As mentioned in this post about the History of the dUltimate design, I knew that I should be able to fit everything on 120 faces, but the next question is "How?". How to fit all these numbers over multiple faces while keeping everything readable. One option would be to write ALL values for each dice type on each face. But that would result in a dice design the size of a house and would be unusable. So how do we remove te duplicates?
For this I printed out a disdyakis triacontahedron plano. Then I simply started looking at this shape and marking lines, faces, and corners.. grouping things together and assigning them with the several dice types.
The first few were easy: d20 requires 6 faces grouped together and they resembled triangles exactly like you would expect on a d20. d10 followed soon after simply by being in the same space as the d20 triangles, only it occurs twice across the die.
Next up was the d12. After roaming Wikipedia's pages on Platonic Solids and checking out how many edges, and corners each of the shapes have I noticed that the icosahedron (d20) has 12 corners. This made me realize that we could put the d12 on the corners of the d20 and simply mark all surrounding faces as such. With d12 out of the way the d6 would simply use the same trick as the d10: double occurrence across the die. See image of paper mockup below:
The second mockup
The d8 and d4 were a bigger problem. I tried combining several d12 /d20 faces to create areas across the die that could be used for d8. But since 8 doesn't fit evenly in 12 or 20, this was not such an easy task.
After many attempt I simply started drawing lines across the die and noticed that there were straight lines that went around the circumference of the shape all in the same plane. Doing some more lines suddenly I had "cut" the die in 8 separate pieces. I cut the die on the X, Y, and Z axis from the center. I'm not sure why this hadn't dawned on me sooner... but there it was!
I fiddled with this concept a bit and created the next paper mockup...
...and it worked!
I could name a dice type, roll the die, read the result and had a perfectly fair die!
However there were some obvious issues, these were the biggest at that time:
The size was too large to comfortably read things on the 'other' side of die;
The d8/d4 was absolutely confusing to use;
Numbers spanned multiple faces, which made it difficult to apply, but also difficult to scale.
Though, some enthusiasm got the best of me and I wanted some of these 3d printed.
The first prototype
Basically the same as the second mockup. I modelled everything using CAD and after many attempt created the first prototype which I went and painted red:
This was SO much better than the paper mockup. Not only could this properly roll, but it also had some nice weight to it which actually made me feel like I was wielding a die. The size was large enough to fill your hand, but not too large to become unusable. We're getting somewhere!
I was pretty happy with the d20 and d10. They were easy to sample and read. Same for the d12 and d6 (which was very understandable with the pips.)
The d8 and d4 however were still not so good. The rule at this point was: "Look at the topmost d4/d8 value." However due to close proximity and sometimes the closest one was in another pentagram this was deemed too confusing and had to be improved, but I still had no good solution.
The second prototype
Without having solved all the issues I wanted to try a different 3d printing supplier while at the same time decreasing the size a bit more. Nothing was changed in the design, but there was still this feeling that something had to change.
Especially now that the size was 'only' 50mm in diameter, the digits that crossed the edges suddenly became less 'sharp'. If only there was a was to fix that issue completely.
This post is getting long enough already. Next time we'll continue with the third prototype - that's were the challenge of producibility comes into play!