52

weeks

Week 47 - 3D Printed Fidget Spinner

November 29, 2023

This week I made my first 3D print, a fidget spinner.

Tools I used:

Context

I’ve recently had the opportunity to be around a lot of 3D printers and people who know how to use them (more details on this to come next year). I’ve always been curious about printing in 3D, and I think this is going to be a growing part of the next few decades of manufacturing. After dipping my toes in the last few weeks, I liked it enough to purchase my own 3D printer (it should arrive this week or next). Until then, I decided to try my hand at printing a few toys to learn about the process.

Process

Thingiverse is a treasure trove of models that people have designed and printed, including many models that people have ‘remixed’ ie taken someone’s model and made it their own. A couple of fun ones that stood out to me at first pass:

parts of the process, I plan to spend more time on the ideation, design, and modeling at a later point. With that in mind, I picked something of low to medium complexity that had some moving parts: a fidget spinner. The particular one I chose was a remix by user Psuemno that had 4 sections, 3 outer rings and a system of planetary gears inside.

It was easy to download the STL file and then view it in Rhino:

The next step is to ‘slice’ the model. There are many different software that do this, typically every printer brand has their own software. One of the most common and user friendly ones is UltiMaker Cura.

I discovered that this is the part of 3D printing that still has a lot of room for improvement in the user experience. There are hundreds of possible settings which all meaningfully affect the print, and if you don’t get them correct can cause an equal number of unexpected and often print-ruining results (and sometimes the print only fails halfway through or later, which can be 5-20 hours in!).

that it seems like there is inevitably a bunch of trial and error, and several failed prints, before you get a good result. That’s part of the fun!

In the slicer software, once you’ve sliced the model, you can see the path of the print nozzle. I guess you can use this to anticipate mistakes, I’m not that advanced yet.

So, after finally understanding and tweaking the settings (this time I was using an Anycubic Kobra Max printer), I created my first print!

It worked great! Almost. The inner rings spun like butter (and I find it incredible that it’s possible to print those moving parts in place, no additional assembly required!) and I had a lot of fun spinning them endlessly.

However, the planetary gears in the inside were fused with the next outer ring and could not turn. This was because I shrunk the model by 50% as I wanted it to fit into my hand, and the nozzle width was therefore not small enough to capture the distance between each gear and the walls they needed to slot into.

In the image above, I drew a line from a point on the gear to the outer tooth and it measured ~0.5mm. The nozzle head I was using was 0.2mm width. Since it’s printing the walls on both sides, rule of thumb is to leave more than twice the nozzle width in space between walls if you don’t want them to touch, as the filament can expand when it’s heated or cooling, and the printer only has a certain precision. The point I measured above is not the smallest distance, it gets much tighter at certain points, hence the gears were too close and weren’t free to spin.

What’s the solution? As with most problems, there are many; the approach I chose was to shrink the gears relative to the outer rings.

  • I isolated a gear, then used ‘cage edit’ to shrink it only in XY plane while maintaining the same height in the Z plane.

[](/images/posts/52w-week-47-3d-printed-fidget-spinner/93ad55af-069f-4a1a-a1cf-62abb7bf71ba_500x469.gif

  • Rather than shrinking each gear individually, it was easier to do it once and then copy it 10 times using ‘ArrayPolar’, which lets you make a copy with rotation around a central point.
  • Then I used a clipping plane to scroll through each layer to make sure the new modified gears weren’t intersecting the outer rings, and rotated them if necessary

Several prints later, I finally managed a print that … still only kinda worked!

Learnings

  • 3D printing has come a long way in the last few years, but it’s still a very involved process, with thousands of variables and settings, lots that can go wrong, and much frustration and tinkering required. That’s part of why it still isn’t more widespread, but these seem like mostly solvable hardware and cost problems that will continue to improve year over year

  • Seeing something generated digitally translated into a 3D object is immensely gratifying

Next steps

  • Design and model my own printable object from scratch

  • Explore 3D jewelry design

  • Scan my head and print a planter of it

  • Scan my face and print myself new glasses frames