So. I’m trying to figure out how to best epoxy the 128 magnets into the rotor can. Right now I’m kicking around 3 or 4 different ideas. I’m hoping writing about it will help me settle on one. Otherwise, I feel like I’m wasting time, sitting in D-Lab thinking as GoPro films a time-lapse of me shifting in place.

Epic Epoxying of Magets

MAGETS! No! Not the Magets!
Mess of clamps

This is the result after yesterday’s mess of clamps.

I scraped off some of the epoxy residue left on the inner face of the magnets so the 3d printed magnet support ring will slide smoothly inside.


3d printed magnet support cylinder fitting snuggly in between magnets.

So, what’s the next step? Well, the next step is to epoxy all the magnets (preferably in the least amount of time, which includes the time I spend thinking about it [Meta-design! Ahhh!]).

Magnetic field scribbles with Dana

If I hadn’t talked with Dana about magnets today for 3 hours, I would have epoxy’d the other 8 magnets completing all the magnets of one direction of the Halbach array. But! I did talk about magnets today with Dana, particularly about my derp mistake last night (why two rectangular magnets stacked, in a stable state, into a rectangle twice as long results in the magnets’ magnetic field pointing in opposite directions [something I definitely don’t want going on inside my motor]). We finally figured it out, something really cool about the balance between the strength of the magnetic field and the volume in which it occupies that determines whether magnets will attract and settle in a weak, but stable state, or repel. We learned more about magnets by reasoning from their fundamental characteristics than anything we learned about them in 8.02. Anyway, that’s for another post (which I won’t ever write because time! Haha! like I have that just piled up in my time safe).

Point is, while talking with Dana about my strategy for epoxying magnets, we thought of some other strategies that could possibly finish the job in 1 or 2 days.

Strategy 1: CNC’d 

The idea is to CNC a circle similar to the 3d printed magnet support cylinder, but have a cutout 2 magnets wide so a magnet stack could be pushed in radially. The aluminum circle would align off the first 8 magnets already epoxy’d, and it would rotate around as I started filling in magnets through the slot. The CNC’d aluminum would then have enough strength to hold ALL of the magnets in to cure overnight, and I could clamp the two plates, from my broaching jig, on top and bottom to force the two magnet layers together (since they repel as discussed with Dana) without clamps, holding the magnets in radially, getting in the way.

  • All magnets at once
  • Residual epoxy fills next magnet’s cavity
  • Can clamp plates on top/bottom to hold magnet layers together

However! The force necessary to constrain the magnets radially is very high! I definitely want to use a clamp, otherwise the friction force between the CNC’d circle and the 62 magnets would be too much to spin it to the 63rd magnet’s slot (it would probably just spin inside the bore’d out wheel, because that friction fit isn’t very tight)

  • Too much friction


I decided I can’t shove the magnets in axially (from the top), using the 3d printed magnet support to hold them all in place, because all the epoxy would squish out resulting in a very thin layer actually adhering the magnets to the rotor (which would ultimately fail under the vibrations and loads from the wheel). That leaves radial shoving (pushing the rectangular surface of the magnets into the groove).