One of the best parts of Combat Robotics is running into unusual problems and
coming up with interesting solutions to them.
is a great robot, but it was literally hitting the other robots so hard that the magnets inside the
S28-150 MagMotor were coming unglued and sliding around the inside
of the motor. I have several theories as to why this is happening, such as sheer impact forces or differential
thermal expansion between the magnets and the can causing the existing adhesive to fail, but the bottom
line is, the problem has to be fixed.
Surprisingly, even after the magnets go walkabout, the motors continue to run reasonably
well! However, it's not a good thing, and I had to find a way to glue them back in place and
keep them there.
After much consultation on the
I decided that a decent approach would be to use some Scotch 2216 slow-curing epoxy to glue the
magnets back on. 2216 is strong, heat-resistant, bonds just about anything to just about
anything else, and best of all, I had some!
big problem was clearly going to
be getting the magnets into the right positions. S28-150 MagMotors use Neodymium
rare-earth magnets that are hellaciously strong; they clamp onto the motor can and
each other like remoras on sharks. To get them positioned to an accuracy of 0.01"
or better was clearly going to require some sort of jig to hold them.
My initial thought was to mill a jig out of UHMW, but then I had a brainstorm.
Quick as a flash, I yanked out my trusty bag of
This stuff looks and feels just like UHMW or Delrin, with a twist - pop it in hot
(150-degree) water and it turns soft, translucent and moldable - like plastic
playdough. It's great for prototyping parts. So I squished a blob of this
into the bottom of a MagMotor can with correctly positioned magnets and made a template.
It worked perfectly; I now had a nifty mold that would hold the magnets in
exactly the right position relative to each other. So I sanded and cleaned the back of
the magnets and the motor can (I used denatured alcohol and cotton swabs to for the final
cleaning), mixed and slathered on the epoxy, and slid the magnets
down the can into the slots in the mould. Note: it's important to mount the magnets
properly; there are two types of opposite polarity (one type has a red mark on it), and
adjacent magnets must have opposite polarity. So around the rim you mount dot, no dot, dot,
no dot. Or to put it another way, dots across from each other.
I also put a little fillet of epoxy along the top of the magnets.
After giving the epoxy a day to cure (it's slow cure, remember -- it actually takes
a whole week to get to full strength), I tried to slide out the form. Whoops, it was stuck,
either just tightly locked by friction or actually glued. No problem! I just popped the
whole kit-and-kaboodle into some hot water to soften the ShapeLock, and out it slid. My
hunch is that the hot water expanded the motor casing just enough (as opposed to the ShapeLock
softening enough to slide out), but since I'd softened the ShapeLock anyway, I let it get
fully soft and remade the mold.
It's really neat stuff!
After having disrupted both kitchen and dinner with my magnetic cooking, I went back
upstairs, mounted a second set of loose magnets in a second can, and put a thick fillet of
epoxy around the perimeter of each of the magnets in the first can I worked on. My theory is
that these fillets will act as berms to hold the magnets in place even if they do manage to
come unglued. Of course, if the berms come loose, they might jam up the motor themselves!
I'll repair at least one motor without the berms to test that hypothesis.
MagMotor Trivia and Tips
During this project, I've learned a few things that I'll document here so that
hopefully the next poor schlub who has to mess around with his mangled magmotor can
benefit from my mistakes. I'd also like to take the opportunity to thank the long-suffering
people at SatCon (makers of the MagMotors) for tirelessly answering my usually-dumb questions.
That's right, I looked like an idiot, so you don't have to!
So, in no particular order:
When rebolting the motor together, using a long, thin T-handle hex-key to probe and find
one of the bolt holes makes it much easier to get the other bolt aligned and seated. See one
of the photos for an example.
Remember to scribe a line on the motor can and endbell as soon as you get your MagMotor
to mark the neutrally-timed alignment.
When reassembling the motor, I find it best to reassemble it as neutrally-timed, and to
alternate tightening the bolts 1/4 turn at a time to help "seat" everything. Then if you
want to retime the motor, loosen the bolts, rotate the can, and retighten them the same way.
Screwing down the brush-holder caps too tight can jam the motor. They should only be
tightened until they are flush with the brush-holder housings. A bit of removable LockTite
is probably a good idea.
The two circular, bent spring-rings both go between the front armature bearing
and the frontplate. If you add an encoder to the motor, only one is used.
The bolts that connect the frontplate to the endbell and clamp the motor together are
3" 10-32 size. I have taken to replacing the stock bolts with stronger cap screws from
McMaster-Carr. The McMaster part
number is 91251A360.