February 15, 2015

Finishing Up the Coreless Motor

First, some photos taken with one of Bayley's magical cameras:

Sorry I didn't focus it well enough to achieve the full 25 megapixels of resolution, Bayley

After last round, the only thing left to make was a ring to separate the two halves of the rotor.   I CNC milled these two rings out of solid 10mm plate, using a 1/8" carbide endmill.  (Very) Subtractive manufacturing for the win.

As you can see here, I really didn't leave much clearance room around the stator.  If I had made the motor any bigger, however, it would have been too large to machine on the MITERS CNC mill, which is limited to 7" of Y-axis travel.

A problem with the stacked-parallel-plates construction is that it is really hard to actually get the stacked plates completely parallel.  I spent way too long trying to get rid of the little bit of skew between the two halves of the rotor.  I was able to make it acceptable, but not nearly as perfect as I would have liked.

As predicted, with both halves of the rotor installed, the circulating-current drag I noticed before disappears.  Here's the motor spinning up at 24 V on a generic hobbyking airplane controller:

At full speed on 24 volts, the motor draws 2.4 amps.  Not too bad, really.  The spin-down time is satisfyingly long as well.  

Drill-scoping across the leads put the motor at 97 rpm/volt, which is a bit higher than I expected.  Maybe the lost torque constant comes from my modeling of the motor about its average radius, or the magnets being a bit weaker than FEMM thought.  Still, not too bad for a first try.  If I could figure out how to easily wind it with wire, moving to a more Servodisc-style winding pattern could give a pretty decent bump in torque constant while still allowing the stator to be constant thickness.

Line-to line resistance is 100 mΩ, or 50 mΩ per phase.  So in terms of motor constant (Kt/√resistance), this thing comes a bit ahead of the venerable 80-100 "Melon" airplane motor.  I just measured one of the 170 rpm/volt flavor at 50 mΩ line-to line.  Realistically, in its current form it's probably very thermally limited, but this is still a pretty sweet result.  Now I just need to find some derpy vehicle to attach it to.


  1. That. Was. Awesome!
    Questions though:
    1. Why did you make the rotor halves turn the axle? Would it not have been easier to have a static axle and run the wires out through that? How did you get the stator stationary? Is the left side bearing static?
    2. Why machine two rings and not simply one?
    Thank you for sharing!

    1. Stationary axle would certainly have been a simpler, but it's often nice to have a shaft coming out of the motor for attaching gears/sprockets for power transmission.

      2 rings for ease of machining on a 3-axis mill.

  2. Hi Ben your work is highly appreciate ! At your age ! Amazing to say the least certainly this project ! Hopefully one day I will build one like your but bigger to drive my FAT BIKE ! Greetings from Spain ; ) http://www.teslapower.rocks/

  3. Hey man, great job!
    I was wondering about the power weight relation of this particular kind of motor. Seems like yours have a maximum of 24V*2.4A = 57.6 W right?
    Is this the maximum power you can run it? What about efficiency?

    Depending on your answers I may do my master thesis on a motor like this one. By the way, the link for the thesis you posted is broken.

  4. Hey Ben,

    Have you had the chance to test it under sine wave control? If so, would you happen to have any footage of that? I'd love to know how much noise it makes because I'm actually trying to see if I can make myself something similar, and it must be very very quiet.

    Thank you