This is something I've been wanting to do for awhile. I have a small desktop machine that I built over a year ago, and it's needed a reasonable spindle. I had the idea of using a brushless motor to run one, but at the time I didn't have the facilities to make it happen... Now, after having acquired a bunch of brushless motors and ESCs for other projects, I have the parts lying around to make it happen.

For the spindle itself, I have a 1/2" shank ER16 collet chuck that I got off ebay. It's mounted in a couple of ABEC-1 ball bearings rated to 16k rpm in a stainless housing that I bored. I don't yet have a way to preload the bearings, it's just turning free in them, but I'll probably use a shaft collar and a shim. The spindle housing is mounted in a flat of aluminum with a couple holes in it that came with another small spindle that I bought off ebay for this purpose, but which had an oddball collet that I couldn't really use.

The original spindle had a little geared DC motor, which is what I originally tried to use. The gearbox hao it's output shaft non-concentric with the outside of the motor, and belt tension was achieved by rotating the motor. Unfortunately, that motor turns far too slowly for my application.

That's why I started looking into the brushless motors again. I wanted to keep the same eccentric tension system that I'm using now, so I turned a small aluminum cup with an OD that matches the OD of the old geared motor, and thin 1/32" walls. I then drilled holes on the bottom of the cup that match the mounting holes in the bottom of the motor, but I drilled the pattern 0.056" off center, so that I can rotate the whole cup in the mount and the motor shaft is eccentric to the mount, allowing me to tension the belt.

The motor I'm using is on the small side, but I had it lying around. Here's info on the particular motor I'm using:

http://www.bphobbies.com/view.asp?id...7&pid=B1488466

180W isn't a ridiculous amount of power, but for this little machine it'll be alright for proving out the brushless concept. Just hooking up an ESC with an RC tx/rx to run it, it runs the spindle quite well. I'm going to run it off a 13.6V 20A bench supply. It's a 1400kv motor, so it should develop around 19k rpm. I have it driving the spindle through a 12:60 timing belt setup, so the 5:1 reduction will give me around 4,000 rpm in theory. In practice it'll turn a bit slower than that due to losses, but I think that's a good starting point. That'll be at max speed, of course; the ESC will allow me to control the speed down from there.

These work via PWM 3-phase DC power, and max torque is developed pretty much at any reasonable speed.

Control and power is going to be the biggest issue, and it shouldn't be a major deal. I'm using a cheap chinese 25A speed controller with this motor, which is more than adequate, as the motor shouldn't draw more than 15-18A. It'll likely burn up if I load it any more heavily than that.

The speed controllers expect an RC servo pulsetrain, as they're intended to be hooked directly to an RC receiver. I'll need to make a small microcontroller circuit to provide this pulsetrain. Also, the speed controllers have some intelligence to them, and they require you to provide an arming signal of 0% speed for a couple of seconds before they'll arm and allow you to spin the motor. My microcontroller will accept an input signal like you'd feed a VFD so my G540 can run the speed for me. The microcontroller will then handle converting that to the servo pulsetrain and deal with all of the arming logic.


Here's some pics of the spindle with the motor mounted. I'll get some video of the spindle running later; I'm quite pleased with the runout on this spindle!