I finally have a micro brushless outrunner motor that works. This is the result of a collaboration between Graham Stabler and I. Along the way it spawned the N20 magnet replacement tuneup idea of Jochen's and recently done by Michael (Epilot).

The heart of the idea (not original to us) is to use an existing 3-pole stator and then machine a outrunner to go on it and hold more and higher quality N45 grade NdFeb magnets. This required obtaining custom magnets to our specs. After analyzing it with a program Graham wrote to help, we settled on 5x1.5x1mm magnets and then ordered 600 of them to be made up in China.

Then, I machined the outrunner itself. The flux ring (can) is machined from 4130 cromemoly tubing on the inside and outside to the appropriate diameters. The front bell (with a lot of input from Graham) is machined from aluminum rod and has a lip that allows it to be inserted and glued in the steel tube. The "snout" has two Didel brass bearings at each end. The front of the snout has a flange to allow a CF prop to be glued to it. The outrunner will be held on the stator shaft with a brass washer and a press fit short length of tubing.

The stator will be held in the back with some sort of machined or wood bracket with a hole in it. The stator in this motor is from an M20. Future motors may be made for N20 stators as well.

The brushless ESC is the new Phoenix 10 from Castle creations. The circuit board portion weighs just ~2g and much thinner wires will be substituted.

There is a lot of work to be done. The important thing is that it runs on the bench with the Phoenix controller. I haven't even attached a prop to it yet. Nor have I determined which M20 stator to use. I'll probably use either the Toytronics or Kenway M20-HV's as I'll probably use this in a 2-cell plane since it will have to haul around a somewhat heavy speed controller.

Thanks on this go to Graham for helping me understand a lot of the concepts along the way. Matt Keennon was also a very useful sounding board for ideas.

Gordon

Looks nice - I assume you turned that up on your new Chinese lathe?
The air gap looks pretty large though. Does it work at all? Whats the stator from?

Trizza,

The air gap is large as I machined it before I had the magnets and was guessing. This is the proof of concept motor. The next one will have better air gaps and be much refined, and look better. Yes it is from my Chinese lathe.

This particular stator is from an M20-UHV as it's the cheapest of the M20's if one is going to gut them.

Gordon

Wow, its smaller than I thought! How much is the weight as is? Is the can a press fit to that front plate?

What sort of controller are you planning? I'm sure a super-light one could be built if one were willing to use hall sensors.

I'm watching with great interest here..

finally... someone's at least TRYING this!

nice... I take it you're aiming for ~7-10W? what's the current weight?

I'm sure castle could do a run of thin wire p-10's... maybe even drop some fet's and make it a p-2.5 or something? controller weight is fine... sky hooks rx i guess? need something narrow band that can drive standard servos on at least one port...

nice work...

How the ballbearings have fit? Or is it made without ballbearings?
I tried make one from GWS motor parts, but broblem was bearings. More specs, please!

Gordon and Graham,

fantastic job! I am curious to see your next version. If it does at this size what the bigger outrunners did to their can equivalents, you will open up a totally new world here! If now someone could get motivated to do a real small print ESC....

Nice one Gordon!

Can't wait to see TLB2 (Tiny Little Brushless)

Smaller controllers are definately possible but not the easiest thing to develope. I too wish "someone" would get motivated

Graham

Castle will not do a special controller (I've already talked to Patrick). Soldering on thin wires is so easy that there is no point in him doing that for us. The weight of this particular motor is ~3.5g, a bit less than an M20 at 3.8g. The can fits on a flange machined in the front bell, which is then simply glued in place.

Glad you guys are as excited about this as I am.

Gordon

wow v nice, how does it run?, ive seen it done before with a speed 280 but it suffered vibration with certain magnet setups cause by the coils pulling the can in towards itself. possibly not a problem on the scale your working on

Nice! I have no doubt it will work. You will probably be doing a lot of armature rewinding to find the right combination of wire diameter and number of turns. I have a feeling thicker wire/less turns are needed.

Try and get the airgap as small as possible on the next motor - it makes a big difference.

I think I have to get out the lathe and make some motors myself. That way we can make some comparisons.

Michael

PS: And Gordon: Sharpen that tool next time you turn steel

If you guys would point me in the right direction
I'll solder up a controller for it. I have no idea what
makes it spin but can learn real quick.
I dont have anything else to do till the mail comes today.
And then till tomorow if there is nothiing in the mail.
I can program a pic if that is all that is needed I have no hall sensors unless there are some in CD Rom motors or floppy or harddrive motors. I have some bar magnets like that and a couple M20s no flux ring though I can run out to a shop and make one. Call me motivated. Better yet instead of me making another motor could the controller be tested on a brushless or is it different?
Am I way off here do you just need a simple circuit like the one in the RCmicroflight DIY brushless article.


Billy

Billy, making a brushless controller is harder than you think. It involves a lot of A/D conversion. Just to start the motor running is not a simple thing. Once running you have to monitor the back EMF from the motor to keep the firing sequence in sync with the rotation. The timing needs to be right for the type of motor you use.

If you feel up to the challenge, go ahead, but I think you will discover why there are relatively few people who develop brushless controllers. We need to find a way to get Ulf Herder (Schulze) interested in micro models

The Castle Creations ESC sounds good if it is only 2 grams without wires. I believe that is lighter than the 4A Jeti.

Michael

A/D conversion is one of my specialties.

I do not have a PIC with an A/D converter on it yet
but will order one soon. Which one would be best for this job?
Maybe use one of the PICs to control an A/D converter.
I would like to work on it with one of the reprogramable ones.
I just got into PIC programming but it is similar to learning to program the gameboy I dont like the fact that you can not have nested functions(this means I can not turn it into a webcrawling spider ) with the ones I have learned about so far but I can deal with it. Sounds like something that can not be done in a day but I'll get started learning about it.

Billy

Michael,
You have to understand that this is about the 3rd metal object I've machined on my lathe. And at the time I did this one all I had was the crappy tools that came with the lathe, and no grinder to sharpen them on, or even cutting oil (used WD40 for cutting oil when doing this). Since then I've laid in a supply of tools and materials etc to machine with and from. This one is definitely a learning experience, and quite ugly. However, it runs!

Graham and I have intended to go for a 1mm or less air gap since the beginning. We know it makes a difference. But, without the magnets on hand, or the right size materials to machine from this is what was possible. I did not have solid rod then. And, the tubing I had allowed pretty much this size.

Billy,
From what I understand people like Schulze and Castle Creations and others worked a long time on the programming for their brushless controllers. Apparently the ability to start a motor that is not spinning, in a gentle manner, is quite difficult. The controller I'm using is sensorless, so no Hall Sensors. Magnets moving past a stator winding act like a generator and the controller monitors the increase and decrease in this voltage coming back from that wire and determines where the stator is positioned so it knows when to send a pulse of current. There's way more to it than that and I know I'll not be the person to write that code. Object-oriented statistical programming is more my area of expertise.

Gordon