# Magnetic gear box

### Help Support HomeBuiltAirplanes.com:

##### Well-Known Member
I'm building a prototype magnetic reduction box (10:1) with a friend at work.
He's building a small turbine based on an automotive turbocharger with an external combustor, and looking at how to take power off a shaft spinning at 100,000 rpm.

If the prototype works we'll try for 20:1 with a designed power of 1 kW.
Theoretically it should have similar losses to a transformer: hysteresis in the cores.

It's just a hobby project, but we should learn something.

Here's our inspiration:

http://www.magnomatics.com/pages/media/pdd-gear-video.htm

http://www.magnomatics.com/pages/technology/low-ratio-magnetic-gears.htm

#### wsimpso1

##### Super Moderator
Staff member
Log Member
Cool stuff. What is their weight metric and how well do they do on weight?

#### rv6ejguy

##### Well-Known Member
Pretty neat, thanks for sharing. Interested to see how it works out.

##### Well-Known Member
Cool stuff. What is their weight metric and how well do they do on weight?
I have no idea!

Our machinist showed me the mag reduction video - I was thinking of a synchronous motor-generator pair with more poles on the motor side.

I've guesstimated a 1" diameter rotor, with an overall diameter of about 2", and we've ordered a bunch of magnets.

Goals:
See if we can make it work.
Measure the stall torque at which it slips.
Try and estimate losses.

The "planetary" bars will be soft iron, but I expect there will be eddy current losses, and that they really should be laminated strips as in transformers, motors, alternators etc.

To be honest, it's something completely new to me, and I thought I knew a bit about magnetism.
As soon as I saw it, we just had to build one.

#### pictsidhe

##### Well-Known Member
Forget iron, the losses will be horrific at turbo frequency. Have a look for electronics cores. There are materials good up to MHz available off the shelf as toroids that could be machined. Etch them with ferric chloride after machining to get rid of surfaces stresses. Those are very lossy again.

If I was going to gear down a turbi, I'd use rollers rather than gears.

#### Pops

##### Well-Known Member
Log Member
One of my grandsons is just 22 years old and works for the company in this field and received a Patent last year. Very smart, been in a school for gifted children since 8 years old. He is 6' 5" and 250 lbs, made the JMR to fit him since him and another grandson who will be owning the JMR someday. My other grandson received a share in a Luscombe when he was 16 years old.

#### RonL

##### Well-Known Member
HBA Supporter
Magnetic couplings have been around a long time, but this is a bit different and interesting.
The first thing that comes to mind in your stated mission, would be a heat buildup that makes the magnets less and less efficient.

#### Dusan

##### Well-Known Member
There is a caveat: there is a slip above certain torque. Using ferrite cores the slip torque will be much lower than silicon steel lamination as used in transformers. For higher torque, the diameter needs to be bigger.

#### Pops

##### Well-Known Member
Log Member
Sort of like a magnetic coupling used on an electric motor.

#### Jay Kempf

##### Curmudgeon in Training (CIT)

We built one of these a couple years back. Took a lot of head scratching to figure out how the gear ratio is actually achieved. We looked at combining this with an electric motor stage to get an entire prop with reduction drive in one brush-less unit. There is no limit to how you can stack these things. Not sure if it is strong, reliable or light enough to be practical. Easy to build prototypes out of 3D printed parts, bearing balls and magnets from Amazon.com

##### Well-Known Member
One of my grandsons is just 22 years old and works for the company in this field and received a Patent last year. Very smart, been in a school for gifted children since 8 years old. He is 6' 5" and 250 lbs, made the JMR to fit him since him and another grandson who will be owning the JMR someday. My other grandson received a share in a Luscombe when he was 16 years old.
Ask him when they are going to make a PSRU

It's actually probably not the answer for a PSRU anyway : the main reason we are looking at it is for the high gear ratios achievable.

#### pictsidhe

##### Well-Known Member
This is an offshoot from universty research, so there is probably public domain info on it. Seems easier tha reverse engineering it, anyway...

##### Well-Known Member
This is an offshoot from universty research, so there is probably public domain info on it. Seems easier tha reverse engineering it, anyway...
I have a grasp of the theory involved (I think), but for actually putting numbers to eg. torque limits a prototype is required, as so much depends on magnet strength, core material, air gaps, detailed geometry (e.g. wedges or arc segments) etc.

re core material:

The magnetic field in each bar in the "cage" reverses once (or twice) per input revolution, so about 1 kHz : comfortably within transformer frequencies.
Once we have some numbers from the prototype, I'll think about core materials. Ductile iron will do for now.

#### pictsidhe

##### Well-Known Member
Flux density will need to be low at 1kHz to avoid excess heat in iron. The laminations won't be orientated for structural strength. Seems like a difficult and inefficient way to do it compared to machining bits from a toroid to me. Have a look at the hysteresis loss of various materials before you use iron...

##### Well-Known Member
Flux density will need to be low at 1kHz to avoid excess heat in iron. The laminations won't be orientated for structural strength. Seems like a difficult and inefficient way to do it compared to machining bits from a toroid to me. Have a look at the hysteresis loss of various materials before you use iron...
The prototype will probably never see over 3000 rpm input.
I'm not setting out to machine multiple bars from a sintered core for the proof-of-concept, when low-hysterisis ductile iron will do.

The turbine version would ideally have something like 20:1 reduction and higher temperature magnets, too.

Look at 65-45-12

https://www.daycounter.com/LabBook/Physical-Properties-Ductile-Iron.phtml

#### pictsidhe

##### Well-Known Member
You can chexk the eddy and hysteresis loss without building a whole motor.

Make a single bar. Wrap an electromagnet round. Measure its resistance so you can calculate copper loss. Estimate AC voltage to flux density relationship. Power it yo with AC. 60Hz from a transformer will be quite informative. Measure the power dissipation and current. Calculate copper loss from current and resistance and subtract. Now you have iron loss at whatever flux density you aimed at.

#### Sockmonkey

##### Well-Known Member

We built one of these a couple years back. Took a lot of head scratching to figure out how the gear ratio is actually achieved. We looked at combining this with an electric motor stage to get an entire prop with reduction drive in one brush-less unit. There is no limit to how you can stack these things. Not sure if it is strong, reliable or light enough to be practical. Easy to build prototypes out of 3D printed parts, bearing balls and magnets from Amazon.com
Huh, that sounds like it could do both the job of a torque converter and reduction gear for a car with an automatic transmission.

#### TFF

##### Well-Known Member
For the big question, on a unit that is big enough for an engine gearbox, will my compass or flux gate point straight at it so I am always flying north? That of course is if I am not flying a pusher so I would be flying south then.

##### Well-Known Member
For the big question, on a unit that is big enough for an engine gearbox, will my compass or flux gate point straight at it so I am always flying north? That of course is if I am not flying a pusher so I would be flying south then.
The magnetic field rotates, so I think if you keep up a continuous barrel roll you'll be fine.

#### TFF

##### Well-Known Member
Ah. Which part do I roll to, inside or outside magnets?

2