Dual rotor axial flux motor design

[rhitee05]

The two-rotor design has good and bad aspects for the bearings. The good is that the magnetic forces will nominally cancel each other out as far as the bearings are concerned - the internal braces between rotors will take care of those loads and there will be net zero force on the entire assembly. The bad is that it's an unstable equilibrium, such that a slight shift to either side will unbalance the forces in such a way that the imbalance wants to get larger.

I'm not an ME and I certainly don't know much about bearing design. But, given the above, it seems like you'd want bearings that are a) precise, and b) capable of handling the unbalanced forces due to modest misalignment. So you would need bearings with at least some thrust capability, but probably not enough to withstand the entire magnetic attractive force.

[Kingfish]

I can help with bearings…
For most needs, follow examples by the automotive industry; they use tapered bearings for axles. These can be externally greased ala trailer and hub axles, or sealed (maintenance-free). There are three companies that I have dealt with in the past, all fabricating reliable product.

• ABT
• SKF
• Timken

The issue generally is with supply; pick a rep that is prompt with low shipping costs in your neck of the woods. My local SKF supplier took about a month to send me the replacement hub bearings, whereas one supplier delivered SKF seals from half-way across the country in three days. Personally, I enjoy calling up the reps and negotiating with the sales because they often have great insight on matching product with the application and saving a lot of time and expense in the process.

Know before Ordering
You will need to spec out your axial load, radial loads (static and dynamic), maximum shaft speed, maximum temp, ID (shaft diameter), OD, and allowable thickness (affecting cup design). It’s a little bit cumbersome with the wide selection, though I would lean towards both Automotive and Fuel-Efficient designs. Agreed that you will want the sealed bearings.

Regardless which manufacturer you select, I am certain that tapered bearings will suit your design well for precise positioning
Good hunting, KF

[Miles]

Thanks guys.

I'm not sure I could get a tapered roller bearing small enough... I'm only using a 12mm shaft.

I guess the first thing is to estimate the maximum axial force that's likely to be experienced.
Some bearing pre-load should take care of axial movement.
Need to find a way to balance the two gaps accurately.
It may be that deep groove radials will be sufficient.

Otherwise, I was thinking to use a pair of these 40deg. angular contact bearings:
http://medias.ina.de/medias/hp.ec.br.pr ... A7?clrsb=1

[Miles]

This is the smallest tapered roller bearing listed:
http://medias.schaeffler.de/medias/hp.e ... I4?clrsb=1

I could go to 15mm hollow shaft........

[bearing]

I think the angular contacts are better for your application. They should have less drag.

[Miles]

The other possibility is an opposed pair of spindle bearings. You can get 15 deg. or 25 deg. contact angle in the small sizes.

[Kingfish]

Timken makes a single-taper bearing that goes down to 8mm, though I think you are wise to upsize the axle for more common product as it is bound to be less expensive

If weight is an issue, you could make use of hollow shaft stock. Just an idea...

~KF

[Miles]

So they do: http://www.timken.com/EN-US/products/be ... up%29.aspx

[Miles]

Crikey!

Core cross-sectional area has gone up 1.375x

Volume of copper has gone up 1.45x

Number of turns down from 6t to 4t

2 phases length down to 3 metres.

[fechter]

36W of wasted heat will barely get warm. And that's at full power. Who needs cooling?

[Miles]

Who needs cooling?

Quite.

You have to add the parasitic losses to the 36W but if I use G-O steel and 0.178mm laminations, they should be pretty low.

[Miles]

If we don't need the through venting, we can pack more copper in and get:

[Miles]

Km would then be 1.21

Weight has crept up to 1.4kg because of the additional material, so specific Km is 0.87

Scorpion S-5545 10t Delta: http://www.scorpionsystem.com/catalog/m ... 545-150KV/

Weight 1.026 Kg

Kv 150 rpm/V

Kt 0.0637 Nm/A

Rm 0.03 ohms

Km is 0.0637 / SqRt 0.03 = 0.37

[ 0.37 / 1.026 = 0.36 ]


Astro 3210 8t Wye: http://www.astroflight.com/pdfs/3210WEB.pdf

Weight 1.0 kg

Kv 169 rpm/V

Kt 0.0565 Nm/A

Rm 0.08 ohms

Km is 0.0565 / SqRt 0.08 = 0.20


Ok, the Scorpion is in the lead....

It's not now...

[bearing]

Number of turns down from 6t to 4t

2 phases length down to 3 metres.

With the same kV?

[Miles]

Yes, that was the assumption. There's quite a lot more iron, now...... We'll see.....

[Miles]

Just came across this one:

Design of an axial flux machine
for an in-wheel motor application
Masters thesis by Christian Du-Bar.

[Miles]

Arnold Magnetics Thin Gauge chart:

[fechter]

What about asymmetrical pole pieces?

I realize this is way 'outside the box', but there could be some benefit at high loads. Normally the 'teeth' are centered on the windings. The teeth need to be there to maximize the active area and still have room for copper. Everybody always centers them. I guess if the motor needs to behave exactly the same in both directions, this may be the right thing to do. Most of us run the motor one way or the other most of the time and rarely need to reverse it.

When the iron starts to saturate at high loads, it will be almost like it 'shrinks' or has reduced active area. By offsetting the teeth, this effect can be minimized. The iron in the attracting side of the tooth will saturate first, so shift the tooth over so more of it is on the repelling side. You'll still have the same room for copper, but instead of an "I" shaped cross section, it would be more "C" shaped. At light loads where the iron is not close to saturation, the change won't make much difference, but at high loads the active area should 'shrink' less.

One disadvantage is you need to build the motor for a particular rotation direction and reversing it would require flipping the stator around.

I think this may have more benefit on a radial motor than an axial one, as the teeth get really skinny at the tips.

Maybe not worth the gain, but an interesting idea.
I don't expect MFEA models to accurately reflect this concept.

[Miles]

That's a cunning idea.... It's a pity that it doesn't fit in with my 2 speed retro-direct gearbox concept, which does require the motor to operate in both directions.....

[Miles]

Carpenter Hiperco 50A losses graph:

[kenkad]

fechter,
Your comment about the tooth offset for single direction rotation is interesting. Some two years ago, I had asked Shane Colton (when he was deep into dual rotor axial flux design) about this type of thing and if a better solution might be to 'skew' the offset of the dual rotors a small amount (maybe somehere between 1-3 degrees). He was going to try that to see if it also minimized cogging, but, I do not think it finally worked into his efforts. At least he never indicated that there was any noticable effect. I personally believe that it would be more noticable as a minimization of cogging. Since I am not a motor designer, this may be way off base. This is on my list (rotor skewing) of things I want to test as well as magnet orientation of the adjacent magnets on the rotor (not the typical N/S in/out orientation). Isn't learning fun?
kenkad

[Kingfish]

Some months back someone posted an induction motor that had renderings, photos, or animations of a C-shaped (as opposed to skewed) laminated stator by some big-named company (possibly GM or Honda). The stator had big single-copper wire turns. My understanding was the skewing and C-shape is designed to reduce torque ripple, although I find the affect upon saturation interesting and appealing. Unfortunately most of us do not have advanced-level modeling software and full-time researchers figuring out the proper mechanical offsets per each lamination.

Checking my lotto numbers, KF

[Miles]

Just looking at the Sineton dual stator radial design (below, top left):
http://www.sineton.com/web/index-10.html

[Miles]

N.B.
"Core losses are substantially reduced with the use of a grain oriented silicon steel."
http://www.sineton.com/web/index-33.html

It's not clear how they take advantage of GO steel with the above topology...

[Miles]

Latest version: