best permanent magnet core materials 1

[gmeast]

Hello,

I'm new here. There are many companies involved in research or outright production of magnetic couplings. I am involved in an engineering project regarding a radial magnetic torsional coupling. There is an internal radial group of PM's in proximity to an external radial group of PM's, separated by a 1/16" austenitic (paramagnetic, nonmagnetic) stainless steel cylinder (gap). The PM's are even-numbered, 1 for 1, N-S-N-S, etc. on the internal and external radial group. There is no oscillating field. I'm regarding the fields as "static". What would be the best "back' material for the PM groups ... ie cast iron, some steel alloy, I don't think anything like MU metal is required though it has a relative permeability of 1,000,000 compared to the next best of 200,000 for pure Hydrogen annealed Iron.

I have tested a linear version of the coupling using NdFeB disc magnets stuck to 'dumb' mild steel strips (3/4" X what looks like 10 ga galv mild steel from Home Depot). Absolutely NO leakage. The back side of the strips can't even pick up a small paper clip! Seems like the PM field is adequately 'steered' from PM to PM as I said N-S-N-S, etc.

Thanks for any input,

gmeast

 

[logbook]

I am not convinced that I fully understand what you are trying to do. A drawing/sketch would help a lot. The question you have to ask (yourself) is what are you trying to achieve with the backing material? Is it really high permeability or is it really high saturation flux density. Any sort of iron is going to be good on relatively permeability. Saturation flux density is another matter.

magnetic saturation

Whilst magnet technology has improved dramatically in recent history, the magnetic irons are still in the dark ages. The best on saturation flux density is Hyperco 50
Hyperco
This stuff (2% vanadium steel) has been around since 1930 (from memory). At 2.4T saturation it is the best stuff around. If you don't need that extra 30% then any low carbon steel will do.

 

[logbook]

Opps. Senior moment. Hyperco is COBALT steel. It is 49% cobalt and that is the key thing. (How come this site doesn't have an EDIT button.)

 

[gmeast]

Hi logbook,

Thanks for the reply. I have uploaded a sketch that should clarify things.

gmeast

 

[logbook]

Interesting. Thanks for the sketch. I don't see any stainless steel. Is the 1/16th inch gap all stainless steel, rubbing tightly against one set of magnets?

What you have is like a synchronous motor. Basically almost no starting torque. If the prime mover drives the shaft too hard (fast) at startup the coupling will not be able to accelerate the load and the coupling won't hold. It then will be unable to start. However the stainless steel will act like an induction motor rotor and will give some starting and running torque. It is uncertain if this will be enough to get close enough to the drive shaft speed to lock.

If the total air gap is only 1/16th inch the requirements for the rest of the magnetic circuit are easy. Don't bother with Vanadium Permendur (aka Hyperco). Just use any old iron/steel. (NOT mumetal/radiometal). The iron/steel doesn't even have to be more than a few mm thick.

I bet gluing the magnets in place will be the biggest problem. The pull force could be immense!

 

[logbook]

I bet gluing the magnets in place will be the biggest problem. The pull force could be immense!

I have changed my mind about this. The magnets will probably prefer to sit in their recesses rather than jump across the gap. However, I think the recesses need to be quite shallow. If the magnets are say 2mm thick, the recesses probably need to be between 0.5mm and 1mm. Otherwise the field will take a shortcut (fringe) rather than jump the gap.

 

[gmeast]

Hi again logbook,

I see your point about the recesses for the magnets. The fields will choose a shortcut path for sure. Yes ... your assumption is correct ... the "gap" in the sketch is occupied by a stainless steel sleeve. I'm using a stainless steel alloy that produces virtually 'zero' eddy current loss. This coupling only serves to 'lock together' two members (the ring and the hub) that already have a very low relative angular velocity. The hub can be translated axially within the sleeve so it can be re-positioned under any number of rings running on the outside of the stainless sleeve. So, the sleeve isn't in tight contact with either the hub or the rings. The sleeve isolates two environments ... the hub's environment and the rings' environment. The sleeve is mostly just free floating.

As time and testing proceeds, I will be more than happy to share specifics of this beast if you're interested.

I'm uploading a short video showing the "no-loss" stainless steel material I'm using for the sleeve. My YouTube channel is "gmeast".

Thanks for your input ... especially about the magnet capture recesses.

gmeast

 

[gmeast]

Hi logbook,

Here's the link to the quick video showing the low loss Stainless Steel I'm using for the sleeve. It just shows the simplest of tests for Eddy Currents:

https://www.youtube.com/watch?v=_OY30sA6was

Thanks again for your input,

gmeast

 

[logbook]

It just shows the simplest of tests for Eddy Currents

Awesome!
I have supermagnets. magnets
I bought squares of aluminium, brass, copper and steel to try that exact test and there was absolutely no eddy current effect.
Then I saw what you had done and put TWO magnets side by side to get an intense field at the join. Suddenly it works! I can't thank you enough.

Brass is pretty poor on this test because of its higher resistivity. HOWEVER, looking at induction motor characteristics, a higher sheet resistance might just change the speed at which the peak torque appears.

As time and testing proceeds, I will be more than happy to share specifics of this beast if you're interested

Definitely interested.

 

[Sparweb]

Gmeast,
Interesting project.
I've done about 1/2 of what you want to do:

http://www.sparweb.ca

Magnetic brakes already exist, similar to what you want to build. Basically the proximity of a magnetic disk to a solid steel disk induces such strong eddy currents that a large rotor's spinning kinetic energy can be converted to heat in a very short period of time. Then it has to sit and cool down for an hour, of course.

I want to go back to the magnetic coupling idea. Basically you want to feed power into the central shaft, and have it turn the outer cylinder via magnetic coupling of the magnets, right? The outer shaft (in my imagining of the coupling) would drive an output shaft, which allows a load to be driven up to a certain amount of torque. If the load over-torques the coupling, the magnets will slip, thereby preventing mechanical failure.

If that premise is clear and correct, then you want to have the maximum magnetic attraction between poles as possible. Each pole being a magnet face with one polarity (either N or S) facing the opposite pole of the opposing magnet). Lines of flux must cross the gap with no deviation and at a high intensity.
You do not want to embed the magnets into grooves, as you've shown in your diagram.

If you want to simulate some design ideas I would recommend FEMM,

http://www.femm.info/wiki/HomePage

Heat dissipation when the coupling is slipping will be a major design issue for you. Once the heat passes the curie point, your magnets will loose their strength and the coupling torque will drop. Wrapping the rotor in the stainless steel band may help shed heat, but you will probably find that tricky to do safely, and maybe not as effective as you hope it to be. The outer rings will need cooling fins, like an induction motor, and the inner rotor may benefit from hollow holes through its core to allow air to pass through, too.

Oh yeah, and I think it's called "Carpenter", a supplier of specialty metals, including some with high magnetic permeability. But plain old electrical iron will do for your first go, and probably way way cheaper, too.


STF

 

[gmeast]

Hi SparWeb,

As you can see, logbook already made me painfully aware that the magnets must NOT be embedded as deeply as I have shown in my sketch. As you have deduced, what I am exploring is a torsional coupling more than a brake. Also, yes ... I am also planning to take advantage of the 'break-away' characteristics of such a coupling to prevent over stressing other components. In application (except for a 'break-away' scenario) there will never be high relative motion between the hub and whatever ring it is going to couple to.

I put together some NeFeB magnets and some dumb galvanized mild steel strips from HomeDepot as an initial look-see of a linear model of the beast. The mild steel strips are only about 3/32" thick, but I'm astounded at how well that material 'steers' the field of the magnets. There's virtually NO LEAKAGE from the back side!!!. After I post this reply, I will make another short video of my linear iteration an post that also to my Youtube.

I appreciate the input from you guys. I'm glad I stumbled into this forum ... dumb luck!!!

gmeast

 

[gmeast]

The latest accompanying video is at:

https://www.youtube.com/watch?v=CGwvorhIs7M&list=PLd5QXXXUSBRaUusW29cnCS8GpEMX-czEM

gmeast

 

[gmeast]

Well I took the advise of those who commented and made a design change to the test fixture which suggested an approach to a near-final design ... or at least 'a step closer'. So I've attached two sketches ... one is just a mod of my initial one and the second is a little more comprehensive scheme for securing the magnets without burying them in recesses. In that design, it is show some non-magnetic clamping blocks that will add some eddy current drag to help an coupling transitions .... I actually want some drag to act kind of like a 'synchro' in a manual transmission. They also act a bearing shoes.Thanks for viewing,

gmeast

 

[gmeast]

this sketch should have come through first.

 

[logbook]

I am not too happy with the wedge design concept. The induction/eddy-current startup concept seems fine. It is the mechanics of the wedges and the tapers on the magnets that bother me. You can buy any size of rectangular Neo magnets without paying tooling charges. But you now have two different types of custom magnets with tapers. Is that expensive?

The wedge is only tight when it has a magnet either side. I guess you have to assemble the whole thing loosely, then go round and round, tightening them up a little here and there until they are all tight. (I am assuming CR SHCS is cold rolled socket head cap screw).

You have to tolerance the whole thing so the wedges are guaranteed to be sub-flush in order not to use up your air-gap tolerance. Because there are so many parts, the tolerance build-up is going to be unpleasant (or you need very tightly toleranced parts.) But the wedge can’t sit on the hub as you have shown. It is over-constrained. You need a definite gap under each wedge in order to get it to grip at the sides.

That’s just my off-the-cuff thoughts on first inspection.

 

[gmeast]

Hi logbook,

The sketch you're referring to is NOT of the test fixture. What you're looking at and commenting on is a near-final design. In production, both the wedges and the magnets are sintering candidates and the cost would be low. I come from the gas turbine R&D field and always look at certain production considerations early on in any development program.

If you look above immediately after the post you responded to, you'll see the actual test fixture which has the shallow recesses for the magnets. The magnets I've chosen are simple 1/2" X 1/2" X 1/4", magnetized 1/2" square face to 1/2" square face.

I wanted to attach both sketches to the same post, but it didn't turn out that way. Following is the sketch belonging to the test fixture:

 

[gmeast]

Hi logbook,

Thanks for replying,

gmeast

 

[logbook]

What you're looking at and commenting on is a near-final design. In production, both the wedges and the magnets are sintering candidates and the cost would be low

Fair enough.

In case you, or anyone else, is interested, I just published some results on simple eddy current testing.

Eddy current braking on your kitchen work top

 

[gmeast]

Hi logbook,

Thanks for the pdf on the eddy current kitchen presentation. Isn't it amazing how easy it can seem to be to "bias" data. The political idiots in every government uses various techniques of 'data biasing' to further their respective agendas. It happens here in the U.S. all the time. Thanks for the post,

Greg

P.S. I sent out for quotes today on my 'Coupling Test Fixture'.

 

[logbook]

P.S. I sent out for quotes today on my 'Coupling Test Fixture'.

(Coudn't find an emoticon for waiting eagerly. On another site they have an animated popcorn-eating cartoon character.)