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

 

[MagBen]

Of course, the scale measured the force, not the real torque (different unit). By saying that I meant the torque in gmeast's set up where the torque = shear force x radius. since the radius is a constant, so torque is proportional to shear force, so the reltionship holds.

AnSYS Maxwell is a popular one for electromagnetic simulation. I am used to Magneto 2D and Amperes 3D softwares by Integrated Engineering Software. SOMSOL is powerful tool, but maybe too general for such kind of torque simulation.

 

[logbook]

I meant the torque in gmeast's set up where the torque = shear force x radius

Ahh! Yes. Now we agree.

 

[gmeast]

Hi logbook and MagBen,

I'm going to construct a linear test fixture to accurately measure the shear force of two magnets. logbook, your setup has inspired me. MagBen, your analogy is also correct in how the measurement of a single pair of magnets put in shear can be projected onto my coupling investigation. What I need to do is account for the gap between the magnets ... like the one that actually exists in the coupling. This is easy to do. I need only keep the magnets separated by some little brass needles that have a diameter the size of the gap. This might need to be a horizontal setup with a cage for the needles and a support for the cage. I'm uploading a sketch of it.

gmeast

 

[logbook]

GM, without wishing to deflate your “inspired” state, what you would measure with your setup is not representative of your actual task by quite a bit. If you had two magnets on top and two underneath, with iron backing for each pair, then that would be more representative. At the moment you have a huge air gap for the “unused” pole faces of the magnets. The flux needs to get around from one side to the other. When you do that the field will be massively increased.

By the way, my tiny little 2mm thick magnets were 3.5mm apart. If you get your monster magnets 1mm apart the compression force is going to be huge!

Why not just complete the assembly of your rotor and stator and test that? Or are you trying to save a few days before the other parts arrive?

 

[gmeast]

Hi logbook,

You're right ... don't know what I was thinking. When I build it, I'll do exactly as you have pointed out ... a pair (or maybe 3) above with iron backing and a pair (or maybe 3) below with iron backing. I guess what I was actually trying to show was a practical method for representing the gap with 'frictionless' brass needle bearings so there would be no sliding frictional losses but only rolling losses to impact the results. What I've started is nearing completion as I assemble the frame for testing the coupling ... which consists of a 2" slab of wood, a couple of spacers, the moment arm, an overarm supporting the Hub's axle shaft end ... well that's most of it.

The linear test fixture mechanism will be used for extrapolating the results of the first Ring and Hub tests. It will be much less expensive than a new Ring and Hub for every iteration. Thanks for pointing out my mess-up.

gmeast

 

[logbook]

Design info for permanent magnets K & J

N42 curve

Taking each magnet assembly as a complete magnetic circuit we have 2 off 1/4" magnets and a 1/16" inch air gap. That makes 1/2" of magnet per 1/16" inch of air. I get a permeance coefficient of 8 to draw on the N42 curve, giving an impressive 1.2T field strength. Given 1/2" square magnets you don't want the iron cross-section in the direction of the flux to be less than half of that or it will be saturating. In other words your backing steel wants to be at least 3/8" thick.

 

[logbook]

ok GM, I'm going to put my money where my mouth is, apart from there being no money involved for me

I am going to predict the peak static shear force on your jig as being nominally 8 lbf/magnet pair.
Because of my uncertainty I need limits from 2.4 lbf/magnet pair to 24 lbf/magnet pair.

But I will be disappointed if it is not within the range 4 lbf/magnet pair to 16 lbf/magnet pair.

 

[MagBen]

@logbook:
I am curious about the big force range, 2.4-24 lbs, i.e. max/min force ratio = 10. One uncertainty I can think of is the permeance coefficient(Pc), due to the reluctance of the backing steel, and the joined gaps between magnet and steel, Pc must be less than 8. While the min of Pc should be larger than that for an open circuit, which is 1.25. correspondingly B= 0.7T from BH curve of N42.
Assuming the force is proportional to B^2, given the B range of 0.7T to 1.2T, max/min force ratio = 2.9.
Does that mean you have other uncertainties?

 

[logbook]

MagBen, the uncertainty I have put down is x3 in either direction from a very dodgy starting point. I used the K & J magnet calculator to get the pull force between two magnets with steel backings. I have simply said the maximum transverse force will be somewhere near that value. I have made that statement based on using the same K & J calculator on my magnet setup and comparing the transverse force I measured with the normal force from the calculation. The whole process is dodgy, hence my large uncertainty.

If you would like to throw your hat into the ring and predict the peak static shear force I am all ears.

 

[MagBen]

It will be complicated to calculate the shear force without FEA/BEA. This force is a function of displacement, as you demonstrated, it will be zero when the pair of magnets are perfectly aligned.

 

[gmeast]

Hi all,

Well I joined everything together and I was pleased with the test results. I uploaded two pictures of the test setup. One shows the entire test fixture and the other is a (blurry) close-up shot of the gap. The over-all shot shows the moment arm at 12" (1 foot) ... it's labeled on the arm. 40 ft-lbs of torque ... yes ... that builds to 40 lbs at its highest force perpendicular to the arm at that 12" location.

gmeast

 

[gmeast]

OK ... I guess someone will need to show me how to attat more than one picture to my posts. so here's the 'gap' picture.

gmeast

 

[logbook]

Check my maths here …

You have 40lbf on a 12 inch lever.
Measuring and scaling from the photo it seems that the magnet gap is at about 2.8 inch radius from the axle.

I therefore get 40*(12/2.8) = 171 lbf maximum shear force for the 18 magnet pairs
= 9.5 lbf/magnet pair.

And I predicted 8lbf/magnet pair

I’m a genius

or just plain lucky

 

[MagBen]

logbook,
I would say you are a genius and in good luck as well.
Transverse (normal, or lateral) force = 0 when pull force is maximized (perfect alignement), the max transverse is less than the max full force, somewhere half of that value when the displacement cause a normal vector, per K & J, NOT near that max pull force.

The pull force between two steel backings (Case 2, still an open circuit) should much smaller that that for the close circuit. Actually the pull force between two magnets without any steel plates (case 3) is even higher (9.91 vs 8.71 lbs). Using 12KG as a good estimation of field at the gap in the setup, the pull force is calculated as F = .577 B[sup]2[/sup] x S = .577 x 12 x 12 x .5x.5 = 20.77 lbs. so the lateral force per pair = 10.4 lb.

Coincidently, you underestimated pull force by half, but then overestimated lateral force also by half. and then, you got the right answer. Anyway nobody can deny you made a perfect prediction!!

 

[MagBen]

I donot have a formula to calculate the transerve force, but the attached is an excerpt from K&J, the plot shows that max lateral force is about half of the pull force (no friction force here).

 

[gmeast]

Hi logbook,

You're a 'lucky genius'. The center of the shear line was designed to be on a 5 inch diameter circle ... which is what it is. The dimensions of the hub and ring (each) were backed off from that circle to create the gap and the recess locations based on the dimensions of the magnets. I'm glad you guys caught me on that 'recess thing' on the original design ... the field lines would probably have shorted out or at best produced lousy results. Thanks, and nice prediction and calcs.

I can now confidently scale for the real thing. Happy Holidays to all,

gmeast