magnet design 101 - saturate iron with a ferrite magnet?

[capslock]

We've had various discussions here on the advantages of AlNiCo, Neodymium and field coils. One thing that crystallized here was that it is advantageous to saturate the iron return system because that makes a modulation of the iron and magnet magnetization impossible, and it also keeps the pole piece from acting like a core that increases the inductance of the voice coil.

One point that I found in the literature was that ferrite magnets had too low a flux density to saturate iron (1.0-1.2 vs. 1.3-1.5 T). Nobody has challenged this view.

While this statement is certainly true if the magnet and the iron share the same cross section, I wonder if it is possible to saturate a part of the return system by letting it have a smaller cross section than the rest of the return system and the magnet. Where the iron becomes thinner, the field lines move together und the flux density increases. If the iron becomes thin enough, it will sature. Any "extra" field lines will have to go through the air until the system get back to its original thickness.

The gap should be placed in the thin area so that all iron that has contact with the field of the VC is saturated. Also, this will keep the stray field spread out, so that we won't have a strong assymetric stray field around the gap. The thin area should, however, not extend too far because this will keep the iron barely in saturation, making it easier for the VC's field to drive the iron out of saturation.

Any flaw in my reasoning? How come it's not done?

[phase_accurate]

I think it is already done more often than not !
If you look at the cross section of the magnet and the cross section of the air gap the former is usually significantly larger than the latter one for the average driver with ferrite magnet.

I think the WHOLE iron part of the magnetic loop must be saturated to take any advantage of saturation.

Regards

Charles

[capslock]

Hi Charles,

theoretically, it should be sufficient to saturate anything that interacts with field of the VC. So the upper pole plate (both sides of the gap) should be saturated, and also the upper few cm of the pole piece.

In a standard system, if the inner side of the gap is just saturated, the outer side is probably not. Maybe a pole plate that becomes thinner as the radius increases is the solution?

Greetings,

Eric

[phase_accurate]

Maybe a pole plate that becomes thinner as the radius increases is the solution?

Quite a clever idea !

Regards

Charles

[Bill F.]

Hi Guys,

Yes, it certainly is possible to achieve very high gap flux density with ceramic mags. Just to see what was possible, I recently created a finite element magnetic simulation of a ceramic-magnet motor with a steel and permandure return circuit. I achieved 2 tesla across the gap! But it took a *big* grade-8 magnet. I'll try to post a pic of the simulation after lunch.

theoretically, it should be sufficient to saturate anything that interacts with field of the VC. So the upper pole plate (both sides of the gap) should be saturated, and also the upper few cm of the pole piece.

Yes, I believe this is true.

In a standard system, if the inner side of the gap is just saturated, the outer side is probably not.

Unless you're thinking of a tapered T-shaped pole piece, the reverse is often true. Yes, the diameter of the top plate is larger than that of the pole piece, but the top plate is thin and projects most of the system flux from its inner perimeter. Therefore, while its inner perimeter may be saturated, the flux field expands vertically as it crosses the gap, often creating a trapezoidial field and not necessarily saturating the surface of the pole piece.

Here's a sim I did of my best guess at Adire Audio's (still secret!) XBL2 topology. Neither the pole nor top plate are saturated in this example, but you can see what I mean about the way flux spreads out between the top plate and pole. Incidentally, here's here's the thread where I expose this secret motor, if you're interested!

I think the WHOLE iron part of the magnetic loop must be saturated to take any advantage of saturation.

This would be ideal if it was practically achievable, but the trouble is that as iron approaches saturation, flux lines begin to project out from the surfaces and escape. To achieve full-circuit saturation, I think you'd need to spec a ridiculously large NdFeB magnet that would waste so much flux that it would distort a TV from across a room. I believe a more practical ideal is saturation of both sides of the gap.

Bill

[capslock]

Originally posted by Bill F.
Unless you're thinking of a tapered T-shaped pole piece, the reverse is often true

Hi Bill,

at last somebody who has played with these things! Yes, I was thinking about T-shaped pole piece. What do you mean by "tapered", i.e. what would be growing successively thinner, the vertical column of the T or the top?

What software do you use for FE modelling? I am a little surprised about the calibration. Your ferrite magnet seems to have a pretty uniform flux of below 0.36 T. I have not studied magnet catalogues, but I am told that regular ferrite can be magnetized at about 1 T, maybe even a little higher if magentization is done with a return system installed. What is the grade 8 vs. grade 5 stuff you are talking about in your linked thread?

Similarly, the iron seems to reach almost 2.0 T where I thought the limit was somewhere around 1.5 T. Is this a special alloy? If so, does it get used in common drivers?

Regards,

Eric

[Bill F.]

Hi Eric,

By tapered, I mean that the ears of the T taper down with increasing diameter toward the gap. If they are not tapered, the flux-conducting area is greater at the gap than where the T ears join the cylinder of the pole piece, creating a flux bottleneck that prevents saturation at the gap.

The program I use is FEMM. I haven't taken the time to check the values FEMM specs for ceramic magnets, I just plugged 'em in. My understanding is that there are two broad categories of industrial ceramic magnetic material--grade 5 and 8. The material properties in FEMM are easily modified if I need to.

The steel alloy I specd for the XBL2 motor sim is 1006, a very low-carbon steel with good magnetic properties. I believe P-Audio uses it in some models, and probably other manufacturers as well.

The simulation wasn't supposed to accurately reflect the flux numbers in an actual XBL2 driver, just the shape of the BL curve that results from the topology.

Driver design is my favorite hobby right now, in case you couldn't tell!

Bill

[capslock]

Originally posted by Bill F.
Hi Eric,

By tapered, I mean that the ears of the T taper down with increasing diameter toward the gap. If they are not tapered, the flux-conducting area is greater at the gap than where the T ears join the cylinder of the pole piece, creating a flux bottleneck that prevents saturation at the gap.

Ok, I am getting the point. Perimeter increases with r, so thickness has to decrease with r to kepp the cross section constant. So my idea isn't new

Of course, one could continue the tapering with the top plate, or at least the inner half inch of the top plate. Has it been done, with either the pole piece or pole piece + top plate?


So if acutal magnets are 3x stronger than you assumed in your calculation, you could easily saturate large parts of your geometry...

I gather FEMM is not exactly freeware?

Greetings,

Eric

[RobM]

Originally posted by capslock
I gather FEMM is not exactly freeware?

In fact, it is!

http://femm.berlios.de/

[Bill F.]

Of course, one could continue the tapering with the top plate, or at least the inner half inch of the top plate. Has it been done, with either the pole piece or pole piece + top plate?

Of course, the effective cross-sectional area of a flat top plate already increases with r, but this is still a good way to ensure that your return circuit nears saturation only at the gap, thereby allowing it to efficiently conduct the flux.

I gather FEMM is not exactly freeware?

I see Rob gave you the good news while I was writing. Yes, FEMM is free and the best thing since sliced bread! Check out the FEMM forum on Yahoo to come up to speed. If you check the names on many of the posts, you'll notice that many audio industry who's-whos use this jewel of a program. Babb, Adire, Aura, etc.