Radially-oriented magnetization of Ring Magnets

[DAlbertson]

Engineers,

I'm building a set of small (12mm stroke, 20mm diameter) linear actuators for an after-hours toy. This is only my 2nd magnetics project, so I need to ask for your tolerance.

The DC actuators are based on Lorentz Forces with a small spring return. The coil is wrapped on a steel (1018) core rod, and does not move. The armature is a set of 5 rings, glued end-to-end, in this top-to-bottom order:
1. top steel pole piece,
2. top PM oriented axially North-down,
3. center steel pole piece,
4. bottom PM oriented axially North-up,
5. bottom steel pole piece.

The coil/core assembly slides into the center ID hole of the armature assembly.

The armature dimensions are such that the center steel pole piece is located over the coil winding (gap to steel rod about 4mm) , and the top and bottom steel pole pieces are over the bare end sections of the core rod (gap to steel rod about 0.5mm).

This works OK (at least, on the computer ) but it's kind of a poor solution because I lose about half the flux coming from the center pole piece: half goes "inward", passing through the coil and generating Lorentz Force (actually, Laplace force since it's a conductor / electrons interaction) but the other half goes "outward", doing me no good at all. In fact, it creates two problems:
1) bigger magnets and a heavier armature
2) magnetic interference with adjacently mounted actuators

I'd like to change the armature to a single PM ring magnet with radially-oriented magnetization, inserted into a steel tube, with top and bottom steel pole pieces to keep the 2 gaps to the core as small as possible. That would change the design from 4 magnetic circuits down to 2, and the steel tube would additionally shield (short) the flux from tending to leave the actuator and creating interference.

Now, I've got some questions if that's OK:

1) I can't find ring magnets with radially-oriented magnetization such that the ID is South, and the OD is North (or vice-versa). I suspect that it's a fabrication problem rather than a physical impossiblity; can anyone tell me?

2) I have found ring-section magnets that are like the magnets I described above, but cut radially into 60 degree or 45 degree sections. So, if I took 6 of the 60 degree sections, and slid them into a steel tube, will I end up with the equivalent of a radially-oriented ring magnet?

3) If there is a better way to acomplish this, please let me know!

Thank you all very much!

David Albertson

 

[RHMagster]

Hello DA...

It does sound as if the radial orientation, radially oriented magnets, would do you better than the first design described. The most common material used to produce these rings (as for my own experience goes) is Neodymium.

Dimension of the rings play a major role in whether they can be produced for your design requirements. Wall thickness is very important for pressing of the ring. This being a ratio compared to its lenght and over all diameter.

As you mentioned, a steel sleeve over the OD of the ring will direct and concentrate the flux return to the center making it a more efficient circuit.

I will look for more info on Radially oriented design concerns. I have this data at home from previous employment as Magnetic Design Engineer for large USA manufacturing plant.
My most recent work was in the design of systems for magnetizing these and other complex shapes of SmCo, Neo, Ferrite and Alnico materials and have fixturing to magnetize radial rings at home. I can put you in touch with some manufacturers who press radial rings as well.
I'll gather some info and post in the next few days.

 

[EdStainless]

It depends on the field strength that you are after. In order to get high strength materials you will be building rings from segments. I have done this in Alnico 8 and from various RE alloys.
When you make rings that have radial orientation, the orientation isn't real crisp. So you don't get full potential properties. But as you have found out, in many cases it is plenty good since the alternatives are difficult.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[RHMagster]

Hello again,

I have found, using Neo, rings of this nature have always performed as expected. Also found them to out perform segmented arcs in that you do not have the loss with solid radial rings as you do with the segments.
Segments allow for alternate return paths other than that of the radial ring. In many application they are cheaper to manufacture and assemble into end product. Of course, like all components, they have their place.

Here are a couple of sources for aquiring radial rings. Hope they can help.

http://www.vacuumschmelze.de/dynamic/en/

http://www.quad-tech.com/quadrantmagnetics.html

 

[MagMike]

Hello David,

You are on the right track, and you've been getting good advice from RHMagster & EdStainless. When I get a request for a radially oriented ring magnet, I usually recommend going with arc segment construction (8 pieces minimum if possible) since it is very difficult to source small quantities of solid rings.

One big advantage of making the ring out of segments is that one is allowed to use full strength SmCo or NdFeB magnets. Most solid radially oriented rings are only mid-range in strength.

The only thing I wanted to add is that there is no reason to orient the magnets (items 2 & 4 in your list) of opposite directions. With proper fixturing and adhesives, they can be assembled to get more flux (in the proper direction) out of your assembly. A radial ring would be best, but this might get you enough performance.

Best regards & good luck,
Mike

 

[DAlbertson]

Wow, thanks! This is exactly the kind of information I need!

If I may, let me direct comments and questions to each of you and then ask some general questions at the end.

RHMagster - Thanks for the advice and pointers to the two PM shops. I'm writing an inquiry to them both, and will send it off tomorrow.

1. As a sanity check, would you please give me your opinion on whether the following ring can be built: ID=13mm, OD=16.4mm, Axial Length=11mm. Is this buildable? I've not locked down the design yet, but it will be somewhere right around these numbers.
2. From MagMike's and EdStainless's comments I gather that the Energy Product is limited by the ring configuration. Typically, what is the best that can be done there?
3. Is there a desirable "ideal" geometry for a ring of roughly the dimensions I mentioned? I don't have a lot of free variables to play with, but if you could tell me some ratios (ID to length, or whatever) it would give me something to work towards.


EdStainless - I appreciate your comments on segmentation very much. I've got a couple questions:

1. On the magnetization of segmented arcs: in general, for segmented arc magnets, is the vector of magnetization simply a constant, like in a bar magnet, or does it's direction change synchronously with the arc of the physical magnet material? OK, ugly sentence. If you looked down the assembled collection of segments axially, would you "see" a circle of magnetic flux or a n-polygon of flux, where n is the number of segments. Still an ugly sentence, but the best I can do. It obviously matters to the model, but if n is large (per MagMike: n>8), and given fringing effects, the segmented approximation of a circle may work fine. I can certainly see where it would be cheaper!
2. What about getting even cheaper -)) and approximating the circle with n=16 rectangular bar magnets, instead of the arc segment magnets?

MagMike - Thank you for the advice on segmented arcs.

1. Using my ID=13mm estimate, if I went to n=16 segments I would end up with each segment having an interior arc length of 2.55mm, which is still large enough that I could deal with it without special handling. Is this a reasonable thing to ask the PM shops for, or is it silly?

2. Thank you for reading my description of my existing axial-orientation actuator, and I appreciate your pointing out the discrepancy with the inverted polarity between magnets #2 and #4. However, I think the problem is only that I was not clear in my description. The center, #3 pole piece is the only one focusing flux on the coil itself. The #1 and #5 pole pieces are adjacent to the core rod, on either side of the coil winding. By having the #2 Upper PM's North face pointed down into the #3 pole piece, and the #4 Lower PM's North face pointed up into the same #3 pole piece, then all the flux headed into the #3 pole piece goes at right angles, either into the coil winding (good) or free space away from the actuator (bad). I think this is correct, but I have been wrong 32 times since breakfast and this would simply make it 33!

If I haven't run out of good will yet, I have a couple more questions that I'd like to ask of the group in general:

1. I've been using 416 stainless in my models as an approximation of 410, because I cannot find B-H data or Hc for 410. I've checked AZOM, the S/S institute, and a couple of big stainless suppliers. Do any of you have a pointer for me on where I could get it?

2. My understanding is that NdFeB is running about $60/lb, licensed. Now, even if I built a bunch of these, and used them as Christmas presents for the next 5 years, I'm still not going to need more than a couple pounds of NdFeB. So, can you tell me what to expect for dollar adders on these sorts of prototype quantities (100 units)? $300 in setup fees is OK. 5 grand to machine a mold in A2 is not. That sort of thing. Just so I know roughly what to expect when the vendors start talking money.

Thank you all very much! I appreciate your expertise and patience very much.

Best Regards,
David

 

[RHMagster]

Hello All,

The dims you propose would be a size suitable for production. the wall thk. is of most concern when pressing this part. Too thk/Too thin presents problems during sintering. This due to shrinkage cracking and warping.

Cost - for high volumes these can prove to be less costly. This due to assembly cost of the segmented arcs. The more the arcs per unit, the more cost associated to assembly.

Low volume - you may find a supplier who has production of current rings. Close to the size you require, adjust your design to accomodate (if possible) and reap the cheaper cost otherwise they will be expensive.

One such company is Hitachi. Try giving them a request for quote and leave the size somewhat open. See if they can get close to your need with a current production item produced. They may be able to give you close to what you seek with out tooling charges etc. Check China suppliers as well. Quadrant Magnetics brokers China manufacturing, tooling there (China) is, on the most part, cheaper.

I recently spoke to VAC, asked if they produce rings currently. They may have some proto-types close to the size you require. I asked them to look into qty on hand and size of. Will let you know what I learn from them.

In my experience in making / energizing / testing of magnet mat'l I've seen properties in the 90% range. This is made possible by good pressing control, mainly alignment fields and pressing methods. Testing of these mat'l's, you must account for the radial alignment of the part. Special test equipment is required to find true properties of the finished product. Std Helmholtz or other test methods may not supply you with these true properties.
Saturation of the product falls to this as well. A true Radial alignment must be saturated with a true radial field thus a true radial field must be measured with fixturing capable of reading radial aligned product.

I applaud MagMike's view of circuit design and believe, based on qty required, that the circuit should be studied and could be made to focus more flux into the desired area.
Sometimes a design review can save big $ and provide better product.

OK, DA... I will let you know more as I learn more. Hope this helps.

This is the best part of ENG-Tips forum and have been a big fan. Read many a good tips from MAgMike and EdStainless. over the years. Keep up the good work.

 

[EdStainless]

Even though the blocks in a segmented ring have straight line orientation the field ends up being very close to uniform. Remember the bolcks are wedge shaped and the inside is curved. We used 7 or 8 or 9 segments for rings up to about 1.25" OD. Above that there are easier ways to do it and uniformity isn't as big of an issue.

In small quantity you can build lots of different configurations with segmented rings.
I built some that were actually 5 layers, stainless, radial magnet N in, axial magnet, radial magnet S in, stainless.

There are lots of possabilities.


= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[gphatch]

Daido Steel makes some very nice radially oriented Neo magnets.

Gareth P. Hatch, Ph.D.
Director of Technology
Dexter Magnetic Technologies

http://www.dextermag.com