Do we have any magnet/magnetism gurus here?

[cvanc]

I have a couple questions that I'd like help with, is there anyone here who can talk fluent magnet?  Thanks.

[RandallMcRee]

shoot your question out, don't be shy...

If it has anything to do with perpetual motion, however, don't expect a reply.

[Zero999]

I don't know about gurus, but as far as I'm aware, most higher level electrical/electronic courses contain a section on magnetics, so a lot of people here who do engineering professionally, will have more than a basic clue about magnetics and there will be some experts too.

[ArthurDent]

shoot your question out, don't be shy...

If it has anything to do with perpetual motion, however, don't expect a reply.

I like the rhyming.   

[cvanc]

All right, in we go 

I want to build a large flat "sheet" magnet.  A mathematical plane.  It needs to be strong, have a smooth even field, and be North on the front side and South on the back side.

Think maybe 50 square feet.  Can it be accomplished at this size?  Without breaking the bank?

Thanks.

[duak]

I would start with two steel sheets separated by permanent magnets.  The steel sheets form the structure and smooth out the field.  The strength and uniformity of the field will be a function of the type and number of magnets, their spacing and arrangement, and the thickness of the sheet steel.  The array could be broken into tiles to make it easier to construct and move.

[T3sl4co1l]

At what field strength?  At the surface, or what volume?

It's not a very reasonable thing.  Because field has to go somewhere, you can't simply have a region of uniform field, and ignore what's outside it.  Normally you'd have pole pieces outside the area for instance -- look at the design of cyclotrons in the 50s and 60s.

Is there an X-Y problem here -- what's your application?

Tim

[RandallMcRee]

What Tim said.

W/out breaking the bank? Its a big ask unless you are talking about the Earth's magnetic field...

Technology does exist to create a sort of planar magnet, it is called a halbach array. https://en.wikipedia.org/wiki/Halbach_array

There are some neat applications for that (google produces a lot of interesting ones).

Either as duak says, above, or perhaps two huge differently polarized halbach magnets, see
https://engineerdog.com/2015/03/12/why-do-refrigerator-magnets-only-stick-on-one-side/
you should be able to print one with a north field out and another with south out and put the non-magnetic planes adjacent. Done.

[strawberry]

how a magnetic sphere with poles on inner and outer face can work?

[Zero999]

Are you talking about somethign like this?

[ejeffrey]

how a magnetic sphere with poles on inner and outer face can work?

It can't.  The field lines have to get from the north to the south pole.  That's also the problem  T3sl4co1l was explaining about the flat plane.  The field lines in the center have to go *much* farther to get back to the other side than the ones at the edge.  It isn't mathematically impossible like the sphere question but it is get very hard very quickly as your requirements on area, field strength, and uniformity go up.

Its much more practical to create a uniform field in a gap between pole pieces than on the surface of a plate.

[RandallMcRee]

Are you talking about somethign like this?
(Attachment Link)

Yeah, thanks for illustrating. So reviewing the halbach orientation it seems that there is not predominantly north or south, so not red and blue. Instead we have a linear orientation. So, we could, for example have one side oriented in one direction and the flip side oriented 90 degrees to it.

Perhaps that would be useful to the OP. Hard to say.

[cvanc]

Are you talking about somethign like this?
(Attachment Link)

Yes.  Imagine this the size of a sheet of plywood.  Thickness is undefined.

Can a Halbach array be made such that the entire face of the plane is North?  The asymmetric pattern would be helpful as I only care about the field on one side.

Thanks everyone.

[ejeffrey]

Are you talking about somethign like this?
(Attachment Link)

Yes.  Imagine this the size of a sheet of plywood.  Thickness is undefined.

Can a Halbach array be made such that the entire face of the plane is North?  The asymmetric pattern would be helpful as I only care about the field on one side.

Thanks everyone.

No. The whole point of the halbach array is that the alternating polarity allow closing flux loops to stay local.  It is basically a bunch of horseshoe magnets stacked in a row.

If you make a plate with all north on one side the field has to return to the back.  A uniformly magnetized plate will always fall to half at the edge and 1/4 at the corner so you will need nonuniform magnetization to get an approximately uniform field.

[IanB]

I want to build a large flat "sheet" magnet.  A mathematical plane.  It needs to be strong, have a smooth even field, and be North on the front side and South on the back side.

Take any small magnet and measure the field at one of its poles. You will find the field is strongest at the center. For example, anything you stick to the edge of the magnet will tend to get pulled to the center. So even a small, ordinary magnet doesn't have an even field over its very small area of flat surface. Now imagine scaling up the small magnet to ever larger sizes. As you scale it up the uneven field will scale with it.

From an engineering perspective, what you are asking would seem to be impossible, or at least, hugely infeasible.

[Marco]

This seems to be about ribbon speakers ... but ribbon speakers don't have a planar field coming from below the ribbon, it wouldn't work. Ribbon speakers need the magnetic field lines to be parallel to the ribbon, Lorentz force and right hand rule and all.

I know the Apogee ribbon speaker has a massive "ribbon" which makes it impossible to achieve that with magnets just on the side of the ribbon (like ribbon tweeters have) but they can't cheat physics ... if the force has to move the ribbon front to back, then the magnetic field lines have to come from the side (on average).

What you could do is have the ribbon metallized in sections, and run the current up and down on alternating sections. Then have alternating strips of N/S magnets below, along the insulated parts of the ribbon. It won't be an even field, but on average the magnetic field will have a component running parallel to the ribbon and at a right angle to the current and thus create a force front to back. The forces on all the sections will point the same way too.

[T3sl4co1l]

OP refuses to answer about their application, so I guess we should assume they don't know what it is; and therefore the answer is trivial (it doesn't matter).  Or if they are unable to tell, but the same conclusion applies, at least as far as public discussion is concerned.

Tim

[CatalinaWOW]

This breaks the bank for all but the mega rich, but you could cheat a bit given your ill defined requirements.  Build a large C shape magnet with dimensions of the pole pieces on the scale you describe and spacing between them as you need.  Then a flat sheet of virtually any soft magnetic material placed between the poles would assume the configuration you are looking for.   

This approach does illustrate in a different way the difficulty of achieving what you want.

[Marco]

Well he was talking about ribbon speakers in another thread, so it's a safe guess that's what this was for. Ribbon tweeters have the magnets as in pic, but that doesn't work with a full range speaker ... the magnets would have to be way too big.

Apogee segments the ribbon into zigzag pieces. So I assume they work just as I suggested, with the bar magnets below the ribbon not all being the same polarity, but alternating such that they create field lines which run somewhat parallel with the ribbon. Alternating in polarity as the direction of the current alternates so the force stays pointing the same way. The field will also have large components perpendicular to the ribbon, but those are mostly irrelevant, they will create forces in plane with the ribbon.

The magnet cvanc wanted to create would create forces only in plane with the ribbon ... that's not going to be a very good speaker.

[ejeffrey]

Yeah, that should work and would be a good application of a halbach array