Experiments with solid-state magnets

[JoeStrout]

According to Bussard 1991, magnetic field strengths of 0.1 to 0.5 T are all that's needed for a small polywell reactor. That's well within the range achievable by permanent magnets. Permanent magnets offer a number of advantages over electromagnets:

• no ohmic heating issues; easy to run in steady-state rather than pulsed mode
• no flexing of the system as you turn the current on and off
• no huge power supplies needed to drive the electromagnets

I realize that you probably can't scale permanent magnets up to the size needed for a break-even reactor, but I do think you could do a lot of useful work with them, far more easily than with electromagnets. Of course magnets that strong would have to be handled with great care when assembling the machine... but once assembled, the only power needed would be for the electron and ion guns.

Thoughts?

[pstudier]

One could do an experiment similar to the Levitated Dipole, see http://en.wikipedia.org/wiki/Levitated_Dipole . However, it would have a loss cone similar to a mirror machine because the magnetic field lines go through the solid magnet. The earth is a dipole and has a plasma around it called the Van Allen radiation belts, see http://en.wikipedia.org/wiki/Van_Allen_radiation_belt . Sorry if this doubles posts, the last post disappeared.

[JoeStrout]

One could do an experiment similar to the Levitated Dipole, see http://en.wikipedia.org/wiki/Levitated_Dipole . However, it would have a loss cone similar to a mirror machine because the magnetic field lines go through the solid magnet.

Why would the field lines go through a permanent magnet any more than they go through a same-shaped electromagnet?

Granted, toroidal permanent magnets aren't common, but they are possible (and even commercially available).

[pstudier]

Magnetic materials are composed of small dipole magnet regions called domains. Each domain is a small dipole. If one lines up these domains in a chunk to point in the same direction, then this chunk becomes a dipole magnet. There is no way to add a bunch of dipoles to make a magnetic field that does not go through the material. If you don't believe me, buy some magnets, and some iron filings, and a piece of paper, and let us know how it goes.

I have seen ring magnets, but if you lie one flat on a table, then the top will be one polarity and the bottom will be the opposite polarity, with the lines going through the material. Flat refrigerator magnets often have more complicated fields, but the lines always go through the material.

[MSimon]

One could do an experiment similar to the Levitated Dipole, see http://en.wikipedia.org/wiki/Levitated_Dipole . However, it would have a loss cone similar to a mirror machine because the magnetic field lines go through the solid magnet. The earth is a dipole and has a plasma around it called the Van Allen radiation belts, see http://en.wikipedia.org/wiki/Van_Allen_radiation_belt . Sorry if this doubles posts, the last post disappeared.

Exactly!

It is also why adding flux enhancers (ferro magnetic material) to the coils will not help. The flux lines wind up in all the wrong places.

Only currents develop the right kind of field. And they are not perfect because the wires are discrete. It kind of spreads the field near the coils where you would prefer the most uniformity.

However, with strong enough fields they are good enough.

The best is a single round wire. I can't wait to do superconductors.

[MSimon]

Joe,

Magnetic fields go around coils. They go through magnetic material.

Look up the field around a single wire vs the field around a magnet.

[JoeStrout]

I have seen ring magnets, but if you lie one flat on a table, then the top will be one polarity and the bottom will be the opposite polarity, with the lines going through the material.

I understand, but I think this may be an artifact of manufacturing convenience (these magnets are made by placing the heated material in a very strong magnetic field, which from the magnet's perspective is aligned straight up and down). I don't think it's a physical necessity.

As a current is passed through the coil, small magnetic regions within the material, called magnetic domains, align with the applied field, causing the magnetic field strength to increase.

I see no theoretical reason why a permanent magnet couldn't have its magnetic domains lined up in the same way as what the current causes. Indeed, in remanent magnetism, exactly that happens — some of the domains remain lined up even after the current is shut off. Seems to me that these would be producing a field of the same shape as when the current is on (albeit weaker).

So if this really is a manufacturing problem, there may be ways around it. I'll dig deeper and see what I can find out. Thanks to all for the feedback!

[JoeStrout]

OK, I think I've got this figured out.

In a typical cylindrical permanent magnet, with N at one end and S at the other, the dipoles are all in line with the axis and the equivalent electrical current would be one that goes around the outside of the cylinder, the round way.

But we want a magnet whose field matches that of a current travelling down the axis. In that case, the dipoles are all lined up with the circumference of the cylinder; drawn as arrows, they'd be chasing each other around the axis. It's hard to define a "north" or "south" pole for such a magnet as a whole, but at any spot on the surface, north points (say) clockwise around the circumference, and south points counter-clockwise.

I can see how this would be hard to manufacture in a single piece. But there's no reason why it needs to be made in a single piece; some really interesting and useful magnet geometries are made by combining simpler pieces (like a Halbach array for example).

In this case, you could approximate the polywell coil fields by combining lots of thin wedge-shaped permanent magnets. Each would be magnetized in the standard way, N on one side and S on the other. Stack these up so that they form a cylinder, and now you've got a combined field just like a section of coil, which goes around the cylinder without intersecting it.

Of course that's just a cylindrical section; you'd need to combine more of these in the other direction to make a full torus. But in a way, that's good; many small identical magnets will be a lot cheaper than a few enormous ones.

But certainly it's no longer a matter of going out and buying some ring magnets... to do a good job, you'd probably need to have the wedges custom-made, as they're an unusual shape. Still might be worth it, though, for all the problems it avoids.

[Nanos]

Would you need the magnets to be wedges though, what if you used standard shape ones and have fillers inbetween which was cheaper to produce ?

I'm also keen on the permant magnet idea myself, and wonder about if one might look towards a screw effect using magnetic force.

[MSimon]

OK, I think I've got this figured out.

In a typical cylindrical permanent magnet, with N at one end and S at the other, the dipoles are all in line with the axis and the equivalent electrical current would be one that goes around the outside of the cylinder, the round way.

But we want a magnet whose field matches that of a current travelling down the axis. In that case, the dipoles are all lined up with the circumference of the cylinder; drawn as arrows, they'd be chasing each other around the axis. It's hard to define a "north" or "south" pole for such a magnet as a whole, but at any spot on the surface, north points (say) clockwise around the circumference, and south points counter-clockwise.

I can see how this would be hard to manufacture in a single piece. But there's no reason why it needs to be made in a single piece; some really interesting and useful magnet geometries are made by combining simpler pieces (like a Halbach array for example).

In this case, you could approximate the polywell coil fields by combining lots of thin wedge-shaped permanent magnets. Each would be magnetized in the standard way, N on one side and S on the other. Stack these up so that they form a cylinder, and now you've got a combined field just like a section of coil, which goes around the cylinder without intersecting it.

Of course that's just a cylindrical section; you'd need to combine more of these in the other direction to make a full torus. But in a way, that's good; many small identical magnets will be a lot cheaper than a few enormous ones.

But certainly it's no longer a matter of going out and buying some ring magnets... to do a good job, you'd probably need to have the wedges custom-made, as they're an unusual shape. Still might be worth it, though, for all the problems it avoids.

You are almost there.

In ferromagnets the field must go through the material or there is no field. There is no way around it. The aligned matter creates the field.

The essence is that we want the fields to go around the grid. Coils of wire are the only way.

[JoeStrout]

In ferromagnets the field must go through the material or there is no field. There is no way around it. The aligned matter creates the field.

That's true of electromagnets too; the currents simply align the domains in the material, which creates the field.

But the key is that fields are additive; all the magnetic domains add up their fields to generate the overall field. This applies just as well when you use lots of small magnets put together; their fields add up. In the configuration I described, the field lines (of the combined field) would indeed coil around the combined cylinder (or torus), not go through it. Yes, there are field lines inside the material as well, but they don't go outside the material, and the ones outside it don't go inside.

The essence is that we want the fields to go around the grid. Coils of wire are the only way.

Any field that can be generated by an electromagnet, can (at least in principle) be generated by a permanent one, and vice versa. The magnetic domains don't care whether they're being lined up by a current, or because they're frozen into position.

I'm going away for a couple of days, but next week I'll look into generating some simulations, which may be more convincing than words.

[pstudier]

It is not possible to have a magnetic field line that does not enclose current. It is no more possible to have field lines from a permanent magnet that do not go through it than it is to have a closed field line from an electromagnet that does not enclose one of the wires.

It follows from Maxwell's equation. If one integrates along a magnetic field line, then this must equal the current flowing through any surface that has the field line as its boundary. One can always find a surface that does not enclose the permanent magnet or the coil of the electromagnet, unless they are interlocking. That is, unless the loop of current interlocks the magnetic field line.

[MSimon]

In ferromagnets the field must go through the material or there is no field. There is no way around it. The aligned matter creates the field.

That's true of electromagnets too; the currents simply align the domains in the material, which creates the field.

But the key is that fields are additive; all the magnetic domains add up their fields to generate the overall field. This applies just as well when you use lots of small magnets put together; their fields add up. In the configuration I described, the field lines (of the combined field) would indeed coil around the combined cylinder (or torus), not go through it. Yes, there are field lines inside the material as well, but they don't go outside the material, and the ones outside it don't go inside.

The essence is that we want the fields to go around the grid. Coils of wire are the only way.

Any field that can be generated by an electromagnet, can (at least in principle) be generated by a permanent one, and vice versa. The magnetic domains don't care whether they're being lined up by a current, or because they're frozen into position.

I'm going away for a couple of days, but next week I'll look into generating some simulations, which may be more convincing than words.

Nope,

With coils of wire there is only the alignment of space. Just as the field OUTSIDE the magnet aligns space. Now this is not strictly true for real coils. There will be SOME intersection of the fields with the coils. However, that is not a primary effect.

[cuddihy]

In ferromagnets the field must go through the material or there is no field. There is no way around it. The aligned matter creates the field.

That's true of electromagnets too; the currents simply align the domains in the material, which creates the field.

But the key is that fields are additive; all the magnetic domains add up their fields to generate the overall field. This applies just as well when you use lots of small magnets put together; their fields add up. In the configuration I described, the field lines (of the combined field) would indeed coil around the combined cylinder (or torus), not go through it. Yes, there are field lines inside the material as well, but they don't go outside the material, and the ones outside it don't go inside.

The essence is that we want the fields to go around the grid. Coils of wire are the only way.

Any field that can be generated by an electromagnet, can (at least in principle) be generated by a permanent one, and vice versa. The magnetic domains don't care whether they're being lined up by a current, or because they're frozen into position.

I'm going away for a couple of days, but next week I'll look into generating some simulations, which may be more convincing than words.

It seems no one thinks this is possible -- Joe, why not just put up a CAD drawing of what such a magnet piece would look like.

[MSimon]

In ferromagnets the field must go through the material or there is no field. There is no way around it. The aligned matter creates the field.

That's true of electromagnets too; the currents simply align the domains in the material, which creates the field.

But the key is that fields are additive; all the magnetic domains add up their fields to generate the overall field. This applies just as well when you use lots of small magnets put together; their fields add up. In the configuration I described, the field lines (of the combined field) would indeed coil around the combined cylinder (or torus), not go through it. Yes, there are field lines inside the material as well, but they don't go outside the material, and the ones outside it don't go inside.

The essence is that we want the fields to go around the grid. Coils of wire are the only way.

Any field that can be generated by an electromagnet, can (at least in principle) be generated by a permanent one, and vice versa. The magnetic domains don't care whether they're being lined up by a current, or because they're frozen into position.

I'm going away for a couple of days, but next week I'll look into generating some simulations, which may be more convincing than words.

It seems no one thinks this is possible -- Joe, why not just put up a CAD drawing of what such a magnet piece would look like.

It is actually worse than that. Dr. B tried the experiment and got burns on the magnets just where theory predicts.