[arobertson1 (OP)]

I don't understand why everybody is trying to make a donut shaped reactor? Why spin hot plasma round in a circle and pray that it doesn't become unstable and hit the sides? Wouldn't it be simpler to have the plasma in the middle like a sphere (kind of like the sun!) and hold it in place by repulsion from another enclosing outer shell. You could inject it from all round the enclosing shell with fuel - kind of like a colander. The initial trigger could be done with lasers. It seems like a lot of over engineering to me...

[MikeS]

Can't say as I really know the answer but is probably to do with the problems with building magnets to operate around a sphere.  A pipe (with the ends connected) is much easier. 

[Geezer]

This one is spherical

http://en.wikipedia.org/wiki/National_Ignition_Facility

[JP]

Not everyone is making a donut-shaped reactor.  There's an entire field of fusion research on what's called inertial confinement fusion, where a pellet of fuseable material is held in place and hit with high-powered laser pulses.  

I think what you're talking about is magnetic confinement fusion.  I had to go to the wiki to look up why donuts (known as toroids in geometry) are used.  Essentially, they use magnetic fields to confine a moving plasma.  An efficient way to do this is to build a solenoid, which is just a wire coiled up like a spring, with current flowing through it.  A plasma inside the coil won't be able to escape out the sides, but might escape out the ends.  A way to prevent this is to bend the coil so it's ends connect, forming a donut, which is why that's commonly used.

In other words, the magnetic field produced by that donut shape is the right kind to keep a plasma trapped within the body of the donut.

http://en.wikipedia.org/wiki/Magnetic_confinement_fusion#Compact_toroids

[Geezer]

You mean it's not because they were designed by this World famous nukeular engineer?

[MikeS]

Not everyone is making a donut-shaped reactor.  There's an entire field of fusion research on what's called inertial confinement fusion, where a pellet of fuseable material is held in place and hit with high-powered laser pulses.  

I think what you're talking about is magnetic confinement fusion.  I had to go to the wiki to look up why donuts (known as toroids in geometry) are used.  Essentially, they use magnetic fields to confine a moving plasma.  An efficient way to do this is to build a solenoid, which is just a wire coiled up like a spring, with current flowing through it.  A plasma inside the coil won't be able to escape out the sides, but might escape out the ends.  A way to prevent this is to bend the coil so it's ends connect, forming a donut, which is why that's commonly used.

In other words, the magnetic field produced by that donut shape is the right kind to keep a plasma trapped within the body of the donut.

http://en.wikipedia.org/wiki/Magnetic_confinement_fusion#Compact_toroids

arobertson1
Was obviously talking about plasma reactors cos that's what he said.  In which case my answer was correct.

[CliffordK]

I haven't read the design specs, but there are potentially a few advantages of a toroid, or donut shaped apparatus.

It would have been modeled to some extent from the super colliders.  One way to add energy to particles is to accelerate them using magnetic coils.  And it is likely easier to manage the input and output of a particle stream than a hot ball of plasma.  Can one make a magnetic centrifuge to separate out the fusion products?

One of the major problems with a sphere is that a magnetic monopole is inherently unstable.  Perhaps a particle cloud could act as a second pole inside the sphere, but a uniform magnetic field inside of a sphere would risk cancelling itself out.

[imatfaal]

One of the major problems with a sphere is that a magnetic monopole is inherently unstable.

I thought magnetic monopoles were only hypothetical at present

[CliffordK]

One of the major problems with a sphere is that a magnetic monopole is inherently unstable.

I thought magnetic monopoles were only hypothetical at present

Precisely. 

I'm not quite sure about magnets and ions.  One note suggested that positive ions migrate towards the magnetic North Pole, and thus would be repelled by the magnetic South pole.

So, if you wanted to keep the positive ions in the center of the sphere, you would point the south pole of every magnet inwards.  And, thus effectively have a monopole.  In the absence of other forces, the magnets should cancel themselves out.

If you have particles orbiting a toroid.  Centrifugal force would tend to keep the particles from colliding with the center wall.  And, thus you would add a magnet that was repulsive on the outer wall, and attractive along the inner wall.  And, you would avoid the monopole problem.

I don't know if there are limits on the magnets on a sphere problem.

For example, if you had a sphere that was 1 mile in diameter, the local polar interactions might be more important than non-local interactions.

[damocles]

The answer is probably something to do with combing hairy donuts and spherical dogs.

In a fusion reactor the material undergoing fusion must be contained.  There is nothing by way of a normal container that could work -- any material known vapourises at much lower temperatures. But a plasma of ions can be contained by magnetic fields.

In a magnetic field, a moving charged particle can be made to follow a circular path. With a sphere, it is impossible to comb it without getting a crown or a spike -- circulating charged particles will necessarily bump into each other somewhere. But in a donut you can get completely orderly circulation without clashes -- either in small circles passing through the hole, or large circles passing around the hole, or diagonal spirals that wind in and out through the hole.

A donut shape was chosen for the original attempts at building tokamak prototype reactors because that is the simplest conceivable shape for containment of a plasma in a wall-less situation.

[imatfaal]

Damocles - I like that suggestion.  You can "comb" a torus such that over the whole surface there is a continuous tangential vector field.

[AlmostHuman]

I'm not quite sure about magnets and ions.  One note suggested that positive ions migrate towards the magnetic North Pole, and thus would be repelled by the magnetic South pole.

This depends on their trajectory. Lorentz force is defined by B and v vectors (Magnetic field B and charged particle velocity v)...
Or something else had proven this wrong.

Magnetic monopoles are way sci-fi-ish right now, but I hope someone invents/discovers it in near future.

[imatfaal]

Magnetic monopoles are way sci-fi-ish right now, but I hope someone invents/discovers it in near future.

AlmostH - they are a bit more than that.  they are closely tied into the theory of quantised electromagnetic fields.  dirac predicted that the presence of the magnetic monopole would require a quantised electrical charge - well we do have quantised charge, now that is by no means proof (ie the logic goes in the "wrong direction" )- but it is suggestive.  http://en.wikipedia.org/wiki/Magnetic_monopole

[CliffordK]

I was thinking about this a bit today.

A magnet doesn't attract or repel a ion. 

Moving ions create magnetic fields that can be manipulated with magnets.  So, rather than compressing a ball, they use the toroid to manipulate a moving stream of ions.

[Lab Rat]

In a fusion reactor the material undergoing fusion must be contained.  There is nothing by way of a normal container that could work -- any material known vapourises at much lower temperatures. But a plasma of ions can be contained by magnetic fields.

Yes, adding on to what damocles was saying, NASA's best ceramics would vaporize instantly on contact with plasma.  Plasma is 40,000,000K or 39,999,727ºC (71,999,540.6ºF), which is a temperature much hotter than the sun (about 7,000 times more, to be exact (depending on where you look)). See below for conversions:

5778K-273=5505ºC
5505ºC x 1.8=9909+32=9941ºF