A torus mag confinement?

[rogerlam]

I have been thinking if it is possible to achieve mag confinement with a torus configuration(yes, the ugly torus!).

Since polywell mag field can't completely stop electron from going out from the gaps. Maybe we can have a seamless mag field with torus:

Imagine a conductive tube, if you pass current along the tube a circular mag field will be produced inside. If you shoot ebeam through the center of the tube they will experience either an outward dispersive lorenz force or an inward contractive lorenz force depends on which direction you fire them. To confine electron of cause we make them contract into a tight beam this way. Connect the two end of the tube and you got a torus with ebeam going in circle in the center.

The advantage is one don't have to consider electron loss through the gap in mag grid. Problem is once election slows down(collision with ion) the confinement would not be as good resulting in some loss. Maybe you can have a ebeam gun to replenish or find a way to accelerate electrons.

[jtolson]

Like the pinch effect but without the instablilites if run the current on the outside not in the plasma.

[pstudier]

This sounds like a very old idea called a Stellarator. Tokomaks worked better so Stellarators fell out of favor.

See http://en.wikipedia.org/wiki/Stellarator

[TallDave]

I have been thinking if it is possible to achieve mag confinement with a torus configuration(yes, the ugly torus!).

Since polywell mag field can't completely stop electron from going out from the gaps. Maybe we can have a seamless mag field with torus:

Imagine a conductive tube, if you pass current along the tube a circular mag field will be produced inside. If you shoot ebeam through the center of the tube they will experience either an outward dispersive lorenz force or an inward contractive lorenz force depends on which direction you fire them. To confine electron of cause we make them contract into a tight beam this way. Connect the two end of the tube and you got a torus with ebeam going in circle in the center.

The advantage is one don't have to consider electron loss through the gap in mag grid. Problem is once election slows down(collision with ion) the confinement would not be as good resulting in some loss. Maybe you can have a ebeam gun to replenish or find a way to accelerate electrons.

IIRC the reason this won't work in an IEC setup (confining electrons magnetically, confining/accelerating ions with the electrostatic charge on the electrons) is that the "bottom" of the electrostatic well would be in the hole in the center of the torus.

[hanelyp]

IIRC the reason this won't work in an IEC setup (confining electrons magnetically, confining/accelerating ions with the electrostatic charge on the electrons) is that the "bottom" of the electrostatic well would be in the hole in the center of the torus.

No, the bottom of the well would be inside the plasma inside the torus. The hill inside the center of the torus may be shorter than the depth to the outside, but it is a hill. Add in a Faraday cage lining the torus and you get even depth all directions from the plasma.

I see a few problems with a toroidal configuration:
- injecting electrons without a cooperative magnetic path. The better electron confinement is a double edged sword.
- in a magnetic toroid charged particle want to spiral to one side unless the field has a twist. The tokomak does this with plasma current.
- with no passive magnetic well, plasma has little to correct bulk drift to one side.

On a related idea...
The tokomak, from accounts I've read, is derived from the toroidal pinch by increasing an external stabilizing field until the pinch is suppressed. I'm thinking, what if you attempted a toroidal discharge in a charged plasma. Would the electrostatic forces prevent the pinch and associated instabilities?

[rogerlam]

I see a few problems with a toroidal configuration:
- injecting electrons without a cooperative magnetic path. The better electron confinement is a double edged sword.
- in a magnetic toroid charged particle want to spiral to one side unless the field has a twist. The tokomak does this with plasma current.
- with no passive magnetic well, plasma has little to correct bulk drift to one side.

You mean inject electrons would oscillate right? Maybe there are ways to accelerate slow electron up to speed within the torus?

Can't follow you on the other points though. I suppose the electron in the toroid will build up large electrostatic field that the ion won't care about mag field. This is how polywell works anyway.

[TallDave]

No, the bottom of the well would be inside the plasma inside the torus. The hill inside the center of the torus may be shorter than the depth to the outside, but it is a hill. Add in a Faraday cage lining the torus and you get even depth all directions from the plasma.

Yah, I guess you could. But then you've got a weakened, weirdly-shaped well, with the "bottom" spread around a doughnut, and the force spread across a shape that isn't friendly to the distribution of the electrostatic charge force, which is spherical.

One of Bussard's points in the Google Tech video is that the problem with tokamaks is that unlike stars, they use right-angle forces to confine the plasma. If you spread the electrostatic force around a torus, you re-introduce that problem, because the ions are going to see a lot of charge attraction at angles.

[93143]

Without the Faraday shielding effect of some kind of shell (which admittedly is probably necessary in any real system), the hole in the centre is the (flat) bottom of the potential well. It's not a hill. Even with shielding, the torus would have to be fairly skinny for the electrostatic confinement to be any good radially.

This in itself is not a showstopper, but combined with the other points mentioned here it seems that toroidal IEC is not optimal.

EDIT: Technically the 2D case would be a cylinder, not a ring. This invalidates my conclusion. Sorry.

[hanelyp]

One of Bussard's points in the Google Tech video is that the problem with tokamaks is that unlike stars, they use right-angle forces to confine the plasma.

The polywell uses those same right-angle forces to confine electrons. But electrons may be confined over a smaller volume with a weaker field than ions at similar energy. Also, the polywell configuration is not subject to some instabilities that plague the tokomak.

If you spread the electrostatic force around a torus, you re-introduce that problem, because the ions are going to see a lot of charge attraction at angles.

Could you clarify?

Without the Faraday shielding effect of some kind of shell (which admittedly is probably necessary in any real system), the hole in the centre is the (flat) bottom of the potential well. It's not a hill.

Don't assume that the flat potential effect inside a spherical shell applies to a ring. It doesn't.
Consider a test particle in the plane of a charged ring. Integrate the forces of the ring on the test particle over a circle centered on the test particle. For any dTheta, a distant portion of the ring has more total charge in proportion to r, but force is ~Q/r^2. Thus total force is integral over a circle of k/r dTheta, net force towards the near portion of the ring.

This in itself is not a showstopper, but combined with the other points mentioned here it seems that toroidal IEC is not optimal.

Quite possible. A point attractor for the ions does seem preferable.

[rogerlam]

Ok, so instead of a fat donut it may look more like a skinny onion ring. I can live with that.

And it is replacing a zero dimensional well (point) with a less optimal one dimensional well, in exchange for less electron lost(?) and no need for mag grid->simpler. Is it a positive trade off?

[TallDave]

If you spread the electrostatic force around a torus, you re-introduce that problem, because the ions are going to see a lot of charge attraction at angles.

Could you clarify?

Sure, just picture an ion being injected. In a spherical IEC setup, it only sees one point as the "bottom" and moves straight toward it. In a torus, there is a line "bottom" so it sees the charge from many points, most of which are at an angle to where you presumbly want it to go (straight "down").

Also, unless you have a long enough torus that the curvature is very small, I think the "bottom" is going to tend to focus inward, toward the inner walls of your torus. The Faraday cage doesn't eliminate all line-of-sight attraction within the torus, just most of it.

And I wonder about the beam collision set-up in a torus. I assume you're injecting them from along the inner and outer sides of the tube so they meet in the middle of the tube, but the skinnier you make the torus the harder that is, and the wider it is the bigger the previous problem becomes.

It might work, but it seems non-optimal at first glance. But it would be interesting to see if you could succesfully eliminate the electron losses this way.

[93143]

@hanleyp: I'll think it through some time when I'm less busy. It does, however, sound like you're right. The correct 2D case would be a cylinder, not a ring. I was in too much of a hurry...