Page 5 of 11

Posted: Thu May 21, 2009 5:59 am
by MSimon
chrismb wrote:I still don't understand (and have never had an answer when posting the suggestion) why a toroidal solenoid would not do the whole job better. If the whole way Polywell works is some 'multi-species' collective behaviour of ions and electrons, rather than the flat density of both that Art and I expect, then;

why not confine electrons around the central major radius of a toroidal solenoid, then inject the ions into it along poloidal radii? In poloidal cross-section, this would look exactly like any of the diagrams of Polywell, just without the cusps!! If the annealing process works as advertised, then this will deal with ions scattered 'toroidally', just as Polywell should do for those ions scattered at some radius away from the centre.
Isn't that a tokamak type design? Dr. B was rather familiar with those (he worked on them before he came up with the Polywell concept). If he did not do it that way he must have had a reason.

BTW where in the toroid do you put the accelerator grids? And how do you shield them? And how do you create the well?

Posted: Thu May 21, 2009 1:49 pm
by gblaze42
chrismb wrote:I still don't understand (and have never had an answer when posting the suggestion) why a toroidal solenoid would not do the whole job better. If the whole way Polywell works is some 'multi-species' collective behaviour of ions and electrons, rather than the flat density of both that Art and I expect, then;

why not confine electrons around the central major radius of a toroidal solenoid, then inject the ions into it along poloidal radii? In poloidal cross-section, this would look exactly like any of the diagrams of Polywell, just without the cusps!! If the annealing process works as advertised, then this will deal with ions scattered 'toroidally', just as Polywell should do for those ions scattered at some radius away from the centre.
How would you inject the electrons?


Edited - just saw MSimons last post, basically what he said

Posted: Thu May 21, 2009 5:06 pm
by chrismb
MSimon wrote: Isn't that a tokamak type design?
Just 'cos it's toroidal, it don't make it a tokamak!

I'm thinking more like Lavrent'ev/Artsimovich's ATOLL experiment.

Where to put the guns? As I say, take the 2-D section diagram of a Polywell and extrude it into a torus (or long cylinder if you like - proof of principle and all that) rather than rotate it around a diametric axis to get a Polywell. Thus, you will then immediately see suitable locations for all parts for a toroidal format.

Posted: Thu May 21, 2009 5:21 pm
by KitemanSA
Sorry, not picturing it. Sketch?

Posted: Thu May 21, 2009 5:28 pm
by BSPhysics
As I say, take the 2-D section diagram of a Polywell and extrude it into a torus
Isn't the magic of the polywell its ability to recirculate electrons and ions? It seems that making the torus would demand a larger volume, a less symmetric shape and more ways for electrons to escape the system. I wouldn't think a torus could trap or recirculate particles as well as a whiffle ball. FUSION LIKES SPHERES!

BS

Posted: Thu May 21, 2009 5:53 pm
by Barry Kirk
It could be argued that we know a shape that doesn't work... A torus... and we do know a shape that does work... a Sphere.

It may be that other shapes work, but we don't know yet.

Posted: Thu May 21, 2009 6:36 pm
by chrismb
BSPhysics wrote:I wouldn't think a torus could trap or recirculate particles as well as a whiffle ball.
That's funny! Made me laugh.
Barry Kirk wrote:It could be argued that we know a shape that doesn't work... A torus... and we do know a shape that does work... a Sphere.
What do you mean, a torus doesn't work?
KitemanSA wrote:Sorry, not picturing it. Sketch?
You're thinking on it too hard. OK, so the Polywell was conceived from a fusor where the grid was taken out, and in place of it a magnetic field was put around it to hold electrons in place to act as a -ve grid. [Are we agreed on that tiny point, at least, or is there some endless arcane 'knowledge' going to be invented/dragged up to make Polywell look somehow more mysterious and earth-shattering than that?]

So, quite simply, you can theoretically do a fusor in a toroidal format, thus;

http://www.fusor.net/board/getfile.php? ... tt_id=4983

Now, as per WB taking out the grid and sticking in magnets to recreate the central grid, do the same with this construction.

I do think you are thinking too hard on this. Imagine a circle spinning on a diameter, you get a sphere. Imagine a circle pulled out of the page (and back on itself) you get a cylinder/torus. Now do the same operations to the diagram of Polywell. You'll get a ring of electrons at the torus' major circumference held by the toroidal magnetic surfaces. They will form a ring rather than a small volume, but still they should act to attract ions inwards, along poloidal radials, just like the pic. - if Polywell works.

Posted: Thu May 21, 2009 6:52 pm
by Barry Kirk
hmmmm...

Welll....

If I understand what your saying, with the Torus shaped fusor, Fusion occurs on the circumference of a circle located at the center of the minor radius instead of at a point. Well that does improve the volumetric efficiency quite a bit.

However, the picture you provided was for a gridded fusor. The genious of the Polywell is to Magnetically shield the grid and then to allow electron recirculation.

All the volumetric efficiency in the world won't help if your losing too many electrons to collisions with the grid.

I'm not sure what geometry you would need to allow electron recirc.

Perhaps wrapping the magnets around the torus in a helical pattern with space between the wraps for electron recirc might do the trick?

Posted: Thu May 21, 2009 6:59 pm
by chrismb
Barry Kirk wrote: The genious of the Polywell is to Magnetically shield the grid and then to allow electron recirculation.
But in a Polywell, that statement only applies because there are cusps through which the electrons get lost and need to be recovered. In a toridal solenoid there ARE NO cusps through which electrons can get lost!

fields

Posted: Thu May 21, 2009 7:13 pm
by bcglorf
chrismb wrote:
Barry Kirk wrote: The genious of the Polywell is to Magnetically shield the grid and then to allow electron recirculation.
But in a Polywell, that statement only applies because there are cusps through which the electrons get lost and need to be recovered. In a toridal solenoid there ARE NO cusps through which electrons can get lost!
Isn't the benefit from a spherical arrangement about trading off magnetic confinement for electrostatic confinement?

It's a given that a Polywell's magnetic confinement is worse than a toroidal magnetic field. The gain is that the electrons that are contained in the centre don't experience a net electric force from a uiniform shell of ions introduced around them. In a toroid that isn't going to be true and electrons won't just be pulling ions in, but the ions will be equally pulling the electrons out leaving you with little more than straight magnetic confinement.

Of course, this is way over my head so if that was ignorant and stupid just dismiss me and carry on while I try and follow along as best I can :).

Re: fields

Posted: Thu May 21, 2009 7:30 pm
by chrismb
bcglorf wrote:The gain is that the electrons that are contained in the centre don't experience a net electric force from a uiniform shell of ions introduced around them.
OK, this is good thinking and beginning to drill down into a little more detail. In a 'perfect' toroid, if electrons were held at the poloidal centre (the major circumference of the torus) then, because the inner part of that circle is smaller than the outer, the electrostatic forces would be asymmetric if the ions were to be static at the outer limits of their travel. (They aren't, of course, see lower down.)

But, of course, the whole geometry need not be exactly like that and the poloidal cross-section can become a little elliptical to compensate, or the electrons need not end up at the dead centre of the poloidal cross-section. Or, if you don't buy any of that, you could make the hole thing a figure of 8 doughnut and so the charge would bunch up on the inside of the curves and spread out on the outer. Also keep in mind that the ion charges are not meant to be static at what would be the 'inner' and 'outer' sides of a poloidal cross-section, so each time an ion reciprocates about a poloidal diameter, if it is heading in a direction to the centre of the whole torus then effectively the ion density should become more compressed on the inner side (i.e half of the reciprocating ions should be either side of the centre, whichever way you look at it).

[Bear in mind none of those solutions can work for a tokamak because a tokamak relies on rotational translation of ions and needs to rotate them around poloidally aswell (the more polodial rotations per toroidal rotation, the higher the stability). None of this is relevant to what I am describing.]

Posted: Thu May 21, 2009 7:35 pm
by Barry Kirk
The polywell has 2 major advantages over the Tokamak.

1) It is only trying to directly confine electrons not ions. Electrons are a lot easier to confine.

2) It is not using a Maxwellian distribution for energy.

By flipping a Tokamak from ion confinement to electron confinement, you gain one of those advantages. As for the other one, I'm not so sure.

Posted: Thu May 21, 2009 7:48 pm
by 93143
The toroidal configuration has problems with plasma instability. One of the reasons Polywell is supposed to be better is the inherent MHD stability of the field configuration. With a wiffleball, you push outwards and the field strengthens. With a toroid, you push outwards and the field weakens.

With a torus, you lose all the advantages of the "wiffleball" configuration, since you can't run at high beta. The electrons are always on field lines, so they aren't truly 'confined'; by turning a neutral 5 keV plasma into a non-neutral 50 keV plasma, you've just taken ITER's confinement troubles and magnified them substantially. Aneutronic fusion of any description is probably out, and even if it's not, direct conversion will be a nightmare.

I suppose it could work, but if EMC2 can get the transport worked out for Polywell it should be vastly superior to any toroidal system.

Re: fields

Posted: Thu May 21, 2009 7:50 pm
by bcglorf
chrismb wrote:
bcglorf wrote:The gain is that the electrons that are contained in the centre don't experience a net electric force from a uiniform shell of ions introduced around them.
OK, this is good thinking and beginning to drill down into a little more detail. In a 'perfect' toroid, if electrons were held at the poloidal centre (the major circumference of the torus) then, because the inner part of that circle is smaller than the outer, the electrostatic forces would be asymmetric if the ions were to be static at the outer limits of their travel. (They aren't, of course, see lower down.)

But, of course, the whole geometry need not be exactly like that and the poloidal cross-section can become a little elliptical to compensate, or the electrons need not end up at the dead centre of the poloidal cross-section. Or, if you don't buy any of that, you could make the hole thing a figure of 8 doughnut and so the charge would bunch up on the inside of the curves and spread out on the outer. Also keep in mind that the ion charges are not meant to be static at what would be the 'inner' and 'outer' sides of a poloidal cross-section, so each time an ion reciprocates about a poloidal diameter, if it is heading in a direction to the centre of the whole torus then effectively the ion density should become more compressed on the inner side (i.e half of the reciprocating ions should be either side of the centre, whichever way you look at it).

[Bear in mind none of those solutions can work for a tokamak because a tokamak relies on rotational translation of ions and needs to rotate them around poloidally aswell (the more polodial rotations per toroidal rotation, the higher the stability). None of this is relevant to what I am describing.]
It sounds to me like all the same arguments for minimizing losses from magnetic confinement in non-toroidal shapes. The bottom line being magnetic confinement in anything but a toroid is going to be worse than a toroid. Similarly electrostatic confinement with anything but a sphere of charge is going to be worse.

I still see any toroidal electron cloud being quickly torn apart by any ions you try to use it to confine and ending up with a magnetically confined neutral plasma, and then you might as well have a tokamak. But as evidenced previously I'm way out of my league with this.

Posted: Thu May 21, 2009 8:17 pm
by MSimon
I still don't see how you get an electrostatic field in a torus.