Page 1 of 43
magrid configuration brainstorming
Posted: Fri Nov 09, 2007 2:56 am
What other variations of magrids might work besides a polyhedron? How about two coils parallel, with either same or opposite direction of B-field, or a single coil?
The single coil would be like the Levitated Dipole Experiment, except that electron transport outward should be minimized by the charge of the coil.
The north-north or south-south configuration with two coils would be kinda like a polywell, but with one big line cusp and then the normal coil-center cusps. It might be able to create a wiffle-ball.
The north-on-south config would be a magnetic bottle or magnetic mirror. This has demonstrated confinement ability; a problem might be that the ions might be confined as well. This and the above set-up both resemble the U. of Wisc. beam-bunching operation with the coil position.
These would all have to rely heavily on recirculation; the single-coil is the ultimate recirculator. They might not have sufficient confinement to achieve a high electron density in the center and low density outside (although they might have better B-shielding of the magrid). How do you think the transparence of the grids to alphas in these configs would stack up against polyhedrons?
Posted: Fri Nov 09, 2007 4:45 am
The levitated dipole would have most of the electrons on the outside, like the Van Allen belts of the earth. With the Van Allen belts, there is zero density in the center, because the earth is there. See http://en.wikipedia.org/wiki/Van_Allen_radiation_belt
Posted: Fri Nov 09, 2007 4:47 am
Good questions!! I spent a lot of time looking at how to place an arbitrary number of coils around a sphere. For 4, 6 or 12 you can get them all equal and uniformly distributed, but no other combination will work (1 and 2 can be "equally distributed, but my other rule was they touched).
That led me to ask "why uniformly distributed?" and then the question "why on a sphere?"
and the answer is "it's easier!" This is why I like the virtual polywell - we can ask lots of questions and try all kinds of different experiments in theory. If it doesn't work in theory, there's no point in trying it for real.
One of the main questions I'd like to answer even for the 6 coil setup is "what radius is optimal for circulation?" If the coils actually touch, it makes it hard for the electron current to flow past the coil. If the coils are too small, they don't create any well to speak of. So there has to be an optimal radius of the coil. How do we find it? Lots and lots of simulations. It will be a lot easier to run 50 simulations than to build 50 real experiments.
However - we have to build at least 2 or 3 experiments to verify the simulations. Simulations help point the way, but they can't possibly tell you everything.
The more ideas we come up with to try in theory, the better chance we have of finding a great way to build something real. I think it'd be fun to try to model your questions and see what happens!
Posted: Fri Nov 09, 2007 5:20 am
Dr Bussard is quoted as saying that he patented the polyhedral configuration because it was the only one that could ever work. I firmly believe he was correct. People have been trying for years to design magnetic fields for colliding particle reactors, as evident in these patents
US517083 - Method For Forming Magnetic Fields
US5034183 - Apparatus For Colliding Nuclear Particle Beams Using Ring Magnets
US4202725 - Converging Beam Fusion System
Some work was done in the 70's trying to use penning traps for fusion confinement, but it all comes down to the losses. Yes you can trap some electrons, but when it comes to what we are trying to do anything other than the polyhedral cusp just has too many losses. But i dont want to discourage you solo, keep thinking, you could stumble upon something we missed.
It a good thing you bought up coil geometry Dr Mike. I was thinking about it a few days ago.
WB6's minor torus diameter = 1/5 of major torus diameter.
It seems to closely follow the ideal geometry of a Brooks Coil.
Its geometry is minor diameter = 1/4 of Major diameter. Its obvious this gets tweaked to the 1/5th rule to allow for the circular conformal coil geometry to minimise electron losses.
Posted: Fri Nov 09, 2007 6:59 am
@pstudier: perhaps, but a high-potential toroid should create a virtual potential in its center, I think. That would tend to concentrate the electrons there. But then, I guess if the potential were positive, you wouldn't have the negative potential needed to confine the ions. Or would you? It seems like that argument, if true would work against the traditional polywell idea too. Eh, oh well, you are probably right.
@dr.mike: so it looks like the two-coil approaches might be worth modeling. How much trouble would it be to mod your model? Well, heck, it really wouldn't take much; just comment out the four coils on the sides and leave two opposing coils, and then switch up the polarity to get the "magnetic bottle." The single-octant approach should do fine for that.
@Keegan: thanks for breaking it to me nicely! I had no idea people had done that much research on this kind of thing. Here's a nutty thought that I wrote up before I saw your comment, so I guess I'll post it anyway:
What if you merged the coils where the line cusps are, so then there are no cusps? The current through the junction would be zero, so that's no good. But you could change the thing up, so you have a polyhedron with an even number of edges meeting at each vertex (an octahedron for example) and you had a current running through each edge of the thing. (A truncated cube has three edges at each corner.) Each face would apparently have a two currents beside it flowing counterclockwise, and one flowing clockwise, or vice versa, it looks like. Potentially, there could be one or two faces with all the current in the same direction. I'm afraid that the lopsided faces will result in problems, either a lack of shielding for the grid or a lack of confinement for the electrons. And I bet it's geometrically impossible to make all the currents flow the same direction around each face. Yeah, that seems pretty obvious.
I've no idea whether that is a good idea or a load of crap. It's well past my bedtime
I guess it depends on whether those lopsided faces are worth anything.
Posted: Fri Nov 09, 2007 5:42 pm
Solo wrote:I've no idea whether that is a good idea or a load of crap.
Thus the whole of polywell design to extract energy from fusion. I hope for the former and fear the latter.
Posted: Fri Nov 09, 2007 6:33 pm
@solo It would be trivial to do 2 coils as you point out. Definitely worth investigating if only to prove why it won't work!
@keegan I have definitely not gotten that far! I'm thinking more about placement geometry than coil geometry. If a delta function sized coil doesn't work, there's no point in worrying about coil geometry.
One way I think about recirculation is to imagine cold electrons on the B field lines. They just follow the field. If your field hits metal, you lose electrons. A uniform polyhedral field will always work. You have rings of current with rings of B field and rings of electrons. Nice and clean. Think of it topolgically as a "knot" problem. I suspect Bussard is right, it's the only thing that will work. It's a darn good place to start!
Posted: Fri Nov 09, 2007 7:32 pm
@scareduck: I was speaking there about my idea, not about Bussard's whole plan! Sorry I was unclear.
After reading the latest paper here:
http://ecow.engr.wisc.edu/cgi-bin/getbi ... nl0107.pdf
I really think the polywell could be workable.
@drmike: yeah, the polywell certainly has simplicity going for it. This topology-meets-physics is interesting!
Posted: Fri Nov 09, 2007 11:17 pm
I probably don't entirely comprehend the problem, but I thought Dr Bussard (in the video?) said something about equatorial line cusps being very bad for e losses. If you have 2 coils, how are you not going to have a line cusp where the fields meet?
Posted: Sat Nov 10, 2007 3:44 am
John, you are right. Cusps are where the loss problems crop up. As I understand it, there's a trade off between bring the coils close enough together to trap the electrons without making them bump into the coil. But every machine is going to have cusps. I am just trying to throw out some ideas. I'm really not sure I understand the whole cusp business myself.
Posted: Sat Nov 10, 2007 3:17 pm
If you have one coil, the magnetic field looks like a torus. It is straight in the center of the coil and highly curved away from it. If you take two of these and smash them so the fields in the center cancel, the field has to squash away somewhere else. Since it is already curving around the coil, the magnetic field moves away from the center and towards its own coil.
Think of the magnetic field as a flexable material. It flows and bends, but it can't go away.
So if you smash several fields together, the fields have to smash thru some where.
The regions between the coils must have very high fields because that's the only place for them to go. If the electrons have low temperature they will stick to the field lines and circulate around the coils. If they have high temperature they will not stick to the field lines so well and hit the MaGrid.
This is where some modeling can help. Understanding the trade off between current density, temperature allowed and MaGrid physical size will be the difference between building something that works and something that doesn't. Lots of short repetitive pulses will allow lower temperatures, but it is harder on the equipment. It may well be that crummy coils and weaker fields will work better for a longer life machine because it would allow for higher electron temperatures (and more circulation = less losses).
With two coils, you have a lot of circulation, but not much trapping. That's a different trade off. I think it is definitly worth investigating!
Posted: Mon Nov 12, 2007 3:27 am
On the topic of alternate coil configurations, I thought I'd illustrate an idea I've been kicking around. The wiffleball would be in the center. This was inspired by a proposed permanent magnet configuration.
Posted: Mon Nov 12, 2007 12:54 pm
It seem quite interesting!!!. Congratulations!! I am not an expert but maybe it could reduce even more the electron loosses because the recirculated electrons will go always around the inner part of each magnet. If the electron lifetime is increased it would be a perfect candidate to get net power. What about the Magrid and the WB in the cusps? Is it maintained fine? Opinions?
Bussard metions, I think in Google video, that a good confinement is reached if in all the border regions of each magnet the polarity is alternated with the one in front ( North-South-North-South ) to help creating the WB. Have you check it for your device? The opinion of some expert is needed.
Posted: Thu Nov 15, 2007 9:23 pm
Another configuration to chew on:
The core principle of the polywell is that electrons, with their high charge to mass ratio, are much easier to contain. The resultant charge well can then be used to contain ions. What if the electrons were contained in some other magnetic configuration, like the tokomak? Having no cusps, this toroid might contain the electrons better, unless there's some instability I'm not familiar with. The secondary ion containment still works. But I expect the electrons will thermalize. Then again, I'm not sure they wouldn't in a polywell.
Posted: Thu Nov 15, 2007 10:11 pm
The trouble with a tokamak configuration is that the centre of charge for the electron distribution is outside the confinement area (ie: in the hole in the middle of the donut).
There could be some sort of equilibrium with ions in a generally tighter ring than the electrons but with the distribution overlapping, but something feels off about that idea (and I'm apparently too tired to figure out what). It would be fine for attracting ions perpendicular to the plane of the torus, but I'm not so sure it would be a good idea for radial confinement...