magrid configuration brainstorming

[rjaypeters]

If you used the same short segment design found here, in higher order polyhedra, you might actually come up with the ultimate polywell design.

IIRC, happyjack27 liked that design also.

To take that thought and run with it...


The gapped extension pointing out to the lower right is a crude representation of a superconducting switch. Blue conductors lead current away from the "switch." After the halfway point, magenta conductors lead current back to the switch.

I've made this a single coil design because single-coil is my current emphasis.

I think I've answered MSimon's question about magnetic field for single-coil configurations. Cooling is relatively easy in that the coolant enters and exits via the support structure. Support structure and actual configuration of the coils...can't really be investigated until someone simulates a proposal.

Since the dodecahedron seems to have good performance, I'll look at that next. It will be single-coil.

P.S. I kept the group-of-four conductor extension, coolant and structural supports, but now I'm thinking that might be a waste of good superconducting cable. OTOH, the current carried by the conductor provides magnetic field to keep particles away from the supports...

P.P.S. I doubt there is an ultimate polywell design, just as there is no ultimate jet engine design. We design specifically for what we need.

P.P.P.S. I'm not sorry about creating so many concepts and showing them here. Among other things, it helps maintain my sanity and is a much better use of my time than some alternatives.

[Aero]

Does that solve the "Nub" problem by magnetic shielding? It looks like it might, but I would like to see a magnetic field map of this configuration. Intuition is not enough.

[hanelyp]

In most details I like that design. I only see one problem: an electron that finds it's way through the cusp and into the channel between a quartet of feed and support lines will see no electric field to recirculate it back into the core until it escapes from the channel or reaches the far end. This might be addressed if a potential gradient can reasonably be maintained along the length of the supports.

[rjaypeters]

Another aspect of the support of the single-coil designs: I think the outer case of all the supports and "ring" segments have the same charge. My reasoning: since there is only one conductor, it must be contained in some case, I assume metal*. Since this conductor case passes through the cooling case(s), the conductor case distributes charge to the cooling cases.

I can imagine insulators between the conductor case and the cooling cases where the conductor case penetrates. But it would be fun to build.

Am I thinking correctly? And will enough electrons finding their way between the supports be a problem? Only a simulation will tell.

*Maybe not a good assumption.

[rjaypeters]

Dodecahedron, single-coil, short-segment. Hideousness...





This is the result of trying to cram 49 N into a 9.8 N bag! The view is only of half of the coolant casings and looking almost directly down the gap between a group-of-four casings.

The assumption in the pictures is the standard 1m radius sphere with 0.2m diameter cooling casings bending with a minimum radius of 0.26m. The bending radii actually overlap.

Next I'll be using a 2m radius sphere.

[imaginatium]

In most details I like that design. I only see one problem: an electron that finds it's way through the cusp and into the channel between a quartet of feed and support lines will see no electric field to recirculate it back into the core until it escapes from the channel or reaches the far end. This might be addressed if a potential gradient can reasonably be maintained along the length of the supports.

The big question is how many electrons could find there way down that channel, and then once in the channel, how likely is it that they would escape? The entire length the of the support should be magnetically shielded, since the conductor runs down it's length, but between them are 4 line cusps running the entire length. Any electron that made it down the entire length, would slam into the chamber wall, but I would expect, that many would escape out the line cusps, along it's length, and recirculate back into the wiffleball.

It would be interesting to see the effect of splaying the supports 33 degrees.

[rjaypeters]

2m radius dodecahedron, 5 x 2 x 12 = 120 supports:


Support tubes removed:


I took some liberties with the tube bundle spacing to simplify my task. Chamber diameter ~8m.

[D Tibbets]

I like the 8 faced curved magrid design. The variation that might be more manageable, is to have each coil separate. That way the outgoing arms can serve as both the path for the current carrying magnet wires, coolant flows, etc. Having each magnet separate, with it's own separate cooling flows, would ease the coolant flow requirements (per magnet instead of all magnets in series) Also, the extensions serve as the support standoffs. The positive potential could be maintained on the metal shell of the magrid itself, while the extending arms could be ceramic insulators.
The corner cusps would taper into a narrow sheath clear to the vacuum chamber wall. This might be a problem, or it may have advantages for direct conversion, recirculation, electron gun focusing,etc. These extension/standoffs would be fully magnetically insulated clear to the vacuum vessel wall. If a port was placed at the end of these corner sheaths, escaping electrons and ions might be fed directly into a baffled vacuum pump or ion diversion / concentrator port into a vacuum pump. For these cusps at least, you might end up in maintaining a tolerable non arcing background chamber pressure, with higher Wiffleball densities, The system would be less demanding on the vacuum pumping, or rather, it would improve the effective vacuum pumping capacity, thus allowing pushing the Wiffleball density a small (?) amount. And, as fusion rate proceeds as the square of the density, it might allow for significantly increased fusion rates at the same B field strength and volume.

Dan Tibbets

[rjaypeters]

...The variation that might be more manageable, is to have each coil separate. That way the outgoing arms can serve as both the path for the current carrying magnet wires, coolant flows, etc. Having each magnet separate, with it's own separate cooling flows, would ease the coolant flow requirements (per magnet instead of all magnets in series)

The intent is to show one current-carrying coil (if that is technologically possible - no one has answered my previous question about splicing SC cable), but the cooling flows are arranged in any fashion we find desirable. One, two, three, etc. passes through the "core" may be possible before the coolants need to be recooled and sent back through the core. I'm just not showing the exterior coolant paths because, once again, I don't have any idea how many passes around the core are possible before recooling is necessary.

Also, the extensions serve as the support standoffs. The positive potential could be maintained on the metal shell of the magrid itself, while the extending arms could be ceramic insulators.

Works for me, but...

The corner cusps would taper into a narrow sheath clear to the vacuum chamber wall. This might be a problem, or it may have advantages for direct conversion, recirculation, electron gun focusing,etc. These extension/standoffs would be fully magnetically insulated clear to the vacuum vessel wall. If a port was placed at the end of these corner sheaths, escaping electrons and ions might be fed directly into a baffled vacuum pump or ion diversion / concentrator port into a vacuum pump.

I think I see your point about separating the coils, but won't the magnetic field exactly between the "turn-arounds" be zero? Viz:

We are looking at the turn-around, but end-on:

Blue arrows show current direction at the turn-arounds.

And if we keep the metal case almost all of the way to the chamber wall, won't the electric field between the four stand-offs also be zero? And then we will get the apertures you envision?

If I have missed your points, please let me know.

I like this semi-octahedron much more than the dodecahedron monstrosity, if only for simplicity's sake. I'm going to look back and see if we have a simulation of the spherized Tombo to see how well it behaved from a confinement perspective.

[rjaypeters]

Re: Dodecahedron single-coil with many supports.

Here's the 7m diameter vacuum chamber:

Cheers!

I'm slightly tempted to produce the segment list for this monstrosity, if only for the benefit for the Asian lurkers out there who have taken happyjack27's code and are reporting the results to their respective industrial establishments. On second thought, never mind, they've probably already reverse-engineered it.

[D Tibbets]

My intention was to have each leg penetrating through the wall of the vacuum chamber. Outside the vacuum chamber, the magnet segment is supplied with electrical surface current, magnetic current, and coolent plumbing individually. The leg extending to (and through) the vacuum chamber wall is electrically insulated, so there would not be any potential on the extension surfaces themselves*. The only electrostatic interaction of the electrons trapped in this sheath would be to be reversed and recirculated by the positive charge on the metal surface of the actual magrid, or to fly to and hit the vacuum chamber wall (or a vacuum pump port located there).

The heat load on the entire magrid in a working machine may be from a few MW to a few dozen MW or more. Things inside the pipe will be crowded. My supposition is that it will be difficult to force enough coolant through this fixed pipe to cool one magnet, yet alone 6,8, or more. Remember, if a superconductor is used, it will not be difficult to keep it cooled by itself, but insulating against and carrying away the heat from the X-rays, electrons, ions, and possibly neutrons that will impart the outside of the magrid will impart a huge amount of heat into the magrid.

*M. Simon mentioned that insulated standoffs (the extended tubes in this case) would not have a potential intentionally applied to to it, but because escaping electrons could reach the surface and would secondarily impart a negative charge to this surface (it is floating in electrical terminology). This would provide some electrostatic shielding against further electrons reaching them. These magnetized standoffs/ tubes, would have the benefit of the magnetic shielding, and electrostatic shielding added to that due to the ExB drift of some of the electrons through the magnetic field to the insulated surface (not grounded, or connected to the positive potential on the magrid proper). Also, need to consider inducing a positive charge to the surface of the insulated pipe (Gauss's Law stuff, I think), but I am assuming the electrons dominate over ions, and the electrons leak through the magnetic shielding to directly contact the surface at a rate that dominates over other effects. When you consider the high energy (with associated much larger gyroradius) fusion ions exiting the cusps, the picture is more complicated, and is either a problem or an advantage depending on what clever direct conversion schemes might be developed.

Note: The Superconductor could be one continous strip/ tube assembly weaving through the larger multiple insulating and cooling pipes/shells contained within the Pipes, with the turn arounds actually being outside the vacuum vessel wall, while the cooling. the other support pipes/ shells are isolated to a single magnet.

Dan Tibbets

[rjaypeters]

Okay. I just like the single-coil, if it can be built...

Edit: I looked back and the spherized Tombo was never submitted for simulation. AFAIK it hasn't been simulated.

Topologically, the closest coil configuration to a spherized Tombo is a tetrahedron which has terrible confinement (according to happyjack27). Let's be careful about loving the spherized Tombo too much before a simulation gives an idea about performance.

[KitemanSA]

My intention was to have each leg penetrating through the wall of the vacuum chamber.

This would suggest that the highly positive MaGrid would reach the ground state chamber wall, in which case there would be little or no circulation, right?

[rjaypeters]

The leg extending to (and through) the vacuum chamber wall is electrically insulated, so there would not be any potential on the extension surfaces themselves*

I'm the one who wants the extensions to have conductive cases. But even my ignorant self knows they need to have insulators separating them from the vacuum chamber walls.

[rjaypeters]

One technique to reduce the numbers of extensions and core segments is to depend on symmetry (in this case bilateral). I wanted to combine core segments for the dodecahedron single-coil, but there isn't bilateral or tri-lateral symmetry to exploit.

The classic WB has some bilateral symmetry that might be useful. NB the +Z and -Z extensions are only paired:



I think the +Z and -Z coils can be included in the single-coil scheme with SC cable outside the vacuum chamber, but I have limited ambition this morning to investigate*. From previous correspondence, I think (hope?) the paired extensions attract each and so increase their effective strength.

If the paired extensions don't work out, we can go to quadruple extensions all around, but it starts to get messier again. We could go part of the way by quadricating(?) the +x and -X coils to include the +Z and -Z coils in quadruple groups. I'll probably do that later today.

As you can see from the core detail, I haven't optimized the geometry for the coil close approaches and I have taken the usual minor liberty with core to segment transitions to simplify my work.

* Just as my head hit the pillow last night, this concept popped into my head! I'm just glad I was able to go to sleep, sheesh!