Spherical and hierarchic grids

Discuss how polywell fusion works; share theoretical questions and answers.

Moderators: tonybarry, MSimon

ankovacs
Posts: 23
Joined: Mon Nov 24, 2008 6:32 am
Location: Espoo, Finland

Spherical and hierarchic grids

Post by ankovacs »

I am puzzled by the persistence of using 'almost polyhedron' shaped grids, instead of more spherical geometries.

Here is a link to a grid arrangement I thought up, made of 3 rings:
http://www.broadbit.net/download/Spheri ... ometry.pdf
The descriptions lists some advantages of it, compared to WB-6; can anyone suggest any reason to remain with the 'almost cube'/'almost dodecahedron' track?

An other possibility is the spherical 'tornado trap' arrangement:
http://ps-div.web.cern.ch/ps-div/ECRIS9 ... /Zorin.pdf

Then the question that puzzles me is whether some concentric arrangement of grids could produce a concentric sequence of 'wiffle-ball' surfaces, with limited cusp holes. See the more elaborate description of that question here:
http://www.broadbit.net/download/Hierar ... nement.pdf
I would be very interested in computer simulation to validate whether such arrangement is feasible.

If above proposal proves to be a possibility, then the 'cusp hole size debate' that went on this forum is less relevant: concentric grids keep electron density sufficiently low outside to prevent Paschen arcing, and the magnetic field arrangement prevents heated electrons from flying out to outer wall.

The main feasibility issue then to resolve will be the balancing of synchrotron radiation loss from electrons that are heated by ions.

MSimon
Posts: 14334
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:

Post by MSimon »

The descriptions lists some advantages of it, compared to WB-6; can anyone suggest any reason to remain with the 'almost cube'/'almost dodecahedron' track?


There must be no metal at magnetic field zeros or where the slope of the field is perpendicular to the coil form. Those are places where there will be electron losses.

In addition it is not clear how the geometry you propose will allow wiffle ball formation/closure.
Engineering is the art of making what you want from what you can get at a profit.

hanelyp
Posts: 2261
Joined: Fri Oct 26, 2007 8:50 pm

Post by hanelyp »

The proposed octahedron configuration would allow formation of a wiffleball. But as MSimon points out, it has coils running through field null points. Cusps must be open to allow electron recirculation. That could be rectified by using 4 loops on half the faces of the octahedron. Also, an octahedron is farther from a sphere than the truncated cube the wb-6/7 config is based on.

ankovacs
Posts: 23
Joined: Mon Nov 24, 2008 6:32 am
Location: Espoo, Finland

Post by ankovacs »

as MSimon points out, it has coils running through field null points.
True, but the space surrounding these points has non-zero field; as shown in the simulation diagram approaching electrons are steered away. Or do you think electrons could still hit these points for some reason?
That could be rectified by using 4 loops on half the faces of the octahedron
I would be interested to see a drawing of that, as I do not clearly understand what you mean there.
an octahedron is farther from a sphere than the truncated cube
That is true on one hand, but on the other hand the enclosure of my proposal is more spherical and fields of nearby coils do not work against each other. So the question is the balancing of these effects.

What do you think about the feasibility of the formation of multiple hierarchic wiffleballs? Does anyone have an argument showing why it would not be feasible?

MSimon
Posts: 14334
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:

Post by MSimon »

True, but the space surrounding these points has non-zero field; as shown in the simulation diagram approaching electrons are steered away. Or do you think electrons could still hit these points for some reason?
Well yes. The corner cusps in the truncated octahedron (cube) have fields (very strong fields) all around them and yet they are a loss mechanism (were it not for "recirculation" or as I prefer oscillation).

In your set up there is only loss - to the metal.
Engineering is the art of making what you want from what you can get at a profit.

ankovacs
Posts: 23
Joined: Mon Nov 24, 2008 6:32 am
Location: Espoo, Finland

Post by ankovacs »

Cusp loss and recirculation is the same mechanism in all setups. In fact the distance for the recirculating electron to travel is shorter in my proposal than with WB-6/7. (with similar gird dimension) That is significant because recirculation ratio depends on mean free path.

Whether electrons hit the metal is a different question, I do not see yet why electrons would hit metal in my proposal.

ankovacs
Posts: 23
Joined: Mon Nov 24, 2008 6:32 am
Location: Espoo, Finland

Post by ankovacs »

I made an update to the grid geometry analysis:
http://www.broadbit.net/download/Spheri ... try_v2.pdf

A small change in the 'spherical octahedron' grid prevents junctions at zero magnetic field, without making any significant difference to the inner field lines. (just to settle the worries about metal at zero magnetic field)

Also, the last part shows how the Polywell grid is enclosed into a Tornado trap at zero potential. I expect that this provides a major help with preventing Paschen arcing.

Hopefully this geometric arrangement brings it closer to validate the Polywell fusion concept.

spaccemonkey
Posts: 10
Joined: Mon Dec 08, 2008 3:26 pm

Post by spaccemonkey »

ankovacs says in http://www.broadbit.net/download/Spheri ... try_v2.pdf
In summary, slow electrons are turned back by towards center by electric force, and fast electrons are
reflected magnetically
The charge on the magnetic grid is not relevant to the behavior of a charged particle inside it's volume, so the only thing confining the electrons is the magnetic field. I would suggest starting with the thread "Ureka, my new understanding of the Polywell" to help get your head around the basic idea of how this device is supposed to operate.

Do keep in mind that the magrid is positively charged and consists of electromagnets.

hanelyp
Posts: 2261
Joined: Fri Oct 26, 2007 8:50 pm

Post by hanelyp »

An image I posted to the magrid brainstorming thread:
Image
Same basic config as the cube polywell, but with the coils bent to be more 'spherical'. Is this something like what you're trying to get at?

ankovacs
Posts: 23
Joined: Mon Nov 24, 2008 6:32 am
Location: Espoo, Finland

Post by ankovacs »

spaccemonkey says: The charge on the magnetic grid is not relevant to the behavior of a charged particle inside it's volume
I was talking about turning back electrons moving outwards, when they are OUTSIDE the Polywell grid. It seems my English was not clear enough - don't assume right away that I am missing the concept :wink:
hanelyp says: Is this something like what you're trying to get at?
Yes, we are on the same track. I am sure there are even more variations on this theme.
What should be investigated is the proper balance between very symmetric Polywell grid (when cusp sizes are expected to be smaller but field lines stretching far out before returning) and a less symmetric but more tightly recirculating Polywell grid.
As an additional requirement, the Polywell arrangement should complement well an outer Tornado-trap in my view, which is emphasizing the point on tight recirculation.
3-D Vlasov simulators would help to clarify if current computers were powerful enough to do it.

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

hanelyp wrote:An image I posted to the magrid brainstorming thread:
Image
Same basic config as the cube polywell, but with the coils bent to be more 'spherical'. Is this something like what you're trying to get at?
This is much like something I have been working on. If the sides of the "square" coils actually followed the curvature of the sphere rather than having two straight segments, and the connections (the yellowish tubes) went radially from the vertices for a small distance before crossing between coils, I think we would have something.

First, we could use a slightly modified Bitter magnet. As an alternative, it may be fairly easy to lay up a high temperature superconductive magnet similar to the modified Bitter magnet. The ribbons of HTSC might be almost as easy to lay up as the ribbons of copper for the Bitter magnet.

Second, If I understand the forces created by the coils and virtual coils, this will minimize the stresses in the magnets.

Third, with the crossovers configured as suggested, they could be coated with a thermally conductive electrical insulator an thereby become negatively charged. This should limit the electron loss to the crossovers.

Could you modify this graphic as described?
Thanks

classicpenny
Posts: 106
Joined: Thu Feb 14, 2008 5:50 pm
Location: Port Angeles WA USA
Contact:

Post by classicpenny »

ankovacs, your configuration reminds me of Tombo's first octahedron drawing back on May 31, in "Magrid Configuration Brainstorming."

viewtopic.php?t=289&postdays=0&postorder=asc&start=60

The thing that always bothered me about that drawing was the directions of the magnetic fields produced by the coils on the opposite faces of the octahedron: if you apply the right-hand rule to the coils on the opposite faces, the resulting magnetic fields do NOT oppose each other. This seems to me to be a very serious problem - indeed I find it very hard to imagine how a Wiffleball could be produced under these circumstances.

Bill Flint

blaisepascal
Posts: 191
Joined: Thu Jun 05, 2008 3:57 am
Location: Ithaca, NY
Contact:

Post by blaisepascal »

classicpenny wrote:The thing that always bothered me about that drawing was the directions of the magnetic fields produced by the coils on the opposite faces of the octahedron: if you apply the right-hand rule to the coils on the opposite faces, the resulting magnetic fields do NOT oppose each other. This seems to me to be a very serious problem - indeed I find it very hard to imagine how a Wiffleball could be produced under these circumstances.
I'm not sure it's necessary for opposing faces to have opposing magnetic fields. Consider the analogous 2D case: six equally-spaced magnets, 3 pointing in, 3 pointing out, alternating around the center. If you plot the field lines, there is still no field in the center, and the lines go to the adjacent magnets, not to the opposite side.

An octahedral system like ankovacs originally described is topologically identical to a tetrahedral polywell, (in much the same way that the early cuboctahedron reactors Bussard build are topologically identical to the WB-6/7 cube reactors).

ankovacs
Posts: 23
Joined: Mon Nov 24, 2008 6:32 am
Location: Espoo, Finland

Post by ankovacs »

Yes, the criteria is to be able to pair arcs in such way that their magnetic field cancels out at the center of sphere.
It does not matter if opposing 'faces' have same current orientation on edges (they are not planar anyway), as long as above cancellation can be done.

classicpenny
Posts: 106
Joined: Thu Feb 14, 2008 5:50 pm
Location: Port Angeles WA USA
Contact:

Post by classicpenny »

blaisepascal wrote:
classicpenny wrote:The thing that always bothered me about that drawing was the directions of the magnetic fields produced by the coils on the opposite faces of the octahedron: if you apply the right-hand rule to the coils on the opposite faces, the resulting magnetic fields do NOT oppose each other. This seems to me to be a very serious problem - indeed I find it very hard to imagine how a Wiffleball could be produced under these circumstances.
Consider the analogous 2D case: six equally-spaced magnets, 3 pointing in, 3 pointing out, alternating around the center. If you plot the field lines, there is still no field in the center, and the lines go to the adjacent magnets, not to the opposite side.
True. But draw the little magnetic field circles around the each coil cross-section in your proposed 2D case: the fields cancel in the spaces between the magnets. In the same thread I referenced earlier, about Tombo's octahedral design Dr. Mike said, "The easiest way for me to think about coils and fields is to draw circles around the wires. Just like an electric circuit needs a complete loop to work, and magnetic field needs a complete loop to exist. If you make a "cyclone fence" with wires, the magnetic fields will cancel in the small hole between the connecting wires. That's way too big a leak."

Post Reply