Near Spherical Magrid

[imaginatium]

Over in magrid configuration brainstorming and writing (I think) about the "Cubist" WB:

I thought the idea of squaring out the coils like in that picture was to reduce the "size" of the funny cusps in the corner? It would increase the lengths of the line cusps between coils, but to me it would seem to decrease the overall cusp area, which should thus decrease losses/increase confinement ...

Rotating the filetted pentagons by thirty-six degrees will increasing the size of the semi-triangular regions between three coils, compared to a "standard" dodecahedron with straight torii. Does krenshala's quotation apply here? I though we were trying to reduce the triangular regions?

Give me some time and I'll do some measurements.

Rotating the filetted pentagons by thirty-six degrees will increasing the size of the semi-triangular regions between three coils, compared to a "standard" dodecahedron with straight torii. Does krenshala's quotation apply here? I though we were trying to reduce the triangular regions?

Absolutely not. Well, Icarus seems to be, but that is directly in opposition to what Dr. B. wanted and patented. You need the triangular regions to make a wiffle-ball in a polywell.

The triangular regions can be either real magnets or virtual, but the OUT field is needed.

rjaypeters:

Just to put what kiteman said in different terms; what we are trying to minimize is areas where coil segments are parallel. Those segments are where the line cups appear. Since it is mechanically impossible to completely eliminate any trace of parallel lines, the goal is to
make the line cusp leakage as small as possible, by bringing the parallel lines as close as possible and making them as small as possible. these "squeezed" parallel segments are the "funny cusps". The center of regular polygons, have the most symmetrically balanced fields, this is where point cups occur, even though it is a wider opening.

[KitemanSA]

rjaypeters:

Just to put what kiteman said in different terms; what we are trying to minimize is areas where coil segments are parallel. Those segments are where the line cups appear. Since it is mechanically impossible to completely eliminate any trace of parallel lines, the goal is to
make the line cusp leakage as small as possible, by bringing the parallel lines as close as possible and making them as small as possible. these "squeezed" parallel segments are the "funny cusps". The center of regular polygons, have the most symmetrically balanced fields, this is where point cups occur, even though it is a wider opening.

Yup, and many thanks.

[mattman]

Hello,

New Polywell Post Explaining Rider's theoretical arguements against the Polywell Here:

http://thepolywellblog.blogspot.com/201 ... ument.html

Criticism is Welcomed, will be listened too, and used to improve this post.

[rjaypeters]

I understand a "baseball" plasma device has a coil that looks something like this:


With a confinement volume that looks something like this (in blue):


I further understand the baseball devices won't get us where we want to go. However, what I posted on 20 Aug isn't exactly the same as the images above. See here:


I wonder if the combined magnetic fields of two coils might be beneficial. I have no clue what, if any, the confinement volume would look like.

From my limited reading, the plasma leaks from the wide gaps in the coil. Can the coils to changed to help the situation?

[D Tibbets]

I don't think the baseball stitching works (as per Bussard). Doubling the magnets like you did doesn't help the distorted and elongated point cusps and adds additional line cusps, a net loss.
Your illustration showing a possible Wiffleball border should probably be re drawn to show the borders ~equal distance from a magnet surface. The spurs would be longer. This does show that the containment volume would be far from a sphere, which is bad. The extent of this 'badness' depends on your assumptions. My assumption is that there needs to be fairly symmetrical magnet surfaces on opposite sides of the center. Others think only an average symmetry is needed.

If you take the two bowled grids and place them with the inner curves facing each other, you would have two elongated point cusps, and two 'virtual point cusps', and two 'funny cusps' (one on each end). This is even simpler than the 4 grid arrangement that started this thread. Symmetry would be reasonable. The problem might be (as I visualize it) that the real and virtual point cusps would be just as long, but the width (horizontal plane) would be twice as wide (or more once you consider the inverse square law) as in the 4 grid sample. In this incarnation I see one line cusp, narrowest at the 'funny' near meeting points, then flaring out into the spiked football shape of the 'virtual point cusp'. Compared to a classic opposing magnet mirror machine the equatorial line cusp, instead of being equally wide along the entire circumference, is pinched down at the top and bottom, so the total area of this cusp is reduced (to a useful amount?). The down side is that the true point cusps are widened in one dimension. If there is a net and useful gain in confinement is uncertain.

In the four grid design you have 4 true elongated point cusps, no 'virtual face point ' cusps, and two equatorial line cusps that flare out into a small 'virtual point' cusp at the top and bottom. I think the advantage is that these line cusps, while they are longer, are as narrow as possible, within the limits necessary for clearance of the electron gyroradii, and recirculation. The widening of the cusps on the ends into 'virtual point' cusps (or what I often refer to as corner cusps) is limited due to the less separation of the magnet surfaces in that design. The magnetic field surface description is completely different in these two samples. In the two grid design the 'virtual point' cusps are on two of the sides, while the short 'funny cusps' are on the ends. In the four grid design, tho 'virtual point' cusps are on the ends, while the long (but narrow) line or 'funny' cusps are on the sides.

The tradeoffs of these two types of cusps is the basis of arguments on this thread. The considerations of the 'virtual point ' cusps having loss areas comparable to equivalent true point cusps seems unrealistic. This understanding is of course based on two dimensional visualizations. In three dimensions the 'virtual point' cusps may indeed have some additional (?) pinching of the line cusps that enter them, allowing for the corners to be considered as independent from the line/ funny cusps. As you increase the faces this effect may become maximum (if it exists at all). I do not think it would apply to the two grid design above. The other thing is that the corner 'virtual point' cusps are closer to the magnets, so the fields are stronger than the face centered true point cusps, so this offsets the cusp area in the triangular shaped cusps (using WB6 as an example) that tail off into the line/ funny cusps, so their losses can be comparable to the true point cusps. Again, I think this would not apply to the two grid design above.

Dan Tibbets

[Aero]

You can find information re: baseball stitch like wound magnets here:

http://www.ripplon.com/BiotSavart/examp ... index.html

Includes well depth information.

[rjaypeters]

I was also bothered by the lack of sphericity of the double baseball stitch coils, but I went in a different direction. I added what I call "pinholes":

.

I will think carefully about what D_Tibbets wrote:

If you take the two bowled grids and place them with the inner curves facing each other, you would have two elongated point cusps, and two 'virtual point cusps', and two 'funny cusps' (one on each end).

and create some pictures.

[happyjack27]

I was also bothered by the lack of sphericity of the double baseball stitch coils, but I went in a different direction. I added what I call "pinholes":

that's the route i went in my idea. i took it a step further though. notice if you make it symetric - i.e. equalize the penninsulas (for lack of a better word) you get a pair of conjoined four-leaf clovers. i discovered a somewhat simpler and more symmetric version: conjoined 3-leaf clovers. what i called a "single-coil octahedral magrid" in the brainstorming thread: viewtopic.php?t=289&postdays=0&postorder=asc&start=375

that was actually the motivation for it. it's essentially adding a degree of "geometric tesselation" to the baseball seam idea, thus squeezing out the big line cusps. notice that as you widen the the pennisulas and push out one of them, it morphs seemlessly back into the baseball coil.

(also the picture there shows it with experimental "cusp disruptor" magnets added.)

[rjaypeters]

Still, I wasn't satisfied and extended the pinholes:



So, I think I've got the sphericity concept/requirement covered. But it still looks a little funny.

[rjaypeters]

Then I realized I could connect the extended pinholes and end up with four coils:



And these are seriously bowed magrids.

From the moment I decided to add pinholes to connecting the pinholes was about five seconds of thought...and about a week to model (what with pesky reality intruding).

I'm having difficulty visualizing happyjack27's "single-coil octahedral magrid", so I'll model it after I turn around a coil on the dual baseball stitch.

The other thing I need to do is make sure my next models are 3m diameter with 0.2 diameter coils. The above models were created before I started using standard dimensions.

[DeltaV]

Rjay, which CAD software are you using for your models?

[rjaypeters]

ViaCAD version 7 published by Punch! Software.

http://www.punchcad.com/

[rjaypeters]

If you take the two bowled grids and place them with the inner curves facing each other, you would have two elongated point cusps, and two 'virtual point cusps', and two 'funny cusps' (one on each end).

So, I turned the red coil around and ended up with two unsatisfactory intersections because both coils are mapped onto the same sphere.



Then I created a larger diameter sphere for the red coil:
Is this what you had in mind?

[D Tibbets]

No...

Just draw the original 4 grid bowed magrids like you did initially for this thread, remove two of them and stretch the remaining two to compensate.

Dan Tibbets

[rjaypeters]

Okay.