Talk-Polywell.org

Or, as ladajo named them, "spikeyballs":
http://homepage.univie.ac.at/alexandra. ... dodec.html
http://homepage.univie.ac.at/herwig.hau ... s-2010.pdf

Can you do one with an image magnet representing the plasma? This looks similat to KCDodd's bagshot.

To be fair, I think my now 7 year old son gets credit.

These look good, but what about the cusp in the center of the field? There would seem to be a way to mix two polyhedrals to get the effect.

Good find.

KitemanSA wrote:Can you do one with an image magnet representing the plasma?

These are just exact expressions for surfaces which superficially resemble what one might expect for Polywell surfaces of equal <insert relevant physical parameter here>, not anything based on field solutions. With jiggled constants to approximate (soon to appear?) experimental data they might have some minor utility for rough estimates of <???>.

DeltaV wrote:These are just exact expressions for surfaces which superficially resemble what one might expect for Polywell surfaces of equal <insert relevant physical parameter here>, not anything based on field solutions.

Aw dang!

ladajo wrote:but what about the cusp in the center of the field? There would seem to be a way to mix two polyhedrals to get the effect.

I'm not sure what you mean. The rhombic dodecahedral star seems to capture all 14 point cusps of the 6-coil cube magrid. Do you mean center of the magrid "sphere"?

duplicate post - deleted

Yes it does. But how would you account for coil spacing verses field strength? I was thinking that if you combined two shapes as a merged one, maybe you could do it.

ladajo wrote:Yes it does. But how would you account for coil spacing verses field strength?

Not including the leading "1" term, I count 9 coefficients* in the rhombic dodecahedral star equation. I think those can be tweaked to approximate what you're referring to. In other words, shaping the parts of the surface that are rotationally symmetric about the coil/cusp axes.

If I had time during the next few weeks, I'd code up a Mathematica demo with 9 sliders for the coefficients.

* Or, 14 unique integers making up the rational numbers which make up the 9 coefficients.