Can you answer these questions?

[Robthebob]

I'm planning to do my PhD on polywell geometry.

My plan is to test performance improvement in polywells by only changing the geometry. There are at least a couple questions I need answering. I understand that maybe no one in the world can answer some of these questions.

1. To keep the power needed for the magnets the same, if you have more magnets, you'll have to distribute the power appropriately. So you have more magnets but the B field strength of each magnets is weaker, how does this trade off work?

2. How does the geometric factor (a rather silly term, because you can design polywells base on 1 shape in different ways) scale? Is it a constant? How does it change as the size of the machine and/or B field increase? Maybe the terms are coupled? A very basic question to ask actually is how big should I make the cores?

3. Are electron recirculation important even in bigger machines with high fields? (this has to do with design, we know with smaller machines, electron recirculation is a big deal, but if at bigger machines with high fields, electrons that escape from the center of the machine may not be able to return back into the center. If geometry improves electron recirculation and that's the measured numbers actually shows, it may be false hope for bigger machines to expect the same type of improvement)

Also, are there any other features of polywell that may reduce in effectiveness as you go up in size and field strength?

Obviously my experiment can answer all these questions (and then some) if I build a bunch of polywells with varying sizes (4 different order geometries with similar design philosophies x 4 sizes = 16 of them. I'm only one person with no money.... )

4. There is obviously a practical ratio of B field strength to the size of the machine. What is this ratio? (I could just play around with the cube one first before doing anything else)

5. Design tips? Dimensions and their ratios?

6. What kind of geometries to do?

7. how crude of polywells can these be?

Also other questions I havent thought of yet. This is for real. If you can give me a list of all polywell and related papers I should read, that would be awesome.

[hanelyp]

Re: magnet power, Assuming resistive magnets, more magnets while keeping the same magnet power means less power for each magnet. But fewer amp-turns are needed to get the same magnetic field at the center of smaller magnets. But smaller diameter magnets may have a smaller minor diameter to fit the turns.

lots of configuration ideas posted in the magrid configuration brainstorming thread. A few more in an album I have, https://plus.google.com/photos/11157407 ... 1271460865 I can get you the models used to render any of those if the geometry is of interest.

[asdfuogh]

Hey, I was wondering where you're doing your PhD because I'm interested in knowing which university is willing to use funding for Polywell research?

[KitemanSA]

I'm planning to do my PhD on polywell geometry.

My plan is to test performance improvement in polywells by only changing the geometry. There are at least a couple questions I need answering. I understand that maybe no one in the world can answer some of these questions.

1. To keep the power needed for the magnets the same, if you have more magnets, you'll have to distribute the power appropriately. So you have more magnets but the B field strength of each magnets is weaker, how does this trade off work?
Why would you really care? In the final design, the magnet power would be zero anyway (superconductive) and in a research unit, magnet power is easy to increase. Right?
2. How does the geometric factor (a rather silly term, because you can design polywells base on 1 shape in different ways) scale? Is it a constant? How does it change as the size of the machine and/or B field increase? Maybe the terms are coupled? A very basic question to ask actually is how big should I make the cores?
Sorry, not quite sure what you are asking. I am told that, all else being equal, the machine output scales with the cube of the size and the fourth of the B.
3. Are electron recirculation important even in bigger machines with high fields? (this has to do with design, we know with smaller machines, electron recirculation is a big deal, but if at bigger machines with high fields, electrons that escape from the center of the machine may not be able to return back into the center. If geometry improves electron recirculation and that's the measured numbers actually shows, it may be false hope for bigger machines to expect the same type of improvement)
I have always been of the opinion that if the electron is following a field line out, it will follow it back in unless it hits something physical. ICBW.
Also, are there any other features of polywell that may reduce in effectiveness as you go up in size and field strength?

Obviously my experiment can answer all these questions (and then some) if I build a bunch of polywells with varying sizes (4 different order geometries with similar design philosophies x 4 sizes = 16 of them. I'm only one person with no money.... )

4. There is obviously a practical ratio of B field strength to the size of the machine. What is this ratio? (I could just play around with the cube one first before doing anything else)

5. Design tips? Dimensions and their ratios?

6. What kind of geometries to do?
I would not suggest you do the 16 units. Please try to complete what Dr. B. wanted to do before his death. Do a comparison between the WB6 round planform and a square planform cubeoctahedron. Is you can make three units, it would, in my mind, be a toss up between the higher order polyhedron and the bow sided square planform.
7. how crude of polywells can these be?
Just keep all else equal. Also, try to keep the cross-section conformal and try to keep metal out of the cusps as much as possible.
Also other questions I havent thought of yet. This is for real. If you can give me a list of all polywell and related papers I should read, that would be awesome.

If you have not visited the askmar site, do so expeditiously.
http://www.askmar.com/Fusion.html

[Robthebob]

1. I think I still need to apply some sort of standard. I'll be measuring confinement improvement, and I dont want to comparing apples to oranges. The question is really to ask by what standard across 2 different machines should I use?

2. it may scale like that, but I think there are definitely different terms depending on the machine confinement. Hell, machines of different geometries may scale differently.

3. I was just concerned that if they have a hard time getting electrons into the machine, once an electron leaves the machine and tries to recirculate back in, it may be difficult doing that. While mirrors can keep electrons in, they may keep the electrons out. I'm also not convinced that somehow with the WB effect, it keeps the electrons in but doesnt keep the electrons out. I've heard of the funnel analogy, but how does that even work?

6. So just make a cube one and a cubeoctahedron and compare them? (what you mean round planform and square planform and bow sided square planform?

7. crude polywells cost less money, I'm not made of money, that's why I asked.

I also have another question. should I study high beta conditions? I feel like I have to see how different geometries affect the WB effect.

Also, there were some papers I can find on google scholar (by krull and others I think), I'm just wondering if I'm missing important stuff. I think I need to call the sydney group at some point and ask them stuff...

[KitemanSA]

6. So just make a cube one and a cubeoctahedron and compare them? (what you mean round planform and square planform and bow sided square planform?

...

Also, there were some papers I can find on google scholar (by krull and others I think), I'm just wondering if I'm missing important stuff. I think I need to call the sydney group at some point and ask them stuff...

All the units I have proposed are variations on the cubeoctahedron theme except the "higher order polyhedron" I mentioned. The WB6, WB7, & WB8 are all approximate cubeoctahedrons. They all had round planforms, i.e. they had magnets in the shape of toruses, round cross-section and round planform. Trouble is that the round planform messes with the pure cubeoctahedron ideal Polywell shape. Take the toruses and make their major radius circle into a square and you will have square planform magnets. Take the sides of the squares and pull them out in the middle until they lie along the surface of a sphere and you have a bow-sided square planform MaGrid.

As far as I know, ALL important papers are at askmar.com .

[Robthebob]

whoa... um... i dont, on the top of my head, know what field configurations square currents make....

[KitemanSA]

Like round ones but with wrinkles and without the line-like cusps. The question is which effects the results more. Dr. B seemed to think that the elimination of the loss controlling line-like cusp would dominate to the tune of a 3 - 5 times advantage. That is the question. Does it?

If you haven't yet done so, please read Dr. B's Valencia paper. You may also want to contact EMC2 and ask for a copy of their final report on WB-6 which was released by mistake onto the web several years back. Those that were lucky enough to download it before it was removed seem to be honoring EMC2's wishes and treating it as proprietary. But hey, maybe they will let you have a copy.

[mvanwink5]

I believe that report version is back up on EMC2.org. Is this the one?

http://www.emc2fusion.org/2006-9%20IAC%20Paper.pdf

[KitemanSA]

That is the Valencia paper. The report has more detail and data.

[Robthebob]

looks like I've been ignorant and misunderstood some stuff. I need to be able to visualize the fields of a polywell with primary square loop faces.

So to proceed, I can just build 2 of them, one of primary circular loop faces and the other one of square loop faces. While I'm not entirely convinced that this simple square loop face one would have a WB effect in the same fashion as the circular loop face one, it's at least not an entirely different geometric shape.

(So after I properly understood some stuff, I'm pretty unconvinced that going from circular loops to square loops is going to improve performance by 5 times.)

I still want to build a polywell of higher order shape. That answers a bunch of questions actually... do you have a design for a higher order shape that I should try and build?

BTW, even without the support of emc2, I'll proceed...

[hanelyp]

if you want to try higher order shapes, you can start with the basic 12 coil form

In this specific image, the magnet diameters are set so they cover half the sphere, with the image magnets covering the other half.

[KitemanSA]

(So after I properly understood some stuff, I'm pretty unconvinced that going from circular loops to square loops is going to improve performance by 5 times.)...

I would strongly suggest you contact EMC2 and ask if they would allow you, with proper NDAs, to have a copy of the WB6 Final Report. I found it convincing that the potential was there. It has to do with the length of the loss governing line-like cusp between magnets. The square planform has a MUCH shorter cusp length.

PS: 3 to 5 times IIRC.

[KitemanSA]

if you want to try higher order shapes, you can start with the basic 12 coil form

In this specific image, the magnet diameters are set so they cover half the sphere, with the image magnets covering the other half.

If Dr. B. is correct about the square planform improvement, then this should be made with pentagonal planform magnets.

[Robthebob]

the 12 face higher order polywell with pentagon faces will have to be compared to the square face one in order to avoid comparing apples and oranges...

I'll probably start with just the original design, then square faces, and hopefully if it wont be too bad, then 12 face higher order polywell with pentagon faces.

but i feel like my measurements and perhaps comparisons wont be too good for any form of prediction, only whether or not one design is better than the other, not even how much better or worse.

experimental science without theory is like not doing science...