Alpha collector geometry idea...

[93143]

I've been wondering if something like a mass spectrometer design could be used to seperate the alphas by energy into three distinct bands aimed at three distinct collectors. So far all my ideas along this line require impractically large vacum vessels, but still, the mass spectrometer seperates particles by mass, and surely a similar idea could seperate same mass particles by velocity ?

That's exactly what I've been trying to do (except I think it's two bands, not three - the low band (two alphas per fusion) has a spread in energy, but the high band (one alpha per fusion) is pretty uniform). I figured I could use lateral electric fields and the magrid shadows to do it in a reasonable amount of space, where using magnetic fields would be more difficult. I'm lazy/busy with research, so I haven't done all that much calculating...

I suppose you could use a parabolic electric reflector to collimate all the alphas into a beam, like in a DFP drive - but then you increase the magrid losses because some of the alphas make more than one pass. Plus you need a very long vacuum chamber (though it doesn't have to be particularly big in the other two dimensions)...

Or you could do as was suggested earlier - chuck the repeller spike and try to collimate the alphas in each gap by shaping the electric field. Somehow that sounds fishy to me, but it's late at night so I can't tell. Besides, that one DOES require a really large (and funny-shaped) vacuum chamber to pull off magnetic separation...

Wasting the extra energy from the high-band alpha as heat isn't all that bad - it's only a little over half the expected magrid loss...

[MSimon]

The collector supports can't be made conducting because you need half a megavolt between the low-band and high-band collectors, and half a megavolt across a conductor constitutes a short circuit. You could make intermediate parts of them conducting, but they wouldn't be invisible - they'd be at some uniform electric potential, which could be useful or not depending on the value of said potential.

The magrid supports can't be made conducting for electron loss reasons, or so I gather. They're insulators so that as electrons hit them they build up a negative charge, keeping other electrons away.

It all depends on what you mean by conducting. I had in mind many gig-ohms. The point is to make the field along the supports mirror the field in the vacuum.

Let me add a bit about vacuum tubes. If the glass wall of the tube manages to get significant trapped charge at a given point, the currents attracted can melt the wall of the tube. So insulators have the problem of uneven trapped charges. Just letting the electrons collect may not be a good idea.

Any way we will see what the real problems are when we build some continuous operation test reactors.

I can hardly wait.

[93143]

It all depends on what you mean by conducting. I had in mind many gig-ohms. The point is to make the field along the supports mirror the field in the vacuum.

Interesting. Possibly difficult to balance if it's subjected to significant alpha flux, but I see what you mean.

Let me add a bit about vacuum tubes. If the glass wall of the tube manages to get significant trapped charge at a given point, the currents attracted can melt the wall of the tube. So insulators have the problem of uneven trapped charges. Just letting the electrons collect may not be a good idea.

I'm just going by what was done with WB-6 - which admittedly was pulsed at fairly low power.

You know, we might want to use direct conversion no matter what reaction we run - both the protons and tritons from D-D fusion are hot enough to escape the wiffleball easily, as are the alphas from D-T. Better yet, D-T only emits alphas at 3.5 MeV (unless I'm mistaken), so you'd probably get around 10% extra thermal efficiency just by using the outer wall as a collector.

On the other hand, maybe we should be thinking divertors here. It's bad if the neutrals from direct conversion just drift back and forth until they happen to hit the intake of a vacuum pump (or, more likely, an unfortunate 20 keV proton in the core). If a configuration could be devised that would sweep/suck the alpha stream around a corner and out of the main vacuum chamber before stopping and neutralizing it, mitigating neutral backwash would be far easier. I'll think about that...

[ravingdave]

It all depends on what you mean by conducting. I had in mind many gig-ohms. The point is to make the field along the supports mirror the field in the vacuum.

Interesting. Possibly difficult to balance if it's subjected to significant alpha flux, but I see what you mean.

Let me add a bit about vacuum tubes. If the glass wall of the tube manages to get significant trapped charge at a given point, the currents attracted can melt the wall of the tube. So insulators have the problem of uneven trapped charges. Just letting the electrons collect may not be a good idea.

I'm just going by what was done with WB-6 - which admittedly was pulsed at fairly low power.

You know, we might want to use direct conversion no matter what reaction we run - both the protons and tritons from D-D fusion are hot enough to escape the wiffleball easily, as are the alphas from D-T. Better yet, D-T only emits alphas at 3.5 MeV (unless I'm mistaken), so you'd probably get around 10% extra thermal efficiency just by using the outer wall as a collector.

On the other hand, maybe we should be thinking divertors here. It's bad if the neutrals from direct conversion just drift back and forth until they happen to hit the intake of a vacuum pump (or, more likely, an unfortunate 20 keV proton in the core). If a configuration could be devised that would sweep/suck the alpha stream around a corner and out of the main vacuum chamber before stopping and neutralizing it, mitigating neutral backwash would be far easier. I'll think about that...

This is a joke, so please don't take it seriously.

Spin the collector sphere. Centrifugal force will sling the neutrals out!


David

[tombo]

OK, I finally got my alpha collector ideas together.
Here is my take on a low-band alpha particle separator/collector.
This is one unit cell of the design.

http://i299.photobucket.com/albums/mm31 ... Alphas.jpg

It shows 2 plates both at a potential just below the low end of the range of the low band alphas.
This potential slows the alphas to allow them to turn faster.
The high band alphas are still traveling very fast and go right on through to the outer wall collector.
One plate is at a higher potential to push the slowed alphas toward the collector plate.
The collector plate is sized to catch the worst case low band alphas. The large 100keV spread of the alphas is a problem and perhaps a show stopper.
The high band alphas are only deflected slightly.
But, they are turned far enough that maybe a quarter of them will impact the collecting plate causing heating problems.
This shows the shadows of the coils (3.2m coil spacing (from the 1.6m radius number), 8” thick coils, 0.33m radius fusion core)
http://i299.photobucket.com/albums/mm31 ... Sheet1.jpg
You can see that the shadows are not big enough to be very useful.
This shows the overall assembly with the shadows poking through. (It is a bit confusing.)
http://i299.photobucket.com/albums/mm31 ... Sheet2.jpg
This shows the overall assembly without the shadows from 3 angles.
http://i299.photobucket.com/albums/mm31 ... Sheet3.jpg
What it shows is a set of 89 conical plates of the same dimensions as the unit cell above. They are on the latitude lines and are aligned with the sphere center and with one pair of coils.
They alternate voltage (1.18 MV & 1.28 MV) and function (repeller & collector).
I can see the fusion core clearly from any angle, but I am only showing 3 angles.
This means that the plates are relatively transparent.

Read the notes on the unit cell drawing.
They say that the concept will basically work but there are show stoppers in the details.
They also describe a possible way to improve it that involves following the alpha paths if they miss the collecting grid.
The saving grace would be if the low band alphas have a much narrower spectrum than the one I found in another thread and used.
Oh well, back to letting them all slam into the outer wall collector and settling for a higher thermal power fraction.

[Betruger]

Tombo, what tools do you use for those drawings? I've never done any, and it's currently not in my curriculum, but I'd like to learn how to on my own.

[tombo]

P.S. The separation plates are close enough to the coils that there will be a significant B field contribution to the alpha motion especially once the E field slows them down.
It may be possible to finesse a better separation by taking advantage of the B fields or they may just make things worse.
My model is purely electrostatic, and has not been checked by anyone else.

Betruger,
I use AutoCAD Inventor which is very powerful and very expensive.
Each of those drawings has many hours into it.
That is why it sometimes takes me weeks between significant postings.
I like it because it is parametric. i.e. I can use a mathematical model in excell to drive the dimensions in my Inventor model.
Off the top of my head I would suggest for a beginner AutoCAD LT, but that's just me as I used straight AutoCAD for years and years so it is second nature to me.

Which CAD program you learn first depends on your goals.

1) Are you trying to develop a marketable skill? Then choose AutoCAD or one of the others that you see in the help wanted ads (pro-engineer, solidworks).
These tend to be expensive but sometimes they have inexpensive student versions.

2) Are you trying to design things on your own, for your own use, on a shoestring budget? (You know, minor little things like fusion generators.)
Then choose one of the free programs (open source is a nice philosophy).
I must admit that I am not conversant with their characteristics because I have and use AutoCAD and don't feel like going up the learning curve of another one without a good reason.

3) Also your goals are important. Are you trying to make pretty pictures or animations or physics models or machine models or shop drawings?

4) There is a short discussion here on another thread that has some recommendations. (try a google site search like "site:talk-polywell.org cad") (cocreate,brl-cad,qcad,blender,... )
(see discussions here viewtopic.php?t=474&sid=5367bff2cb7db85 ... 02b9d41a6c & here viewtopic.php?t=406&sid=d2a7ee8febdf4b5 ... 471c02184e)

5) Your operating system makes a difference too. Are you on windows or linux or mac?
My architect student son recommends Google Sketchup as a starting cad tool that is free, 3D and seems to have a quick learning curve. There are a number of free linux cad's.

I once received a free demo disk called "the monkey wrench conspiracy" that packaged a tutorial with a game and a crippleware version of the cad program they were trying to sell (Think3). It was very similar in operation to Inventor. It doesn't look like they still offer it, but it might be available. I used it to try to entice my little kids into learning a little about cad, but they soon found the bypass to get them back to the 1st person shooter part of it, alas.

So you see even a simple question like this has no simple answer.
And, I'm sure you will get more opinions too.

Conclusion:
Pick one that is quick and easy (like google sketchup), install it, get started making things with it, then move on if you need more capabilities.

[Betruger]

Thanks.. My employer bought me a copy of Solidworks, and I'll probably get to play with Autocad in the computer labs when I get back to school this fall.

[MSimon]

I have used Google Sketchup. It is not bad. It seems weak in dimensioning.

[domoarrigato]

I was looking at the diagram tombo posted which shows extensive shadowing from the magrid and have a (probably) dumb question. Seems to me like the reacting core of a functional unit will be non-zero in size. The shadowing assumption in the diagram shows the rays as all radiating from the exact center. Since the confined area is likely to fill a (hopefully) large part of the area inside the magrid, wont the alphas be coming off at all angles from every point in the core in a frequency proportional to the local well depth (or some other parameter) It will be like a giant porcupine shaped emitter.

I'm even less certain about this next bit, but:

If I'm right about the first part, then that also makes collection of the alphas more difficult, since the geometry of the collection device will have to be some compromise solution that maximizes the conversion of the kinetic energy of the alphas. You can't assume they will all be traveling - or even a large %age - parallel to the electric field of the collector. to the extent that the alphas are traveling off axis, they will be deflected and end up in places they shouldn't - and won't drive the collection current.

[tombo]

Yes you are right.
Somewhere (else?) I posted a picture or link to a picture showing the fully shadowed area for a reacting core diameter posted by MSimon(?).
When you find it you will see that the volume in full shadow is very small and very close to the Magrid.
Yes, it makes the alpha deflector geometry even worse, and is part of why I have pretty much dropped the idea of alpha energy separation.

[eros]

How about magnetic fields. Dos they bend alfas path and how much is effect?

Can field catch alfas circulating wider and make circulation to collide magnets surface?

Maybe talked thing but not yet find that..

[Quaoar]

How about magnetic fields. Dos they bend alfas path and how much is effect?

Can field catch alfas circulating wider and make circulation to collide magnets surface?

Maybe talked thing but not yet find that..

In a future where the HT-superconductor technoligy will be more mature, could be possible to use magrid coils with stronger magnetic fields. In this case, the same fields of the potential well will be used also to screen the coils ad to separate the two bands of alpha, with the low band trajectory confined near the center of the coils and the high band more laterally?

If not, in a more far future, where we will be more confident with plasma windows technology,

http://en.wikipedia.org/wiki/Plasma_window

will be theorically possible to use a plasma window as a grid instead of a solid grid?
In this case can we simply use a plasma double layer to separate the low alpha band plasma grid and the high alpha band plasma grid?
Plasma grids would be also very nice for spacecraft application (I'm not a phisicist but a SF writer): we can imagine and open frame polywell design, where the most part of the waste heat is radiated directly in the space.

If not, it is possible to use another aneutronic fuel like lithium-7 and hydrogen that release two alpha particle of the same energy?

Thanks to all
Quaoar