Electron recirculation

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

Moderators: tonybarry, MSimon

hanelyp
Posts: 2257
Joined: Fri Oct 26, 2007 8:50 pm

Post by hanelyp »

I think you need a negative grid between the magid and the direct conversion shell. Otherwise electrons outside the magrid will see the direct conversion shell and zip straight for it.

tombo
Posts: 334
Joined: Fri May 09, 2008 1:10 am
Location: Washington USA

Post by tombo »

The direct conversion anode sure looks like a Van de Graaff generator top end to me.
But instead of putting charges of a few volts at time onto its id, we are putting charges onto the id at 1.25 MV.
What voltage will we see on the outside? Pretty high, I think.

Ummm Wacky Idea:
I wonder if this effect could be reversed for use as a voltage reducer.
The anode might work better if it is convex toward the center of the PW. (like the B field)
That way the alphas would strike the convex surface where its voltage is highest.

Potentially a lower voltage could then be withdrawn from the concave surface.
-Tom Boydston-
"If we knew what we were doing, it wouldn’t be called research, would it?" ~Albert Einstein

93143
Posts: 1134
Joined: Fri Oct 19, 2007 7:51 pm

Post by 93143 »

hanelyp wrote:I think you need a negative grid between the magid and the direct conversion shell. Otherwise electrons outside the magrid will see the direct conversion shell and zip straight for it.
It doesn't matter what the charge on the outer shell is - not to the electrons, anyway. Once they're inside it, they can't see it due to Gauss' Law. The same is true of alphas, of course, which is why you need a trap grid - the core (magrid+wiffleball) is positive and repels alphas, so something has to start attracting them before they hit the wall. There doesn't need to be a huge vacuum chamber - just enough so that the risk of arcing is low at the required voltage.

The only thing the charge on the outer shell does is determine the voltage everything has to float at.

I'm pretty sure I've understood this. Arguments welcome.

I don't thiink a van de Graff generator is a very useful analogy for the direct conversion system. It's more like a particle accelerator in reverse. The current is supplied by the impinging charge, but the power comes from the kinetic energy of the alphas, hence the operating voltage is pretty much fixed by the particle kinetic energy in eV divided by the particle ionization number. The optimum shape is a sphere, to a first approximation (fancy energy-band diversion schemes notwithstanding) because the alphas stream out in all directions equally, and you want them all to be as slow as possible when they finally touch the wall (without, of course, failing to reach it and falling back into the core). If you just let the alphas hit the wall without a decelerator grid, you have a thermal machine.

TheRadicalModerate
Posts: 145
Joined: Thu Oct 04, 2007 4:19 pm
Location: Austin, TX
Contact:

Post by TheRadicalModerate »

93143 wrote:It doesn't matter what the charge on the outer shell is - not to the electrons, anyway. Once they're inside it, they can't see it due to Gauss' Law. The same is true of alphas, of course, which is why you need a trap grid - the core (magrid+wiffleball) is positive and repels alphas, so something has to start attracting them before they hit the wall. There doesn't need to be a huge vacuum chamber - just enough so that the risk of arcing is low at the required voltage.

The only thing the charge on the outer shell does is determine the voltage everything has to float at.

I'm pretty sure I've understood this. Arguments welcome.
Well, you've at least got the electric field from the electrons in the wiffleball. Since we're charging up the magrid to increase the well depth for the ions (we are, right?), I'll assert that the direct conversion charge will also be felt. They're both essentially the same geometry.

93143
Posts: 1134
Joined: Fri Oct 19, 2007 7:51 pm

Post by 93143 »

TheRadicalModerate wrote:Well, you've at least got the electric field from the electrons in the wiffleball. Since we're charging up the magrid to increase the well depth for the ions (we are, right?), I'll assert that the direct conversion charge will also be felt. They're both essentially the same geometry.
The magrid is positive, yes. That's where all the electron energy comes from - you need this to deepen the well, so that electrons don't just sit there right next to the grid. Ions are injected inside the grid and don't see it at all due to Gauss' Law (repulsion from the near side is exactly cancelled by repulsion from the much larger but further away far side - this is a property of an inverse-square force and is the reason a Dyson sphere doesn't have gravity on its inner surface - although with gravity there's no equivalent to charge equalization in a conductor so the enclosure has to be spherical and the Faraday cage effect doesn't work). The ions do see the magnetically-confined ball of electrons (which don't see the magrid charge any more either) and the result is familiar to all of us.

dch24
Posts: 142
Joined: Sat Oct 27, 2007 10:43 pm

Post by dch24 »

93143 wrote:The magrid is positive, yes. That's where all the electron energy comes from - you need this to deepen the well, so that electrons don't just sit there right next to the grid. Ions are injected inside the grid and don't see it at all due to Gauss' Law (repulsion from the near side is exactly cancelled by repulsion from the much larger but further away far side - this is a property of an inverse-square force and is the reason a Dyson sphere doesn't have gravity on its inner surface - although with gravity there's no equivalent to charge equalization in a conductor so the enclosure has to be spherical and the Faraday cage effect doesn't work). The ions do see the magnetically-confined ball of electrons (which don't see the magrid charge any more either) and the result is familiar to all of us.
Let me see if I understand the questions...

1. The ions don't see the positively charged MaGrid (Gauss' Law, sounds good), but they do see the direct conversion sphere further out? What am I missing?

The direct conversion sphere would be able to attract or repel things from the outside, just like the charge on the MaGrid, but inside it has no effect on a charged particle, so it wouldn't do any good as a "particle decelerator" for outbound Helium Alphas.

2. Assuming the direct conversion works kind of like a charged plate, how far away does it need to be? In other words, what are the spacing requirements to prevent electron loss?

The direct conversion grid repels Helium Alphas (thus collecting their energy), but of course we take into account the effect it will have on recirculating electrons. Some electrons will be lost, but I think the percentage can be minimized by making the direct conversion grid large enough. The Helium Alphas will be repelled by the MaGrid (drawing some power from it) but the energy is conserved and will be recaptured when they are slowed by the direct conversion grid. The electrons will be attracted by the MaGrid and the direct conversion grid, so the direct conversion grid needs to be large enough that it only steals a small percentage of the recirculating electrons (specifically, the electrons with a very high velocity as they leak out the cusps). The lower-velocity electrons stay close to the MaGrid and fall back into it.

93143
Posts: 1134
Joined: Fri Oct 19, 2007 7:51 pm

Post by 93143 »

1) They don't see the direct conversion sphere. That's why you need a trap grid, so that once the alpha particles have passed it on the way to the wall, it attracts them and decelerates them to a stop.

Naturally, if any electrons get past the trap grid, they get fired into the outer wall at ~2 MeV.

2) I don't know - you've pretty much summed it up. You'd need to calculate how fast the upper X% of electrons are going when they leave the magrid, and how much distance they would need to turn around in the combined electric and magnetic field from the grid. (I'm not currently equipped to do that calculation.) Then put the trap grid outside that. Make sure there's enough space between the trap grid and the wall to prevent arcing.

This system assumes that all the alphas come off at exactly the same energy, which is not true. We'll have to pull tricks with the trap grid geometry or something. Multiple grids? Magnetic fields?

drmike
Posts: 825
Joined: Sat Jul 14, 2007 11:54 pm
Contact:

Post by drmike »

93143 - you have it right. To create accelerating fields you need multiple grids. What you want to build is a spherical triode. The cathode is in the center, the anode is on the outside and the control grid is in the middle. You can add more control grids if you want.

If you add a control grid that has more negative voltage than the expected majority of electrons from the cathode, you can prevent most of the electron current from reaching the anode.

If you want to do BOE calcs, look up "Child-Langmuir law" of diodes. For example: http://www.answers.com/topic/child-lang ... technology

dch24
Posts: 142
Joined: Sat Oct 27, 2007 10:43 pm

Post by dch24 »

Thanks for the helpful answers! :)

MSimon
Posts: 14333
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:

Post by MSimon »

2. Assuming the direct conversion works kind of like a charged plate, how far away does it need to be? In other words, what are the spacing requirements to prevent electron loss?
That strictly depends on what kinds of field gradients you can maintain before arc over.

I think 10KV/cm is a sure thing. 20KV/cm is possible. Around 30KV/cm you probably are starting to look at some problems. By 50KV/cm steady state you are getting to the edge of what is doable.
Engineering is the art of making what you want from what you can get at a profit.

hanelyp
Posts: 2257
Joined: Fri Oct 26, 2007 8:50 pm

Post by hanelyp »

93143 wrote:
hanelyp wrote:I think you need a negative grid between the magid and the direct conversion shell. Otherwise electrons outside the magrid will see the direct conversion shell and zip straight for it.
It doesn't matter what the charge on the outer shell is - not to the electrons, anyway. Once they're inside it, they can't see it due to Gauss' Law. The same is true of alphas, of course, which is why you need a trap grid - the core (magrid+wiffleball) is positive and repels alphas, so something has to start attracting them before they hit the wall. There doesn't need to be a huge vacuum chamber - just enough so that the risk of arcing is low at the required voltage.
Perhaps I described it a bit sloppy. Without the grid between the magid and the outer shell, the electrons outside will see the magid will see the magrid at a MV or so lower electrical potential than the surrounding and rush for the higher electric potential. A trap grid isn't needed for direct conversion as the alphas would be similarly attracted to the magrid if the trap grid was missing.

A point repeated many times in basic circuit theory class, because so many had a tough time grasping it, is that ground, 0V potential, is a reference point. We could validly assign ground to the alpha collecting grid, the trap grid, the magrid, or any other convenient point in the device. Just so long as everyone concerned knew where ground was assigned. Electric potential is not an absolute, but a relative measure.

tombo
Posts: 334
Joined: Fri May 09, 2008 1:10 am
Location: Washington USA

Post by tombo »

We could validly assign ground to the alpha collecting grid, the trap grid, the magrid, or any other convenient point in the device.
For safety I think of putting the conversion shell and the vacuum tank at V=0 and at earth ground.

For separating the different energy alphas I'm thinking of something like a mas spectrometer.
That would require collimating the alphas coming from each cusp and focusing them say with an electrostatic lens.
That would not be simple and would require 14 of them for the truncube or 32 of them for the trundodecahedron (or the 8x4hedron).
But it would only need to focus one species at about 3 energy levels.
I think this is probably something more for version 2.0
-Tom Boydston-
"If we knew what we were doing, it wouldn’t be called research, would it?" ~Albert Einstein

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

Post by blaisepascal »

tombo wrote:That would require collimating the alphas coming from each cusp and focusing them say with an electrostatic lens.
As I understand it, current theory is that the alphas will have too much energy to be seriously steered by the B-Fields. That's why there's another thread going on about how to shape the coils to present a minimal radial cross-section and thus minimize losses due to alphas slamming into the coils.

tombo
Posts: 334
Joined: Fri May 09, 2008 1:10 am
Location: Washington USA

Post by tombo »

That's why I suggested E fields.
It can be fitted to any coil shape. (I hope)

I'm thinking of starting with electrodes placed in the shadows of all the coils.
An electrostatic lens nominally consists of 3 charged tubes in series with the center one having the opposite charge from the end ones.
No, I don't know if it can even be made to work.
Even if it does I expect it will increase the diameter of the machine to the point where it will only fit in stationary installations. (or in space :wink:)

The crux of the problem is how far can the alphas be bent toward the centerline of the coil with electrodes in the umbra-shadow of the coil.
They need to be bent enough so they won't hit the portion of the electrode that extends into the penumbra.
Only a careful modeling will tell us whether it will work or not.

p.s. A side benefit would be more space to put things without getting them hit by alphas
-Tom Boydston-
"If we knew what we were doing, it wouldn’t be called research, would it?" ~Albert Einstein

drmike
Posts: 825
Joined: Sat Jul 14, 2007 11:54 pm
Contact:

Post by drmike »

Let's hope we have high energy alphas to worry about!!
:wink:

Post Reply