Electron recirculation

[charliem]

What makes you think that just because an ion reaches its apogee (so to speak) at the radius of a particular grid, that it will get picked up and neutralized by that grid?

Two concentric spherical surfaces of different radius and at different potentials create an almost perfect axial E-field.

If we use grids instead of surfaces the field is still quite radial at distances in the mesh oppening order or higher, but at shorter distances the E-field geometry is dependant on the grids geometry.

An example, if we had an alpha leaving the magrid with 2.40 MeV of energy/speed, G2 at 1.18 MV, and G3 at 1,23 MV, that alpha would completely stop between G2 and G3 and then start to fall back. But the local field between G2 and G3 dont have to be radial, so it's possible to make that alpha accelerate not only towards the center of the device but also sideways, til it collide with G2.

Ways to do this, I see two and I suppose there are more. A simple one would be to aproximate G2 and G3 so the field gradient between them were higher than outside this region. Another, it would also help making G2 from ribon (instead of wire) twisting and positioning it so its plane is parallel to the machine axis at every point, that way this grid cross-section -seen from the magrid- would not be very much affected (I think that you proposed something in that line in other thread).

And by the way, good luck tomorrow with your meeting.

Carlos.

[93143]

If we use grids instead of surfaces the field is still quite radial at distances in the mesh oppening order or higher, but at shorter distances the E-field geometry is dependant on the grids geometry.

That's true, but a 98% transparent grid is going to have a lot of very wide spaces in between the wires. I doubt very much that the transverse field would be anywhere near high enough to actually collect the alphas before the stopping voltage hauled them back in. It's just not a big enough target; even if the transverse field was high enough, most of the alphas would still miss.

Another, it would also help making G2 from ribon (instead of wire) twisting and positioning it so its plane is parallel to the machine axis at every point, that way this grid cross-section -seen from the magrid- would not be very much affected (I think that you proposed something in that line in other thread).

I did. I included the idea of radially deep plates with repeller spikes (fat and blunt enough, of course, to prevent arcing) in between them, because intuitively I don't believe that using wire grids will come close to working. Technically I suppose you don't need to have the radially-directed plates extend right to the wall structurally along their entire base length, and if you made them deep enough radially (but how deep is deep enough?) you might dispense with the need for the repeller spike.

This means that with your suggested modification, I think your idea would probably work. If I'm not mistaken it would still require a bigger reactor chamber than mine, but that's essentially due to the absence of the repeller, which is a minor detail. I'm not sure I've adequately thought this through anyway, and I certainly haven't done any math...

And by the way, good luck tomorrow with your meeting. :D

Thanks.