Well, all this playing around is making me appreciate the problems some more! Even when I set the MaGrid voltage to zero, I only get 0.15 microseconds/L (where L is the distance from center of Polywell to center of coil) for confinement.
Pure magnetic and pure electrostatic can't work. I'm willing to bet static combinations can't work either. The whole thing has to be dynamic.
Why can't pure static fields work? Because the charged particles always see a hill to go down and they constantly pick up energy. The phasing has to be right so the "down hill" direction is always towards the center. Sounds like POPS, so I guess I'm on the right track...
Virtual Polywell
The simple magnetic bottle confinemenet in my simulations is poor as well.
Here you can see how the confinement scales with the strength of the magnetic field in my simulations:
http://www.mare.ee/indrek/ephi/confine1/
This is a pure magnetic bottle simulation, no electric fields present.
Comparably at 60K amperes I get 0.43 microseconds/L.
The wiffleball is supposed to dramatically increase the confinement but I'm not seeing that in my current models. It's early days though and I haven't yet perfected my unconventional model. But it might just turn out to be that recirculation is more important than the wiffleball for confinement.
- Indrek
Here you can see how the confinement scales with the strength of the magnetic field in my simulations:
http://www.mare.ee/indrek/ephi/confine1/
This is a pure magnetic bottle simulation, no electric fields present.
Comparably at 60K amperes I get 0.43 microseconds/L.
The wiffleball is supposed to dramatically increase the confinement but I'm not seeing that in my current models. It's early days though and I haven't yet perfected my unconventional model. But it might just turn out to be that recirculation is more important than the wiffleball for confinement.
- Indrek
I was at 200k Amp-turns. I'm including the E field too. I'd say we are in the same ball park, using totally different models. If the order of magnitude is similar, then we are close enough.
I'm pretty sure your calculations are better than mine, I've got too many approximations in the way I handle the data.
I don't see how it would work if the confinement is that short. The whole point of the coil is to keep the electrons around so you have a high potential for the ions. If the electrons leave, you only have the grid, and you are kind of back to the Farnsworth fusor.
I think I'll take a giant step backwards and look at simple physics of a dynamic system. Static external fields can't work given what I see here, so looking at dynamics makes sense. Ball lightning works, can't hurt to go find out why.
I'm pretty sure your calculations are better than mine, I've got too many approximations in the way I handle the data.
I don't see how it would work if the confinement is that short. The whole point of the coil is to keep the electrons around so you have a high potential for the ions. If the electrons leave, you only have the grid, and you are kind of back to the Farnsworth fusor.
I think I'll take a giant step backwards and look at simple physics of a dynamic system. Static external fields can't work given what I see here, so looking at dynamics makes sense. Ball lightning works, can't hurt to go find out why.
I don't think you can see it all happen until you get lots of particles in the system.
We do in fact know that a well forms. This has been known since the 1920s or 1930s. And proven with laser diagnostics by the Japanese in 1999 I think.
If I'm not mistaken it is called the Debye effect. Or the double layer effect.
No magnetic field is needed to get well formation. However you do need ions to balance the charge (quasi neutral).
So if you don't see well formation with particles, it is back to the drawing board time.
BTW have you looked at the news section? WB-7 has produced first plasma.
We do in fact know that a well forms. This has been known since the 1920s or 1930s. And proven with laser diagnostics by the Japanese in 1999 I think.
If I'm not mistaken it is called the Debye effect. Or the double layer effect.
No magnetic field is needed to get well formation. However you do need ions to balance the charge (quasi neutral).
So if you don't see well formation with particles, it is back to the drawing board time.
BTW have you looked at the news section? WB-7 has produced first plasma.
Engineering is the art of making what you want from what you can get at a profit.
As long as they come back it is an inventory problem, not a loss problem.JohnP wrote:I thought you want to have the electrons spend most of their time in the wiffleball. If there's too much leakage, even with recirculation, the ions are going to go somewhere other than towards the middle.What about recirculation? The electrons come straight back in.
Engineering is the art of making what you want from what you can get at a profit.
In my simulations well formation doesn't seem to be a problem. With the quasi-neutral model even, electrons still do spend most of their time in the middle. And if I crank up the electron current then the electrons will indeed form a quite symmetrical sphere.
Still I'm not finished yet so. Will see. Maybe things will be better.
- Indrek
Still I'm not finished yet so. Will see. Maybe things will be better.
- Indrek
If you can keep electrons in the vacuum and away from walls you prevent a lot of materials problems. You can also improve efficiency.
A well for ions means a hill for electrons. You lose energy if you have to keep pushing the electrons back up a hill.
I'm sure that it can work, but I'm worried that the magnetic field levels have to be way to big for effective confinement.
I can't wait to read the reports from WB-7. Hopefully they will write up Bussard's notes along with their report and we can learn a whole lot of stuff we just don't know about.
A well for ions means a hill for electrons. You lose energy if you have to keep pushing the electrons back up a hill.
I'm sure that it can work, but I'm worried that the magnetic field levels have to be way to big for effective confinement.
I can't wait to read the reports from WB-7. Hopefully they will write up Bussard's notes along with their report and we can learn a whole lot of stuff we just don't know about.
You need an electron population inside the magrid to produce the electric potential which confines the fuel ions. Electrons outside the magrid have little effect on this electric well. But so far as electrons circulating outside the magrid are lost through such factors as collision with support structure, it is desired to confine electrons inside with the magnetic well.JohnP wrote:I thought you want to have the electrons spend most of their time in the wiffleball. If there's too much leakage, even with recirculation, the ions are going to go somewhere other than towards the middle.
Indrek, looking at the animation of simulation results, the cusps where coils kiss don't look like they're leaking, something I was a bit concerned about. That being the case, struts between the coils WB6 style look reasonable.
Could we get plots of electrons passing through various cross sections? Plane of a coil and the symmetry planes would be good for plotting where it's relatively clear to place supports.
Could we get plots of electrons passing through various cross sections? Plane of a coil and the symmetry planes would be good for plotting where it's relatively clear to place supports.
Well, I took a giant step back and found a lot of stupid ideas
But it led me to an idea which I haven't seen bandied about, so I tried one that looks pretty good. I simply changed the E field on the MaGrid from static to be cos(w*t). And I went from 37000 steps before loss to 500000 steps and no sign of loss. Here's a plot of the velocity space:
The step size is .005 (unitless time, nu = e*mu0*I/(4*pi*me*L) * time ) and I just multiplied the MaGrid voltage by cos(5*nu). Whether or not we can swing the MaGrid at 50kV +/- is another story, but it sure looks like a great way to help with electron confinement!!
Here is the particle track:
The MaGrids are by definition 1.0 away from center. So an electron with 3% of the energy of the peak of grid voltage (in eV) is clearly trapped in the center. And unlike the staic case, it never gains enough energy to escape.
It's a long way from here to a power plant, but it sure looks good on paper!
But it led me to an idea which I haven't seen bandied about, so I tried one that looks pretty good. I simply changed the E field on the MaGrid from static to be cos(w*t). And I went from 37000 steps before loss to 500000 steps and no sign of loss. Here's a plot of the velocity space:
The step size is .005 (unitless time, nu = e*mu0*I/(4*pi*me*L) * time ) and I just multiplied the MaGrid voltage by cos(5*nu). Whether or not we can swing the MaGrid at 50kV +/- is another story, but it sure looks like a great way to help with electron confinement!!
Here is the particle track:
The MaGrids are by definition 1.0 away from center. So an electron with 3% of the energy of the peak of grid voltage (in eV) is clearly trapped in the center. And unlike the staic case, it never gains enough energy to escape.
It's a long way from here to a power plant, but it sure looks good on paper!
We can swing the MAgrid 50 KV.
DC Bias with RF superimposed. Not a problem. I'm a 50KW RF guy. Eimac Tubes. AM modulation. Give me a tube spec and I can design an RF converter. Diplexers to combine a number of tubes. Class C for 60% to 70% efficiency. Class E and we can go above 90% efficiency. However Class E is fairly narrow band in terms of frequency so I'd start with Class C for experimental work.
Even better once we get this started it should be possible to bleed some RF off the operating machine to feed the grid. After all what we have is a giant radio tube. In such a case it might only take a few KW of pilot power to get things up to speed.
And phase locked to the desired freq. I'm a phase locked loop expert.
Doc. This is really hot. Shows great promise.
DC Bias with RF superimposed. Not a problem. I'm a 50KW RF guy. Eimac Tubes. AM modulation. Give me a tube spec and I can design an RF converter. Diplexers to combine a number of tubes. Class C for 60% to 70% efficiency. Class E and we can go above 90% efficiency. However Class E is fairly narrow band in terms of frequency so I'd start with Class C for experimental work.
Even better once we get this started it should be possible to bleed some RF off the operating machine to feed the grid. After all what we have is a giant radio tube. In such a case it might only take a few KW of pilot power to get things up to speed.
And phase locked to the desired freq. I'm a phase locked loop expert.
Doc. This is really hot. Shows great promise.
Engineering is the art of making what you want from what you can get at a profit.