Near Spherical Magrid

[rjaypeters]

... all these whacky coil shaped...

I have this faint hope, if they have a minute to spare, the polite folks at EMC2 look at this thread, look at some of the things I have created and ROF and L.

[BenTC]

Actually, a gap of ~ 5-10 mm would be closer to the separation desired based on the several gyroradii of the electrons at perhaps 20-50,000 KeV energies.

What effect does the electric field have on gyroradius?
The formula and description on wikipediaonly mention magnetic field.

[KitemanSA]

Actually, a gap of ~ 5-10 mm would be closer to the separation desired based on the several gyroradii of the electrons at perhaps 20-50,000 KeV energies.

What effect does the electric field have on gyroradius?
The formula and description on wikipediaonly mention magnetic field.

The electric field (kV) causes energy gain (keV) which is related to velocity and the equation has a velocity term in it.

:<math>r_g = \frac{m v_{\perp} }{|q| B}</math>
where
:*<math>r_g \ </math> is the gyroradius,
:*<math>m \ </math> is the mass of the charged particle,
:*<math>v_{\perp}</math> is the velocity component perpendicular to the direction of the magnetic field,
:*<math>q \ </math> is the charge of the particle, and
:*<math>B \ </math> is the constant magnetic field.

[D Tibbets]

Or to put in in a different (more hand waving way).
As the positive potential on the magrid goes up, the entering electrons are accelerated to higher velocities/ kinetic energy. As the velocity goes up, the gyro radius in any given magnetic field goes up. If an electron is exiting (or entering for that matter) a cusp, it is in a magnetic domain (except for the infinitely small border exactly between the two opposing magnetic fields. On either side of this null area the magnetic fields increase so rapidly that it is reasonable to consider some arbitrary value between the center of the cusp and the surface of the magnet and use this as the magnetic field strength that feeds into the gyroradii equation. So in a purely mono energetic system the gyroradii is a specific value (say 1 mm). If the drive potential is doubled, the gyroradii would also double (I'm assuming it is a linear effect)

The need for ~ 3-5 (up to ~ 10 maximum) gyroradii separation at a given drive potential and magnetic field strength is because the system is not pure. There is some allowable upscattering in radial speed. Also there are local electrostatic (electrodynamic?) forces at play, ExB diffusion across the fields, etc. that have to be allowed for if you desire to recirculate almost all of the electrons. A gap of >10 gryrradii may be even better for recirculation (without the electrons brushing against the magnetic coil surfaces), but Bussard mentioned that this is past the breakeven point between recirculation and cusp confinement (not cusp confinement in terms of power loss (recirculation takes care of that), but in terms of the Wiffleball trapping factor/ internal density that needs to be maintained for useful fusion).

Dan Tibbets

[D Tibbets]

I might as well add the speculation that in WB6, the nubs may have been the dominant and limiting surface that initiated crippling arcing. They were not magnetically sihielded to any significant extent, thus they were the closest surface that the electrons could ground on. The work with WB7 may have emphasized this. The significance is that arcing is dependent on the density * potential * distance.
With the nubs shielded or moved to more remote locations (further out) or replaced with electrostatically or magnetically shielded standoffs, the distance to the potential grounding sink could be raised, perhaps to significant amounts. This means that the seperations between the magnets could be widened, so that there would be less electron losses, especially in the vunerable 'funny' / closest approach area cusps. The recirculation efficiency would improve while the cusp confinement would worsen, but be compensated by the improved recirculation before the electron could reach far enough outside the grid to reach an unshielded surface. Essentially Wiffleball trapping would remain ~ the same while input electron losses would fall. This would improve the energy balance, perhaps to the extent that the performance predicted by WB6 might be exceeded by as much as a 1000 fold! Err, well, perhaps closer to 2-3 fold. . Still, that would make a large difference for some applications. As mentioned before though. This may be moot, if machine size is limited by thermal wall loading issues (assuming most of the heating comes from fusion derived heat, not input waste heat). It might not make any useful difference for D-D fusion, but it could help where the maximal net Q is smaller, like with P-B11 fusion. You are not increasing the fusion rate, but you are decreasing the input losses and resultant heating that needs to be overcome before climbing into high Q efficiencies.

Another way to look at it, is that the density (Wiffleball trapping) may fall some, but the drive voltage and dependent fusion crossection could be raised (with out net increase in input power- voltage increased, but current decreased) a disproportionate amount.

Dan Tibbets

[rjaypeters]

Friday night (for me anyway) follies. The Twister, probably impossible to build (and who would want to?):

[KitemanSA]

Friday night (for me anyway) follies. The Twister, probably impossible to build (and who would want to?):

Now if the red and blue go one way (say positive at the north PHYSICAL pole) and the yellow and green go the other (negative at the north pole) this might work. How well remains to be seen, but it seems to meet the requirements (alternating fields around each vertex). Hmmm.

[D Tibbets]

If KitemanSA's view holds, the 'funny' cusps are too numerous. If my view holds, there are too many corner or vertual or near 'point' cusps. The cusp widths in the open rectangle cusps is proportional to the seperation of the magnets (top and bottom), but the length of these cusps is proportional to the magnets on the ends of these regions. The cusp would not be a round or triangular hole, but a rectangle. The system is approaching the configuration of a simple mirror machine, except that you cave converted a single equatorial line cusp into a series of line cusps that add up to the similar area (?). The multiple physical magnet tubes may decrease the cumulative resultant cusp widths, but the many funny cusps would probably push the net effect to a cusp area similar or worse than a mirror machine.

Also, keep in mind that if recirculation can be made good enough, the effective confinement time may approach that of a cuspless machine, where cross field transport comes close to dominating. In this case, the the crossfield transport losses (ExB drift mostly) is proportional to the magnet surface area. A lot of grids is a penalty. Admittedly, this would be a very optimistic scenario for recirculation. The 2008 patent application mentions that a recirculation factor of 10X to 100X might be possible, and that at 100X the effective cusp confinement ~ = the cross field transport. I don't know if this 10-100X improvement uses WB4 or WB6 as the baseline. If WB4 is the baseline, then WB6 is already at the ~ 10X level.

Dan Tibbets

[rjaypeters]

"Meridian" 6 coil:



Notes: The usual dimensions. Red current up, blue current down, if you like.

I've been admiring the simplicity of D Tibbets two-coil designs (after the Friday folly you can understand why) and this popped up.

[happyjack27]

i wrote some code to do mixed particle em-simulation on a GPU. you can see a run here: http://www.youtube.com/watch?v=x5zPjqashAc

next step i'm working on adding a static field. it'll work off a list of 3-d line segments w/ relative current and voltage. so file input format would be like this:

x1,y1,z1, x2,y2,z2, relative_current_density(per meter),relative_voltage_density(per meter)
x1,y1,z1, x2,y2,z2, relative_current_density(per meter),relative_voltage_density(per meter)

the numbers being floating point decimals. current going from point 1 to point 2. the current and voltage are relative to a value you can change in the program, so you can just set them all to 1.0. if you don't have those last two numbers on the line i'll just have the program assume as much.

if there's a way to get what you draw into this format, well it'd be easy enough to just plug in the filename.

[KitemanSA]

"Meridian" 6 coil:

Isn't this essentially Bussard's ram jet?

[rjaypeters]

If you say so. I haven't seen Dr. Bussard's ramjet. I have heard of preferential "openings" for extracting thrust. Is that what you mean?

[KitemanSA]

If you say so. I haven't seen Dr. Bussard's ramjet. I have heard of preferential "openings" for extracting thrust. Is that what you mean?

I think so. If the collected plasma is "slowed" (relative to the ship) and compressed and introduced with little excess velocity into the south end... or maybe not. Hmm. Wait a second... hmm. This needs more thought. Anyone have interesting (and PERTAINANT) comments?

[D Tibbets]

KitemanSA, it sounds like you may be thinking of the Bussard ram scoop. This is basically an open ended (funnel if you like) magnetic field that would collect and concentrate interstellar hydrogen gas from a surface are of many, many square kilometers into a small collection port, there it would then be burned to create thrust much greater than the drag created by the collection process. This was a workaround to avoid carrying the fuel with you. It was not a power source by itself. Later analysis showed this system might work up to ~ 20% of the speed of light, though there might be some tricks that might allow greater speeds. The details of how the hydrogen gas was ionized, so that it could be collected, and how you would manage to fuse it were not obvious. P-P fusion is not a viable option. Perhaps CNO catalyzed fusion might work. Also, I don't know whether a net gain in speed is possible if a fission reactor provided the power and heating of the fuel (reaction mass).

Dan Tibbets

[KitemanSA]

I suspect your ram scoop is the same thing as the widely known ramjet. In wikipedia, ramscoop re-directs to ramjet.

The Bussard ramjet is a theoretical method of spacecraft propulsion proposed in 1960 by the physicist Robert W. Bussard, popularized by Larry Niven in his Known Space series of books, and referred to by Carl Sagan in the television series and book Cosmos.

As I understand it, there is a magnetic scoop, analogous to ram-jet's diffuser, and a "combustion" section and a nozzle. The design above may be appropriate for the "combustion" section. Yes, I know it doesn't ACTUALLY combust!
I am not really sure it makes a whole heep of difference if there are "openings" like those. A standard Polywell ala WB Dodec or something may be the best combustor.
The way I figure it though is to take along solid B and pick up a lot of H along the way. A small part of the H would provide the "p" for the reaction and the rest whould be coolant/ extra reaction mass like the nitrogen in the atmosphere with a ram-jet.