rcain wrote:
just scanned through that 2nd paper: following caught my eye::
http://pop.aip.org/resource/1/phpaen/v1 ... ypassSSO=1{on electron injection}... A fraction of the injected beam with velocities outside the loss cone are reflected from the Polywell™ faces. This idea predicts that there will be a threshold point where a potential well can no longer form since a substantial portion of the injected electron beam is now reflected and no longer enters the device, explaining the observed phenomena.
looks like the same injection issues EMC2/US Navy were dealing with on the FOI request thread. all to be expected.
interesting question: is it possible to make an 'asymmetric' loss-cone? (i'm assuming you'd need multiple coils somehow...)?
I'm not certain, but the question of electron injection efficiency my have different parameters in an ion gun and a gas puff machine. In an ion gun machine the ions are injected while the gas puffer design has the ions created inside (or near the surface of the Wiffleball.). This means the electrons population is generated in different ways. In the gas puffer, one injected hot electron may create through ionization of a neutral gass up to ~ 100 cool electrons (10,000 ev/ 100 eV ratio).
This still leaves the issue of heating the cool secondary electrons. The energy for this comes from the injected hot electrons. Issues of confinement time may be very important in comparison to injection efficiency. Bussard estimated potential well depths 0f appx 80 percent of the injection energy. This may represent the injection inneficiency. A statement that a percentage of the electrons being lost to injection does not mean much with out a stated expected number (1%, 10%, 50%, 99%?).
WB 5 and WB6 comparisons may be enlightening. WB5 was approaching deep potentail wells, but it required exponential increases in electron input currents. How much of the inefficiencies in WB5 was due to input inefficiencies vs confinement inefficiencies is not reported. But in WB6 the improvement in confinement seemed to far outweigh the input requirements, ie: the improved confinement was much more important than input efficiencies. That recirculation efficiency of hot electrons was ~ 90% efficient also implies that under certain conditions this input method at least was highly efficient.
Also, in the patent application, it is given that geometry and e- gun distance outside the cusp is important both for preventing cusp plugging, and to maintain acceptable injection efficiency. It is evident that Bussard, etel considered the issue, whether an engineering solution is in place is another matter.
Also, I suspect the high voltage e- gun injection is different from the low voltage e- guns. With the high voltage e-guns the electrons are at eg- 12000 volts and the beam spreading of the electron beam is dependant on that energy begining from the e-gun surface (or very near it). With the low voltage e-guns the electrostatic/ scattering -beam spreading is less (?), as most of the acceleration comes when the electrons approach the +12000 V magrid surface. And, I am guessing that most of the acceleration occurs while the electrons are deep in the steep portion of the cusp. In effect I see the hieh energy e-gun having a greater spread of the e-beam, and thus more mirroring away from the cusp. With the magrid acceleration of the low energy electrons from a low voltage e-gun, I see the electric fields focusing the electrons deeper into the cusp . In other words the high energy e-guns spray out electrons and only those that hit the cusp nearly head on will enter. With the initially low energy electrons, the electrostatic charge on the magrid surfaces surrounding the cusp does a better job of focusing the electrons towards the center of the cusp before magnetic mirroring comes into play.
This could also be looked at as the high voltage e- guns spraying electrons out in all directions at high energy/ speeds, while the low voltage e-guns also spray the electrons out in all directions but at much slower speeds. The electrostatic charge surrounding the magrid accelerates the electrons towards the center of the cusps. In effect it is the difference between a poorly and a well collimated electron beam. With the high energy electron gun (magrid grounded), a similar effect may be achieved by designing the e- dun to emit a well collimated electron beam focused on the cusp, verses the bruit force, uncollimated e- gun represented by a light bulb hot filament
Knobs would be e-gun design, accelerating potential at the e-gun or the magrid, distance from the cusp, and geometry of the cusp. Also , I believe that a grounded magrid (with high voltage e-guns) might stop escaping highe energy electrons, but if they then recirculated back into the machine, they would be cold ( not re-accelerated by the high + potential on the magrid. For recirculation purposes you need the high positive potential on the magrid. This does not preclude high negative potentials on the s- guns, but does imply that some best compromize may need to be explored.
PS: Concerning corner cusps vs face centered cusps for injection sites may also be significant as mentioned by Kiteman. It seams that I recall corner cusps being the preferred injection site for the e-guns.
The corner cusps are presumably tighter with steeper loss cones and this might favor higher injection efficiencies compared to the face centered point cusps.
Dan Tibbets
To error is human... and I'm very human.
