cusp field on point-like grids and their relevance to IEC

[guest]

Hi,

I've been reading the Bussard's patent on a IEC, but im stuck at getting grip on the following concept: the cusp fields caused by point structures of polyhedral grids tend to acummulate particles of the appropiate charge, but why is this desiderable?

Greetings,

[hellblazer]

It's not desirable, but it is unavoidable. The point cusps are less lossy than other magnetic configurations - you can just minimize and stragegically locate the leaks, you can't eliminate them.

Bussard's magnetic confinement systems confine the electrons, not the ions. The ions are confined by the electrons.

Also, note the cage is held at the negative potential of the electrons, so the confinement is magneto-electrostatic. Not pure magnetic confinement.

[guest]

ok, but why it is not desiderable? from where to where the electrons leak?

[hellblazer]

The electrons leak out the axis of the point cusp. The point cusps look like ice cream cones, where the wide part is at the center of the confinement chamber. The pointy end is at the periphery, along a radius. The cone has a hole in it - it's not a closed system. So, electrons trapped inside the system with the just right momentum go bouncing down the channel and leave the cone through the end.

In Bussard's system, electrons are injected through the point cusps into the confinement system, eventually they leak out through the same cusps. The electrons trapped form a virtual cathode from their space charge. This virtual cathode is what confines the positively charged hydrogen ions.

It's not desirable, as it takes power to replace the electrons lost. Systems that have huge electron losses through the cusps will take a lot of amperage to keep the ions confined. Systems that have low losses through the cusps will take a correspondingly smaller amount of power to keep the same number of ions confined.

In an ideal, lossless system, you could inject the electrons in. They would stay trapped, and you could then keep injecting ions into the system. They would eventually fuse (or leave the system though other losses) and then be replaced by new ions that are injected. The only power input would then be the initial electron charge at the voltage level desired, and then the power needed to ionize and inject the hydrogen ions.

In reality, since there are electron losses out the cusps, you need to continually inject electrons at the desired voltage level. Thus power is continually drawn by the system.

The key is to minimize the power loss.

[3l]

The electrons carry away energy that could be used to heat the plasma and maintain fusion. Almost all the xrays of inertial confiment come from errant electrons that hit the shell of the device making radiation. So the less xray the better your containment is.

Fusion is fun!
Larry Leins
Fusion Tech

[DaveC]

One can get an idea about what cusp fields do and how they are made, by considering a section view of a regular cage type grid. Make the cut plane so that the grid wires are seen in cross-section. Then sketch in the equipotentials. Finally cross the potential lines at right angles to show fieldlines. You will see a channel develops between each section of the grid electrodes. since the grid wires are all the same in potential, field lines go outward toward the adjacent gird wire, and then turn to go radially out of the grid cage. It is along these ray-like field lines that we see the "rays" in the demo fusor modes. There will one ray exiting at the center of each spherical rectangle formed by the rings of the cage.

The rays seem to be a major energy "leak"..in the fusor.

Dave Cooper

[Richard Hull]

Dave is right on about the leaks in the star mode rays.

I always felt it was sad that the very thing that makes the geodesic fusor action so captivating to watch is also its downfall in efficiency. Electron losses in the fusor are rather tremendous.

Richard Hull

[hellblazer]

My understanding is that these are the actual input channels to the fusor. If you check out the research in http://wwwrsphysse.anu.edu.au/~smc112/r ... /paper.pdf , the data seems to imply that these are the modes of oscillation in the ion beams generated by the grid accelleration. Lord knows I could be misinterpreting this paper....

I don't quite understand how these could be power losses, other than due to the ion beams slamming into background neutrals, giving up power to charge exchange.

Just wondering what the power loss mechanism is...

[LehighUBoy]

The points operate as injection and ion loss points, similar to Bussard's injection of electrons for virtual grids.

-garrett

[hellblazer]

I understand that, but I wouldn't think the loss should be the major power drain in the system. I would have thought the loss due to ion impact on the cathode would be far higher. Has anyone quantified the loss through these channels?

[hellblazer]

Thought about this a bit more. Let's say that due to scattering and momentum transfer, there is ion loss though these cusps. The ions would have to have enough energy to hit the fusor chamber and become neutralized in order to leave the system. Otherwise, it will be attracted back into the system. So my admittedly meager understanding is that even though there's a lot of activity on these channels, the primary loss there would be collisions with neutrals, not ions escaping from the system as a power loss.

In Bussard's case, when the electrons leak through the cusp, they are repelled by the negative field on the structure and attracted to the ground potential of the chamber, so they represent a clear power loss.

I just don't see the same mechanism in the normal IEC case, unless the ions have enough energy to hit the chamber wall before being attracted back into the center.

Again, just trying to understand... Sorry if these are rather foolish questions.

[Richard Hull]

The absolute greatest power loss in the system are electron losses. The ion impact on the cathode is one the most MINIMAL of all system losses!!!!!! The middle loss would most likely be ionization and neutral collisional losses.
Electrons are created throughout the fusor volume, end of statement. Those that to not create ions via collision in that volume WILL crash into the outer shell! END of statement. All (100%) of the ions created in the volume WILL head towards the inner grid or cathode and only a small fraction will impact the cathode. Some will crash into neutrals.

The fusor bleeds to death by electron losses.

Richard Hull

[hellblazer]

Ah! Many thanks. I was neglecting the negative half of the equation.

[3l]

Don't worry Hal:

There have been many a soul tortured by the pesty electron.....Plenty more anguish to go!
Hey but you know this up front now.
Academics with fancy diplomas lining the wall have done no better.... We may be rabble in arms but we didn't promise that this fusion stuff would only take ten years and 10 billion....no wait a minute 15 billion...
Kids in college already? Ok make it twenty billion.

Larry Leins
Fusion Tech

[hellblazer]

This has been nothing but a great experience. Learning a lot and having a great time. Many thanks to all... Particularly the rabble in arms.

[DaveC]

Not to belabor a point here, but may I emphasize that just little bit of electrostatics goes quite a long way towards explaining the basic behavior of the simple fusor.

We can with great impunity ignore collisions with neutrals, ions and etc, since we are working at pressures low enough that mean free paths are in cm to meters. (Can't ignore them forever, since we ARE trying to arrange for at least the Deuterium ions to collide.)

Thus we can , (and everyone should do it at least once), make simple potential plots by hand around a scale drawing or sketch of the fusor grids. If you are good in 3D art, by all means do it in 3D. All you need to get started is the fact that equipotentials are always parallel to the surface of a conductor. So we can sketch in some equipotential lines tightly around grid wires, and then smoothly draw others with larger radii. When the potential lines are about to intersect or overlap others, we just make a continuous larger curve letting the shape develop from the combined contours. It is not really important how accurate these are, just that we see the general shapes.

Next create the field lines by drawing so they intersect potential contour lines at right angles. These are the lines of electrostatic force that move ions one way and electrons the other. If you are the slightest bit accurate in your sketch, you will begin to see the field structure that gives the "rays" their shape.

If you play around a little bit, with fewer or more electrode rings, you can see what happens to the fields. The more cathode rings, the less defined the rays become, because the voltage differences are reduced. Fewer rings and the voltage differences are greater, hence stronger rays.

One thought that I have been mulling over, is to redirect the electrons/ions in the rays, so that they return (via the next adjacent ray ??) to the center. Something like the recirculation that Farnsworth envisioned. In principle this could be done either electrostatically, or magnetically. Magnetic bending would not perturb the static grid fields, whereas electrostatic deflection out near the fusor's outer shell, wil distort the grid fields somewhat.

The extent to which this can be done directly affects the net energy input to the fusor. You need one deflector for each ray., but the required magnetic fields are not too large, (a few hundred gauss max. to bend electrons). But the ions won't be bent much at all, so this would separate ions from electrons.

Again, take the time to do some simple analysis, by hand at least, and if you have access to either a 2D or 3D electrostatic program (check the local university), you may be able to get some time to do a real computer model in much more detail. It will help enormously in understanding what we arre trying to do in the Fusor.

Dave Cooper.

[hellblazer]

For those looking for some FREE electrostatics modelling programs:

EStat 5.0 from FieldPrecision
http://www.fieldp.com/estat/estat.html

CPO-2D and 3D
http://www.electronoptics.com/downloads.htm

These programs have limits in the free versions, but you can easily model subsections to get the feel for what the fields look like.

[roktengr]

You may also want to check out this program:
http://www.ansoft.com/maxwellsv/
It is free and is very easy to learn. It does (in 2D) both magnetic and electrostatic fields with material effects if you wish. It's the best I have found for the price!

I just need a good plasma analysis or partcle trajectory add-on to this. I looked at Vector Fields and Integrated Engineering (Lorentz), but they both are very expensive and don't distinguish between big companies and one-man shops. I did not care for Field Precision since it did not import 3D. I would be willing to pay $5000 for 'point and click' analysis package that imports 3D geometry (I can export almost anything from ProEngineer.)

If anyone knows how to get one of these cheap, let me know.
Jim

[hellblazer]

Yea, the biggest problem in this field is the construction UI. If you can manage to model the problem with their 3D or 2D tools, then it's fine. What they do have is incredibly non-intuitive and downright painful. One would think with all the work that has gone on in CAD that they would have integrated with some 3D interchange format like STL or even openGL or something...

[DaveC]

Here's two more programs that MAY be of interest.

There is the Classic 2D modeler POISSON at the Los Alamos Accelerator Group Site. It is also not too user friendly in setting up the electrodes. Need to input the coordinates of the electrode system, which is tedious and error prone. But Poisson is FREE (your tax dollars at work) and has an excellent Physics course in Electrostatics and Magnetostatics and a lot of EM and specialized applications basics. Good reference stuff. James Billen maintains the program and it has had some significant upgrades. One project on their books is to provide some type of CAD interface to ease the setup.

The second program is not free, but quite reasonable. It is SimIon 7.0 Available at the Idaho National Resesearch labs site. or just do Google on SimION.
If you are an educator it is (or was) available at about a 90% discount. Full price is around $495 US. The discount price is about $50. I use this and it is fairly intuitive.. it is very good once you get the potential arrays set up. I used this program to track a single electron around the interior of the Fusor for something like 250 loops, before it collided with the walls.

The ability to import a CAD file is why you pay the big money, sad to say..

Dave Cooper

[LehighUBoy]

Dave Cooper wrote:
> The second program is not free, but quite reasonable. It is SimIon 7.0 Available at the Idaho National Resesearch labs site. or just do Google on SimION.
> If you are an educator it is (or was) available at about a 90% discount. Full price is around $495 US. The discount price is about $50. I use this and it is fairly intuitive.. it is very good once you get the potential arrays set up. I used this program to track a single electron around the interior of the Fusor for something like 250 loops, before it collided with the walls.

Are you willing to post the SimION file that modeled an electron in a Fusor?

-garrett

[DaveC]

Sure -be glad to post it ... I wanted to do it earlier.. but .. was not sure how to export it.. Was much easier than I thought.

Will do that in the near future.

Dave Cooper.