Electron interactions with the magnetic field

[MSimon]

MSimon: is there any way we could make a sticky at the top of the "theory" forum to house links to relevant documents, or possibly citations of documents that are not available electronically?

DW: that's what you woul think. Apparently, nobody actually tried Bussard's exact idea, and we don't know enough to bridge the gap from what they did try in order to understand this.

See what you all can make out of this:

http://mr-fusion.hellblazer.com/pdfs/ma ... nement.pdf

I am using IEC Fusion Technology Blog as a repository for pdfs. I can also post papers that are in text form.

http://iecfusiontech.blogspot.com/

I'm not sure about a sticky because some one has to maintain it - delete OT posts etc.

Over at IFT I have full control of the front page and am already maintaining it.

If you have a pdf that should be included just leave a comment there at any post or send me an e-mail.

[spaccemonkey]

Looking at the videos of the single electron simulations, the paths reminded me more of a marble in a bowl. So, would it be fair to say that the whiffle ball is more of an asymptotic potential well where the energy to escape through a cusp is sufficiently high that few electrons have that much energy. (once you know what it is, couldn't you use a Boltzmann distribution to calculate what percentage of the population would have the Ke?) Also, this explains why the magnets have to have a gap between them, the parabolic paths would cause them to collide with the magnet otherwise.

[MSimon]

The actual distribution is non-Maxwellian. However there may be a small Boltzmann distribution on top of a central energy.

Of course there is a lot we don't know.

[TallDave]

The field lines passing closest to the line between two coils will make loops so big that they intersect the wall.

OK, but does it make sense to expect the electrons to follow that path all the way to the wall? My understanding is that they follow the field lines only due to the magnets pushing them at right angles to where they are trying to go: the Magrid, with its large positive charge relative to the walls. Shouldn't they jump off those field lines onto a path that loops more tightly if they're being pushed directly away from the Magrid? Seems there should be competing forces there and just plotting the field lines wouldn't be sufficient to predict electron behavior as they leave cusps.

I think M Simon was alluding to something similar when he talked about "oscillation" for electrons that can't overcome the electrostatic incline to get to the wall. That would suggest the size of this loss could be controlled by the charge on the Magrid relative to the walls.

Again I have to go back to the empirical evidence, such as it is assuming we believe Bussard: if WB-6/7 had the same losses from cusps as WB-5 (with electrons just going into the walls instead of the box) they shouldn't have performed any better.

Hopefully we'll get some data that better resolves this.

When I was at college, we used vacuum tubes with patches of phosphor paint to make the electron paths visible...

Heh, well in the earlier machines we got nice clear black marks on the devices to show us where the electrons were going.

[alexjrgreen]

We can do just as good a job these days with magnetic field measuring eqpt. Or for that matter in a static situation with field simulators.

Once you get a lot of plasma and electron guns etc. things get complicated.

The strange thing about this thread is that it's clear the basics are not agreed on.

Leave aside the ions and the plasma for a moment. Let's have a simple lab model that shows exactly where the electrons go.

[bcglorf]

We can do just as good a job these days with magnetic field measuring eqpt. Or for that matter in a static situation with field simulators.

Once you get a lot of plasma and electron guns etc. things get complicated.

The strange thing about this thread is that it's clear the basics are not agreed on.

Leave aside the ions and the plasma for a moment. Let's have a simple lab model that shows exactly where the electrons go.

That is what Nebel just finished as WB-7, the navy is currently reviewing those results.

[alexjrgreen]

The strange thing about this thread is that it's clear the basics are not agreed on.

Leave aside the ions and the plasma for a moment. Let's have a simple lab model that shows exactly where the electrons go.

That is what Nebel just finished as WB-7, the navy is currently reviewing those results.

That might be over-optimistic, since the Navy just gave EMC2 a small contract to do more research on "Spatially Resolved Plasma Densities/Particle Energies"... I seem to recall the phrase "a simple 9-dimensional optimization problem".

Art spent months on this forum trying to get straight answers about what particles go where. Even with WB7 running like a top in his lab, Rick had a challenge to meet him half way.

What I'm suggesting is something simpler. Something you could exhibit at a Science Fair to show electron interactions with the magnetic field in a polywell...

[MSimon]

There was some work done at MIT on this. Electrostatic confinement only:

http://ssl.mit.edu/publications/theses/ ... Thomas.pdf

http://ssl.mit.edu/publications/theses/ ... chCarl.pdf

and of course fusors.

Also some private work I'm aware of on the Bussard Configuration with a magnetic and electrostatic field.

The one thing they all seem to have in common is that the plasma seems to self oscillate and the oscillation produces "spikes" that oscillate from the central region to outside the grid and back.

Which is why I think recirculation is the wrong term. I have always liked oscillation.

What is required IMO is a test reactor capable of continuous operation (seconds to minutes) where measurements are not difficult and can be done unambiguously.

Which is why I did my BOE LN2 cooled Cu coil BFR design. The big cost for such a design is the coil power supplies and the grid power supplies. With the grid power supplies being by far the largest cost. The reason being that start up current is estimated to be in the 10 to 15 amp range at 50 KV. So to give some margin the design would have to be around 100 KV at 25 Amps. About 2.5 MW net. Probably around 3.5 MW total including losses cooling pumps etc. Coil power would run about 1 MW. With about 1.5 MW total. Both supples would be rampable with either voltage or current control with a useful bandwidth of 1 to 3 KHz (35 KHz switching with standard 1 KV switching modules phased so that the effective switching frequency would be 35KHz * the number of modules with modules able to be seriesed or paralleled.) The supplies would be water cooled with IGFET or MOSFET switching elements depending on trade offs. One standard module for both the coil and HV supplies. The HV supply elements would be immersed in transformer oil for compactness and immediate cooling of the transformer oil. With water cooling of the transformer oil in a secondary loop. Also water or oil cooling of the coil supply depending again on trade offs.

Estimated cost of the whole rig would run from $5 to $25 million depending on engineering costs and the design of custom transformers/cooling loops/testing etc.

The LN2 cooled BFR itself could be run in pulsed mode while the power supplies were being designed and built.

The HV power supply could be reused for a superconducting coil full scale BFR.

[drmike]

The strange thing about this thread is that it's clear the basics are not agreed on.

Leave aside the ions and the plasma for a moment. Let's have a simple lab model that shows exactly where the electrons go.

The problem is that a plasma is a non-linear system. Even if you understand a simple
subset of the problem, you don't know much about the superset of interactions. A current
flow that is reasonably sized will create a magnetic field that bucks the external field. That
will cause particles to move differently and in make the current change direction.
Sometimes that's called turbulence, and sometimes it's called a disruption.

Being able to understand the simple stuff is important, and being able to build on complexity
is really important. That's why a long program of basic research on all forms of fusion
really ought to be in place. We can learn a lot from all kinds of configurations, and then
start to combine them in ways that will be predictable.

But I agree that it'd be really nice to have a system of diagnostics in a lot of experiments
that would tell us a lot of the basic details so there's no guessing at anything.

[KeithChard]

It appears to me that the effects of the B field within the wiffleball are being ignored in the discussion of electron losses. If one looks at Indrek's visualisations of the field lines using the method of images and based on a perfectly spherical wiffleball, it is apparent that the Bfield WITHIN the wiffleball makes a significant contribution to confining the electrons. I concede that these visualisations are only a first approximation to the real situation but they do give a significant indication of an effect which deters most of the electrons from even entering the Bfield outwith the wiffleball.

This confinement will clearly be imperfect but it will have a favourable effect when estimating the numbers of electrons escaping.

[icarus]

It appears to me that the effects of the B field within the wiffleball are being ignored in the discussion of electron losses. If one looks at Indrek's visualisations of the field lines using the method of images and based on a perfectly spherical wiffleball, it is apparent that the Bfield WITHIN the wiffleball makes a significant contribution to confining the electrons.

Keith: thanks for raising this again. It has been my contention from when I first proposed the spherical inversion model that the wiffle-ball internal B-field is the real reason for good electron containment.

While the field far inside the electron wiffle-ball plasma will not look like the simplistic images solution, the field very near the wiffle-ball boundary can not look very much different, just by the continuity requirement that the B-field lines will be tangential to the boundary, on both sides of the boundary.

Furthermore, if you follow the internal field lines along the boundary towards the cusps on the inside then they turn radially inwards and away from the cusps on hyperbolic trajectories. It will only be a very small percentage of electrons that are coming up out of the plasma with very high radial energy and directed exactly along the radial cusp line that escape, most of the particles hanging out near the boundary never even "see" the cusp hole as their energy is contained in azimuthal components of the field by the wiffle-ball surface.

The more spherical the wiffle-ball surface the better will be the electron plasma containment. Also, the harder the plasma pushes back against the magnetic field of the coils (i.e. higher pressure) the better the containment. Of course, it also has more energy to contain so there is probably some optimum condition to be reached here between mag. field strength, electron pressure, rate of cusp losses and power required, etc.

Edit: here's the links to those who need pictures (I know I do) to recall what we're talking about.
viewtopic.php?t=650&postdays=0&postorder=asc&start=15
http://www.mare.ee/indrek/ephi/images.pdf
viewtopic.php?t=650&postdays=0&postorder=asc&start=30

[Art Carlson]

The method of images is a technique to easily calculate the fields associated with certain boundary conditions. There is no implication that the internal fields match those from the images, or even that there are any internal fields at all.

For example, you can use an image charge located in a conducting sphere to calculate the electric field from a point charge near the sphere. The actual field in the sphere is zero.

[MSimon]

Sometimes that's called turbulence, and sometimes it's called a disruption.

I rather prefer the term reaction. Plasma reacts.

[KeithChard]

Art: The image technique is a useful qualitative and quantitative tool. which in this case has given a useful insight into a probable structure for the fields inside the wiffleball. It is most unlikely to be a correct solution as it stands, but it is capable of being usefully elaborated and it has already increased my understanding of the situation, which was pretty much zero before. The exercise of a little common sense should prevent us from falling into the sort of trap you suggest and I am sure that there are plenty of aware members of this forum to keep a check on that sort of thing happening.

[icarus]

Art you said

or even that there are any internal fields at all.

What would be the basis for thinking that there is zero magnetic field inside a structure such as the wiffle-ball?

How would the electron plasma not carry some kind of field along with it? These would actually be some particularly difficult to achieve formation requirements (initial conditions), would it not?