Design FAQ issues - discussion thread

Discuss the technical details of an "open source" community-driven design of a polywell reactor.

Moderators: tonybarry, MSimon

Post Reply
WizWom
Posts: 371
Joined: Fri May 07, 2010 1:00 pm
Location: St Joseph, MO
Contact:

Design FAQ issues - discussion thread

Post by WizWom »

http://www.ohiovr.com/polywell-faq/inde ... =Main_Page

I suggest adding:
The basic construction of a polywell is to make a box will all the faces' being a short length solenoid with the magnetic field pointing "in". Then electrons are shoved into the box. These electrons are the "potential well." Ions which we want to fuse are pushed in next, and are attracted to the electrons electrostatically. This way, containment is separate from ion density, the decoupling which makes this so exciting.

"Cusps" are where the magnetic field MUST escape - try to close them (by, say, truncating the corners) and you just generate more cusps. The vacuum chamber shown in the WB6 picture has barrel shaped devices outside the machine at the corners - these are the cusp electron collectors. They would be the ground side of the electron production guns.

The "line cusps" description is good. Bussard calls them "corner spacing line-like-cusps" in http://emc2fusion.org/RsltsNFnlConclFmI ... 120602.pdf - it might be worth mentioning that so people don't get confused. Every edge will have this going on. Bussard's solution is wonderfully elegant.

The recirculating electrons will follow the magnetic line they are on. This will bring them to a face, not a cusp. All the magnetic field lines going into the box go in through faces, they then come out the cusps.

"Recirculate" means they go an and out of the CORE: the "reactor" would be the whole vacuum chamber. This is a minor quibble, but important. This is why I posed the question about space operations.

SI Magnetic field is Tesla. Tesla unit is kg/A/s/s
Gauss is CGS for Magnetic field. Gauss unit is Mx/cm^2 - 10,000 G=1T
Wandering Kernel of Happiness

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Re: Design FAQ issues - discussion thread

Post by D Tibbets »

WizWom wrote:http://www.ohiovr.com/polywell-faq/inde ... =Main_Page

I suggest adding:
The basic construction of a polywell is to make a box will all the faces' being a short length solenoid with the magnetic field pointing "in". Then electrons are shoved into the box. These electrons are the "potential well." Ions which we want to fuse are pushed in next, and are attracted to the electrons electrostatically. This way, containment is separate from ion density, the decoupling which makes this so exciting.

"Cusps" are where the magnetic field MUST escape - try to close them (by, say, truncating the corners) and you just generate more cusps. The vacuum chamber shown in the WB6 picture has barrel shaped devices outside the machine at the corners - these are the cusp electron collectors. They would be the ground side of the electron production guns.

The "line cusps" description is good. Bussard calls them "corner spacing line-like-cusps" in http://emc2fusion.org/RsltsNFnlConclFmI ... 120602.pdf - it might be worth mentioning that so people don't get confused. Every edge will have this going on. Bussard's solution is wonderfully elegant.

The recirculating electrons will follow the magnetic line they are on. This will bring them to a face, not a cusp. All the magnetic field lines going into the box go in through faces, they then come out the cusps.

"Recirculate" means they go an and out of the CORE: the "reactor" would be the whole vacuum chamber. This is a minor quibble, but important. This is why I posed the question about space operations.

SI Magnetic field is Tesla. Tesla unit is kg/A/s/s
Gauss is CGS for Magnetic field. Gauss unit is Mx/cm^2 - 10,000 G=1T
Ion density is not decoupled from containment per say. The ions are decoupled from the Magnetic containment. The electrons that permit this decoupling are contained magnetically, and this mitigates the disadvantages of trying to magnetically contain ions directly.

The devices outside the magrid are electron guns, not electron collectors. The gas puffer in this area might be considered as an electron collector as I believe it is at ground (?). Undesirable but unavoidable. I believe an ion gun penetrating to the Wiffleball would be a much worse problem.

The relative contribution of the magnetic field lines , the electron- electron collisions and the electron - ion collisions produces a very complex system. Electrons that are dominated by trapping on a magnetic field line will mostly travel along the magnetic field line till they reverse on their own (mirror reflection or bouncing) or they follow the magnetic field line outside the magrid where the potential on the grid reverses them (most of them). As the electrons are bouncing back and forth on the field lines, the original direction quickly become insignificant, so it does not matter whether the North or South poles of the magnets are pointing in, just so long as they all point the same way. The dominate magnetic losses are electrons escaping through cusps- they do not hit the magnets (if the magnet cases are round, etc.) There are other methods of electron transport losses to the magnets but these are not dominate.

The number of cusps that work best depends on how well the corner cusps avoid linear cusp characteristics. It is a trade off, as more faces/ cusps are used the losses per cusp hopefully decreases. It depends on the balance. Also, Bussard mentioned that going to more magnets could increase performance 3-5X. How much of this might be due to better cusp confinement and how much due to improved quasi sphericity of the Wiffleball is unknown.

The "Core" is an arbitrary definition. The magrid interior could be considered the core of the entire machine. But, I believe that Bussard referenced the core as being the central portion of the magrid confined space where the ions are converging to the maximum extent, thus creating beneficial boosts in effective density, well above the average density inside the Wiffleball, and higher likelihood of beam- beam collisions. In this regard a plasma that is thermalized in the radial direction will not exhibit a core, central convergence, virtual anode, etc. The extent of this core is arbitrary. A perfect (and impossible)confluence would increase the fusion rate to almost infinity. No con fluency would result in a fusion rate based on the average density within the Wiffleball, further compromised by significantly fewer effective beam- beam collisions. The target size of the core is based on a compromise between what is desirable and what is obtainable. The existence of a core (convergence) also has significant consequences on mechanisms that effect the thermalization rate of the system.


Dan Tibbets
To error is human... and I'm very human.

WizWom
Posts: 371
Joined: Fri May 07, 2010 1:00 pm
Location: St Joseph, MO
Contact:

Post by WizWom »

As the electrons are bouncing back and forth on the field lines, the original direction quickly become insignificant, so it does not matter whether the North or South poles of the magnets are pointing in, just so long as they all point the same way.
Of course it matters which way the fields point. You want to maximize the area where electrons are pushed in, and minimize the area where they come out. If you switched the magnet directions (by reversing their current) you would invert your areas; the CUSPS would be the confining, and the confinement ratio would invert, too. You'd spend a lot of power, for no effect.

"cusp" is NOT arbitrary. The direction of the magnetic fields is VERY important.

Ions go the opposite direction from electrons - of course you want your Ion injectors to be where the electrons GO.

The "corner cusps avoiding linear cusp characteristics" is nonsense. The entirety of the structure not in the inside of a circle of wire will be where the magnetic field lines come out. They will be evenly spread across the AREA they can come out. It's just the way magnetics works.

As for the electron and ion secondary fields: very important, I agree. The currents in the coils, though, should have the predominant magnetic effect. mostly because the direction of travel of the majority of electrons in the well will be random.

As to grounding to the vacuum vessel wall: it's going to happen. Of course some of the magnetic field lines will intersect the wall; when you design a solenoid, you treat the magnetic field outside as 0 because it can spread over the infinity of space. I'd love to see the math for calculating what portion of the magnetic field will recirculate within a given radius, but I'm not so confident in my integration that I'll attempt it.
Wandering Kernel of Happiness

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Re: Design FAQ issues - discussion thread

Post by KitemanSA »

Love to. Would you be a bit more specific about where?
If you could copy the text from the FAQ into a {quote="FAQ"} box and then edit it into another quote box, that would be WONDERFUL! And thank you for taking an interest. :D

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

Oh and WizWom, would you delete this post from the "sticky"?

WizWom
Posts: 371
Joined: Fri May 07, 2010 1:00 pm
Location: St Joseph, MO
Contact:

Post by WizWom »

Cusp: A hole in the magnetic containment field.
Should change to:
Cusp: places where the magnetic field lines exit the core.
How is magnetic field strength measured- Gauss, Tesla, Amp Turns?
Answer: Gauss or Tesla. Gauss is typically used for small magnetic fields, such as the Earth's, since 10000 Gauss = 1 Tesla
Recirculate: To exit and re-enter the reactor. In the Polywell, electrons that exit a cusp are restrained by the electric field on the MaGrid. Unless they have been upscattered, these electrons do not have enough energy to reach the containment chamber wall so they are returned, usually along the same path they exited, to the reactor core. If they Have been upscattered, the electrons will reach the wall, but in climbing the potential, they will return much of their energy to the system.
change "Unless they have been upscattered, these electrons do not have enough energy to reach the containment chamber wall so they are returned, usually along the same path they exited, to the reactor core."" to "The majority of electrons return, following the field lines, through a face to the reactor core."

The reason for this is that the "upscattering" just kicks them out of the core; their magnetic field line which they cycle along will be random, and independent of their energy (their energy determines the gyro radius, which comes into play in a completely different part of the design).

Add the first paragraph, about the basic design, before the FAQ. You could leave off the exciting bit if it's outside the tone you want.
Wandering Kernel of Happiness

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Post by D Tibbets »

Indeed the electrons in this area may stick to magnetic field lines and recirculate through another cusp so long as there enough room for these magnetic lines to curve around befor hitting the wall. This might require a very large vacuum vessel which would complicate vacuum pumpinmg concerns. Also, upscattered electrons are bad electrons. You want to get rid of them if your energy budget allows it. So ideally you want your magnetic field or some other mechanism to remove these hot electrons. Outsied the magrid, the electrons have their own potential well (the positive charge on the magrid) to climb out of. If they are two fast they never stop and reverse (ignoring various magnetic contributions. With this arguement the electrons recirculate through their own cusps, or they climb outwards along the magnetic field lines untill the are intentionally allowed to intersect the wall. Thias is probably overly simplistic, but again represents one of the compromises that are important to the machine.

Dan Tibbets
To error is human... and I'm very human.

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

WizWom wrote:
Cusp: A hole in the magnetic containment field.
Should change to:
Cusp: places where the magnetic field lines exit the core.
Thank you for the other edits. I will insert them soon. I do however have a concern with your re-definition of "cusp". In the case of a "funny cusp" there is no magnetic field to exit the core. "Line" and "Point" cusps, true, but not "funny". In each case however, a cusp presents a hole in a containment field, I think.

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

OK:

"Recirculate" has been slightly edited:

Some additional lead-in material has been added to the begining of the main page. Included is a link to the wikipedia "Polywell" entry.

The "tesla, gauss" question from the definitions section has been moved and edited with a bunch of stuff from wikipedia.

Check it out and comment si vous please!

WizWom
Posts: 371
Joined: Fri May 07, 2010 1:00 pm
Location: St Joseph, MO
Contact:

Post by WizWom »

You've confused your magnetic and electrical fields.

Magnetic field constrain the electrons. The magrid makes magnetic fields.
The polywell consists of .... Within the MaGrid, magnetic fields confine most of the electrons and those that escape are retained by the electric field
Recirculate: To exit and re-enter the core (MaGrid). In the Polywell, electrons that exit a cusp are restrained by the electric field on the MaGrid.
Just change "electric field" to "magnetic field" both places, and you'll be good.

You see, the electrons are actually pushing apart - because of their electrical sign matching. They have to be pushed just as hard back by the magnetic field. The Magrid itself is neutral - if it carried a charge, it would quickly be neutralized by either the electrons or the ions.

Charges make electric fields just by being. Moving charges make magnetic fields (even permanent magnets, from the moving electrons in atomic orbitals).
Wandering Kernel of Happiness

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Post by KitemanSA »

Sorry to disagree with you but I believe the sentences are correct as stated. First, please be aware that the MaGrid is NOT neutral, not by a LONG shot. It was charged to ~+12kV for WB-6 and comparable values for WB7 AFAIK. WB8.1 may need to get to ~+110kV to drive the pB11 reaction.

INSIDE the MaGrid, it is the B field that constrains the electrons, OUTSIDE it is the multi-kV positive charge of the MaGrid that pulls the elctrons back, not the B field.

If the chamber were big enough and the MaGrid charge were not there, the electrons might travel out and around and back in due to the B field, but they never get that far AFAICT.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Post by D Tibbets »

WizWom wrote:You've confused your magnetic and electrical fields.

Magnetic field constrain the electrons. The magrid makes magnetic fields.
The polywell consists of .... Within the MaGrid, magnetic fields confine most of the electrons and those that escape are retained by the electric field
Recirculate: To exit and re-enter the core (MaGrid). In the Polywell, electrons that exit a cusp are restrained by the electric field on the MaGrid.
Just change "electric field" to "magnetic field" both places, and you'll be good.

You see, the electrons are actually pushing apart - because of their electrical sign matching. They have to be pushed just as hard back by the magnetic field. The Magrid itself is neutral - if it carried a charge, it would quickly be neutralized by either the electrons or the ions.

Charges make electric fields just by being. Moving charges make magnetic fields (even permanent magnets, from the moving electrons in atomic orbitals).
Wrong. The magnetic fields indeed confine the electrons internal to the magrids, except when they escape through cusps where there is a chance of recirculating them, or they are transported across the magnetic fields to impact the magnet casings. In this case they are lost and and can be ignored for the purposes of this discussion. This transport is tolerable for electrons, but presumably not for ions- thus the trick of decoupling of ion containment from the magnetic fields.

Look at Bussard's description of WB6. The electron guns were essentially low voltage car headlight filaments pointed at a cusp, putting out a lot of current, but perhaps only 12 volts. The electrons were then accelerated by a ~ 12000V positive potential on the magrid. The magrid could be at ground with the high voltage on the electron guns, and Bussard said they used both methods in earlier models (probably especially in the closed box machines). But I believe he had concluded that having the magrid at high positive potential had the most advantage, especially with recirculation.
The electrons are injected at ~ 12,000 volts, but due to some inefficiencies, the resultant potential well attains ~ 10,000 volts. It is this potential that pushes the electrons out a cusp. Once outside the electron sees the electric potential on the grid again (Gauss Law considerations) and is reversed and accelerated back through the cusp. If the electron is upscattered above 12,000 eV it will be decelerated by the potential on the magrid, but not stopped, and it will continue on the magnetic field line till it hits the wall (or some other interesting fate associated with the magnetic fields). Charged particles will spiral around magnetic field lines and reverse direction (bounce or mirror reflect) and this is one of the processes that Penning traps use to confine charged particles, but it is inadequate for effective confinement in fusion reactors (why they have tried to use electrostatic repellers, etc to augment the mirror confinement in Penning traps). I have not seen what minor contribution this might have in a Polywell, but I assume it is so far behind the recirculation by electrostatic means, that it is essentially ignored. The clever aspect of the Pollywell recirculation is that it does so without any (most?) of the complications introduced by repellers that sit in the middle of the cusps. Bussard tried to make these work, most recently in WB5, and finally had the eureka moment that he describes in his Google talk.

And the reason the electrons (or ions) do not quickly neutralize on the magrid casings (irregardless or the potential on the casings) is because the magnetic fields act as insulation. This is one reason why the conformal round casings on WB6 was superior to the square casings of WB4. The square corners of WB4 was like a wire where the insulation was worn away. The grids in WB6, WB7, etc. are insulated well against the electrons. Not so well against the ions. That is why (in part) the excess contained electrons are so important. They essentially add another layer of insulation against the ions reaching the magrid casings. In an ideal system (ignoring fuel ion upscattering, and fission produced ions) the magnetic field is completely insignificant to the ions as they are ideally contained purely by electrostatic means.

PS: Again remember Gauss Law consideration. Ignoring high frequency oscillations that might be introduced or evolve in the system, neither electrons nor ions see the potential of the magrid casing. Irregardless of the charge- positive, negative or neutral, the grid is essentially outside the electrostatic universe that the charged particles exist in. The only way the charged particles would hit the casing (from the inside) is due to electrostatic effects of the particles interacting with themselves and/or inertial conciderations. The magnetic fields insulate against these effects. Outside the magrid, the insulation also has to work to insulate against the potential difference between the charged particles and the magrid casing

Dan Tibbets
To error is human... and I'm very human.

Post Reply