Does the Gyroradius of an Alpha fit in a cusp?

Discuss how polywell fusion works; share theoretical questions and answers.

Moderators: tonybarry, MSimon

Joseph Chikva
Posts: 2039
Joined: Sat Apr 02, 2011 4:30 am

Post by Joseph Chikva »

rcain wrote:
.. and the electrostatic and magnetic fields produced by(/reflected/translated by) the 2-species inertial plasma itself.
That is not externally applied field.
And where is self-magnetic field in Polywell if not referring the local non-uniformities caused by instabilities?

TallDave
Posts: 3140
Joined: Wed Jul 25, 2007 7:12 pm
Contact:

Post by TallDave »

rcain/ladajo: Noble efforts, I salute you. Keep in mind, electron behavior is stochastic according to Rick, which is probably why he says two-stream is a non-issue.
TallDave wrote: I'm scratching my head thinking what metric they might have used to control the e-current to keep things at or near beta=1 in WB-8. Density interferometry? Just a wild guess...
I mean, considering that Bussard's team was originally using PMT tubes, and just assuming the flash of light was at beta=1 (!).

Also, the well blew out so fast you would think they would need a pretty darn fast control mechanism. I wonder if laser density interferometry would even be fast enough. Maybe we can find some device specs. I should poke MSimon.

Or, maybe they ramp up the e-current slowly. Losses should tighten by 1e3 as they move into the WB regime. That makes me wonder what they would need for power supplies. I'm guessing it's no longer a capacitor bank for what I assume is at least a multi-second run, but my EE fu isn't strong enough to say. And did we ever calculate the inputs on a WB-8? That should be doable...
Last edited by TallDave on Tue Jun 07, 2011 1:28 pm, edited 1 time in total.
n*kBolt*Te = B**2/(2*mu0) and B^.25 loss scaling? Or not so much? Hopefully we'll know soon...

rcain
Posts: 992
Joined: Mon Apr 14, 2008 2:43 pm
Contact:

Post by rcain »

Joseph Chikva wrote:
rcain wrote:
.. and the electrostatic and magnetic fields produced by(/reflected/translated by) the 2-species inertial plasma itself.
That is not externally applied field.
And where is self-magnetic field in Polywell if not referring the local non-uniformities caused by instabilities?
true. and true again.
so there you have it. (nonuniformities Vs uniformities, instabilities Vs stabilities - there's always a flip side, otherwise nothing would change). it seems not to be such a problem so far, in any case. but maybe we will learn more.

Joseph Chikva
Posts: 2039
Joined: Sat Apr 02, 2011 4:30 am

Post by Joseph Chikva »

rcain wrote:
Joseph Chikva wrote:
rcain wrote:
.. and the electrostatic and magnetic fields produced by(/reflected/translated by) the 2-species inertial plasma itself.
That is not externally applied field.
And where is self-magnetic field in Polywell if not referring the local non-uniformities caused by instabilities?
true. and true again.
so there you have it. (nonuniformities Vs uniformities, instabilities Vs stabilities - there's always a flip side, otherwise nothing would change). it seems not to be such a problem so far, in any case. but maybe we will learn more.
Good luck.

ladajo
Posts: 6258
Joined: Thu Sep 17, 2009 11:18 pm
Location: North East Coast

Post by ladajo »

TallDave wrote:rcain/ladajo: Noble efforts, I salute you. Keep in mind, electron behavior is stochastic according to Rick, which is probably why he says two-stream is a non-issue.
TallDave wrote: I'm scratching my head thinking what metric they might have used to control the e-current to keep things at or near beta=1 in WB-8. Density interferometry? Just a wild guess...
I mean, considering that Bussard's team was originally using PMT tubes, and just assuming the flash of light was at beta=1 (!).

Also, the well blew out so fast you would think they would need a pretty darn fast control mechanism. I wonder if laser density interferometry would even be fast enough. Maybe we can find some device specs. I should poke MSimon.

Or, maybe they ramp up the e-current slowly. Losses should tighten by 1e3 as they move into the WB regime. That makes me wonder what they would need for power supplies. I'm guessing it's no longer a capacitor bank for what I assume is at least a multi-second run, but my EE fu isn't strong enough to say. And did we ever calculate the inputs on a WB-8? That should be doable...
My understanding is that they used WB7 and 7.1 to not only confirm WB6 but also to explore and confirm WB conditions. In short, they now have a good handle on formation and mechanisms leading to WB and Beta=1.

Now, hopefully they are working with WB8 and improved Ion injection, well beyond puff gas, which IMHO would result in fast blow outs, vice the ability to dial up a gun or set of guns...

rcain
Posts: 992
Joined: Mon Apr 14, 2008 2:43 pm
Contact:

Post by rcain »

ladajo wrote:
TallDave wrote:rcain/ladajo: Noble efforts, I salute you. Keep in mind, electron behavior is stochastic according to Rick, which is probably why he says two-stream is a non-issue.
TallDave wrote: I'm scratching my head thinking what metric they might have used to control the e-current to keep things at or near beta=1 in WB-8. Density interferometry? Just a wild guess...
I mean, considering that Bussard's team was originally using PMT tubes, and just assuming the flash of light was at beta=1 (!).

Also, the well blew out so fast you would think they would need a pretty darn fast control mechanism. I wonder if laser density interferometry would even be fast enough. Maybe we can find some device specs. I should poke MSimon.

Or, maybe they ramp up the e-current slowly. Losses should tighten by 1e3 as they move into the WB regime. That makes me wonder what they would need for power supplies. I'm guessing it's no longer a capacitor bank for what I assume is at least a multi-second run, but my EE fu isn't strong enough to say. And did we ever calculate the inputs on a WB-8? That should be doable...
My understanding is that they used WB7 and 7.1 to not only confirm WB6 but also to explore and confirm WB conditions. In short, they now have a good handle on formation and mechanisms leading to WB and Beta=1.

Now, hopefully they are working with WB8 and improved Ion injection, well beyond puff gas, which IMHO would result in fast blow outs, vice the ability to dial up a gun or set of guns...
faster, (when/if it blows) because of increased core pressure; guns providing a)greater uphill entry past magrid into the core, b) more accurate control of entry energy and direction, etc.

it will be very interesting to see what stochastic processes emerge - some of HappyJack's simulations give hints i think.

re: control current - back-emf on the magrid coils perhaps?

ladajo
Posts: 6258
Joined: Thu Sep 17, 2009 11:18 pm
Location: North East Coast

Post by ladajo »

Would not the back EMF require the Plasma to have significant relative motion to the coils? I am not sure how that would play given the shape of the plasma.

It is a prickly problem at the least. It may well be better to measure input current of the guns against the loss rates.

Maybe there is a way to measure an axis of plasma diameter and use that as the feedback mechanism.

rcain
Posts: 992
Joined: Mon Apr 14, 2008 2:43 pm
Contact:

Post by rcain »

ladajo wrote:Would not the back EMF require the Plasma to have significant relative motion to the coils? I am not sure how that would play given the shape of the plasma....
- yes, but tiny rate of change should be detectable - filter, rectify, integrate, bridge. will get a good signal out i think; (core shrinks, meter goes one way, core grows meter goes the other way). is what we want i think.

i don't know how superconducting coils perform in such a set up. maybe separate embedded sensing coils (if they are not in harms way).

ladajo
Posts: 6258
Joined: Thu Sep 17, 2009 11:18 pm
Location: North East Coast

Post by ladajo »

rcain wrote:
ladajo wrote:Would not the back EMF require the Plasma to have significant relative motion to the coils? I am not sure how that would play given the shape of the plasma....
- yes, but tiny rate of change should be detectable - filter, rectify, integrate, bridge. will get a good signal out i think; (core shrinks, meter goes one way, core grows meter goes the other way). is what we want i think.

i don't know how superconducting coils perform in such a set up. maybe separate embedded sensing coils (if they are not in harms way).
That raises a good point, regarding the Superconducting coils being in a persistant state. I'll have to think on that.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

rcain wrote: - yes, but tiny rate of change should be detectable - filter, rectify, integrate, bridge. will get a good signal out i think; (core shrinks, meter goes one way, core grows meter goes the other way). is what we want i think.
but the core is rarely perfectly spherical symmetric. it's more of a spiky ball until you "blow it up" (like a balloon). would be better to get readings at different angles and then maybe graph it out on a computer screen. seeing the rough shape of the core in addition to just the size would help to find beta=1. imo.

rcain
Posts: 992
Joined: Mon Apr 14, 2008 2:43 pm
Contact:

Post by rcain »

happyjack27 wrote:
rcain wrote: - yes, but tiny rate of change should be detectable - filter, rectify, integrate, bridge. will get a good signal out i think; (core shrinks, meter goes one way, core grows meter goes the other way). is what we want i think.
but the core is rarely perfectly spherical symmetric. it's more of a spiky ball until you "blow it up" (like a balloon). would be better to get readings at different angles and then maybe graph it out on a computer screen. seeing the rough shape of the core in addition to just the size would help to find beta=1. imo.
yes, thats all possible also i suppose - scanning laser or UV camera + edge detection, (+ Bayesian training on all available data maybe).

there i was trying to think wholesome analogue thoughts, only to be reminded its all digital already, so why bother.

ps. HappyJack - in all your most excellent projections of the Polywell SIM, did you ever produce a simple spectral graph of resonant modes? (ion energy or dB Vs freq)?

pps. sorry, i think this has gone way off topic again.

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

Post by D Tibbets »

TallDave wrote:Dan -- FWIW, it's been strongly implied WB-8 was going to handle continuous operation. So the EMC2 guys probably have a better handle on that by now.

I wonder what they're using for the inputs on that PID loop. It sounded like on WB-6/7, they were just blowing the cusps out every time. I'm scratching my head thinking what metric they might have used to control the e-current to keep things at or near beta=1 in WB-8. Density interferometry? Just a wild guess...

You know what would be great? If we had another spreadsheet for WB-8 like the one someone dug up for the WB-7 project. Anyone remember how we got that, or know if it's available for WB-8? Might tell us some things.
I'm not sure how they will manage maintainence of Beta=1. I'm thinking that the balance has feedback, so long as there is a modest excess elecron current, the cusps will open up slightly(hopefully slightly) as Beta=1 es exceeded, and the excess electrons will spill out . There may be some oscillation about this point, again hopefully slight. I don't know if they can fine tune it with fast responding e-guns, or possibly microwaves to heat (increase KE of the electrons slightly) to dampen the oscillations.

With WB 6, and 7(?), the e-guns were powered by batteries, and the magrid potential driven by capacitor discharge (I don't know if WB7 had a power supply that was powerful enough to forgo the poorly controllable capacitors). For the short durations both the electron current and capacitor driven potential was probably fairly constant.
The B- field was ramped up so that Beta was swept through. I don't know if there was enough transition time for any oscillations to form. Would the transient sweep time very near Beta= 1 be long enough to observe any oscillations, if they exist? If they have any data, it is among a probably boat load of unpublished data (or rather in today's terms - a thumb drive load of data) :wink:

This ossilation may not be all bad. It might be a convenient way to introduce / manage some POPS effects

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

hanelyp
Posts: 2261
Joined: Fri Oct 26, 2007 8:50 pm

Post by hanelyp »

Maintaining beta=1 should be fairly straight forward.

The plasma is a very good conductor. Magnetic fields penetrate good conductors slowly, inducing eddy currents that oppose the magnetic field. The magnetic field penetrating plasma vs. plasma penetrating the magnetic field is a matter of reference frame. As long as more plasma/electrons are injected at the same rate as leakage, the front at beta=1 should remain stationary.

The magnetic field and plasma both being compressible, should they be at unequal pressure at the interface (beta != 1), the interface will shift to equalize pressure. If the interface shifts inward too far, the containment degenerates from wiffleball mode to cusp mode. If it shifts outward too far, you get a plasma blowout.

Inside the interface layer, plasma pressure will be greater. Outside, magnetic pressure will be greater.

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

Post by D Tibbets »

Joseph Chikva wrote:
D Tibbets wrote:I've been talking about looses- charged particles with with their associated KE reaching the magnet surfaces or exiting through a cusp.
And I am talking about that may be that some of those charged particles will have KE more than you estimate.
And by this reason reachable number density will be much less than calculated for condition beta=1.
Of course ions will be upscattered. Annealing is one process to limit this. Another is that as the ions speed up their time to fusion becomes closer to thei thermalizing times- the coulomb collision rate decreases, as the fusion rate increases. This would decrease the proportion of progressively upscattered ions. Also, always remember that when the ions are scattered above the potential well confinement, they are no longer electrostatically confined (actually there is some variability here depending on how deep in the potential well the ions are introduced). They will escape after a few thousand passes when they hit a cusp. This is a energy loss , but not as bad as it might seem. The key of course is how many of the well behaved and upscattering ions fuse before escaping with their upscattered energy and any energy gained from the magrid. My impression is that a few thousand passes may be all that is needed for a good likelyhood of fusion, so few of these more reactive upscattered ions would escape before fusion. So the losses while not zero, is considerably less than losses from electron escape, bremsstrulung, etc.

This is one of the advantages of the Polywell density advantage, the ion lifetime to fusion is much shorter than a Tokamak (perhaps by as much as a million, or a more conservative number of ~ 60,000 which Nebel mentioned). There is much less time for thermalization.

What, you argue that the increased density also accelerates the coulomb collisions? Well, certainly. But, if there is any significant ion confluence (focus) towards the center, many of the coulomb collisions will occur near the dense core, but, that is limited by the dwell time of the fast ions in this portion of the machine. The fast ions will have a longer Coulomb MFP, while the mean path length to fusion will be shorter. Near the edge the fusion rate among the slower ions will be less, and the Coulomb collisions will be greater- much greater as the edge is reached. In this area the coulomb collisions will be at their greatest frequency, but the average temperature will be so much lower than in the core, the unavoidable thermalization will dominate. But, the Maxwell distribution will only be a small (?) fraction of the velocity the ion will gain as it again falls down the potential well. This is 'annealing'.

Concerning electron two stream instability, it is one of at least several influences on beam cohesion and electron thermalization. Again, describing the injected electron streams as beams is a convenient simplification. Even on their first pass the electrons are dispersed by magnetic effects as they pass through the cusps. This is at least in part why only ~80-85% of the drive voltage will be transformed into a potential well. How much radial persistence the electrons maintain during their primary confinement life time is uncertain. Certainly bouncing off of the magnetic field imparts some angular momentum on each pass. Few of the electrons will pass straight through the center, or stop just short of the center and reverse. Most will glance off the central region at some angle. Some feel that the electrons will quickly assume a cloud that concentrates near the Wiffleball border. I'm uncertain, but I think the situation may be some where in between. There is a large current of new or recirculated electron which are initially within perhaps ~ 20 degrees of radial on their first pass. If this current is ~ 1000- 10,000 Amps, that would be ~ 10^22 or 10^23 new or reconditioned electrons per second.

Additionally, with ions present, their greater momentum will result in electrons being dragged towards the center.. In the Google talk Bussard mentioned that the potential well from the electrons was square, which to me means most of the electrons were accumulating near the edge. Once ions were introduced, the dragging resulted in a more parabolic well.
How can electrons near the edge accelerate ions towards the center? Easy, provided that the ions are introduced further out than most of the electrons. Gauss's law takes care of the rest.

Comparing the Polywell to flares on the Suns surface is wrong. These magnetic structures are loops that are concave towards the center. Polywells are claimed to not suffer from this MHD instability due to convex fields towards the center.
This would be a much better match to Tokamaks and their macro instabilities.

Some resorces that describes some of the stability issues are below. I leave it to you to Google them for the actual links.


Collisional Equilibration
Robert W. Bussard, EMC2




The Polywell: A Spherically Convergent
Ion Focus Concept

page 5 of this paper does not give much pertinent infromation to your two stream criticism, but it does demonstrate that such issues were not ignored.

"'Electron instability processes will undoubtedly be an
issue in this device. As the density rises, the wavelengths
for beam instabilities shrink and may play a role in the
electron behavior. If these instabilities lead to electron
thermalization, they may be unimportant to the gross
energetics. If they lead to cavitons and the production
of superthermal electrons, they can increase the elect
tron loss rate. If they lead to turbulence, which decays
by cascade into low frequency turbulence, they can
change the ion angular momentum, which we have
noted will affect the ion convergence radius."




ELECTRON TRANSIT TIME
IN CENTRAL VIRTUAL ANODE WELLSt



Some Physics considerations of the Polywell



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

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

rcain wrote:
happyjack27 wrote:
but the core is rarely perfectly spherical symmetric. it's more of a spiky ball until you "blow it up" (like a balloon). would be better to get readings at different angles and then maybe graph it out on a computer screen. seeing the rough shape of the core in addition to just the size would help to find beta=1. imo.
yes, thats all possible also i suppose - scanning laser or UV camera + edge detection, (+ Bayesian training on all available data maybe).

there i was trying to think wholesome analogue thoughts, only to be reminded its all digital already, so why bother.

ps. HappyJack - in all your most excellent projections of the Polywell SIM, did you ever produce a simple spectral graph of resonant modes? (ion energy or dB Vs freq)?

pps. sorry, i think this has gone way off topic again.
no, i never made a fourier transform view. great idea, though. as always, if anyone wants to take a crack at it, the code's public on sourceforge.

Post Reply