Magnetically Shielded Fusor Grids--Why Won't This Work?

[chrismb]

So, to summarize the last 3 pages of this melee, we have three fundamental sources of loss in a fusor:

1) Fuel ions thermalizing.

2) Grid discharges to the wall.

3) Ion collisions with the grid.

Of these, chrismb is putting up a spirited defense that #3 doesn't swamp the other 2 sources of loss, and therefore magnetically shielding the grid doesn't help that much.

Yes. exactly. Also to add to the list - 4) sputtered anode (shell/wall) contaminants will suck up 'the juice' as they are born right at the top of the potential well, 5) 'fast' neutrals [not necessarily when the ions are going fastest, but as they slow down], formed by the re-combination of ions with electrons, plus associated recombination EM radiation. 6) formation of fast charged molecules, that won't move quick enough for fusion and so lower efficiency.

I would've subdivided into losses due to the central plasmoid, and other ionisation issues, but your categories of 1 and 2 are good general descriptors.

So, two questions (well, rhetorical questions, with argument attached):

First, chrismb, you mentioned that the fueld ions thermalize "with the background". What background? As far as I can tell, the only background consists of monoenergetic ions, discharing electrons, and the grid. So the only source of thermalization ought to be from coulomb scattering near the center of the device. I would further suspect that thermalization would be considerably reduced from a single-species D-D reaction, since you're unlikely to get big momentum transfers when everything's the same mass.

Second, can't we minimize electron discharge by a) pre-ionizing the ions, rather than relying on corona effects, and b) simply making the vacuum chamber bigger?

Background media; as Art says for polywell-type, but fusors operate in the units to tens of microns range. The fusion output actually comes from fast ions colliding with the background nucleii. The fast-fast ion collisions at the centre are so rare that they account for millionths of the actual neutron output in a fusor. A couple of the amateur guys are trying to change that scenario by creating continuous ion sources rather than relying on the discharge regime of a typical 'two elecrode' fusor. As the vacuum is pumped down, the power delivery becomes more and more erratic because of the smaller numbers of viable ions in the reaction chamber, so you've got to rely on some other way of forming ions. All good ideas...we've yet to see fusion as a result.

Pre-ionisation; if possible, this could be a good idea. But the fusor relies on a static electric field to accelerate ions, and if the ion density is too high then the efields will become screened by the resultant plasma.

Hence, for a standard two-electrode fusor, upper and lower pressures are pretty much defined for you.

What magnetic fields could do for you is limit what you describe as 'Grid discharges to the wall'. A given magnetic field will cause electrons to drift sideways in a ExB drift (and wouldn't get to the wall if the 'critical magentron condition' exists) but won't affect the heavier ions which'd be free to carry on their then-slightly-elliptical reciprocations.

[D Tibbets]

[quote="chrismb]

.... , but fusors operate in the units to tens of microns range. The fusion output actually comes from fast ions colliding with the background nucleii. The fast-fast ion collisions at the centre are so rare that they account for millionths of the actual neutron output in a fusor. A couple of the amateur guys are trying to change that scenario by creating continuous ion sources rather than relying on the discharge regime of a typical 'two elecrode' fusor. As the vacuum is pumped down, the power delivery becomes more and more erratic because of the smaller numbers of viable ions in the reaction chamber, so you've got to rely on some other way of forming ions. All good ideas...we've yet to see fusion as a result.

Pre-ionisation; if possible, this could be a good idea. But the fusor relies on a static electric field to accelerate ions, and if the ion density is too high then the efields will become screened by the resultant plasma.

Hence, for a standard two-electrode fusor, upper and lower pressures are pretty much defined for you.

What magnetic fields could do for you is limit what you describe as 'Grid discharges to the wall'. A given magnetic field will cause electrons to drift sideways in a ExB drift (and wouldn't get to the wall if the 'critical magentron condition' exists) but won't affect the heavier ions which'd be free to carry on their then-slightly-elliptical reciprocations.

Actually, in gridded fusors, the gas in the chamber is ionized to a miner extent (perhaps 1%?). So the dominate target (other than structure) for any fast ions would be a neutral gas molecule. But those collisions dominating by a factor of millions? I don't know of any report that substantiates this.
I don't know if any amatuer fusors have used ion guns in a neutron producing fusor, but Hirsch used them in his fusor that produced ~ 10e10 fusions per second at ~ 120,000 volts. I don't know if any other profesionals like at the Univ. of Wisconsin , etc. have done it .

As Bussard has stated the Polywell has to be dominated by ions by a factor of at least a thousand to one. Or, at least that is my interpretation of claims of needing > 1000:1 inside to outside ratios of fusion fuel to achieve net power production while avoiding system killing arcing. Any persisting neutral gas (if a gas puffer- not being ionized by the electric field and high speed electrons/secondary electrons on the first pass, if an ion gun then only recombined gas molecules would exist beyond the background level), or recombined neutrals would ignor the magnetic fields and would diffuse evenly throughout the chamber (ie- if persistant, neutrals will have a 1:1 ratio inside and outside the magrid).


Dan Tibbets

[D Tibbets]

A couple (my) impressions taken from the last few pages.
Comparing the performance of a gridded fusor that lasted 1 millisecond to a Polywell is misleading. The only way it could be compared is in the timeframe of the surviving ions. If the ions are traveling an average speed of ~ 100 million cm per second and survives for ~ 100 passes befor hitting the grid, the lifetime in a gridded fusor would be eg:~ 10 microseconds. In the same size Polywell the life time might be as high as 10 milliseconds so a direct comparison could only be valid in timescales of a few microseconds. Another way to look at it, is in a an equivalent pulsing mode- the gridded fusor could provide a 10 microsecnd fusion pulse and repeat every 10 milliseconds (or repeat every 10 microseconds at 1000 times the input power). The Polywell fusion pulse would be 10 milliseconds long, repeating every 10 milliseconds. In this comparison (all other things being equal) the Polywell would have a 1000 fold improvement in power balance. Fusion rate is dependent on cross section, MFP, survival time,and density of reacting particles. The Polywell has a thousand fold advantage in survival time (confinement time). The density is also increased by a significant amount. I don't know how much better confluence or focusing is improved, but the lack of significant neutrals that would be moving/scattering randomely, and a physical grid to intercept the ions and provide polluting sputtered nonfuel elements can only help the initial and sustained radial directions of the ions. I don't believe anyone could reasonably deny this. Of course wether this prolongs thermalization to any Signficant degree is a different matter. And, this is before any arguments about anneling at the edges.

If you take chrisMB's claim that electrons streaming to the surface loses 60 times as much energy as ion's hitting the grid, then magnetic confinement of the electrons yields another ~ 60 fold improvement in efficiency. And this would be in the thermalized model that he is argueing for.
Being generous, this adds up to to ~ 100,000 fold improvement in the Polywell. The rest (~ 10,000X)of the needed improvement in efficiency would need to come from the B4 R3 scaling.

A comparison of the Polywell (WB6) and a gridded fusor at 10,000 volts:
WB6 yielded ~ 500,000,000 neutrons/s.
Gridded Farnsworth Hirsch type Fusor yields ~ 10,000 neutrons/s *
The ratio is similar to my derived number.

* My guess for gridded fusor output is based on ballpark numbers obtained by amatuer fusioneers at ~ 10-12,000 volts.

[EDIT] The above discussion doesn't account for the increased current input of the WB6, but it should only shift the baseline numbers, not effect the rates.
Also, my feeble understanding of Rider's assesment of the shortcommings of IEC fusors is based mostly on bremsstrahlung radiation. A different issue which has been addressed by Bussard (again there is debate wether this has been successfully addressed).

An example of fusor output at these lower voltages-

http://www.fusor.net/board/view.php?bn= ... ction&st=0


Dan Tibbets

[rcain]

Heh, did you see the chart on p53? Ten passes! Ouch.

Hi TallDave

Good link (http://ssl.mit.edu/publications/theses/ ... Thomas.pdf) - thanks for that.

I havent finished reading it yet, but lots of juicy theory, for (hard) gridded devices at any rate. Very interesting analysis of multi-grid and ion focussing - ie. what we have been discussing on other threads about ion-core vs ion-background vs beam-beam collisions. populaton 'bunching' very interesting.

If I read correctly the 10 passes in the graph the author 'artificially' imposes here since it approximates to single grid device with a single 'grid-stalk' soaking them all up from circulation in short measure. inclusion of symetrical 'dummy stalk' improves matters greatly (so his theory suggests).

this was written 2007 and suggests much experimental measurement that could/should be undertaken. would be nice to see. anyone got a spare fusor?

[TallDave]

Yeah, it's a good general background as well. You can see pretty easily the many barriers to fusors as reactors: grid losses, space charge density limits, thermalization, etc.

It's sort of amazing a 10-pass fusor could produce any detectable fusion at all.

[D Tibbets]

Heh, did you see the chart on p53? Ten passes! Ouch.

Hi TallDave

Good link (http://ssl.mit.edu/publications/theses/ ... Thomas.pdf) - thanks for that.

I havent finished reading it yet, but lots of juicy theory, for (hard) gridded devices at any rate. Very interesting analysis of multi-grid and ion focussing - ie. what we have been discussing on other threads about ion-core vs ion-background vs beam-beam collisions. populaton 'bunching' very interesting. ) .

If I read correctly the 10 passes in the graph the author 'artificially' imposes here since it approximates to single grid device with a single 'grid-stalk' soaking them all up from circulation in short measure. inclusion of symetrical 'dummy stalk' improves matters greatly (so his theory suggests).

this was written 2007 and suggests much experimental measurement that could/should be undertaken. would be nice to see. anyone got a spare fusor?

While most amatuer fusors operate at Q's in the region of1/1billion, on page 21 the author of the link above mentions that Hirsch's machine was ~1/10,000. This would serve as a much more forgiving comparison of how much improvement a Pollywell needs to reach a Q of 1 compared to the best gridded fusor. Keep in mind that Hirsch's fusor was only perhaps 20-30 cm wide, so even without magnetic scaling, and only radius scaling the Hirsch fusor could reach break even (perhaps even before it reached ITER size) If The X-Ray Losses Could Be Controlled, and that pesky real cathode could survive.
This ignors the electron loss scaling which would probably be larger than in a Polywell. So perhaps the Hirsch fusor would need to grow to several times ITER's size to reach break even (again, conviently ignoring Rider's paper that emphasizes the x- ray losses preventing this). There are no slow electrons in the center of a Hirsch machine, like there are susposed to be in the Polywell. It might be better to say an Elmor Tuck type gridded fusor might reach break even at very large scales as it should have similar central electron behavior to a Polywell (except for lousy electron lifetimes/containment).


Dan Tibbets

[TheRadicalModerate]

Why do fusors have to run a higher pressure? Why can't they run at low enough pressures that there's very little ion background?

[chrismb]

Why do fusors have to run a higher pressure? Why can't they run at low enough pressures that there's very little ion background?

'higer pressure' is relative. They operate in what is technically termed 'medium vacuum' for the simple reason that the basic two electrode fusor is just like a neon light, and if the pressure gets too low, so no discharge occurs. The current of a power supply into a fusor becomes progressively more and more unstable, until it becomes very difficult to maintain any semblance of 'control'. As I mentioned before, some fusioneers have been progressing alternative ion forming techniques to press further into lower pressures, but bear in mind that they will be relying on fast-fast ion fusions, rather than ion-background fusions, and there are precious little fast-fast in evidence with existing experiments. So, just like Polywell, they're hunting a dream of a spherical focus at the centre which, in my reading of results, has never been yet demonstrated.

[D Tibbets]

Why do fusors have to run a higher pressure? Why can't they run at low enough pressures that there's very little ion background?

'higer pressure' is relative. They operate in what is technically termed 'medium vacuum' for the simple reason that the basic two electrode fusor is just like a neon light, and if the pressure gets too low, so no discharge occurs. The current of a power supply into a fusor becomes progressively more and more unstable, until it becomes very difficult to maintain any semblance of 'control'. As I mentioned before, some fusioneers have been progressing alternative ion forming techniques to press further into lower pressures, but bear in mind that they will be relying on fast-fast ion fusions, rather than ion-background fusions, and there are precious little fast-fast in evidence with existing experiments. So, just like Polywell, they're hunting a dream of a spherical focus at the centre which, in my reading of results, has never been yet demonstrated.

While it is not Earth shaking, this reference claims convergence in a D-D gridded fusor with 10% of all fusions occuring in the core. I'm guessing that thier defined core makes up ~ 1 % or less (ball park figure) of the total volume, which would mean the fusion rate in this core would be 10 times higher than the average throughout the fusor. This is in a fusor with a grid in the way, and lots of neutrals. In a machine without a physical grid, and rare neutrals*, I reason that this converged central fusion rate, as a percentage of the total (whigh I assume is predominatly fast-fast ion collisions in this example), would be higher, perhaps much higher in a Polywell type machine.

http://iec.neep.wisc.edu/usjapan/UW-IEC ... ashley.pdf

*As stated by Bussard,rare neutrals compared to ions is a prerequsite for the Polywell maintaining adiquite density inside for net fusion power while preventing arcing outside.

Dan Tibbets


Dan Tibbets

[TallDave]

The B^4 scaling only applies to beam-beam collisions, so power scaling in WB-8 could tell us a lot in that regard.

We know there is a well from the laser fluouroscopy, so there's reason to think there's focus.

[MSimon]

Dan T.,

Thanks for that. Added to IECF sidebar.

[chrismb]

I agree this work demonstrates some fast-fast fusion.

I do not understand how they concluded that 10% were fast-fast, however, when the window of the chamber they were looking at only covers 5% of the reaction volume!?

It requires to also know the total reaction from the remaining 95% of the reaction volume to derive the percentage. If the remainder of the reaction volume produces 95% of all the reactions, then you get a very different answer!!

Of course, the proton detector is also looking 'along a beam' and if a 'whole beam length' is eclipsed (the disc doesn't block just the centre point, but a length across the chamber) then that will also impact the results. The off-set eclipsing disc will only block small lenths of 'prime' fast-background reaction regions, whereas the central obstruction will block a full reciprocation length.

Intrinsically, it does look like there is fast-fast going on, but I'd like to see more on how the results were handled to arrive at this 10% value.

[D Tibbets]

I agree this work demonstrates some fast-fast fusion.

I do not understand how they concluded that 10% were fast-fast, however, when the window of the chamber they were looking at only covers 5% of the reaction volume!?

It requires to also know the total reaction from the remaining 95% of the reaction volume to derive the percentage. If the remainder of the reaction volume produces 95% of all the reactions, then you get a very different answer!!

Of course, the proton detector is also looking 'along a beam' and if a 'whole beam length' is eclipsed (the disc doesn't block just the centre point, but a length across the chamber) then that will also impact the results. The off-set eclipsing disc will only block small lenths of 'prime' fast-background reaction regions, whereas the central obstruction will block a full reciprocation length.

Intrinsically, it does look like there is fast-fast going on, but I'd like to see more on how the results were handled to arrive at this 10% value.

I agree more details about thier methods, measurements, calculations and conclusions would be nice (even if I would just scan the math). I havn't seen a writeup from Google search other than this short presentation piece.
One point, I don't know how it affects the dynamics of thier measurements, but thier proton detecter was inside the Fusor so I think it would be seeing all the fusion protons except for those hitting the walls or grids or mask first.


Dan Tibbets

[D Tibbets]

I agree this work demonstrates some fast-fast fusion.

I do not understand how they concluded that 10% were fast-fast, however, when the window of the chamber they were looking at only covers 5% of the reaction volume!?

It requires to also know the total reaction from the remaining 95% of the reaction volume to derive the percentage. If the remainder of the reaction volume produces 95% of all the reactions, then you get a very different answer!!

Of course, the proton detector is also looking 'along a beam' and if a 'whole beam length' is eclipsed (the disc doesn't block just the centre point, but a length across the chamber) then that will also impact the results. The off-set eclipsing disc will only block small lenths of 'prime' fast-background reaction regions, whereas the central obstruction will block a full reciprocation length.

Intrinsically, it does look like there is fast-fast going on, but I'd like to see more on how the results were handled to arrive at this 10% value.

I agree more details about thier methods, measurements, calculations and conclusions would be nice (even if I would just scan the math). I havn't seen a writeup from Google search other than this short presentation piece.
One point, I don't know how it affects the dynamics of thier measurements, but thier proton detecter was inside the Fusor so I think it would be seeing all the fusion protons except for those hitting the walls or grids or mask first.

Thanks to a post by Dustin on the fusor.net site, this link provides more details and interesting information. Perhaps (if my perception of the neutral population in the Polywell being scant is correct) the relative lack of neutrals in the Polywell has a much more profound effect on efficiencies than has been appreciated. [EDIT] For instance, allowing the B e4 scaling mentioned by TallDave above.

http://mr-fusion.hellblazer.com/pdfs/fu ... -focus.pdf


Dan Tibbets

[chrismb]

Thanks to a post by Dustin on the fusor.net site, this link provides more details and interesting information. Perhaps (if my perception of the neutral population in the Polywell being scant is correct) the relative lack of neutrals in the Polywell has a much more profound effect on efficiencies than has been appreciated.

http://mr-fusion.hellblazer.com/pdfs/fu ... -focus.pdf


Dan Tibbets

Skipping to the summary; "Proton collimation experiments verify that the diffuse fast neutral population provides most of the observed fusion production, especially for the glow discharge modes (P > 100 mPa)....These results are inconsistent with the expectations from the multiple well model or any other model invoking a centralized source of reactivity."

....agreed!!...