Measurable Polywell Fusion at low Voltages

[D Tibbets]

In several threads ChrisMB has asked why the Polywell (WB6) hasn't shown neutron outputs at substantially lower potential wells than 10,000 volts. If the Polywell has such an advantage over gridded Farnsworth fusors, it is not unreasonable to ask this as they can show fusion at a little over 10,000 volts. But closer inspection shows why such is not expermentally reasonable. In the fusors the neutrons are counted for many minutes to accumulate their few neutron counts which are then converted to perhaps a few thousand fusions per second. Remember that the neutron counts in the WB6 tests converts to ~ 500,000,000 neutrons per second. So at similar voltages the WB6 was ~ 10-100,000 times better at producing fusions ( +/- ~ 30%). Also, keep in mind that this was at a feeble 1000 Gauss magnetic field strength. What does that have to do with detecting neutrons at lower drive voltages? Well ...

Extrapolating from cross section graphs and a few other resources. it looks like the D-D fusion cross section is perhaps ~ 0.1 X at 5,000 volts and ~ 0.01 X at 2-3,000 volts, compared to 10,000 volts energies.. Neutron counters are not very efficient. They may detect only 1 out of several thousand neutrons that pass through them. Add to that the brief duration of the WB6 tests (~0.25 milliseconds) and the distance to the detectors. the advertised output of ~ 500,000,000 neutrons per second was determined by only ~ 3 actual neutron counts during the test.
If 5,000 volts was used they would have detected only ~0.3 neutron counts. This is obviously below the sensitivity of their setup, so asking for counts at these low energies as proof that the Polywell performs as claimed is practically impossible.
And, actually, acurate counts at any voltage corresponding to energies that can be compared to the cross section curves should be equally valid. Not having the sensitivity to detect something at lower outputs does not imply anything other than the limits of expermantal setup.

Now WB 7.1 may be a different story (if only we could see the data ). If they had longer run times, perhaps slightly stronger B fields, slightly larger size (?), neutron detectors closer to the magrid, etc. they might have detected neutrons at 5,000 or even 2,000 volts, if they felt inclined to run at these voltages. But, it is irrelevant to the fusion efficiency question. It might be interesting to better measure the cross sections at these relative low energies and contribute to more precise information for the curves, but again, it would not add much, if anything to the Polywell fusion physics, except it might be interesting to compare the KeV dependent thermalization rates to the resulting fusion rates. Actually pushing the voltage to perhaps 15-20,000 volts would be better as that would increase the cross section dependent fusion rate, leading to more neutron counts and more statistical confidence and reproducibility of the data.

Dan Tibbets

[TallDave]

Interesting points Dan. I'm still not sure we get longer runs with WB-7.1, but we'd definitely get more measurements if they have microsecond time domain information.

[chrismb]

In several threads ChrisMB has asked why the Polywell (WB6) hasn't shown neutron outputs at substantially lower potential wells than 10,000 volts.

I am not sure I have said that. Please excuse me if I have mislead you with any ambiguities.

What I have aimed to say is that if a given configuration of Polywell, or in fact of any fusion reactor whatsoever, has the capacity to seriously pump out useful amounts of power then it must be able to produce measurable amounts of DD fusion at lower voltages.

So, for example, if you set up an experiment and you do not get measurable (by which I mean 10^~6 neutrons/s) at either 500eV for thermal or fast-ion-into-fast-ion reactors, or 2kV for beam-target devices, then scaling up the collision energies to max DD peak will not give you useful power outputs. This point I am making is just about the scaling of fusion reactions with collision energy.

Now on your particular point, it is valid. If you were to only get a 10^~6 reaction rate for a few ms, then, indeed, your statistics of neutron detection would give you some wide uncertainties, for sure. But a similar same calc can be done at whatever collions energy you *do* begin to get measurable fusion reactions. For example, if your reactor is putting out 10^9 neuts/sec for DD at 10kV, then the maximum this reactor can put out, if you up the collision energy alone, is around 10^12 because the peak reactivity for DD is only 3 oom or so above that to be found at 10kV, and 10^12/s maximum neutron output is of no interest for power.

Bottom line corollary is this: If you are in the business, or hobby, of trying to build your own over-unity fusion reactor, then you don't need to design [thus, pay] for anything more than a few kV in prototype form because if you don't get measurable neutrons out of DD at a few kV, then your reactor just ain't gonna do much at higher voltages. (NB: Even if you do get neutrons at a few kV, then this is still not necessarily a proof that it'll work, but without this characteristic, then forget it!)

So, to recap, you can only expect a DD reactor to get to around 3 oom output above whatever you can get at 10kV. This is also true for DT because, similarly, its peak is 3 oom above that to be found at 10kV. Whereas, the peak output for DD is around 12 oom above that to be found at a few kV, so if you get 10^6 neuts at a few kV then you might extrapolate to 10^18 at peak.

(Just to note; if you run p11B, then its peak is 11 oom above that at 10kV so if anyone actually gets 10^6 alphas per sec at 10kV then that reactor will go really steaming if the voltage is cranked up.)

[D Tibbets]

ChrisMB, everything you say is true. Though using 500 eV in WB6 would produce a lot less than 1 million neutrons per second. By extending the cross section graph at 1000 eV the rate would be ~ 100x less than at ~ 2,000 eV and 500 eV might be ~ 100 X smaller again. So that would be ~ 1,000,000 times less neutrons per second compared to 10,000 eV. Or, if the WB6 counts were accurate, you would expect ~ 500 neutrons per second. In less than 1 ms, that would be only ~ 0.5 neutrons. With the inefficiencies of the neutron counters you would only have ~ 1 chance in twenty thousand of detecting any neutrons. The cross section scaling you mention is valid. But that is the point. The performance of WB 6 was very good at these levels (as claimed by Bussard). This is only the baseline measurement. Similar scaled results at higher or lower potential wells would be the same, as you say.

The important point is that this is the claimed results for a feeble machine. The important questions are the magnetic and size scaling along with the cost scaling. The claimed B^4 r^3 power scaling is the critical items, not the cross section scaling. The claimed loss scaling of B^1/4 r^2 is the other important concideration that needs conformation. If these claimed scaling laws are accurate, then other conciderations like improved quasi sphericity, improved recirculation, and POPS like effects are icing on the cake.

The single data point for fusion in WB6 says very little about the scaling. In a way, it is fortunate that EMC2 was able to reach the B field and potential well necessary to detect the borderline neutron production vs detectability they did. You can consider this as fortunate, of as clever. Certainly they had information and modeling from other machines that helped to define their expectations and hopes.

I'm believe thermalization times issues are reflected in the predicted loss scaling. This may be the critical issue for D-D fusion. You can add bremsstrulung issues for P-B11 fusion.

Dan Tibbets

[Tom Ligon]

I'm not sure WB6 did not show fusion much lower, at a well depth of 5 kV. They had one run at that condition and it produced one count.

Which also says I'm equally not sure of the contrary, statistically speaking.

I'd love to know if WB7 spent time in this realm, producing a count somewhat better than 1 +/- 1.

I'm fairly sure my fusor does produce fusion when run at around 8 kV with deuterium, using a neon sign transformer. I've just never seen it produce any count above background while running on that supply. The world has enough experiments that claim barely detectable and statistically insignificant fusion. Useful devices make it unambiguously, or preferably dangerously unambiguously.

[D Tibbets]

I should add that one of the machines- the solid copper block I think (I can't remember it's name) reportedly produced ~ 1,000,000 neutrons (or fusions?) per second with ~ 35000 Gauss B fields, and only ~ 300 Volts potential wells.

Lets see... B field ~ 35X larger so B^4 scaling would be ~1,000,000 X. Potential well of ~ 300 volts. A very rough estimate would put the crossection at perhaps 1/1,000,000 to 1/100,000,000 X (?). Within the obvious uncertainty in the crossection at these low energies, and the smaller size of the machine, the numbers could be consistent with the scaling predictions.
Has anyone else claimed measurable D-D neutron production at such very low energies?

This result may be the real indicator of the gross fusion potential of the Polywell. The WB 6 results may be more important for addressing acceptable loss concerns, ie: reaching breakeven at some point. (Also, the mechanisms nessisary to achieve deep potential wells).

Dan Tibbets

[chrismb]

I should add that one of the machines- the solid copper block I think (I can't remember it's name) reportedly produced ~ 1,000,000 neutrons (or fusions?) per second with ~ 35000 Gauss B fields, and only ~ 300 Volts potential wells.

Never say never, but I will take this one with a pinch of salt. 300eV is running too much on the downward slope of reactivity that I would expect to see next to nothing down there.

But I think the issue would revolve around what '300eV well' means. I mean, if there is a kV 'well' of the opposite polarity somewhere else nearby then of course you may well have the environment for the necessary acceleration potential. If you are saying that neutrons were observed and that no ion collision would have been in excess of 300eV CoM then that'd be a tough call. If you are arguing fast-fast could have been present and 1200eV collisions were possible... well, I suppose there may be a few squeaks of plausibility but I would suggest there would still be 'everything to play for' [even if your recollections are correct] because of the statistical limits you and Tom are, correctly, raising.

Let us not forget that if a Polywell cannot run continuously then the few joules it might make during pulses would be unlikely to provide any payback for the magnetic field energy alone, let alone the other ancillary parts. It is a problem conveniently forgotten of tokamaks as well, so let's be clear on the issue; the 20 MJ of energy produced in the most productive tokamak pulse ever run did not pay back the 1GJ magnetic field energy necessary to contain it for the couple of seconds it took to generate that 20 MJ of neutrons. If a fusion reactor that operates for a few seconds is orders of magnitude out of the running on magnetic energy payback, then polywell operating for a ms or two won't cut it either.

As I have posited, it is not just detectable neutrons at a few kV necessary to demonstrate the most basic qualification for 'viability', it also has to be a continuous reaction, so the first post of this thread is somewhat moot.

[D Tibbets]

As with Tokamaks, the energy input into the magnets can be mostly ignored if superconductors are used and the magnets are kept at constant strength. Even without superconductors the Polywell is claimed doable, probably ay larger sizes. A P{olywell or other machine could be pulsed in a number of ways- individually, or in combination.

1}Magnetic fields can be varied- unattractive if superconductors are used.

2) pulsed fuel supply

3) Pulsed energy drive. In the polywell this could be the electron current or the driving potential. This would complicate the maintainance of B= 1 trapping and thus the pumping concerns to maintain adiquate vacuums.

Pulsing by itself can provide useful power, as indicated by some forms of FRC, DPF, ICF, etc

As far as disbelieving that 300volts wells (600 V equivalent if beam- beam? I don't know where your 1200 volts comes from) could produce fusion, you are arguing on one hand that such measurements would be a positive indication, then you dismiss the actual claimed results, despite my guesstimate that it would be well within the predicted B^4 r^3 scaling.

If you are suspicious of the WB6 experiments barely detecting ~500 million neutrons per second, while the low voltage copper block test detected 1,000,000 neutrons per second, consider that these were different experimental setups. The neutron detectors may have been closer, and test runs may have lasted orders of magnitude longer.

If you are suspicious due to the sparsity of accessible date, well, that's a valid criticism.

WB6 neutron results, B fields, and potential wells are public. While WB3, WB4, and WB5 and other machines neutron results are not quantitatively known, and for the sake of this argument, the voltages they were run at (specifically the potential wells and B fields they were run at) are publically unknown. Such information, which should be aviable to insiders, would presumably reinforce the scaling predictions.

Dan Tibbets

[chrismb]

As with Tokamaks, the energy input into the magnets can be mostly ignored if superconductors are used and the magnets are kept at constant strength.

'Fraid not. Remember your basic MHD training and recall that the magnetic field and the plasma becomes one, so when the plasma collapses so does the magnetic field, and that energy is then dumped along with the plasma energy. A constant strength magnetic field means a constant plasma means a constant reaction [if there is one to be had].

As far as disbelieving that 300volts wells (600 V equivalent if beam- beam? I don't know where your 1200 volts comes from)

CoM, etc.. A conversation had many times. v^2 and all that.

If you are suspicious due to the sparsity of accessible date, well, that's a valid criticism.

Exactly. But then some.. I want to see how this data is measured. There are far too many ways to measure a plasma wrongly - that's what this new Navy contract has been about, to build a device from which some solid plasma data can be measured.

[Roger]

Great thread.

[D Tibbets]

As with Tokamaks, the energy input into the magnets can be mostly ignored if superconductors are used and the magnets are kept at constant strength.

'Fraid not. Remember your basic MHD training and recall that the magnetic field and the plasma becomes one, so when the plasma collapses so does the magnetic field, and that energy is then dumped along with the plasma energy. A constant strength magnetic field means a constant plasma means a constant reaction [if there is one to be had].

Yes and no... Take my comments for what they are worth, as I know little about plasma physics and magnetic interactions. Tokamak plasmas once they reach ignition can be a run away process- it gets hotter, fuses faster, gets hotter, feedback magnetic effects (I've heard the magnatized plasma in a Tokamak acts much like the secondary windings of a transformer),etc. till the magnetic containment fails or some other instability becomes intolorable. But, I don't know if this implies the superconducting wires suddenly dump their current. Even if this occurs to some degree, a (very) large capacitor bank could store this energy for the next cycle. This would be difficult and complicated, but from diascussions in other threads the danger of rapidly draining (quenching) superconducting magnet currents would need to be accommodated as a a safety feature anyway in order to prevent large explosions. Also, I'm not sure the fusion plasma magnetic field needs to collapse once the fusion fuel is exausted. The plasma would still be very hot and well insulated. This B field plasma would lose its energy through radiation or through presumably controlled diverters that are again presumably coupled to other controls and feeds to maintain the process. Diverters are one area of the steady state (?) Tokamak that might be a show stopper. Without diverters there is no way to get the accumulated charged fusion products out without collapsing the magnetic fields, or at least weakening them so that the thermalized charged particals (hopefully mosty fusion products at this stage) can be drained off. The difficulty in a steady state machine is how to do this preferentially for the fusion products. The Polywell avoids this difficulty as the fusion products are not confined to the working plasma. those cusps are useful after all, and makes a steady state operation much simpler to achieve.

I'm also uncertain that an ignition machine needs to be a run away process. Careful fuel starvation and external energy inputs might allow prolonged ~ steady state operation (if you can get your diverters to work). In perhaps different ways, this is the end result of controls in a supercritical fission reactor.

I guess the situation is simplier in a Polywell as it is a power amplifier, not an ignition machine. Also, the plasma is not magnatized (no dominate current direction) so magnetic feedbacks would be less complicated.

Dan Tibbets

[chrismb]

This B field plasma would lose its energy through radiation or through presumably controlled diverters that are again presumably coupled to other controls and feeds to maintain the process.

No. The energy dissipation of the collapsing magnetic field is a very mechanical process. In the case of JET, the whole toroidal vacuum vessel, around 100 tons of it, jumps up a centimetre or so. You can hear it as a 'thump' in the control room some distance away (and behind metres of concrete) as it 'falls back to earth'!

It is very much a magnetic version of a capacitor. There *is* 1GJ of magnetic energy induced into the volume of the chamber, and when that drops back down to no-field (when the plasma collapses), then that 1GJ has to go somewhere. I guess the actual process of back-reaction is by inductive effects back into the coils, and the surrounding magnetic structures.

In the case of ITER, it is 50 GJ of magnetic enegy. So each time the magnetic field collapses in ITER, it is the same energy dissipation as you would get if you set off 10 tonnes of TNT inside the vessel (and probably on a similar time scale!).

It would be nice to create the scenario in which the magnetic field slowly ramps down, but if the plasma is not sustained (as it never is [yet] in a tokamak) then as the plasma hits the side of the chamber and cools, so the magnetic field is shed very quickly. As you suggest, a slow release of that energy and some means to recover that energy *would be nice*, but at the moment they are not yet in full control of the plasma anyway, so the rest of the speculations on what may-or-may-not be possible is moot.

The corollary is this;
- if JET can produce 20MW of neutrons and its magnetic field is 1GJ, then it has to run for at least 150 seconds to repay just the magnetic field energy (or show some means to be able to recover the magnetic energy content of the plasma)(assuming 30% conversion efficiency back into electrical power for the mag fields).
- in ITER, this figure is around 350 seconds, yet the experiment is only aiming to produce 400 second pulses at best, so it would appear to be an experiment that is unlikely to achieve its end objective, even if it works to plan.
- in Polywell, consider that a 1 m radius volume with 10T in it has a magnetic energy content of 160MJ, so if you're planning on running it in 1 ms pulses then it will need to put out 160 GW during those pulses just to pay for the magnetic field.

I have two predictions; 1) that viable fusion reactors will employ both magnetic and electric confinement means*, 2) that the requirements on generating the necessary magnetic fields (for (1)) are such that it means pulsed devices are non-viable.

*(notwithstanding that these are the same, but in different inertial frames - it seems odd to miss the opportunity to use two forms of 'force' when neither seem to be sufficient on their own)

[WizWom]

- in Polywell, consider that a 1 m radius volume with 10T in it has a magnetic energy content of 160MJ, so if you're planning on running it in 1 ms pulses then it will need to put out 160 GW during those pulses just to pay for the magnetic field.

I have two predictions; 1) that viable fusion reactors will employ both magnetic and electric confinement means*, 2) that the requirements on generating the necessary magnetic fields (for (1)) are such that it means pulsed devices are non-viable.

Except that "blowout" in the nearly neutral plasma won't have this magnetic field collapse effect. Because the magnetic field is there only to guide the electrons away from the grid.

That's the decoupling that Dr. Bussard talked about. Because polywells don't rely on the ions or electrons for any of the magnetic field, and the containment is of the electron well, which the only way for that to neutralize is by sending beta > 1. It's a different beast than a tokamak, which requires a charged plasma to contribute to its own confinement.

[chrismb]

Except that "blowout" in the nearly neutral plasma won't have this magnetic field collapse effect. Because the magnetic field is there only to guide the electrons away from the grid.

In a perfect conductor, the total magnetic flux through any surface is a constant. In a plasma which is nearly perfectly conducting, this means that the form of the plasma is tied to the magnetic fields, and vice-versa. Neutrality or otherwise is immaterial.

In every description I have heard of of a Polywell, the electrons sound pretty mobile and I would regard them as being essentially perfect conducting. The thing is, it is just so bloomin' difficult to tell what is meant to be what. I've never yet been able to conjour up a consistent view of what the plasma in a Polywell is supposed to be, and once I take a view I am told that's not it, so I go through a few cycles of re-interpretation and end up with a new conclusion only to be told that's not it either.

The ions clearly have near perfect conductivity - I am constantly told there are no impediments to their motions as they slide down the electric field to the electrons. So I guess the only thing I can ask here is why you think the electrons won't have low conductivity? What are the transport barriers that stop them swapping with their neighbours, and if there are such mechanisms then why do they circulate around the magrid?

Are we talking here about electrons that have high conductivity that do not respond to electric fields, are mobile sometimes, but not on other occasions? Or low conductivity electrons that respond to electric fields and can exert a dia-magnetic field without thermalising collisionality? [This isn't my mess of a desription to sort out!!.......]

I would hold any claims that the internal ions and electrons behave in a dissimilar way to what one might expect from a conventional plasma as exactly that - a claim. It is for the experiments to prove otherwise. I would suggest it is wise to simply presume the electrons are as any other electron plasma description, for now at least, and presume they do, indeed, have low conductivity.

[rcain]

...
I have two predictions; 1) that viable fusion reactors will employ both magnetic and electric confinement means*, 2) that the requirements on generating the necessary magnetic fields (for (1)) are such that it means pulsed devices are non-viable.

*(notwithstanding that these are the same, but in different inertial frames - it seems odd to miss the opportunity to use two forms of 'force' when neither seem to be sufficient on their own)

Good thread and some very useful/true points Chris. However, I disagree with your 2nd final conclusion.

By my reasoning, a pulsed solution may be the only way we able to cost-effectively/practically sustain/frame(?) the densities and d/dt's required.

In particular, how about a reciprocating 'pair' of Polywells, back to back, 180 degress out of phase: as one dumps the other powers up. ie. we dont wasts all that energy we have been trying so hard to produce.

On your other point, on our (mutual/universal) lack of understanding/agreement on plasma structure and dynamics::

...
What are the transport barriers that stop them swapping with their neighbours, and if there are such mechanisms then why do they circulate around the magrid?

Surely the bit you're missing is the I in IEC, yes?