EMC2 has published a polywell preprint on arXiv

Point out news stories, on the net or in mainstream media, related to polywell fusion.

Moderators: tonybarry, MSimon

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

Re: EMC2 has published a polywell preprint on arXiv

Post by D Tibbets »

The electron thermalization process is of course proceeding. The real question is how fast it is proceeding, Increased density speeds it up, increased machine size increases the oppertunity, and increased temperature slows the process. How all three combine, along with probably several other considerations, determines the answer.

Increased opportunity in larger machines is due to increased transit distance. If confinement is defined in number of transits, then the time to complete a transit is the product of average velocity and distance across the machine. At the same transit confinement number, a 1.5 meter machine would have a confinement time ten times greater than a 0.15 meter machine. Note that average velocity is perhaps tricky, as it is a gradient and also depends how radial from the center the motions are.

The only detailed analysis that I know of is a paper by Bussard. I have referenced it before and it .seemed to satisfy Bussard.

As for thermalization in this machine. The measured x-rays came only from electrons with >2000 eV due to the cutoff filter. As the number of relative low energy ions and electrons from the plasma guns was vastly greater than the ~ 7000 eV e- gun electrons (assume a ~ 1,000,000 to 1 ratio), means the average temperature of the total charged particles did not increase much over the lifetime of the test. The e-gun electrons were the only contributor to the X-rays. That the E-guns were operating before the plasma burst, and the increased plateau values were ~ consistent with the density measurements, it is reasonable to assume that the signal came from increased confinement and thus density of the injected high energy electrons. To work the themalization time had to be longer, perhaps much linger than the confinement time. This places a limitation on the injected electron thermalization time minimum possible. How this would extrapolate to other conditions is complex, but this is at least one example of minimal thermalization time of the electrons under these specific conditions.

The temperature of the injected plasma was presumably somewhere between 1000 and 2000 eV. This is suggested by the quantity introduced and the reported ~ 700 MW of power from the plasma gun. Perhaps about 30 to 40 thousands of plasma amps was injected. Compare this to the e- gun current of ~ 2 amps. Assuming much of the E-guns electrons managed to enter the machine, the electron population would be ~ 10^22 at an energy of ~ 1500 eV, and ~ 10^19 would be ~ 7000 eV (at most). Once mixed and thermalized the average temperature would be ~1510 eV. This is well below the x-ray cutoff filter passage window. So thermalization between the two populations of electrons has to be significantly longer. This is of course an over simplified analysis, but I think serves my argument. It does raise questions about the WB6 situation. The high energy electrons ionized the injected gas and the secondary electrons were heated by the hot electrons. This suggests significant thermalization between the two populations. The conditions are much different between these two machines and requires analysis well beyond my pay scale. That is why I have referred to a detailed analysis. Keep in mind the thermalization has two components. The time it takes for the two populations to near a common average energy, and the time it takes the outliers-mostly up scattered high energy tail, to build to their Maxwellian statistic population. This is a slower process as the higher energy electrons have longer MFP and less frequent Coulomb interactions.

Monoenergetic does not have to imply a very narrow energy distribution. But, I think it does have to imply a significant truncation of the high energy Maxwellian tail. The name is perhaps misleading, but I don't know of any better name.

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

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

Re: EMC2 has published a polywell preprint on arXiv

Post by D Tibbets »

dnavas wrote:
D Tibbets wrote: Note that this performance / breakeven is ~ 10 times worse than projected for WB100. What is different? That is hard to say as no numbers were provided about confinement, recirculation, confluence, etc.
The way I read it, there was no recirculation. Lifetime of 300 bounce confinement, and the profusion of measuring devices along the cusps make me think that recirculation wasn't being attempted. That would affect the numbers by a factor of ~10. That said, high B remained for about an order of magnitude more than 7ns * 300 bounces would predict, so I'm not sure what to conclude from that. Also, 300 is far lower than reported previously (1k and 10k?) and it's not clear why that is either.

If I'm viewing the videos correctly, it looks like highest losses are out of the faces, and I thought recirculation was less certain for such losses?

Corrections welcome -- it's been a long time, and I'm no physicist....
The losses out of selected cusps depends on the distance from the magnet cans to the center of the cusp and geometry. In this small machine with relative wide spacing between the magnets and relative small spacing from the can surface to the center of the rings suggests to me that the losses were weighted more towards the corner cusps when compared to a WB6 size and relative spacings. Just how much of a difference is another question. As such I wonder if in this instance a face centered point cusp was the best target for the E-gun. The simulations may have had different dimensions than this machine so it is perhaps misleading to compare them.

Recirculation should not care which cusp the electrons escape through. If there is no obstructions, like nubs, they will decelerate and stop provided the decelerating voltage is great enough to overcome the escaping electron kinetic energy. Once that happens and ignoring ExB drift and unshielded surfaces the will fall back towards the machine interior, much like the virgin electrons from the E-guns. WB6 report claims there was 90% efficiency. The remaining 10 percent of the potentially recirculating electrons were lost by hitting a nub or ExB drifting to the magnet can surface. Or, if upscattered their outward momentum was not overcome and they continued outward till they hit an external surface (which would happen rather than looping around a field line to another cusp because they designed it this way). I don't know but I guess that these recirculated electrons do not travel far past the magrid midplane radius before stopping and reversing. Perhaps a few centimeters at least and a dozen cm at most (in WB6), It is important that the electrons do not slow to a stop too close to the magrid midplane radius because of problems of cold electrons plugging the cusps as was found in WB5.

This machine (we need a name) had all of the injected electrons (not the plasma gun electrons) accelerated by the potential on the E-gun extractors, reported as 7,000 volts. This implies there was no charge on the magrid (grounded). This would seriously impeded recirculation. An electron with up to 7,000 eV outbound KE would be stopped by a +7,000 volts on the magrid , but at 0 volts there would be very little deceleration- thus trivial or non existent recirculation.

The ~ 90 percent electron cusp injection represented by recirculation may be an indicator of the maximal performance possible. This machine probably didn't reach 50% efficiency, and WB6 was probably much worse than this. Recirculation suggests that the majority of electrons that escape are traveling perpendicular to the cusp and upon reversal reapproach the cusp on the most ideal vector, there is not much distance for much beam spreading. This distance for the e-guns is mentioned in the patent application, and I wonder why this E- gun was placed so far away. There are multiple considerations and this may be the best compromise, but
still, I wonder...

7 micro seconds was not the cusp confinement time. It was the ~ duration of the plasma gun firing. Using seven passes , the machine diameter and the speed of the electrons of perhaps ~ 6 million M/s, gives Distance per pass (~ 15 cm) * number of passes (7)= distance= ~ 100 cm or ~ 1 M.
1 M/ speed (6,000,000 M/s gives cusp confinement time of ~ 0.16 microseconds. Given a confinement increase to ~ 300 passes (~40 times more) would give a confinement time 0.16 *40 = ~ 6.4 micro seconds. ~7 microseconds would be the Wiffleball confinement time, not the cusp confinement time.

Further number crunching of e-gun input suggests that the high voltage electrons optimistically entered at ~ 1 amp of current or ~ 6*10^19 electrons per second. With a confinement time ~ 1/100,000 of this the high energy electron content at any given time with steady state Wiffleball conditions would be ~ 6*10^14. The lower energy plasma gun electrons injected would be at a density near ~ the peak pulse input divided by the decay/ escape rate. Over the time frame of several confinement times at Wiffleball condition multiplied by some fudge factor for losses as the Wiffleball was forming. This might be an order of magnitude or 2 drop over the time of the test. So the plasma density from the plasma gun may have been ~ 10^ 20/ machine volume. So at any given time during the test the neutral plasma was ~ 1 million times greater than the e-gun derived electrons. It is complaint with the 1 ppm that is the target for the electron excess to create the potential well. My earlier post was actually off by several orders of magnitude for this ratio. This is probably due to the quantity of plasma gun (and proportionate loss of e-gun electrons) losses during Wiffleball inflation/ formation. After the pulse there were no new plasma gun electron input, but there was continuous e-gun electron input, and these electrons built up over time due to the improved confinement, until a new equilibrium was reached. this was not a constant plateau though because there was not enough injection to maintain the steady state. I don't know of the magnitude of the electron shortfall, or how soon the confined ion losses would become significant. Perhaps recirculation and/ or improved cusp confinement baseline could push it over the edge . But, for the money, this under powered and/or under performing machine with good instrumentation got the desired job done.

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

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

Re: EMC2 has published a polywell preprint on arXiv

Post by ladajo »

There's no evidence that it has any effect on the potential well or the main plasma population
You need to go back and review how the machine is designed to function.
E-gun power is a direct effect on potential well, and indirect on plasma population.

As for naming the machine, how about "Bob"?
Okay, just kidding. How about, "Mini-B". With "B" meaning Ball and/or Beta=1?

As for all the speculation, I would posit that there is more than likely more data that was not released in this initial paper from Dr. Park & Co.
What all this really begs is the "When is what's next?" question.
The development of atomic power, though it could confer unimaginable blessings on mankind, is something that is dreaded by the owners of coal mines and oil wells. (Hazlitt)
What I want to do is to look up C. . . . I call him the Forgotten Man. (Sumner)

mvanwink5
Posts: 2143
Joined: Wed Jul 01, 2009 5:07 am
Location: N.C. Mountains

Re: EMC2 has published a polywell preprint on arXiv

Post by mvanwink5 »

I guess Convergent Scientific Inc. just needs to get some gargantuan Godzilla guns and take a look at Beta = 1. Surely they know by now about the news?
Counting the days to commercial fusion. It is not that long now.

dnavas
Posts: 84
Joined: Sun Nov 11, 2007 3:59 am

Re: EMC2 has published a polywell preprint on arXiv

Post by dnavas »

dnavas wrote:If I'm viewing the videos correctly, it looks like highest losses are out of the faces, and I thought recirculation was less certain for such losses?
D Tibbets wrote:Recirculation should not care which cusp the electrons escape through.
I'm sure I'm missing about six years of evolution in understanding here, but sadly the search for recirculation to refresh my poor memory led me to your thread back in 2008 titled "Does Recirculation real?" (I'll resist the temptation to link directly, as that will just make that thread's page rank in our search system even higher :| )
There are enough competing ideas in that short thread, though, that I might be able to read it in a number of different ways :)
D Tibbets wrote:This implies there was no charge on the magrid (grounded).
I had completely missed that. Is this an experimental choice, the result of the problems they encountered on their last machine, something else? [Lots of questions this little paper raises]
dnavas wrote:That said, high B remained for about an order of magnitude more than 7ns * 300 bounces would predict
D Tibbets wrote:7 micro seconds was not the cusp confinement time.
Err, 7ns. From the paper:
This is because an electron beam at 7 keV has a transit time
of 7 ns to move across the cusp system, compared to several
microseconds during which the high B state is maintained.
I also missed the bit about the plasma gun firing length. One of the things I was worried about was that as they varied the ion input energy, the width of the pulse might have been changing too. Is there contra-evidence? (I think this is a similar concern to the electron thermalization issue brought up a little bit back, but asked about the ion species instead.)

Also, what do you think they mean by the 300 number anyway? "at least 300" is a little vague. Clearly it doesn't mean exactly what it says ("an electron will be contained for at minimum 300 passes"), but is it a lower bound on the mean? the lower bound of the first std.dev??

AcesHigh
Posts: 655
Joined: Wed Mar 25, 2009 3:59 am

Re: EMC2 has published a polywell preprint on arXiv

Post by AcesHigh »

a few days ago, on NextBigFuture's article about Focus Fusion, a comment by Sebtal:
Pollywells don't seem likely to work at all for net energy on a physics basis - they get better than expected confinement and scaling through self organising kinetic effects that are unlikely to work in the higher density regimes they need to get to.
I asked if this new Polywell new changed his opinion...

his answer
Yes, it changes my opinion of it - which was based on the dominant narrative of re-circulating beams, a
completely different regime - but I should stress that is starting from a low bar.
I thought it highly implausible and the physics deeply suspect based on those descriptions. Operating in this regime I would elevate it to "not implausible" - which is not the same as "plausible".
To put in context, they have demonstrated some evidence of a mechanism where they could realistically contain a dense plasma - note they may be mistaken, that happens a lot, I'm afraid I'm not in a position to evaluate whether they are kidding themselves but lets assume that the result stands up to scrutiny and is replicated.
They then need to demonstrate they can then contain a dense plasma. Then they need to demonstrate that it can scale to the necessary regimes.
It's a long way from a fusion reactor - and the fact is that even if they have demonstrated the mechanism to sustain particle confinement, the mechanism may not scale up, and other things may lead to "anomalous" (i.e. not well understood) losses of heat from transport mechanisms they were not aware of and do not manifest on this scale when they try to scale up.

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

Re: EMC2 has published a polywell preprint on arXiv

Post by ladajo »

So basically, he has no real idea on what they have been doing over the last years, and is basing his opinion on prior hearsay.

His wishy-washy wandering critique could apply to any number of initiatives.
The development of atomic power, though it could confer unimaginable blessings on mankind, is something that is dreaded by the owners of coal mines and oil wells. (Hazlitt)
What I want to do is to look up C. . . . I call him the Forgotten Man. (Sumner)

choff
Posts: 2447
Joined: Thu Nov 08, 2007 5:02 am
Location: Vancouver, Canada

Re: EMC2 has published a polywell preprint on arXiv

Post by choff »

Do we know the actual timeframe for the experiments run with this smaller wiffleball machine. The presumption seems to be they went to this after hitting the wall with startup on WB8. It could have been done over any number of the last six years as a side experiment.
CHoff

MSimon
Posts: 14334
Joined: Mon Jul 16, 2007 7:37 pm
Location: Rockford, Illinois
Contact:

Re: EMC2 has published a polywell preprint on arXiv

Post by MSimon »

Stubby wrote: http://www.dailykos.com/story/2014/06/0 ... early-2006

word is spreading

mvanwink5 and tom quoted in blog

which one of you is Roger Fox? :lol:
Expect something from Alan Boyle. He has been in contact with me. Thanks deltav for the heads up given me early on.
viewtopic.php?p=112781#p112781

BTW Roger Fox used to be a regular here. I sent him an e-mail thanking him for the KOS post.
Engineering is the art of making what you want from what you can get at a profit.

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

Re: EMC2 has published a polywell preprint on arXiv

Post by D Tibbets »

The mention of ~7 nanoseconds for the transit time (one pass) in this machine is just wrong, unless I am far off in my understanding. [EDIT- Indeed, I am wrong, see subsequent posts :oops: ] The KE of the electrons of somewhere between 2 and 7 KeV implies speed of eg: 5,000,000 M/s . To travel ~ 15 cm/ 0.15 M requires 0.03 microseconds or 30 nanoseconds for one pass to occur. And, ~ 0.2 microseconds for 7 passes and about 8 microseconds for 300 passes. This is based on a given speed for a 10 KeV electron of 10 million M/s. It also assumes the speed is constant for this average KE and that the path goes through the center on each pass. The average speed of the plasma gun injected electrons and the e-guns injected electrons will vary from this speed but I doubt that it would be more than a factor of two either way, because the velocity only changes as the square root of the energy.

Another number crunching exercise is to consider the injection energy of 7 KeV (for the electrons of intrest) versus 12 KeV in WB6. Assume about a 20 percent drop off from space charge resistance, then the numbers are reduced to ~5.5 KeV and ~ 10 KeV in the compared machines. Assume the average energy is 1/2 of the peak energy. So the KE average energy would be 2.7 KeV and 5 KeV.
The comparative KE is 2.7 / 5 or 0.54. The square root of this is 0,73 or ~0.7. ).7 *10 million M/s gives an average speed of ~7 million M/s in this machine. You can adjust the confinement time accordingly. Do keep in mind though that this a moving target as the confinement passes change and energy loss occurs through radiative processes. Then throw in the degree of thermalization that occurs between the two electron populations and the permutations of the formula becomes very complex. Essentially all of these permutations tends to reduce the energy of the electrons of intrest. On the other side the continuous injection new energetic electrons impedes this decay, but only to the extent that the continous injection keeps up with particle losses.

Dr Parks, etel had to consider all of these factors and design the machine to give the desired/ required conditions for the nessisary makeup of losses (within some brief time limited test interval) to maintain the energetic electron population for long enough and hot enough to give definitive bremsstruhlung output numbers that distinctly showed the difference between regular cusp confinement and Wiffleball/ Beta=1 confinement. That they did so with the budget restraints requiring significant compromises in size and design, is impressive.

PS: Along with the demonstration of the Wiffleball, that they achieved at least a temporary density of well over 10^21 charged particles with this mini machine with modest B fields, goes a long way (I think) towards reaching the densities needed for useful fusion output. If there was some anticipated or unanticipated road blocks to high density operations they were not apparent in this machine.

PS #2: Mmm... Mini B for a name... Not bad. Change it to Mini ME for small machine Magnetic Electrostatic is more catchy and a more obvious plagiarizing of a Hollywood name. 8)

Dan Tibbets
Last edited by D Tibbets on Thu Jun 12, 2014 2:12 am, edited 1 time in total.
To error is human... and I'm very human.

tauntaun_rider
Posts: 11
Joined: Thu Jul 18, 2013 6:27 pm

Re: EMC2 has published a polywell preprint on arXiv

Post by tauntaun_rider »

D Tibbets wrote:The KE of the electrons of somewhere between 2 and 7 KeV implies speed of eg: 5,000,000 M/s [....] This is based on a given speed for a 10 KeV electron of 10 million M/s.
I get 50,000,000 m/s for a 7 keV electron and 59,000,000 m/s for 10 keV. That makes a 7ns transit for a 7keV electron across a distance of 34cm, which looks like the size of the device.

dnavas
Posts: 84
Joined: Sun Nov 11, 2007 3:59 am

Re: EMC2 has published a polywell preprint on arXiv

Post by dnavas »

tauntaun_rider wrote:
D Tibbets wrote:The KE of the electrons of somewhere between 2 and 7 KeV implies speed of eg: 5,000,000 M/s [....]
I get 50,000,000 m/s for a 7 keV electron.
Yeah, that agrees with my simple math as well:
v = sqrt(KE * 2 / m)
v = sqrt(7kev * 1e3ev/kev * 1.6e-19j/ev * 2 / 9.11e-31kg) = 50e6m/s

but then the electrons turn around, so average speed is likely halfish of that. The enscribing cube face is doc'd as being 21.6cm, so the interior would be <19cm (subtract out 2x minor radius of doc'd 1.3cm).

.19/25e6 = 7.7ns That's close enough for me.

mattman
Posts: 459
Joined: Tue May 27, 2008 11:14 pm

Re: EMC2 has published a polywell preprint on arXiv

Post by mattman »

Hello,

Dan: Thanks.

The Navy has a great theoretical underpinning for the work: Citation #2: James Tuck, Los Alamos, Nature, 1960.

From the Abstract:

...One way to get around this difficulty is to have no magnetic field in the plasma, so that the radiation becomes only a surface phenomenon instead of a bulk one. Of all the magnetic confining geometries at present known, theoretically only one has enough stability to hold a pure (that is, beta = 8pinkT/B2 = 1) plasma. This is the picket fence or cusped geometry. Unfortunately, according to [Dr. Harold] Grad, the cusps will be very leaky...

James Tuck is very famous inside the fusion community. Using his paper in Nature as underpinning gives this work credentials. Dr. Tuck devoted a good portion of his career to the "Picket Fence" idea:

Image


Looks like this cusp confinement is making a comeback from the grave
Last edited by mattman on Wed Jun 11, 2014 1:01 am, edited 4 times in total.

mattman
Posts: 459
Joined: Tue May 27, 2008 11:14 pm

Re: EMC2 has published a polywell preprint on arXiv

Post by mattman »

I applied the Biot-Savart law to the Navy Device.

My calculations come out about 74% of the field listed in paper.

Not bad.


Image


Anyone see an error with these numbers?

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

Re: EMC2 has published a polywell preprint on arXiv

Post by D Tibbets »

tauntaun_rider wrote:
D Tibbets wrote:The KE of the electrons of somewhere between 2 and 7 KeV implies speed of eg: 5,000,000 M/s [....] This is based on a given speed for a 10 KeV electron of 10 million M/s.
I get 50,000,000 m/s for a 7 keV electron and 59,000,000 m/s for 10 keV. That makes a 7ns transit for a 7keV electron across a distance of 34cm, which looks like the size of the device.


Using the answer by King Freshy at this link I came up with an electron speed of 60,000,000 M/S for a KE of 10 KeV. His solution is simplest as you do not have to convert units. Just plug in the rest mass for the electron in eV.

https://answers.yahoo.com/question/inde ... 824AANdbtu

KE= 1/2 M V^2
10,000eV = (1/2) (500,000 eV) V^2
10,000eV/ (0.5* 500,000 eV) = V^2
0.04= V^2
0.2=V
This number is the speed of the electron expressed as a fraction of the speed of light
0.2 * c
0.2 *300,000,000 M/s
~60,000,000 M/S



So my number was obviously erroneous. I had just used Bussard's number . I had wondered if he was referencing the top speed or if he was referencing the average speed, but I had not pursued it further.

This brings up the further question of what is the average speed? The simple answer is that speed associated with a KE of 5,000eV. Either I recalled wrong, or Bussard was mistaken.
But, what is the average speed over the electrons travel across the machine? It may not be so simple. The mean speed would be at 5,000 eV , Velocity of ~ 42,000,000 M/S. Or would it be 60,000,000 M/s divided by two? For the mean to equal the average there can be no bias in the acceleration. But there is a bias. This is represented by the potential well. At startup, the potential well is almost square for the ions with most of the ion acceleration occuring on the edge . This is represented by the electrons clustering at the edge .As such the electron potential well would also have almost all of their acceleration in this area (except in the opposite direction). As pointed out in the Google Talk though, once ions are introduced the potential well is deformed and assumes a parabolic shape. The reason given is that because of the difference in their momentum the ions easily tug the electrons along (but not to the extent where the plasma would be considered closely coupled), while the opposite is not true. The square potential well would only be present in the pure electron mode before the gas puffer in WB6 or the plasma gun firing in this machine.

The parabolic shape can vary a lot so the detailed , and probably complex derivation of the acceleration curve for the electrons is an open question. I will assume that Bussard knew what he was talking about and came up with the 1 billion cm/S (10 million M/S) average electron speed for WB6.

With some assumptions this matches the 100,000 transits claimed for WB6. With a claimed confinement time (with recirculation) of ~ 2 milliseconds. The WB6 machine was ~ 30 cm wide. Assume the Wiffleball diameter was ~ 20 cm. 100,000 passes of 20 cm each = 20,000 M traveled in 0.002 seconds, or 10 million M/S.
So, I was wrong, except I wasn't. I was right for the wrong reasons. :) I think that this potential well parabolic shape would hold for any energy and density so long as the ion to electron ratio stayed the same, at least up until some point where collisions become so common that things are messed up. I think Bussard, or Nebel or someone said that this close coupling does not occur in fusion plasmas, weather you are talking about Tokamac or Polywell expected conditions of temperature and density. The average electron speed would be the same fraction of the peak speed for any machine, so WB6 numbers can be applied to the Mini Me machine.

So my approximate numbers given in the previous post hold, though only because I didn't know what I was talking about! :oops:

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

Post Reply