Why Not Just Build The Darn Thing

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

Moderators: tonybarry, MSimon

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

Post by MSimon »

chrismb wrote:
MSimon wrote: Well I tried. You multiply a bunch of numbers to together and come up with an answer. I did that and I came up with 2. So right away there is a fundamental difference.
Ah! Fair enough. I've played butter-fingers on the calculator. Yup, the calc comes out as 2.6W. Given that the pulse length assumed was 250E-6s, so that is a total output of 650uJ for the pulse.

Apologies for the confusion. Do check the arithmetic, that's why I pin it up line by line for these type of corrections.

Getting back to the question of how long is the residence time for an ion is before a fusion event, you see that if it were, say, 10,000 reciprocations instead of my 2.5 billion reciprocations, so that calculation would be x250,000 and so the fusion power out would have been 650kW. I'm pretty sure it wasn't, hence I'm pretty sure the residence time is a lot longer than 10,000 reciprocations both for theoretical and experimental reasons.
Fair enough. It still doesn't matter. As long as the Output power is sufficiently greater than the input power.

So how about a thought experiment? With large GWB factors expected in a 10T machine (~ 1 million) and assume for convenience 3 million reciprocations for fusion and further 100 KeV the cost of acceleration.

Which says that you lose 2 X 100 KeV particles for every fusion (total 400 KeV - two lost particles - one fusion - two more lost) to produce an 8 MeV event. A gain of roughly 20. Not counting other losses.

It is the "other losses" that concern me. And of course long term (seconds) stability of the reactor.
Engineering is the art of making what you want from what you can get at a profit.

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

Post by MSimon »

Some calculations:

1/[(1E20/m^3).(2E-30m^2)]=5billion metres

1/x * y = distance

x = (1E20/m^3)

y = (2E-30m^2)

Now please: tell what x and y are and how did you derive 1/x*y = z
Engineering is the art of making what you want from what you can get at a profit.

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

Post by KitemanSA »

MSimon wrote:Some calculations:

1/[(1E20/m^3).(2E-30m^2)]=5billion metres
1/x * y = distance
The first equation looks more along the line of 1/[x*y]=distance which doesn't look right. If it was your way, why didn't he just put distance = y/x ?

chrismb
Posts: 3161
Joined: Sat Dec 13, 2008 6:00 pm

Post by chrismb »

x is particle density and y is reaction cross section. The mean free path to a reaction is 1/(x.y). It is a general formula for any interactions and is straightforwardly derived from the very nature of 'density' and 'cross-section', if you think about it.

chrismb
Posts: 3161
Joined: Sat Dec 13, 2008 6:00 pm

Post by chrismb »

MSimon wrote: So how about a thought experiment? With large GWB factors expected in a 10T machine (~ 1 million) and assume for convenience 3 million reciprocations for fusion and further 100 KeV the cost of acceleration.

Which says that you lose 2 X 100 KeV particles for every fusion (total 400 KeV - two lost particles - one fusion - two more lost) to produce an 8 MeV event. A gain of roughly 20. Not counting other losses.

It is the "other losses" that concern me. And of course long term (seconds) stability of the reactor.
Sure. Why not. As you say, the "other losses" are the issue. If there are no high energy particle losses through thermalisation, great, we're on to a winner!

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

Post by D Tibbets »

chrismb wrote:
MSimon wrote: Well I tried. You multiply a bunch of numbers to together and come up with an answer. I did that and I came up with 2. So right away there is a fundamental difference.
Ah! Fair enough. I've played butter-fingers on the calculator. Yup, the calc comes out as 2.6W. Given that the pulse length assumed was 250E-6s, so that is a total output of 650uJ for the pulse.

Apologies for the confusion. Do check the arithmetic, that's why I pin it up line by line for these type of corrections.

Getting back to the question of how long is the residence time for an ion is before a fusion event, you see that if it were, say, 10,000 reciprocations instead of my 2.5 billion reciprocations, so that calculation would be x250,000 and so the fusion power out would have been 650kW. I'm pretty sure it wasn't, hence I'm pretty sure the residence time is a lot longer than 10,000 reciprocations both for theoretical and experimental reasons.
Just to add some more confusion to confusing numbers-
CB, you quoted output for WB6 of 650 micro J for the pulse. Do you mean 650 micro watts? That number would be close to the ~ 1 billion fusion per second =~ few hundred micro watts which were the reported results for WB6. In any case the new numbers would allow the expermental results to be well within the limits allowed by your predictions.


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

chrismb
Posts: 3161
Joined: Sat Dec 13, 2008 6:00 pm

Post by chrismb »

D Tibbets wrote: Just to add some more confusion to confusing numbers-
CB, you quoted output for WB6 of 650 micro J for the pulse. Do you mean 650 micro watts? That number would be close to the ~ 1 billion fusion per second =~ few hundred micro watts which were the reported results for WB6. In any case the new numbers would allow the expermental results to be well within the limits allowed by your predictions.


Dan Tibbets
No. I mean 650uJ per pulse. The prediction for 2.5E16 ions each undergoing a fusion every 7000s is 3.5E12 (3.5 trillion) reactions per second. If the rate of reactions is some billion/s, then either the time-to-fusion is even more than 7000s or the total number of ions participating is less than 2.5E16, or some mix thereof. Remember I was being wholly generous with my numbers to show that my estimate for the residence time for an ion to undergo fusion was 7000s even whilst being overly optimistic. You should expect my generosity of figures to end up with a higher figure than measured experimentally.

I would suggest the figures are probably more like time-to-fusion = 14000s (because I was giving the highest experimental measures for cross-section, experimental measures which differ by ~100% at these low fusing energies) and around 1E14 for the total ion participation as only some 1% of the electrical discharge is likely to actually get ions moving on fusible trajectories. Then the numbers fit.

I hope you see how bad that comes out - only a thousandth of the electrical energy put in actually gets ions moving, and then they have to reciprocate around for 4 hours before a fusion event. You then get reaction rates as seen, in the billions per second. Maybe it gives a hint as to why I think fast neutrals into the chamber wall is a strong contender for a goodly fraction of the overall reaction rate.

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

Post by D Tibbets »

chrismb wrote:
D Tibbets wrote: Just to add some more confusion to confusing numbers-
CB, you quoted output for WB6 of 650 micro J for the pulse. Do you mean 650 micro watts? That number would be close to the ~ 1 billion fusion per second =~ few hundred micro watts which were the reported results for WB6. In any case the new numbers would allow the expermental results to be well within the limits allowed by your predictions.


Dan Tibbets
No. I mean 650uJ per pulse. The prediction for 2.5E16 ions each undergoing a fusion every 7000s is 3.5E12 (3.5 trillion) reactions per second. If the rate of reactions is some billion/s, then either the time-to-fusion is even more than 7000s or the total number of ions participating is less than 2.5E16, or some mix thereof. Remember I was being wholly generous with my numbers to show that my estimate for the residence time for an ion to undergo fusion was 7000s even whilst being overly optimistic. You should expect my generosity of figures to end up with a higher figure than measured experimentally.

I would suggest the figures are probably more like time-to-fusion = 14000s (because I was giving the highest experimental measures for cross-section, experimental measures which differ by ~100% at these low fusing energies) and around 1E14 for the total ion participation as only some 1% of the electrical discharge is likely to actually get ions moving on fusible trajectories. Then the numbers fit.

I hope you see how bad that comes out - only a thousandth of the electrical energy put in actually gets ions moving, and then they have to reciprocate around for 4 hours before a fusion event. You then get reaction rates as seen, in the billions per second. Maybe it gives a hint as to why I think fast neutrals into the chamber wall is a strong contender for a goodly fraction of the overall reaction rate.
OK, I figured my watts- joules argument was foolish, but it did serve to reenforce my understanding.

Some questions-
The 10,000 ion reciprications to a significant chance for a fusion event has been tossed around. But, what is significant- 99% chance, or 50%, or 1%, or 0.01%?. A 1% chance would be equivbalent to ~ 1 million reciprications for a very high probability of fusion (assuming a linier relationship). I assume your ~ 2.5 billion reciprications is the mean (50%) chance, so for a 1% chance 50 million reciprications would be needed (?).

Where does your 1/1000th efficiency of induced ion energies come from. Is it based on thermal tail in Tokamak type Maxwellian plasmas? I'm guessing that so long as the electron current is high enough to maintain the potential well, almost all ions in the Polywell reach the kinetic energy of the potential well. That is one of the claimed advantages of the Polywell over Maxwellian systems.

There is a difference between energy lost and ions lost. If an ion is lost near the top of it's potential well, the energy loss may be trivial. A thousand low energy ions may be lost, but the energy loss may add up to that in only one high energy ion. What happens to the low energy ion once it exits the magrid is confusing. Wouldn't it be accelerated away from the magrid, regaining it's kinetic energy? If this is irrelavent, then the numbers of low energy ions escaping is only important in regards to how fast you can pump them out of the system. I'm assuming the positively charged magrid is 'neutralized' by the nearly as strong negative electron charge inside the magrid so that the escaped ions do not pick up too much energy while outside the magrid.

Again, I understand that neutrals will be relatively rare inside the magrid so beam- neutral fusion collisions will be insignificant, and there are no (minimal) surfaces exposed inside the magrid so beam - target fusions will also be insignificant (as opposed to gridded fusors in which the deuterium loaded wires can serve as a target). So, where in the Polywell can beam -neutral or beam - target fusions contribute to the total?

Comparing the Polywell against gridded fusors: If the ions can recripricate 100,000 times before escaping and hitting something compared to a gridded fusor's 100 reciprication life time (in a very good fusor) then there is a 1000 fold advantage. In a fusor the ion loses most of it's energy, which has to be replaced in a newly ionized ion harvested from the background neutrals. In the polywell, assume only 1% of the energy is lost. That is another 100 fold advantage. Assumeing the Polywell can have a 1000 fold advantage in the density it can maintain in the reaction space (this does not concider convergance/ focusing which might effectively increase this density even more). This adds up to a 100,000,000 fold advantage for the Polywell. This is nearing the area of breakeven. Then, add the claimed advantages of monoenergetic ion energies (higher percentage of ions at the desired kinetic energy compared to the fusors where the ion velocities ar more diffuse (and lower average) due to dominate neutral and structure collisions ), confluence, possibly higher reciprication numbers/ density from even higher Wiffleball traping factors. At least from a hand waving perspective, doesn't this support arguments that the Polywell can exceed breakeven once certain size and magnetic conditions are met?
Bremsstrulung , if controlable as claimed, and other losses would not preclude breakeven, only influence the size and magnetic strength needed.

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

chrismb
Posts: 3161
Joined: Sat Dec 13, 2008 6:00 pm

Post by chrismb »

I'm not aware I've ever said Polywell won't work. I've just raised some of the issues with the theory of its operation so as I can understand the logic of why they aren't seen as issues. If what you're saying holds together then, sure, it means we're on to a winner and the experimental results will speak for themselves.

(Those are some pretty big 'ifs' though, do you not think?)

I'm afraid I don't see why the thing won't theramlise almost immediately, though, and I've never understood why 'annealing' doesn't happen in a fusor if it is going to work in Polywell. If these things aren't uniquely resolved by Polywell's mechanisms then your 'ifs' will unravel, so we wait for the evidence one way or the other....

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

Post by MSimon »

chrismb wrote:I'm not aware I've ever said Polywell won't work. I've just raised some of the issues with the theory of its operation so as I can understand the logic of why they aren't seen as issues. If what you're saying holds together then, sure, it means we're on to a winner and the experimental results will speak for themselves.

(Those are some pretty big 'ifs' though, do you not think?)

I'm afraid I don't see why the thing won't theramlise almost immediately, though, and I've never understood why 'annealing' doesn't happen in a fusor if it is going to work in Polywell. If these things aren't uniquely resolved by Polywell's mechanisms then your 'ifs' will unravel, so we wait for the evidence one way or the other....
I think magnetic confinement makes the difference and the density vs radial distance that happens in such a configuration.

But as you point out - it is hard to tease out on strictly theoretical grounds if the device will work as a commercial net power machine. About 9 months to the next contract way point. We should see a blizzard of new government paper around then. If the current experiments show promise.
Engineering is the art of making what you want from what you can get at a profit.

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

Post by mvanwink5 »

Simon, re: "About 9 months to the next contract way point. We should see a blizzard of new government paper around then. If the current experiments show promise."

Any chance of intermediate rumors? :)

Cheers (yes, I know you are patient!)
Counting the days to commercial fusion. It is not that long now.

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

Post by MSimon »

mvanwink5 wrote:Simon, re: "About 9 months to the next contract way point. We should see a blizzard of new government paper around then. If the current experiments show promise."

Any chance of intermediate rumors? :)

Cheers (yes, I know you are patient!)
If Rick has anything useful to say he says it. Either in a public forum or here.

My estimation is that he reports when he has some slack time. I'm guessing that he has been busy for the last few months.
Engineering is the art of making what you want from what you can get at a profit.

krenshala
Posts: 914
Joined: Wed Jul 16, 2008 4:20 pm
Location: Austin, TX, NorAm, Sol III

Post by krenshala »

While I'd love to hear more (current) news about the experimentation and what they've found, the fact he's busy enough to not really have time to post here has to be considered a good thing for the Polywell.

Helius
Posts: 465
Joined: Sun Oct 21, 2007 9:48 pm
Location: Syracuse, New York

Post by Helius »

krenshala wrote:While I'd love to hear more (current) news about the experimentation and what they've found, the fact he's busy enough to not really have time to post here has to be considered a good thing for the Polywell.
It's a good thing for the *science* of it; Maybe he's documenting on how it's a blowout... Time will tell.

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

Post by TallDave »

MSimon wrote: My estimation is that he reports when he has some slack time. I'm guessing that he has been busy for the last few months.
Ooooooh, I hope so..............

Do you think WB-8 could be constructed already?

Maybe they're doing a test run right now....

/drool

Post Reply