Google Polywell Fusion Counter

[jmc]

Anyway, you're quite right, in the end it doesn't matter how fast you progress, but whether you land on target or not. Competitive tokamak power plants have always been within the extrapolation uncertainties. Now we're getting down to the wire and that attractive extrapolation scenario is getting thinner and thinner. Some people think it has already disappeared. Others think it is worth one more shot to know for sure.

If you want to talk about fusion-fission hybrids, JET already HAS hit the required plasma performance and the diverter problems in JET are manageable. All that's really required is continuous current drive (again probably possible on a JET sized reactor) infact, they're engaging in a fully non-inductive current drive campaign as we speak, we won't get long plasmas until a JET equivalent with long-pulse neutral beams and super conducting toroidal field coils is built though.

One more shot? I certainly don't think if ITER fails we should close down the tokamak programme, all the failure of ITER should mean is we play around with some new advanced concepts for another few decades before trying again. And there are advanced concept out there, detached super-X divertors, edge impurity seeding to reduce flux on the divertor. Current drive methods haven't been optimised yet, lithium could boost performance and allow a for a smaller machine, liquid metal limters etc.
Needless to say if a super conductor was found which could operate at 30 Tesla at 100K say, everything would change for fusion, reactors could be made much smaller for example.

Even if ITER does work there are still plenty of rabbits left in the tokamak programme's hat. Whether it will be shut down for political reasons is another issue.

[Skipjack]

From what I saw spheromaks looks promising too...
They are like Tokamaks but smaller and -so the claims- more efficient.

[KitemanSA]

Oooh! Can I play, too?

Sure, but play nice. Some of those kids aren't as well edicated as you!

Don't know how accurate your component estimates are (I'd rather hear Rick Nebel's figures.), …

Me too, but what are ya gonna do?

but we can take them as a jumping off point.

Geronimo!

First let me check the power figure. I suppose we are talking about magnets with maybe 1 m diameter and 0.5 m bore. That should translate to R_plasma ~ 0.5 m.

Actually, more like 3m diameter for the WB-100, 30cm for the WB-7B (Bussard's WB-7, not the current one) but let it slide.

I would expect losses from 12 line cusps, …

Should’t be any line cusps, though the current circular plan-form coils make a “funny cusp” that looks too much like one.

each somewhat less than a quarter circumference long,

I’d say WAY less, given the squared plan-form coils proposed by Dr. B.

and an equivalent loss from the point cusps.

Ok, but I have no basis to accept or deny this.

That makes an effective length of about 2*12*(2pi*R/4), call it 15 m.

I’ll give you half of that , for the point cusps, but I don’t think the “funny cusps” are anywhere near that long.

Let's guess B = 1 T, so rho_e at 100 keV is around 1 mm.

Ok. I’ll presume you know this well enough that this is true.

Multiply that times 8 (the fourth root of the deuteron/electron mass ratio) for the hybrid gyroradius, assuming standard theory holds. Makes 8 mm cusp width.

Why in the world would you do this? Everything I have read about the Polywell says the ions don’t get anywhere near the MaGrid. The only things near the cusps are electrons.

Ouch. That's a gaping wound. Good thing we're only building an experiment, not a power reactor.

Not so gaping, but even then, so what? If the electrons are all returned by the MaGrid, there are no losses.

What's the power loss through the cusp, again assuming standard theory (or worst case, depending on your point of view)? An area of 0,1 m^2. An energy density of (1 T)^2/(2mu_0) ~ 4e5 J/m^3. A sound speed of about 2e6 m/s. Adds up to 100 GW. We're gonna need a bigger boat!

Using 1/8th the width, and ½ the length, and assuming 99%+ return efficiency on the electrons, we are down to ~the 40MW that Dr. B wrote of.

Did I make a big mistake somewhere? Possible. (Excuse: I'm sick, remember?) If not, then you need 10,000 of your power supplies (You could build ITER for that money!), to be sure you can definitively test the theory.

If your assumptions are correct, then your conclusion is reasonable. I just don’t think your assumptions are correct. I am hopeful that time will tell SOON!

Or you can just spend 10 million in the hope that the thing will work 10,000 times better than the theory says.

MSimon says $10M, me, I don't know. But I already got 1,600 times better than your theory (8x2x100). Multiply by the ~5 times improvement Dr.B expects from a less-quasi-more-spherical dodecahedral Magrid, and there you have it!

Actually, I had all kinds of clever things to say about your cost estimate, but considering the above result, it seems kind of pointless.

My cost estimate? The only thing I remember dicussing cost wise is the potential to do the squared coil WB-7 via non-for-profit funding. In another thread, someone suggested I was a pollyanne, but I just think I'm a polyweller.

[KitemanSA]

My latest hit count is 15,700. Are we losing ground? Someone said it had been 17,000.

[MSimon]

Again, I may be missing something, but everything I have seen suggests Polywell should work, and if it does, would get there bf&c.

I guess you haven't been reading my posts. I've spent a good deal of effort over the last six months explaining why, given everything I know about plasma physics, the polywell shouldn't work.

Can anyone articulate an experiment that would falsify either proposition? I mean something cheaper than building a fully-operational Wiffleball-N?

I suppose I'm asking if you physics boffins agree on the characteristics required of plasma in order for Polywell to work or not? Then how can we demonstrate plasma lacks one of them?

Spend the money.

[Art Carlson]

Again, I may be missing something, but everything I have seen suggests Polywell should work, and if it does, would get there bf&c.

I guess you haven't been reading my posts. I've spent a good deal of effort over the last six months explaining why, given everything I know about plasma physics, the polywell shouldn't work.

Can anyone articulate an experiment that would falsify either proposition? I mean something cheaper than building a fully-operational Wiffleball-N?

I suppose I'm asking if you physics boffins agree on the characteristics required of plasma in order for Polywell to work or not? Then how can we demonstrate plasma lacks one of them?

Good question. Unfortunately, I don't see any cheap way to test whiffle ball theory without building a whiffle ball. We really only need a single point cusp or a short section of line cusp, but it's hard to get that without building at least a spindle cusp, which should only be about a factor of 3 less power-hungry than the cubic job. We presumably need beta=1 and a cusp thin compared to the device dimensions. Then all my calculations above apply. We could trim a bit more off the cost by going to still lower temperature, say 10 eV at rock bottom, but atomic physics, especially with argon operation, will make the physics very dirty.

10 or 15 years ago, the tokamak community thought long and hard about building a simple machine to test the physics of divertors at realistic power levels. It turned out there was no way to do it exactly right. The best way to simulate a divertor plasma is with a divertor, even if you are not interested in the tokamak attached to it.

[Art Carlson]

As long as almost everything that goes out the holes comes back in (electrons) and if you have a way to prevent upscattering then it MIGHT work. The theory is you have to pump in electrons to make up for the losses.

What evidence do we have for such a statement? Bussard's publications are so full of holes that they are worthless. Nebel is saying something to the effect that there are no show-stoppers, but that is considerably less than saying he has evidence that the standard theory breaks down. Until he publishes, we have nothing we can use from him either. I have given some theoretical arguments why I don't think it should work. If I'm wrong, then somebody should be able to sketch out a density/potential/field distribution with the desired properties. Nobody has done that, so there is zero theoretical evidence. "If these holes don't leak like other holes leak, then we have it made!" Yippee.

[Art Carlson]

I don't know where the idea came from that tokamaks haven't been successful. The triple product has doubled every 3 years for decades. Breakeven (Q=1) has been achieved. That's impressive in my book. No other concept has come close.

Moore's law has the density of ICs doubling every two-ish years... at the same (or lower) price. Tokomak seems to double every 3 years at 10 times the price.

p.4 of this PPT shows a graph you might find interesting entitled "Progress in Fusion Energy has Outpaced Computer Speed".

[MSimon]

First let me check the power figure. I suppose we are talking about magnets with maybe 1 m diameter and 0.5 m bore. That should translate to R_plasma ~ 0.5 m.

Actually, more like 3m diameter for the WB-100, 30cm for the WB-7B (Bussard's WB-7, not the current one) but let it slide.

Art is right on this point. MRI magnets. Modified. COTS as much as possible.

[MSimon]

Nebel is saying something to the effect that there are no show-stoppers, but that is considerably less than saying he has evidence that the standard theory breaks down.

Rick also stated that the experiment works as expected. Just as Dr. B. predicted. That is a little stronger than "no show stoppers".

[Billy Catringer]

I have looked at all the fusion reactor designs I could find. None of them seem to be all that promising. Nearly all of them look like something Rube Goldberg dreamed up while he was high on a mix of THC, steroids and California wine. American boffins have been working on a way to fuse D-D or D-T since what, 1953? US taxpayers have shelled out several hundred billion dollars by now and I don't know how much money other countries have spent, but you can bet it is a VLCC load of red ink.

Here is the real rub as I see it. Let's say that something like ITER eventually works. Will we make use of it? I think the answer to that question is a resounding, "NO!" Why? Because we will play hob ever maintaining one of these monsters. Worse, I don't think you could build one that will ever earn its initial investment back.

The maximum time between major turnarounds on one of these beasts is going to be nine months to a year at best. Mind you, we are not talking about just radioactive stuff here, we are talking about the use of a very aggressive alkali metal that will have tritium bubbling out of it. Fixing one of these monsters ain't gonna be no fun. If one of them loses that molten lithium blanket while in operation, we are going to be in for some real excitement.

You know what else? It looks to me as though we will need a fairly good sized fission plant or coal fired plant to crank one of these big ITER derived units. What do we do with this monstrous great auxiliary plant? Shut it down until we need to restart the fusion plant? The more I look at these things, the less I like them. The bankers for danged sure ain't gonna like 'em.

So what about Doctor Bussard's whiffle-ball design? He said that the science is done and that it is time to start the engineering. Okay, but it looks like some pretty hairy engineering to me. Just controlling it is going to be tough, even with computers and modern instrumentation. But here is the big but. If it works, especially if it can take advantage of the p-11B reaction, its sins are easily forgiven. If we spend $200 million only to find out that it cannot be made to work, so what? We have already spent hundreds of billions on something that we can already see will NEVER be used.

All things considered, I have no trouble taking Doctor Bussard at his word. If he was wrong, he was honestly mistaken and our money is being thrown down so many rat holes already that we can easily afford the $200 million he said is needed to sort the thing out. The Space Shuttle program was an honest mistake and a single shuttle cost something like $2 billion back in the 1980's. We can easily afford a few billion on a BFR. Considering what is at stake, a BFR plant is cheap at $200 billion in R&D costs.

[MSimon]

Just controlling it is going to be tough, even with computers and modern instrumentation.

Nah. I've been working on controller design for about a year now. D-D will be very easy: short term (30 uS) fluctuations under 50% - control the high voltage. Longer term (100mS to S) vary the reactant inventory in the reactor.

I can do 100 M samples per second with a 16 bit A to D at a not unreasonable cost. I can buy or make out of FPGA (depending) a processor fast enough to to handle that kind of data rate. Which means I can update my power supply PID algorithm every microsecond. Given a 3 db roll off of the power supplies at 3 KHz that means I can very easily get 300 new outputs in one cycle of 3 KHz. Very smooth control indeed. Control accuracy could be out 12 bits or better (10 is probably enough).

If/when we go to pB11 I have some ideas for tighter control of the power supply voltage so operation on the resonance peak becomes possible.

Gas valving for D-D will be a relative snap. H2 the same. What is going to cost is B11 feed. No one has a idea that seems workable. And then you have the problem of B11 plating out on every surface.

[Billy Catringer]

Opening and shutting valves quickly enough might well be done easily, but what about shutting down and restarting the pumps? As best as I can understand this beast, you have to keep the feed very precisely balanced with the exhaust while making certain that all the electrical stuff does its job.

Don't get me wrong, this is great news if all the bases have been covered and I'm tickled to death.

As for the boron, I think we will have to resort to one of the boranes and a helicon or some other microwave transmitter and wave guide. When I looked up Boron I was shocked at what a truly refractory material it is.

[Tom Ligon]

Billy,

The pumping system would probably be throttled to a desired steady state, likely wide open. Neutral gas is the enemy of the system and you want to get rid of all you can. A typical temperature chamber does something like this, running the refrigeration system full on, or cycles slowly on and off, while doing the fine control with heat.

Introducing neutral gas can be regulated by very fast valves, one example being piezo valves. In principle, you could operate those at tens of megahertz.

Ion guns can be regulated more or less like vacuum tubes, with bandwidths at least to the hundreds of megahertz. Traveling wave tube techniques can get you well into the microwaves.

I've always been interested in some means of clearing out spent ions and electrons. These may also involve some form of electronic control.

[MSimon]

Opening and shutting valves quickly enough might well be done easily, but what about shutting down and restarting the pumps? As best as I can understand this beast, you have to keep the feed very precisely balanced with the exhaust while making certain that all the electrical stuff does its job.

Don't get me wrong, this is great news if all the bases have been covered and I'm tickled to death.

As for the boron, I think we will have to resort to one of the boranes and a helicon or some other microwave transmitter and wave guide. When I looked up Boron I was shocked at what a truly refractory material it is.

Gas inventory in an operating 100 MW reactor is about 1 second. Controlling that will be a snap. A control response of 10 mS should keep the gas within 10% or better of the central value. For steady state operation that might get to around .1%. Going to zero as the errors are integrated out.

I don't like the gaseous boranes much. There is one that boils at 215C I think that should be the least hazardous to handle.

I was thinking boron metal feed with laser vaporization. Much less hazardous. OTOH the fusion soup might produce boranes as a byproduct. We shall see. Maybe.