MS: "If most of the particles are monoenergetic"
I'm more sympathetic with AC's view that it will all thermalise. This seems a much more realistic treatment, but poses what appears at first sight some unsurmountable issues down the road at higher temps. I see no reason at all now that the energy would be maintained in a particle as it sheds its energy through inelastic collisions
MS: "if you operate the machine using D-D with a 50 KV well (60 to 80KV drive depending) the collision temperature (head on) would be equivalent to 11,605 * 200,000. Roughly 2.3 billion degrees. The cross section for D-D at that velocity is about .1 barns. "
I feel it's just an academic detail in this discussion, but actually you'd end up with a collision energy of 50keV+50keV=100keV. If you then go to nuclear collision data, then at that point you'd probably look up the '200keV' data [depends on the inertial frame the data is prepared for] because that is the rest-frame energy for a fast-stationary collision as per the experiment from which it is taken; which is 35millibarns for the 3He+n and another 35mb for the T+p.
MS: "Now if it is really a colliding beam machine - during most of the cycle there would be no radiation. So you only get significant radiation when you are producing fusions."
Not really. The radiative inelastic collisions are down to the behaviour of the relativistic electrons. It won't matter much what the ions are up to.
MS: "Let me add that one of the reasons I love this device is that the engineering problems are both interesting and hellacious."
You must like the tokamak research as well, then. They even have neutron output data that you can study all day and not run out of more to study.
MS: "So I'm really hoping the physics pans out."
I wish it [Poly] well!
best regards,
Chris MB.
A few questions on Polywell facts and figures.
We have been through this before.
A 50 KV beam will have a velocity V. Since it is a non-relativistic velocity two beams will collide at V+V or 2V. Since energy is proportional to V^2 if the collisional velocity is 2V the energy will be proportional to 4V^2 .
You seem rather knowledgeable. I'm surprised you missed that.
The problem with all of this is that we do not have near enough data to even begin engineering. I am in semi-regular contact with Rick Nebel (the last a few days ago) and he hasn't even hinted at his results. As best we can tell from what he posted here is that so far he has seen what he expected i.e. WB-6 has been replicated.
A 50 KV beam will have a velocity V. Since it is a non-relativistic velocity two beams will collide at V+V or 2V. Since energy is proportional to V^2 if the collisional velocity is 2V the energy will be proportional to 4V^2 .
You seem rather knowledgeable. I'm surprised you missed that.
Well it gets complicated. If the ions are dragging the electrons then it is less of a problem.Not really. The radiative inelastic collisions are down to the behaviour of the relativistic electrons. It won't matter much what the ions are up to.
The problem with all of this is that we do not have near enough data to even begin engineering. I am in semi-regular contact with Rick Nebel (the last a few days ago) and he hasn't even hinted at his results. As best we can tell from what he posted here is that so far he has seen what he expected i.e. WB-6 has been replicated.
Engineering is the art of making what you want from what you can get at a profit.
I did miss it. Quite right! End of a long day..MSimon wrote: You seem rather knowledgeable. I'm surprised you missed that.
I do keep making that silly mistake because I tend to think mostly about fast ions into lab-stationary ones, the collision energy for which is a half of the fast ion's energy (because there are 2 x ions come towards the CofM at half the speed, thus is 2 x 1/2^2). So the actual collision energy for a fast into stationary particle is a half of the drive voltage, yet two fast ions it is four times the drive voltage. I simply reversed that logic in my addled brain.
So, your are quite right and I duly retract my error.
-
Art Carlson
- Posts: 794
- Joined: Tue Jun 24, 2008 7:56 am
- Location: Munich, Germany
"Held together"? Equilibrium is not a problem. That's just getting the magnetic pressure to balance the plasma pressure. Global stability is not a problem either. Confinement is the killer. Particles like to leak out the cusps, and they take their energy (at least most of it) with them.chrismb wrote:You are saying that this will be a 6 billion Celsius thermalised plasma, with a central core at 1/10th of atmospheric pressure, and all this is to be held together with some cusping magnetic fields.
Bremsstrahlung power, if you can keep the plasma clean, is only a couple percent of the fusion power. (Of course, that's for the D-T fuel cycle. For the fairy tale fuel cycle of p-B11, bremsstrahlung is greater than fusion power and the game is over.)chrismb wrote:The radiative inelastic collisions, (recombination, brems, etc..) in this setup will be simply ENORMOUS! What do you think would be the power loss from such a configuration? I can't remotely see any mechanism for keeping enough energy in this to maintain any nuclear processes.
Atomic line radiation, except of high-Z impurities, is non-existent at that temperature because you have 100% ionization.
Radiative recombination doesn't happen either.
But, yeah. Those cusps holes are just a bit too big. The fusion output scales up faster than the cusp losses (especially if you assume B can scale up with R), but by the time you get to break-even you would have a humongous machine. You would neither be able to afford to build one, nor would you be able to build a wall that could survive the power load.
Well, OK, my phrase was a little vague - 'held in place' was the intent.Art Carlson wrote: "Held together"? Equilibrium is not a problem.
Ah, I see. Your comment/interest is with a view to Polywell containing thermalised DT!? I can see that, that makes the leap from currently understood physics easier to understand without picking up on so many details.Art Carlson wrote:
Bremsstrahlung power, if you can keep the plasma clean, is only a couple percent of the fusion power. (Of course, that's for the D-T fuel cycle. For the fairy tale fuel cycle of p-B11, bremsstrahlung is greater than fusion power and the game is over.)
The advantage being then, I guess, that this is the next best idea we can hope for next to an impossible spherical magnetic surface. At least some of the losses through the cusps find their way back into the device again.
But though a spherical magnetic surface is impossible, a toroidal magnetic surface is quite fine and easy to accomplish with a toroidal solenoid, which is also convenient as it can be installed outside of the volume. So I don't quite get it, what is the reason for wanting to create a 'quasi-spherical' magnetic surface when a perfectly good toroidal surface can be created with relative ease?
I will add, also, that I have read the various documents on Polywell and following the recent posts I now realise I simply glossed over something that I presumed I did not understand: As it is impossible to create a stable spherical magnetic surface I presumed I had misunderstood the 'wiffleball' effect. But I now gather from these posts that the idea *is* that the Polywell can do the impossible and form a magnetic spherical surface. Do I remain mistaken on this point? Is the idea of the wiffleball that there is a continuous *outer* spherical magnetic surface with nothing inside it?
-
Art Carlson
- Posts: 794
- Joined: Tue Jun 24, 2008 7:56 am
- Location: Munich, Germany
Participating on this forum I keep being reminded (inadvertently) what a brilliant idea the tokamak was. What you would like is to confine a plasma with perpendicular magnetic fields everywhere. The hairy billiard ball theorem of topology tells you that is impossible on a sphere, but it is possible on a torus! As much as the high-beta nature of the polywell is praised (understandably), you might also ask if you get more benefit by filling your confinement volume with as much plasma as possible, or whether it is better to have some internal magnetic insulation. Somewhere around this point we have to start doing detailed physics, and soon thereafter actual experiments. The proof of the pudding is the eating thereof.chrismb wrote:The advantage being then, I guess, that this is the next best idea we can hope for next to an impossible spherical magnetic surface. At least some of the losses through the cusps find their way back into the device again.
But though a spherical magnetic surface is impossible, a toroidal magnetic surface is quite fine and easy to accomplish with a toroidal solenoid, which is also convenient as it can be installed outside of the volume. So I don't quite get it, what is the reason for wanting to create a 'quasi-spherical' magnetic surface when a perfectly good toroidal surface can be created with relative ease?
I will add, also, that I have read the various documents on Polywell and following the recent posts I now realise I simply glossed over something that I presumed I did not understand: As it is impossible to create a stable spherical magnetic surface I presumed I had misunderstood the 'wiffleball' effect. But I now gather from these posts that the idea *is* that the Polywell can do the impossible and form a magnetic spherical surface. Do I remain mistaken on this point? Is the idea of the wiffleball that there is a continuous *outer* spherical magnetic surface with nothing inside it?
The polywell, even if you call it a whiffle ball, can't get around topology. It has to have cusps. Not only that, it has to have line cusps (or something very similar, although I recently figured out that there wouldn't really be much difference even if you could make purely point cusps).
So.... I'm very sorry, but I don't get this at all.
It is my improved understanding now, following the helpful replies to my earlier emails, that the principle means for accelerating ions in the Polywell is that you feed a mass of electrons and ions with more electrons so that it is off-neutrality a little to created an internal field, subject it to a high external e-field forcing an electric field gradient which accelerates ions and then.... constrain all of this in a cusped magnetic field.
On behalf of Peter Debye, I express a little disquiet over the first part of this, but no-one has strictly equivalent actual experimental evidence that it will or won't work.
So that bit doesn't matter. Maybe it'll work, maybe it won't, let's test it. Fine.
The bit I don't understand, then, is why test this in a quasi-spherical cusped magnetic field when you could test this mechanism equally in a reliable and easily generated toroidal magnetic field?
That's not to say "in a tokamak". But a toroidal field can be exploited in many ways. Why not expolit it with the Polywell's ion-acceleration idea?
If the first bit, the mechanism of acceleration, works as labelled then the electrons would migrate to the major circumference (the centre in poloidal cross-section, just as a sphere would look in section) and so you'd get radially poloidal ion accelerations. The "e-field/ion acceleration" part of the Polywell's mechanism is topologically transformable into this configuration, so why bother trying to generate an impossible spherical magnetic surface when it should work for an easily generated toroidal field?
It is my improved understanding now, following the helpful replies to my earlier emails, that the principle means for accelerating ions in the Polywell is that you feed a mass of electrons and ions with more electrons so that it is off-neutrality a little to created an internal field, subject it to a high external e-field forcing an electric field gradient which accelerates ions and then.... constrain all of this in a cusped magnetic field.
On behalf of Peter Debye, I express a little disquiet over the first part of this, but no-one has strictly equivalent actual experimental evidence that it will or won't work.
So that bit doesn't matter. Maybe it'll work, maybe it won't, let's test it. Fine.
The bit I don't understand, then, is why test this in a quasi-spherical cusped magnetic field when you could test this mechanism equally in a reliable and easily generated toroidal magnetic field?
That's not to say "in a tokamak". But a toroidal field can be exploited in many ways. Why not expolit it with the Polywell's ion-acceleration idea?
If the first bit, the mechanism of acceleration, works as labelled then the electrons would migrate to the major circumference (the centre in poloidal cross-section, just as a sphere would look in section) and so you'd get radially poloidal ion accelerations. The "e-field/ion acceleration" part of the Polywell's mechanism is topologically transformable into this configuration, so why bother trying to generate an impossible spherical magnetic surface when it should work for an easily generated toroidal field?
Last edited by chrismb on Tue Dec 16, 2008 12:39 pm, edited 1 time in total.
Art, yes, the tokomak is so brilliant it has taken already 50 years to demonstrate that it might be possible to build one that might one day work, given many more billions. And, oh yeah, it is a globally unstable flow topology. For now, it has proven to be a blind alley of holy grail proportions, how many more resources does one throw at it to prove such a brilliant idea can actually work?
If the wiffleball effect is a true spherical inversion of the magnetic field carried by the plasma, as posited here,
viewtopic.php?t=650&postdays=0&postorder=asc&start=60
then the cusps effectively close, and it is game on with a near perfect field topology for containment being satisfied.
A big "IF" maybe, but any bigger than stabilising a tokomak flow?
If the wiffleball effect is a true spherical inversion of the magnetic field carried by the plasma, as posited here,
viewtopic.php?t=650&postdays=0&postorder=asc&start=60
then the cusps effectively close, and it is game on with a near perfect field topology for containment being satisfied.
A big "IF" maybe, but any bigger than stabilising a tokomak flow?
Last edited by icarus on Tue Dec 16, 2008 12:58 pm, edited 1 time in total.
It is possible you misunderstand my point. The issue is not 'the tokamak'.
The tokamak is a device which exploits toroidal magnetic fields. It isn't *just* a toroidal field - you have to monkey around with other sutff aswell, like generate a toroidal current.
I'm saying that the Polywell can also exploit a toroidal magnetic field. It wouldn't be a tokamak, there would be no [intentional] toroidal currents, nor barrier transport issues, nor divertors, nor helical safety factors, &c. &c.. it would just be those radial currents in poloidal cross-section. Just like the Polywell in section. This would NOT be a tokamak.
I cannot disprove potential experimental issues, like the ion transport, but the one thing I can PROVE is that an uncusped spherical magnetic surface is impossible. This is a definite, well understood mathematical proof, with no degree of experimental uncertainty whatsoever. So why bother with it. Why not exploit the total coverage of a toroidal magnetic surface? Just to emphasise - this would NOT BE a tokamak, it would be a toroidal 'Polywell' (of course, it would be a complete surface, so it would be a 'Uniwell', perhaps?!)
The tokamak is a device which exploits toroidal magnetic fields. It isn't *just* a toroidal field - you have to monkey around with other sutff aswell, like generate a toroidal current.
I'm saying that the Polywell can also exploit a toroidal magnetic field. It wouldn't be a tokamak, there would be no [intentional] toroidal currents, nor barrier transport issues, nor divertors, nor helical safety factors, &c. &c.. it would just be those radial currents in poloidal cross-section. Just like the Polywell in section. This would NOT be a tokamak.
I cannot disprove potential experimental issues, like the ion transport, but the one thing I can PROVE is that an uncusped spherical magnetic surface is impossible. This is a definite, well understood mathematical proof, with no degree of experimental uncertainty whatsoever. So why bother with it. Why not exploit the total coverage of a toroidal magnetic surface? Just to emphasise - this would NOT BE a tokamak, it would be a toroidal 'Polywell' (of course, it would be a complete surface, so it would be a 'Uniwell', perhaps?!)
Chris,
If it is an oscillating beam machine and not a thermal machine a toroidal configuration would be sub-optimal.
Hot random vs fast aimed.
Sadly all we have to go on at this point is hand waving and unverified simulations. The simulations I have seen indicate beam formation. Bussard had an interesting theory about edge and center annealing (reducing the rate of thermalization). I lean more towards bunching effects for annealing. I think some AC drive added to the DC grid voltage (possibly by a power amplifier - possibly by a reasonably high Q LC in series with the DC drive just before it enters the machine) could enhance this effect.
Really though, we are starved for data.
And just for fun:
Plasma Physicist Dr. Nicholas Krall said, "We spent $15 billion dollars studying tokamaks and what we learned about them is that they are no darn good."
Krall is keeing an eye on Polywell developments. He thinks it has promise. But he does have reservations.
If it is an oscillating beam machine and not a thermal machine a toroidal configuration would be sub-optimal.
Hot random vs fast aimed.
Sadly all we have to go on at this point is hand waving and unverified simulations. The simulations I have seen indicate beam formation. Bussard had an interesting theory about edge and center annealing (reducing the rate of thermalization). I lean more towards bunching effects for annealing. I think some AC drive added to the DC grid voltage (possibly by a power amplifier - possibly by a reasonably high Q LC in series with the DC drive just before it enters the machine) could enhance this effect.
Really though, we are starved for data.
And just for fun:
Plasma Physicist Dr. Nicholas Krall said, "We spent $15 billion dollars studying tokamaks and what we learned about them is that they are no darn good."
Krall is keeing an eye on Polywell developments. He thinks it has promise. But he does have reservations.
Engineering is the art of making what you want from what you can get at a profit.
chrismb: your Uniwell!? idea, someone else has already suggested this idea on another thread here somewhere ... start trolling... probably call it a "Toriwell" or a "TokoWell" and you could get some funding from Princeton for it.
What's to stop the plasma flowing around the toroidal axis and flopping and banging around like a tokomak again? Do you know something about stabilising turbulent flows that the legions of tokomak researches have missed?
What's to stop the plasma flowing around the toroidal axis and flopping and banging around like a tokomak again? Do you know something about stabilising turbulent flows that the legions of tokomak researches have missed?
Exactly.
I am talking about a fast-aimed sphere versus a fast-aimed toroid.
I am NOT talking about Polywell versus tokamak.
The 'wiffleball' seems to be regarded as ideal - why? - because it is a continuous magnetic surface. So why not exploit the best magnetic surface possible for the fast-aimed idea, rather than relying on somethin with holes in it? A toroidal volume enclosed by magnetic curfaces does not HAVE to be a tokamak.
I am talking about a fast-aimed sphere versus a fast-aimed toroid.
I am NOT talking about Polywell versus tokamak.
The 'wiffleball' seems to be regarded as ideal - why? - because it is a continuous magnetic surface. So why not exploit the best magnetic surface possible for the fast-aimed idea, rather than relying on somethin with holes in it? A toroidal volume enclosed by magnetic curfaces does not HAVE to be a tokamak.
Why would it? It is a symmetrical system without directional bias (again, quite unlike a tokamak). It is no more nor less likely that the contents of the Polywell spinning around and around than it goes spinning around a toroid!?icarus wrote: What's to stop the plasma flowing around the toroidal axis and flopping and banging around like a tokomak again? Do you know something about stabilising turbulent flows that the legions of tokomak researches have missed?
The surprise to me here, now that I have been disabused of my notion of a central core of electrons as a substitute for a fusor grid, is that to prove the desired ion acceleration mechanism works, this confangled cusped magnetic surface idea is being used in preference to a toroid.
-
Art Carlson
- Posts: 794
- Joined: Tue Jun 24, 2008 7:56 am
- Location: Munich, Germany
But we already know what happens when you put a plasma in a toroidal field without a rotational transform. It accelerates to the outside in short order (just a few times the free-streaming time). This is a bit more than pure mathematics, but it is pretty elementary plasma physics. (And, of course, you are massively unstable to an interchange because the outer field lines have unfavorable curvature everywhere - but your plasma won't even stick around long enough for you to see that, so why worry?)chrismb wrote:It is possible you misunderstand my point. The issue is not 'the tokamak'.
The tokamak is a device which exploits toroidal magnetic fields. It isn't *just* a toroidal field - you have to monkey around with other sutff aswell, like generate a toroidal current.
I'm saying that the Polywell can also exploit a toroidal magnetic field. It wouldn't be a tokamak, there would be no [intentional] toroidal currents, nor barrier transport issues, nor divertors, nor helical safety factors, &c. &c.. it would just be those radial currents in poloidal cross-section. Just like the Polywell in section. This would NOT be a tokamak.
I cannot disprove potential experimental issues, like the ion transport, but the one thing I can PROVE is that an uncusped spherical magnetic surface is impossible. This is a definite, well understood mathematical proof, with no degree of experimental uncertainty whatsoever. So why bother with it. Why not exploit the total coverage of a toroidal magnetic surface? Just to emphasise - this would NOT BE a tokamak, it would be a toroidal 'Polywell' (of course, it would be a complete surface, so it would be a 'Uniwell', perhaps?!)