Fun on the back of an envelope...

Yours and mine. This is where you can gas on about how you think the universe works. To a point, after that we'll expect you to actually test your stuff and report.

Re: Fun on the back of an envelope...

Postby Doug Coulter » Wed Mar 23, 2011 11:59 am

I think I'm confusing people by talking about a couple of different issues together. So let me try and sort it out (please be patient, I already deleted my first try at this).

For one issue, lets assume fusion is purely classical (yeah, I know, but bear with me a minute). If I "hit" one D with another, close enough to "head on" then I get fusion -- every time. In this case my factors are "how hard I hit" and "how close to dead on". Normally, in physics, this is called cross section kind of stuff. But there are some really big assumptions made in that, which I intend to debunk (and it won't even be hard -- in theory). So let's assume the simplest possible situation for the moment -- firing D's one at a time at some target. Since the atoms in a target are much larger than the target zones for "dead on" -- hugely larger, one looks at the probability of actually hitting one by pure chance via a few things. One is simply how close together (dense) the "target zones" are at the surface, or near the surface, assuming that something that "misses" in the first few atomic layers has lost enough energy to the electrons in the target it no longer satisfies the "hard enough" hit criteria. You can only get a solid so dense, and with "random" aim of the incoming D's, you therefore create an interaction rate determined by target density and how fast they are going (if they're going fast enough, they can penetrate more deeply with enough remaining energy to fuse). The "effective" size of the targets is often related as the "cross section" and can be larger than the physical size of the nucleus under favorable conditions, but in the case of DD, it's not a huge effect (nothing like some resonances for neutron capture, for example).

The only way probability enters into this is via the assumption that I can't aim D nuclei well enough, or know where the targets are well enough, to just hit one. This is certainly true of all of the attempts I know of at this point. But that also seems to be because no one has tried -- I smell low hanging fruit. If we make the assumption that I CAN know where the targets are, and that I can build an accurate enough "gun" -- then that bet is completely off. Well, in real life, it seems I can even make a regular rifle with the required accuracy, and in fact have done so, even though there are issues there that make it harder to do than with D nuclei -- wind, barrel vibrations, non identical bullets and gunpowder, varying case capacities -- all these have to be overcome for the rifle, but not so much for the D gun, yet I did it anyway. The limits for the D gun seem to be wavefunction (which turns out not to be a problem so I'm told by "real" physicists I asked), thermal agitation of the gun (lenses) and the projectile, and various sources of vibration between the gun and the target, vibrations (heat) in the target. All these seem pretty susceptible to being handled fine to the level of accuracy required to vastly increase the interaction/fusion rate.

So this one is generally modeled as a probability, but that's only because of the assumption of a bad "gun" and a target you can't identify, so you're reduced to a "shotgun shooting at widely spaced golf balls" kind of situation, the way it's been done so far. But I don't see we need to keep doing that if we get smarter or more ambitious -- and toss out that assumption.
(gotta watch for things that have built in and unstated assumptions in general)

So, that's one whole topic, or set of them.

Now, onto the next.

Given that I have two D's in such and such proximity, for so and so length of time, what's the likelihood they fuse, vs just scatter? How sensitive is this to either parameter, distance and time for starters? This can be calculated I believe (and probably has), and it seems to be a linear function of time, but a very steep (and nonlinear) function of distance. This mostly speaks to the above, as if we have "good aim" we maximize both proximity and time of proximity.

But I am also saying we might be able to affect that parameter in our favor with various forms of orientation or polarization. I do not have a clue at the moment which ones or how much, or even how I'd go about achieving a particular set of parameters for interaction even if I knew which ones mattered. Just the idea that it sort of has to be logically possible. And some actual data that seems to show that it is happening -- and affecting which reaction pathway happens to be taken more frequently at the moment -- meaning the book ratios of the fusion pathways aren't end-all, be-all numbers, but simply what happens when everything is random (thermal) and uncontrolled. So, random in, probability out (GIGO). But what if -- NOT random in?
I think that's a path worth going down, is all I'm trying to say, as it makes the above situation potentially easier too; with the right "polarizations" whatever we wind up meaning by that, our aim might not have to be as good to still get a good rate. In essence, the existing cross sections have a built in assumption of "random" in this, that I intend to investigate the "not random" side of.
That would make current data not wrong, but simply incomplete under the removal of the "random input" assumption.

Further, if we can actually control which pathway is more frequent, the more interesting ones might become more important. Would it not be better if we could do the one that gives ~16 MeV vs ~4MeV? Or, between the two more common ones, maybe I want neutrons today (to activate or breed things) but not tomorrow, so I want one or the other pathway, mainly.

Now it's a pretty strong claim to say I'm able to do this -- all I have is some data that seems to say it happened in my lab (by accident!), and it seems possible to make it happen repeatably in the sense that I can see it drift from one pathway (mainly) to another over some fairly long time-scales. This happened during runs where everything was much more controlled and stable than in other runs in the past, which kind of makes sense -- this has to be soooo critical, that with the best stability I can muster, which is pretty good, it still drifts in and out. This makes sense when you look at typical vibration/rotation frequencies -- 10^22 type numbers, while flight times are in the 10^-6 type numbers. That's a lot of spins on a tomahawk between the thrower and the target to say the least, so under very slight drifts, for awhile they all hit point first, for awhile they all hit some other way....


So we have a couple of topics under consideration here. Do we accept that all the numbers that are surely true for the "random shotgun at golf balls" case are also correct for the "aimed fire" case? I think we don't. And to add one to the first case above, instead of firing D's at some amorphous target, we could either use a crystal and take advantage of its order to know where the targets are (hard, but possible) or simply fire single pairs of D's at one another with enough accuracy to take the "random" factor of flood beams out of the picture.

Assuming we do that (and that it's possible, it seems like it might be), then we can also reduce the accuracy requirement by having the D's get into proximity with the correct "orientation" to make the aim less critical -- they "want" to fuse more in some relative orientation (this is where the real quantum mechanics come in).

At any rate, for our "microwatt club" (should we start one like that), the very original point of this thread was to point out that should we get anywhere near gain via increase possibility of "first try" fusion (which avoids all the issues of recovering from thermalization -- because it all fused in the first place), we won't need squat ion current, in fact, far less than we've been using and talking about for the "shotgun" approach. 10 pA instead of 10 mA....or thereabouts if we are "at breakeven" interaction rates on first passes.

This is a different approach than things like Chris' idea, where he is able (we hope, actually) to re-gather and un-thermalize the ions that merely scattered with little enough energy input to make it worth it for them to "try again", even accepting a very low "first try" rate of fusion. Similar issues apply to the "recirculation" most hope for in a fusor that I just can't seem to measure actually happening here, though Chris' approach seems a heck of a lot better in the main. And of course all of these seem better than merely trying to squish and contain a thermal plasma with forces that only act normally to the direction of motion, and at that, affect electrons and ions differently, with electrons present to waste energy (tokomak).

It's just a plain different way of approaching the problems we know we have. One way tries to undo thermalization caused by crappy aim, this way tries to avoid it in the first place by increasing the first-try fusion rate. We can't yet know which (if any) will pay off till more lab work is done, there are possible gotcha's lurking in either way of looking at things.

At any rate, this is leading me to want to try something with low current, no electrons present (ion trap conditions) with simpler geometry (at first) to explore all this and see what is to be found there. Most ion traps have the ions avoiding collisions, but it seems quite possible to make something like a colliding beam/storage ring kind of thing with appropriate optics to get that aim in the range required to very much up the effective interaction rate -- actually aiming things at one another, rather than two shotguns firing at one another from a distance and hoping some pellets hit -- that case is what leads to thermalization and losses (and the requirement to model it with probability instead of deterministically).

I'm not done with fusors altogether -- there are a few more things I'd like to try there with an aux grid as an ion trap and pre-buncher/focuser, as well as some things to control electrons with a magnetic field. But having taken that a lot farther than I'd hoped (roughly 500x better Q than most), but not as far as I want (and the big fat improvements aren't coming easy anymore) it looks like that was just a step along the way, not the destination. So, a few more things to try with the basic fusor hardware, then I move on to hopefully better things. No way I (and we) won't learn important things doing that, so it's worth the effort. Who knows, what we learn may "back-port" to more conventional designs? It wouldn't be the first time that new knowledge helped in areas that weren't suspected before.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
Posts: 3038
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA


Return to Theories and speculations

Who is online

Users browsing this forum: No registered users and 3 guests