Current Working Theory

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.

Current Working Theory

Postby csnyder » Mon Feb 12, 2018 12:52 pm

Doug,

From the Fusor Forum:

Re: My Visit w Doug Coulter
Post by Doug Coulter » Sat Feb 10, 2018 5:02 pm

That post is a lucid description of your current "theory and speculation". IMO, it would be worthwhile to have as a "sticky" post or have its own sub-forum that only you could post to and edit. You could update it as required if and when the idea changes or becomes more refined. It could sort of become the "credo" post.

I am sure the points you make in that post are expressed in these forums somewhere but not as succinctly as that one post.

FWIW
Charles
User avatar
csnyder
 
Posts: 14
Joined: Thu Feb 03, 2011 2:24 pm
Location: Lexington, NC

Re: Current Working Theory

Postby Doug Coulter » Mon Feb 12, 2018 11:37 pm

Well, here's a copy of that post, FWIW. It's actually a poor paraphrase of the patent disclosure paper I wrote...which is on the big side for here (!) and by now, slightly inaccurate due to further learning.
At least this is a backup of the fusor.net one:
Pretty much what I read above is correct, FWIW. As John and Bob say, there's more on my site. Which is why you don't see me here that much - I'm busy with my own. Thanks to Bill for the head's up here. He's been kind of a silent (to the world) partner and big contributor to the effort.

What I'm trying to do, though it uses AC, is substantially different than Farnsworth's plan and anything having to do with multi-paction at all. (The only things I want to hit are deuterons - each other -, not any part of the tank whatever, and electron multipaction in the enemy here too, though I think one-shot secondaries are important, it's complicated).

The basic idea is that of an RF/DC ion trap, based on similar math to what a quadrupole mass spectrometer uses. Now, over time, these let the trapped ions find orbits that actually avoid one another (lowest energy state for the system) - which isn't what I want, but it's a start, and it's forced recirculation, not "hoped for" - I have a youtube video up showing detection of this. https://www.youtube.com/watch?v=fQYqOsk5QRo&t=2s
This shows some bunches from previous cycles piling up - just not very well - and I'm getting a bit of fusion here even at ~~ 10kv - way out of the noise. For reference, to get that far out of the noise on the same gear here with pure DC - it takes ~20kv (and Q goes up MUCH faster than linear above that). I was working at much lower voltages and power because I didn't have higher, and because I haven't made this remote-controlled yet.

So this is officially "interesting" and "hmm, that's funny" turf.

Part of my scheme includes also controlling the electrons. As Richard has noted in the past, having them all whiz out to the tank walls at high energy is one of the main wastes in a fusor (and a source of obnoxious X rays). The same math applies, but the speeds and feeds are around 60x different, the square root of the charge to mass ratio ratio of the two types of things involved, (yes, the ratio of the ratios) Getting rid of this loss is key - and Bob's reference to a parasitic is what I believe was going on there and doing that for me. The electrons were being yanked back toward the middle when the fast oscillation drove the grid positive...I think.

Since tightening things up and creating the remote, the parasitic stuff went bye bye (I got rid of a huge inductive ground loop...), and everything is much more stable now. It works well as a normal fusor - around 3-5 million neutrons/sec at 50kv and 6 ma would be normal, but of course, that's not all that exciting - it's just what I measure in the old school mode. The high end of the Q range is reached using some external ionizer and a lower gas pressure, but otherwise nothing very special - it's just a good precise build.

Using the fusor as a power triode itself is slick, and might be "the way" once the real numbers are well known and it can be coerced to do what we want pretty precisely, but since it's not handing itself to me on a silver platter other than those couple of times - that kinda destroyed my data acq - I've reverted to just driving the thing with known waveforms (making arbitrary waveforms at 50kv levels ain't a picnic - this is part of what slowed me down). I can't just use the standard math, as we are in a rather strange set of conditions. The ion trap math basically assumes super low charge density - hardly any particles - and space charge when there is lot of charge flying around perturbs the math-assumed conditions one heck of a lot.

Here's a hint of the "shape" of the standard math stuff - we have a different mechanical shape, of course, but that's not all - we traverse from molecular flow to viscous flow depending on where in a cycle we are! https://en.wikipedia.org/wiki/Quadrupole_ion_trap This math doesn't take that into account at all...but where it's right, it's right. It's also why polluting the fuel or going lower in voltage won't have the desired effect to lower output for testing - an ion trap or anything dynamic similar needs to change frequencies in concert with voltage, and the mass of the entities is one of the tuning parameters.

So, we search, the hard way again. Just getting to be able to do that, and the roof falls in on me - literally. It'll be fixed soon, it's just a pain. Those who won't be stopped...can't be.

The basic underlying idea here is to drive recirculation.
I have to oversimplify this at first - those who know, give me a little slack for a few words.

1. Assume we have a more or less uniform cloud of D+ ions in our tank. Assume the pressure is such that we're in molecular flow status.

2. We take the grid negative, starting them in towards the grid center.

3. To make them arrive all at the same time (bunched), we don't use a sine wave on the grid - we use an exponential increase to bunch them. We are trying to get more or less the same kind of bunching one gets in a klystron or for that matter a drift tube accelerator, but in one half-cycle.

4. As they come in, we get denser and denser, and we leave the molecular flow regime, and they start significantly repelling one another as well.
(This means we can't just use some sort of simple waveform, as we need to compensate the D+ charge effects on the net field which is no longer the same as what we think we are applying - it's the sum of our drive and the charge flying around in there). To help keep them bunched, we have to increase the negative volts on the grid even more - we want the tail-end charlies to be going faster than the ones that started closer in so they all get to the center at once.

5. As they get near the grid edge the Coulomb repulsion starts to be a real major problem - slows things down, spreads the beams we've created with our grid's electrostatic lens. To help overcome this, we now want electrons in there, stat. Luckily, they are much easier to move around, being lighter for the same charge. And they're following the ions fairly closely anyway, since the ions, particularly as they get near the grid, cancel our applied field, so the electrons don't see the negative grid field, but the one made by the ion concentration.

6. So, making a virtue of necessity - we flip the polarity of the grid right then - the ions are at the edge or inside already - they don't care, their trajectory is set, they don't have far to go, and won't be perturbed much. The flipping of the polarity perhaps drives them in a littler harder, but that's not the main thing - what it does is pull in tons of electrons, which help neutralize the Coulomb repulsion. The electrons move around 60x as fast (see above) so this can be done...

7. Now, some fuse and some don't, and everything flies back out towards the tank walls. A proper tuned circuit (this is for later on..don't need it yet) can absorb the kinetic energy as things leave and put it back in the power supply, again, somewhat klystron-like. This phenomenon is also mentioned by some of the accelerator labs when they slow down particles on purpose - you have to put the energy somewhere (charge induction). It's also how my faraday probes picked up charge bunches in my tests. The idea here would be to recover the energy invested in the ones that didn't fuse, and let them slow down and re-thermalize near the tank walls instead of wasting the input energy bashing them (and being neutralized - Richard was correct that most ions formed by the grid itself are already almost at the bottom of the potential well).
At low energy, this thermalization - randomization - helps us have a uniform starting place for the next cycle. We already got our money back, so to speak, this is just loose change energy helping us have a nice even starting cloud again.

8 Go to #1 and repeat. We always had a spring-mass system. Now we're driving it instead of hoping it will magically oscillate (recirculate) on its own.

There are cases in the literature of diodes (what a fusor kind of is) oscillating with tuned circuits because of transit time effects, but that's rare, flaky, and not high power. Most tech realizes you need an active element with power gain in there someplace...

Almost like an IC engine or Otto cycle -

(To the objections I can hear in my head from people who know this better - trust me, the electrons that get jammed into the tank walls early on - and are lost at LOW energy, are later released again at or near the grid as the D+ bunch arrives and a few hit and cause secondary emission and are therefore available as described above, again... Details like that...Also details like where/when we go from effectively molecular to viscous flow - which can be on part controlled by the basic pressure and in part by how well we get bunching to work - there's a lot of little demons in details I'm leaving out - I know)

This is what I think happened here, more or less by accident. A two-frequency oscillation with added DC is a pretty close approximation to what the ion trap theory math would predict would work for a 2 species system (D+ and e-). It's my best guess, at any rate.

An interesting tidbit is that at relative velocities in the range that electrons reach at ~ 50kv, the electron De Broglie wavelength becomes commensurate in size to the Schrodinger wavelength of a deuteron(otherwise stated, the distance they have to be within for a high probability of fusion to happen)...One might hope that at this point, the electrons are sufficiently "small" to actually get between deuterons on a small enough scale that they can still "see" each other's leaking strong force and tunnel into fusion, which otherwise would be (and we note, in standard fusors IS - quite rare).

So, nothing like multipaction, though I think secondary emission does have a part in this. Sorry Richard (and Philo). You were right about a metric ton of other stuff, though. One thing not otherwise mentioned here is Richard's intuition that most of the ions in a self-ionizing fusor are created near the grid and near the bottom of the potential well - I tested and verified this myself.

Really, the main advantage I or anyone has over Philo - we have much better data acq stuff. When I replicated some of what he claimed, that others couldn't replicate, and thus thought him nuts - at least I had proof I - and others - could believe in, and a ballpark reading of stuff he had no way of seeing at all - the oscillation frequencies (which, on examination of his gear, would be just about where I'm poking around now, looking at his lash up with parasitic L and C in mind). I note no pictures of oscilloscopes in Philo's work - or video recorders pointed at them. It might require fewer brains with help like that(!).

I have to say that something else Farnsworth mentioned - later also mentioned by U of Wis - was the use of BN in feedthroughs. This was actually the major enabling piece of technology that got rid of all the noise that was contaminating our attempts at precise measurements. Those who've followed my work for awhile know we thrashed hard on HV feedthroughs and did a lot of revising. All these had short lifetimes and always had a bunch of micro-arc noises they made as well - that EMI fusors are famous for largely comes from that. The "dog that didn't bark" is that once we went to the current design - that all stopped, and rev 1.000 of that feed through is still in there, rock solid, and noise free - no more RFI crap from a disintegrating thing. Seems super hot hydrogen will take apart anything but BN or pure sapphire. At current prices (and machinability), I'll take the BN - The design of the FT deserves it's own post - it's not just a chunk of BN - pyrex is also used *as a conductor* for field shaping but...in terms of apparatus, that was actually the biggest thing, after all the stuff everyone here already has down. I guess I should document that further, it's a pretty big deal.

And as luck would have it - those micro-arc noises were also what provided the signal I had to have to get some important data on the rest...but once I had that, were a serious problem... That stimulus signal - very fast rise time pulses at 10's of kV amplitude, would have been (And later, was) a terribly difficult thing to create any other way, to see the impulse response of the charged particle system.

Further, those micro arc noises, oscillation, and other similar things were what made possible the observation that anything that perturbs the fusor off that nice stable equilibrium mode everyone seems to shoot for - increase the Q! However, things that happen "by accident" are rarely as good as things you can do on purpose, once you know what to do. The accidents just give you the "heads up".

If this pans out, I'll also have to give Richard some serious credit for motivation. Nothing gets me going like some authority telling me I'll never pull it off. Thanks!

Funny old world, eh?
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: 3515
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 1 guest