We are finally doing it, no ifs, ands, or buts!

Imagine a regular array of target nuclei in say, a lithium fluoride crystal. At some magnification you'd see them as the size of golf balls, about a third mile apart -- in a plasma they'd be considerably less dense, or even farther apart for the size of each nucleus, and really moving around; we will use solid matter. I am, for the moment neglecting that anything above absolute zero is moving around a bit. In my workups though, this turns out to the be scaling limit to how many at a time can be made to interact, with confidence, per shot. Once a shot has occurred, the target will be hot, so either it or the beam has to move for the next shot, or a long cool-down period would be incurred.

The current approaches are worse (6 degrees of freedom to split input energy between, random directions, radiation losses) than shooting at this from the sky with a shotgun, as it is very expensive in terms of power used to accelerate all the projectiles that miss a target -- nearly all of them. But even in normal real life, I own several bench-rest guns that can do that, one shot at a time, and that with all sorts of accuracy-robbing problems that protons don't necessarily have (me in the control loop, wind!, etc). So, if one could do this with bullets in the wind, why not hit nuclei with protons when there is effectively no wind, no serious barrel vibration, and so on? The reaction I favor is P -> 7Li gives two alphas with about 16 MeV of energy between them.

And no neutrons. I consider this a feature.

The alphas could be collected either by collision with the tank walls and using the heat, or one might get some of the energy via charge collection, at fairly high voltages. I've got 5 textbooks or more that have various (obviously sloppy) estimates of the actual cross sections vs voltages, the above is an estimate assuming no capture resonance effects, based on simple sizes -- in other words, those golf balls might turn into beach balls at the right energy for an interaction resonance, which in turn would make this easier to do than this back of envelope calculation suggests. Until I find a reference I can trust, the plan is to simply measure the cross sections myself.

In a nice, dense, solid crystal, once you know where a few (3 or more, measuring this would be noisy) targets are, you know where they all are, in relation. So why not focus protons on such a substrate such that only ones that had a chance of interacting directly with a nucleus were accelerated and focused? Think of a bunch of protons in a pixel sort of array, just one per pixel in the limit, with lots of blank space in between. One would select the protons "going my way" via the equivalent of a shadow mask, such as was used in CRTs. This way the rejected protons only have the ionization energy expended on them, about 13ev. Only ones that would hit get past the selection device(s) to get whatever acceleration turns out to be needed, which is what uses the power input (however you do the acceleration).

Let's assume we can do this at around 250 keV proton energy. That's 64:1 gain, and we get our 250keV back too!

Our hit-to-miss ratio doesn't have to be perfect to have system gain. Remember, nearly all energy input also heats the tank you do it in, so you get that back along with whatever you get from fast alpha rays. You of course would lose some in conversion assuming current technologies, but you wouldn't have to have much gain from fusion compared to 64 to break even. Gain of a few would do. I'm just trying to heat and power my house, after all. You'd have to make your own. I will deliberately expose all the details I know myself, so you (and everyone else) can do just that. If it works, of course, but for fun I will also show the failures, as they tend to be instructive as well as sometimes humerous.

I can think of several ways of coming up with protons that all have about the same energy and direction, of which spalling them off another crystal is a possibility. The ion source / mechanical shadow mask seems barely possible for small arrays, but also very difficult to make a size that will fit into a vacuum tank. Perhaps an array of nanotubes aligned in some matrix could be used to get the hole size to spacing ratio to match the target? A lens could then focus protons from whatever source to the right scale during acceleration, with deliberate astigmatism to compensate for the space charge repulsion.

Why has no one looked hard at this before? If you're looking for a rant or conspiracy theory, look elsewhere. DOE is really the department of weaponry, so if it doesn't make neutrons to breed weapons grade fuel, they're not interested, it's not really under their mandate. All their "laser fusion" stuff has been advertised as green, but is really related to nuclear weapons stewardship programs, ways to avoid doing illegal under treaty tests. They are at most guilty of false advertising, trying to get good green PR on what is essentially a weapons program. Maybe they have indeed looked at this, and that's why those cross sections on this reaction don't appear anywhere I can find, so far...

Power companies? As I will fill in more details about later, this cannot scale to gigawatt centralized subscription model power plants for fundamental physics reasons. As I currently see it, the upper limit is around 10kw gross per setup, enough for a home, but not enough for them to make money on, and in their terms, a threat to their business model. OK, a small rant -- when I was selling solar power installations, local power company employees tore down my adverts and defaced them (we caught them at it and talked to them). I have no idea if this was company policy, but clearly they saw it as a threat to be stopped, whether individually or corporately. As far as I know, these attacks had no actual effect. The real problem with solar power is that you have to put your money where your mouth is, upfront, as I have -- no new tech is needed, my systems have been solidly running for 25+ years, with some of the tech stuff being that old, and still working. Sadly, most people would rather talk than do, and get the bling (on credit, of course) before investing in what's really important. Some have even asked me how I live without all that stuff -- the answer is "easily, and less distracted, and no debt on anything at all". Those Joneses will be trying to keep up with me at some point...

I have had some offers of support, from scientists working for other governments (our friends in the chilly upper right of the map), but only after I do all the grunt work to prove it can work -- proof of concept. Time will tell if I accept any of these somewhat hypocritical offers for support. If you want "in" for the payoff, you'll have to take some of the risk too! Only fair. If I had proof of concept working, I could get plenty of money/support from just about anywhere.

It's very tempting to say, (so I will)

"Follow my quest, as I figure out how to do this, and do it".

Any actual help is appreciated. And put away your cynicism and-or wallet, I don't need either one as I have plenty of both in-house... What I could use is some advice from someone who has been there, done that, re accurately putting particles where they are supposed to go. Someone with physics/math creds would be nice too. That could save loads of time and money, and time is the main factor here, as I am getting no younger. The good people on the Fusor Forum have also helped with scrounging (they are the MASTERS of this), and this help is greatly appreciated. I'll put up a credits page soon.

The pictures below describe what I am presently working on. What I am doing now is developing the basic techniques so as to be able to answer questions such as "how would I know if I DID hit a Li nucleus?". So, I'm hitting up every descriptive physics source I can to learn as fast as I can, which so far, is mostly books I am duplicating techniques from, or finding out why I can't, yet. As we used to say at C-Lab, baby steps are sometimes the fastest way.

This is my first really good experimental vacuum system. I got most of this from Pfeiffer and Lesker after giving up on scrounging all the stuff, as that was just taking too long for my goals. I must say the Pfeiffer Mass Spectrometer is a very worthwhile learning tool, without something like it one is flying pretty blind as to what is actually in the tank. Hint: H2O is the enemy, mostly. The Farnsworth crowd are in for a big shock when they get one of these and discover that the desired gas just goes down into the pump, leaving the other non-fusing stuff behind in the tank. Even the Tokomak boys recently discovered how huge an amount of their energy loss was from tiny contaminations, in this case heavy metal ions from the beam stops.

Here is a cylindrical design Farnsworth fusor I did for fun, and to get into the "plasma club" on the forum. In this case, there is no fusion going on, as far as I know, the gas content isn't correct for that. Strictly for show and proof of concept. And oh yeah, fun, a guiding principle around here. Remember that what you see is atomic electron energy transitions at the visible range only (about 1eV), in other words the tail end of the losses, and only the visible ones at that. The real action would be happening at X-ray sorts of energies, which we can't see, in this picture anyway. To paraphrase, "You can't judge a fusor by looking at its poisser". Under some conditions this setup generates a visible beam of ions out of the near end that streams to the hole in the upper right, which has a vacuum gage in it that has a large external H field pointing into the tank. Gas is HeNeAr in some mix I thought pretty at the time, at about a tenth millibar total, and we are running several kV at around 20-50ma here. I used NiCr wire at first, but this sputtered badly onto everything else. I now use Ti, which at least acts as a getter when it is sputtered. The other stuff is for other experiments, I try to do several for each system teardown. Sticking down from the top is what's left of an old metal vacuum tube after cutting off the top and making a hole in the plate. Yes, this works and I am getting several mA emission from it after rejuvenating the cathode in vacuum, or about the original specs for this tube. At least a couple of books from Phillips and from Terman say this cannot be done. Wrong, here it is, and it's pretty easy, almost the same as the original activation process. More when I know more. I've even restored a coated filament from a 5u4-class tube after it had been in atmosphere for 6 months. Evidently, at least in the tubes I've tried, the coating is far thicker than need be, and creating a new surface can be done a few times with little loss.

Here is an e-beam setup I played with, using a 6ak5 type tube this time. I got copious X-Rays from this beam hitting a copper target at the back of the tank. I am using a CCFL inverter from Digi-Key followed by a Cockroft-Walton multiplier to get up to about 35kv on the target, with only about 6 watts of power available. This is a good thing for safety! My survey meter goes nuts when this is powered up. But these low voltage X-rays are pretty easy to stop, even one more layer of glass over the window does it. What this picture doesn't show is the modification to the tube, other than the missing glass. I made a hole in the tube plate for the electrons to come out of on the side towards the target, and played with various grids voltages, around the normal tube specs, with success. This is done in a pretty good vacuum, on the order of 10E-7 mbar or so. You can see an old CRT I intend to swipe the electron gun out of in the background. I will just cut the tube, saving me the effort of making a nice one from scratch. I may even keep the deflection plates for fun and games later.

I had been planning to do it all in the above system, but a friend from the Forum (that's you, BillF) found this tank at a metal scrap yard for the price of the stainless steel by weight. Far nicer! Most of the extremely valuble stuff had been stripped off, but this is a great start to a really nice pro system, and this saved a few tens of thousands of bucks getting there. I've since bought a new turbo and dry backing pump from Pfeiffer for it, along with a nice door from Lesker, and various other stuff. This was one heck of a complex dish to wash to vacuum specs! Sadly, I lunched my Mass Spec head transferring it here, but it will be sent back to the factory and repaired, now that I know what ills lurk in the smaller system. The smaller system will now be relegated to messy stuff, like evaporation and sputtering, and my vacuum tube hobby, as well as small scale testing. But for now I only afford one fast access door (they cost over a grand), so whatever system it is on gets the most action. The big tank is now back on its cart indoors, and doing real work. The system base pressure is a couple times 10E-8 mbar (20E-11 atm) at this time, with no real bakeout yet, thank heavens the huge base gasket does not leak! All the gaskets, flanges, feedthroughs, and windows cost quite a bit, but it's all there now, whew. I am using a 512 l/sec turbo and a dry piston pump for this, and that system from Pfeiffer takes no prisoners as later pictures of experiments will show. This system attains pretty good vacuum even with a small deliberate gas leak, and I have to throttle down the turbo to keep the pressure up for some things, which the nice controller allows me to do. Saves the price of a big UHV valve, no small thing. Also saves power, important when one is running on solar panels, as we are. Although the turbo draws lots during spin up, it only takes a few watts to keep it spinning in a decent vacuum, as it should be. Nice design, Pfeiffer!

Here's a picture of the inside of the big tank, looking through the nice door. You can see an extra piece of borosilicate glass I've added that is held in by home made clips. If this gets trashed, a replacement is cheap at McMaster-Carr (link to vendors here), and for minor stuff, it can be dunked in the acid tank for cleaning, something I would not like to expose the door to, not to mention that it would need a new copper ga$ket each time. I got the evaporation filament from Lesker, they are relatively inexpensive, and I am running it on a large surplus transformer/variac combination. You can see some type C thermocouples I made from wire I got at NANMAC to monitor things. That feedthrough really hurt the wallet, but is important. A little further on, this and other things will be interfaced to the PIC based data logger I designed for this (put link here). The idea is that it will be opsys independent on the PC side (I run linux, mainly), and the logger board doesn't cost much to replace if some EMP nails it. (link to pic project here) In real use, the filament area is covered with a shield I made from a cut up coffee can, to prevent deposits on everything else in the tank -- I nearly ruined some things before I found out how important that is. The tank lighting here is from two "stab in" heaters, really big quartz/halogen lamps, that came with the tank. I only used two of the four I have, as it seems the reason someone surplussed this tank is that they badly overheated it and caused all manner of hard-to-find leaks with the original 4 heaters. All the original copper gaskets were oxidized black, and even the tank itself was "blued" from heat. Their loss, my gain. The ring stand was donated by BillF, and I then copper plated it to reduce outgassing and rust issues. I've since made some target holders for it, and made some pretty nice first surface mirrors on very thin microscope cover glass with this setup. These mirrors are light enough to be levitated with the SciToys pyrolitic graphite and magnet kit for some interesting applications.

Here is a cyclotron prototype I am working on now. I am trying to use NMR to measure the field I'm getting inside the pole pieces. For that, it seems I will need a better signal generator than I currently have, as I'm not sure within a factor of two what I am getting and my current generator has lots of harmonic output so I cannot tell from that signal. This will be used to measure cross sections, using the Dee walls as the target. This eliminates the requirement to make very high voltages and get them into the tank. The magnets are grade 42 NFeB, which should make on the order of 12,000 gauss in that gap, more or less. It would be nice to know the value so I can adjust the design to get what I want there (more is better, up to the point of saturating the iron pole pieces). In use, this will have water cooling, as the Curie temperature of these magnets is quite low, and they're not cheap enough to not care about wrecking them. The current pole pieces are 1.5" diameter and the gap is about .25" high; more H field is better for more energy at the Dee wall limit. This is all constrained by what will fit through the tank door, or I would not have bothered to round off the corners on the magnet structure. I have not made the Dees as yet, but they will be made of pure Ti metal. This way, I can use it with deuterium for DD reactions, as D sticks to Ti well, or flash them over with Li for measuring the P->Li reaction. I have not yet decided to apply the cooling to the Dees or the iron pole piece extensions, either is a hassle, just a different hassle, but will be required. One thing you learn fast in vacuum work is how poorly things lose heat in a vacuum unless they are hot enough to radiate it away well, or it is conducted away efficiently by something. In this case, that'd be well after the magnets are toast. I consider this a real innovation in cyclotron design. Rather than the 100's of tons of water-cooled external electromagnets (which had to be external due to sheer size), this fits entirely into a small tank. Because we don't need tank walls inside the pole pieces, they can be closer together, which in turn reduces the size of the required magnets. This might be scaled far smaller, in fact, but this size is convenient for me to work with as is -- my machining errors won't matter as much at this size as they would were it much smaller. Current uses of cyclotrons are in hospitals for activating radioactive tracers with short half-lives, and the costs for those is simply crazy. This looks a lot cheaper to me, at least sans the probably-required tort insurance, so maybe it has commercial applications besides what I'm using it for. We'll see how much energy we can get, it will be determined by the field strength, but even with a pessimistic estimate, it will be enough for measuring the cross sections I am interested in.

No one gets this far this fast alone. I've had much help from good people, including those on the Fusor Forum, and my private band of local (more or less) good guys, who are:

Dale Jorgensen -- programming and general engineering here. The best.
Frank Gentges -- wisdom, equipment -- he provided the Ham transmitter that will drive the cyclotron.
Joe Sousa -- fantasic sideways thinking engineer. The proverbial other set of eyes.
Bill Fain -- finds things I can't. Money is no good if you can't find the stuff you need!
Richard Hull (forum) -- for putting up with my weird ideas and encouragement.
THANK YOU!!!!!!!!

If you have bandwidth to spare, can play .mov files and want to see a movie of an analog gas bargraph I was playing with try this. If someone can tell me how to edit these files (from a Kodak camera that used to belong to an Apple person) on my Ubuntu setup, I'd love to know -- many times they are just too big for putting up on the web. This one got up there by accident, so I said "what the heck". There appear to be no settings in the camera to change compression/size or anything related.

Stay tuned, I honestly do plan to update this frequently. For one thing, there's no picture of the main tank in action yet! And it's truly a thing of beauty to a certain type, but also very hard to photgraph well due to having stuff on all sides and underneath. Soon, I promise!

Useful documents

Back to Projects

Coulter's Smithing Home

Here's how to contact us:

Contact information