Magnetron Ion Source (Olson Fusor)

[tolson1118]

I finally got all the parts and put a Magnetron Ion Source together. Much like Jon Rosenstiel's, it's based off Takamatsu’s ion source minus the complexities of water cooling.

The NdFeB magnets I'm using are 0.375” X 0.375" Cylindrical magnets. Grade N38SH for higher temperatures.

Cathode is 1" diameter aluminum rod bored out to accept 3 magnets stacked, the top was tapped and sealed.

The BeO feedthrough, was machined and silver soldered on to a 2-3/4" CF flange from Lesker. A bit of an overkill, but it works nicely. The D2 feed goes thru a hole bored and tapped thru the CF flange.

I'm running the source at around -1100v @ around 5mA

Seems to be increasing neutron counts by roughly 10%
Switching on the ion source causes a very small increase in fusor power draw, however the grid does seem to notice (increase in temp).
With my newly completed texlium He3 Detector, I noticed it was still producing a neutron count (lower of course) at 5 mTorr with no plasma in sight.
My running Pressure is down to around 12 mTorr now.


Pic1: Ion source disassembled
Pic 2: Ion Souce assembled
Pic 3: Ion source on Fusor
Pic 4: Inside pic at low voltage and higher pressures, ion source ring, much like in Takamatsu’s paper.

Oh, the ion source is not protruding so much into the chamber anymore, like the picture shows,

-Thiago Olson

[tligon]

Gorgeous rig! I've seen pics of it before, but I always enjoy seeing the parts.

I was really suprised that mine seemed to have a nearly constant neutron output when I held constant voltage and the pressure varied over a range indicated by a TC gage of maybe 5 to 20 millitor, with the current going up and down with pressure. At the bottom end, sometimes you could only tell it was still lit in a dark room with dark-adapted eyes. I didn't have a camera on mine, and human eyes are exquisitely adapted to blue light.

[Carl Willis]

Hi Thiago,

Very pretty photo of the entire discharge. You can see the magnetron glow very distinctly.

Watch out that you don't dissipate too much power in the magnetron cathode. NdFeB will reach its Curie point at temperatures around 80 C, and the -H grade doesn't buy you much head room over this. Perhaps for these uncooled sources (yours and Jon's) a SmCo magnet would be preferable (pricier, but likely to survive the heat).

Looking at your discharge, I think you SHOULD get better results if you can drop the pressure (the ion source, of course, will help you do this). The glow looks more like 50-100 microns to my eye, regardless of what your gauge is reading. I think your system can probably go lower with success.

Good work.

-Carl

[tolson1118]

thanks,

Good eye Carl, for the ease of a picture the photo was taken around 80mTorr at low voltage.
I like the Idea of SmCo, it should be much better temperature wise, and has only about a 3/5 field strength reduction. Which should be worth it, as it can withstand much higher temperatures. I'll look into it, I can just increase the magnet diameter a bit to compensate for the field loss.
Need to test it again to see how much the source lowered my operating pressure.

Tom, your results make sense. The lower your vacuum level, the more efficient your fusor will be, as the ions will have less gas to impede their acceleration. At higher pressures one compensates for the efficiency losses by increasing the wattage(current) going into the fusor. This mostly happens on its own, as current automatically tends to rise as pressure rises.
So running at the low pressure requires less wattage the keep the same neutron count, it is then possible to increase the wattage (and neutrons) without burning up your grid. I'm hoping that one or more ion sources will be able to lower the operating pressure enough to allow higher accelerating voltages without the grid over heating.


-Thiago

[Jon Rosenstiel]

Very nice, Thiago. Your picture of the ionization "halo" is awesome.

I'm curious about your power supplies. (Fusor and ion source) Are they filtered or unfiltered? Lots of ripple, or not?

Jon Rosenstiel

[Richard Hull]

You are slowly working your fusor to a plus ultra system. I would think an old alnico cylindical magnet would work OK. I have obtained a few from old varian TWT's and some HP noise sources at hamfests. Alnico is very pricey.

Richard Hull

[tolson1118]

Jon, My fusor is running off a rectified x-ray transformer, max ~ 65kv, 200ma, using a saturable core inductor for current regulation, not sure what the ripple would be.

The ion source is currently being run off a 10kv, 6ma Spellman with low ripple, I think .001%RMS. I am eventually going to have it running off a much smaller supply, with possibly more ripple, so I can use the spellman for something else. I have a small 2kv, 5ma dc driven unit that I am going to try to replace it with.

How's your new spellman going?


After demagnetizing the NdFeB magnets in my source on a long run (halo diminishes), I decided to buy some new magnets.

I ended up finding a somewhat OK deal on Smco magnets, so I'm going to give them a try. The Alnico's are really good for temperature, but I ended up going with the Smco's because of their strength, and their curie point is past the melting point of my aluminum ion source anyway.
With 2 magnets stacked the cylinder will be .475x1" with around a 65% feild reduction as compared to the same size in NdFeB.
hopfully it will work out.

-Thiago

[Jon Rosenstiel]

Thiago,

Oh-yeah, I forgot that you had acquired Carl’s original supply. Sounds pretty much like my x-ray supply. (Cept’ for the saturable core inductor). If you haven’t added any filter caps then you’ve no doubt got a fair amount of ripple.

I was asking about ripple because when I power up my ion source, (at ~13 mA), my fusor’s power draw (and ripple) increase but my instrumentation is telling me nothing has changed. See rather lengthy post here:
viewtopic.php?f=12&t=4990#p32298
This makes it almost impossible for me to accurately access what effect the ion source has on neutron output.

I killed the Spellman…see my post in the High Voltage forum.

Let us know how the SmCo magnets work out.

Jon Rosenstiel

[DaveC]

Thiago - Very nice work. Thank you for sharing your results and those great pictures.

Been mulling over Jon's observation that despite the ion gun operating, the overall fusor current - voltage characteristics are not particularly different.

Assuming the problem is NOT with instrumentation, then this suggests that for some gas density regimes, an ion gun supply may not have an advantage over the standard glow discharge approach.... which is to be expected, I think.

An ion gun has its maximum advantage at low gas densities, where the mean free path of the electron would exceed fusor shell to grid, or electron gun dimensions. In these cases, the strong magnetic field bends the electrons into spiral ( or even closed circular) paths greatly increasing their path lengths and giving more collision opportunities At low gas densities, the magnetic field can give large increases in ion current. As gas densities increase, this advantage fades away.

Also, since the magnetice field scales with the square root of ion gun voltage, the actual field strength is somewhat a matter of choice. Higher gun voltages, require larger magnetic fields.

There's a tradeoff between the ion gun voltage, magnet field strength and pressure that minimizes the ion gun power (and hence the heat).

At larger gas densities, I would expect the ion gun advantages to diminish as far as gross fusor currents are concerned. But there are other advantages which I will mention below.

As pressures increase, Townsend (or avalanche) processes will dominate and the electrons supplied by the inner grid will be capable of creating more gas ions than our power supplies can feed. This, incidentally is why you need the ballast inductance in the HV PS. As ion currents increase expontentially, the magnetic field in the inductor stores more and more of the input energy, until the ion current rush is damped, then as the field collapses, the inductor provides a voltage kick (-L(di/dt) ) which relights the plasma and things oscillate around a more or less stable value. Without the ballasting inductance, the devices would probably self destruct, since plasma resistance is inversley proportional to temperature, there is no real limit to where this can go, with unlimited input power. (Maybe to the center of the earth..???)

One major advantage the ion gun brings to the fusor, is the ability to provide most of the ions at the same energy. Thus the ion current from an ion gun may be identical to that of an un-gunned fusor. But, all of the ions from the ion gun are reaching the central area with full fusor voltage,( i.e.: mono-energetic, sort of) , whereas in the std fusor, they arrive with a wide range of energies, only some of which are capable of providing fusion.

A second benefit, will be seen when the ion gun is properly positioned in the fusor shell. High energy ions are easier to aim, and will have a greater percentage missing the inner grid (at least on the first pass through). This lowers grid dissipation, and life becomes somewhat "better".

Dave Cooper