RF for new ion source

For the design and construction details of ion guns, necessary for more advanced designs and lower vacuums.
User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 11:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: RF for new ion source

Post by Carl Willis » Mon Sep 08, 2008 7:49 am

I have a "real" discharge coupler put together now, and another eye-candy discharge pic. Dustin asked about the coupling, and previously it just consisted of a piece of loosely coiled hookup wire. Not particularly effective. Now it is a well-constructed piece of hardware, some kind of pi network or L network depending on how you choose to analyze it.

The coupling inductor can be a single turn or hairpin of copper sheet as shown, about 5" long; or may alternately be 1-3 turns of wire at ~1" diameter or less. The tuning capacitor typically develops very high voltage and carries high current--it needs to be made well and have some space between the plates. The input loading capacitor is not as critical. A choke is shown around the braid of the cable leading up to the coupler. I found that the coupler could be tuned to positions where a lot of sauce got coupled onto the outside of the coax, resulting in very erratic, irrepeatable performance. The choke is a low-frequency ferrite transformer core in my case. It essentially kills the problem.

In the photo is an inductively-coupled, electrodeless air discharge, probably about 0.1 torr. The inductive hairpin can be seen as well as its clear relationship to the bright inductive plasma. The SWR was 2:1 and the forward power about 60 watts.

On another note, my ion sources are almost done. Some details on those coming shortly.

-Carl
Attachments
schemat_Ltuner.jpg
schemat_Ltuner.jpg (36.09 KiB) Viewed 2455 times
icp.JPG
icp.JPG (249.93 KiB) Viewed 2455 times
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277

Dustinit
Posts: 257
Joined: Wed Mar 19, 2008 12:02 am
Real name:

Re: RF for new ion source

Post by Dustinit » Mon Sep 08, 2008 8:27 am

That looks cool.
Can you wrap the inductor around the glass
and still match to it?
Dustin.

User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 11:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: RF for new ion source (NEW DRIVER AMP)

Post by Carl Willis » Tue Feb 17, 2009 6:06 am

My most recent post about this ion source itself was here:

viewtopic.php?f=6&t=2910#p12420

I just added a new amplifier stage to drive the AM-6155. I had been driving it with an old Mini-Circuits ZHL-2, which only managed about a watt output, max. This got me a maximum output of ~75 W from the AM-6155, which was running in Class-A, linear mode in order to enable that kind of output. The new stage is a broadband, linear amplifier good for about 10 watts out. It uses a Polyfet SQ202, a push-pull MOSFET amp, in the test circuit described in the manufacturer's application note here:

http://www.polyfet.com/tbplt/tb165.pdf

I did not make the circuit board itself--that was made by Jim Potter, who builds the RF hardware for the accelerator company I work for. These amps are a standard part of the low-level signal train for what ultimately is a ~hundred kilowatt power plant.

The amp I built, shown in the photo, had over 15 dB small-signal gain all the way from 30-500 MHz when I checked it on the network analyzer at work. At 200 MHz, it was 15.3 dB. When I got home, I tried running the AM-6155 with it. Even with the amp in "tune" mode, with the grid bias probably putting it in Class C, I can get close to 200 W out with ~6 W out of the SQ202 driver. Flipping the bias switch back to the "operate" setting for a second appears to get about 0.5 kW, at a great expense in terms of efficiency. I know some in the ham radio community like to push this amplifier, but I will be running it CW and don't need all that sauce. I'm happy with 200 W and the improved efficiency of a less linear bias point.

That's all for tonight!

-Carl
Attachments
RF_sq202pa.JPG
RF_sq202pa.JPG (133.59 KiB) Viewed 2455 times
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277

steve_rb
Posts: 133
Joined: Sun Oct 18, 2009 8:18 am
Real name: Steve Robinson
Location: Tehran University

Re: RF for new ion source

Post by steve_rb » Mon Oct 19, 2009 5:46 am

Hey Carl,

This looks great. I am very interested to try this. It looks very simple and exactly what I was looking for to get a sence of a plasma as starting step for building an ICP ion source myself . Could you list exact parts specifications along with key assembling instruction steps.

Steve Robinson

User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 11:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: RF for new ion source

Post by Carl Willis » Mon Oct 19, 2009 5:39 pm

Hi Steve,

Thanks for your interest. The most detailed information I have is probably in the top drawing shown here: http://carlwillis.wordpress.com/2009/02 ... on-source/

You'll have to adapt the idea to your own fabrication capabilities and make your own decisions about how to build it. But the entire design principle is simplicity and non-precision construction. This ion source seems to be able to do about 1 mA, tops, in its current configuration and drive level. The copper parts sputter quite a bit and copper is also known to promote the formation of molecular ions that are undesirable for nuclear reactions; if I built this again I would replace all copper parts with a non-magnetic stainless steel. You might also check out Andrew Seltzman's posts describing his ECR source or his more recent magnetically-enhanced DC ion source for different ideas.

-Carl
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277

User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 7:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: RF for new ion source

Post by Doug Coulter » Mon Oct 19, 2009 11:27 pm

Carl,
Nice work! The RF design here reminds me a bit of swatting a fly with an elephant, but it will work (obviously). Of course, that's because my own source is a much simpler one based on a magnetron from an oven ;~). I run that with about 20-50 watts *input* (not counting filament power, which is 30w).

Class C operation will also cut your cooling and plate draw needs, but will need significantly more grid drive to get to.

You'll find this works much better with quartz tubing than pyrex, the loss of which is getting to be quite high by 200 mhz (much less the 2.45 ghz I run). Changing to quartz on my setup made a world of difference for the better -- it runs for hours now and barely gets warm. If you need quartz tubing, get in touch, I've got a stock in various sizes, or can direct you to where I get it.

I also had copper sputtering troubles with mine (the extraction electrode of course), and have switched to something that doesn't sputter as well. I'm now using magnesium, but Al would probably be nearly as good, based on sputter rates from the old "procedures" book. At any rate, it's easy enough to clean off -- gun cleaning supplies are cheap and they specialize in copper removal chemistry. (ammonia or acetic acid work, but the pro copper removers add some things that work even better). With stainless, it will take longer to sputter to failure, but...will be lots harder to get off to fix it too.

I find that in my ion sources, an ECR magnet tuned right is worth quite a lot. For the same everything else I can run 2 orders magnitude lower pressure, or at the same pressure, *far* less power -- which in turn helps with sputtering, heat and all the rest. I've not yet made the promised measurements on atomic vs diatomic ions, but I will, and various literature says go ahead and do the magnet thing, it's better. In my case, I made a yoke that fits outside the rest with magnets on it, and one pole piece slides to adjust the field. It's not real critical, but moderately so. I happen to have gotten a magnetometer to measure this, and found that the simulations and other online resources are a bit optimistic about the net field you're going to get where it counts (eg in the middle of the gas tube) -- they tend to be off by about 20%, and that's more error than this process tolerates. In that range, though, you can just make a thing where you can slide one pole piece back and forth, and you quickly find the sweet spot. In your case, you'll need a mag field 200/2450 of what I'm using (about 980 gauss), which will be almost too easy to get -- and stray lab fields will mess you up somewhat if you're not careful about that.
Attachments
uWaveCloseup.jpg
uWaveCloseup.jpg (47.44 KiB) Viewed 2455 times
Why guess when you can know? Measure!

Post Reply