RF ion source v2.0

[Carl Willis]

I have modified my Kiss-Koltay inspired RF ion source to use 2450-MHz microwave power rather than 200 MHz. Additional improvements have been to replace most copper parts with stainless steel and to adopt better feedthroughs for gas and the extraction voltage. I'm not sure yet what this all means for results, but the beam-target neutron generator into which this is operating produced about 8000 cpm on my little He-3 tube on the first run, beating the previous source and with significantly less current actually delivered. I suspect a good tuneup will bring much better results, so I'm excited.

Original design was discussed here:
viewtopic.php?f=12&t=5006#p32314

Some updates and data are at the following links:
viewtopic.php?f=12&t=5006#p32960
viewtopic.php?f=12&t=5008#p32316

The beam-target neutron generator using this ion source was discussed here:
viewtopic.php?f=15&t=7273#p49145

Little has changed with the source itself. I picked a more robust, more economical HV feedthrough for the extraction supply. I got rid of the troublesome capillary gas feed and adopted 1/4" VCR fittings, using a 5-micron laser-drilled gasket for deuterium conductance control. Copper has been eliminated from the extraction electrodes, seemingly to good effect in limiting evaporation and sputtering. A list of component parts may be found below. The basic design requires a lathe and a drill to make, as well as soft silver solder and silver braze, but nothing requires notable precision. This time around I used the lathe to ensure coaxial alignment of the extraction electrode during brazing. Exact dimensions depend on how parts fit together and are best left to individuals to determine in their own projects.

The new RF system is a cheap Chinese magnetron driven by two low-wattage MOTs salvaged from junked ovens. One MOT is run directly off the switched AC line and provides the ~3V / 10A or so for the maggie's heater, which in ordinary service is largely heated by back-bombardment of the cathode, but this can't be relied upon when the anode supply is variable. The other MOT is operated in the classic level-shifted ferroresonant anode supply with a cap and rectifier. Its primary is fed from a 0.9-kVA 0-120V variac. Thus, crude control of the output power is achieved with one knob. The maggie is clamped into a piece of aluminum WR284 waveguide with the antenna 1/4 guide wavelength from one shorted end (note: guide wavelength must be calculated from waveguide dimensions and is greater than free-space wavelength). A hole is drilled for the ion source's discharge tube 1/4 guide wavelength from the opposite shorted end. To tune the system, a classic 3-stub tuner is interposed between maggie and load. The stub tuner consists of three #10-32 brass screws spaced 1/4 guide wavelength. Tuning protocol is usually iterative, starting with the screw closest to the load and proceeding back to the source, and this scheme works well here This system is easy to tune decisively. Attachment to the waveguide of the brass nuts for the tuner, aluminum end fixtures, and a copper pipe choke around the discharge tube is accomplished with a soft aluminum solder (Muggy Weld Super Alloy #1).

If components are sourced with reasonable care, it should be possible to do this project for under $200. (That does not include gas delivery, extraction supply, consumable supplies like solder, or tools like drills and torches that are necessary).

Parts list: Ion source
-Extraction terminal: Ceramaseal 0215-01-B brazeable low-profile terminal ($28.93)
-Extraction supply feedthrough: MDC #9421011 15A / 5 kV nickel feedthrough ($31.00)
-Extraction cathode body: 3/8” OD thick wall SS-316 instrument tubing, Swagelok (~$6 / foot at local distributor), turn downstream end to fit extraction terminal for brazing
-Extraction cathode aperture: Lathe-turned from 0.625” SS-304 bar stock (Metals Supermarket)
-Extraction anode: Lathe-turned from 0.625” SS-304 bar stock (Metals Supermarket)
-Extraction cathode flange: Double-sided 2-3/4” CFF (Kurt Lesker DFF275X000 or equivalent, $40.20), lathe and drill modifications
-Gas delivery flange: Standard 2-3/4” CFF (Kurt Lesker F0275X000N or equivalent, $15.00), lathe and drill modifications
-Discharge tube seal: ž” compression fitting (Kurt Lesker B-075-K or equivalent, $34.00)
-Discharge tube: 19mm standard-wall borosilicate, custom OR ž” borosilicate test tube
-Magnets: N40-grade 1” long x ź” square lengthwise-magnetized NdFeB magnets (17) (EBay pricing)
-Waveguide: silver-plated aluminum WR284 (Ebay), end plates soldered in using Muggy Weld Super Alloy #1
-3-stub tuner: ź” guide-wavelength-spaced #10-32 brass screws, in brass nuts soldered to the Al waveguide using Muggy Weld Super Alloy #1
-Magnetron: Galanz M24FA-410A ($14.99 Amazon.com)

Parts list: Gas delivery
-Gas flow restrictor: Lenox Laser SS-4-VCR-2 w/ 5 micron hole ($39.50)
-Gas valve: Nupro ź-turn bellows-sealed shutoff valve ($20.50 on eBay)
-Gas regulator: Premier Industries custom regulator (brass, single-stage, hydrogen w / ss membrane) ($90.00 in 2006)

Photos below show various stages of construction and initial operation. Captions:
(0,1) A stainless cap from bar stock has been brazed into the tube that forms the extraction cathode, and the extraction channel is being drilled in the lathe.
(2) The extractor is brazed into the Ceramaseal terminal, using live center for alignment
(3,4,5) Views of assembled ion source
(6) Gas delivery system: D2 cylinder, regulator, laser-drilled orifice, shutoff valve, VCR hose
(7) Waveguide, showing three-stub tuner and choke for discharge tube
(8) Inside maggie control box
(9) Initial test and tune of microwave circuit, using a sealed tube at 15 torr as load
(10) Neutron generator with microwave circuit and discharge-tube magnet in place
(11) The rosy glow of plasma during operation
(12) Closer look into the waveguide during operation. Magnetron antenna feedthrough glowing red from heater power; behind it are the three tuning stubs; at the far end is I.S. discharge tube
(13) Beam entering the target assembly at about -50 kV, accompanied by a slight thread of unsuppressed electrons. This is a neutron-producing beam.

-Carl

[DaveC]

Hi Carl -

I can tell you're havin' fun. Nicely done... great pictures. We have one of those cute little lathes in the lab at work. Very useful, for lots of things.

Thanks for posting.

Dave Cooper

[George Schmermund]

Nice project, Carl. This makes me want to plug-in some machine tools and stop playing with these back patio projects!

[Steven Sesselmann]

Carl,

Excellent work, you are entering a productive phase....

Would like to know more about the wave guide tuning screws, how does it work and why is it a requirement?

Steven

[John Futter]

Nice work Carl.
You should mention the mesh over the open waveguide port so that others use it. It is not too noticeable in your pics.
I know that your final setup is different from the discharge tube pic but you would want a second waveguide beyond cutoff to stop leakage around the discharge tube (ie the vacuum system side)


Steven -Microwave triple stub tuners allow one to match the characteristic impedance of the wavguide and load to the generator (the magnetron in this case). It is possible to shift reactive loads so that they display a perfect resistive match to the generator.
If Carl doesn't mind I'll post a few pics of what actually is achievable @ 2450MHz with a good triple stub tuner on my Vector network analyser, probably in a new post.

[Chris Bradley]

Carl,

Fine work.

What is your interpretation of the last image?

[Richard Hull]

Great work Carl! It is nice to see this rather moribund forum come to life every now and then with sweet projects like this.

Richard Hull

[Carl Willis]

Thanks for the appreciative comments.

Steven: The three-stub tuner has been a fixture of microwave impedance matching for some time. With three stubs, it is possible to match any load, while single or double stubs can only match certain loads. In its better embodiment it actually is constructed with stubs (i.e. pieces of waveguide) branching from the main guide, with sliding shorts to adjust the admittance of each. Mine is a lossy, boganized offshoot that is nonetheless quite common, employing screws rather than stubs. The theoretical justification is implicit in the Smith chart presentation in the letter attached.

John: Yes, microwave safety is important and I made no mention of it. This is an electrically closed guide (with perforated aluminum at one end, hence the interior photograph), and all holes are substantially below cutoff including those for the discharge tube. It is true that the discharge tube itself can in principle conduct and radiate some microwave energy outside of the guide. The copper-pipe choke is intended to prevent that. I can't detect much leakage with this particular discharge.

Chris: The last photo just shows the deuteron beam passing from the ion source (at right) into the target (at left), which is at -50 kV. A few unsuppressed electrons form a wisp at the center of the beam.

-Carl