I am working on a new ion source (two identical ones, actually) to be implemented on my next fusor or other small accelerator projects. The design will use a magnetically-enhanced RF electrodeless discharge in a 3/4" Pyrex tube, and borrows heavily from Kiss and Koltay whose old paper is attached below.
There are many variants on the RF ion source, but what I favor in the Kiss version versus others, like the earlier Thonemann variant, is that the extraction is applied not across the discharge tube longitudinally, but via a negatively-charged puller electrode supported beneath the grounded mounting flange. In a fusor, such an arrangement means that particles enter the main discharge with an energy slightly lower than the main potential, and can be turned around at the walls of the fusor rather than run into them. My interpretation is centered on a 3/4" glass-to-metal adapter and a machined double-sided Conflat flange. More later on this.
Today I worked on an RF system to provide juice for the discharge at 200 MHz. I have a big old FAA amplifier, called an AM-6155, that I modified to suit the purpose. Some cathode bias resistors were removed, some grid bias components added, and I added my own shunt-feed components for the tube high voltage. Today it saw "first light," and produced 15-20 watts into a magnetically-coupled discharge in a 3/4" mercury vapor tube. Drive power into the AM-6155 was probably about half a watt. I intend to drive with 10 watts and extract at least 50 watts per each ion source in use. Unfortunately the amp has some funny behaviors that maybe the more RF-skilled folks will recognize and explain to me:
-Output power is low, but operating plate current is very high: in excess of 100 mA at a plate voltage of 2 kV. The idle plate current I have set with a zener on the grid to about 60 mA. (Another zener can be switched in at the front panel to nearly cut off the tube. All this works fine). What I don't understand is why I have so much plate current, clearly associated with drive power, and yet such a sorry power yield. I gotta figure out what's going on there.
-I can't tune this thing by watching for a plate current dip as I tune the plate resonator. Plate current is uniformly high no matter where the resonator is set. Maybe this problem is the same as the last one.
Pics show the setup causing a discharge in a mercury vapor lamp. The "antenna" is a single turn of wire, split opposite the coaxial feedpoint, with little capacitive tabs on the cut ends. Works great and a piece of cake to build. The other photo is of the amplifier per se, taken out of the power supply crate. The long pipe is the plate resonator. Tube is at left.
-Carl
|
2008-01-09 02:15 RF for new ion source 
(Carl Willis) [Latest: 2009-19-10 18:27]
(Carl Willis)