- The parts before assembly
This is the kit of parts before we cut the final pipe and put it all together. Jon cut the nice slits in the end of the magnetron adapter by hand with a dremel -- but it looks like it might have come off of one of the machines they are so straight and perfect.
- Hardline details
Here's some detail on the hardline as we made it. The center conductor is 1/8" OD copper pipe, drilled and tapped in one end for the 4-40 adjustment screw. The other end has a 6/32 nut tapped on to provide a flat surface for soldering to the end of the magnetron "antenna". While coupling isn't the most critical adjustment in this, it's good to have it adjustable, and our first light on the thing had it too short, so we were glad to be able to do that. I filed the screw head into a sphere spinning it in the drill press, which works well to cut down arcing issues and make it a better more uniform line.
- Inside the cavity
Here's a shot looking into the end of the cavity. The proper protrusion of the hardline conductor is about 3/16" inch when things are working sweetly. We put a turn of teflon about 1/2" wide just inside the end of the outer conductor of the feedline, as my other one needs that to not have arcing issues. This one never showed the least sign of that, even running at roughly double the normal operating power.
This time, we were a little smarter and drilled all the holes a tiny amount oversize so the quartz would pass easily, and rather than putting in the gas through this (with the problems I had with that on rev 1) we just put an adjustable "pusher" electrode in the pipe cap we turned from brass. On the first try, we accidentally had it too long -- and it was resonant with the RF so things acted really funny. Shortening it 1/4 inch or so took care of that one, and it's up a lot closer to the action, and so should work better than rev 1 did.
Another improvement was the use of another turned copper plug for the non tuning end cap, instead of a normal pipe cap. This means we didn't have to slot the main cavity, but just drill through it, or notch a pipe cap. It was made a perfect friction fit by cutting it to zero tolerance with the pipe ID, then upsetting a couple spots with a centerpunch to raise the metal a little. No solder was needed for either end cap to get them tight and RF-good.
The last and most important improvement was moving the feed line down about 1/4" closer to the ground end of the resonator rod. This lowered the impedance and voltage at that point, which is good -- when there was phase shift between the drive and the rod before, we had arcs during adjustment, and sometimes even when adjusted. That just "went away" for this one.
