Enter the Bread-slicer

Dual 300 pF variable capacitor. Plate spacing is about 1/2".
That's a half-dollar coin on the workbench. This sucker
is BIG. Breakdown voltage is probably around 14 kV.

I've been having trouble with my 80m shunt-feed matching network. Since I moved from Floyd County, I found that the SWR shifts when I transmit more than about 20-25 watts. The Elecraft K3 will work just fine at it's tuning level of 5 watts and find a perfect match. But, when more than about 25 watts is applied, the SWR shifts.

This has made the shunt-fed tower unusable on 80m. This is especially bad, since the tower is my only 80m antenna at the moment in Gwinnett County. I've written about my pursuit of 5-band DXCC. I need an 80m antenna.

The question in this case -- what causes the shift? No doubt it is due to arcing. Initially, I thought the 80m network had the same problem as I had on 160m -- that the inductor was arcing between turns. I originally used enameled wire to wind both inductors. Over time, with the heating and cooling of the coils caused the windings to rub against another, which eventually caused the 160m inductor to arc between turns.

I re-wound the 160m inductor with insulated wire back in 2006, right at the start of the ARRL 160m contest. I haven't had any trouble with arcing. It was behaving much like the 80m network was.  Re-winding the 80m inductor was easy. Unfortunately, it didn't solve the problem.

There's not much else to the matching network. It's a simple L network with a series inductor and a shunt capacitor. It might be the capacitor, it might be some part of the switching network. How to find out?

What I needed was a capacitor with greater breakdown voltage. The variable capacitor I used on 80m is a 15-250 pF unit that was the plate-tuning cap in a DX-40. A survey of likely substitutes found nothing that suitable in my junk box. I'd need to buy a replacement. But first, I needed to know if the capacitor was the problem.

I did have one part in my junk box that I might use. The Bread-slicer. You see, when my friend Mike (now W1YM) moved away from Atlanta, he gave me a whole bunch of ham-related stuff, with instructions to sell off what I couldn't use and fund my tower project. And that's exactly what I did. Among the items he gave me were a couple of very large variable capacitors.

One of them was far to small to work -- only about 15 pF max. The other, however, was perfect. It's a dual 300 pF cap with a plate game around 1/4 inch. Breakdown voltage for this cap is over 10 kV. I call it the Bread-slicer, since it looks large enough to slice an entire loaf of bread at once. It would be perfect for a test.

The Bread-slicer in position at the base of
the tower. Note that I don't even have a
knob large enough to fit on the shaft.

The problem with this cap is that it is huge. It's literally bigger than a breadbox. No way it would fit in the 6" cube of the NEMA box that houses the 160m and 80m matching networks. But, it would prove to me if the variable cap I was using was the failing part.

So, one very damp evening this week, I put the matching network back on the tower without the 80m capacitor and jumped the Bread-slicer in the circuit. Lacking any knobs large enough to fit on the shaft of this enormous capacitor, I used a pair of vice-grips.

With some quick work with the MFJ-259 antenna analyzer, I found a 1:1 match right around 3800 kHz - just as expected. Hooking up the coax and stepping back inside, I tuned the K3 to the 80m shunt-feed, and gave it a quick full-power test.

No SWR shifts. Yes, it's definitely the variable capacitor that was arcing. Since I can't use the Bread-slicer on a permanent basis, I've already ordered another part. It's only 80 pF, but has a 4 kV breakdown voltage. And, it should fit in the NEMA box. Once I get it, we'll see how well it works.