> Solid-state relays around high power RF tend to be problematic.
That turns out to be the answer. After posting my original message, I noticed
that it did NOT happen when I transmitted into a dummy load, only an antenna.
The acutating voltage comes from a 78L05 regulator which is bypassed both on
input and output. The wire to the relay box is shielded. In the relay box, I
added a pi-section filter in the actuating circuit. It didn't help.
Then I realized that there was probably a high RF potential between the
actuating circuit and the line terminals. I put a couple of 0.001 mf 3KV
capacitors from the line terminals to ground, and it solved the problem!
I presume that the RF was shutting down the relay, causing the supplies to draw
more current, as Phil suggested.