A friend of mine just gave me an AB763 amp to look at. He said "one of the tubes was sparking." The amp (I built it) is about six months old and the JJs are new. It's a Super Reverb circuit running with a solid state rectifier. Took it home, pulled the tubes and checked on my tester (Hickok TV-D10/U mutual conductance tester. The tube in question passed all shorts tests, but when I hit the button to measure transconductance the needle bottomed out and the tube glowed lavender. With tubes out, I went through the amp and all seemed OK. Put a fresh set of JJs in, brought it up slow on the variac and the numbers on the 6L6s are fine. They are consistent with the plate, screen grid, control grid and cathode voltages specified on the schematic. Maybe 10-15 volts high overall, but otherwise consistent. Currently the tubes are running about 478 on the plates, with dissipation around 40ma on one and 38ma on the socket that had the bad tube. So they're biased right around 60 percent or so. Don't want to bias them any colder for a few reasons ... tone and the fact that the plate voltage will rise are the chief ones. At 60 percent and 478 plate voltage though amp has been running smooth, stable and fine for hours. What would cause that internal glow? Might I just have gotten a bad JJ? I've read about JJs handling 500VDC on the plates no problem, as long as they're not biased hot. I think my biasing is pretty conservative at 60 percent so I'm wondering if it was just a tube that decided to quit on its own. Any thoughts?