This was one of the bummer stories. I fixed it and returned it to the client, then a week later it ate a brand new set of tubes. I had seen the bias voltage drop out when I had it before, but thought that I had addressed the problem – the ground connection for the low-voltage supply had been cracked and compromised, and I fixed that. The amp ran for two hours at maximum power on my bench. But I hadn’t cleared the ENTIRE problem.
In the picture you can see two electrolytic caps on a small board to the left. That board is the low-voltage supply, fed by the orange PT secondary wires. It produces dual-rail +/- 43VDC on the output. Those two caps are the filters for each rail, immediately after each half-wave rectifier. The bias supply is tapped off of the negative rail of this supply.
The arrow on the right points to the bias control and a set of zener regulators that knock the +/-43VDC down to +/-16VDC for use in the preamp. They have their own filter caps.
In this case, the two reservoir caps on the left had cracked solder joints. When the amp would heat up, they would lose contact, causing the bias supply to see huge 60Hz half-wave ripple instead of real, flat DC. This meant that the actual grid bias on the power tubes went from -34V of clean DC to an ‘average DC’ of -18V of unfiltered half-wave rectified AC. Instant redplating on all four tubes.
What obscured the issue is that there was still enough DC left in front of the zener regulators that the +/-16V regulated output was completely unaffected. The preamp continued operating normally. This is a case where a seemingly rational assumption – that everything downstream of a faulty supply will be affected by its malfunction – was completely wrong, and cost a lot of unnecessary time and money as a result.
Just played a show in Columbia, SC with this new band called Ivadell (members of Darkentries, Apart) and their rigs are good as hell and loud as the dickens. A few FYAW followers among their ranks as well. Check them out, they just blew us all away!