Just over 10 years later, and I am encountering similar behavior in a PL-400 Series II I recently acquired. The unit had a blown 10 amp main fuse, but intact 8 amp rail fuses.
I replaced the mains fuse, connected a variac and DBT, and proceded to troubleshoot according to the Service Manual, though the symptom does not exactly match anything in it.
The symptom is no DC offset, no abnormal DBT behavior in-spec bias on both channels, and undistorted output waveform on both channels that represents the applied input as viewed via an oscilloscope with NO LOAD.
When an 8-Ohm load is attached, the DBT appears almost fully on and never dims, and about +7 VDC is present on both channels as measured with a DMM. I do not leave the unit in this state long enough to see what an input signal does to it. During power off, I see as much as 31 VDC on the outputs after several seconds have elapsed, before the mains filters drain off.
Also, if I power the unit up without a load and subsequently connect an 8 ohm load to it, the DBT remains as if the load was unattached during power up.
I removed and tested every transistor and died on the rear of the unit as will as every one on the driver board. I have measured the 0.33 ohm emitter and collector resistors (the amp is a quasi), as well as the bias resistors for both channels
The rail filter capacitors were a mix of an original and a mallory replacement. Both had 1978 date codes. The original one was measuring low and the Mallory replacement was at the limit of its working voltage. I replaced them with 12000 uF units as well the Bridge rectifier with items from White Oak, so I could add bridge snubber caps and connections for a to-be-installed-afterwards Watts Abundant Speaker Protect Relay Kit.
I replaced the original op-amps with TL072's purchased from Mouser, as well as the 15 volt zener diodes and correct value (7.5K vs 1.8K) ballast resistors. and 8-Pin DIP sockets while I was at it. I replaced the large carbon composite resistors on the driver board with flameproof metal film equivalents. and the 4 electolytic capacitors with higher voltage and thermally rated nichicons. (Two of them tested leaky and under capacitance). I readjusted the bias to 375 mV, on both channels, after over 5 mins because I had no confidence that the original settings had not been altered somewhere along the line. I verified that the variac voltage was not sagging below its required value.
The behavior persists after making all these changes.
I have the "Watts Abundant" kit as well as a White Oak Driver board kit, but the instructions for the WA lead me to refrain from installing it until after I have eliminated this incorrect power up/down behavior.
Any help would be most appreciated.