An unexpected and unwanted road trip has come up and preparing for it has been consuming my thoughts and my time. I must leave in 48 hours and will be gone for a week. Consequently I cannot pursue things with this amplifier untill I return.
At present, all 8 output transistors are removed - 4 from the left channel and 4 from the right. With absolutely no output transistors in place, the DBT glows steadily when the variac reaches about 50 volts.
Jumper between C and E of Q404 is still in place.
If it will be of any benefit, I am prepared to reinstall whatever output transistors I am advised to and resume voltage checks with variac at full power and DBT in place. I suspect this will cause the DBT to glow at it's maximum brightness. And perhaps blow the fuses.
Regarding normal background current draw.
Before the bias pot let go and caused whatever damage has occurred, the DBT went dark immediately after amplifier power was switched on. At idle, there did not seem to be enough current draw to make the DBT produce any visible glow. At least not under normal room lighting. In a dark room, there may have been a perceptible, feeble glow but I did not observe it under that condition.
If there are any further suggestions, I will gratefully accept them but cannot act on any until I get back home. I appreciate the patience.
At present, all 8 output transistors are removed - 4 from the left channel and 4 from the right. With absolutely no output transistors in place, the DBT glows steadily when the variac reaches about 50 volts.
Jumper between C and E of Q404 is still in place.
If it will be of any benefit, I am prepared to reinstall whatever output transistors I am advised to and resume voltage checks with variac at full power and DBT in place. I suspect this will cause the DBT to glow at it's maximum brightness. And perhaps blow the fuses.
Regarding normal background current draw.
Before the bias pot let go and caused whatever damage has occurred, the DBT went dark immediately after amplifier power was switched on. At idle, there did not seem to be enough current draw to make the DBT produce any visible glow. At least not under normal room lighting. In a dark room, there may have been a perceptible, feeble glow but I did not observe it under that condition.
If there are any further suggestions, I will gratefully accept them but cannot act on any until I get back home. I appreciate the patience.
It sounds like something else is amiss to me.
I would look at the voltage on each main rail, the positive and negative rails which would go to the collector of the output transistors. Each rail should be equal and opposite (polarity wise). See what voltage each comes up to when the bulb is bright.
Do not fit the outputs, no point at this stage and it might confuse things.
Perhaps the original faults have taken a diode out in a bridge rectifier
That suggests something has happened beside a normal output stage issue. Check the rail voltages first with the bulb lit and report back what you measure.
I returned from my road trip yesterday. I got some very sad news at my destination so it was not a pleasant one.
I am hesitant to take that action for fear of causing gross damage from sustained current draw. As mentioned, the DBT (40 watt) is fairly bright when the variac is at approximately 50V. But it is not at it's brightest.
How bright should the DBT be? Or, at what voltage should the variac be at to measure voltages at those output transistor collector connection points?
Should I increase the voltage past 50 volts before making the measurements?
I am hesitant to take that action for fear of causing gross damage from sustained current draw. As mentioned, the DBT (40 watt) is fairly bright when the variac is at approximately 50V. But it is not at it's brightest.
How bright should the DBT be? Or, at what voltage should the variac be at to measure voltages at those output transistor collector connection points?
Should I increase the voltage past 50 volts before making the measurements?
I'm sorry to hear that news 
The DBT is there to protect. If a diode in the bridge is failed for example then nothing bad will happen, if the output stage were conducting (if the transistors were fitted) then again nothing bad happens because the current is easily within the capabilities of the transistors.
Take it to the point the bulb stops increasing in brightness and/or the point the rails stop increasing. A 40w bulb is pretty low in the scheme of things, a 60 or 100 would be more the thing to use for a power amp... just don't overthink it at this point, just try and get a handle on what is going on and where the current flow is occurring.
The DBT is there to protect. If a diode in the bridge is failed for example then nothing bad will happen, if the output stage were conducting (if the transistors were fitted) then again nothing bad happens because the current is easily within the capabilities of the transistors.
Take it to the point the bulb stops increasing in brightness and/or the point the rails stop increasing. A 40w bulb is pretty low in the scheme of things, a 60 or 100 would be more the thing to use for a power amp... just don't overthink it at this point, just try and get a handle on what is going on and where the current flow is occurring.
I now understood what was shorting, causing DBT to glow.
It was collectors of the output transistor traces shorting to ground. My fault for not seeing it before but it hit me when I was about to try voltage checks again with a 100W DBT. I eliminated the shorts (there were several) and now DBT behaves properly - goes dim immeduiately after amp is turned on.
Output transistors in this amp are TO220 so a metal tab as collector. There is a screw that anchors the transistor and it is those screws that attaches the PC board to the heat sink. When I removed the output transistors, I replaced the screws so that the heatsink and PC board remained coupled. Without an insulator behind the TO220's, collector traces and metal heat sink were in direct contact.
So no screws in place and the DBT is happy.
Since I now know why the DBT was staying bright and since it no longer does, should I begin replacing the output transistors? I would install them one at a time and check the DBT after each one was put into place.
It was collectors of the output transistor traces shorting to ground. My fault for not seeing it before but it hit me when I was about to try voltage checks again with a 100W DBT. I eliminated the shorts (there were several) and now DBT behaves properly - goes dim immeduiately after amp is turned on.
Output transistors in this amp are TO220 so a metal tab as collector. There is a screw that anchors the transistor and it is those screws that attaches the PC board to the heat sink. When I removed the output transistors, I replaced the screws so that the heatsink and PC board remained coupled. Without an insulator behind the TO220's, collector traces and metal heat sink were in direct contact.
So no screws in place and the DBT is happy.
Since I now know why the DBT was staying bright and since it no longer does, should I begin replacing the output transistors? I would install them one at a time and check the DBT after each one was put into place.
I did not achieve being "happy with it". What transpired is this:
I began replacing output transistors, checking the DBT (now 100W) after each installation. Sorry, but I saw your advise to install NPN/PNP pairs after i began. Power was turnned on after each output was installed and DBT went dark immediately after power turn on.
The last two outputs left to install were Q307 and Q309 - both NPN.
FWIW, the original NPN output transistors were 2SC1060 and the original PNP output transistors were 2SA670.
In 2021, NPN output transistors became MJE15032G and PNP output transistors became MJE15033G.
I did not think emitter resistors need be changed but am open to be corrected.
So, the first of the last two remaining output transistors I installed was Q307 As before, DBT went dark and stayed dark as it had with the other outputs. However, after several seconds, R551 incinerated in a spectacular fasion - flame & smoke. DBT remained dark.
The service manual described R551 as 47 ohms and the diagram depicts it as the emitter resistor for Q551. Q551 is described as Ripple Filter. I pulled Q551, checked it and it was damaged.
The Q551 transistor I removed from the board was a 2SA735. The service manual says Q551 is a 2SA678. There are other 2SA678's on the board so I do not think the substitute originated at the factory. I guess it was done at a later time.
Searching for 2SA735 data produced none that I could find. I am going to assume it has the same pinout as 2SA678 because it was in place for the entire time the amplifier worked while in my possession.
I am not knowledgeble enough to imagine why it failed and burned up R551 in the process.
My understanding is that KSA1015 can replace 2SA678 and I have those at hand. Also at hand are 1/2W 47 ohm resistors to replace R551.
Would it be wise to install those two components and continue the output transistor replacement?
That ripple filter only feeds into high impedance nodes. It will also feed the same points on the other channel.
I would say only a dead short on that line could do this, the 33k and 3.3k resistors it feeds can't draw anywhere near enough current through the 47 ohm to make it burn.
It has to be either a physical short on the output side of the 47 ohm or C551, which is the 47uF in the ripple filter is faulty or (if you have replaced it before) it is fitted wrong way around.
Measure across the cap and see if you read a short. If not replace the cap anyway. If the cap has been replaced before do not assume the PCB marking are correct. Check by looking. The positive end of the cap goes to ground for this one.
The KSA1015 should be fine but in any case, this part of the circuit is 'only' a filter and the amp should still work even if the transistor was linked out. Voltage and voltage out are essentially the same, it just filter ripple and noise.
I would say only a dead short on that line could do this, the 33k and 3.3k resistors it feeds can't draw anywhere near enough current through the 47 ohm to make it burn.
It has to be either a physical short on the output side of the 47 ohm or C551, which is the 47uF in the ripple filter is faulty or (if you have replaced it before) it is fitted wrong way around.
Measure across the cap and see if you read a short. If not replace the cap anyway. If the cap has been replaced before do not assume the PCB marking are correct. Check by looking. The positive end of the cap goes to ground for this one.
The KSA1015 should be fine but in any case, this part of the circuit is 'only' a filter and the amp should still work even if the transistor was linked out. Voltage and voltage out are essentially the same, it just filter ripple and noise.