Ampeg G110 problem

[Bishop]

I have an Ampeg G-110 that sounds good for a few minutes, then starts popping (like a low E string hitting the metal on a pickup) while still sounding good, then after a while it sounds bad and fades out.

I believe this is related to the fact that the 24VDC part of the output section starts at 35V and slowly climbs to 48V, which is I think when it fades out. I can't figure out why.

R52 was discolored. I changed this and R53. Do these set the 24V supply (see next question)?

The base of Q9 is around 37V to start and also rises. Am I right to assume that this should be closer to 24V?

I replaced Q8, Q9, and Q6 with no change. I also replaced C34 and C35 with no change. Q7 measures fine with my meter, so does the diode. If I continue shot-gunning those would be next. Resistors all measure correctly.

Q5 has normal static voltage, the odd voltage starts on the other side of R48

Q6 and Q7 base and emitters have normal voltages. Both collectors have the 37V and rising problem.

Odd note: I can't measure anything in the signal path up to Q5 base without major squealing.

I could use some help if anyone want to take a crack. The part of the schematic I'm focused on is top right.

Thanks!

[The Dude]

I wouldn't keep shotgunning. It's not really an effective way to fix the thing. This amp works with a single 48V supply instead of the more typical plus and minus supplies. Q8 and Q9 are the outputs and if they are both operating correctly and conducting evenly, the midpoint is 24V at the output. That is why C35 coupling cap is necessary. Without it, there would be 24VDC at the speaker. Either something is shorting on the high side or something is going open on the low side causing your 24V midpoint to go high. Let's start with voltages on the output transistors. Can you measure and post those? If you can, just write them on the schematic and repost. That way we have a running measurement diagram. Remember, it does not have to be a bad part. It could be an intermittent connection issue or something of the sort. That is why shotgunning is a bad idea. You could throw parts at it all day and not fix it.

[J M Fahey]

I suspect C32 is leaky.
Just for testing, remove it and recheck you have proper 24V at the output or rising 35V as before.
Second suspicion is around R46, it should have 22.5V on the top leg and a few millivolts on the bottom leg.
On R 50: right leg 0mV ; left leg 8.3mV ... post what you actually measure.

For every value I ask please post 2 results: one with amp just turned on (cold) , second 10 or 20 minutes later or when voltage drifts from 35V (which is already wrong) to 45/48V (which is worse):

We might have a thermal problem.

Finger test transistors from Q5 to Q9 .

Only Q7 has reason to be warm (definitely not finger scorching though) because it dissipates respectable 1.6W at idle.

[Bishop]

I wouldn't keep shotgunning. It's not really an effective way to fix the thing. This amp works with a single 48V supply instead of the more typical plus and minus supplies. Q8 and Q9 are the outputs and if they are both operating correctly and conducting evenly, the midpoint is 24V at the output. That is why C35 coupling cap is necessary. Without it, there would be 24VDC at the speaker. Either something is shorting on the high side or something is going open on the low side causing your 24V midpoint to go high. Let's start with voltages on the output transistors. Can you measure and post those? If you can, just write them on the schematic and repost. That way we have a running measurement diagram. Remember, it does not have to be a bad part. It could be an intermittent connection issue or something of the sort. That is why shotgunning is a bad idea. You could throw parts at it all day and not fix it.

I hear ya. Maybe I used the wrong wording, there was some thought to my process, it was just misguided by lack of knowledge. I'd rather swap a few parts than waste anyone's time on the forum.

[Bishop]

I suspect C32 is leaky.
Just for testing, remove it and recheck you have proper 24V at the output or rising 35V as before.
Second suspicion is around R46, it should have 22.5V on the top leg and a few millivolts on the bottom leg.
On R 50: right leg 0mV ; left leg 8.3mV ... post what you actually measure.

For every value I ask please post 2 results: one with amp just turned on (cold) , second 10 or 20 minutes later or when voltage drifts from 35V (which is already wrong) to 45/48V (which is worse):

We might have a thermal problem.

Finger test transistors from Q5 to Q9 .

Only Q7 has reason to be warm (definitely not finger scorching though) because it dissipates respectable 1.6W at idle.

Thank you sir... C32 it was. Leaky tantalum. 25V holding steady. Replaced it with an E-cap, we'll see if that was that was the whole problem, but I am pretty confident.

Thanks both for the reply's, much appreciated.