Krell amplifier repair

[jpcar]

Hi,

I'm trying to trace through the power amp section of a Krell hifi amp. The right channel is working fine but the left is outputting a very low signal - about 10% of the good channel.
Schematic is attached.
I fed a 500Hz sine 80mVpp into the input with a signal generator. All voltages I've marked are peak to peak.
On the good channel, before R171 reads 80mV and then after is 2V. On the bad channel this point only gains to 120mV.
Can anyone offer advice as to where I should be looking next?

Thanks!

[g1]

That's not the signal path but the feedback path.  It's weak there because of the low output of the power amp.
The signal path is from the collector of Q80 to the bases of Q75 and Q76.

Compare the voltage at Q75 & 76 collectors to the good side.  Also check the DC idle voltages at E,B, & C of Q75 & 76.

[Jazz P Bass]

That is an odd if not interesting amplifier circuit.

Each polarity side is basically a mirror image.

Cute 'bias mute' circuit.

[jpcar]

Ah! Thanks for the help! I was looking in the wrong place.

So,
On the bad channel, the signal is at the base of Q80 but then it’s very small at collector of Q80. This is also happening at Q82 (the other polarity). I thought it’s unlikely to be both transistors (Q80, 82) failing but I swapped them with the good channel anyway. No difference. The same channel is still bad.

Q79,80,81,82 are all getting + & - 60vdc on their collectors. And Q75,76,77,78 getting + & - 60vdc on their emitters.

Is there something a miss around Q83, 84?
All the voltages look the same on both channels.

Any thoughts..

Thanks

[jpcar]

I'm stumped with this.

The signal comes in. Is present at the base of Q80 & Q82, and then drops to a very low signal at the collectors of Q80 & Q82. The very low signal is at the base of Q75,76 and Q77, 78 and gets amplified back up at the collectors of Q75,76,77,78.
On the other (working) channel, the signal at Q80,82's collectors is about the same as their bases, so by the time it gets to the Q75,76,77,78's collectors it's many times larger.

I just can't figure out why this is happening on both Q80 and Q82. I swapped those transistors with the good channel just in case, but it's exactly the same.

All transistors are getting their power from the +60 & -60 rails.

[g1]

Is there something a miss around Q83, 84?
All the voltages look the same on both channels.

DC voltages are all the same for both channels?  Even Q80 emitter and base?
If you don't see any differences I would maybe try swapping out Q83 and 84 with the good channel.  Those big W42 and W92 transistors seem to have their fair share of issues.

[jpcar]

All DC voltages are the same for both channels. I have now swapped Q79,80, 81,82 and 83,84 from the good channel to the bad channel and the problem remains the same. Very low signal output from the collectors of Q80 & 82. I tested all the DC voltages around Q83, 84 and the resistors and diodes around there. Both channels are the same.
I tried removing Q75, 76 to see if they were somehow were pulling down the signal at C of Q80. It remained the same. The only difference was slightly higher DC voltage because two less transistors were being powered.
Is there any way to isolate part of the circuit to see if it's working? Like just apply signal to the base of Q80 or Q82 to see if one side works? I'm not sure if that's a bad idea. Everything is so interconnected. Not sure where to go from here.

[jpcar]

I tried removing Q82 so that the signal just flows through one side of the amp.
On the good channel, I see a signal that’s amplifying only the positive half of the sine wave and not the negative, as expected. But on the bad channel, I’m seeing the full wave, just not amplified much. Looks exactly the same as if Q82 was there.
Next I’ll try removing Q80 and see if I get the same results.

[g1]

I think the results will be the same.

Can you post DC idle voltages for Q73 & Q74 bad channel, and good channel.

[jpcar]

DC idle voltages for Q73 and Q74 are the same for both channels.

Q73:
C= 66v, B= 1.7v, E= 1.1v

And for Q74:
C= -66v, B= -1.8v, E= -1.2v

[jpcar]

In an attempt to get further with this...

I removed a few transistors so I can try to isolate the problem. Q73, 75, 76. So now the signal is not getting any further.
The results are good channel in green and bad channel in red. Blue is the same for both. These are all 500Hz sine signal with peak to peak measures.
I tested DC voltages and all the transistors are getting same voltages on the good and bad channels. I tested all the resistors and they all seem good. 

In the schematic I'm just showing the results on one side but the other side (bottom half) has exactly the same results.

There's a strange signal (200mv) at Q79 base that continues all the way through the circuit out to the speaker terminals.

I have switched out Q80, 83, 79 (and Q82, 84, 81) with the good channel.

Any advice on how to hunt down the problem? I'm not sure I'm approaching this the right way.

[Jazz P Bass]

Not too sure removing parts is the way forward on this.
That 200 mv signal is coming back on the feedback path.

[jpcar]

I found the problem!

You're right though. I don't think this was the best way to approach it but I couldn't think of another way to isolate the problem.

The fault was C78, 79. One of them was almost shorting.
I don't know why I didn't suspect these sooner! They looked ok. Not bulging or leaky. Never trust a capacitor I suppose! I thought I checked them but I guess I didn't check them properly.
I learned a lot.
The problem was on both sides of the channel so I suppose I should have focused on what's common to both. That area from the bases of Q79 and Q81.

Thanks for all your help!

[Jazz P Bass]

My hat is off to you.
So the feedback path was the issue.

[g1]

Well done sticking with it.  Many would have given up on it.