Wow! That input stage has more feedback than a Billy Graham rally...
Q1 and Q2 are the ones to replace. Also, R51 should be tied to ground through an electrolytic cap of some sort...if the cap is old and leaky, then it can also cause excessive offset and should be replaced. Then again..because the differential pair is included inside the feedback loop the coupling cap might be unnecessary.
The opamp at the input looks like a standard 8-pin DIP package. Too much offset from the opamp could possibly be so bad that the differential pair (Q1, Q2) can't compensate. Plus, it's likely old and of questionable sonic character. I suspect that excessive offset from the opamp (due to large bias currents) as well as mismatched transistors are causing the offsets you see. On this amp, I'd replace the opamps first, with something like an OPA604 or an OPA134 (FET input opamps with very very low bias currents) might reduce or even eliminate the problem, and is what I'd do first. At the very least they'd improve the sound. I question the necessity of C5 (I think it's C5) from the opamp output to the inverting input. If I changed the opamp I'd want to check for stability with a scope and square wave. The cap should go if at all possible, but stability needs to be looked at both with the cap and without it.
If the opamp swap didn't get the offset low enough, then Q1 and Q2 should be replaced with matched pairs of high gain transistors. If you can't match transistors, then simply use transistors with such high gain that base currents are as low as possible (no, you can't use Darlington's). I use Zetex ZTX795A's for input stages like this...
Q1 and Q2 are the ones to replace. Also, R51 should be tied to ground through an electrolytic cap of some sort...if the cap is old and leaky, then it can also cause excessive offset and should be replaced. Then again..because the differential pair is included inside the feedback loop the coupling cap might be unnecessary.
The opamp at the input looks like a standard 8-pin DIP package. Too much offset from the opamp could possibly be so bad that the differential pair (Q1, Q2) can't compensate. Plus, it's likely old and of questionable sonic character. I suspect that excessive offset from the opamp (due to large bias currents) as well as mismatched transistors are causing the offsets you see. On this amp, I'd replace the opamps first, with something like an OPA604 or an OPA134 (FET input opamps with very very low bias currents) might reduce or even eliminate the problem, and is what I'd do first. At the very least they'd improve the sound. I question the necessity of C5 (I think it's C5) from the opamp output to the inverting input. If I changed the opamp I'd want to check for stability with a scope and square wave. The cap should go if at all possible, but stability needs to be looked at both with the cap and without it.
If the opamp swap didn't get the offset low enough, then Q1 and Q2 should be replaced with matched pairs of high gain transistors. If you can't match transistors, then simply use transistors with such high gain that base currents are as low as possible (no, you can't use Darlington's). I use Zetex ZTX795A's for input stages like this...
awprint: