Epiphone Firefly 30 DSP hum

[DrLev]

Anyone here ever worked on an Epiphone Firefly 30 DSP ?
I have one in for repair. The problem is hum.
It's caused by an imbalance of the idle current at the output transistors.
Q34 base to emitter -0.38 volts
Q36 base to emitter +0.52 volts
I have checked every component and replaced all the transistors to no avail.
I noted the schematic shows the voltage rails as +42V and -42V whereas the rails on this read +30V and -30V. The transformer is 22V-0-22V.
I see a schematic for an Epiphone Triggerman 100 DSP on this forum.
It is the exact same schematic as the Firefly 30 DSP.
I can only assume Gibson used the same board and used different transformers for different output power.
I powered up this board with a variable transformer and brought the rails up to +42V and -42V. The outputs became more balanced and the hum disappeared.
Also I can balance the outputs and get rid of the hum if I change R323 from 330R to 470R. I was wondering why I would have to do that and if I leave that change would it be stressing any of the components.
I would appreciate any thoughts or ideas on this.
I don't know how to upload the schematic here but it is exactly the same as the Triggerman 100 DSP here on forum except the name.

[g1]

Also I can balance the outputs and get rid of the hum if I change R323 from 330R to 470R. I was wondering why I would have to do that and if I leave that change would it be stressing any of the components.

You tell us where you came up with the idea to change that resistor. 
Likely what is happening is that you are compensating for a damaged or out-of-spec component, which is not how you want to do the repair.
2 things happen when you adjust R323, you alter the balance of the differential pair (Q312,Q313), and you change the bias of Q39, which affects the conduction of the output devices.

You did not mention whether there was any DC on the output, or the voltages of the drivers (Q310,Q315).

Differential circuits can be very picky about out of spec parts.  In some modern circuits, replacing the differential pair with new parts from the same batch will cure issues like this.
But it may also be a fault somewhere else in the circuit. 
In any case, modifying component values is not a good fix, if it was a poor design, all amps of this model would have complaints about the issue.

[DrLev]

Thank you for replying. I will do some voltage readings tomorrow.

[DrLev]

I did some voltage checks today.

20 millivolts at speaker output with or without speaker connected.


Q315 voltages referenced to ground

With speaker connected
Base -0.915 volts
Collector -30.4 volts
Emitter -0.377 volts

With speaker disconnected
Base -0.974 volts
Collector -30.4 volts
Emitter -0.435

Q310 voltages referenced to ground
With speaker connected
Base +1.07 volts
Collector +30.4 volts
Emitter +0.535 volts

With speaker disconnected
Base +1.00 volts
Collector +30.4 volts
Emitter +0.467 volts

I becomes more balanced when the speaker is disconnected.
Would that indicate the problem is the output transistors are out of balance even though I did change them?

I would appreciate any help.
Thanks.

[g1]

Measure voltage across R320, with and without speaker.  It will be millivolts DC, so use whatever range is most accurate.
Then do same for R321.

[DrLev]

Thank you for getting back to me.
R320
With speaker
1.5 millivolts
Without speaker
0.0 millivolts

R321
With speaker
0.0 millivolts
Without speaker
0.0 millivolts

[g1]

Try removing or disconnecting one end of R35 (10K), then repeat those last measurements.

[DrLev]

Thanks again.

R35 disconnected.

R320
With speaker 1.2 millivolts
Without speaker 0.0 millivolts

R321
With speaker 0.0 millivolts
Without speaker 0.0 millivolts

[g1]

It seems like Q34 is not conducting at all, and Q36 just barely.

edit:  I think I went the wrong way with R35, to increase idle current of the outputs, you need to reduce R35. 
If you have a 10K pot you could wire it up as a rheostat and start with a 10K setting, then slowly reduce til you see a few millivolts across each of R320 and R321.  From there once you have a good setting, you can measure the pot resistance and install a resistor of that value for R35.
If you don't have a pot to use, you can do trial and error with values less than 10K for R35, or add a parallel resistor if that is easier than replacing it.
The way it is set up, R35 and R318 in parallel are making about 900 ohms.  You want to slowly decrease that to get the outputs conducting.

[DrLev]

Thanks again for your reply.

Here are the results.

Combined resistance of R35 and R318 decreased to 790ohms

Q34 base to emitter -0.572 volts
3.0 millivolts across R321

Q36 base to emitter +0.566 volts
4.6 millivolts across R320

The hum is still there.
Might be a little louder.
If I reduce the resistance lower than 790 ohms the hum gets a lot louder.
As I sweep the resistance,using a pot,from 900 ohms down to 790 ohms the hum does not reduce at any point.

I had another SA1695 so I changed it.
All the voltages stayed the same.
I don't have another SC4468.
I had already replaced both of these before.
I'm just wondering if they are all unbalanced.