Originally Posted by hopeless
But does it sound that bad?
You could just bypass the first, unity gain stage, and see what happens (to the sound).
No, it doesn't sound as bad as it looks. It's just a bit dull, probably because of the very obsolete TL082. I think the cure will be:
1. to use a single opamp on a dual BrownDog, just keeping the gain stage by hardwiring the input to the pot, I can see no need to keep the input buffering stage
2. remove C2, it just **cks up the signal
3. bypass C1, there's absolutely no reason for this if JFET input opamps, or modern bipolar input opams with low bias currents are used
4. bypass C4, there's no need for this either for the same reason as in point 3, who wants electrolytic caps to mock up the audio signal?
5. bypass R10, I want a low output impedance, this resistor will increase crosstalk and reduce the damping factor
6. increase R7 to 1M to better match the 100k pot and JFET input amps
7. remove the zobel (R11 and C3). If the amp is unstable without zobel, I'll change the values to 47R/47nF instead. 10R/10nF is mocking up the signal.
8. depending on the results, I'll probably reduce the input R5 and feedback R13/R14 resistor values to reduce the noise
I was thinking of bypassing the DAC opamp, so I'll have two OPA627's to go into the headphone amp.