Originally Posted by tangent
Do you have a fairly sensitive meter? If so, give the AC voltage measurements of the input signal -- both before and after the pot -- and both output signals. Also give the R3 and R4 values. Do this with a simple sine wave.
High crosstalk is usually a grounding problem, which often comes down to the power supply. What power supply are you using?
How should i measure this? If i measure from ground to either input r/l before the pot, i get exactly 0.7V AC on both channels. After the pot, from ground to R1, i get nothing realy? Just tipping it with one lead, i get 1.49 on left and 1.44 on right. i don't know if this is a valid measurment. same if for the output.. from ground to l/r nothing, just tapping r or l 1.39 V AC. But that cannot be a correct measurment. All done with a 400hz sine wave.
for the psu i use your yjps at 16.7V output. i could not find a problem within the psu, but U1 gets pretty damn hot.
I just checked some of the voltages.. from ground to the op amps v+/v-, each rised by 0.1v per chip. left: 6.11V each, ground: 6.21V each, right: 6.31V each. Also.. from ground to the left output i get 1.7mV, right: 1.6mV and ground... 180mV.. something is wrong with that but what? I also swapped the op-amps... still the same issue.
I found an error! I'm feeling kinda stupid now, but i forgot to add C6. I don't know why, populated the board 2-3 years ago. Now each channel is about 1.6mV. The crosstalk is gone and so is a grounding problem (touching the headphone metal 3.5 to 6.3mm adapter). Whats left is that left/right problem. I mean that the right channel is "louder"/starts earlier.
Some pictures of my build:
Edited by nini_knoxville - 10/31/13 at 3:29pm