Headphone amp -- Oscillations fixed!

[riggler]

This amp sounds great!!

I was about to start desoldering components when I decided to go through the circuit one more time. Checking for continuity / non-continuity, I found that at the left channel gain pot leads I was getting continuity, even with the pot turned all the way down. Hmm...

So I thought maybe the pot wasn't tracking perfectly down there. Well I desoldered the pot to try another and I still had continuity! Turns out there was a short in the cabling where two nearby jackets melted together. Couldn't even see it. It looked totally fine. So fixing that resolved everything with the intermittent noise. I also re-routed my grounds per PRR's advice, except I don't see a ground point on the schematic around the 634 "rail caps"; did you mean at the junction of the 47pF and the 2.2k resistor? Didn't seem right.

Anyway--
The right channel is ever so *slightly* hotter than the left, just a touch. Can I simply add a resistor after the right channel pot to balance things out? (Or I could measure all of my resistors and replace some to get better matching if possible - not sure the best route to take... they were all very very close to begin with)

THANKS ALL FOR YOUR HELP! I have learned through this process for sure! I really do appreciate it and hope that someone else can benefit from the dialog in this thread.

 

[Samuel Groner]

Is it possible these are too big?

No, 220 nF is fine. Even 1 uF would probably work, although more might be critical.

Can I simply add a resistor after the right channel pot to balance things out?

A better way would be to parallel the 3k3 resistor which goes to ground next to the pot (darn, how can one draw a schematic without designators..?) with another resistor in the left channel. This makes the left channel gain go up independent of the pot position. A shunt resistor after the pot alters the gain in dependence of the pot position, i.e. there's just one pot position where it leads to the correct gain change.

Samuel

 

[riggler]

Samuel,

Are you saying to add a resistor in series somewhere in the left channel signal path, and then run a set of leads from either side of it to either side of the 3.3k on the right? If so, how do you determine the value?
(or even a good starting point) Thanks!

 

[Samuel Groner]

Are you saying to add a resistor in series somewhere in the left channel signal path, and then run a set of leads from either side of it to either side of the 3k3 on the right?

No! You don't place it somewhere and then run leads across the whole board. You place it right next (that's less than a quater of an inch for me) to the 3k3 and use the part leads to connect it to the 3k3. ;D

If so, how do you determine the value?

First of all you need to know the level difference; then it is simply a matter of looking up the opamp configuration and the parallel resistor math. The value won't be anything less than 33k (ten times 3k3) unless levels are way off (and then there's something else wrong).

Samuel