Amp design question - Page 6

So the ground driver will look like this?

yes I believe the RC could help with high frequency stability the Buf634 would like to see a low value R at the input at higher frequencies - I would use physically small C - for low inductance - down to 1 nF value if the package can be smaller - 1 uF thru-hole Caps will have self resonant frequency too low in bigger packages with longer leads

even in the R,L multiloop circuits op amp circuit could benefit; op amp outputs look inductive at frequencies below/near their loop gain intercept - beyond their GBW they probably drop off due to parasitic device C

I wouldn't recommend using only the pot wiper for any critical setting - a parallel resistor divider would hold the input to reasonable tolerance if the wiper bounces or fails open

the confusion I have is with the op_gnd not being the same gnd symbol as the rest - it is hard to read layout, interconnection intent from just a schematic

I see what you mean... I'll ditch the pot and just match the resistors closely then.

The Zobel components would be SMD.

I usually use a ground plane for my designs, is there any reason

I should change to a star ground rather than a ground plane?

I know it is hard to see in my schematic, but C1/C2 are supposed to say .1uF (1206 smd)

a common gnd plane for the signal, feedback gnd  "branches" can be very good, it would still be a correct application of the star gnd principle to run the active gnd buffer's output as a trace to the hole for the TRS gnd which connects to the plane

the signal and feedback reference gnds have very low currents so mixing them in a very low resistance plane will usually be fine - the higher current load return (summed, both channels) could still cause some drop and crosstalk in the plane resistance - confining the load return current to the separate trace keeps it from interfering with the signal, feedback gnd reference

the R,L channel amps ps local bypass caps can go to the plane as well, the +,- power traces supplying them should be routed close together (twisted in a point-to-point build) - a full hybrid gnd could include series lossy inductors in the R,L amp chip's runs to the bulk ps caps - unlikely to matter much at audio unless you're building a DAC/amp combo with possible digital noise problems

this advice is for nearly theoretical improvements - differences would be at levels that require careful instrumentation to measure

 a simple gnd plane will be quite good enough for the low currents and small size of most headphone amps

R1 is an adjustable resistor (potentiometer, trimmer...).

I was going to use it to fine tune the center voltage.

I'm going to match the divider resistors instead.

@ JCX

This is the amp I intend to do next...do you see an problems with it?

I included the Source and Output resistors because it's easier to jumper

over them than to add them later...