Does the amp supply current need to return to the amp?

[rgrayton]

Or otherwise stated, where is the proper place to return the load ground: PS or amp. I started a thread a while back on this, but the holiday got a discussion going and I could not convince my EE father the PS cap supply the electrons, bo the load return is best returned closer to them and not on the amp board. If they need to get bcak to the amp, then returning them to the PS is wrong and viceversa.
-BG

[richie00boy]

As the central ground in any system should be the star point of the PSU cap bank, all grounds (yes load grounds as well) must return there. Returning them to the amp pcb induces an error signal as the amp ground reference is then modulated by the load current.

[jcx]

the most accurate/least distortion requires signal, feedback and speaker load return ground to all be the same "point"

since little current flows in signal and feedback they can be a little separated from the load return but adding trace length is asking for magnetic field induced distortion from the 1/2 wave rectified currents flowing in the +/- power supply traces in a class AB amp

nearly all of the layouts I see here get power distribution wrong, the "symetrical" layout of +/- pwr on opposite sides of the board is the worst from a coupling standpoint

+/- pwr trace sould be as close together as possible and both even closer to the pwr gnd which should also be the feedback&signal ground on the board to minimize trace length and loop areas

as to insisting on sar ground at the ps; why would adding 100 mV ps gnd return ir drop matter to a transformer isolating 110/220 V rms?

the use of multiple separate amp boards connected to a single supply is better delt with by using a much lower impedance ground connection between the boards, not by lengthening the gnd path - unless you've got balanced inputs the multiple boards are going to share a common input signal gnd anyway

ground loop current in your pwr wiring where you can control loop area with twisting or star quad construction seems much better to me than having the gnd loop extend outside the amp to the input signal cable gnds, back to the source pcb/chassis - which is what happens with ps star gnd and no local low Z gnd connection between pwr amp boards

[rgrayton]

jcx,
Thanks for your inputs. This is my first amp and I have read a lot of stuff that puts makes use of a "PS star ground", but never really heard anyone "explain" this type of arrangement like your comments otherwise.
-BG