A PSU with 4 bridge

[scottjoplin]

Where does the signal ground connect in your circuit?

 

[zjjwwa]

Where does the signal ground connect in your circuit?

If that is a question to me, I would say: to the clean section of respective ground.
Basically, all typical, diligent rules of ground sequencing still apply.
The only difference being, that we have a Left_Ground, and an "independent" Right_Ground.
Keeping in mind that there are "dirty grounds" and "clean grounds", that need to be carefully sequenced, so as not to induce design-originating hum,
I would envision the following ground returns sequencing (starting dirty, nearest PSU, and ending clean, far off):

Connected to the Left_Ground of the PSU:
1). Left_Zobel_Return; then:
2). Left_Rail_Bypass_Returns; then:
3). Left_Speaker_Return; then:
4). Left_nfb_and_front_end_return.

Connected to the Right_Ground of the PSU:
1). Right_Zobel_Return; then:
2). Right_Rail_Bypass_Returns; then:
3). Right_Speaker_Return; then:
4). Right_nfb_and_front_end_return.

Aforementioned sequencing can be executed as per preference of the designer, being either in the form of a left_star_ground and right_star_ground, or a left_bus_ground and right_bus_ground.

The main idea is that the Lefts and the Rights do not touch each other. Anywhere, with the only exception being the two series connected hum breaker resistors within my proposed PSU.

 

[Mark Whitney]

The front_end_return left and right are joined together by the equipment connected to inputs of the amp (See post #32).
Consider what happens when the positive and negative currents to the amp are not equal, like in the case of a class AB amp.
Variations in the diodes and resistors is also a possible cause of problems.
What does the charging current look like?

 

[DF96]

As he clearly cannot see the snags in this design despite prompting from us, and presumably is puzzled that nobody else does it this way, I suggest he builds it and discovers for himself what will happen.

 

[zjjwwa]

As he clearly cannot see the snags in this design despite prompting from us, and presumably is puzzled that nobody else does it this way, I suggest he builds it and discovers for himself what will happen.

OK. I sense that I am in gross error here somewhere, reading your opinions. Need to think about this, sleep over it. Maybe make some simulation, if I manage to model it somehow ...
Thanks for alerting me about the pitfalls in it.

 

[DF96]

Your design will work fine if both polarities of both channels draw exactly the same current as each other at all times. Note I say same current, not same average current. Any current imbalance forces PSU currents through wherever else you join the left and right signal grounds.