Yes, the whole issue of grounding is very difficult with high-gain high-power amplifiers and, when you consider the voltage and current gains involved, it is surprising that any audio amplifier can be made stable/have a good frequency response/have low distortion.
My view, based on years of messing with audio amps of all classes and types, and having all sorts of malfunctions, is that there are many facets to the problem and likewise there are many measures that need to be taken to get a fast but stable, optimally damped audio amplifier. You can get a clue by looking inside an audiophile amplifier where there are many extra bits and pieces to sort the performance.
One of the problems is that you can not only get instability in the amplifier overall but also in the sub amplifiers- even a humble emitter follower can oscillate.
In terms of layout, my feeling is that the radio frequency (RF) men have it. Quite simply, with RF circuits, the layout is either perfect or the circuit does not work.
So what approach do the RF men use: star point, ground planes, ground and signal ordering, screening, decoupling, and impedance matching. All of these, in my opinion, apply to audio amplifiers, with the exception impedance matching. In audio amplifiers impedance mismatch is used to give low distortion and good efficiency. For example, the output impedance of most audio power amplifiers is essentially zero ohms, but the speaker load is typically 8 Ohms. By contrast, the voltage amplification stage (VAS), in most high-end amplifiers has a high output impedance to drive the output stage with a constant current to reduce the inherent distortions in the output stage.
But there is a further consideration: ground current minimization. This means aiming at zero ground current, both direct and induced, and with careful design/layout you can get close to this with split supply line amplifiers. With single supply line amps you haven't got a chance, by definition. But you can minimize the effects of ground currents by having low resistance earth paths with hefty ground planes and wires.
You can inadvertently introduce ground current, not only by parasitic coupling but also by decoupling capacitors and, if you are not careful, you can inject noise and hash from the supply lines into the ground line- especially troublesome with with none linear amplifiers like class AB and D. This is a major advantage of class A where the supply line signal still resembles the audio signal but, with class AB, the supply line signal is a witch's brew of half wave rectification and xover artifacts.
Where star points (and signal ordering) are fundamentally important is in the power supply; the power supply has a surprising effect on practically every aspect of audio amplifiers. Also, with power supplies and high current paths in the amplifier, low resistance wiring is important. Similarly, low effective series resistance (ESR) capacitors are beneficial. I think Tony Stewart will advocate this.
Here endeth the sermon.
spec
).
.
