Unless something is horrifically wrong you're unlikely to be able to hear any difference. But you'll be able to measure the differences in the form of higher distortion, especially towards the end of the audio spectrum. You're also likely to see higher jitter in the digital circuitry with suboptimal grounding.
Tom
Tom
This was exactly my learning a few years back. The speaker return must go back to the amp module and from there a single module 0V cable back to the PSU.
The feedback resistor must be laid over the ‘quiet’ ground to the top of the HBR (ie the incoming signal ground). All of the tracks associated with the feedback network must be laid over the quiet ground. This minimises the loop area.
It goes without saying the front end has to be compact, again to minimise loop area.
For the grounds on the amp module, I bring the speaker return in on the opposite side of the 0V connection to the PSU that the module dirty ground and quiet ground come to. This keeps the speaker signal currents out of the module grounds and prevents common impedance coupling.
You'll get even better performance with a ground plane where the quiet stuff is to one side and the noisy stuff is on the other side. The point is to separate the current flows while maintaining a low ground impedance.
You can, of course, separate the current with separate traces, but that increases the ground impedance, which hurts performance. So there's a tradeoff there.
Tom
You can, of course, separate the current with separate traces, but that increases the ground impedance, which hurts performance. So there's a tradeoff there.
Tom
