#dantimax

Cool, glad I could help! As it turned out, my power supply was not very clean (I wound up replacing it with a 5V wall wart), which may have made the problem worse than the average case. I think whether of not you should keep the AGND and PGND tied together is going to depend on the larger context of the circuit – like, what is the bigger grounding scheme, and what kind of power supply you are running. But cutting that tie is certainly one avenue to explore, if you are having noise problems.Original post about Dantimax RelVol3 ground issues here.

The low level noise in the monitor controller persisted, even after the improvements I made in the grounding scheme. When I had the box open this week, I took a closer look at the power supply, and tried to find the sources of the noise. As I mentioned, this power supply was a 5V supply salvaged from an old MIDI rack unit... it wasn't designed to be super clean, as you would want in an audio circuit. And, even though the audio path here is passive, the Dantimax relay-controlled attenuator is still running the 5V lines right next to the audio lines. And, the audio shares a common ground (now a star ground) with the power ground. I read up on some more grounding info, in these papers by Jensen and Rane. I found one potential ground loop in the power supply – the earth of the 3-prong plug, which I had run to my star/chassis ground point, was also running to the ground of the DC supply (ie on the secondary side of the power transformer) via a hidden PCB trace I had missed. I eliminated that  "extra" trace, and played around with just leaving the DC supply floating, or running a separate wire to the star ground. These different schemes had different levels of noise, some much higher than before. Leaving the DC supply floating, for example, was not an improvement in output noise! Since I'm not an expert in these things, I decided to try replacing the whole supply with a 5V wall wart. This made a significant improvement in noise, so out goes the internal supply! I like to know exactly what is the source of a problem, but in this case, after several hours of troubleshooting, I'll take the improvement with the wall wart supply, even if I'm not completely sure whether the fault lay in the quality of the MIDI power supply, or my grounding scheme...