[Dan Banquer]

"The only problem with this type of grounding scheme, particularly for manufacturers of DAC's etc.. is emissions.  EMI can be effectively reduced by grounding circuit-boards to the chassis at multiple points.  While I am not a fan of this, sometimes it is necessary in order to pass FCC testing."

This is the classic mistake that most people make. What's really needed here is good localized grounding on the PC board along with a star ground. Dac's are really not high speed units, and therefore do not qualify for that type of grounding.  I should know, I along with others have done this without issues from the agencys.
There is no shortage of information on these issues if one decides to  look at back issues of EE times as well as other engineering sources.

             d.b.

[Dan Banquer]

) "Signal return paths - these are often confused with chassis ground.  I even have modded a Sony 400-disk CDP where some of the signals were forced to flow through the chassis because of a break in the circuit board ground plane.  The break was only connected when the board was screwed to the chassis.  This shows that even large companies make this mistake.  There are instances of changing the chassis from steel to copper and suddenly it sounds better - there were probably signal currents flowing through the chassis, a bozo no-no.
5) Power return paths - power currents should not flow in the chassis either."

Bingo! this is a major problem because major companies do just that and they open themselves up wide for noise and interference issues. Now think of it this way Steve: if you did do what is outlined in the original post and everyone followed it, would the caps at the input for ESD be necessary?
If you follow my logic here then you begin to realize that bad grounding schemes are the root of the problem.

"If isolation is impossible, or compromises sound quality, then I would agree that floating the DC-common can be a useful strategy.  However, one must be aware that potentials can develop between the ground of the connectors and the chassis, depending on the phase and circuits of the AC wiring."

There is no floating of the DC common. If you read the original post carefully I am using the center point earth ground at the return of the supply of the pre amp. This grounds all of the signal to earth at one point so there is no floating DC. If it was floating than yes voltage potentials between the "floating ground" and earth can rise to near 120 volts. However, and please note this carefully, none of this hazardous because the impedance's involved are so high that no shock hazard exists. I'm living proof of that because I would be dead by now.

One final issue. Manufacturers could easily use the system I presented with one simple addition. Have a switch that connects the return of the power supply to the chassis. This should satisfy the emissions and safety issue, and let the customer set the earth ground where needed.

I am convinced that this issue alone is responsible for the bulk of "synergy" issues, noise issues and interference issues and a good deal of the cable/shield controversy.
         d.b.