Great reference, Jim (Jea48), as I'd expect from Mr. Whitlock. And all of the sections you referenced make perfect sense to me. Thanks!
Amazing that it took until the last few years for an explanation and measured verification to finally be presented for "what drives 99% of all ground loops."
I note that rate of change of current is indicated on page 31 as being a significant variable, which reinforces the significance of what I said earlier about not extrapolating results obtained with one amplifier class (i.e. A or AB or D) to other classes.
With respect to the comparison chart on page 35, I wonder what the results would be for the situation that probably exists in many older homes where the metal conduit IS the safety ground connection, for example where older two-prong outlets have been replaced over the years with three-prong outlets, or where a three to two prong adapter is being used, with its ground connected to the cover plate screw. I would guess the results of his test in those situations would be fairly good. I assume, though, that modern code doesn't permit installation of conduit containing only two conductors. Am I correct about that?
I note that the BrickWall I had suggested to Margot is based on the series mode non-MOV technology that is recommended in the paper. And I seem to recall that there may have been some kind of relation between BrickWall and the SurgeX brand he specifically mentions.
Regarding the EP-2050 Dodgealum suggested, having looked at the datasheet and other info on it provided at their website it certainly seems to be a good product, that would be a worthwhile investment in many cases. But as a whole-house protector and noise reducer it appears that it won't do anything to reduce the effects of noise generated within the system, certainly in an application that utilizes a single dedicated line. Also, I note that it lists at $730, plus the cost of professional installation. Also, I would note that the absence of audible noise does not necessarily signify that noise is not a problem. Digital noise has very substantial frequency content far above the audible frequency range, and that high frequency content can have effects within the audible spectrum when introduced into analog circuitry as a result of intermodulation with signal and other effects.
The bottom line would seem to be that given the success many audiophiles achieve with both multiple dedicated line and single dedicated line approaches, and given the system dependency and technical unpredictability of the tradeoff between the possibilities of ground loop issues and inter-component noise coupling, it seems clear to me that there is no one size fits all answer, and in general there is no approach that can be determined to be optimal apart from experimentally.
Best regards,
-- Al