Quote Originally Posted by skeptic
that the term input here is meant to be the AC power input?

Do you have any measured or calculated range of K factors for different PCs? Doesn't K1 and K2 to a large extent depend on the internal topology of wire and component layout in the amplifier and source, and the physical arrangement of the interconnect wire and the PC external to the components? So doesn't this signal appear in effect as the residual hum and noise?

External decoupling then could better be accomplished by an additional shield on the signal cable than on the PC because unlike the PC, heat dissipation is not a consideration. Artificial means to shield power cords is not only a violation of UL unless designed in by the manufacturer and UL listed such as in the case of MC or BX cable or THWN/THHN run in flexible metal conduit such as greenfield but it can cause cable overheating and the risk of fire or shock.

It is clear that the neutral conductor referred in NEC as "the grounded electrode" is current carrying and is grounded usually at the service entrance to the premesis. The defeating of the safety ground if one is provided in the manufacturer's power cord is not only dangerous and illegal, it is useless because under normal conditions ground current does not flow through that conductor. Similarly installing "ferrite rings" on the ground conductor is therefore equally useless as it is dangerous. Installing ferrite rings on the phase and neutral conductors may marginally increase the overall power input circuit inductance but is probably all but useless as well.
Input is the audio input.

I have no measured K values.

K1 is absolutely dependent on the line cord geometry, external wire dressing, and amp internal power circuitry dressing. K2 is totally dependent on the internal ground loop configuration within the source box and the amp box...how the first loop currents are routed affects the coupling to the input circuitry.

I kept your warnings about pc mods intact, as they are worth repeating..

Shielding will be absolutely worthless, as any shield will NOT eliminate the external flux..copper is useless to remove this coupling. Mu metal will enhance the field strength around the pc..the only possibility would be to use mu metal to divert the flux around the rest of the ground loop, so none gets trapped in the loop.

The only way to reduce the PC field generation is to coaxially run hot and neutral..This will completely eliminate external fields for the pc, but is entirely against all code..running the wires in a conduit won't work because there is no cylindrical symmetry for external field cancellation.

Same with amp internals....even a star ground is worthless for this.

The signal is a result of amp draw, which varies with signal..if you could draw dc off the supply cap, the xfmr draw would produce the signal at the amp input.

Perhaps using a load to pull dc current off both caps, and monitoring the output terminals with no input signal? That would show haversine coupling constants, forcing that error term to be visible at the outputs, without the confounding issue of the input signal..