Why Power Cables Affect Sound

@atmasphere

 I’ve seen a power cord make a difference of nearly 30% of output power out of a power amplifier. I could also see that that was caused by a voltage drop across the power cord.

and

You can measure differences in output power, output impedance and distortion on many power amps just by changing the power cord- and many of these differences are simply caused by voltage drop.

and finally,

There is more to it than voltage drop though. It also has to do with bandwidth of the power cord.......

 My questions are not whether a power cord can affect the sound of an amplifier (or other gear) since I have heard the differences, but rather,
1. What characteristics of a power cord would affect voltage (i.e., cause a voltage drop) - is that primarily a function of resistance and wire gauge, or something else, and
2. What characteristics of a power cord would affect bandwidth?  

Thank you @jea48 for your post with comments from Ralph, Nelson, and OLC.

I had a couple of questions about Ralph’s post yesterday but did not see an answer and then the thread seemed to spiral down a rabbit hole. Ralph’s response from your post covers my questions and Nelson’s comments and linked information address the amplifier end of things. BTW, OLC is posting again at Audio Asylum since he apparently no longer has conflicts of interest between his day job and posting on a public forum.

My questions from yesterday;

1. What characteristics of a power cord would affect voltage (i.e., cause a voltage drop) - is that primarily a function of resistance and wire gauge, or something else, and
2. What characteristics of a power cord would affect bandwidth?

Based on Ralph’s comments, it seems both the voltage drop and bandwidth issues are affected by the amount of current available to the power supply. This implies, bigger (heavier gauge) wire is better for PCs, and particularly when supplying power to large amplifiers, no? Both issues seem to be related to wire gauge size.

Shielding is another issue that comes up, with some believing PCs sound better when unshielded, which seems to contrast with the findings of OLC.

The connector thing makes sense that they should be substantial enough not to heat up upon use.

Therefore, it seems we are back to wire gauge, geometry, shielding, and connectors as being the factors affecting PC performance (notice I left out marketing, pixie dust, magic beads, and directionality (sorry Geoff)). Common sense would imply the PC should be at least as large as the household wire feeding the circuit (mine is 10 awg for 20A circuits and my DIY cords that feed two large mono amps are 7 awg). The conductors are twisted to reduce inductance, the cords are shielded to resist interference, and the connectors are robust and use actual copper for the connection. Don’t forget the quality of the wall outlets everything is plugged into. Somebody mentioned a direct connection rather than using an IEC and I agree. I had a direct connection on an amp Steve McCormack upgraded for me, but later I added an IEC so I could try different PCs.

The possibility of making DIY PCs from NM Cable (i.e., Romex) is interesting, particularly as Ralph believes that cable works well for power distribution. I looked into why Romex is not approved for use outside of the wall and the primary reason the code doesn’t allow it is that the covering is not considered robust enough to resist damage that might affect the wires - but it is suitable for the set it and forget it in-wall use. Another concern is that the large solid core wire in Romex may not be as resistant to multiple bending events. There is also some difference of opinion as to whether Romex can be used inside of conduit, with some seeming to think this is ok while others caution there could be a heat issue (inside of the conduit) that might affect the wire. Aside from those safety considerations, it might be fun to make up a couple of 10 awg Romex PCs to try on my big power amps - just to hear how they sound (but of course I would unplug them from the wall when not in use).

@elizabeth 

mitch2.. One part of you query is about what else can be in the powercord that is not being measured" (totally parphrasing, if in error, yell at me)

No yelling necessary Elizabeth, but that was not actually the question.  My questions were in response to Ralph's post from 1-11 at 1:55pm where he pointed out;

I’ve seen a power cord make a difference of nearly 30% of output power out of a power amplifier. I could also see that that was caused by a voltage drop across the power cord.

and

You can measure differences in output power, output impedance and distortion on many power amps just by changing the power cord- and many of these differences are simply caused by voltage drop.

and

There is more to it than voltage drop though. It also has to do with bandwidth of the power cord as it has to be able to provide current at high frequencies because rectifiers in power supplies often commutate (switch on and off) for very short periods of time, and if the current gets limited during that current pulse the power supply can’t charge completely.

I am curious then whether simply the wire gauge of the PC is the primary factor affecting the incidence and severity of voltage drop, which then affects power output and bandwidth.

I am not implying that is the only factor affecting power cord performance but, I am skeptical there is much more than wire gauge, geometry, wire type, shielding, and connectors that would make a significant difference.   For wire type and geometery, I suspect OFE/OFHC C101 copper and a twisted pair should be just fine, although a star-quad geometry would further lower inductance.  I would be interested in knowing what other parameters (than those listed) would make a meaningful difference in a power cord.