Do powercords make a difference? A Test.

It certainly is not just about the noise. Power cables have an effect on other aspects of the sound as well. Not to belittle your experiment... it's very interesting. But, you are only comparing one criteria by which we judge cables. I'd gladly take a little noise along with a great sounding cable, rather than a quiet cable with lesser sound.

Research and read everything you can find about how Caelin Gabriel of Shunyata Research has approached this subject. Here is an interview done by 6Moons: http://www.6moons.com/industryfeatures/caelin/caelin.html
It seems the ability to analyze the effect of PC design demands a bit more than what you suggest.

I have a dedicated circuit in my board for the audio stuff. I use an MIT plug receptacle. So it's pretty well controlled. I mean, the fact is the cords sounded different. And the PS Audio was quieter. So I use that as the primary now.

I'm not sure noise is the main point of a power amp cord. ITs more about ability to deliver current efficiently with a power amp.

For source level gear, digital ones in particular, noise levels are a much bigger issue that a power cord might help address.

Pangea ac-14, not ac-9, is the model designed for digital/line level/noise applications, whereas ac9 is more about the current delivery for a power amp.

Its better to consider how cord designs are different to address different issues at hand (noise levels in source gear and pre-amps versus current delivery for poweramps) rather than assume all power cords are designed to solve the same problem in the interest of better sound quality.

Also, in the case of noise levels, noise issues are likely not merely just more harmless random noise that might be detected separately from the musical signal, but rather more structured, non random, forms of noise at certain frequencies, etc. that might be more difficult to measure and distinguish yet have a more detrimental effect on the sound by introducing specific distortions. A good pair of ears probably have a better chance of detecting a difference than any measurements, unless the noise/distortion is quite significant.

Of course its very hard to communicate and bank on what one hears in any particular case, so the results may never be reliably reproduced by others.

I think the key is to collect enough data points from many distinct test cases by many users over time to be able to discern anything conclusive about these things. There is statistical power in numbers that a single case or even a larer set of uniquely biased cases can never achieve.

The interview with Caelin is really great! Highly recommended.