Hey PJ,
I realize that the variable subjective perceptions of powercords is quite frustrating. And frankly, I'm unsure what makes a good powercord. But I do know what can make a bad pc. Sadly, the result of my own dumbass attempts at building a better mousetrap.
I can only suggest the following -
1. Get some Volex 17604/5 -
http://www.carlton-bates.com/cb/invsrch/invframe.asp?srchfield=stkmfgpart&vendor=ALL&codevend=ALL&sval=17604&submit1=Go2. Compare them by DBT or whatever protocol you desire to a generic 16ga cord on a class AB poweramp, something with a changing current draw.
This should remove anecdotal reports and that dreaded Hawthorne effect.
3. Call Unicable (unicable.com) and cadge [you've got to be able to talk the talk] a Unicable 6450. This is their 14ga shielded cord, nominally equivalent to the Volex.
4. Repeat #2 above, between the Volex and Unicable. (yes, when I did this, I actually replaced the moulded plugs on both with Marincos to eliminate that variable)
5. Out of frustration, perform a circumcision on both cables to discern WTF would account for the vastly differing, subjective perceptions. Also measure all parameters between all legs of the circuit.
I certainly understand why a trained engineer would dismiss this 'nonsense' of 'mo betta' powercords. But a simple empirical experiment will show you that it really does matter, in a repeatable, statistically significant way.
Based on this minimal empirical experimentation, it will also become clear why those cords that use multiple twisted pair or star-quad to increase accross the line capacitance and/or EMI rejection have to give up shielding