Power cord? Why?

When performing this test, are you driving speakers or dummy loads ? If you are using speakers and performing the tests simultaneously, my guess is that the speakers would have to be phenomenally well matched in order to do this. So well matched that i don't think it could be achieved without great time and expense.

As far as using dummy loads go, they will not "modulate" the power supplies like the signal being fed into a speaker would. Nor would they provide the reactance necessarily to simulate the actual conditions that the output stage has to deal with during normal use.

I would like to know what you intend to use as a test signal, the amplitude of signal that you intend to drive the amplifier up to and if the amps will be connected to the same identical unfiltered power source ?

I bring all of this up as these specific amps are very load sensitive ( vary their frequency response / output characteristics as impedance is changed ) and their distortion characteristics vary drastically as frequency rises. As such, using some type of wideband or sweeping tone that is varied in amplitude with a reactive dynamic load attached should REALLY produce some interesting results. I would LOVE to have some computer assisted test equipment that was capable of performing such a test. This is the kind of testing that could REALLY advance the SOTA of audio gear. That is, if one knew how to interpret the data provided and manipulate the circuitry to correct for the flaws detected.

If you are talking about doing a "null test" where equal but opposite signals are fed into each amp and calibrated to produce a "zero beat", you would have to do a baseline run with "stock" cords first and then with one of the cords replaced or modified. I'm sure that BOTH tests ( stock and modified cords ) would provide very sufficient "nulling" ( at least -60 dB ) for the scientist in you. The question is, if there are measurable differences in "nulling", would the differences be audible on a finely tuned system with a good "listener" at the helm ???

If you really want to do this correctly, you'll have to perform a baseline on the amplifier prior to cord modification and then afterwards. That is, unless the two amps measure and respond to signals in identical manner. They would need to do this regardless of load, frequency or amplitude of signal. Relying on two "similarly constructed" mass produced amplifiers of antique vintage to provide identical levels of lab grade reproduction would be asking a bit much in my book. Sean
>

My question would be that after performing a test like described above, and having the possible result of no measurable difference, but they still sound different, what do you do then? Do you conclude that the testing methods are flawed, or that your hearing perception is flawed? This is what this subject really comes down to, isn't it?

My position is that if you want to spend the evening running test signals through your power cords into an analyzer, pick the one that measures best. If you want to spend the evening listening to music, pick the one that sounds the best. I really think that this simple approach is the best one.

Tom hits the nail right on the head. Of course measurements have validity, but nothing can take the place of listening as a yardstick of sound quality.

sean and others...

1. About the A/D. Of course I meant 100K samples/second.

2. The amp load will be Magnepan MG1.6 speakers, 4ohms ,and with impedance characteristics like a dummy load. (Of course I will listen, but that is not part of my test plan).

3. The input signal will be real music, of various spectral characteristics. I find that real music works best for most things like setting speaker levels, and subwoofer gain and phase.

4. I thought that my initial setup measurements using two super cords provided the reference for comparison. Of course I could do it with the original cords as well.

5. Some of you guys didn't understand the objective of my test. I am not (at this time) evaluating audio quality. I am testing to see if the electrical signal applied to the speaker changes in any way when the power cord is replaced. If it does, then listening tests would be in order. (And I will apoligize to all those I have offended). Viridian..I agree with you about listening being the best test of sonic quality, but I just don't want to fool myself, which is easy to do. If there is no change, listening tests might be fun but can prove nothing).
My speakers only "know" about what comes in to them as an electrical signal from the power amp. Maybe your speakers have psychic powers and can discern the type of power cord in use. That's too far out for me.

6. By the way, when I said that the original power cords were "crummy" I really meant it. They don't even have a 3prong (grounded) plug...but maybe that was by design. I will need to try the super cord with and without ground.

7. The Kenwood LO7Ms may be old, but they were top rated by T.A.S and Stereophile. They are comparable to the best available today. (Besides they are what I have on hand. If you have Krells, or the like, feel free to use my procedure).

8. One procedural detail...I will make the electrical hookup such that I can switch cords without shutting down power, or interrupting the signal measurement process.

9. Hey this could be a money-maker! If I find a power cord effect I can sell the data to power cord manufacturers. If not, they can pay me to keep it under wraps.

10. Finally...(for now)...how about some suggestions about the design of the super power cord(s).

TWL wrote: "I'm currently using a large capacity lead acid automotive battery with 650 cold cranking amps, located about 24 inches from my amplifier. I have the terminals directly connected to the power input barrier strip on the amp, with 12ga OFC copper stranded wire."

24 inches of ANY wire represents a large inductive impedance. You need what is referred to as "decoupling caps", which essentially decouple the power source from the load by providing local energy storage. This means low ESR (equivalent series resistance) capacitors located as close as possible to the load. And a combination of HF caps (.01 UFd to .1 UFd) and some swamping caps (1 UFd to 100 Ufd) are needed to supply current at a broad frequency of transient currents. It would not hurt to have even a 1000Ufd close-by as well.