My Road to Damascus

by Paul Kaplan

To be honest, three years ago in 2007, I never thought power cords made much of a difference. After all, all that was required is that they deliver adequate power when called upon to do so, nothing more, nothing less. I used Volex 17604 cords, and industrial 14ga. Shielded cable that provided an excellent upgrade to the typical OEM cord supplied with component and were quite inexpensive, at less than $15. Actually, in comparison to a number of $100+ cords, I preferred the Volex's.

I had built some power cords, learning some important lessons by building what turned out to be some very poorly performing power cords. Most importantly, I learned some important lessons about technical characteristics that should be minimized, as well as those that should be maximized. But my real interest was in power conditioning, specifically for medical use such as laparoscopic surgery, filmless radiology and nuclear medicine, etc… which has its own specific requirements. My own audio system provided the test bed to evaluate my designs, as well as a business opportunity to bootstrap capital for the exceedingly expensive necessity of getting ETL approval for medical devices. More than a year of effort and many an hour with signal generator and spectrum analyzer and I was ready to test my conditioner—consisting of my conditioner fed by a Volex plug and cord with the conditioners output feeding the other half of the cord terminated with it IEC plug. I had tested it against some very good bespoke conditioner both fed by and feeding similarly bespoke power cords.

So I took the cord over to my friends. He had (and still does) have the best system I’ve ever heard. And well he should, as at retail it represents over $250,000. And we proceeded to compare my conditioner/cord against his bespoke cord on his Reimyo CD player. I’ll not name his cord, but I will say it graced a lovely lady’s back in ads in both Absolute Sound and Stereophile, and caused substantial outrage among the more puritanical. All I can say was that his cord stomped my efforts (although the bass from mine was better). Nor was it relevant that his cord retailed for $6,000+. I learned something critical; no matter how good your power conditioning, power cords matter.

And so it was back to the lab, literally. I started looking at what was going with these darn cables, in addition to my previous power conditioning measurements. Those constant parameters, both direct and parasitic, resistance, capacitance, inductance, impedance at differing frequencies weren't problematic to measure and deal with. But my computer interfaced digital oscilloscope and signal generator, functioning as my spectrum analyzer, limited further exploration of dynamic behavior, resonances, to 10 megahertz.

So I call the local engineering school, and inquire about renting lab time to use their spectrum and bridge analyzers, and/or hiring a grad student for some research on power filters (I wasn’t about to mention cables). We discussed in some detail what I’d previously done, and what I was seeking to explore, and was told that for insurance they couldn’t rent lab time, but I could use their lab if I’d teach a lab course for senior undergrad EE majors.

Teaching the course was a hoot. (I really do think the students benefitted. They learned that one could actually make a multi-pole low pass filter using passive analog devices, rather than their previous digital-only understanding. And at least they did learn to solder.)

While my access to sophisticated measurement tools confirmed much of my “lower resolution,” lower frequency investigations, it also confirmed that measurements don’t account for much of a cables performance. This isn’t to say that ultimately metrics won't be found that correlate more accurately with performance, or that one can’t make a horrible cable based on known measurements.

But to make a really excellent cable, one must combine technical knowledge with tedious, empirical evaluation. You’ve got to build, listen, make another with a single specific change, listen, evaluate, decide what characteristics may account for a given measureable and/or subjective change, and build yet another to hopefully verify. Repeat until done.

Just listen.