Kevin Hayes: Knowledge from the Glass Age
My father trained as a BS/EE in 1951, and that put a lot of old electronic texts and old test gear in my hands during my formative years. Playing around with them built up knowledge that came in useful later. In the late 1980s, I was working on my master's, and later my doctorate, at Duke University, and simultaneously trying to build up my stereo system—which was frustrating, because I couldn't find anything I liked. I was shopping in the $3000 component range, which was good money at the time. Finally, I decided to test out some ideas I had and build my own power amplifier.
Phillips: So you built your own amp because you were disgusted with your choices as a consumer?
Hayes: Disgusted is too strong a word. Not pleased is closer. It was an emotional reaction. As a child, I found music was exciting and big to listen to. I want to recreate that experience in reproduced sound. Nothing I'd auditioned did that for me, so I began to take some of the information I had absorbed and apply it to circuit design. Over a period of three years, I evolved the prototypes for what became the PA45 and the PA90.
When I had constructed what I thought was a pretty good-sounding unit, I invited a friend over to listen, a hi-fi dealer, and watched him as I played it. His mouth fell open and he exclaimed, "You built that!?"
At that point, I decided to go to the Summer CES and see what the industry looked like. I loved music, I loved designing, and I loved vacuum tubes—so it had always been kind of a dream to work in this field.
Phillips: So what was that first amplifier? Triode, pentode?
Hayes: It was a utility player—it was originally designed to function either in pentode ultralinear or triode connections. It ran a pair of KT77s per channel with a 12AX7 driver and a 12AU7 input tube. It was voiced to have about 8dB feedback. The trickiest part for any experimenter to come up with is the output transformer. The prototype originally breathed around a Dynaco A470 transformer, which rapidly became the limiting factor. But it was a very good starting place.
That first amp had a very seductive character; the midrange could just draw you in, and it sounded very dimensional. The final production amplifiers were more extended and more detailed, but weren't quite as magic in the mids—or at least that's how I remember it.
It took another six or seven months to reach the final commercial embodiment, and that was a very frustrating time. However, it set the tone for what we do at VAC, which is to spend a prolonged period of time listening to every possible seemingly minor change that can be rung on the completed circuit.
I once spent two months listening to tube brands—I listened to approximately 290 different permutations, out of which 280 just did not work sonically. That drove home a very important lesson about listening and checking component selections—everything doesn't sound the same, even if the engineers tell you that it should.
Late last year, I heard a Renaissance 70/70 that had come off the production line and it sounded bad. So I pulled a second one off the line, it sounded bad. And a third one—they all sounded bad in the same way. At that point we stopped production on the 70 and checked our listening notes for the last one that sounded normal (so we could account for any made between those sessions). We did not assemble or ship any Ren 70s for eight weeks, while we traced the problem: our respooler had substituted a wire for the one we had originally tested, and the difference—between two brands of wire that were mil-spec'd to the same tolerances—was quite audible! You can't take anything for granted. It was painful holding off customers for those two months, but it was the right thing to do.
Phillips: Tell us about the PA80/80.
Hayes: It has elements of both the PA90 and the Renaissance series in its heritage. The input circuitry is similar in its topology to that used in the Renaissance amps, which is basically a refined Williamson-type topology. That's a design that almost always, to my ears at least, sounds good. The output stage is a conventional partial triode, or ultralinear, connection, which is more akin to the original PA90. The PA80/80 relies upon a printed circuit board, where the Renaissance amps utilize point to point hand-wiring—there's an effect on sound quality, of course, but there are major cost benefits as well. In terms of component quality, they're actually comparable units. We use the same coupling caps and transformer construction techniques as in the Renaissance 140.
We originally envisioned the PA80 as a $2000 amplifier, but I wasn't happy until I achieved a certain level of sound quality—which demanded parts and construction techniques that, unfortunately, make it a $2800 amplifier. I have a hard time compromising on sound.
Phillips: The Renaissance series products allow the end-user to dial in varying amounts of loop negative feedback, but you've opted for a set level on the PA80/80.