Nelson Pass: Simple Sounds Better Page 2
Norton: Do you have any particular priorities, other than the obvious one of sound quality, when you approach the design of an amplifier?
Pass: Well, I suppose everyone, when they settle down to do creative work, likes to think of themselves as artists. In that regard I see myself as primarily a circuit topologist. That's what interests me—it's the basis of my patents, and it forms the central thrust of my design effort. I like very simple topologies—-the simpler you can make an amplifier, the more likely there is be good correspondence between the sonic performance and what you measure on a bench. The more complex, the less likely that is to occur. Having been through other designers' amplifiers for years and years now, I've come to learn that the amplifiers everyone regards as classics—the ones that stand out over the years—if you look at them you find that they were fairly simple topologies. And they were elegant—they are elegant. So I strive for that.
But at the same time there's quite a bit of room for creativity. The dynamic bias was one such example, I think. The Stasis amplifier certainly stands out as probably the premiere example of a simple, creative topology that's done a very good job and stood the test of time. It's my fundamental belief that if the idea is correct, if the topology makes real sense, the amplifier will tend to function right off the drawing board with very little extra work required. So as a rule, we don't spend a lot of time tweaking the individual values of capacitors and such. You certainly have to go in and make all of those adjustments, but I've seen other manufacturers who start out with a design, then tweak it at length in an exhaustive series of adjustments until they finally get what they want. I'm a lot happier with a design if it rapidly converges on proper performance. If it doesn't, I think there's something fundamentally wrong with the approach.
Norton: I'm curious about the gestation process you go through designing an amplifier. How do you normally operate from, say, a clean sheet of paper? Though this is probably never the case because you're always building on your previous efforts...
Pass: Building on previous designs is an economic requirement. That is to say, there has to be a significant amount of continuity in how a designer's series of designs evolves. It has a lot to do with the success factor of the company. If you have a concept that you've taken to market and sold people on and which they're buying, you can't simply dump it and introduce a radically new idea every year or so. The instability would kill the company economically. So we look at what the market needs out of us. It needs an evolutionary approach to design, and that is what you see.
When I'm alone with a blank piece of paper, almost anything can come up. Quite a few of the ideas are perfectly workable, but they're not appropriate. If it's a truly good idea we'll find a way of slowly introducing it without a lot of disruption. But continuity is very important when you're turning millions of dollars in product. And the dealers don't want a radically different product every year.
Having said that—well, I find that ideas just kind of bubble to the surface. You don't push the river; it flows by itself. Basically the ideas take time to ferment. Quite a few of the things that we've done were thought of by myself or occasionally other people for quite a long time before the time was right to get serious about them. The Ion-Cloud loudspeaker had been bubbling around for years; the Phantom Acoustic Shadow is a product that we were talking about ten years ago. And there are a whole host of as yet unused ideas. Their time will come. There will be a time in which the company is poised to use them; there will be a point at which the market might be more receptive to an idea. It's a lot easier, as you might understand, to develop products which conform to popularly held ideas. High-current, high-powered class-A amplifiers are a perfect example. There's every good reason why they'd be popular! But to go out and pioneer a product is quite a difficult and expensive thing to do. And again, economic considerations keep you from doing quite a few things that are interesting.
Norton: You discussed your interest in circuit topology. It used to be thought—and still is outside the high end—that circuit topology, the actual circuitry itself and its layout, was the all-important parameter in amplifier design. Then along came a few rebels ten or so years ago arguing that parts—not just quality parts per se, but type-specific resistors, capacitors, and whatnot—were at least as important, if not more so. What are your feelings on the tradeoffs involved in the actual topology or circuitry vs the parts used to realize it?
Pass: They're both important. Going beyond the recognition that high-quality parts give high-quality results, all other things being equal, the other thing is that the highest-quality construction and components represent the appropriate embodiment of what I consider to be exquisite ideas. But in terms of capacitors, wires, and such, first off, there's a lot in that. And there are the differences that people hear; it's highly appropriate that these items should be pursued.
There are times when I feel that there is a fair amount of overemphasis in those areas. Example: your capacitors, resistors, and all the elements that people are popularly replacing in, say, a power amplifier or preamp. When you look at the kinds of performance variations and the degradation offered by, say, a reasonable-quality capacitor, and compare that to, say, what a transistor might do, you recognize that the distortions of the transistor are going to be much higher than what you're going to be able to measure with a capacitor. The variations in tolerance are enormously higher. I mean, you're talking ±100% tolerances on transistors.
There is a reason why we go to the trouble of selecting devices as extensively as we do: When a manufacturer makes them in batches, he can't offer a selection process in any sort of economical sense. So there hasn't been enough emphasis on the quality of the semiconductors. For the most part they have an even greater influence in parts quality. People feel that they understand capacitors and wire better. It's true that a semiconductor is a fairly unfathomable black box with three pins, but I think that's where most of the action is. We use as high a quality parts as we consider to be reasonable, and always in the context of what we think people are willing to spend money on.
Norton: You mentioned earlier that you feel that a circuit topology is best when it comes together rapidly. At that point, do some manufacturers tend to tweak more than you would feel to be desirable? At what point do you give up designing for available parts values and decide to bite the bullet and have custom parts made? You use custom transformers and capacitors in some of your products—I'd guess that price has a lot to do with this.
Pass: Well, the transformers are all custom, and the heatsinks and metalwork are fabricated for us to our spec. We generally pick, for power supplies, computer-grade capacitors—I prefer Mallory as a vendor—and we try to pick power-supply capacitors that are reasonably common, because you can have supply problems in those areas. The lead times are very long. Twenty-week lead times do exist there, and if you find yourself with problems either in the quality or availability of a batch of capacitors, 20 weeks is a long time to not ship any product. For all intents and purposes, there's really not any such thing as a custom power-supply capacitor, and there really isn't all that much in the way of custom signal capacitors either. I mean, a couple of outfits are doing some interesting things. I don't put a lot of emphasis there. We do use WonderCaps in one of the preamplifiers, but as a rule I'm very happy with polystyrenes if I can get them, and polypropylenes and polycarbonates.
Norton: Is 20 weeks about typical for the parts you use in your products?
Pass: That's the outside. Resistors can be obtained in a matter of a very few weeks. Twenty weeks is about as far out as it gets.
Norton: Is parts supply one of the reasons for the lag—sometimes over a year—between the showing of a product at a CES and the time the product actually appears in the dealer's soundroom?