John Ulrick

It was with regret that I heard John Ulrick had passed away on May 20, 2015 due to complications from cancer. With Arnie Nudell and Cary Christie, John was one of the founders of loudspeaker manufacturer Infinity, a company that, with Audio Research, Magnepan, Mark Levinson, and Threshold, epitomized the nascent High End that emerged in the early 1970s. After leaving Infinity, John Ulrick started Spectron, to manufacture class-D amplifiers.

Just by chance, I met with the John Ulrick, in Los Angeles in late 1987, when he was doing some design consultancy on a switch-mode power amplifier to be used with the Sumo Samson subwoofer. As I had my Walkman Pro with me, I took the opportunity to tape some background from John about the birth of Infinity and about switching/pulse-width-modulated/class-D amplifiers—boy, can this man talk about switching amplifiers! The natural kickoff question was, How did Infinity get going?

John Ulrick: In the mid-'60s, Arnie Nudell and I were working in the laser laboratory at Litton Industries, developing a system to be used for ranging for bombing systems, where you point a 4.5-million-watt laser at a target with a mirror. I was the electronic engineer in the group, Arnie was the physicist, and there was another guy who was doing systems engineering. Arnie and the other guy were off working on making the laser switch work, which they were successful at—at that time it was very state-of-the-art to make these things really high-power—and my part was to design the electronics: the amplifiers that amplified the pulses, the power supplies, and the servo systems.

In an aircraft, you can't use linear [conventional] amplifiers because they're very inefficient; we were doing some of the very first servos and switching amplifiers. I was testing a system, just running the pointing mirror at 20Hz, and Arnie walks in. "God, that's great!" he says. "Couldn't we use that switching amp to drive a woofer?" "Well, of course," I said.

Both of us had an interest in hi-fi. I had designed the amplifiers I had at home, and Arnie had been playing around with horns and things, so we both went off the library and got a copy of Acoustical Engineering, by Harry Olson. We started working on it, [not just to] make something for ourselves, but I had an interest in starting a business...We went to Gene Czerwinski [of Cerwin Vega] to get a woofer. We had him wind a second voice-coil on it, so we actually got the woofer and an amplifier working. It was very crude, but we got it working.

About the same time, I met two guys who worked at Atomics International who were going to build electrostatic panels. So we bought the panels from them, designed a crossover with a common bass unit, and put together the Servo Statik system. And that's how we actually started Infinity. During the next year [while we were looking for people to invest in the company], we developed the Servo Statik 1, and sold a few systems that were so crude you wouldn't believe it. But we did sell them. And we developed the 2000, which was a 12" woofer with electrostatic tweeters; then we developed another speaker, the 1000. We already had a product line in our minds. So we got a building out in Chatsworth, and I quit Litton in August of '68. I have to tell you, the first day I was there I just sat at my desk—I didn't know what the hell to do. I was an engineer, so I started designing more circuits! Finally, Arnie went nuts, and a week later he also quit Litton.

We started going out and calling on dealers, and from that point built the business up. That's how it got going, using the Servo Statik technology to build the company.

John Atkinson: Was the woofer amplifier in the Servo Statik 1 a switching design, then?

Ulrick: No, not at all. We used linear amplifiers, about 150 watts per channel, in the servo bass. We started out to do it with switching amplifiers, but it was just too much new technology to pump in all at once.

We showed the first full-range Infinity switching amplifiers, as far as I can remember, in 1974 at the CES. They were very rudimentary amplifiers—I think they were like 125 watts per channel—but they worked. Bill Johnson [of Audio Research, footnote 1] had his first vacuum-tube amps out, and they sounded different from the Phase Linear 400s and the other amplifiers that were out then. We didn't say our amplifiers were better; we just said, "Look, transistor amps sound one way, vacuum-tube amps sound another, and our amps sound different from either of those other amps, because they're working on a different principle."

Atkinson: A PWM amplifier can be a very efficient amplifier.

Ulrick: Yes, it is. It's the right way to build an amplifier. If you had perfect components, it would be a perfect amplifier. If the FET switches were perfect, without any on-resistance, and the integrating inductors didn't have any loss in them, it would be 100% efficient...

Atkinson: All the power from the AC line goes into the load.

Ulrick: Exactly. It's the only way you could do 100% efficiency. Ultimately, in the future, it's the way amplifiers will be built; but it's very, very difficult to do. And they're not cheap.

What you can do [to make them less expensive] is to rectify the 60Hz AC line. With a couple of diodes, a couple of capacitors, you get ±160V DC rails. You make a switch by having two field-effect transistors, one connected to the +160 and one connected to the –160, and [modulate the width of the FETs' output switching pulses with the audio signal]. We're switching these days at 500kHz! The junction between those two FETs goes to a very small inductor, like 100 microhenries, that performs the integrating and smoothing [of the output pulses to give an audio-bandwidth analog output signal]. But driving the FET switches is harder than you imagine. You have to switch the FETs on with 15 amp pulses. And off. And that's not easy to do. FETs [with their high gate capacitance] are much harder to drive than bipolar transistors.

As well as efficiency, PWM amplifiers have other advantages. In a practical linear amplifier, the delay time through the amplifier is pushing at least a half a microsecond—500 nanoseconds—to about 2000ns. Okay. In a PWM amp, the delay time though the loop is about 200ns! Now, as a practical matter, what that means is that the negative feedback loop is responsive to changes in the music several times faster than a linear amp typically is. Now, most people get around that by not having much loop feedback in the first place. But loop feedback, if you could use it, is very desirable; it's just that you really can't use it in linear amplifiers. So most designers who design amplifiers to sound good try to use as little negative feedback as they can.

Now, another thing is that if you watch music on an oscilloscope, power is not what's important in an amplifier. What's important is how much voltage you can put into a speaker. And [a PWM amplifier] with 160V rails can put 125–135V into a speaker on real high transients—the snap of a drum, for example. So [a switching amplifier] can put a pretty clean, very-high-power transient into a speaker without a lot of screwy things going on.

Atkinson: You still need the power in the sense that the amplifier has to be able to supply the instantaneous current to maintain that voltage.

Ulrick: Yes, and that's another thing that's important. A switching amplifier is stupid about where it's delivering current. If the output signal is down at –100V, following a bass note, for example, it can happen right at that moment that the drummer strikes the snare drum, asking for a positive pulse. In those circumstances, a linear amplifier would probably current-limit to protect its output devices, or just run out of gas. A switching amplifier doesn't mind going from –100V all the way up to +100V, to deliver that kind of a pulse. So it sounds different in transient response.

Atkinson: Why is it you feel that PWM amps have really never been commercially successful?

Ulrick: Technically, it's real hard to do...nobody else is doing it (footnote 2). I just talked to [a patent attorney] and asked him if anybody else was working on audio switching amplifiers. He said no, but some guys over at Infinity were building them about five or six years ago. I said, "That's me!"

Atkinson: Sony marketed a PWM amplifier, the TA-N88, in 1978.

Ulrick: Yeah, they used a FET switch. It was rather crude; they took a triangle into a comparator. It was clever, but it was low power—100 watts—and didn't sound very good...

The Infinity switching amplifiers had very mixed reviews—some people loved the sound while others hated it. The amplifiers were very controversial. They were also not the most reliable in the world. It wasn't horrible, but they had problems—they were real sticky, and we'd have some amplifiers that would go out and come back and play ping-pong with us.

We sold quite a few of the Infinity switching amplifiers, then we came up with a preamplifier that we sold with it. Our success was that we really built things at Infinity the way we wanted to build things for ourselves.

Atkinson: When did you finish the relationship with Infinity?

Ulrick: 1979. We all sold our stock in '74 to Eastern Air Devices, who then became Electro/Audio Dynamics...The guys at EAD were building motors—it was more of an industrial kind of a company—and wanted to get into the audio business. It turned out they really weren't that good at it...but they helped us credit-wise, and they helped us with a management structure. In those early days, we had a lot of problems with the company. See, Arnie and I couldn't decide who was going to be president, so we flipped a coin. I won the coin toss, so the first year I was president, then he was president. Every year we alternated. Well, that's no way to run a company; finally, EAD said, you guys figure out who's going to be president. I decided that I'd rather work on my amps, so Arnie became the permanent president. At the end of my contract, I went out to do other things.



Footnote 1: Bill Johnson richly deserves his success...he builds beautiful equipment. He sticks with the old-world vacuum tubes and does a magnificent job with them. But those days are probably numbered.—John Ulrick

Footnote 2: This may have been true in 1987, but there have been a few in-car PWM amplifiers sent to market in the early 1990s, including one from Infinity.—John Atkinson

COMMENTS
bornie's picture

For John, selling Spectron amps, back in the day. He and his wife, Toni were lovely people! They are both terribly missed! RIP.

Allen Fant's picture

A wonderful contributor (infinity) to this hobby of ours!
R.I.P.

sodapopman's picture

John and Toni were amazing people. RIP. Such beautiful souls and contributors to our love for audio. John was brilliant and Toni supported his love and passion for design. I wish that Manh had kept Spectron Audio alive...

X