At the 2006 Audio Electronics Society conference in San Francisco, Bruno Putzeys came face-to-face with the odd cult of celebrity that surrounds a select few audio amplifier designers.
”A guy came up to me talking about me without realizing I was me,” he laughs. ”Then he saw my name tag, and he said, ’Oh, awesome meeting you.’ He shook my hand, and then he hurried off!”
In one of the few disciplines of electrical engineering known to make otherwise rational people rage or rave, Putzeys is a star of growing magnitude. He is widely regarded as the leading designer of a type of audio amplifier known as class-D. Also known as switching amplifiers, these ultraefficient models are already dominant in multichannel sound systems, portable media players, cellphones, car stereos, and computers. Lately they’ve made significant inroads into the ostentatious world of high-end audio, where a component can cost tens of thousands of dollars. Their success there is due, mostly, to Putzeys.
In 2001, while working at Philips Applied Technologies in Leuven, Belgium, Putzeys designed a compact, versatile class-D amplifier module that he called UcD, for ”Universal class-D.” Over the past few years, dozens of amplifier models, with prices ranging from US $500 to $8500, have been built around Putzeys’s modules, which are now manufactured under license by Hypex Electronics of Groningen, Netherlands. The amps have received mostly ecstatic reviews.
For Putzeys, the success of the UcD boards has conferred a measure of professional freedom that’s pretty rare for a 34-year-old EE. In May 2005, he followed his modules to Hypex, where he is now the chief tech guru. He lives and works in a two-story building in a picturesque suburb of Leuven. The first floor is his laboratory, where he works most days, and it’s a hands-on audiophile’s dream. There are a couple of big, brightly lit lab benches, strewn with toroidal transformers, power-supply boards, and multimeters. Signal generators, high-end oscilloscopes, and spectrum analyzers abound. In one corner is a tall glass case housing a collection of unusual vacuum tubes. There’s also an assortment of high-end audio components, including some prototypes and other one-of-a-kind pieces that some audiophiles would trade a limb to get.
Above the laboratory is his living space: a bedroom, a small kitchen, and a spacious living room with his main audio setup dominating one wall. Along another wall in the living room are shelves with an eclectic mix of a couple of hundred CDs: Johnny Cash, Art of Noise, Jacques Brel, and Baaba Maal, the Senegalese singer who is a particular favorite of Putzeys’s. There’s no TV anywhere on the premises.
Putzeys credits his father, Raymond, for many things: his introduction to audio, his choice of college, his choice of career, even his occupational tactics. When Bruno was around 10 years old, his father, rekindling a long-standing interest in audio, began building amplifiers from schematics. Then, when Bruno was 16, a friend of his father’s visited the Putzeys home in Herent, Belgium, with an amplifier built around two pairs of EL84 vacuum tubes in push-pull configuration. The young Putzeys was bowled over. ”I thought, this thing can give me a sense of the music that the other amps can’t.” He taught himself how to design electronic circuits with tubes, a really odd activity for a teenager in the late 1980s, when personal computers were ascendant.
”Nowadays, all you basically learn in engineering school is how to use these chips,” he says. ”You don’t really get to learn the basics. Even though now I don’t really consider tubes a serious alternative for high-quality audio reproduction anymore, learning them was an important stimulus for me.”
After high school, he enrolled at the National Radio and Film Technical Institute in Brussels, from which his father had graduated in 1963 (the school has since been merged into the De Nayer Instituut). While there, he became intrigued by class-D audio, and did his thesis on it. After he graduated in 1995 with a bachelor’s degree in electrical engineering, Philips, the Dutch electronics giant, which had sponsored his thesis work, offered him a job. He recalls that initial job as ”tedious:” he designed conventional (class AB) amplifier products around modules manufactured by Philips partner Sanyo. And he itched to get back into class-D.
Within a few years, he’d succeeded in ”annoying all the managers” until one of them agreed to let him take charge of the testing of a class-D audio IC being designed elsewhere in the company for a television set. He didn’t think much of the chip. ”If you give me one month,” he told his bosses, ”I will make something out of discrete parts that will be much better.”
He had no idea how he was going to deliver on his promise in such a short period. But ”if I had asked for more time they certainly would have said no,” he points out. ”If I didn’t make that promise, it would have been back to the Sanyo modules, and that was no choice at all.”
In the back of his mind was a somewhat similar stunt his father had pulled off, many years before. Working as an engineer at CP Clare Corp., now owned by IXYS Corp., his father declared that he could shrink magnetic relays small enough to put them in a dual in-line package (DIP) that would fit nicely on a printed-circuit board or a DIP socket. In their bright-blue packages, these parts have been ubiquitous for years. But in 1971, his father’s co-workers thought it would be impossible to miniaturize the magnetic components that much. The elder Putzeys showed them otherwise, and achieved a significant and lucrative breakthrough.
”It’s a habit I copied from him,” the younger Putzeys says. ”First promise something without really knowing how to make it work. You should have a good idea you can make it work, a gut feeling, without knowing how you’re going to do it.”
It’s not all he learned from his father, who was subsequently promoted through a series of sales and management jobs—and spent much of the rest of his career pining for the workbench. ”When I was 10 years old, he told me to promise him I’d never accept any promotion that would take me out of the lab,” Putzeys recalls.
But getting back to his own brash promise: he made good on it, of course. In three weeks he built a 25-watt class-D amp with better performance figures than those of the IC-based amplifier the four-man Philips team had labored over for two years.
His supervisors were impressed enough to let him keep working on class-D amplifiers. His early designs wound up in a Philips plasma TV, and in products sold by other companies under agreements with Philips, including an active speaker system from Microsoft and a home theater system from Marantz. In 2001, a Philips executive, George Aerts, secured some research money so that Putzeys could further refine his class-D amp design. The goal was an amp module that would be easy to manufacture and suitable for a wide variety of audio applications, from mass-market to high-end. It had to be compact and have the same output impedance and power-supply requirements as a conventional amp. And it had to be cheaper and sound better. ”The idea was that there would be no excuse for not using it,” Putzeys recalls.
This time, it took him eight months. He went through four generations of circuit boards without listening to any of them. Instead, he connected each board to an audio analyzer and then rejected it because the results on the analyzer weren’t what he wanted.
The fifth iteration, though, looked good. Just before Christmas 2001, he brought a pair of the amps home and connected them to the speakers in his living room. He put on a CD of Spanish classical music and selected a song by the 18th-century composer Juan Francés de Iribarren, ”Viendo que Jil, Hizo Rayo.” He settled back in a chair and listened. It took him just a few seconds to reach a conclusion: ”Straight in the bull’s-eye.”
At a visitor’s request, he re-creates the event, with the very same CD and stereo components. The music begins to flow from the speakers, and Putzeys’s eyes seem to unfocus, like he’s lost in thought. There’s a little grin on his face. The sound really is remarkable—extremely transparent, neutral, and precise. And yet there’s a lovely warmth and force in the female vocals, and not the slightest trace of harshness.
”It was a defining moment,” Putzeys says of that experience. ”Until then, I had subscribed to a lot of audio folklore, namely that measurements don’t matter.” He had basically moved on from his adolescent infatuation with vacuum-tube audio. ”Tube designers often build and see if they like it,” he continues. ”You never get to question what it is you want to hear. In my case, what I want to hear is music and nothing else.”
In an engineer’s universe, Philips would have swiftly embraced Putzeys’s UcD module, incorporating it into countless products. In the actual universe, the module basically fell in the cracks at the giant company. But in April 2003, an enterprising young entrepreneur named Jan-Peter van Amerongen visited Putzeys at Philips. Some years before, van Amerongen had started Hypex Electronics to supply amplifiers and other gear to makers of active speakers and to recording studios. He had heard great things about the UcD.
Ironically, the one thing he didn’t want to hear, at least initially, was music amplified by the module itself. ”The only thing he wanted to see was the output signal on an oscilloscope,” Putzeys recalls. ”He looked at it, and in about one minute he said, ’Okay, I want to buy a license.’ He had seen so many dreadful outputs, full of RF hash. He could tell from the signal whether it was well designed.”
Not long after, Putzeys left Philips for Hypex, where he has pretty much free rein to explore the boundaries of class-D. Just ”for fun,” he recently designed an audio amplifier with 0.0003 percent total harmonic distortion, at full power, amplifying a 20-kilohertz signal. That figure is more than 1000 times better than some very good solid-state amps. In fact, it’s an improvement that no human ear can detect, as Putzeys acknowledges.
But that figure is also about 30 000 times better than that of some tube amps—a difference that’s not beyond the ability of human ears to detect. Putzeys rejects the idea, which most tube-amp enthusiasts take for granted, that it’s okay—desirable, even—for an amplifier to ”color” the music it is reproducing.
”Audio is not supposed to be art,” he insists. ”Making music is art. Getting it from the CD to the listener should not be art.”
His views were forged in part during his one and only foray into the world of music production, in 1998. He was visiting some friends in Finland who were in a folk music group. They happened to be recording a CD while he was there, with a producer who handled mostly hard rock and roll. The Finnish folkies fired the producer, because ”he was drunk all the time and he knew nothing about folk music.” So Putzeys took over. ”I didn’t know anything about production, but I knew I couldn’t do worse than that guy.”
In the end he decided he wasn’t cut out to be a music producer. But he also had an insight into the debate, endless among audiophiles, about coloration. ”If it sounds good,” he says, ”the right place to put it in is in a recording or mastering studio. There, an engineer can add the coloration that is appropriate to the music. You get an artistically informed choice about when the coloration is used, and when it isn’t. It’s better than using it indiscriminately all the time”—as would be the case if your amplifier added the coloration.
Avoidance of coloration seems to be a general principle with him. As he brews a cup of green tea in his kitchenette, a visitor tells him about the peach- and grapefruit- and mint-flavored green teas on offer in U.S. supermarkets. He winces.
The existence of flavored teas notwithstanding, life is pretty good. Besides his gig at Hypex, he is also one-fourth of a Netherlands-based start-up company called Grimm Audio, in Utrecht. It specializes in very high-end components for recording studios. Between Hypex and Grimm, he has a full schedule of design work, speaking engagements at conferences around the world, and sales visits to recording studios in North America and Europe.
Outside of work, he finds time to maintain a long-distance relationship—very long distance—with a woman who lives in Rwanda. He lectures at his alma mater and at conferences around the world. And when he can, he also listens to his own stereo setup, usually at night. The system is beautiful in the way that only a DIY audiophile’s can be; no rich poseur would ever wind up with something like this.
The disc player is a 13-year-old unit he built himself. He’s just using it as a transport; its digital output goes to a digital-to-analog converter he designed and built five years ago. This D-to-A converter uses all discrete components (no chips) in the analog section, including the actual conversion. His preamplifier is based on a small headphone amplifier he built for Philips when he was there. His power amplifiers, small, elegant, and shiny, are a pair of monoblock units (a separate unit for the left and the right channel); they were prototypes for the MP150 amplifier from the Netherlands audio company Kharma International. Inside each box is, of course, a UcD module. His radio tuner is a 1970s model from the German company Wega (which was bought by Sony in 1975 and later dissolved). His father found the tuner in the trash some years ago and Bruno fixed it up.
He designed and built his speakers, too. They’re a meter and a half tall; each one has 4 woofers, two midrange units, and a tweeter. Some of the speaker elements are installed backward, to cancel distortion. Next to each speaker is circuit with a bank of enormous capacitors; these are crossover circuits, which didn’t quite fit in the speakers.
The system is marvelous, if ungainly. And it nicely exemplifies Putzeys’s own credo about hi-fi components.
”Stereo replay never actually reproduces a musical event,” he says. ”The only thing you can hope for is a credible illusion. But it can be a very nice illusion.”
To Probe Further
A shorter version of this article is available at http://www.spectrum.ieee.org/feb08/5924.