An Interview with David Reich

by Richard Hardesty

Richard Hardesty, Audio Perfectionist Journal: Dave, when we first met many years ago you were extolling the virtues of the Classé amplifier that you had carried into my store. Since then you’ve been associated with some of the most respected companies in audio—Classé, McCormack and Theta—and you’ve been responsible for some of the best solid-state amplifiers I’ve heard.

You could have been designing Mars excursion vehicles or less-interesting electrical devices but you chose to create fine audio components instead. What got you started in this industry?

David Reich, Theta Digital: First and foremost, my love of music. Secondly, the technology to reproduce it. I’ll never forget the first time I heard stereo. As a young teenager, I was not aware that my family’s two-speaker record player was capable of it, so my older brother and I had continually bought mono LPs. When he brought home Jimi Hendrix’s “Are You Experienced” we realized the jacket said it was a stereo LP. (Richard: you may have to explain the software nomenclature to younger readers.) The store didn’t have it in mono, so we begrudgingly kept it. Upon listening to it, with my head between the two speakers, I experienced “sound placed in space,” and was absolutely amazed. I was astounded at the possibility of instruments and voices being purposely placed across a soundstage. I had to know more about how this was done!

I was further indoctrinated by building venerable Dyna kits and obtaining my first real system. A few years later, while I was visiting the same brother who was attending S.U.N.Y. in Buffalo, NY, we came upon Transcendental Audio, a true high-end store that professed the likes of Decca cartridges on manual turntables, tube electronics, and electrostatic loudspeakers. It was an epiphany. Another enriching experience was attending the Recording Institute at the Eastman School of Music, in Rochester, NY. There I had the opportunity to learn electronic theory from the likes of Walt Jung (author of the classic IC Op Amp Cookbook), and rub shoulders with top recording engineers while we learned the ropes on state-of-the-art recording consoles.

Back in Montreal, I found work at a high-end store that sold Dayton-Wright electronics and electrostatic speakers. They were rather, uh, shall we say, temperamental, and I was sent to the Dayton-Wright factory in Toronto for a week to learn how to service them. And that, Richard, was the turning point for me. When I walked through the door of that plant, it may as well have been a time-portal. I was transported into a world of music, technology, and fascinating people that would propel me into the industry we called high-end audio. So I really have to thank Michael Wright for giving me the opportunity, though I didn’t start amplifier research until about four years later. Dayton-Wright had been sold and within two years, regretfully, closed (some may say thankfully—but I beg to differ).

It was actually while working for a musical instrument amp company that I had some time to begin amplifier research. Why amplifiers, you might ask? Well, there really wasn’t much out there to power a highly modified stacked pair of Dayton-Wrights, with an impedance of about one-quarter of an ohm. And I figured if I could accomplish that, it would sound that much better on conventional speakers. A few years later, was it any wonder why the one-ohm Apogee Scintillas and the early Classé amps were a match made in heaven? The Scintillas were a piece of cake compared to what I had designed my amps to drive. My research led to a very small start-up company, which was actually the forerunner to Classé. I sold about sixty 25-watt stereo, pure Class A amplifiers to local dealers within Quebec, and some in Toronto. That was around 1979-80. After that, I refined the design and called it DR-2. Classé was formed in 1981, and the DR-2 was its first product.

Amplifiers really can make or break the sound of an audio system. Why are they so important to the sound we hear?

Component audio is a funny thing. I don’t think anyone would consider buying a car without a motor. Yet that’s exactly what we do when buying loudspeakers. We leave the job of supplying the horsepower to the amplifier. It’s a difficult task we take for granted now, given that, theoretically, the loudspeaker is part of the amplifier circuit. It’s taken decades for us to learn how to minimize the changing effects speakers will have on our amplifiers, and with credit to most designers, we’ve pretty much accomplished that.

Now don’t go thinking I therefore favor internally-powered speakers. Powering the woofers is fine, but I haven’t heard much else I care for. I believe it’s best left to those who specialize in it, and it appears most speaker manufacturers agree. Now with the onset of (truly) digital amplifiers, the varying loads are once again presenting problems, but at least we can rely on DSPs to manage part of that problem.

So, to answer your question, the loudspeaker may have the ability to move rapidly, at many frequencies, with wonderful accuracy, but it won’t emit a ppp note without the amplifier that precedes it. With that come the speed, stops and starts, harmonic structure, nuance, soundstage cues, and anomalies that the amplifier possesses. Start with a bad amplifier and good speakers, and I guarantee you will have bad sound. Start with a good amplifier and good speakers, and you may have good sound.

Today’s amplifiers utilize vacuum tubes, bipolar or field-effect transistors, and various combinations of these devices. Can you give us your impression of the technical advantages and disadvantages of these components and the sonic differences between them?

The simplest, and most boring, explanation of these devices is well known: tubes and FETs are voltage devices; bipolars are current devices. But that’s not my answer because, let’s face it, you can’t have one without the other. Every voltage has current along with it, and vice versa. So what’s really going on? I prefer to look at what these devices do in amplifiers.

Tubes are the greatest seducers in audio. They don’t generate too much ugly stuff, and when they do add something it’s sooo pretty! Who hasn’t been carried away by the ethereal effects of a luscious tube preamp or power amp? Alas, for me, it would always end up as a too much of a good thing affair. I couldn’t ignore the extras, and particularly in power amps, the often-flaccid bottom was not acceptable to me. Looking at the elements of glowing tubes always made me think of mechanical reverb units. I couldn’t help but wonder what all the vibrations were adding to the sound! That, combined with the device variability, and its declining performance over time, provided too many parameters I couldn’t control, so I opted for solid-state devices, and challenged myself to make music with them.

Initially I worked exclusively with bipolars. I liked their finite operating parameters, and found that if you manipulated them in heavy enough Class A, you could eliminate their nasty characteristics. Once I got rid of the bright, edgy hardness of a bipolar, a wealth of harmonic structure and holographic information revealed itself. The sound of the early Classé designs was always compared to tubes, yet they were done entirely with bipolars.

Back then, MOSFETs were just coming along, and numerous designers jumped over to them. They were technically forgiving, so an average design with them yielded reasonably musical results. But in an output stage, for me, they lacked current.

MOSFETs are much better now, but I still use bipolars in the output stage of amplifiers. This stage, after all, is purely for amplification of current. All the voltage amplification has already been done in prior stages, so why not let the bipolars do what they do best! Besides, I am a total current and power hog when it comes to output stages. The possibility of blown output devices leaves a very bad taste in my mouth, so I design these stages very conservatively, with gobs of output current capability. And I think they sound better that way. You can hear an output stage that is working too hard, as opposed to one that’s coasting. I’ll take the latter. Of course, you do not want more devices than necessary, so there are trade-offs that must be carefully evaluated.

Theta builds solid-state amplifiers using FETs and bipolar devices. Why do you use one type of transistor in the input section and another type for drivers/gain and yet another type for the output stage?

The previous answer partially addresses this, particularly for the output stage. We haven’t mentioned JFETs. They don’t get much attention because they are almost exclusively small signal devices; little guys that can’t handle too much voltage. But do they sing! Small devices tend to sound superior, anyway. I recall a Nelson Pass design: I believe it was a 10-watt amplifier, supposedly built with 100 small signal devices. A sweeter sound was never heard! Their low voltage amplification is incredibly linear, making them ideal for input stages—and preamps, for that matter. And there are tricks, such as cascoding, that allow you to implement them in higher voltage circuits. The MOSFETs we touched on above have earned a solid position in the second stage and driver stage of my amplifiers. Their input impedance, like JFETs, is high, so they don’t load down preceding stages, and they set up the signal for the output stage beautifully. JFETs, MOSFETs, bipolars. It’s a popular configuration in many good amplifiers.

In your opinion, which sections of an amplifier have the greatest effect on the sound we hear?

(continued)

Page Number: 1 | 2