An Interview with David Reich — continued

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

By the sheer fact that any given section of an amplifier can mess up its sound, I am forced to give equal strength to all sections of the product. The power supply, input, driver, and output stages are so dependent upon each other that, while individually addressed, they operate as a whole system. The effects of each section must be evaluated with those that precede and follow it. The only exception is the output stage, the performance of which depends more heavily on the power supply.

The amplifier power supply is especially important. What makes a good one? If cost didn’t have to be considered how could you make a better power supply?

I presume you are referring to the supply that powers the output stage, and not the low current supply that takes care of the previous stages (that, too, is important). Conventional, unregulated, linear supplies, that most commonly power output stages, are made up of three basic elements: magnetics, rectification, and filtering—or the transformer, rectifier(s), and capacitors. The variables of a power transformer for an audio product are so complex that, in terms of massaging those variables to achieve great sonic performance, it easily competes with the amplifier itself. And be sure that we are talking about just a power transformer here, not an audio transformer, which may be even harder to design. If it sounds like I’m paying my respects to transformer craftsmen, I am. A great transformer’s design goes way beyond its electrical voltage and current requirements.

Rectifiers are important also. In Theta’s Dreadnaughts, because each channel module has its own capacitor bank, it is also able to have its own bridge rectifier. Power and speed are important here, as are some other less familiar parameters. Now capacitors are lots of fun, and we’ve seen trends go from giant ones as big as your neighborhood transformer sitting atop a pole, to multitudes of tiny ones, and everything in between. And don’t get me started on bypassing. But designers have come full circle on this, I believe, and, in general, multiples of moderately sized caps work well. One of them per output device is even better, if you have the space and budget for it. Credit Steve McCormack for that, who was the first to do it.

As for making a better power supply, that’s what we did in the Citadel. Power supplies took a giant leap forward when we added “choke-input filters,” as discussed by Charlie Hansen in the previous issue [#10] of Audio Perfectionist Journal. The effect is nothing less than synergistic, resulting in power supply lines so clean that the output stage thinks it’s a preamp. But these chokes have to be big enough, or dynamics may be restricted. In Citadel they are almost as large as the power transformer itself.
Beyond input chokes, that leaves only the fully regulated, high current supply, which effectively requires its own output stage as large as the audio stage. Back to the Levinson ML-2, or perhaps some Krells for that, I think. Results have been mixed, some feeling it left the bass overly taut, and the sound restricted. In the other direction, there have been cost-effective designs with minimalist supplies, building a “loose” supply with just enough oomph to give a lively and dynamic sound, if not scrutinized too closely.

Theta’s amplifiers have no global feedback but some of your previous designs have used modest amounts of negative feedback. What’s your current position on this controversial subject?

Right now, I have to say I love the way our amps sound without feedback. Feedback is, after all, a case of trying to correct for something that’s already happened. There’s a freedom from hash and grain without it that just seems to be saying, “Hey, I’m running free, and I don’t have to double back and cover my tracks.” I’m not negating good sounding amplifiers that use some feedback. Hey, I reserve the right to do so in a future design if I see fit. I will even go out on a limb here and say, contrary to popular opinion, I think there are some excellent sounding op amps, and I have employed them successfully in past designs. Perhaps keeping the overall loop within chip size helps. But that said, zero feedback is where I am right now. And Theta’s amps are very happy that way.

Theta makes several amplifier models that share similar circuitry. Can you tell us about how they differ and how these differences affect the sound?

It might be easier to first discuss their similarities. They all have the following in common:

1) True, fully balanced circuitry. What a pleasure this is to work with. Mirrored circuitry that can’t be beat for any common mode rejection of noise, distortion, EMI and RFI. It’s just so quiet, and helps give you a beautifully “black” background from which pure signals can spring. The way we do it, even our single-ended input benefits. With today’s digital sources, it helps achieve greater dynamic range, too.

2) Three-stage circuitry topology, with zero overall feedback. JFETs, MOSFETs, and bipolars are used for the input, driver, and output stages, respectively.

3) Linear, high current power supplies for output stages. Fully regulated, discreet MOSFET circuits for all preceding stages.

I think that covers the main common features. There are some component differences between models. The single-channel module in Enterprise is very similar to that in Dreadnaught, but the loosening of price restraints gave me the opportunity to include items from my wish list that could not go in the multichannel amplifier. The modules in the Dreadnaught II (225W) and Enterprise (300W) both use toroidal transformers. Power is shared by two of them in the Dreadnaught and one in the Enterprise. So they are quite similar, and you could say the Enterprise is a more powerful Dreadnaught. But the latter also benefits from being a monoblock, which brings it closer to our flagship Citadel.

The stereo module for the Dreadnaught (2 x 100W) and each channel of the Intrepid (100W) are similar. While the low frequency limit of all our amplifiers is below 1Hz, they are not DC coupled, and capacitor quality bears sonic significance here. Because of size constraints, these two models use different, more compact capacitors than the other models. The Dreadnaught stereo module must share the same power supply as its higher powered brother, so the power is set by limiting the regulated supply lines on the actual card. The Intrepid has its own unique EI core transformer.

Citadel (400W) stands above all our other models, sonically and literally. I have discussed its power supply above. Many other components have been sonically selected for their specific application. These include custom-made film capacitors, Vishay precision resistors, solid copper buss bars, Litz/Teflon Cardas wires, and a custom-made output connector. They all combine to give Citadel the status it has been awarded by reviewers and consumers alike. Of course this makes for a successful product, but moreover, as a designer, it provides invaluable feedback as to how well my applied design techniques are tied to the way I hear things.

Theta Digital built its reputation on unique digital-to-analog converters and surround sound processors. Today the company makes some of the finest analog amplifiers available. What do you see in the future?

I see both (digital and analog) technologies continuing and merging, which comes from the fact that each technology now has a solid footing. They no longer are mortal enemies that prefer to exist at the exclusionary expense of the other. I find it fascinating that major semiconductor companies that offer both are extolling the virtues of analog as well as digital right now, because digital has been an easy sell for so long, while the virtues of analog were nearly forgotten. Closer to home, Theta is doing research on digital amplifiers now, future products that will be welcomed for their small size, great efficiency, and good sonics.

It is my task, and Theta's mission, to bring a sense of immersion realism to these new technologies, integrating them with the rest of Theta's highly evolved designs. The standard we have is high enough that a lot of the work is in getting the new technology up to the threshold of the old, so that we're actually making improvements instead of slipping backwards by going with the newest way of doing things. Theta's constant offering of an upgrade path has to stay an upgrade path, after all, not a sideways or a downward one! This keeps me honest—and it certainly keeps me busy.

Thanks for an informative discussion of amplifiers, Dave. I’m sure our readers will learn a lot, as I have.

Page Number: 1 | 2