Solid State Power Amp Reviews

Stuart Taylor was working in a sound studio in Toronto in 1972 when he purchased one of the very first Bryston amplifiers. Over the years he kept up a close relationship with the Russell brothers, who ran Bryston, and in 1988 came to work for them. Taylor's first assignment was to redesign Bryston's four-channel Home Theater amplifier, the 8B. Though the job began as a routine fine-tuning of a product before it was to be reintroduced to the market, it quickly led to major redesigns of all Bryston amplifiers.

Chris Russell, Vice President of Engineering, asked Taylor to design an improved layout for the 8B that would allow its four transformers to operate within a single chassis within Bryston's performance limits of noise and distortion. Taylor first reorganized the 8B's parts layout, placing the transformers vertically against the inner front panel. Then he methodically tracked down all sources of noise by reworking all the grounds, finding tiny millivolt ground loops and eliminating them. Using new distortion-measuring devices, Taylor looked at the signature of the noise and gradually "massaged it out" by eliminating point-to-point wiring, reorganizing component placement, and changing the gain structure of the overall amplifier.

Taylor also found that he could cut high-frequency distortion in half by running a shielded wires rather than pcb traces. By changing or eliminating most internal wiring, he was able to reduce crosstalk, hum, and noise. Later, he reduced chassis depth in the redesigned 4B-ST and 7B-ST to shorten the path lengths between input/output connectors and main amplifier circuit boards.

A key change involved reworking the input circuitry to derive some of the amplifier's gain from Bryston's proprietary input buffer. It was evident to Taylor that the NRB series' balanced inputs sounded better than the unbalanced...and the balanced input was the only input using Bryston's proprietary buffer. Applying it to all inputs on the ST Series, Taylor was able to reduce noise by lowering the source impedance of the signal inside the amplifier. This also reduced distortion!

Stuart Taylor and Chris Russell explained why this design approach improves their amplifiers' performance: Imagine a hypothetical amplifier with an open-loop gain of 1000, open-loop distortion of 1%, and a required closed-loop gain of 25. Because the ratio of closed-loop to open-loop gain (and distortion) is 1:40, the expected closed-loop distortion will be 0.025%. If one then cascades two of these amplifiers, the closed-loop gain need now be only 5; the total gain is 5 x 5 = 25, same as before. The overall distortion, however, is now 0.005% in each amp, adding geometrically to about 0.007%, or several times lower than before.

Although these calculations are hypothetical, and actual results are influenced by considerations of stability, bandwidth, and gain, the ST Series shows a noisefloor reduction of 6-8dB compared with the NRB series, and a decrease in distortion by a factor of three. In contrast to designs that use brute-force feedback, the ST designs are said to show no tendency to increased distortion at higher frequencies.

Taylor notes that the balanced input buffer in the NRB-series amplifiers required 100% feedback to achieve unity gain. The ST Series, which provides a gain of 2x in the input buffer, requires less feedback in the buffer and less feedback in the amplifier modules that follow. This addresses concerns from both RD and myself that the new ST Series' gain structure might complicate the signal path over previous designs. Taylor's answer was simple and persuasive: The ST Series adds no additional feedback or complications. It just reduces noise and distortion.

The success of Taylor's ST redesign was rewarded when the ST Series logo was applied to the faceplates and rear panels of the new 1994 Bryston products. In addition, each ST Series internal circuit board has the letters "S" and "T" highlighted on the Bryston name label, just as an artist might sign a painting.—Larry Greenhill