If you have been writing about audio for a long time, there’s a delicate trade-off, between spinning a tale out of every product, and starting your next review with “this is the 374th amplifier I have ever tested.” However, in the case of the Ayre AX-5 twenty amplifier, this dichotomy is resolved because there is a genuine tale to be told in its development, and its performance is good enough that you don’t need to be scrabbling round for filler.
The big story here is the VGT (variable gain transconductance) approach used in the preamplifier stage, and first seen in the company’s KX-R preamplifier. To understand precisely why this is such a big story, you need to step outside the received audio wisdom for a moment and think in first principles. If you do this, the way a line level signal passes from source to power amplifier actually becomes a little screwy: attenuating a signal to taste before passing the signal to a gain circuit to boost it to amplifier-friendly loudness is – at best – the tail wagging the dog. As the amplifier has a fixed level of noise, the maximum signal-to-noise ratio is only achieved at full volume. The other analogue option is to use a passive preamplifier. This has the advantage of not boosting the gain (so the signal-to-noise ratio of the input signal remains the same), but often causes as many issues as it resolves in terms of cable length having an influence over sound quality. Also, the volume level of the system is ultimately governed by the output voltage and impedance of the source component, which is taking matters too far in the other direction.
Basically, Ayre’s VGT circuit uses a pair of double-pole silver-contact rotary switches, connected to an array of precision, hand-selected resistors. This is not a stepped attenuator, however: these resistors alter the transconductance of two pairs of complementary JFETs in the power amp’s input gain stage. In essence, the volume control is influencing the power amplifier, leaving any signals passing through the preamplifier entirely unattenuated, meaning no changes whatsoever to the signal-to-noise ratio during the line-level stages (where comparatively small changes make large differences). This makes the signal path extremely simple, but dictates the kind of switchgear used in that signal path (very low resistance and linear switching is needed), with FET switches instead of relays in the input selection because that’s the best sounding line up Ayre has found for the circuit. Most amps are not designed to these lengths, and the end result sounds good, but doesn’t come cheap from a bill-of-materials standing. By its inherent characteristics, VGT cannot work in anything other than a zero-feedback environment and, as is the case with all Ayre amps, it’s run as a balanced circuit – there are two single-ended inputs, but think of them in the same way you might think of salad at a fast-food joint.
At the other end of the circuit, Ayre goes right back to the beginnings of solid-state technology, using a form of gateable bridge network of bipolar transistors. Ayre’s Charles Hansen is something of a student of amplifier design and utilised Richard Baker’s ‘diamond circuit’, so called because the original notes written by Baker in the 1950s show the circuit drawn in a diamond-shaped configuration, even before there were the components to use in that circuit. The diamond buffer is not uncommon as the output buffer stage is several high-performance IC op-amps, and even a few discrete preamplifier designs, but is extremely rare in a power amplifier output stage. The reason this circuit is rarely found in power amplifiers is that the driver stage must operate at the identical bias current as the output stage itself.