I'd better show a picture to illustrate what this means:
- the red and blue lines are the supply rail voltages, as seen by the output stage (one half of the amp output shown);
- the green line is the output signal, close to full-scale output in the example;
- the area in the middle, outlined in yellow is the 'Class A' operating region.
When the output signal is within that area, operation is in pure Class A. Outside of it - the opposite half of the output stage is switched off. Within - both sides are in a conducting state.
The beauty of this setup is that it widens the achievable 'Class A' region given the same size heatsinks while improving overall efficiency, as the 'voltage overhead' is minimal compared to the same circuit with fixed supply rail voltages. Crossover distortion is 'non-existant', at least below what is measurable given today's equipment, and certainly way way below the perceptual threshold.
Furthermore, as the supply voltage is low during idling and at low output levels, bias can be driven up, thereby increasing the Class A region while keeping heat output (losses) manageable. A distinct advantage of our technology (the DX Drive) is the way bias is set and maintained, practically eliminating its dependancy on temperature and/or supply voltage variations.
This power supply setup also increases the usable part of the 'Safe Operating Area' of the output devices and their reliability/life expectancy. It allows 'insane' power output levels as the output devices can be used up to and even beyond the breakdown voltage limits in a fixed supply scenario - the non-conducting half sees much less voltage during the off-state.
***
Bob Carver is a very talented engineer whom I deeply respect. Many of his ideas have inspired my work over the years. The way the 'tracking power supply' principle is implemented in my circuits is radically different from his design, however, for engineering reasons.
