Back in the mid-1900s, amplifier classes were defined as Class-A, Class-B, Class-D and several others (Class-C, -G, -S, etc).
Class-A, Class-AB, and Class-D are the primary topologies found in modern amplifiers, with the other classes largely variations of the three fundamental topologies.
Let’s briefly look at the basics of each design.
Figure 1 (below) shows an amplifier output stage. A resistor (RL) is the load. However, real loads are loudspeakers, and present a reactance - not a pure resistance - to the amplifier’s output.
A reactive load can sink or source current, independent of instantaneous voltage. An amplifier’s ability to control its output into a reactive load contributes significantly to its sonic character. Other factors, such as linearity and transient response, also play roles in sound quality.
Also note that the active devices in Figure 1 are shown as bipolar transistors. This is not always the case, although the majority of amplifiers are still outfitted with such devices due to their maturity and low cost. MOSFETs and vacuum tubes are examples of alternate output power devices.
All classes of amplifiers can be divided into two broad categories: linear and switching amplifiers. Class-A and Class-AB designs, along with their variants, are linear, while Class-D designs are switching.
Figure 1: Basic amplifier output stage. (click to enlarge)
In fact, the formal definition of a Class-D amplifier requires only that the output devices be used as switches. Many fictitious class designations have been created by switching amplifier manufacturers, but they are variations of Class-D amplifiers. Other non-class names have been given to amplifier topologies, such as BCA, which is a Class-D topology, and BASH, which is a Class-AB topology.
Figure 1 also depicts a half-bridge topology, with bipolar power supplies. Other variations exist, such as bridging, which drives both sides of the load using two output stages for double the effective output voltage swing.
Multiple output stages connected in parallel (to share the output current) are also used in some cases to allow inexpensive transistors in high-power designs. Whatever the “trick,” these special adaptations still fall within the three main classes of audio amplifiers, and we will refer to them simply as “variations.”
If linear amplifiers were 100 percent efficient, there would be no reason for this discussion. It’s generally assumed that there is a compromise between efficiency and performance (qualitative or quantitative), and that one must trade performance for efficiency based on the application, but in light of recent technology advances, this is a generalization. The missing factor here is cost, which can be loosely correlated to efficiency and performance.