Crossover Distortion - Distortion Mechanism

Distortion Mechanism

The image shows a typical class-B emitter-follower complementary output stage. Under no signal conditions, the output is exactly mid-way between the supplies (i.e., at 0 V). When this is the case, the base-emitter bias of both the transistors is zero, so they are in the cut-off region where the transistors are not conducting.

Consider a positive going swing: As long as the input is less than the required forward V_BE drop (≈ 0.65 V) of the upper NPN transistor, it will remain off or conduct very little - this is the same as a diode operation as far as the base circuit is concerned, and the output voltage does not follow the input (the lower PNP transistor is still off because its base-emitter diode is being reverse biased by the positive going input). The same applies for the lower transistor but for a negative going input. Thus, between about ±0.65 V of input, the output voltage is not a true replica or amplified version of the input, and we can see that as a "kink" in the output waveform near 0 V (or where one transistor stops conducting and the other starts). This kink is the most pronounced form of crossover distortion, and it becomes more evident and intrusive when the output voltage swing is reduced.

Less pronounced forms of distortion may be observed in this circuit as well. An emitter-follower will have a voltage gain of just under 1. In the circuit shown, the NPN emitter-follower and the PNP emitter-follower will generally have very slightly different voltage gains, leading to slightly different gains above and below ground. Other more subtle forms of crossover distortion, stemming from slight differences between the PNP and NPN devices, exist as well.