TABLE OF CONTENTS A full and detailed account of semiconductor amplifier design since its beginnings would be a book in itself - and a most fascinating volume it would be. This is not that book, but I still feel obliged to give a very brief account of how amplifier design has evolved in the last three or four decades.
Valve amplifiers, working in push-pull Class-A or AB1, and perforce transformer-coupled to the load, were dominant until the early 1960s, when truly dependable transistors could be made at a reasonable price. Designs using germanium devices appeared first, but suffered severely from the vulnerability of germanium to even moderately high temperatures; the term thermal runaway was born. At first all silicon power transistors were NPN, and for a time most transistor amplifiers relied on input and output transformers for push-pull operation of the power output stage. These transformers were as always heavy, bulky, expensive, and non-linear, and added insult to injury as their LF and HF phase-shifts severely limited the amount of negative feedback that could be safely applied.
The advent of the transformerless Lin configuration[1], with what became known as a quasi-complementary output stage, disposed of a good many problems. Since modestly capable PNP driver transistors were available, the power output devices could both be NPN, and still work in push-pull. It was realised that a transformer was not required for impedance matching between power transistors and 8 ? loudspeakers.
Proper complementary power devices appeared in the...
