Error-feedback for amplifier linearization

Abstract

A novel general linearization topology called 'error-feedback' is presented and applied to the linearization of amplifiers. Error-feedback uses the difference/error between (linearly) scaled input and output signal in its feedback loop to reduce nonlinearity. It does not rely on main-amplifier gain for linearization as does traditional feedback. Instead an auxiliary amplifier is introduced into the feedback loop, leaving the main amplifier's linear gain unaffected. Results from theory, and simulation as well as measured results complement each other. They show that linearization takes place over a continuous band. The topology shows prospects for concurrent multi-band linearization. Internal delay and phase shifts influence design limits and overall performance.