Comparison of linearization techniques for differential amplifiers in integrated circuit design

Due to the high intermodulation and adjacent channel radiation performance requirements, state of the art mobile communication systems ask for extremely linear transfer characteristics of the input, output and mixer stages of the transceivers. This requirement can be met by using high power or applying linearization techniques. Several methods like the emitter-coupled pair with emitter degeneration, the linearized cross-quad circuit and the concept of employing plurality of emitter-coupled pairs have been proposed to reduce the nonlinearity of the emitter-coupled pair. But they have been mainly characterized by their total harmonic distortion for audio applications. For RF-applications the l-dB gain compression point and the third-order intercept point are widely used. The authors consider that none of these figures of merit is appropriate to compare the performance of various linearization techniques. They define a signal-to-intermodulation ratio taking into account intermodulation products up to the 7th-order. This ratio is a much more significant parameter representing the circuit performance. They present the intermodulation behaviour of three linearization concepts compared with that of the conventional emitter-coupled pair