Modelling spectral regrowth and the effect of packet size

Abstract

Traditionally nonlinearity has been characterized by two-tone intermodulation distortion. However, for digital radio, modelling spectral regrowth is much more important. There is currently much interest in modelling spectral regrowth in microwave power amplifiers for wireless communications. The ability to predict spectral regrowth or, more specifically, adjacent channel power ratio (ACPR), has a direct impact on design to achieve acceptable linearity performance. Since there is a trade-off between linearity and efficiency the amount of back-off can significantly affect power amplifier efficiency and thus, for example, talk-time for a mobile handset. Many wireless communications systems, such as GSM, use time domain multiple access (TDMA) schemes, sending signals in bursts (or packets). Using a behavioural model which can predict ACPR given an estimate of input signal statistics, it is shown that ACPR is highly dependent on the input signal statistics and it is dependent on the length of the packet.