Model Matching Approach in RF Power Amplifier Linearization

Abstract

KurzfassungRadio frequency( RF) power amplifiers (PA) are the most power consuming components of am obile communications unit. They are used to convert the DC power from the battery into RF power delivered to the antenna. In ac ell phone it becomes very important to use highly efficient power amplifiers, such as Class Ca nd Class EP As, to increase the talk time whichisdirectly proportional to the battery life. On the other hand, these RF PAsa re inherently nonlinear and produce spectral regrowth and other undesirable effects. Therefore, to exploit their high efficiency, it is desirable to employl inearization techniques to linearize their overall response. Linear model matching linearization techniques are investigated in this work to compensate for PA nonlinearities. Thea pplication of theses techniques results in a controller architecture that delivers excellent linearity performance of compensated Class CP Am odels, making Figure 1.6-1 shows the result of aT woTone Test. Themagnitude response plots of the uncompensated and compensated models are shown. Thet hird order products in the compensated case are suppressed significantly, at least by 18 dB and the fifth order intermodulation products are suppressed by 16 dB.Thei mprovement of noise behavior is shown in Figure 1.6-2, considering the result of the white noise test. Thepower spectraofthe input and of the output of the uncompensated and compensated models are shown. Thes pectrum of the output band-limited noise, in the compensated case, preserved the deep notch in the middle of the spectrum, whereas in the uncompensated case the notch wasfi lled-in significantly.T his test also provest he linear behavior of the compensated power amplifier.[1] Prof. Dr.M ario Magaña from the Oregon State University,U SA, has been at the University of Applied Science Offenburg in the summer semester 2008 for research and teaching.