Rf Power Amplifier Linearization Techniques

The use of spectrum-efficient digital linear modulation techniques and multi-carrier radio communication systems place stringent linearity requirements on the power amplifier, PA, in order to maintain modulation fidelity and avoid interfering with communication systems in the adjacent channels. Though conventional output power back-off can improve the linearity of a power amplifier, it severely degrades the power efficiency and increases the capacity requirement of the amplifier.In the past, many circuit-level and system-level linearization schemes have been presented to solve the competing linearity and energy efficiency requirements. Analog pre-distortion stands out because of its simple structure, low cost, medium linearity, high efficiency compared with feed-forward and suitable for stand-alone integration. Device-level linearization techniques have been standing as a promising area in the solution of the power amplifier, linearity versus efficiency trade-off, thanks to recent advances on the understanding of the device's large-signal distortion behavior. In all cases, achieving the desired power amplifier performance is only possible due to the appropriate understanding of the power amplifier behavior, especially the intermodulation distortion behavior, from a device, circuit or system perspective.Based on an improved nonlinear drain-source current model, analytic expressions for the power of the second and third-order intermodulation products, for common- source amplifier are derived. Both the indirect contributions of the second-order Taylor series coefficients to the third-order intermodulation products and the contributions of the cross-terms to overall intermodulation products are included. Based on the analytic expressions of the IM products, intermodulation distortion (IMD) sweet-spot and its dependencies on Taylor series coefficients, gate bias, drain bias, temperature, input signal levels and loads are analyzed. Under large signal condition, the IMD behavior of the amplifier is explored. Better efficiency may be achieved with similar linearity as in class A. At high frequency, the contribution of nonlinear capacitor plays an important role and must be considered. It may smoothen the IMD behavior making only one less pronounced sweet-spot appear. However, proper tuning of the harmonic terminating impedances might restore the performance. A true second harmonic short-circuit configuration, which taking account of the parasitic inductance of the FET drain, is presented.For analog pre-distortion, the operating principle is described and applied to the analysis of its ideal performances. In order to improve the performance, we analyze some basic performance degradation factors, including high-order nonlinearity components, the amplitude and phase error, delay mismatch effects between the pre-distorter and the power amplifier that needed pre-distorted. Some memory effects manifest themselves by producing asymmetric intermodulation distortion products on either side of the signal have been provided and solutions to that are also given. With a deep understanding of the self-biasing mechanism of the diode, a wide-band pre-distortion linearizer with waveform control is designed to improve the linearity performace.