A Study on Performance Enhancement of Base Station Power Amplifier

DoctorAs the wireless communication systems evolve, the systems need to transmit a large amount of data. To manage the increased data within a limited available frequency spectrum, the systems adopt spectrally efficient modulation schemes, such as wideband code division multiple access (WCDMA), long-term evolution (LTE), and mobile world interoperability for microwave access (m-WiMAX). These modulated signals have large peak-to-average power ratio (PAPR), because the envelope of these signals varies rapidly. Therefore, the power amplifiers need to be able to operate at the back-off power level with a high efficiency. Meanwhile, the transmitters should amplify these signals with high efficiency and linearity to lower the power consumption and preserve the signal quality. Also, the PAs need to be reconfigured properly for the variation of the output power, because the total data usage can vary depending on the time of a day. For this purpose, this dissertation presents various techniques to improve the efficiency, linearity, and gain characteristics of the power amplifiers. Especially, this dissertation presents the efficiency and gain improvement methods for the Doherty amplifier and the unit cell PA. Four major works toward the enhanced RF PAs are explored in this dissertation. First, for a femto-cell base station, a highly efficient symmetric Doherty PA is proposed and implemented. The Doherty PA is designed using monolithic microwave integrated circuit (MMIC) process for PA with a small size. As a highly efficient transmitter, the Doherty technique is investigated and analyzed to improve its performance. In the PA architecture, the nonlinear capacitance of the device affects the input match and the gain characteristic can be different according to the input matching condition. To enhance the gain, a new input impedance matching concept for a proper load modulation is proposed. Second, Doherty PA, which is optimized at back-off power, is introduced. To get high efficiency for amplification of modulated signals, the efficiency at a back-off power level should be high rather than at the peak power region. An optimized offset line and output matching circuit of the carrier PA are proposed not to affect the PA performance at the back-off power level. Also the efficiency is improved at back-off output power, while the peak output power is maintained. Third, bias adaptation of a Doherty PA is presented for high efficient operation. Gate bias adaptation and drain bias control are analyzed respectively. To maximize the output power and efficiency of a Doherty power amplifier, a new gate bias adaptation method of the carrier and peaking PAs with the control profiles of the bias voltages is proposed. Then, an average power tracking (APT) Doherty PA is analyzed in terms of its biasing voltage condition, efficiency, and output power. For the optimized operation of APT Doherty PA, the drain and gate bias voltages are optimized at different output power conditions. Fourth, an optimized Doherty power amplifier is designed using all topologies described so far for high efficiency at a high power. To increase the gain at the average power level, the input matching impedance is optimized at back-off power level. In conventional design, due to the phase variation of the peaking PA with power level, the conventional offset line does not produce the proper load modulation. To solve the problem, additional offset lines are adopted and the efficiency is increased at the modulation region. Fifth, an optimal design of a highly efficient power amplifier is described using independent fundamental and second harmonic impedance control technique. In fabrication of a power amplifier, a tuning method is indispensable because the simulation models of the device and capacitor have somewhat difference with the actual device. To achieve high drain efficiency, the fundamental and harmonic impedances need to be accurately optimized. The matching circuit of the PA adopts the independent harmonic control circuit using the characteristic of a quarter-wavelength microstrip line