Advertisement

In this article, a new Doherty amplifier using high-efficiency inverse class F and class F amplifiers is presented. The characteristics of inverse class F and class F amplifiers have been analyzed for their bias conditions and load impedances in order to identify their suitability to be the carrier or peaking amplifier in a Doherty amplifier. An inverse class F amplifier was employed for the carrier amplifier due to its superior linearity and efficiency performance over a wide range of load impedances. The class F amplifier was used for the peaking amplifier by virtue of its larger gain expansion characteristics. The proposed Doherty amplifier was configured using the implemented inverse class F and class F amplifiers for the 1 GHz band and evaluated using two-tone and down-link WCDMA signals. A state-of-the-art efficiency performance was achieved at a given linearity level: A power-added efficiency of 50.9 percent at a third-order intermodulation distortion level of -30 dBc from the two-tone test and 45.2 percent at an adjacent leakage power ratio of -30 dBc from the WCDMA test, respectively.

Wide-band signals with a high peak-to-average ratio (PAR) have been extensively adopted in modern wireless communication systems. For radio-frequency circuit designers, the efficiency versus linearity trade-off turns into one of the most important design considerations. Doherty amplifiers, which employ a load impedance modulation technique, have been adopted for use in base station power amplifiers due to their simple structure, good linearity and high efficiency at average output power levels.^1-6 For high efficiency, various class F and inverse class F (or class F^-1) amplifiers have been analyzed and implemented. Most of the previous work regarding class F or class F^-1 amplifiers has been designed to enhance efficiency.^7-9 However, these amplifiers inherently have poor linearity due to excessive harmonics in the voltage or current waveforms.¹⁰

S. Goto, et al proposed a Doherty amplifier with a combination of class F and inverse class F amplifiers.¹¹ They reported high efficiency at an output power back-off of 10 dB for two-tone excitation. However, the efficiency and linearity rapidly degraded after the peak point, which can result in significant performance degradation for the signals having much higher peak-to-average power ratio (PAR or PAPR), such as the down-link wide-band code division multiple access (WCDMA) signals. J. Kim, et al proposed a saturated Doherty amplifier using class F amplifiers.¹² They reported very high efficiency but poor linearity. For the WCDMA signal, the reported performance could not come down to an adjacent channel leakage power ratio (ACLR) of -30 dBc. The ACLR performance got even worse for the lower output power level.