Comprehensive study of Class‐‐E/F2 and inverse Class‐‐F power amplifiers for mm‐Wave systems utilizing 130 nm CMOS process

Abstract

AbstractThis paper presents a comprehensive study of Class‐‐E/F2 and inverse Class‐‐F power amplifiers (PAs) designed for 35–42 GHz millimeter‐wave applications, utilizing a 130 nm CMOS process. The proposed RF PAs are well suited for wide frequency‐band millimeter‐wave and 5G radio transmitters. A comprehensive study of limiting factors in amplifier efficiency is presented in this paper. The study encompasses an analysis of all the factors that restrict the efficiency of the switched power amplifier. The first proposed power amplifier is based on the Class E/F2 architecture, comprising a parallel capacitor, second harmonic resonance circuit, and output matching network. Conversely, the second suggested power amplifier is constructed using the inverse Class‐‐F topology, which includes harmonic termination and matching networks. The harmonic termination circuit incorporates a second harmonic resonance network and utilizes a parasitic capacitor to control the harmonic components. Both the proposed Class E/F2 and inverse Class‐‐F architectures are employed to reshape the drain current and voltage waveforms, aiming to reduce the overlap between them and, consequently, improve efficiency. The suggested power amplifiers comprise a driver Class‐‐AB stage and a power stage constructed based on either the Class E/F2 architecture or inverse Class‐‐F topology, utilizing harmonic termination networks at the output load. Two new designs of high‐Q factor on‐chip finger capacitors have been implemented to improve efficiency and radio frequency performance. Achieving input matching and inter‐stage matching is facilitated by employing two novel on‐chip transformers designed for maximum power transfer. Additionally, an inter‐stage matching inductor is utilized in a cascode configuration to enhance the overall RF performance. The on‐chip transformers, inductors, and finger capacitors are designed using the HFSS software program. S‐parameter files (SNP files) of the designed on‐chip components are extracted and inserted into the simulation tool to ensure accurate results. The proposed Class E/F2 power amplifier achieves a constant power of 14.9 dBm, an extreme power added efficiency (PAE) of 11.7%, and a maximum gain of 13.74 dB. In contrast, the suggested inverse Class‐‐F power amplifier attains a constant power of 15.4 dBm, a peak PAE of 12.6%, and a maximum gain of 14.8 dB. The DC power consumption is 73 and 66 mW for the proposed Class E/F2 and inverse Class‐‐F PAs, respectively, while their active sizes are 0.14 and 0.2 mm2.