线性化器
预失真
高电子迁移率晶体管
单片微波集成电路
静脉曲张
拓扑(电路)
放大器
Ku波段
宽带
电子工程
电气工程
晶体管
数学
工程类
物理
电容
电压
CMOS芯片
电极
量子力学
作者
Dawei Zhang,Wenli Fu,Xiangke Deng,Xin Xu,Bo Zhang,Hongxi Yu,Kaixue Ma
标识
DOI:10.1109/tmtt.2022.3219404
摘要
A new design method of microwave monolithic integrated circuit (MMIC) predistortion linearizer compatible with the standard gallium arsenide (GaAs) high-electron-mobility transistor (HEMT) technology is proposed for wideband cancellation of power amplifier (PA)'s frequency-dependent amplitude-modulation (AM)/phase-modulation (PM) distortion. Adopting the dual-branch vector synthetic topology, a modified reflective topology is proposed to improve the extracted nonlinear signal with higher gain expansion dynamic range, and the reversed-biased HEMT-based varactor is deployed to realize variable phase tuning capacity. Phase and time-delay allocation between the linear and nonlinear branches is analyzed to achieve arbitrary monotone frequency-dependent phase predistortion performance. Four linearizer MMICs are designed and fabricated using a commercial 0.15- ${\mu }\text{m}$ GaAs HEMT process based on the proposed method and design procedure, in which the $X$ - and $Q$ -band chips are dedicated for positive phase conversion with the increase in input power, and the $Ku$ - and $K$ -band chips are for negative phase conversion. The $X$ -, $Ku$ -, and $Q$ -band designs are capable of exhibiting monotone increasing phase conversion with frequency, while the $K$ -band design exhibits monotone decreasing phase conversion with frequency. All these results are validated by on-chip measurement. Moreover, measurement conforms that consistent predistortion performance can be maintained by tuning the biasing voltage under different temperatures; static and dynamic measurements are performed with the designed linearizers applying to PA modules, validating the wideband cancellation capacity of frequency-dependent AM/PM distortion and the linearity improvement under wideband modulated signal excitation.
科研通智能强力驱动
Strongly Powered by AbleSci AI