Nonlinear Characterization and Distortion Mitigation in Six-Port Modulator

Abstract

This paper presents the characterization of six-port modulator for various hardware impairments with scheme for their digital correction. Six-port modulator is low-cost and wideband, which is easily scalable to mm-wave frequencies due to its simplified architecture. However, similar to quadrature modulator, six-port modulator also has few impairments resulting into distortion in terms of output spectrum with adjacent channel leakage, I/Q imperfections, and carrier leakage. These hardware imperfections and impairments are characterized, and the digital predistortion is then applied at baseband data to mitigate the distortion appearing at the output of six-port modulator. A neural network model is employed as inverse model in digital predistorter due to its inherent structure capable of modeling cross-interference, nonlinearity, and dc bias. The proposed characterization and distortion mitigation system is tested for modulating 64- and 256-quadrature amplitude modulation (QAM) data with the speed of 100 and 400 Mbps, respectively. These are modulated to radio frequency carrier at 2 GHz. Without a digital correction, the adjacent channel leakage ratio and error vector magnitude (EVM) does not qualify the industry standard. After the digital correction, the EVM of value as high as 20.14% in case of 256-QAM is reduced to 2.15%, which comply with the industry requirements.