An FFT-Based High SNR MWC Back-End Architecture with Analog Low-Pass Filter Compensation

• Published in: Chinese Journal of Electronics Vol. 29; no. 3; pp. 563 - 573
• Main Authors: Liu, Sujuan, Liu, Lei
• Format: Journal Article
• Language: English
• Published: Published by the IET on behalf of the CIE 01.05.2020
• Subjects: advanced MWC; analog filter compensating method; analog low‐pass filter compensation; back‐end architecture; compensating operation; Compensation; Compressed sensing; computation loads; computational complexity; conventional back‐end; conventional MWC backend; different filters; digital signal; digital signal processing chips; expanding operation; Fast Fourier transform; fast Fourier transforms; FFT‐based high SNR MWC; frequency point; ideal MWC; improved back‐end; low‐pass filters; Modulated wideband converter; nonideal filters; Non‐ideal low‐pass filters; proper length; recovery Signal‐to‐noise ratio; recovery SNR; signal processing; signal sampling; fast Fourier transforms; signal processing; recovery SNR; Modulated wideband converter; Fast Fourier transform; Compensation; analog low-pass filter compensation; signal sampling; conventional MWC backend; low-pass filters; different filters; frequency point; nonideal filters; proper length; computation loads; back-end architecture; improved back-end; Non-ideal low-pass filters; recovery Signal-to-noise ratio; digital signal; compensating operation; ideal MWC; FFT-based high SNR MWC; conventional back-end; expanding operation; advanced MWC; digital signal processing chips; Compressed sensing; computational complexity; analog filter compensating method
• ISSN: 1022-4653; 2075-5597
• DOI: 10.1049/cje.2020.03.012

We study the impacts of non-ideal filters in Modulated wideband converter (MWC). An analog filter compensating method is presented. The limitations of the conventional back-end of the advanced MWC is analyzed and an improved back-end based on Fast Fourier transform (FFT) is proposed. By setting proper length of the digital signal processed each time, the proposed system increases the recovery Signal-to-noise ratio (SNR) and significantly reduces the computation loads. The compensating operation is accurate to each frequency point and the expanding operation is simplified to a splitting of the frequency points. Numerical simulations demonstrate our proposed system against the ideal MWC with different filters. The results indicate that the recovery SNR of over 70 decibel is achieved with almost only one-tenth the computation complexity of the conventional MWC backend.