Improved Quantization Noise Suppression for Low-order Sigma-delta over Fiber

Next-generation applications in communications require a low-cost mobile fronthaul (MFH) network architecture. We propose two quantization noise suppression schemes for low-order sigma-delta modulation (SDM) over fiber, subtracting either (1) one-bit sigma-delta modulation or, (2) analog noise modulation, of the in-band noise from the SDM signal. The resulting signals, unequal PAM-4 or quasi-one-bit respectively, allow the original signal to be recovered using only a simple bandpass filter, significantly reducing the overall system design and cost. The noise component is small relative to the desired signal such that intensity-modulation/direct-detection (IM-DD) can be used for low-cost distributed massive multiple-input multiple-output (M-MIMO) applications. Experimental results show EVM performance improvements of 256-QAM OFDM by 4.6 dB and 6.9 dB respectively, along with minimum optical dynamic ranges of 5 dB and 7 dB.