We.3.A.4.pdf ECOC Technical Digest © 2012 OSA Low-Complexity DSP Using Undersampling for Heterodyne Receivers in Coherent Passive Optical Access Networks S. Dris (1) , P. Bakopoulos (1) , I. Lazarou (1) , B. Schrenk (1) , H. Avramopoulos (1) (1) School of Electrical & Computer Engineering, National Technical University of Athens, Iroon Polytechniou 9, 15773 Zografou, Greece, sdris@mail.ntua.gr Abstract The first application of bandpass sampling in a heterodyne coherent optical communication system is presented. An undersampling receiver achieving a 50% reduction in ADC rate is developed in DSP and its penalty-free operation verified in a 76km coherent UDWDM-PON scenario. Introduction Coherent heterodyne detection has emerged as a promising candidate for the evolution of access networks, as it offers increased receiver sensitivity to mitigate large splitting losses, higher spectral efficiency through the use of advanced modulation formats, as well as simpler optical hardware at the Optical Network Unit (ONU) compared to intradyne reception. To reduce cost, the same ONU laser is used both as the local oscillator (LO) as well as the upstream transmitter, implying that the LO is sufficiently detuned from the downstream to avoid the adverse effects of Rayleigh backscattering (RB) (Fig. 1a). The two main drawbacks associated with the presence of an intermediate frequency (IF) in such a scheme are the need to accommodate a large optoelectronic bandwidth, and the high sampling rate (f s ) required at the ADC to satisfy the Nyquist theorem (>2·(f c +B/2) for an IF frequency f c and signal bandwidth B). While the former cannot be easily avoided, it is possible to employ advanced sampling theory concepts to bring f s arbitrarily close to the theoretical minimum (2·B) needed to reconstruct the modulating baseband signal (a process often termed undersampling or bandpass sampling 1 ). However, this requires careful selection of the carrier and sampling oscillator frequencies f c and f s , which in turn must be precise enough to avoid detrimental aliasing effects. This poses the question of whether such a scheme can be used in an optical coherent communication system, where fast (and relatively large) frequency fluctuations occur due to normalized laser linewidths that are typically as high as ∆v·T s ≈10 -3 (symbol period, T s ). In this work we implement a simple heterodyne demodulator in DSP and show, for the first time to the best of our knowledge, that bandpass sampling can be effectively employed in coherent optical receivers to significantly relax the rate and memory requirements on the front- end digitizer, despite frequency-instability issues associated with commercial-off-the-shelf lasers. In adhering to the low-complexity nature of ONU hardware, complex digital down-conversion (DDC) to the baseband is achieved with a Hilbert transform filter. Thus a single FIR filter replaces the two required in a conventional quadrature DDC, roughly halving the number of multiply-accumulate operations needed at this stage. Moreover, no dispersion compensation or adaptive equalization is included. In addition, we validate the performance of the proposed digital receiver in a realistic application scenario. Using the developed DSP at the ONU, a 3 GBaud (6 GHz bandwidth) heterodyne-detected QPSK signal residing on a 9.3 GHz IF is sampled at a rate of only 12.5 GSa/s, with negligible performance penalty in a 76-km coherent ultra-dense WDM-PON (UDWDM-PON) architecture. This is very close to the theoretical minimum (2·B=12 GSa/s), and is half the rate required when performing conventional Nyquist sampling on the same signal (2·(9.3+3)=24.6 GSa/s). In addition, it represents a ~2.9-fold (1.4-fold) rate reduction (normalized to the symbol rate) compared to NSN’s state-of-the-art implementation 2 , where 0.311 (0.622) Gbaud heterodyne signals were sampled at 3.7 GSa/s. Fig. 1: (a) Allocation of upstream, local oscillator (LO) and downstream in an access network and buildup of RB. (b) Illustrating the spectral replications appearing at intervals of f s after digitization of the original analog signal at a sub-Nyquist rate, <2·(f c +B/2).