We demonstrate a single-ended colorless coherent receiver using symmetric 3x3 couplers for optical hybrids. We show that the receiver can achieve colorless reception of fifty-five 112-Gb/s polarization-division-multiplexed quadrature-phase-shift-keyed (PDM-QPSK) channels with less than 1-dB penalty in the back-to-back operation. The receiver also works well in a long-haul wavelength-division-multiplexed (WDM) transmission system over 2560-km TrueWave® REACH fiber.
© 2012 OSA
Colorless channel drop capabilities can significantly increase the flexibility and efficiency of reconfigurable optical networks. Colorless channel drop nodes can be realized using reconfigurable wavelength demultiplexers constructed of tunable optical filters or of wavelength selective switches (WSS). Alternatively, colorless receivers that do not require any optical filtering mechanisms to separate individual wavelength-division-multiplexed (WDM) channels can be used [1,2].
In recent years, research on high capacity optical communications has shifted to digital coherent detection due to its potential for high spectral efficiency and sensitivity [3,4]. Balanced coherent receivers with a high common-mode-rejection ratio (CMRR) and/or with a high power difference between signal and local oscillator (LO) can be used as colorless receivers without the need for demultiplexing filters, since they suppress the direct-detection terms of all incident WDM channels [5–7]. In contrast, using simpler single-ended receiver front-ends has been shown to yield a 4-dB penalty for colorless detection of 16 WDM channels .
In this paper, which extends the results of , we study the colorless detection performance of coherent receivers and explore ways to simplify the optical front-end while maintaining colorless operation. We do so by replacing the 90° optical hybrid with a simple 3x3 coupler. Optical couplers have been used in coherent optical communications for phase diversity receivers [9–14], and an asymmetric 1:2:2 coupler has been used recently as a 90° hybrid . Here, we demonstrate the use of a simple symmetric 3x3 coupler in conjunction with three single-ended photo-detectors (instead of an optical hybrid with two pairs of balanced detectors). We show that this receiver structure can completely eliminate the direct-detection terms, providing a simple and potentially cost effective alternative to a balanced coherent receiver . Our experimental results show that the receiver can achieve colorless detection of fifty-five 112-Gb/s polarization-division-multiplexed quadrature-phase-shift-keyed (PDM-QPSK) channels at less than 1-dB penalty. Almost no penalty is found for LO-to-signal-power-ratios (LOSPRs) as low as 0 dB. We also demonstrate transmission of sixteen 112-Gb/s PDM-QPSK WDM channels at 50-GHz channel spacing over 2560-km TrueWave® REACH fiber using this receiver.
2. Symmetric 3 x 3 couplers in optical hybrid operation
Optical 3 x 3 couplers were widely used as optical hybrids in coherent optical receivers in the 1980s [5,7,9–11]. As shown in Fig. 1 , with the optical signal and LO fields Es and EL at the input of the coupler, the three output fields can be expressed as 
The three output currents after the photo-detectors areEq. (4) is the direct-detection term and the second term is the beat term, with a 2π/3 phase difference between its three components. In a WDM system of N channels, , where Esi is the optical signal field of channel i. The direct-detection signal term is and can become relatively large compared to the beat term, especially if there are many WDM channels incident to the receiver. In this case, since the direct detection term occupies the same spectral region as the information-bearing beat term, the direct-detection term must be suppressed. In a single-ended receiver this suppression is done by choosing a sufficiently large LOSPR. In a balanced receiver, suppression is further improved by a large CMRR of the opto-electronic front-end. In the receiver proposed here, inphase (I) and quadrature (Q) components are obtained from the three detected currents with the following simple operations,
Like in a conventional balanced receiver, the direct-detection signal term and the LO term (including the LO relative intensity noise (RIN)) are automatically eliminated by performing the above operations.
3. Coherent receiver architectures
Figure 2 shows two architectures of a polarization and phase diversity coherent receiver using single-ended photo-detectors based on symmetric 3x3 couplers. In the architecture shown in Fig. 2(a), the operations to obtain the I/Q components from the three photocurrents are performed with simple analog scaling and subtraction circuits followed by two analog-to-digital converters (ADCs), in analogy to a single-ended or balanced coherent receiver using a 90° hybrid. In the architecture shown in Fig. 2(b), three ADCs are used to individually convert the three detected signal components to digital signals, and the operations given in Eq. (5) are performed digitally. Compared to the architecture in Fig. 2(a), this requires two more ADCs, but offers the advantage of all-digital (and hence easy to calibrate) compensation using simple digital signal processing (DSP). As an additional benefit, the DSP in the right architecture has access to the direct-detection signal term ( after eliminating the LO with a direct-current (DC) blocker). This term contains the directly detected signals of all WDM channels and can be potentially used for cross-phase modulation (XPM) compensation .
4. Experiments and results
4.1 Back-to-back experiment
The experimental setup for a back-to-back test of the receiver is shown in Fig. 3 . Sixteen odd channels (from 1574.13 nm to 1586.62 nm) and 39 even channels (from 1570.42 nm to 1603.17 nm) in the L-band were combined with a power combiner and WDM multiplexer, respectively, and modulated by two nested Mach-Zehnder modulators (MZMs) to generate 28-Gbaud (56-Gb/s) QPSK signals (as we only had one 100-GHz WDM multiplexer, only 16 odd channels were used with a power combiner). The channel under test was derived from an external cavity laser (ECL) and the other channels from distributed feedback (DFB) lasers. Each of the I/Q drive signals for the modulators was formed by electronically multiplexing four delay-decorrelated copies of a 7-Gb/s 215-1 pseudo-random bit-sequence (PRBS). The odd and even channels were combined with a 50/100-GHz interleaver (ITL). A polarization multiplexer (PolMux) with 290 symbols delay between the two polarization tributaries was used to generate the 112-Gb/s PDM-QPSK signals for both odd and even channels. After amplified spontaneous emission (ASE) noise loading, a wavelength and bandwidth tunable filter was used to select the number of incident channels going into the receiver. Signal and LO power were adjusted by variable optical attenuators (VOAs) and are combined with two symmetric 3x3 fiber couplers, whose parameters are given in Table 1 . The receiver structure was the same as that shown in Fig. 2(b) except that two polarization controllers (PCs) were used to align the polarizations of the signal and LO. The six signal components from the photo-detectors were captured by two 4-channel 40-GSamples/s real-time oscilloscopes with 16-GHz bandwidth. The captured signal was digitally processed offline, using the constant modulus algorithm (CMA) for polarization demultiplexing and inter-symbol interference mitigation. Carrier phase estimation was performed using the Viterbi-Viterbi algorithm, and the bit-error rate (BER) was evaluated with differential decoding.
We first studied the impact of the number of incident channels on receiver performance. We fixed the LO power at 21 dBm and adjusted the bandwidth of the tunable filter to vary the number of channels going into the receiver (the filter was bypassed for 55 channels). The signal power was adjusted to keep the per channel power fixed at 1 dBm. The results are given in Fig. 4(a) . It shows that there is almost no penalty for colorlessly receiving 16 channels, and less than 1-dB penalty for colorlessly receiving 55 channels. Note that in the offline processing, we tried imbalance compensation, but little improvement was obtained. Therefore we believe that the impact of noticeable differences among the coupling ratios in the 3x3 couplers on the receiver performance is small and the penalty for receiving 55 channels is caused by the reduced effective number of bits (ENoB) of the ADCs for the beat term in the oscilloscopes, as the ENoB is shared by both the direct-detection term and the beat term in Eq. (4).
We also tested the effect of LOSPR on receiver performance. In this test, the tunable filter selected one channel and the optical-signal-to-noise ratio (OSNR) was fixed at 16.5 dB. The signal power was adjusted to keep the sum of signal and LO powers (in dBm) constant for different LOSPR levels so that the power of the beat term remained the same (see Eq. (4). As shown in Fig. 4(b), the receiver showed no noticeable penalty with LOSPRs of down to 0 dB.
Figure 5 shows plots of captured I1 versus I2 signals without any signal processing for different cases. Plots of I1 versus I3 and I2 versus I3 are similar. As shown in Fig. 5(a), at a high LOSPR, the signal and LO beat term is much larger than the direct-detection term, and the 2π/3 phase difference between I1 and I2 is evident from the tilt of the elliptical signal cloud. When the LOSPR is small and there are many channels, the direct-detection term becomes larger, and one cannot see the 2π/3 phase difference between I1 and I2 in the captured clouds any more, as illustrated in Figs. 5(b) and 5(c).
We next tested the impact of imbalance among the three ports on the performance of the receiver. In this test, we put additional 2-dB and 3-dB attenuators at output 1 of the symmetric couplers but before the photo-detectors for x and y polarizations, respectively. We used 16 channels, which all passed the tunable filter and went to the receiver. The LO power was kept at 21-dBm and we measured BER versus OSNR using two methods. In the first method we did not compensate for any imbalance in offline processing, while in the second method the imbalance was compensated using offline DSP by simply setting the mean square values of the three currents equal before further signal processing. The results are depicted in Fig. 6 . For comparison, the result without any additionally added imbalance among the three ports is also given in the figure. The figure shows that the large induced imbalance causes more than 5-dB penalty if left uncompensated, but after imbalance compensation using DSP, the penalty is reduced to about 0.5 dB. As discussed above, this penalty is due to reduced ENoB of the ADCs for the beat term of I1.
We then compared the performance of this receiver with a balanced and with a single-ended coherent receiver using 90° hybrids. The results are given in Fig. 7 , which depicts the BER versus the number of incident channels for the three receivers at 16.5-dB OSNR. It shows that the single-ended receiver using a 3x3 coupler as a hybrid can achieve similar performance as the balanced receiver using a full 90° hybrid. For the single-ended receiver using a 90° hybrid, with 16 channels incident to the receiver, the Q2-factor penalty exceeds 6 dB.
4.2 Transmission experiment
It has been shown that the impact of incident channels on the performance of a coherent receiver also depends on factors such as accumulated chromatic dispersion (CD) . In order to study the effects of CD and nonlinearities on the colorless reception performance of the proposed receiver, we tested the receiver in a 112-Gb/s PDM-QPSK long-haul WDM transmission system. The transmission experiment was performed in a 4x80-km dispersion-managed all-Raman amplified TrueWave® REACH fiber span recirculating loop. The TrueWave® REACH fiber has the CD and CD slope coefficients of 8.8 ps/nm-km and 0.0417 ps/km-nm2 at 1580nm respectively. The loss coefficient and Aeff of fiber at 1580nm are 0.196 dB/km and 57 μm2, respectively. The experimental setup is illustrated in Fig. 8 . To emulate legacy transport systems, the dispersion of each span was compensated by dispersion compensation fiber (DCF) with a residual dispersion per span near 30 ps/nm, and a DCF with −300-ps/nm was used as dispersion pre-compensation. Both TureWave® REACH fiber and DCF were backward Raman pumped, with the launch power to the DCF about 2-dB lower than that to the spans. The signal spectrum was flattened by a dynamic grain equalizing filter (DGEF) after each loop and the losses of the loop and DGEF were compensated by an erbium-doped-fiber amplifier (EDFA). In this experiment, sixteen 112-Gb/s PDM-QPSK channels (1577.44 nm to 1583.69 nm) at 50-GHz channel spacing were used. The channel under test (a center channel) was derived from an ECL and the other channels from DFB lasers. Both the transmitter and receiver were the same as those in Fig. 3. A tunable filter at the receiver was tuned to choose one or all of the 16 channels going to the receiver.
The BER of a center channel versus launch power after 2560-km transmission is shown in Fig. 9 , with the tunable filter passing 1 and 16 channels into the receiver. It shows that the receiver works well for colorless reception in the long haul system, and similar performance is obtained with 1 and 16 channels going into the receiver. At optimum launch power, a BER below the forward-error correction (FEC) threshold of 3.8x10−3 can be achieved after 2560-km dispersion-managed transmission.
We have demonstrated a colorless coherent receiver using simple symmetric 3x3 couplers for optical hybrids and single-ended photo-detectors. We showed that the receiver can achieve colorless reception of fifty-five 112-Gb/s PDM-QPSK channels with less than 1-dB penalty. There is no noticeable penalty with LOSPRs as low as 0 dB. We further demonstrated long-haul transmission of 16x112-Gb/s PDM-QPSK WDM channels at 50-GHz channel spacing over 2560-km of TrueWave® REACH fiber using this receiver.
References and links
1. J. E. Simsarian, J. Gripp, A. H. Gnauck, G. Raybon, and P. J. Winzer, “Fast-tuning 224-Gb/s Intradyne Receiver for Optical Packet Networks,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference, OSA Technical Digest Series (CD) (Optical Society of America, 2010), paper PDPB5.
2. L. E. Nelson, S. L. Woodward, S. Foo, M. Moyer, D. J. S. Beckett, M. O’Sullivan, and P. D. Magill, “Detection of a single 40 Gb/s polarization-multiplexed QPSK channel with a real-time Intradyne receiver in the presence of multiple coincident WDM channels,” J. Lightwave Technol. 28(20), 2933–2943 (2010). [CrossRef]
3. P. J. Winzer, “Beyond 100G ethernet,” IEEE Commun. Mag. 48(7), 26–30 (2010). [CrossRef]
4. K. Roberts, D. Beckett, D. Boertjes, J. Berthold, and C. Laperle, “100G and beyond with digital coherent signal processing,” IEEE Commun. Mag. 48(7), 62–69 (2010). [CrossRef]
5. S. Yamashita and T. Okoshi, “Suppression of common-mode beat noise from optical amplifiers using a balanced receiver,” Electron. Lett. 28(21), 1970–1972 (1992). [CrossRef]
6. V. Carena, V. Curri, P. Poggiolini, and F. Forghieri, “Dynamic range of single-ended detection receivers for 100GE coherent PM-QPSK,” IEEE Photon. Technol. Lett. 20(15), 1281–1283 (2008). [CrossRef]
8. C. Xie, P. J. Winzer, G. Raybon, A. H. Gnauck, B. Zhu, T. Geisler, and B. Edvold, “Colorless Coherent Receiver Using 3x3 Coupler Hybrids and Single-Ended Detection,” in Proceedings of European Conference on Optical Communication, (Geneva, Switzerland, 2011), paper Th.13.B.2.
9. L. G. Kazovsky, P. Meissner, and E. Patzak, “ASK multiport optical Homodyne receivers,” J. Lightwave Technol. 5(6), 770–791 (1987). [CrossRef]
10. J. Pietzsch, “Scattering matrix analysis of 3 x 3 fiber couplers,” J. Lightwave Technol. 7(2), 303–307 (1989). [CrossRef]
11. Y. H. Ja, “Analysis of four-port optical fiber ring and loop resonators using a 3 x 3 fiber coupler and degenerate two-wave mixing,” IEEE J. Quantum Electron. 28(12), 2749–2757 (1992). [CrossRef]
12. G. Nicholson and T. M. Stephens, “Performance analysis of coherent optical phase-diversity receivers with DPSK modulation,” J. Lightwave Technol. 7(2), 393–399 (1989). [CrossRef]
13. I. Bar-David, “Direct Differential Detection of Phase-Shift-Keyed Signals: a Local-Oscillatorless DPSK Receiver,” IEE Proc., Optoelectron. 141(1), 38–42 (1994). [CrossRef]
14. S. J. Savory, G. Gavioli, R. I. Killey, and P. Bayvel, “Electronic compensation of chromatic dispersion using a digital coherent receiver,” Opt. Express 15(5), 2120–2126 (2007). [CrossRef] [PubMed]
15. L. Du and A. Lowery, “Experimental Demonstration of XPM Compensation for CO-OFDM Systems with Periodic Dispersion Maps” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference, OSA Technical Digest Series (CD) (Optical Society of America, 2011), paper OWW2.