Open Access
Add to BibSonomyAbstract
When optical transceivers employing in-phase and quadrature (IQ) modulators and demodulator are used in real transmission systems, it is very important to search optimum bias points and to prevent bias drift from their optimal points due to of temperature variation, stress, or device aging. We demonstrate a simple and cost-effective automatic bias control scheme for optical IQ modulator and demodulator based on RF power and peak voltage detection. The principle of control scheme and effects of bias voltage on monitoring parameters are presented. The dynamic performance of the control scheme, effects of optical signal-to-noise ratio (OSNR) of signal, and effects of temperature variation are also evaluated. For the evaluation of bias control scheme at real-environment, we implement the automatic bias control scheme in 112 Gb/s dual carrier-differential QPSK (DC-DQPSK) transceiver, and investigate its long-term stability performances in a field transmission experiment over 797-km of installed fiber and ROADM.
© 2013 Optical Society of America
1. Introduction
Increasing data traffics require a continuous expansion of network capacity, and advanced modulation formats play a critical role in moving to high-capacity transmission. Up to recent days, to generate multi-level optical signal such as quadrature phase-shift keying (QPSK), quadrature amplitude modulation (QAM) and optical orthogonal frequency division multiplexing (OFDM), most of the advanced modulation formats typically used an integrated LiNbO3 (LN) inphase/quadrature (IQ) modulator structure with two-parallel Mach-Zehnder modulators (MZM) nested in MZ interferometer (MZI) [1,2]. Optimum operating conditions of the optical IQ modulator should be null transmission points in two MZMs for the generation of in-phase/quadrature data, and the quadrature phase (π/2) difference between in-phase and quadrature data. When IQ modulators are used in actual transmitters for advanced modulation formats, it is critically important to prevent bias drift from their optimal points due to temperature change, stress, or device aging. Thus, a simple and cost-effective automatic bias controller which can search proper bias points and track their optimum points is essential for the long-term operation. Several schemes have been reported to control the bias condition of IQ modulator using backward light [3], differential phase monitor [4], asymmetric sinusoidal dithering wave [5,6], or RF power of signal along with optical power [7,8]. These schemes would require additional light, complex constellation monitor, or precise lock-in amplifier module, respectively, and their performances evaluation at real environments were not reported yet.
Detection of advanced modulation format is another important issue for high speed transmission. In particular, direct detection based modulation formats, such as differential 8-ary phase shift keying (D8PSK), dual-polarization-differential quaternary phase-shift keying (DP-DQPSK), and dual carrier-differential quaternary phase-shift keying (DC-DQPSK) have been continuously investigated for metro network and data-center interconnection since these modulation formats could be implemented with low-power consumption [9–11]. At the receiver, a delay interferometer (DLI) whose free-spectral range is equal to symbol rate performs differential demodulation by converting phase modulated signal into intensity modulated signal. To properly decode phase modulated signal, searching proper bias points and tracking their optimum points are critically important for the long-term operation. However, there was few reports of detailed investigation on biasing and tracking of DLI [12].
In this paper, we demonstrate a cost-effective bias control scheme for optical IQ modulator and demodulator. In the suggested scheme, the optimal bias conditions of two MZMs and MZI are obtained by applying low frequency square-wave (~10 kHz) and detecting RF signal power, respectively. Neither additional light nor complex control circuit is required in the scheme. It is easy to distinguish and to estimate independently the bias drift of two MZMs and MZI. Searching proper bias points of optical IQ demodulator and tracking their optimum points were obtained by monitoring peak voltage of demodulated signal. To distinguish I- and Q- data of DQPSK signal, common and differential tuning of DLI is employed. For demonstration and evaluation at real-environment, we implement the automatic bias control scheme in 112 Gb/s dual carrier-differential QPSK (DC-DQPSK) transmitter and receiver, and investigate its long-term stability performances in a field experiment of 112 Gb/s DC-DQPSK signal transmission over 797-km of installed fiber and ROADM.
2. Automatic bias control scheme for optical IQ modulator and demodulator
Figure 1(a) shows the schematic diagram of automatic bias control for IQ modulator. The IQ modulator is composed of an integrated two-parallel MZMs nested in MZI. To search proper bias points and track their optimum conditions, the low frequency (10 kHz) square-wave is alternately applied to the two MZMs (bias 1 and 2) and its RF power is monitored to search proper bias points and track their optimum conditions of them. In this experiment, the amplitude of applied 10 kHz square wave was set to be 0.3 Vpp, and this value was similar with 5% of Vπ voltage. Since the frequency of square wave is less than typical low-frequency cut-off of high-speed receiver, the penalty induced by square-wave dithering could be avoided. For a single push-pull MZM, the output RF power of square-wave is given by
, where VbI,bQ is bias voltage of I- or Q- modulator, and Vπ is half wave voltage. For BPSK signal generation of each MZM, the bias points of two MZMs should set to be at their null transmission points (i.e., VbI,bQ = Vπ). Thus, the detected RF power of 10 kHz square-wave becomes minimum at optimum bias point. Consider the phase bias of MZI in IQ modulator where the phase difference between two BPSK signals is △ϕIQ. Assuming the bias of two MZMs are set to be at their null transmission point, the output optical power of QPSK signal is given byThus, the detected RF power is proportional to cos(△IIQ)2. To obtain π/2 phase difference between I- and Q- data, the detected RF power of QPSK signal should be minimized. Also, no dithering tone is used for MZI (bias 3). At the output of IQ modulator, 1% of output signal is tapped, and then it is applied to the bias controller. The bias controller detects the RF power of square wave less than 10 kHz and RF power of DQPSK signal less than 700 MHz, and these values are used to generate feedback signal of control circuit. Figure 1(b) shows the schematic diagram of IQ demodulator for direct detection. IQ demodulator is composed of two parallel DLIs. To distinguish I- and Q- data of DQPSK signal, common and differential tuning of DLI is employed. The common control is used to tune the phases of I- and Q-arm simultaneously, while keeping the phase difference between I- and Q-arm almost unchanged. The differential control is used to maintain the phase difference between I- and Q-arm with π/2. The 5% of output signal of balanced detector is tapped and applied to voltage peak detector to produce bias control signal of DLI.
Fig. 1 Automatic bias control scheme of optical IQ modulator and demodulator in DQPSK format (a) bias control of optical IQ modulator based on square wave and RF power detection (b) bias control of optical IQ demodulator based on peak voltage detection of demodulated signal.
Figure 2 shows simulated RF power of 10 kHz square-wave and that of 56 Gb/s DQPSK signal with different bias conditions. Since the RF power of square wave follows the transmission characteristics of MZM, the detected RF power of 10 kHz square-wave becomes minimum at their null transmission point, independent of △ϕIQ, as shown in Fig. 2(a). The RF power of 56 Gb/s DQPSK signal also becomes minimum when in-phase and quadrature data has π/2 phase difference, as shown in Fig. 2(b). Thus, we optimize the bias conditions of IQ modulator by minimizing the measured RF power of square-wave and that of QPSK signal, simultaneously. To find optimum bias points, we used a simple hill climbing algorithm.
Fig. 2 Bias control optical IQ modulator. Simulated results for (a) RF power of 10 kHz square-wave versus bias of MZM and (b) RF power of 56 Gb/s QPSK signal with various RF filter bandwidth versus bias of MZI.
Figure 3(a) shows simulated output data pattern of DQPSK signal after balanced detector. The results show that data pattern varies as a function of the phase difference setting of DLI. An incorrect phase difference setting of 0 or π, results in cancelation of some of the signal bits and enhancement of other signal bits. The phase difference setting of π/4 removes the cancelation of signal bits and reduce peak amplitude. Thus, peak voltage of balanced PD output could be used to proper bias setting of DLI. Figure 3(b) shows the measured output voltage of peak detector as a function of applied voltage to DLI. The measured voltage follows sinusoidal curves, and these characteristics are unchanged with respect to different OSNR of received signal. The minimum voltage corresponds to phase difference setting of ± π/4, for non-inverted and inverted data, respectively.
Fig. 3 Bias control optical IQ demodulator, in-phase channel (a) simulated data pattern after balanced detector (b) measured voltage of peak detector as a function of applied voltage.
3. Bit-error rate and dynamic performance
Figure 4(a) shows the measured monitoring signal of two MZMs and MZI in the 56 Gb/s DQPSK signal as a function of the number of feedback iterations. For the modulation, we applied two 28-Gb/s NRZ signals (pattern length = 231-1) to the QPSK modulator. The amplitudes of the NRZ signals were set to be ~2Vπ. The results show that the control loop searches the optimum bias conditions of each MZM and MZI one-by-one, and stabilizes all bias voltages within 45 iterations. We have confirmed the convergence to optimal operating points for many random initial biases and phases. Figure 4(b) shows the measured peak detector voltage of received 56 Gb/s DQPSK signal, where the phase difference of I- and Q-arm of DLI was set to optimum bias condition within 13 iterations. By tuning common and differential controls of DLI, I- and Q-data were recovered simultaneously.
Fig. 4 Measured performances of the automatic bias control scheme; (a) Monitoring voltage for MZMs and MZI and (b) Monitoring voltage for delay interferometer
Figure 5(a) shows the comparison of the two measured BER curves, one obtained by using the proposed automatic bias control scheme and the other by manually adjusting the bias voltages to minimize the BER. The bias voltages of both modulator and demodulator were optimized. There was no difference in the two BER curves and no penalty induced by low-frequency square-wave. The dynamic tracking of biasing point was also evaluated with respect to different OSNR, as shown in Fig. 5(b). The bias control of IQ modulator was independent of signal OSNR, whereas the quality of received signal varied with signal OSNR. However, the bias control of DLI continuously tracked the optimum value regardless of OSNR value after searching optimum bias point. There was no degradation in BER performance. The performance of automatic bias control against temperature change was also investigated, as shown in Fig. 5(c). To change the inside temperature of DQPSK transceiver, the cooling fan was intentionally turned off. In this case, the measured BER was degraded from ~3.5x10−5 to ~1.7x10−2 without bias control of modulator and demodulator. The performance degradation was mostly due to phase drift of DLI. However, the use of automatic bias control suppressed the BER degradation and recovered the original BER level within 35 seconds.
Fig. 5 (a) BER curves of 56 Gb/s DQPSK signal measured by automatic bias control and manually optimizing bias voltages (b) automatic bias tracking measured with different OSNR of signal (c) bias control against temperature variation. We intentionally turned off cooling fan for transceiver to increase inside temperature of DQPSK transceiver
4. Performance evaluation of automatic bias control scheme in field transmission
We implemented the automatic bias control scheme in 112 Gb/s dual carrier-DQPSK transceiver, and evaluated its real-environment long-term performance in field transmission of 112 Gb/s DC-DQPSK signal over 797-km of installed fiber. The DC-DQPSK was composed of two optical carriers, and each carrier was modulated by DQPSK format with 28 Gbaud symbol rate. The transmission experiment was carried out by combining a prototype of 100-Gb/s transponder, ROADM systems, field deployed fibers, and 100GE tester, as shown in Fig. 6(a). A 100GE test set was used to generate 100GE traffic and to analyze the transmission performance. The output of 100GE test set was connected to the 112 Gb/s transponder. The 100GE traffic was wrapped into an OTU4 frame, and connected to DC-DQPSK transceiver. The transmission line was composed of 797 km of SSMF and NZDSF with 9-ps/nm/km dispersion. Span lengths varied from 40 km to 106 km while span loss varied in the range of 17~27 dB. Thus, fiber launching power per carrier was changed at each span from −2 dBm to 8 dBm, as shown in Fig. 6(b). After transmission, the carrier OSNR was measured to be 21.5 dB. The de-multiplexed 112 Gb/s signal was connected to delay interferometer, and the demodulated outputs were converted to electrical signal via balanced photo-detector (BPD). For BER measurement, FEC decoding at the framer was turned off while bit-rate was still 112 Gb/s.
Fig. 6 Field transmission experiment of 112 Gb/s DC-DQPSK signal with automatic bias control (a) experimental setup (b) launching power and loss of each span.
Figure 7(a) shows measured BER of 100GbE traffic. The average BER was measured to be 3.5x10−3, and the BER was quite stable over 19.2 hours. The range of BER variation was very small from 2.4x10−3 to 5.2x10−3. The control stability of modulator and delay interferometer as well as performances of transmission link such as PMD and reflection affected this BER variation. One can see that the control loop is quite robust even in the presence of significant amount of amplitude and phase noise. From these results, we confirmed that the proposed automatic bias control scheme could be used to actual transmitters of advanced modulation formats for long-term operation. To investigate the transmission performance of 112 Gb/s signal, the measured pre-FEC BER was compared with the results of numerical simulation, as shown in Fig. 7(b). The BER-value after 797 km transmission was obtained for various launching power. Since all EDFAs were operated at AGC mode, the fiber launching power of each span was automatically changed as we adjusted the launching power of the first span. The launching power of 112 Gb/s and 10 Gb/s signals were varied from −8 dBm to 6 dBm at each span. The link configuration of numerical simulation was the same as field trial. The results show that the Q-value of field trial was agreed well with the result of numerical simulation. Even when 112 Gb/s DC-DQPSK signal was co-propagated 10 Gb/s NRZ WDM signal neighboring at 100 GHz spacing, the Q-value degradation was as low as 0.5 dB.
Fig. 7 Transmission performance of 112 Gb/s DC-DQPSK signal after 797 km transmission (a) measured pre-FEC BER of 112 Gb/s signal over 18.5 hours (b) comparison of simulated performance of 112 Gb/s DC-DQPSK signal after 797 km transmission
5. Summary
We have demonstrated a simple and cost-effective bias control technique for optical IQ modulator and demodulator. Low frequency (10 kHz) square-wave and RF power of signal were used for IQ modulator, whereas peak voltage detection of demodulated signal was used for IQ demodulator. Neither additional light nor complex control circuit was required in the scheme. The effectiveness of the automatic bias control scheme was demonstrated in 112 Gb/s DC-DQPSK transmitter. The control loop showed that there was no OSNR sensitivity penalty compared with manual optimization of bias conditions, and provided optimal tracking ability against temperature variation. The long-term stability performance was also investigated in real-environment based on field experiment of 112 Gb/s DC-DQPSK signal transmission over 797-km of installed fiber and ROADMs. All the results indicated that the proposed control scheme could be used to actual transmitters of advanced modulation formats for long-term operation.
Acknowledgments
This work was supported by the IT R&D Program of MSIP/KEIT, [10041414, Terabit Optical-Circuit-Packet Converged Switching System Technology Development for Next-Generation Optical Transport Network].
References and links
1. H. S. Chung, S. H. Chang, J. C. Lee, and K. J. Kim., “Dual-Carrier DQPSK based 112 Gb/s signal transmission over 480 km of SMF link carrying 10 Gb/s NRZ channels,” in OFC/NFOEC2012, JW2A.3.
2. H. S. Chung, S. H. Chang, and K. J. Kim, “Effect of IQ mismatch compensation in an optical coherent OFDM receiver,” IEEE Photon. Technol. Lett. 22(5), 308–310 (2010). [CrossRef]
3. K. Sekine, C. Hasegawa, N. Kikuchi, and S. Sasaki, “A novel bias control technique for MZ modulator with monitoring power of backward light for advanced modulation formats,” in OFC2007, OTuH5.
4. H. G. Choi, Y. Takushima, H. Y. Choi, J. H. Chang, and Y. C. Chang, “Modulation format free bias control technique for MZ modulator based on differential phase monitor,” in OFC/NFOEC2011, JWA33.
5. H. Kawakami, E. Yoshida, and Y. Miyamoto, “Auto bias control technique based on asymmetric bias dithering for optical QPSK modulation,” J. Lightwave Technol. 30(7), 962–968 (2012). [CrossRef]
6. H. Kawakami, T. Kobayashi, E. Yoshida, and Y. Miyamoto, “Auto bias control technique for optical 16-QAM transmitter with asymmetric bias dithering,” Opt. Express 21, B308–B312 (2013). [CrossRef] [PubMed]
7. T. Gui, C. Li, Q. Yang, X. Xiao, L. Meng, C. Li, X. Yi, C. Jin, and Z. Li, “Auto bias control technique for optical OFDM transmitter with bias dithering,” Opt. Express 21(5), 5833–5841 (2013). [CrossRef] [PubMed]
8. P. S. Cho and M. Nazarathy, “Bias control for OFDM transmitters,” IEEE Photon. Technol. Lett. 22(14), 1030–1032 (2010). [CrossRef]
9. H. S. Chung, S. H. Chang, J. H. Lee, and K. J. Kim, “Transmission performance comparison of direction -detection based 100 Gb/s modulation formats for metro area optical networks,” ETRI Journal 34(6), 800–806 (2012). [CrossRef]
10. H. Yoon, D. Lee, and N. Park, “Performance Comparison of Optical 8-ary Differential Phase-Shift Keying Systems with Different Electrical Decision Schemes,” Opt. Express 13(2), 371–376 (2005). [CrossRef] [PubMed]
11. Y. Sakamaki, H. Hattri, Y. Nasu, T. Hashimoto, Y. Hashizume, T. Mizuno, T. Goh, and H. Takahashi, “One-chip integrated polarization multiplexed DQPSK demodulator using silica-based planner lighwave circuit technology,” Electron. Lett. 46, 1152–1154 (2010).
12. H. C. Ji, K. J. Park, H. Kim, J. H. Lee, and Y. C. Chung, “A novel frequency-offset monitoring technique for direct-detection DPSK system,” IEEE Photon. Technol. Lett. 18(8), 950–952 (2006). [CrossRef]
