Aggregated 17.125Gb/s real-time end-to-end dual-band optical OFDM (OOFDM) transmissions over 25km SSMF IMDD systems with 7dB receiver sensitivity improvements are experimentally demonstrated, for the first time, by utilizing low-cost transceiver components such as directly modulated 1GHz RSOAs and DACs/ADCs operating at sampling speeds as low as 4GS/s. The demonstrated OOFDM transceivers have both strong adaptability and sufficiently large passband carrier frequency tunability, which enable full use of highly dynamic spectral characteristics of the transmission systems. This results in the achievements of not only excellent performance robustness to variations in system operating conditions but also significantly relaxed requirements on RSOA small-signal modulation bandwidth. It is shown that the aforementioned transmission capacity only varies by <23% over a RSOA-injected optical power variation range as large as 20dB, and that the 1GHz RSOAs can support successful transmissions of adaptively modulated OOFDM signals having bandwidths of 8.5GHz. By taking into account the adopted 25% cyclic prefix and a typical 7.3% FEC overhead, the demonstrated real-time OOFDM transmission systems are capable of conveying 11.6Gb/s user data.
© 2014 Optical Society of America
To meet the rapid increase of end-users’ bandwidth requirements , optical orthogonal frequency division multiplexing (OOFDM) is regarded as a promising “future-proof” candidate technology for next-generation passive optical networks (NG-PONs) beyond time-wavelength division multiplexed (TWDM) NG-PON2, as OOFDM offers inherent strong adaptability to device/system/network imperfections, rich digital signal processing (DSP)-enabled transceiver intelligence and networking functionalities, low-cost potential and good compatibility with existing PONs . In addition to the aforementioned salient features, centrally-controllable universal transceivers are also achievable if use is made of reflective intensity modulators (IMs) in optical network units (ONUs) of OOFDM PONs. To implement reflective intensity modulator-based cost-effective ONUs, the following reflective IMs may be utilized, including, for example, reflective semiconductor optical amplifiers (RSOAs) , reflective electro-absorption modulators (REAMs) [4, 5] and reflective Fabry-Perot lasers [6, 7].
Since RSOA-IMs have the unique advantages including colorlessness, cost-effectiveness, compactness, low power dissipation, large-scale monolithic integration capability and simultaneous functionalities of signal modulation and amplification [3, 8], over the past several years, extensive experimental investigations of the performances of RSOA-IM-based OOFDM PON systems have been reported using offline DSPs, based on which 40Gb/s OOFDM transmissions over 26km standard single-mode fibers (SSMFs) have been achieved .
Making use of a 1.125GHz RSOA-IM, we have recently demonstrated experimentally 7.5Gb/s real-time end-to-end 16-QAM-encoded single-band OOFDM (SB-OOFDM) transmissions over 25km SSMF systems incorporating intensity modulation and direct detection (IMDD) . Compared to SB-OOFDM, multi-band OOFDM (MB-OOFDM) not only considerably reduces the equipment inventory, but also significantly relaxes the requirements of both the key transceiver component bandwidths and requisite DSP complexity [11, 12]. In addition, adaptive bit and power loading on each subcarrier involved in each individual OFDM sub-band can also be applied together with appropriate adjustments of the electrical signal powers of all the sub-bands and the operating conditions of the RSOA-IMs. Furthermore, the RF frequency of an OFDM passband is also tunable according to the system frequency response to ensure that the sub-band always occupies the optimum spectral region regardless of the RSOA-IM operating conditions. The combination of the above-mentioned MB-OOFDM features ensures full use of the available system frequency responses far beyond commercially available RSOA’s 3-dB small-signal modulation bandwidth specified by the manufacturer. Therefore, this offers, in a cost-effective manner, excellent opportunity for further maximizing the aggregated signal transmission capacity and simultaneously improving the performance robustness and transceiver flexibility.
In this paper, we report, for the first time, record-high 17.125Gb/s real-time end-to-end dual-band adaptive OOFDM transmissions over 25km SSMFs with 7dB receiver sensitivity improvements in simple IMDD systems incorporating 1.125GHz RSOA-IMs and 4GS/s digital-to-analogue converters (DACs) and analogue-to-digital converters (ADCs). Experiments also show that the RSOA-IMs can intensity modulate adaptively bit/power-loaded electrical OFDM signals having bandwidths approximately 8 times higher than the 3-dB small-signal modulation bandwidths of the RSOAs. The 1GHz OFDM passband tunability is also experimentally demonstrated, which corresponds to receiver sensitivity variations of <1dB. In addition, excellent robustness of the 17.125Gb/s over 25km SSMF OOFDM transmission performance is also observed over a 20dB dynamic variation range of the CW optical powers injected into the RSOA-IMs.
2. Real-time dual-band OOFDM systems employing directly modulated RSOA-IMs
The RSOA-IM-based real-time end-to-end dual-band OOFDM IMDD SSMF system considered here is shown in Fig. 1, where core field programmable gate array (FPGA)-based OFDM DSP functions and transceiver architectures can be found in [11, 12]. The adopted key transceiver/system parameters are presented in Table 1. For each OFDM sub-band in the transmitter, the real-time DSP functionalities include on-line adaptive bit and power loading of 15 data-carrying subcarriers, on-line adjustable signal clipping level and a 32-point inverse fast Fourier transform (IFFT) with input data being arranged to satisfy the Hermitian symmetry. On the other hand, for each sub-band in the receiver, the corresponding real-time DSP functions consist of automatic symbol alignment, a 32-point FFT, channel equalization and adaptive demodulation.
To simultaneously generate two separate OFDM sub-band signals, independent digital and RF electronics are utilized in each individual sub-band: one electrical sub-band, referred to as the baseband, occupies a spectral region varying from 0 to 2GHz, whilst the second electrical sub-band, termed the passband, is produced by modulating a 6.5GHz RF carrier with another 0-2GHz OFDM baseband signal. This leads to the generation of a double sideband (DSB) passband occupying a spectral region from 4.5 to 8.5GHz.
As seen in Fig. 1, at the transmitter side, the real-valued baseband signal emerging from the 4GS/s@8bits DAC is first amplified and its power is appropriately adjusted. Simultaneously, for the passband, an OFDM signal from another 4GS/s@8bits DAC is also processed via amplification, RF up-conversion and spectral filtering. The two sub-bands are finally combined in a low loss RF multiplexer.
After passing through an optical circulator, a 1550nm CW optical wave supplied by an external cavity laser source is injected, at an optical power of 4.56dBm, into a polarization-insensitive RSOA-IM with a 3dB small-signal modulation bandwidth of 1.125GHz and a 3-dB input optical saturation power of approximately −10dBm. The 8.5GHz 3.4 Vpp electrical analog dual-band OFDM signal and an 80mA DC bias current are combined in a 26.5GHz bias tee and then modulate the CW optical wave in the RSOA-IM operating at a temperature of 13þC. Such a RSOA temperature is not critical for obtaining the system performances presented in the following sections of the paper. The output optical signal is transmitted through a 25km SSMF.
In Fig. 1, the utilization of the local optical light source rather than a remotely injected CW optical wave is to provide sufficiently high optical powers to enable extensive experimental explorations of two key physical mechanisms of interest of the paper: a) strong optical gain saturation-induced enhancements in RSOA-IM modulation bandwidth; and b) maximum RSOA-injected optical power variation ranges, over which the robust OOFDM transmission performance is still practically obtainable. An in-depth understanding of the aforementioned two physical mechanisms provides useful insights into the practical design of RSOA-IM-based colorless ONUs. It should be pointed out that, to practically realize RSOA-IM-based colorless ONUs, remotely injected optical light sources are necessary, whose powers in the ONUs are, however, too low to heavily saturate the RSOA-IM. To address such a challenge, use may be made of an independent SOA in each ONU to amplify the remotely injected optical wave prior to performing optical modulation in the RSOA-IM.
At the receiver, the optical signal passes through a variable optical attenuator to adjust the received optical power (ROP). A 12GHz PIN + TIA or a 10GHz optical filter followed by an 8.5GHz APD is employed to convert the received dual-band OOFDM signal into the electrical domain. To recover the baseband (passband) signal, a down-conversion circuit identical to that reported in [11, 12] is omitted (included). The received baseband signal or the down-converted passband signal passes to the receiver block, where the digital signal emerging from the 4GS/s@8 bit ADC is recovered using the inverse OFDM DSP processes compared to their transmitter counterpart. For both mixers in the passband transmitter and receiver, the local oscillator (LO) signals are derived from the same RF signal source. A variable delay line is employed to correctly align the phase of the receiver’s LO with the phase of the received RF carrier. If an independent LO is employed in the down-conversion circuit of the receiver, the achievement of transmission performances similar to those presented in the paper is still possible when RF carrier frequency offset and phase offset estimations together with sufficiently accurate frequency and phase compensating functionalities are introduced into the receiver. Given the fact that the RF up-conversion and down-conversion circuits have standard configuration complexity and employ off-the-shelf electronics, their cost is, therefore, expected to be low in high volume for cost-sensitive applications.
3. 17.125Gb/s over 25km SSMF transmission performance
To gain a better understanding of the highly dynamic nature of the frequency responses of the RSOA-IM-based SSMF IMDD systems, experimental measurements are first undertaken of the frequency responses of different system configurations operating at various conditions. For the RSOA-IM operating at a bias current of 80mA and a CW injection power of 4.5dBm, Fig. 2(a) shows that when the RSOA-IM is introduced into the 25km SSMF IMDD system, a deep system frequency response lobe occurs, and beyond the lobe’s dip frequency, a second system frequency response peak is also observed, whose amplitude value is higher than the stand-alone RSOA-IM at the same frequency. Such behaviors arise due to the combined effects of fibre chromatic dispersion, intensity modulation-induced RSOA frequency chirp and direct detection in the receiver . When the RSOA-IM is strongly saturated by injecting high optical powers, the corresponding RSOA frequency chirp decreases , thus resulting in the shifting of the dip of the lobe and the slightly lowering of the second system frequency response peak, as seen in Fig. 2(b). In addition, Fig. 2(b) also shows that a high optical injection power shortens the effective RSOA carrier lifetime, thus giving rise to an enhanced 3-dB modulation bandwidth. This agrees very well with theoretical predictions .
Figures 2(a) and 2(b) indicate that, to maximise the aggregated signal transmission capacity and simultaneously improve the performance robustness against variations in both RSOA-IM operating conditions and system configurations, it is extremely critical if use is made of the dual-band OOFDM transceiver with the passband being located around the second system frequency response peak region. It should also be noted that adaptive bit and power loading on each subcarrier involved in the individual OFDM sub-bands is applied together with appropriate adjustments of the electrical signal powers of these two sub-bands. This ensures full use of the available system frequency responses. Furthermore, the RF frequency of the passband is also tunable according to the system operating conditions to ensure that the passband always occupies the optimum region of the system frequency response.
Having discussed the dynamic spectral characteristics of the transmission systems, attention is then given to exploring the maximum transmission performance obtained under the optimum system parameters listed in Table 1. Figure 3(a) shows the normalized frequency response of each sub-band for the entire RSOA-IM-based 25km SSMF IMDD system ranging from the transmitter IFFT input to the receiver FFT output. For the passband case, an effective system frequency response is determined as the effect of passband transmission-induced relative subcarrier attenuation is considered. It can be seen in Fig. 3(a) that similar frequency response roll-offs of approximately 20dB are observed for high frequency subcarriers of both sub-bands. The observed system frequency roll-offs are mainly attributed by the electrical transceiver components [11, 12].
The utilization of adaptive bit and power loading on the 15 information-bearing subcarriers of each sub-band is very effective in compensating for the aforementioned large system frequency response roll-off effect. The resulting adaptively loaded/received subcarrier power profiles and the corresponding online optimized bit loading profiles are presented in Figs. 3(a) and 3(b), respectively. The inter-sub-band intermixing effect and the finite FPGA and DAC power dynamic range-induced imperfect compensation of the system frequency response roll-off leads to the completely dropping of the last 4 subcarriers in the passband, as shown in Fig. 3 (a) and 3(b). Based on Fig. 3(b) and Table 1, it can be easily worked out that the achieved signal bit rates for the baseband and passband are 11.125Gb/s and 6Gb/s, respectively. This gives rise to an aggregated signal transmission capacity of 17.125Gb/s, of which 11.6Gb/s can be utilized to convey user data when taking into account the adopted 25% cyclic prefix and a typical 7.3% FEC overhead.
For both the optical back-to-back and entire 25km SSMF IMDD system configurations, Fig. 4(a) presents the measured bit error rate (BER) performances of both sub-bands as a function of ROP for the cases of employing the PIN and the APD. It can be seen in Fig. 4(a) that, in comparison with the passband, the relatively flat BER developing curves occur for the baseband at high ROPs. This is due to the fact that unwanted inter- and intra-sub-band intermixing frequency products generated upon square-law photon detection in the receiver are predominantly located in the baseband spectral region. In addition, comparing the baseband with the passband, the observed difference in ROP at the adopted FEC limit of 4 × 10−3 for the PIN back-to-back case is a direct result of the residual frequency response roll-off effect caused by the imperfect frequency response roll-off compensation mentioned previously. Moreover, compared to the passband, the large baseband power penalty for the PIN case is mainly contributed by the effect of chromatic dispersion-enhanced RSOA frequency chirp , which is pronounced for the baseband because of its large signal transmission capacity. Furthermore, Fig. 4(a) also shows that the replacement of the PIN by the APD results in an approximately 7dB improvement in receiver sensitivity for both sub-bands. It is also very interesting to note that very similar BERs for both sub-bands occur in Fig. 4(a), indicating that the RSOA-IM can support successful transmissions of adaptively modulated OOFDM signals having their bandwidths far beyond their 3-dB small-signal modulation bandwidths.
The corresponding subcarrier error distributions for both the baseband and passband after 25km SSMF transmission are plotted in Fig. 4(b), in obtaining which experimental measurements are undertaken at their maximum ROPs shown in Fig. 4(a) for the PIN case. In Fig. 4(b) it is clearly observed that the use of adaptive bit and power loading can result in almost uniform BER distributions across all the subcarriers, whose BER variations are as low as ± 8.8% for the baseband and ± 14.7% for the passband. The relatively high BER distribution fluctuation for the passband is due to the residual frequency response roll-off effect discussed above.
The received constellations of the representative subcarriers recorded prior to channel equalization are plotted in Fig. 5 for the baseband and passband. The constellations are measured at their minimum sub-band BERs after transmissions over 25km SSMF for the PIN case. As expected from the results presented in Fig. 3(a), large variations in subcarrier amplitude levels are clearly seen for all these constellations.
It should be noted that decreasing the transmission distance to lengths of <25km causes the system frequency response lobe to reduce its’ depth very quickly and simultaneously shift its’ dip frequency beyond the dual-band OFDM spectral range. The utilization of both adaptive bit/power loading and passband RF frequency tuning ensures that the aggregated 17.125Gb/s signal capacity is almost transmission distance-independent, and that the corresponding power penalties for both sub-bands lower with reducing transmission distance.
4. Transmission performance robustness
From the discussions in Section 3, it is clear that adaptive bit and subcarrier/sub-band power loading offers a simple and effective means of compensating for the system frequency response roll-off effect. Apart from the maximization of the aggregated transmission capacity, the technique also considerably enhances the transceiver flexibility and the performance robustness against variations in RSOA-IM operating conditions, passband RF carrier frequencies and system configurations.
The representative spectra of the dual-band OOFDM signals measured at the output of the PIN + TIA for different passband carrier frequencies are shown in Fig. 6, where the RSOA-IM operates at a bias current of 80mA and a CW optical injection power of 4.5dBm. It can be seen in Figs. 6(a)-6(d) that the developing trends of the passband spectral envelops follow exactly the corresponding system frequency responses illustrated in Fig. 2. The measured baseband and passband BER performances are plotted in Fig. 7 for different passband carrier frequencies. For all these passband carrier frequencies considered, it is shown in Fig. 7 that the corresponding BER curves are very similar to those shown in Fig. 4(a), and that a 1GHz frequency detuning range corresponds to ROP variations of <1dB. This indicates that the present transceiver has at least 1GHz passband carrier frequency tunability. For the dual-band OOFDM transceiver designs, the practically achievable lower boundary of the passband carrier frequency detuning range is mainly determined by the bandwidth of the baseband because the cross-talk effect between these two sub-bands has to be minimized; Whilst the practically achievable upper boundary of the passband carrier frequency detuning range is mainly limited by the roll-off effect of the entire system frequency responses mainly associated with both the bandwidths of the involved RF components such as the RF filters, and the RSOA-IM modulation and frequency chirp characteristics.
To explore the impact of the RSOA-IM gain saturation effect on the system performance, the maximum achievable signal transmission capacities of both sub-bands versus CW optical power injected into the RSOA-IM are plotted in Fig. 8 for the back-to-back and 25km SSMF IMDD system configurations. In obtaining Fig. 8, for a specific optical injection power, extensive optimizations of all other parameters are undertaken until the maximum signal transmission capacity for each individual sub-band is achieved at the adopted FEC limit. These parameters include subcarrier bit/power, electrical sub-band power, RSOA bias/driving current and passband carrier frequency.
It is very interesting to note in Fig. 8 that an almost flat baseband signal capacity is obtainable for a CW optical power variation range as large as 20dB (from 10dBm to −10dBm), over which the passband signal capacity, however, decreases by a factor of approximately 3. The passband capacity reduction is mainly because the less optical gain saturation-induced increase in carrier lifetime narrows the RSOA modulation bandwidths and simultaneously increases the frequency chirp. These effects affect the passband performance more severer than the baseband. It is shown in Fig. 8 that the aggregated signal transmission capacity only varies by <23% over the RSOA-injected optical power variation range as large as 20dB. The observed optical power variation range is approximately 15dB higher than those corresponding to RSOA-based 10Gb/s non-return-zero (NRZ) transmission systems reported in [16, 17].
Aggregated 17.125Gb/s over 25km SSMF real-time end-to-end dual-band adaptive OOFDM transmissions with 7dB receiver sensitivity improvements have been experimentally demonstrated, for the first time, in simple IMDD systems incorporating low-cost 1.125GHz RSOA-IMs and 4GS/s DACs/ADCs. The transceiver’s strong adaptability and large passband carrier frequency tunability enable full use of highly dynamic system spectral characteristics regardless of system operating conditions. This results in the achievements of not only excellent performance robustness to variations in system operating conditions but also significantly relaxed requirements on RSOA small-signal modulation bandwidth. It has been shown that the abovementioned aggregated signal transmission capacity only varies by <23% over a RSOA-injected optical power variation range as large as 20dB, and that the RSOAs can support successful transmissions of adaptively modulated OOFDM signals having bandwidths 8 times higher than their 3-dB small-signal modulation bandwidths.
This work was supported by the PIANO + under the European Commission’s ERA-NET Plus Scheme within the project OCEAN under Grant Agreement 620029, in part by the National Natural Science Foundation of China (61132004) and Shanghai Science and Technology Development Funds: 13JC1402600.
References and links
1. P. Vetter, “Next generation optical access technologies,” in Proceedings of European Conference and Exhibition on Optical Communication (Amsterdam, Netherlands, 2012), paper Tu.3.G.
2. N. Cvijetic, M. Cvijetic, M. Huang, E. Ip, Y. Huang, and T. Wang, “Terabit optical access networks based on WDM-OFDMA-PON,” J. Lightwave Technol. 30(4), 493–503 (2012). [CrossRef]
3. C. H. Yeh, C. W. Chow, H. Y. Chen, and Y. F. Wu, “10-Gbps OFDM upstream rate by using RSOA-ONU with seeding-light for 75 km long-reach PON access,” in Proceedings of Optical Fiber Communication Conference and Exposition and the National Fiber Optic Engineers Conference (California, 2012), paper JTh2A.65. [CrossRef]
4. S. C. Lin, S. L. Lee, C. K. Liu, C. L. Yang, S. C. Ko, T. W. Liaw, and G. Keiser, “Design and demonstration of REAM-based WDM-PONs with remote amplification and channel fault monitoring,” J. Opt. Commun. Netw. 4(4), 336–343 (2012). [CrossRef]
5. Q. Guo and A. V. Tran, “Demonstration of 40-Gb/s WDM-PON system using SOA-REAM and equalization,” IEEE Photon. Technol. Lett. 24(11), 951–953 (2012). [CrossRef]
6. Z. Xu, Y. Yeo, X. Cheng, and E. Kurniawan, “20-Gb/s injection locked FP-LD in a wavelength-division-multiplexing OFDM-PON,” in Proceedings of Optical Fiber Communication Conference and Exposition and the National Fiber Optic Engineers Conference (California, 2012), paper OW4B.3. [CrossRef]
7. D. J. Shin, D. K. Jung, H. S. Shin, J. W. Kwon, S. Hwang, Y. Oh, and C. Shim, “Hybrid WDM/TDM-PON with wavelength-selection-free transmitters,” J. Lightwave Technol. 23(1), 187–195 (2005). [CrossRef]
8. T. Duong, N. Genay, P. Chanclou, B. Charbonnier, A. Pizzinat, and R. Brenot, “Experimental demonstration of 10 Gbit/s for upstream transmission by remote modulation of 1 GHz RSOA using adaptively modulated optical OFDM for WDM-PON single fiber architecture,” in Proceedings of European Conference and Exhibition on Optical Communication (Brussels, 2008), Paper Th.3.F.1. [CrossRef]
9. G. Cossu, F. Bottoni, R. Corsini, M. Presi, and E. Ciaramella, “40 Gb/s single R-SOA transmission by optical equalization and adaptive OFDM,” IEEE Photon. Technol. Lett. 25(21), 2119–2122 (2013). [CrossRef]
10. R. P. Giddings, E. Hugues-Salas, X. Q. Jin, J. L. Wei, and J. M. Tang, “Experimental demonstration of real-time optical OFDM transmission at 7.5 Gb/s over 25-km SSMF using a 1-GHz RSOA,” IEEE Photon. Technol. Lett. 22(11), 745–747 (2010). [CrossRef]
11. E. Hugues-Salas, R. P. Giddings, and J. M. Tang, “First experimental demonstration of real-time adaptive transmission of 20Gb/s dual-band optical OFDM signals over 500m OM2 MMFs,” in Proceedings of Optical Fiber Communication Conference and Exposition and the National Fiber Optic Engineers Conference (California, 2012), paper OTh3A.1.
12. Q. W. Zhang, E. Hugues-Salas, R. P. Giddings, M. Wang, and J. M. Tang, “Experimental demonstrations of record high REAM intensity modulator-enabled 19.25Gb/s real-time end-to-end dual-band optical OFDM colorless transmissions over 25km SSMF IMDD systems,” Opt. Express 21(7), 9167–9179 (2013). [CrossRef] [PubMed]
13. L. A. Neto, D. Erasme, N. Genay, P. Chanclou, Q. Deniel, F. Traore, T. Anfray, R. Hmadou, and C. Aupetit-Berthelemot, “Simple estimation of fiber dispersion and laser chirp parameters using the downhillSimplex fitting algorithm,” J. Lightwave Technol. 31(2), 334–342 (2013). [CrossRef]
14. J. L. Wei, A. Hamié, R. P. Giddings, and J. M. Tang, “Semiconductor optical amplifier-enabled intensity modulation of adaptively modulated optical OFDM signals in SMF-based IMDD systems,” J. Lightwave Technol. 27(16), 3678–3688 (2009). [CrossRef]
15. X. Zheng, X. Q. Jin, R. P. Giddings, J. L. Wei, E. Hugues-Salas, Y. H. Hong, and J. M. Tang, “Negative power penalties of optical OFDM signal transmissions in directly modulated DFB laser-based IMDD systems incorporating negative dispersion fibers,” IEEE Photon. J. 2(4), 532–542 (2010). [CrossRef]
16. I. Papagiannakis, M. Omella, D. Klonidis, J. Kikidis, A. N. Birbas, I. Tomkos, and J. Prat, “Upstream transmission in WDM PONs at 10Gb/s using low bandwidth RSOAs assisted with optical filtering and electronic equalization,” in Proceedings of European Conference and Exhibition on Optical Communication (Brussels, Belgium, 2008), paper We.3.F.3.
17. I. Cano, M. Omela, J. Prat, and P. Poggiolini, “Colorless 10Gb/s extended reach WDM PON with low BW RSOA using MLSE,” in Proceedings of Optical Fiber Communication Conference and Exposition and the National Fiber Optic Engineers Conference (California, 2010), paper OWG2.