Open Access
Add to BibSonomyAbstract
Impairments of inter-symbol interference and beat noise in coherent time-spreading optical code-division-multiple-access are investigated theoretically and experimentally by sweeping the data-rate from 622Mbps up to 10Gbps with 511-chip superstructured fiber Bragg grating. The BER improvement by using optical thresholding technique has been verified in the experiment.
©2005 Optical Society of America
1. Introduction
Optical code division multiple access (OCDMA) is one of the promising candidates for the next-generation broadband access networks other than time-division multiple access (TDMA) and wavelength division multiple access (WDMA) [1]. There would be a possibility to introduce OCDMA into existing gigabit Ethernet passive optical network, so-called GE-PON in the near future, in order to accommodate large number of customers as well as to increase the bandwidth of the uplink. It has unique features of all optical processing, full asynchronous transmission, low latency, soft capacity on demand, protocol transparency, simplified network control as well as increased flexibility of quality-of-service (QoS) control [2–3].
Recently, coherent OCDMA using ultra-short optical pulse is receiving increasing attention with the progress of reliable and compact encoder/decoder devices, such as spatial light phase modulator (SLPM) [4, 5], and superstructured fiber Bragg grating (SSFBG) [6–8]. In coherent OCDMA, encoding and decoding are based on the phase and amplitude of optical field instead of its intensity. The coding can be either directly time-spreading (TS) the ultra-short optical pulse using SSFBG [6–8] or spectral phase-encoded time-spreading (SPECTS) using SLPM [4, 5]. In both cases, the encoded optical pulses will spread in duration of Tcode. The correctly decoded signals (auto-correlation) will be recovered back to a short pulse in SPECTS-OCDMA or spread in duration of 2Tcode in TS-OCDMA, while the incorrectly decoded signals (cross-correlations) spread in durations of Tcode and 2Tcode in SPECTS- and TS-OCDMA, respectively. In the following discussions, we will focus only on TS-OCDMA using SSFBG en/decoder. The SSFBG with phase shift has been used as coherent TS-OCDMA en/decoder with advantages of excellent correlation property, high compactness, and potentially low cost. The SSFBG also exhibits the unique capability to generate ultra-long optical code (OC) [7], which is crucial for OCDMA application that could provide large number of available OC as well as multiple access interferences (MAI) and beat noise suppression [3, 7–9].
Generally saying, the transmission data-rate in OCDMA is restricted by requiring Tbit ≥ 2Tcode to avoid the inter-symbol overlapping [6, 7]. This restriction is illustrated in Fig. 1(a). In this fig., η is the average wing level of the auto-correlation; Tbit is one bit time duration; Tcode=2L FBG/nc is the period of encoded bit “1”, L FBG is the length of SSFBG en/decoder, n is the refractive index of the fiber and c is the light speed in vacuum. For incoherent 2-dimensional OCDMA we have proposed and experimentally demonstrated the data-rate enhancement scheme by challenging the regime of Tbit <2Tcode [9–11]. It has been shown that inter-symbol interference (ISI) noise has the same effect as the multiple-access interference (MAI) and results in the power penalty in the system [9]. In the case of Tbit <2Tcode, as illustrated in Fig. 1(b), the inter-symbol overlapping occurs in the decoded (or even encoded) data stream, which will result in the arising of ISI noises [3, 9]. However, in the coherent TS-OCDMA, the beat between the correctly decoded symbol and the interference bits will arise and predominant in the system [3, 8, 12]. Therefore, the maximum transmission data-rate should be mainly limited by the inter-symbol beat noise.
In this paper, we will experimentally and theoretically investigate the impairments of ISI and inter-symbol beat noise in coherent TS-OCDMA at different data-rate, and show the improvement by using optical thresholding technique.
2. Experimental results and discussion
Figure 2 shows the experimental setup. In the experiment, the 511-chip SSFBG en/decoder (Nchip=511) with binary phase-shift keying Gold code pattern has the chip rate of about 640 Gchip/s, which corresponds to Tcode≈800ps [4]. Therefore, the physically constrained data-rate without symbol overlapping is 622Mbps. The mode-locked laser diode (MLLD) generates optical pulse train with ~10GHz repetition rate and ~2 ps pulsewidth. To change the data rate, the pulse train could be converted into 1/2, 1/4, 1/8 and 1/16 of its original repetition rate by the first LiNiO3-intensity modulator (LN-IM) for 5Gbps, 2.488Gbps, 1.244Gbps as well as 622Mbps transmission, respectively. 223–1 pseudo random binary sequence (PRBS) data patterns at corresponding data-rate were loaded by the second LN-IM modulator. Tunable electrical delay line was used to adjust the relative phase between optical pulse train and data stream. The decoded signal was detected by a 30GHz bandwidth photo-detector and finally measured by the BER tester.
Fig. 2. Experimental setup for demonstration the data-rate flexibility and the limitation with 511-chip SSFBG
Figure 3 shows the measured waveforms of encoded (upper row) and decoded (lower row) signals with different data-rate at Points B and C in Fig. 2, respectively. The symbol overlapping could be observed in the figs. with data-rate higher than 622Mbps. Table 1 summarizes the relationship of Tbit and Tcode, and the number of interference bits m+1 for these data-rates.
Table 1. Number of interference bits for different data-rate
The measured BER performances of different data-rate for one typical 511-chip SSFBG are shown in Fig. 4(a). The opened marks in the fig. are measured at point A in Fig. 2, while the filed marks are measured at Point C. For the decoded signals with data-rates up to 5Gbps, error-free (<10-9) transmissions have been achieved with power penalties comparing to the input signals, while for 10Gbps, BER floor appears at about 10-5. It needs to be note that since we used a different setup for the 10 Gbps signal in the experiment as shown in Fig. 2, therefore the 10G BER curve for the input signal does not follow the same trend as others in this fig.
In Fig. 4(b), the filled circles are the measured results of power penalties vs. data-rate in the experiment. The theoretical results are plotted in the fig. marked by the opened circles as well. In the theoretical calculation, since the ISI has the same effect as MAI in multi-user’s case, the model we have developed in Ref. [3] could be applied here by simply replacing the cross-talk level ξ with η (~1/2Nchip). The ISI distribution is assumed to be Gaussian with variance of m×(Nchip)-2 and the inter-symbol beat noise is also taking as Gaussian. The experimental result agrees reasonably well with the theory: up to 5 Gbps error-free transmission is available with increased power penalty comparing to the input signals. In 10Gbps, error floor appears due to the ISI and inter-symbol beat noise. The discrepancies between theoretical and experimental results are because of that the theoretical calculation is based on an average η value, while the experimental results are taken with Code 2, which is the worst one among the codes we have used and the η is larger than the average value [8].
Figure 5(a) shows the detailed autocorrelation waveforms of 4 different 511-chip codes measured by optical sampling oscilloscope. It is obvious to see that the η level changes from code to code. The above results are taken with code 2, which has higher η than average value. Code 1 has the lowest η because it is one of the preferred M sequences, which is characterized by the perfect auto-correlation profile.
Therefore, the above BER performance should be code dependent. Figure 5(b) shows the measured BERs for the 4 different codes at 10 Gbps. Code 1 has achieved error free while other codes exhibited error floors from 10-7 to 10-5. This result suggests that it is possible to mitigate the inter-symbol beat noise by carefully choosing codes whose auto-correlation wing has dips at the positions overlapping with peaks.
Similar as in the multi-user’s case [3, 8], improvement could also be expected by using optical thresholding technique to reduce ISI [12]. In the experiment, we inserted the super-continuum generation based optical thresholder [8, 13] after OCDMA decoder. The measured BER performances are shown in Fig. 5(c). Here, the different slopes in the BER curves are probably due to differences of the pulsewidth of decoded signal induced by the ISI with different codes. Error free has been achieved for all the 4 codes verifying the improvement of using the optical thresholding. However, the inter-symbol beat noise still remains and dominates the system performance [8].
It needs to be note that in this experiment, we investigated only the TS-OCDMA with single user to focus on the impairment of the ISI. As we have mentioned in the introduction, in the SPECTS-OCDMA, since there is no wing in the auto-correlation [4, 5], the ISI does not exist in single user’s case. However, in multiple users’ case, the ISI issue exists in SPECTS-OCDMA as well as TS-OCDMA. And as it’s shown in the bottom of Fig. 1, the combination of ISI and MAI will result in further performance degradation.
3. Conclusion
In this paper, we investigated the ISI and inter-symbol beat noise in coherent TS-OCDMA. For this purpose, we focused on only single user’s situation to separate ISI and MAI. The maximum transmission data-rate with SSFBG en/decoder is restricted by ISI and inter-symbol beat noise. Data-rate flexibility associated with 511-chip 640Gchip/s SSFBG, whose length physically limits data-rate to 622Mbps to avoid inter-symbol overlapping, is demonstrated. Error-free (BER<10-9) transmissions at data-rate up to 5Gbps have been achieved. However, the ISI and beat noise result in error floor with 10Gbps, which agrees with theoretical prediction. Optical thresholding technique could provide improvement by eliminating the ISI. The inter-symbol beat noise could also be mitigated by carefully choosing codes. However, in multi-user asynchronous OCDMA, these issues could be even worse because of the combination of ISI and MAI.
Fig. 5. Performance of 4 different 511-chip codes at 10 Gbps(a) Autocorrelations measured by optical sampling oscilloscope(b) BER performances without optical thresholder(c) BER performances with optical thresholder
Acknowledgments
The authors would like to thank the anonymous reviewers for their careful reading and valuable comments. The authors would like to thank for Y. Tomiyama and T. Makino of NICT for the technical support in the experiment.
References and links
1 . J. A. Salehi and C. A. Brackett , “ Code division multiple-access techniques in optical fiber networks ,” IEEE Trans. Commun. 37 , 824 – 842 , ( 1989 ). [CrossRef]
2 . K. Kitayama , X. Wang , and H. Sotobayashi , “ State of the art and applications of optical code division multiple access ,” in European Conference on Optical Communication (ECOC’04) , ( Stockholm, Sweden , 2004 ), Tu4.6.1.
3 . X. Wang and K. Kitayama , “ Analysis of beat noise in coherent and incoherent time-spreading OCDMA ,” J. Lightwave Technol. 22 , 2226 – 2235 ( 2004 ). [CrossRef]
4 . Z. Jiang , D. S. Seo , S. D. Yang , D. E. Leaird , R.V. Roussev , C. Langrock , A. M. Fejer , and A. M. Weiner , “ Four-user, 2.5-Gb/s, spectrally coded OCDMA system demonstration using low-power nonlinear processing ”, J. Lightwave Technol. 23 , 143 – 158 ( 2005 ). [CrossRef]
5 . R. P. Scott , W. Cong , K. Li , V. J. Hernandez , B. H. Kolner , J. P. Heritage , and S. J. Ben Yoo , “ Demonstration of an error-free 4×10 Gb/s multiuser SPECTS O-CDMA network testbed ,” IEEE Photonics Technol. Lett. 16 , 2186 – 2188 ( 2004 ). [CrossRef]
6 . P. C. Teh , P. Petropoulos , M. Ibsen , and D. J. Richardson , “ A comparative study of the performance of seven- and 63-chip optical code division multiple-access encoders and decoders based on superstructured fiber Bragg gratings ,” J. Lightwave Technol. 19 , 1352 – 1365 ( 2001 ). [CrossRef]
7 . X. Wang , K. Matsushima , A. Nishiki , N. Wada , F. Kubota , and K. Kitayama , “ High performance optical code generation and recognition using 511-chip 640Gchip/s phase-shifted superstructureed FBG ,” Optics Lett. , 30 , 355 – 357 ( 2005 ). [CrossRef]
8 . X. Wang , N. Wada , T. Hamanaka , A. Nishiki , and K. Kitayama , “ 10-user asynchronous OCDMA transmission experiment with 511-chip SSFBG and SC-based optical thresholder ,” Optical Fiber Communication Conf. (OFC’05 postdeadline) , ( Anaheim, USA , 2005 ), PD 33.
9 . X. Wang and K. T. Chan , “ Enhancement of transmission data rates in incoherent FO-CDMA systems ,” in OptoElectronics Communication Conference (OECC’00) , ( Chiba, Japan , 2000 ), Tech. Dig., 14A2–5,.
10 . S. Kutsuzawa , N. Minato , S. Oshiba , A. Nishiki , and K. Kitayama , “ 10 Gb/s×2 ch signal unrepeated transmission over 100 km of data rate enhanced time-spread/wavelength-hopping OCDM using 2.5-Gb/s-FBG en/decoder ,” IEEE Photon. Tech. Lett. 15 , 317 – 319 ( 2003 ). [CrossRef]
11 . S. Kutsuzawa , N. Minato , K. Sasaki , H. Kobayashi , A. Nishiki , T. Ushikubo , K. Kamijoh , Y. Kamio , N. Wada , and F. Kubota , “ Field demonstration of time-spread/wavelength-hop OCDM using FBG en/decoder ,” Optical Fiber Communication Conference 2005 (OFC /NFOEC 2005) , ( Anaheim, USA , 2005 ), OME77.
12 . J. H. Lee , P. C. Teh , P. Petropoulos , M. Ibsen , and D. J. Richardson , “ A grating-based OCDMA coding-decoding system incorporating a nonlinear optical loop mirror for improved code recognition and noise reduction ”, J. Lightwave Technol. 20 , 36 – 46 ( 2002 ). [CrossRef]
13 . X. Wang , T. Hamanaka , N. Wada , and K. Kitayama , “ Dispersion-flattened-fiber based optical thresholder for multiple-access-interference suppression in OCDMA system ,” Opt. Express 13 , 5499 – 5505 ( 2005 ). [CrossRef] [PubMed]
