10 Gbit/s optical CDMA encoder-decoder BER performance using HNLF thresholder Kebin Li, Wei Cong, V. J. Hernandez, Ryan P. Scott, Jing Cao, Yixue Du, J. P. Heritage, Brian H. Kolner, S. J. B. Yoo Department of Electrical and Computer Engineering, University of California, Davis, California 95616
[email protected]
Abstract: We demonstrate bit-error-rate performance of a 10 Gbit/s optical code division multiple access system based on a spectral-phase encoding scheme. A highly nonlinear fiber based thresholder discriminates between correctly and incorrectly decoded pulses. 2004 Optical Society of America
OCIS codes: (060.4510) Optical Communication; (060.2330) Fiber optics communication
1. Introduction Conventional optical access networks typically use wavelength-division multiplexing (WDM) and/or time-division multiplexing (TDM) techniques which require wavelength or time domain processing. For RF wireless networks, code-division multiple access (CDMA) has become the dominant approach because of significant advantages. For high-capacity access networks, optical code-division multiple access (O-CDMA) may provide significant rewards. Instead of relying on WDM or TDM technologies alone, O-CDMA can utilize optical codes to achieve truly flexible and reconfigurable access of large network capacity. In addition, it has the potential for achieving enhanced security in the optical layer, decentralized network control, and increased reliability and survivability. Different O-CDMA schemes have been proposed using various laser sources, coding schemes and detection methods [1-3]. The Spectral Phase Encoded Time Spreading (SPECTS) scheme exploits relatively simple all-optical pulse shaping to achieve optical encoding and decoding of information. The pulse shaping action of the encoder on ultrashort pulses spreads the pulse in time, and the decoder can reconstruct the original pulse if the matching code is used. Otherwise, the decoder will generate another pulse spread in time that will appear as low intensity noise-like bursts. By introducing optical threshold detection, properly decoded pulses can be selected against the noise. Several different approaches for optical thresholding have been proposed [4-6], however most of them require high peak pulse powers incompatible with current fiber optic communication technologies operating at multigigabit/s data rates. Here, we demonstrate the effectiveness of using highly non-linear fiber (HNLF) in one of these schemes [5] instead of DSF fiber to increase the sensitivity of the thresholder by more than 40 times in order to measure the bit-error-rate (BER) at the 10Gbit/s data rate. 2. Experimental setup SPECTS Encoder/Decoder M1
Fiber
Synthesizer Grating
PPG 10 Gb/s Optical Clock
Data-out Clk-in BERT Clk-in Data
Modulator
EDFA
SLM
L1
M2
L2
Pulse Compressor
10 Gb/s Receiver Clk-out Data
HNLF
EDFA
DCF
Fiber
Grating
SPECTS Encoder SPECTS Decoder
Thresholder Fig. 1. Diagram of experimental setup for Bit-Error-Rate measurement of SPECTS O-CDMA testbed. EDFA; Erbium Doped Fiber Amplifier, DCF; Dispersion Compensation Fiber, HNLF, Highly Nonlinear Fiber, L1 & L2; Lenses, M1 & M2; Mirrors.
The experimental setup is shown in Fig. 1. The mode-locked fiber laser (optical clock) generates 2.4 ps pulses centered at 1550 nm with a repetition rate of 10 GHz. A synchronized LiNbO3 Mach-Zehnder modulator modulates the pulse train at 10 GHz with a 223 – 1 PRBS. The modulated pulses are amplified by an Erbium doped fiber
amplifier (EDFA), and then compressed to 0.4 ps by a nonlinear-fiber-based pulse compressor. This pulse train is spectrally phase encoded by a pulse shaper and is then decoded by an identical pulse shaper, followed by a threshold detection system composed of a dispersion compensated EDFA and 500 m of highly nonlinear fiber (HNLF). Bit error rate testing was performed using a 10 Gb/s receiver with clock recovery and a HP 70843B test set. The inset in Fig.1 shows the details of a SPECTS encoder or decoder. A spatial light phase modulator (SLM) is located in the Fourier plane of a zero dispersion pulse compressor consisting of a pair of diffraction gratings and lenses. This arrangement allows the individual spectral components to be phase modulated (0 or π phase shift) by the SLM with a m-sequence code. 3. Results and discussion To verify the proper operation of the encoder and decoder, we made cross correlation measurements at several key locations in the system, as shown in Fig.2. Notice that the correctly decoded pulse is only slightly broadened when compared with the input pulse (Fig. 2a). While the encoded (Fig. 2b) or incorrectly decoded pulse (Fig. 2c) is broadened approximately 30 times. 0.6
8
FWHM 0.41 ps
6
FWHM 0.58 ps
4 2 0
1.2
b
0.5
Amplitude ( a.u.)
a
Amplitude (a.u.)
Amplitude (a.u)
10
0.4 0.3 0.2 0.1 0.0
0
1
2
3
4
Time Delay (ps)
5
c
1.0 0.8 0.6 0.4 0.2 0.0
0
10
20
30
40
Time Delay (ps)
50
0
10
20
30
40
Time Delay (ps)
Fig. 2. Cross correlation traces (using 0.4 ps pulse as the reference) showing the optical pulse state at various points in the system. (a) Input pulse to the encoder (gray) and decoder output pulse (black). (b) Encoded pulse with a 31 bit m-sequence on the SLM of encoder. (c) Incorrectly decoded output pulse from the decoder.
Even though the incorrectly decoded pulse is very broad, it will be interpreted the same as the short pulse in typical optical receivers because of their limited temporal responses. This mandates the use of a threshold detector which acts as a discriminator. The threshold detector takes advantage of fiber non-linearity which generates additional frequency components through self-phase modulation for shorter (higher peak power) pulses but not for longer ones. Figure 3a shows the input spectrum to the encoder and the decoder output spectrum. The narrowing of the output spectrum is caused by spectral windowing from the SLM. After passing through the HNLF, the spectrum is significantly broadened (Fig.3b) with the properly decoded pulse showing significantly more spectrum at the longer wavelength when compared to the improperly decoded pulse (rejected) spectrum. When a 1 nm bandpass filter is used at 1561 nm, we measured contrast ratios in excess of 7 dB. Figure 3c shows the output spectra from the thresholder for correctly and incorrectly decoded pulses. Fig. 4 shows the bit error rates for the back-to-back signal and the correctly decoded signal that passes through the HNLF-based thresholder. The inset shows the eye diagrams of back-to-back, correctly decoded, and incorrectly decoded signals. It clearly shows that the eye diagram of the correctly decoded signal is open widely while the eye of the incorrectly decoded signal is completely closed. The degradation of the incorrectly decoded signal prevents synchronization for a BER measurement at the available powers. In contrast, less than 10-11 BER was achieved with a correctly decoded signal. The BER of the incorrectly decoded signal is around 0.4 due to the failure of synchronization while the BER of correctly decoded signal below 10-11.
Spectral Power (dBm)
-10 -15 Rejected a -10 b c Encoder -20 -50 Decoded Input -25 -15 -30 -60 -20 -35 -40 -70 -25 Decoder Rejected -45 Output -30 Decoded -50 -80 -55 -35 1520 1540 1560 1580 1520 1540 1560 1580 1520 1540 1560 1580 Wavelength (nm) Wavelength (nm) Wavelength (nm) -40
-5
Fig. 3. Optical spectra at various locations in the system. (a) The encoder input spectrum and the decoder output spectrum. (b) Output spectra of the 500 m long HNLF for correctly and incorrectly decoded pulses. (c) Output spectra of the thresholder for correctly and incorrectly decoded pulses.
Back-to-back
Incorrectly decoded
Correctly decoded
Fig. 4. Bit Error Rate and eye diagrams for back-to-back and decoded signals
4. Conclusion This paper presented 10 Gb/s results achieved on a SPECTS O-CDMA testbed. The experimental results showed successful encoding and decoding, and selective detection of correctly decoded pulses, while rejecting the incorrectly decoded pulses by using a HNLF thresholder. The 10 Gb/s BER were below 10-11 with reasonably low power detection of -14 dBm. Multiple user access network BER performance study is in progress. 5. References [1] J. A. Salehi, A. M. Weiner and J. P. Heritage, “Coherent ultrashort light pulse code-division multiple access communication systems,” J. Lightwave Technol. 8, 478-491 (1990). [2] H. P. Sardesai, C. C. Chang and A. M. Weiner, “A femtosecond code-division multiple-access communication system test bed,” J. Lightwave Technol. 16, 1953-1964 (1998). [3] H. Tsuda, H. Takenouchi, T. Ishii, K. Okamoto, T. Goh, K. Sato, A. Hirano, T. Kurokawa and C. Amano, “Spectral encoding and decoding of 10 Gbit/s femtosecond pulses using high resolution array-waveguide grating,” Electron. Lett. 35, 1186-1188 (1999). [4] D. T. Reid, M. Padgett, C. McCowan, W. Sleat, W. Sibbett, “Light emitting diodes as measurement devices for femtosecond laser pulses,” Opt. Lett. 22, 233-235 (1997). [5] H. P. Sardesai and A. M. Weiner, “Nonlinear fiber-optic receiver for ultrashort pulse code division multiple access communication,” Electron. Lett. 33, 610-611 (1997). [6] Z. Zheng, A. M. Weiner, K.R. Parameswaran, M. H. Chou and M. M. Fejer, “Low-power spectral phase correlator using periodically poled LiNbO3 Waveguide,” J. Photonics Technol. Lett. 18, 376-495 (1997).
This work was supported in part by DARPA and SPAWA under agreement number N6601-02-1-8937..
