Fibers for Next Generation High Spectral Efficiency Undersea Cable Systems Neal S. Bergano and Alexei Pilipetskii Tyco Electronics Subsea Communications
Presenter Profile Alexei Pilipetskii received his M.S degree in physics from Moscow State University in 1985. From 1985 to 1994 he worked at the General Physics Institute in Russia. He received his Ph.D. in 1990 for research in nonlinear fiber optics. From 1994 to 1997 he was with the University of Maryland Baltimore County, where his interest shifted to fiber optic data transmission. Since 1997 he has been with SubCom, where he works on a number of research and development projects. He is currently the director of a research group focusing on next generation technologies for undersea transmission systems. Alexei Pilipetskii Director - System Modeling & Signal Processing Research email: apilipetskii@subcom.com Tel: 732-578-7533
Fibers for Next Generation High Spectral Efficiency Undersea Cable Systems The importance of new linear fiber types Dispersion management in transoceanic length cable systems 10G DPSK transmission High spectral efficiency systems: Will require polarization multiplexed formats The DP-PSK format with coherent receivers Transmission Fiber Figure-Of-Merit (FOM) New transmission results
Advances in Technology New Fiber Types are Important! 100000 10000 1000 100 10 1 Experimental Transmission Capacity (Gb/s) 2.5 5 10 320 100 640 160 40 1,800 3,730 2,560 6,000 155x100G in 60nm 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Year
Why Fibers are So Important: Linear and Nonlinear Transmission Q (db) 18 17 16 15 14 13 12 OSNR limited Ref: Golovchenko, E.A., et. al., OFC 1999, paper ThQ3 Nonlinearity limited experiment simulation -3 0 3 Relative pre-emphasis (db) One of the first Q vs. power curves published: 16x10Gb/s at 7500 km Nonlinearity limits achievable performance (Q-factor), distance and spectral efficiency
Fibers and Dispersion Management First optically amplifier systems were not designed to support WDM Transmission at or close to zero dispersion Deep dispersion management: NZDSF with SMF Suppressed nonlinear effects between WDM channels Further improvement: large effective area NZDSF (~ 70 µm 2 ) 10 Gb/s bit rates, OOK signals, true WDM transmission Dispersion slope compensated systems Increased transmission bandwidth Large effective area SMF (~105 µm 2 ) Ref: A. Gnauck, et al; IEEE JLT, vol. 26, 2009, p1032 N.S. Bergano, IEEE JLT, Vol. 23, 2005, p 4125
NZDSF Dispersion Map Large Area NZDSF NZDSF Large Area NZDSF NZDSF SMF ~ 500 km Large Area NZDSF 70 µm 2 0.1ps/km -nm 2 NZDSF 55 µm 2 0.07ps/km -nm 2 SMF Dispersion Total NZDSF Wavelength Large Area NZDSF True WDM map Performance may vary with dispersion within the transmission band
Dispersion Flattened Map D+ D- D+ D- D+ ~ 500 km D+ 75/110 µm 2 D- 25-35 µm 2 D+ Dispersion Total Wavelength D- Increased linearity Performance equalized across transmission bandwidth
10G DPSK Transmission in NZDSF Dispersion Map Q fa actor (db) 16 15 14 13 12 11 10 Large dispersion RZ-DPSK Low dispersion RZ-OOK Large dispersion 9 1535 1540 1545 1550 1555 1560 1565 Wavelength (nm) Transmission simulation: 9000 km Ref: W. Anderson, et. al., OFC 2005, OthC1 Experiments: 13000 km, Ref: J.-X. Cai., et. al., OFC 2004, PDP34 Properly-built dispersion map optimizes performance
10G DPSK in Undersea Transmission Success of Dispersion Management Modulation format RZ-OOK (10 Gb/s) RZ-DPSK (10 Gb/s) Channel spacing 33 GHz 33 GHz Fiber plant Amplifier spacing Dispersion Flattened Fibers (DFF) DFF 45 км 75 км System length 9000 км 12700 км Combination 3dB in Rx sensitivity, better nonlinear tolerance in pulse overlapped regimes, and better FEC Very difficult to beat this performance at SE up to 0.4 bit/s/hz
The Need for High Spectral Efficiency Will Require Polarization Multiplexed Formats Q-Facto or [db] 15 13 11 9 7 42.8 Gb/s RZ-DQPSK Pol.Mux.- RZ-DBPSK CSRZ-DBPSK -6-3 0 3 6 Pre-Emphasis [db] Ref: J.-X. Cai et. al OFC 08, PDP4 50x42.8 Gb/s, 5200 km, 66.7GHz channel spacing 150 km repeater spacing Polarization Multiplexed format shows superior performance Higher nonlinear tolerance & higher spectral efficiency Favors coherent polarization multiplexed transponders
Coherent Detection with DSP Ref. A. Salamon, et al., MILCOM 03, M. Taylor, PTL 2004, no. 2, pp. 674 676 Transmission Path PBS LO 90 Optical Hybrid 90 Optical Hybrid A/D A/D A/D A/D DSP Coherent detection allows access to the signal field: Polarization de-multiplexing of PDM signals High spectral efficiency in excess of 1 bit/s/hz Digital signal processing at the receiver
(db) Delta Q conference & convention Dispersion Management with Coherent Technology In-line dispersion compensation 3 2 1 0-1 -2 Optimized for coherent Legacy -2 0 2 4 6 Relative Launch Power (db) can be dropped with coherent Rx No in-line dispersion compensation reduces nonlinear penalties and improves OSNR Very simple transmission line design Difficulty: >10 5 ps/nm need to be compensated in DSP for the long undersea cases
Linear Fibers are Important for Coherent Detection Systems Transition from OOK to PSK with 3 db better receiver sensitivity and better FEC resulted in reduction of required power per channel Helped in managing nonlinearity Transition to PDM modulation formats with 40G and 100G rates will result in higher required OSNR and power per channel We ll need more linear fibers
Advanced Fiber Types Better span losses Better nonlinear tolerance Performanc ce Q(dB) Smaller loss Performance target Difference in span loss (db) Higher loss Performa ance Q(dB) Performance target (A eff1 /A eff2 )[db] A eff1 A eff2 Channel power (db) Channel power (db) FOM ( in db) 10 log( Aeff 1 / Aeff 2) + ( SpanLoss1 SpanLoss2)( db)
0.17dB/km, 150µm 2 FOM 2.5dB 0.14 Improved Fibers 50 km Spans 100 km Spans 180 170 160 150 140 130 120 110 100 0.19dB/km, 105µm 2 90 0.15 0.16 0.17 0.18 0.19 0.20 80 0.17dB/km, 150µm 2 FOM 3.5dB FOM ( in db) 10 log( Aeff 1 / Aeff 2) + ( SpanLoss1 SpanLoss2 )( db) 0.14 180 170 160 150 140 130 120 110 100 0.19dB/km, 105µm 2 90 0.15 0.16 0.17 0.18 0.19 0.20 80
The Push to Higher Spectral Efficiency: New Experimental Results 96x100G over 10600 km (300% spectral efficiency) & 400% spectral efficiency over 4400 km Short 52 km amplifier spacing (better performance) New experimental transmission 12 fiber (0.183 db/km, 150 µm 2 area) 100G on 33GHz (300%) 100G on 25GHz (400%) Post-deadline paper OFC 2010 PDPB10 Q-factor [db] 11 10 9 8 Transmission distance at optimal power 400% SE 0 5000 10000 15000 Transmission Distance [km] 300% SE
Conclusions Advances in fiber technology have played a key role in the past 20 years of long-haul transmission progress 10G RZ-DPSK in combination with modern dispersion management have allowed systems with TB/s capacity per fiber pair Next generation systems will need high spectral efficiency: Higher level modulation formats & higher bit rates Requiring high OSNR/channel Resulting in very high launch powers Fiber performance is very important for next generation systems!
2010 conference & convention The 7th International Conference & Convention on Undersea Telecommunications Pacifico Convention Plaza Yokohama & InterContinental The Grand Yokohama 11 ~ 14 May 2010 www.suboptic.org