DispersionManaged Solitons

1
Dispersion-Managed Solitons Chirped RZ What
is the Difference?
presented by Curtis R. Menyuk
University of Maryland Baltimore County Computer
Science and Electrical Engineering
Department Baltimore, MD 21250
2
Contributers
G. M. Carter UMBC E. A. Golovchenko Tyco V.
S. Grigoryan UMBC J. M. Jacob MITRE D.
Marcuse Retired R. Mu Alphion A. N.
Pilipetskii Tyco A. Wai Hong Kong
Polytechnic D. Wang 3M
3
What is a soliton?
A pulse that propagates without dispersing
And why is that important?
Dispersion leads to pulse spreading and
intersymbol interference
4
Chromatic dispersion
Example of intersymbol interference
0
1
1
1
0
0
1
1
propagation distance
Why does this happen? Different
frequency components travel atdifferent speeds
5
Chromatic dispersion
In the frequency domain where is the
Fourier transform of the field envelope Eac
h frequency evolves independently!


Fourier transform
inverse Fourier transform
6
Chromatic dispersion FWHM 20 ps D 0.16
ps/nm-km (normal)
7
Chromatic dispersion FWHM 20 ps D 0.16
ps/nm-km (normal)
Total length 10,000 km
8
Solitons
Optical fibers are nonlinear Soliton solutions
exist
The nonlinearity and the dispersion balance! ?
nonlinear distortion is avoided
9
Solitons
By the early 1990s linear dispersion
compensating modules exist
BUT Nonlinear distortion is still important
Solitons use dispersion to compensate for
nonlinearity!
10
Solitons

• They do have issues
• They must be well separated (5 x pulse
  duration) ? format is spectrally inefficient
• Their central time shifts change during
  collisons ? solitons can jitter out of their
  time windows

11
What format should be used?
In the early 1990s, the contenders were
Non-return to zero (NRZ)
(close to zero dispersion)
Solitons
(anomalous dispersion)
1 1 0 1 0 0 1
12
By the late 1990s Approaches converge
Solitons and NRZ resemble each other
solitons dispersion-managed solitons
NRZ phase- and
amplitude-modulated pulses
13
Single-channel systems with moderate dispersion
14
Dispersion-managed solitons
GOAL A recirculating loop for both NRZ and
soliton signals
???NRZ-style dispersion map
D
Anomalous
Average
Normal
Input
To Receiver
1.2 nm
AO
60/40 Coupler
PC
Filter
Switch
Anomalous
EDFA
Normal
7 km
25 km
15
Dispersion-managed solitons
Periodically stationary pulses still exist
Pulse Width
They stretch as they propagate effectively
lowering the nonlinearity!
Experimental verification
0 km
Pulses exist after 20,000 km
10,000 km
J. M. Jacob, E. A. Golovchenko, A. N.
Pilipetskii, G. M. Carter, and C. R. Menyuk,
IEEE Photon. Technol. Lett., vol. 9, pp. 130132
(1997)
28,000 km
Time (ps)
0
500
16
Dispersion-managed solitons
T. Yu, E. A. Golovchenko, A. N. Pilipetskii, and
C. R. Menyuk, Optics Lett. vol. 22, 723725
(1997)
Solitons can attract leading to errors
in phase
out of phase
17
Comparison of NRZ, RZ, DMS
(100 Gbits/sec, 20 km amplifier spacing)
D. Marcuse and C.R. Menyuk, J. Lightwave
Technol., vol. 17, 564569 (1999).
18
Comparison of NRZ, RZ, DMS
Conceptual view L4 gt L3 gt L2 gt L1
There is no clear distinction between NRZ, RZ,
DMS They naturally transition from one to the
other
?
I
0
anomalous
normal
D ?
19
A plague o both your houses. They have made
worms meat of me. Mercutio in Romeo and
Juliet (Act III, Scene 1, William Shakespeare)
20
Wavelength division multiplexed systems
21
System Configuration
Tyco CRZ system (Bergano, et al. OFC99, PD2)
Post-compensation
Channel
Pre-chirp compensation
Amp.
0
Dispersion
L1
22
System Configuration
CNET DMS system (Le Guen, et al. OFC99, PD4)
Channel
Pre-chirp
Post-compensation
Amp.
Dispersion
0
L2
L1
23
Single channel behavior large dispersion
Pulse evolution in the systems
Intermediate behaviors are different
CNET DMS system
Tyco CRZ system
Pulses are never periodically stationary!
24
Single channel behavior large dispersion
Eye diagrams at the receiver
Final behaviors are identical!
no ASE and no nonlinearity
no ASE but with nonlinearity
with ASE but no nonlinearity
with ASE and with nonlinearity
Tyco CRZ system
Voltage (a.u)
CNET DMS system
Time (ps)
Signal-spontaneous beat noise dominates the errors
25
Single channel behavior large dispersion
Initial chirp vs. minimum pulse duration
Tyco-CRZ
CNET-DMS
1
1
Tout/Tin
simulation
0
0
0.07
1
linear theory
0
0
0
15
0
2
C (GHz/ps)
C (GHz/ps)

• The behavior is qualitatively linear
• Nonlinearity is important in both systems (but
  more important in the CNET system?)

26
Single channel behavior large dispersion
System behavior is quasilinear in both systems

• Errors are dominated by signal-spontaneous beat
  noise
• Final pulse duration is determined by linear
  theory BUT
• Intermediate evolution is different
• Nonlinearity is important

27
Are the Algety solitons really solitons?
The question remains controversial
A view from Professor Haus April 18, 2001
28
7-Channel WDM Systems
Evolution of the spectrum
Tyco CRZ system
CNET DMS system
29
7-Channel WDM Systems
Output at the receiver
Ch. 1
Ch. 4
Ch. 7
Tyco CRZ system
Voltage (a.u)
CNET DMS system
Time (ps)
30
Conclusion

• Tyco and CNET WDM systems are quasilinear
• Nonlinearity plays an important role in both
  systems ?
• DMS and CRZ formats are converging

What about the KDD systems?
31
Whats in a name? That which we call a rose, by
any other name would smell as sweet . Juliet
in Romeo and Juliet (Act II, Scene 2, William
Shakespeare)