Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
Propagation Through Dispersive and Nonlinear SMF
Using Small Signal Analysis
Mihajlo Stefanović
1
, Daniela Milović
2
and Aleksandra Mitić
3
Abstract – The smallsignal analysis which takes into
account both the lowestorder group velocity dispersion
and the fiber nonlinearity influence on the signal
propagation in a singlemode fiber is derived from the
nonlinear Schrödinger equation. Fiber transfer functions
are derived and used for optical pulse propagation thus
avoiding splitstep Fourier method.
Keywords – Small signal analysis, optical fiber dispersion,
fiber nonlinearity, intensity modulation, optical
propagation.
I. INTRODUCTION
Data transmission in optical fiber communication systems
is determined by a number of linear and nonlinear effects. The
importance of both linear and nonlinear effects increases very
rapidly with the capacity of the transmission systems. The
joint action of the linear and nonlinear effects produces a great
variety of optical phenomena in transmission systems. Many
of these effects can easily distort the signal so that the data
will be completely lost. Nevertheless, by an appropriate
design of the system one can greatly reduce the detrimental
consequences of these phenomena.
Today highcapacity systems are limited by a joint
combination of dispersion and nonlinear effects. Dispersion
induced pulse broadening can be detrimental for optical
communication systems. In the nonlinear regime, the
combination of dipersion and nonlinearity can result in a
qualitatively different behaviour. The anomalous dispersion
regime is of considerable interest for the study of nonlinear
effects because in this regime optical fibers can support
solitons through a balance between the dispersive and
nonlinear effects. The intensity dependence of the refractive
index leads to a large number of interesting nonlinear effects.
Selfphase modulation (SPM) refers to the selfinduced phase
shift experienced by an optical field during its propagation in
optical fibers [1].
The impact of the interaction of dispersion and fiber
nonlinearity on the fiber response was investigated
numerically using the splitstep Fourier method. The effects of
1
Mihajlo Stefanovic is with the Faculty of Eletronic Engineering,
Beogradska 14, 18000 Nis, Yugoslavia, email:misa@elfak.ni.ac.yu
2
Daniela Milovic is with the Faculty of Eletronic Engineering,
Beogradska 14, 18000 Nis, Yugoslavia, email:dacha@elfak.ni.ac.yu
3
Aleksandra Mitic is with the Faculty of Eletronic Engineering,
Beogradska 14, 18000 Nis, Yugoslavia,email:alekmi@elfak.ni.ac.yu
the fiber nonlinearity and dispersion on the fiber transfer
function (FTF) was found to be highly desirable to design
properly longhaul optically amplified Gb/s transmission
systems since it avoids the long computation time required by
simulation and allows us to get more insight on the impact of
the fiber nonlinearity on the FTF [2].
II. SMALL SIGNAL ANALYSIS
Signal propagation in a single mode fiber taking into
account both group velocity dispersion (GVD) and the
nonlinear Kerr effect (fiber nonlinearity) is described by the
nonlinear Schrödinger equation (NLSE) [1]:
A A j
t
A j
A
z
A 2
2
2
2
2 2
γ β
α
=
∂
∂
+ +
∂
∂
(1)
where A is the slowly varying envelope amplitude, z is the
longitudinal coordinate of the fiber and α is the power
attenuation coefficient. The GVD is given by β
2
=–λ
2
D/(2πc)
with λ the carrier wavelength, D the dispersion parameter and
c velocity of light in vaccum. γ is nonlinearity coefficient
responsible for (SPM), given by 2πn
2
/(λA
eff
) with n
2
the
nonlinear index coefficient and A
eff
the effective core area.
The amplitude envelope can be written as
( ) ( ) ( )   t z j t z p t z A
T T
, exp , , φ ⋅ = (2)
where p
T
(z,t) and f
T
(z,t) are both real quantities denoting,
respectively, the power and the phase of the optical field. If
we replace the amplitude envelope in terms of power and
phase derivatives and after some algebraic manipulations, the
following set of two coupled nonlinear equation describing
the power and phase of an optical field propagation through a
nonlinear dispersive fiber valid for arbitrary input signal or
noise is optained [3]:
T
T
T
T T T
p
t
p
t t
p
z
p
⋅ −


.

\

∂
∂
+
∂
∂
⋅
∂
∂
=
∂
∂
α
φ φ
β
2
2
2
(3)
(
(
¸
(
¸

.

\

∂
∂
− 
.

\

∂
∂
−
∂
∂
− ⋅ =
∂
∂
2
2
2
2
2
2
2
4 2
8
t
p
t
p
t
p
p
p
p
t
T
T
T T
T
T
T
T
φ β
γ
φ
(4)
To derive the small signal analysis we express the total
power, p
T
(z,t) as the sum of the average power P(z), and a
noise or modulation term, p(z,t) and the phase as the sum of
the average phase Φ(z), and a noise or modulation term φ(z,t).
585
The small signal analysis implies that the average power is
much larger than the noise or modulation term P(z)>> p(z,t)
and the phase modulation or noise is small enough [4,5].
Substituting these components into the set of two coupled
nonlinear equations for the power and phase of the optical
field and neglecting the products of p(z,t) and φ(z,t) and their
derivatives, a set of coupled linear differential equation
governing the power and phase modulation terms is obtained.
Using the normalized power p
N
(z,t) defined as p(z,t)=
p
N
(z,t)exp(αz) and doing Fourier transforms of both sides of
the coupled linear differential equations governing the power
and phase modulation terms, we obtain the set of linear
differential equations [2,3]:
( )
( ) ( ) ω φ ω β
ω
, 0
,
~
2
2
~
z P
z
z p
N
⋅ − =
∂
∂
(5)
( )
( )
( ) ( ) ω α γ
ω β ω φ
, exp
0 4
,
~
2
2
~
z p z
P z
z
N
⋅
(
(
¸
(
¸
− ⋅ + =
∂
∂
(6)
where ( ) ω ,
~
z p
N
and ( ) ω φ ,
~
z are the Fourier transforms of
p
N
(z,t) and φ(z,t), respecitevely, P(0) is the average power at
the fiber input and ω is the angular modulation frequency.
Equation (5) and (6) govern the power and phase fluctuation
evolution along the optical fiber taking into account GVD,
nonlinear Kerr effect and fiber loss.
III. IMIM CONVERSION FUNCTION
Since in IMDD systems the information recovered only
from the optical power amplitude, we focus on the IM
solution. The information signal at the fiber input is carried by
the IM term ( ) ω
in
p
~
, while the transmitter chirp is included by
the PM term ( ) ω φ
~
in
.
For the purpose of theoretically analyzing frequency
transfer function some conversion functions are given in [6].
If we assume that only intensity modulation is present and
there is no phase modulation, we use IMIM conversion
function given in [6] as follows
0 ) , 0 (
~ ) , 0 (
~
) , (
~
=
−
=
ω φ
ω
ω
p
z p
C
N
IM IM
In case of a lossless and linear fiber, equations (5) and (6)
can be solved analitically [5,7].
( ) ( ) ( ) ( ) ω φ
ω β
ω
ω β
ω , 0
2
sin 0 2 , 0
2
cos ,
~
2
2
~
2
2
~
⋅


.

\

⋅ + ⋅


.

\

= P p z p
(7)
( ) ( )
( )
( ) ω
ω β
ω φ
ω β
ω φ , 0
2
sin
0 2
1
, 0
2
cos ,
~
2
2
~
2
2
~
p
P
z ⋅


.

\

⋅ − ⋅


.

\

=
(8)
we will now consider the propagation of optical pulse having
Gaussian amplitude envelope
) ) / ( exp( ) 0 ( ) , 0 (
2
0
T t P t p − =
Optical pulse propagation is then simply simulated using
equation (7) for the case of no phase modulation for typical
standard single mode fiber (SMF) operating at wavelength of
λ=1550 nm. Obtained results are shown graphically on Fig.1.
200
400
5
10
15
0
0.25
0.5
0.75
1
200
400
Fig.1. Spatial domain optical pulse propagation in linear
dispersive regime (D=17 ps/nm/km)
Fig.2. Pulse broadening due to dispersion for different
propagation distances
When the optical power at the fiber input increases,
nonlinear effects cannot be neglected and dispersive nonlinear
propagation must be considered.
In case of a lossless fiber with the effects of fiber nonlinearity
and dispersion, solving the equations (7) and (8) are obtained
( ) ( ) ( )
( )
( ) ( ) ω φ ξ
ξ
ω β
ω ξ ω , 0 sin
0
, 0 cos ,
~
2
2
~ ~
⋅ ⋅ + ⋅ = z
P
p z z p
N N
(9)
( ) ( ) ( )
( ) ( )
( ) ( ) ω ξ
ω β
ξ
ω φ ξ ω φ , 0 sin
0
, 0 cos ,
~
2
2
~ ~
N
p z
P
z z ⋅ ⋅ − ⋅ =
(10)
t(arb.units)

p
N
(
z
,
t
)

t(arb.un.)

p
N
(
z
,
t
)

z(km)
50
100
150
586
where
2
K = ξ
and K
2
=(β
2
ω
2
/2)
2
+β
2
ω
2
γP(0). If K
2
>0 the
solutions are stable. In the normal regime, β
2
>0 and K
2
>0,
both the power and the phase have a stable solution. In the
anomalous regime, β
2
<0 and for modulation frequencies
smaller than the angular critical frequency given by
( )
2
/ 0 4 β γ ω P =
, K
2
<0, power and phase present an unstable
solution; otherwise solution is stable. A stable or unstable
solution (K
2
>0 or K
2
<0, respecitively) depens critically on
whether the GVD is normal or anomalus. The modulation
instability is a phenomenon resulting from the iteration of
dispersion and fiber nonlinearity. In the case of unstable
solution the power or the phase modulation term will grow
exponentially along the optical fiber [7].
For the case of no phase modulation, optical pulse
propagation in spatial domain, for SMF parameters D=17
ps/nm/km and γ=1.05 W
1
km
1
, is shown on Fig.2. Inputu peak
power value is P(0)=16dBm.
Fig.3. Optical pulse propagation in spatial domain through
nonlinear dispersive SMF
Fig.4. Pulse distortion due to dispersion and nonlinearity for
different propagation distances
Equations (7) and (8) as well as equations (9) i (10) may be
written by a matrix notation from which it may be seen that
for any arbitrary phase and intensity modulation (or noise) at
the fiber input, the corresponding phase and intensity
modulation (or noise) at the fiber output, takes into account
the correlation between the phase and intensity modulation
and considers the second order dispersion effects and fiber
nonlinearity, respectively. The conversion matrix is valid for
any laser source, because it is derived without imposing any
relation between phase and intensity of the field at the input of
fiber, except the small signal assumption.
Fig.5. Power peaks versus propagation distance for: dispersion
only (full line) and dispersion and nonlinearity (dashed line)
IV. CONCLUSION
Small signal analysis derived directly from nonlinear
Schrodinger equation is a very usefull approximation that
avoids splitstep Fourier method and thus make the computing
time less. Although it is restricted only to small input signals,
this approximation is valid for any laser source. In this paper
we considered optical pulse propagation for the cases of: 1)
dispersion only and 2) dispersion and nonlinearity both
included. It is shown that dispersion only causes more severe
pulse distortion since nonlinearity compensate pulse
broadening as well as pulse envelope degradation.
REFERENCES
[1] G. P. Agraval, Nonlinear Fiber Opics, Academic Press
Inc., 2
nd
Edition San Diego, 1995.
[2] A.V.T. Cartaxo, B. Wedding and W. Idler, “Influece of
Fiber Nonlinearity on the Phase Noise to Intensity Noise
Conversion in Fiber Transmission: Theoretical and
Experimental Analysis”, Journal of Lightwave
Technology, Vol. 16, N
o
.7, July 1998.
[3] A.V.T. Cartaxo, B. Wedding and W. Idler, “Influence of
Fiber Nonlinearity on the Fiber Transfer Function:
Theoretical and Experimental Analysis”, Journal of
Lightwave Technology, Vol. 17, N
o
10, October 1999.
[4] A.V.T. Cartaxo and J. Morgano, “Rigorous Assessment
of SmallSignal Analysis for Linear and Dispersive
Optical Communication Systems Operating Near Zero
200
400
5
10
15
0
0.25
0.5
0.75
1
200
400
t(arb.un.)

p
N
(
z
,
t
)

z(km)
50
100
150
t(arb.units)

p
N
(
z
,
t
)

587
Dispersion Wavelenght”, Journal of Lightwave
Technology, Vol. 17, N
o
1, January 1999.
[5] J. Wang and K. Petermann, “Small Signal Analysis for
Dispersive Optical Fiber Communication Systems”,
Journal of Lightwave Technology, Vol. 10, N
o
1, January
1992.
[6] F. Ramos and J. Marti, “RF Response of Analog Optical
Links Employing Optical Phase Conjugation”, Journal
of Lightwave Technology, Vol. 19, N
o
6, June 2001.
[7] A. V. T. Cartaxo, “Smallsignal Analysis for Nonlinear
and Dispersive Optical Fibers, and its Application to
Design of Dispersion Supported Transmission Systems
with Optical Dispersion Compensation”, IEE Proc.
Optoelectron., Vol. 146, N
o
5, October 1999.
ω ) ∂z 4 P(0 ) 2 ~ pN(z.t)exp(αz) and doing Fourier transforms of both sides of the coupled linear differential equations governing the power and phase modulation terms.t). equations (5) and (6) can be solved analitically [5. ω ) (9) ξ ~ ~ ~ ξ φ (z . ω ) p N (z. t ) = P(0) exp(−(t / T0 ) 2 ) Optical pulse propagation is then simply simulated using equation (7) for the case of no phase modulation for typical standard single mode fiber (SMF) operating at wavelength of λ=1550 nm. Obtained results are shown graphically on Fig.7]. Using the normalized power pN(z.t) and the phase modulation or noise is small enough [4. ω ) (7) 2 2 ~ β 2ω 2 ~ β 2ω 2 ~ 1 (8) φ (z. ω ) = cos(ξz ) ⋅ p N (0. 25 0 200 5 50 400 ~ ∂ φ (z . P(0) is the average power at the fiber input and ω is the angular modulation frequency. respecitevely. t(arb.) Fig. In case of a lossless fiber with the effects of fiber nonlinearity and dispersion. (5) (6) 1 0. Substituting these components into the set of two coupled nonlinear equations for the power and phase of the optical field and neglecting the products of p(z. ω ) = − β 2ω 2 P(0) ⋅ φ (z . ω ) − ⋅ sin (ξz ) ⋅ p N (0. ω ) = cos 2 ⋅ φ (0. we use IMIM conversion function given in [6] as follows CIM − IM ~ ( z. ω ) = cos 2 ⋅ p (0.units) Fig.t) and φ(z. ω ) and φ (z . we obtain the set of linear differential equations [2. Spatial domain optical pulse propagation in linear dispersive regime (D=17 ps/nm/km) III. nonlinear Kerr effect and fiber loss. we focus on the IM solution.t)= pN(z.t) defined as p(z. IMIM CONVERSION FUNCTION Since in IMDD systems the information recovered only from the optical power amplitude. while the transmitter chirp is included by the PM term φ in (ω ) . If we assume that only intensity modulation is present and there is no phase modulation. ω ) ~ ~ pN(z.t) ~ 15 150 10 100 where p N (z .t) and their derivatives. 75 0.5]. ω ) + 2 P (0) ⋅ sin 2 ⋅ φ (0. ω ) (10) 2 (β 2 ω P(0)) φ ( 0. The information signal at the fiber input is carried by the IM term p in (ω ) .t) t(arb.3]: ~ ∂ p N (z . 5 0.The small signal analysis implies that the average power is much larger than the noise or modulation term P(z)>> p(z. solving the equations (7) and (8) are obtained ~ β 2 ω 2 P (0) ⋅ sin (ξz ) ⋅ φ (0. a set of coupled linear differential equation governing the power and phase modulation terms is obtained. Pulse broadening due to dispersion for different propagation distances When the optical power at the fiber input increases. ω ) β 2ω = + γ ⋅ exp(− αz ) ⋅ p N (z .ω ) pN = ~ p (0. For the purpose of theoretically analyzing frequency transfer function some conversion functions are given in [6]. ω ) − 2 P (0) ⋅ sin 2 ⋅ p (0. ~ β ω2 ~ β ω2 ~ p (z.1.t) and φ(z. ω ) ∂z ~ p(0. ω ) + ~ ~ we will now consider the propagation of optical pulse having Gaussian amplitude envelope 585 . ω ) = cos(ξz ) ⋅ φ (0. nonlinear effects cannot be neglected and dispersive nonlinear propagation must be considered.ω ) = 0 ~ In case of a lossless and linear fiber. ω ) are the Fourier transforms of ~ z(km) pN(z.un.1. Equation (5) and (6) govern the power and phase fluctuation evolution along the optical fiber taking into account GVD.2.
Journal of Lightwave Technology. Nonlinear Fiber Opics.where ξ = 2 2 2 2 2 K 2 and K =(β2ω /2) +β2ω γP(0). Optical pulse propagation in spatial domain through nonlinear dispersive SMF pN(z.2. Journal of Lightwave Technology. for SMF parameters D=17 ps/nm/km and γ=1. Cartaxo. Inputu peak power value is P(0)=16dBm. CONCLUSION Small signal analysis derived directly from nonlinear Schrodinger equation is a very usefull approximation that avoids splitstep Fourier method and thus make the computing time less. Agraval.5. [2] A.T. In this paper we considered optical pulse propagation for the cases of: 1) dispersion only and 2) dispersion and nonlinearity both included. except the small signal assumption. No. 5 0. for any arbitrary phase and intensity modulation (or noise) at the fiber input.units) Fig.3. takes into account the correlation between the phase and intensity modulation and considers the second order dispersion effects and fiber nonlinearity.05 W1km1. Academic Press Inc. “Influence of Fiber Nonlinearity on the Fiber Transfer Function: Theoretical and Experimental Analysis”. Cartaxo and J. B. is shown on Fig. October 1999. respecitively) depens critically on whether the GVD is normal or anomalus. “Influece of Fiber Nonlinearity on the Phase Noise to Intensity Noise Conversion in Fiber Transmission: Theoretical and Experimental Analysis”. It is shown that dispersion only causes more severe pulse distortion since nonlinearity compensate pulse broadening as well as pulse envelope degradation. No 10. otherwise solution is stable.V.t) REFERENCES [1] G. In the normal regime. β2<0 and for modulation frequencies smaller than the angular critical frequency given by 2 ω = 4γP (0) / β 2 .un. A stable or unstable solution (K2>0 or K2<0.. the corresponding phase and intensity modulation (or noise) at the fiber output. B. Idler. Wedding and W. 1 0. β2>0 and K2>0. 25 0 200 5 50 400 pN(z. For the case of no phase modulation. The conversion matrix is valid for any laser source. z(km) t(arb. If K >0 the solutions are stable. Morgano. power and phase present an unstable solution.7.) Fig. Vol. Idler. In the case of unstable solution the power or the phase modulation term will grow exponentially along the optical fiber [7]. [4] A.t) 15 150 10 100 Fig. Power peaks versus propagation distance for: dispersion only (full line) and dispersion and nonlinearity (dashed line) IV. Wedding and W.V. Cartaxo. 75 0. In the anomalous regime. optical pulse propagation in spatial domain. 16.T.T. The modulation instability is a phenomenon resulting from the iteration of dispersion and fiber nonlinearity. K <0. this approximation is valid for any laser source. respectively. “Rigorous Assessment of SmallSignal Analysis for Linear and Dispersive Optical Communication Systems Operating Near Zero t(arb. July 1998. Although it is restricted only to small input signals. both the power and the phase have a stable solution. 17. P. [3] A.4. 2nd Edition San Diego.V. Vol. because it is derived without imposing any relation between phase and intensity of the field at the input of fiber. Pulse distortion due to dispersion and nonlinearity for different propagation distances Equations (7) and (8) as well as equations (9) i (10) may be written by a matrix notation from which it may be seen that 586 . 1995.
and its Application to Design of Dispersion Supported Transmission Systems with Optical Dispersion Compensation”. Petermann. June 2001. January 1999. Vol. 17. Journal of Lightwave Technology. Marti. Wang and K. January 1992. 587 .Optoelectron. Cartaxo. “Small Signal Analysis for Dispersive Optical Fiber Communication Systems”. V. Vol. Vol. “Smallsignal Analysis for Nonlinear and Dispersive Optical Fibers.Dispersion Wavelenght”. [6] F. Journal of Lightwave Technology. Journal of Lightwave Technology. 19. No 1.. No 5. [5] J. “RF Response of Analog Optical Links Employing Optical Phase Conjugation”. Ramos and J. October 1999. No 6. T. No1. 146. [7] A. Vol. IEE Proc. 10.
This action might not be possible to undo. Are you sure you want to continue?