POWER TRANSIENTS IN EDFAS

Figure 5 a) Absorption and (b) emission cross section

The channels were dropped and added at 2 ms and 6 ms, respectively. Figure 6
shows the results for different values of the proportion of surviving channel power to
the total power (which is kept constant). The results of the addition of channels shown
in Figure 6 agrees with Figure 3. The drop of channels in the system tends to cause
an increase in the output power of the surviving channels.

391
POWER TRANSIENTS IN EDFAS

Figure 6 Evolution of the power excursion of the survival wavelength at 1554nm for 7 different values of
the surviving wavelength power as a proportion of the total input power

In a second example, the effect of different pump powers in power excursion was
analyzed. In this case the input power in the surviving channel (1554 nm) and the
modulated channel (1557.2 nm) was the same: -8.67 dBm. The power excursions
were examined for 4 pump powers: 10.3 dBm, 12.5 dBm, 13.9 dBm, and 15.1 dBm.
The results found are shown in Figure 7.

392
 POWER TRANSIENTS IN EDFAS

Figure 7 Power excursion of the surviving wavelength at 1554 nm for four different pump powers values

For increasing pump power values the evolution of the power transient begins from
higher inversion/gain levels and settles to higher gain levels with an increasing in the
ASE causing a decreasing in the power excursions.

The results presented here are in very good agreement with studies in the references.

393
POWER TRANSIENTS IN EDFAS

References:
[1] Emmanuel Desurvire. "Erbium-Doped Fiber Amplifier: Principles and Applications", John Wiley
& Sons.
[2] C. Dimopoulos. "Study of Dynamic Phenomena in WDM Optical Fiber Links and Networks
Based on EDFA". PHD thesis. University of Essex, April 2001

394
 TEMPERATURE DEPENDENCE IN EDFA

Temperature dependence in EDFA

This lesson demonstrates the temperature dependence in the absorption coefficient

of an EDFA.

The temperature dependence in an EDFA manifests through the variation in the

absorption  and gain g coefficients (or absorption  a and emission  e
cross-sections) [1], [2]. Here, to demonstrate the effect of temperature dependence,
simulations are done using different temperatures and a method to characterize the
EDF's absorption coefficient at each temperature is used.

In the first part of this lesson, the method for calculating the EDF's absorption
coefficient is described. The system shown in Figure 1 is used for the characterization
of the absorption coefficient. It consists of a CW laser source with signal power equal
-20dBm and wavelength varying at the range of 1450nm to 1650nm (the CW laser is
in the sweep mode).

The Optical Power Meter Visualizer gives the output power in the EDFA. The Erbium
doped fiber has the physical parameters shown in Figure 2(a) and the original
absorption coefficient loaded is shown in Figure 2(b).

Figure 1 System layout

395
TEMPERATURE DEPENDENCE IN EDFA

Figure 2 (a) Main tab with the parameters used in the simulations (b) Absorption coefficient loaded

396
 TEMPERATURE DEPENDENCE IN EDFA

The following procedure was used to calculate the absorption coefficient from the data
obtained in simulations. To determine the value of the absorption coefficient of the
EDF, simulations for two fiber lengths (5m and 0.5m) were conducted in the
wavelength range of 1450 to 1650nm. Using the corresponding output signal powers
P out    L = 5 m  and P out    L = 0.5m  , the absorption coefficient can be
obtained by the equation

P out    L = 0.5m  – P out    L = 5 m  (1)

    = ---------------------------------------------------------------------------------------------
5 – 0.5

In Figure 3 we compare the original absorption coefficient loaded on the EDF and the
calculated one using Equation 1. You can see the agreement between the two sets of
results. Therefore, this procedure can be used to calculate the absorption coefficient
of an EDFA.

Figure 3 Comparison between the original and the calculated absorption coefficient

In the second part of this lesson, the temperature dependence of the absorption
coefficient is demonstrated. The method used to calculate the absorption coefficient
is used again, but this time, the simulations considered different temperatures. With
this procedure, the temperature dependence in the absorption coefficient is
demonstrated.

397
TEMPERATURE DEPENDENCE IN EDFA

To consider the temperature dependence in the Erbium-doped fiber component,

Include Temperature Effect on the Enhanced tab must be selected, and the
temperature that the cross sections (loaded) were measured at must be provided.

Here, the original cross sections (Figure 2b) were measured at 20C . Finally, user
the current temperature must be defined in order for it to be considered in the
simulation temperature parameter. Figure 4 shows an example using 80C for the
simulation.

Figure 4 Enhanced tab with simulation setup

Simulations were done for three temperatures, – 40C , + 20C , and + 80C . For
each temperature, the absorption coefficients were calculated using Equation 1.
Figure 5 shows the results for the three temperatures simulated. It is possible to see
the differences in the absorption coefficients caused by the changes in temperature.

These results seem in good qualitative agreement with presented studies in [1][2].
The same approach can be taken to analyze temperature dependence in the gain
coefficient. However, the procedure and the system setup used to calculate the gain
coefficient will be different, because it will be necessary to find the fluorescence
spectrum.

398
 TEMPERATURE DEPENDENCE IN EDFA

Figure 5 Absorption coefficient calculated for different temperatures

This variation in the absorption and gain coefficients causes variation in the
performance of the amplifier, gain, and noise was demonstrated in [1] and [2].

399
TEMPERATURE DEPENDENCE IN EDFA

References:
[1] P. F. Wysocki, N. Conti, and D. Holcomb. "Simple Modeling Approach for the Temperature
Dependence of the Gain of Erbium-Doped Fiber Amplifiers". SPIE Conference on Optical
Devices for Fiber Communication, Volume 3847, 1999.
[2] M. Bolshtyansky, P. F. Wysocki, N. Conti. "Model of Temperature Dependence for Gain Shape
of Erbium-Doped Fiber". Journal of Lightwave Technology, volume 18, Issue 11, Dec 2000.

400
 YTTERBIUM-DOPED FIBER AMPLIFIERS

Ytterbium-doped fiber amplifiers

The objective of this lesson is to demonstrate the performance of ytterbium-doped

fiber amplifiers.

First of all, the gain spectra are analyzed for the pump wavelength at 910 nm and with
different input pump powers: 1mW, 10mW and 30mW. The system used in the
simulation is shown in Figure 1.

Figure 1 System layout for gain analysis

Figure 2 shows signal gain spectra obtained for three different pump powers.

The parameters used in this simulation are displayed in Figure 1, and they have
similar values to those presented in [1], except with regard to the values of the
absorption and emission cross-sections. Nevertheless, the results found are in good
agreement with the results presented in [1].

401
YTTERBIUM-DOPED FIBER AMPLIFIERS

Figure 2 Amplifier gain spectra for three different input pump powers

In a second case, the gain spectra are calculated for two different pump wavelengths,
910nm and 975nm. The system layout simulated is showed in Figure 3, and the fiber
parameters used are similar to the parameters used in the previous example.

Figure 3 System layout

The results obtained in the simulation are shown in Figure 4.

402
 YTTERBIUM-DOPED FIBER AMPLIFIERS

Figure 4 Gain spectra obtained for pump wavelengths at 910 nm and 975 nm

In the next case, the Ytterbium-doped fiber is double-clad. To set the ytterbium-doped
fiber component to work as double-clad, the parameter Double-clad fiber
(Enhanced tab) has to be set to True and the value of the pump reference has to be
specified. The system used to simulate the double-clad fiber amplifier counter-
pumped is shown in Figure 5.

Figure 5 Layout of double-clad fiber amplifier counter-pumped system

403
YTTERBIUM-DOPED FIBER AMPLIFIERS

The EDF parameters used in this simulation are similar to those used in [2]. The
forward and backward output power is calculated for input signal power varying from
-30dBm to -30dBm for co-pumped and counter-pumped configurations. Figure 6
shows the results. This figure is equivalent to Figure 2 of [2].

Figure 6 Output power versus input signal power for different pump schemes

References:
[1] Paschotta, R.; Nilsson, J.; Tropper, A.C.; Hanna, D.C., "Ytterbium-doped fiber amplifiers". IEEE
Journal of Quantum Electronics, Vol.: 33, Issue: 7, Jul 1997, Pages: 1049-1056.
[2] Yong Wang; Hong Po, "Dynamic characteristics of double-clad fiber amplifiers for high-power
pulse amplification". IEEE Journal of Lightwave Technology, Vol.: 21, Issue: 10, Oct. 2003,
Pages: 2262- 2270.

404
 SPO OPTIMIZATION—SYSTEM MARGIN

SPO optimization—System margin

OptiSystem can optimize parameters in order to maximize, minimize or to target a

value for results. This can be done by either using MPO or SPO optimizations.
By using the optimization tool you can, for example,
• optimize the fiber length of the EDFA to obtain the maximum gain
• calculate the attenuation/gain to obtain a target Q-Factor
• minimize the BER by optimizing the fiber length of the link.
In this example we show optimization of a single parameter by using SPO. It shows
how to estimate the system margin. This margin shows the amount of power penalty
that may be added to the system to get a defined Q factor or BER. Our target BER is
10-9, or a Q-Factor of 6.The project is given in System margin.osd, shown in
Figure 1.

Figure 1 Project layout

In this project, for illustrative purposes, the subsystem System under test is an
empty system. The optimization will optimize the attenuation parameter of the
Attenuator component to attain Max. Q-Factor of 6. The parameter attenuation will be
the system margin in dB.
To set up the optimization, go to Tools|Optimizations... and insert a SPO
Optimization. Then, select Gain Attaining type of optimization in the Main tab and
set Result tolerance to 0.05. In the Parameter tab, add Attenuation of the
Attenuator component into the Selected list. In the Result tab, add Max. Q factor of
the BER Analyzer into the Selected list.
Figure 2, Figure 3, and Figure 4 show the Main, Parameter and Result tabs of the
optimization set-up.

405
SPO OPTIMIZATION—SYSTEM MARGIN

Figure 2 Main tab of the SPO

406
 SPO OPTIMIZATION—SYSTEM MARGIN

Figure 3 Parameter tab of the SPO

407
SPO OPTIMIZATION—SYSTEM MARGIN

Figure 4 Result tab of the SPO

In order to run the optimizations, go to File > Calculate in the Main menu, and select
the Run optimizations in the Calculation dialog box, which also runs the simulation.

408
 SPO OPTIMIZATION—SYSTEM MARGIN

Figure 5 Enabling optimizations

After the optimization is done, you will see that the system margin is approximately
19 dB for a Q-Factor of 6 (see Figure 6).

409
SPO OPTIMIZATION—SYSTEM MARGIN

Figure 6 Optimization results

410
 SPO OPTIMIZATION—EDFA FIBER LENGTH

SPO optimization—EDFA fiber length

In this example we show optimization of a single parameter by using SPO. It shows

how to optimize the fiber length to maximize the amplifier gain. The project is given in
EDFA fiber length optimizationFig.osd. In order to run the optimizations
with the components shown in Figure 1, the Laser power is -40 dBm. Default EDFA
parameters are used.

Figure 1 Project layout

The optimization will optimize the Length parameter of the EDFA to get maximum
gain as measured by Dual Port WDM Analyzer.
To set up the optimization, go to Tools > Optimizations and insert a SPO
Optimization. Then, select Maximize type of optimization in the Main tab and set
Result tolerance to 0.1. In the Parameter tab, add a Length parameter of the EDFA
component into the Selected list. In the Result tab, add Gain 1 (dB) result of the Dual
Port WDM Analyzer into the Selected list.
In order to run the optimizations, go to File > Calculate from the Main menu, and
enable Run optimizations in the Calculation dialog box, which runs the simulation.
After 11 passes, fiber length for maximum gain is found to be 9.86. In this case, EDFA
gain is about 38.1 dB.

411
SPO OPTIMIZATION—EDFA FIBER LENGTH

Notes:

412
 EDFA — BASIC CONCEPTS

EDFA — Basic concepts

The sample file EDFA Basic Concepts.osd shows the basic characterization of
the Erbium Doped Fiber Amplifier (EDFA). There are three design layouts: Gain
Spectrum, Gain Saturation, and Amplifier Noise. They enable the characterization
of the gain, noise figure, and output power under unsaturated and saturated signal
input regime. These layouts refer to the topics: Gain Spectrum and Bandwidth,
Gain Saturation, and Amplifier Noise.
In this example, the amplifier performance is characterized by using sweeps for the
input parameters of the amplifier. the layouts Gain Spectrum and Amplifier Noise
sweep the signal wavelength, while the layout Gain Saturation sweeps the signal
input power in order to evaluate the amplifier performance with the small signal input
power (~ -30 dBm) and saturated input signal power (~ 0 dBm).
The three layouts created in this project file are shown in Figure 1. The basic layout
used in each layout considers one Erbium-doped fiber stage setup in a co-
propagating pump scheme. Erbium Doped Fiber model was used in this simulation.
A single input signal operating in the C-band wavelength range is set in the laser
source where small and large signal inputs are considered in the simulations.

Figure 1 Three different layouts created in this project file

413
EDFA — BASIC CONCEPTS

For each design layout we have three graphs showing the Output Signal Power,
Gain, and Noise Figure versus the Sweep Parameter. The components settings can
be modified and the simulations repeated in order to analyze the differences observed
in the amplifier performance as a consequence of the change in parameter settings.
The absorption and emission cross section, input parameters which are critical in the
numerical solution of coupled rate and propagating equations, are displayed in
Erbium doped fiber component. Figure 2 shows the cross-section file used in this
project file.
The cross section input files are characteristic to a specific fiber as well as the fiber
dimensions specified in the Er doped fiber dialog box component, shown in Figure 3.
However, it is interesting to change some fiber specifications in order to evaluate how
it can modify the calculated results.

Figure 2 Cross-section file displayed in the Erbium doped fiber component Graphs

414
 EDFA — BASIC CONCEPTS

Figure 3 Er doped fiber specifications

The basic layout used in the calculations to each layout is shown in Figure 4.

Figure 4 Layout used in the basic characterization of the Erbium doped fiber amplifier setup in one fiber
stage and co-pumped by 980 nm laser diode

415
EDFA — BASIC CONCEPTS

Layout Gain Spectrum

In this first layout named Gain Spectrum, the signal wavelength of the CW laser
signal is swept from 1525 nm to 1600 nm, which enables evaluating the spectral gain
and the bandwidth of the amplifier.
The spectral gain calculated to -20 dBm of input signal power is shown in Figure 5.

Figure 5 Gain versus signal wavelength given in [nm] units considering signal input power equal to –20
dBm

An additional graph which can be checked in the Gain Spectrum layout is the output
power versus signal wavelength shown in Figure 6.

416
 EDFA — BASIC CONCEPTS

Figure 6 Signal output power versus signal wavelength given in [nm] units calculated to a signal input
power equal to –20 dBm

417
EDFA — BASIC CONCEPTS

Layout Gain Saturation

The layout Gain Saturation enables evaluating the amplifier performance as a
function of the signal input power. The signal input power is swept from small signal
to large signal regime, i.e. -40 dBm to 0 dBm. The amplifier performance given by
gain, output power and noise figure are respectively shown in Figure 7, Figure 8 and
Figure 9.

Figure 7 Gain versus signal input power

Figure 8 Amplified signal versus signal input power

418
 EDFA — BASIC CONCEPTS

Figure 9 Noise figure versus signal input power

419
EDFA — BASIC CONCEPTS

Layout Amplifier Noise

The signal wavelength is the parameter swept in this layout, which enables
calculating the spectral noise of the amplifier. Signal and pump wavelength as well as
the fiber parameters can be modified for further evaluation of the amplifier
performance.

Figure 10 Noise bins as a function wavelength used in the calculations. Each curve in this graph resulted
from the different iterations

420
 BOOSTER AMPLIFIER

Booster Amplifier

The project file Booster Amplifier.osd shows the characterization of a

booster/power amplifier setup in a single erbium-doped fiber stage heavily doped,
bidirectionally pumped by two 980 nm pump lasers. A large signal input power is
considered in this case, since high signal input helps to produce high output power,
which is necessary to booster/power amplifiers. As a consequence, moderate gain
will be observed in this case. The low noise figure requirement of EDFAs is not so
critical in this case, because an increase in NF can be tolerated. It is important to
mention that typical configurations of booster amplifiers include multiple Er-doped
fiber stages.
Figure 1 shows the layout of the booster amplifier. A bidirectional pump was used to
exemplify the typical pump scheme observed in booster amplifiers. The Optical
Spectrum Analyzer connected at Output Port1 shows the amplified signal obtained
after calculating this sample file. The component Dual Port WDM Analyzer gives the
calculated results generated by all the propagating signals, signals and pump.
The output power, gain and noise figures calculated as a function of signal input
power are shown in Figure 2, Figure 3 and Figure 4. It is possible to observe the gain
being compressed as a function of the signal input increasing on Figure 3.
The noise figure curve calculated sweeping the signal input power to the booster
amplifier utilized the same input parameters as before. There is a region in the curve
in which NF is minimized as a function of the signal input considered. Observing
graphs related with co- and counter-propagating ASE helps to understand how the
noise figure curve appears as this shape.

421
BOOSTER AMPLIFIER

Figure 1 Layout used in the Booster amplifier

Figure 2 Signal output power versus signal input power calculated to the booster amplifier

422
 BOOSTER AMPLIFIER

Figure 3 Gain versus signal input power calculated to the booster amplifier

Figure 4 Noise figure versus signal input power calculated to the booster amplifier

The power of both pump lasers co- and counter-propagating can be modified as well
as the Er-doped fiber length. Fiber specifications can also be modified and the results
compared with the previous case.

423
BOOSTER AMPLIFIER

Notes:

424
 INLINE AMPLIFIER

Inline Amplifier

The project file Inline Amplifier.osd shows the characterization of an inline

amplifier setup in a single erbium-doped fiber stage, pumped by one 980 nm-pump
laser.

A small signal input power is considered in this case, where results with high gain and
small noise figure values are desirable. The signal input power is swept from -40 dBm
to -20 dBm to enable the checking of the gain, noise figure, output power, and OSNR
as a function of signal input power. The signal input power swept is in the layout
"Signal input power" contained in the file Inline Amplifier .osd.

After sweeping iterations over the signal input power, the amplifier performance can
be checked in the graphs shown in Figure 1.

Figure 1 Graphs presenting the inline amplifier performance setup in a co-propagating pump scheme

Two different pump schemes which consider co- and counter-propagating pump are
available, in "Co-pump power" and "Counter-pump power" layouts, shown in Figure 2
and Figure 3. The performance of the inline amplifier with respect to the co- and
counter-propagating pump scheme can be checked in the Gain, Noise Figure, and
Output Power graphs.

425
INLINE AMPLIFIER

Figure 2 Inline amplifier layout considering a co-propagating pump scheme

Figure 3 Inline amplifier layout considering a counter-propagating pump scheme

The amplifier performance of the amplifier setup in a co- and counter-propagating

pump scheme can be compared to the graphs available. Pump wavelength equal to
980 nm was considered in both cases. Gain versus pump power is shown in Figure 4

426
 INLINE AMPLIFIER

considering the co- and counter-pump scheme, which allows for the evaluation of the
most efficient pump scheme to the EDFAs.

Figure 4 Comparing gain performance of an EDFA setup in a co- and counter-propagating pump scheme

Different parameters can be changed in this example, in which the new results are
compared with the previous ones. The wavelength pump power equal to 980 nm,
considered in the three different layouts included in this project file, can be substituted
with 1480 nm, for example. Different pump power, signal wavelength, and fiber
parameters can be considered to perform additional simulations.

427
INLINE AMPLIFIER

Notes:

428
 PREAMPLIFIER

Preamplifier

The project file Preamplifier.osd shows the characterization of an inline amplifier

setup in a single erbium-doped fiber stage, pumped by one 980 nm-pump laser in a
co-pump scheme.
A small signal input power is considered in this case, where results with high gain and
small noise figure values are desirable. The erbium doped fiber length is swept from
6 m to 15 m to enable the checking of the gain, noise figure, output power, and OSNR
as a function of fiber length. The pump wavelength considered in layouts "Pump 980
nm" and "Pump 1480 nm" is 980 nm and 1480 nm.
In both layouts, signal input power is -35 dBm at 1550 nm. The performance of the
preamplifier with respect to the wavelength pump can be checked in the graphs of
gain, noise figure, and output power.
The basic layout used in this preamplifier example is shown in Figure 1, where just
the pump wavelength is changed from 980 nm to 1480 nm in the component Pump
Laser 1.0.

Figure 1 Preamplifier layout

The gain of the preamplifier is shown in Figure 2. Different parameters can be

changed to generate new results to be compared with the previous ones. Different
pump power, signal wavelength, and fiber parameters can be used to perform
additional simulations.

429
PREAMPLIFIER

Figure 2 Gain versus erbium doped fiber length considering 980 nm and 1480 nm as wavelength pump

430
 PUMPING REQUIREMENTS

Pumping requirements

The amplifier performance characterized by the signal output power, gain and noise
figure depends on the pump wavelength. The project file "Pumping
requirements.osd" shown in two different layouts, "Pump power 980 nm" and
"Pump power 1480 nm", details an amplifier pumped by 980 nm and 1480 nm. The
980 nm and 1480 nm pump wavelengths are the most important ones used in EDFAs.
Figure 1 shows the layout setup in a co-propagating pump scheme with 980 nm and
1480 nm wavelength pumps.

Figure 1 Amplifier setup in a co-propagating pump scheme used with 980 nm and 1480 nm wavelength
pump

High amplifier gains in the range 30-40 dB can be obtained in this project file,
sweeping the pump power from 10 mW to 200 mW. The signal input power
considered in this case is -20 dBm. Output power calculated in this project is in the
range 6 - 17 dBm, while the noise figure varied between 3 - 5 dB. Different signal input
power or signal wavelength as well as fiber parameters can be set in this example and
the new results can be compared with previous one. Results obtained with these
layouts are shown in Figure 2. New Erbium-doped fiber data can be used in this case
and results compared. The cross-section or Giles parameters need to be updated as
well as the fiber dimension and doping characteristics.
EDFAs can be designed to operate in such a way that the pump and signal beams
propagate in opposite directions, a configuration referred to as backward pumping to
distinguish it from forward-pumping configuration in which both beams propagate
along the same direction. Backward or counter-propagating pump scheme is

431
PUMPING REQUIREMENTS

exemplified in the layout "Backward pump @ 980 nm". The pump power is swept from
10 to 150 mW and the signal input power is -20 dBm.

Figure 2 Gain versus pump power considering a co-propagating pump schemes @ 980 nm and @ 1480 nm
wavelength pump

The layout considering backward or counter-propagating pump scheme is shown in

Figure 3. Figure 4 shows the gain versus pump power when 980 nm and 1480 nm are
considered as wavelength pump. The influence of the selected pump scheme in the
amplifier performance can be checked considering unsaturated or saturated regime
just changing the signal input power range.

432
 PUMPING REQUIREMENTS

Figure 3 Amplifier setup in counter-propagating pump scheme used with 980 nm and 1480 nm wavelength
pump

Figure 4 Gain versus pump power considering a co-propagating pump schemes @ 980 nm and @ 1480 nm
wavelength pump

433
PUMPING REQUIREMENTS

The layout Bidirectional pump @ 980 nm added to this project file considers the
bidirectional pumping configuration, where the amplifier is pumped in both directions
simultaneously by using to semiconductor pump lasers located at the two fiber ends.
The advantage of this pump scheme can be checked setting small-signal or large
signal input and observing the amplifier performance given by the gain, output power
and noise figure graphs displayed in Views.

434
 TRANSIENT CONTROL IN EDFAS

Transient Control in EDFAs

Our goal here is to demonstrate how to use OptiSystem to simulate transient control
schemes to avoid or minimize the impacts of transient effects in the system
performance.
As is well known, failures in some channels and/or add-drop of channels due to
network reconfiguration can cause power transients that cause error bursts, which are
unacceptable to service providers.
To overcome this problem, several schemes to protect amplified networks against
power transients have been developed. Here we have worked on one control
scheme: EDFA gain clamping, which will be demonstrated in the next section.
In the examples simulated here, we have used a 16-channel WDM system. The
addition and dropping of 8 channels are simulated by modulating the optical signal by
a square wave at a low bit rate of 4K bits/s.
The transmitter setup configuration and the optical signal waveform at modulator
output are shown in Figure 1.
Eight signals are dropped at t = 1.5 ms and added back at t = 5.5 ms. The signal power
for each channel is -20 dBm (10 mW) and the modulated and surviving channels have
the channel spacing of 1.6 nm between each other.
The sample rate used in the signals was enough to allow the delay due feedback loop
to be considered in the simulations.

Figure 1 Transmitter setup configuration with 16 channels, where 8 channels are modulated to represent
add-drop of channels in the system. One modulated optical signal is shown in the optical vizualizer.

435
TRANSIENT CONTROL IN EDFAS

Gain clamping
There are two basic configurations of EDFA gain clamping [1].
In the first configuration, the Fabry-Perot laser structure is used with two band-pass
reflectors placed in the input and output ends of the amplifiers. The laser wavelength
is determined by the center wavelength of the reflectors.
The laser cavity loss is determined by the peak reflectivity of the reflectors and losses
between the two reflectors.
One example of this configuration can be seen in Figure 2
Two components appear in this setup. They were created to allow the
multiwavelength time-driven simulations, and consequently the transient control
simulations.
The "Convert To Samples" component receives the optical channels in this input port
and transforms these block signals in individual samples that are launched forward
sample by sample in the correspondent time step.
The other component is the "Convert From Samples", which does the opposite. It
receives the individual samples of each channel and assembles the different
channels.

Figure 2 Gain-clamping EDFA in a Fabry-Perot laser configuration

We have run the simulation with the drop and add of 8 channels at 1.5 ms and 2.5 ms,
respectively. The EDFA was designed to give a gain of approximately 19 dB for each
channel. Simulation results for the surviving channel at 1540.3 nm can be seen in the
Optical Time Domain Visualizer show in Figure 3 Small oscillations are present
exactly at the point where the channels are dropped and added.

436
 TRANSIENT CONTROL IN EDFAS

Figure 3 Optical surviving channel at 1540.3 nm after the EDFA in a Fabry-Perot laser configuration

In the second configuration, the ring laser structure is used.

Part of the signal and ASE is coupled in one end of the amplifier and fed back to the
amplifier input through a feedback loop. In the feedback path, the signals (including
ASE) are filtered and attenuated to control the lasing conditions.
Figure 4 shows the gain clamping EDFA setups with the feedback loop including a
band-pass filter and an attenuator.

Figure 4 Gain-clamped EDFA in a ring laser configuration

The same add-drop simulation was done for the gain-clamped EDFA ring laser. The
EDFA was designed to give a gain of approximately 18 dB.
Figure 5 presents 4 surviving channels after the EDFA amplifier. We can verify the
oscillations caused by the drop and add of channels and see that the oscillations are
higher and faster when the channels are dropped.

437
TRANSIENT CONTROL IN EDFAS

Figure 5 Output signal power of 4 surviving channels

The simulations results presented here are in agreement with previous results
presented in the literature. Simulations are not limited to these control schemes.
Variations in the previous control schemes can be introduced.

References
[1] A. Yu and M. J. Mahony. "Design and Modeling of Laser-Controlled Erbium-Doped Fiber
Amplifiers". IEEE Journal of Selected Topics in Quantum Electronics, Vol. 3, Issue: 4, Aug. 1997
pp.1013 - 1018.

438
 AMPLIFIER CHARACTERISTICS

Amplifier Characteristics

There are many decisions to make regarding system implementation of specific

amplifiers.
To make these decisions, a group of typical key characteristics must be extracted
from the amplifier. For example, the gain and noise figure versus the input signal
wavelength, input signal power or input pump power.
To extract the key characteristics of the optical amplifier, OptiSystem allows the user
to easily iterate over these parameters and characterize the amplifier design.
The project "Amplifier Characteristics.osd" shows how to obtain the gain, noise figure
and output power versus input power and signal wavelength.
The calculations are based on results from a WDM Analyzer inserted between two
isolators. The user can replace the amplifier between these isolators and obtain the
same curves for different amplifiers.

Figure 1 Project "Amplifier Characteristics.osd"

439
AMPLIFIER CHARACTERISTICS

Loading the project

• Go to the OptiSystem sample files folder, "…Optiwave Software\OptiSystem
OptiSystem 9\Samples\Optical amplifiers".
• Open the Project "Amplifier Characteristics.osd".
This project has three layout versions
• signal wavelength
• input power
• pump power
In the version "Signal Wavelength", the iterations are swept over the laser signal
wavelength and generate the following graphs:
• Gain x Wavelength
• Noise Figure x Wavelength
• Output Power x Wavelength
In the version "Input Power", the iterations are swept over the laser signal power and
generate the following graphs:
• Gain x Input Power
• Noise Figure x Input Power
• Output Power x Input Power
In the version "Pump Power", the iterations are swept over the pump power and
generate the following graphs:
• Gain x Pump Power
• Noise Figure x Pump Power
• Output Power x Pump Power

Running the simulation

• To run the simulation you can go to the File menu and select Calculate. You can
also press Control+F5 or use the calculate button in the toolbar. After you select
Calculate, the calculation dialog box should appear.
• In the calculation dialog box, press the Play button (Figure 2).

440
 AMPLIFIER CHARACTERISTICS

Figure 2 Calculating project

441
AMPLIFIER CHARACTERISTICS

Viewing the results

• Go to the report page (Figure 3).

Figure 3 Report page

Figure 4, Figure 5, and Figure 6 display some of the graphs that are generated for this
simulation.
Observe the Gain x Wavelength, Output Power x Input Power and Gain x Pump
Power for the amplifier.

442
 AMPLIFIER CHARACTERISTICS

Figure 4 Gain x wavelength

Figure 5 Output power x input power

Figure 6 Gain x pump power

443
AMPLIFIER CHARACTERISTICS

Notes:

444
 AUTOMATIC CONTROL SIMULATION

Automatic Control Simulation

OptiSystem allows the user to optimize the layout in a variety of ways.

User-defined optimizations allow the user to select a parameter and then select a
result. The parameter is then manipulated through a number of iterations in order to
hit the target value of the result.
The user can select a target gain or output power that is measured with the WDM
Analyzer and OptiSystem will search for the pump power to achieve the target value.

Loading the project

• Go to the OptiSystem sample files folder, "…Optiwave Software\OptiSystem
9\Samples\Optical amplifiers".
• Open the Project "Automatic Control.osd".
Figure 1 presents the amplifier layout.
This project has three layout versions:
• Automatic Gain Control (AGC)
• Automatic Power Control (APC)
• Automatic Peak Power Control (APPC)

Figure 1 Project "Automatic Control.osd"

445
AUTOMATIC CONTROL SIMULATION

Automatic Gain Control

The automatic gain control (AGC) maintains the EDFA gain at a fixed level during
signal perturbations or changes in the system loss.
The implementation of AGC using OptiSystem requires inserting a WDM Analyzer
between the two points of interest (to calculate the gain) and adding and configuring
optimizations.
Because we want to keep the total gain constant while controlling the pump power,
we need the WDM Analyzer (Figure 1) to calculate the gain and the optimization
engine to calculate the pump power to achieve this gain.

Automatic Power Control

The procedure to obtain Automatic Power Control (APC) is the same as that used to
obtain AGC. However, instead of monitoring the gain, we will monitor the total output
power.
We are assuming that the total output power must be kept constant by controlling the
pump power.
We need to insert a WDM Analyzer in the layout to calculate the signal output power
and set an optimization to calculate the pump power to achieve the target signal
output power.

Automatic Peak Power Control

The procedure to achieve Automatic Peak Power Control (APPC) is the same used
to obtain APC and AGC. However, instead of monitoring the total gain or total output
power, we will monitor the output power of one signal at a particular wavelength.
We are assuming that the total output power must be kept constant by controlling the
pump power. We need a probe to calculate the signal power and an optimization to
calculate the pump power to achieve this signal power.

Optimization
We will demonstrate the setup for the AGC.
For this setup, the user must select the
• parameter to be optimized (the pump power)
• result to be monitored (the total gain)
• range with a target value.
The other two types of control follow the same basic steps.
• To open the simulation setup, you can go to the File menu and select Calculate.
You can also press Control+F5 or use the calculate button in the toolbar. After you
select Calculate, the calculation dialog box should appear.
• In the calculation dialog box, press the 'Optimizations…' button (Figure 2).
• Double-click on the SPO (single parameter optimization) Optimization item.
The SPO optimization parameters are described in Figure 2.

446
 AUTOMATIC CONTROL SIMULATION

The user set the optimization type to goal attaining. The goal is to control the
parameter Power from the Pump Laser to reach the target Total Gain from the WDM
Analyzer.
Pump power varies from 15 to 165 mw, and the gain target is 25 dB.

Figure 2 Optimization set-up for AGC amplifier

447
AUTOMATIC CONTROL SIMULATION

Running the simulation

• To run the simulation, you can go to the File menu and select Calculate. You can
also press Control+F5 or use the calculate button in the toolbar. After you select
Calculate, the calculation dialog box should appear.
• In the calculation dialog box, enable 'Run all optimizations…' (Figure 3).
• Press the Play button.
• Select the Optimization tab in order to monitor the progress of the Optimizations
(Figure 3).

Figure 3 Calculating project

Viewing the results

• Go to the report page.
• Select the tabs AGC, APC or APPC to see the results.
Observe the pump power and gain graphs for the AGC amplifier (Figure 4).
The total signal power for the APC amplifier (Figure 5) and the signal power for the
APPC amplifier (Figure 6).
The AGC provides a constant gain of 25 dB, the APC provides a constant output
power of 10 dBm and the APPC a constant power of 10 dBm at 1530 nm.

448
 AUTOMATIC CONTROL SIMULATION

Figure 4 AGC amplifier, pump power and gain

449
AUTOMATIC CONTROL SIMULATION

Figure 5 APC amplifier, pump and signal power

450
 AUTOMATIC CONTROL SIMULATION

Figure 6 APPC amplifier, pump and signal channel power

451
AUTOMATIC CONTROL SIMULATION

Notes:

452
 ERBIUM-DOPED FIBER LASER

Erbium-doped Fiber Laser

Er-doped fiber lasers (EDFLs) can be viewed as EDFAs operating in the particular
regime where coherent oscillation of ASE occurs due to some feedback means.
A standard definition could be the following: EDFLs are used as sources for coherent
light signal generation, while EDFAs are used as wave-wave amplifiers for coherent
light signal regeneration.
All EDFLs can be pumped with compact, efficient, and sometimes inexpensive laser
diodes. They are compatible with different fibers and fiber optic components used in
communications so they have negligible coupling losses.
Fiber waveguiding and splicing alleviate any mechanical alignment of parts and
provide superior environmental stability.
There are many possible laser cavity designs and configurations.
The tunable EDFL configuration presented here uses an all-fiber ring laser cavity.
Wavelength selectivity can be achieved by using a tunable transmission filter.
This example will show how OptiSystem can simulate laser ring design and ASE
sources.
The setup parameters are very important to obtain a steady state output power. This
means the user must find the minimum number of convergence iterations to obtain
correct results.

453
ERBIUM-DOPED FIBER LASER

Figure 1 Project "Ring Laser.osd"

454
 ERBIUM-DOPED FIBER LASER

Loading the project

• Go to the OptiSystem sample files folder, "…Optiwave Software\OptiSystem
9\Samples\Optical amplifiers".
• Open the Project "Ring Laser.osd".
Figure 1 displays the laser layout.
Observe that the laser wavelength is selected, which changes the filter center
wavelength, and the laser linewidth is selected, which changes the filter bandwidth.
This laser will generate power at 1550 nm with 1 nm of linewidth (Figure 2).

Figure 2 Filter component parameters

455
ERBIUM-DOPED FIBER LASER

The global parameter Iterations (Figure 3) is very important to obtain a steady state
output power. The user must find the minimum number of iterations to obtain correct
results.

Figure 3 Global parameters: iterations

Running the simulation

• To run the simulation you can go to the File menu and select Calculate. You can
also press Control+F5 or use the calculate button in the toolbar. After you select
Calculate, the calculation dialog box should appear.

Viewing results
• Double click on Optical Spectrum Analyzer.
• Increase signal index parameter to 30 (max value - same as global parameter
Iterations).
Figure 4 shows the signal at the coupler output port.
The user can increase the signal index parameter from 0 to 30, and see the signal
changing until it reaches a steady-state value around 25 iterations.

456
 ERBIUM-DOPED FIBER LASER

Figure 4 Signal at the coupler output

457
ERBIUM-DOPED FIBER LASER

Notes:

458
 REFLECTIVE AMPLIFIER

Reflective Amplifier

In Reflective EDFAs, a device (e.g. mirror, grating) is placed at the EDFA output end
to reflect whether the pump or the signal (double-passing the pump or signal through
the EDFA) produces a net gain enhancement.
There are three basic configurations:
• EDFA with reflected pump only
• EDFA with reflected signal only
• EDFA with reflected pump and signal
The configuration presented here is with reflected signal only. The signal is reflected
using a reflective filter.

Loading the project

• Go to the OptiSystem sample files folder, "…Optiwave Software\OptiSystem
9\Samples\Optical amplifiers".
• Open the Project "Reflective Amplifier.osd".
Figure 1 displays the amplifier layout.
Observe that a circulator and a reflective filter are responsible for sending the signal
to the fiber and reflecting it back.

459
REFLECTIVE AMPLIFIER

Figure 1 Project "Reflective Amplifier.osd"

460
 REFLECTIVE AMPLIFIER

Observe that the reflective signal wavelength is selected by changing the filter center
wavelength at the reflective filter properties dialog box (Figure 2).

Figure 2 Filter component parameters

The global parameter Iterations (Figure 3) is very important to obtain a steady state
output power.
The user must find the minimum number of iterations to obtain correct results.

461
REFLECTIVE AMPLIFIER

Figure 3 Global parameters: iterations

Running the simulation

• To run the simulation, you can go to the File menu and select Calculate. You can
also press Control+F5 or use the calculate button in the toolbar. After you select
Calculate, the calculation dialog box should appear.

Viewing results
• Double click on the visualizers (Dual Port WDM Analyzer and Optical Spectrum
Analyzer).
• Increase signal index parameter to 20 (max value - same as global parameter
Iterations).
Figure 4 displays the WDM Analyzer showing the Gain.
The signal can be followed with the OSA tool to verify how the signal is changing
along the amplifier.
Figure 5 shows the signal at the coupler output port.

462
 REFLECTIVE AMPLIFIER

Figure 4 WDM Analyzer

Figure 5 Signal spectrum at the coupler output

463
REFLECTIVE AMPLIFIER

Notes:

464
 SPLIT-BAND AMPLIFIER

Split-Band Amplifier

The gain peak shifts to longer wavelengths as the fiber length increases and the
population inversion decreases.
The bandwidth of an EDFA can be increased by separating the gain spectrum into two
bands, 1530-1560 nm and 1570-1610 nm, with a band-splitting filter. The gain can
then be optimized in the two regions separately.

Loading the project

• Go to the OptiSystem sample files folder, "…Optiwave Software\OptiSystem
9\Samples\Optical amplifiers".
• Open the Project "s Amplifier.osd".
Figure 1 displays the project.
The first path was optimized to work in the 1530-1565 nm region, and the second path
was optimized to work in the 1570-1605 nm region.

Figure 1 Project "Split Band Amplifier.osd"

465
SPLIT-BAND AMPLIFIER

The 3 Port Filter component splits the signal in the two paths.

In Figure 2, the 3 Port filter properties dialog boxes has the center wavelength at
1587.5 nm and 35 nm of bandwidth. The first port will have the rectangular transfer
function and the second will have the complementary transfer function.

Figure 2 Filter component parameters

466
 SPLIT-BAND AMPLIFIER

Running the simulation

• To run the simulation, you can go to the File menu and select Calculate. You can
also press Control+F5 or use the calculate button in the toolbar. After you select
Calculate, the calculation dialog box should appear.

Viewing results
• Double click on the visualizers (Dual Port WDM Analyzer and Optical Spectrum
Analyzer).
• Increase signal index parameter to 7 (max value).
Figure 3 displays the WDM Analyzer showing the Gain.
Note that the gain is approximately 20 dB (1530-1575nm) for the first path, and 11 dB
(1580-1605nm) for the second path.
The signal can be followed with the OSA tool to verify how the signal is changing
along the amplifier.
Figure 4 shows the signal at the coupler output port.

Figure 3 WDM Analyzer

467
SPLIT-BAND AMPLIFIER

Figure 4 Signal spectrum at the coupler output

468
 DYNAMIC AMPLIFIER USING YTTERBIUM-DOPED FIBER

Dynamic Amplifier Using Ytterbium-Doped Fiber

The objective of this lesson is to demonstrate the amplification of a sequence of

Gaussian pulses by an ytterbium-doped fiber amplifier using the dynamic model.
To analyze the dynamic characteristics of the high-power ytterbium-doped double-
clad fiber amplifier considering pulses in repetition rates of KHz, a dynamic ytterbium-
doped fiber model is necessary.
In this case, a sequence of 5 Gaussian pulses with a repetition rate of 100 KHz was
used.
Figure 1 shows the system designed for this simulation. See sample
Ytterbium_Double_Clad_Dynamic_100KHz.osd.

Figure 1 System layout of the fiber amplifier

The input pulse peak power is 50 W and the energy in the pulse is 1uJ. The optical
signal in the amplifier input with the sequence of Gaussian pulses is shown in Figure
2(a)

469
DYNAMIC AMPLIFIER USING YTTERBIUM-DOPED FIBER

Figure 2 Optical signal at the amplifier (a) input and (b) output

(a)

(b)
The optical signal amplified is shown in Figure 2(b).

470
 DYNAMIC AMPLIFIER USING YTTERBIUM-DOPED FIBER

The first output pulse exceeds the successive ones because the energy stored in the
amplifier was not recovered before the successive pulse.

The first output pulse can be seen in Figure 3.

Figure 3 First output pulse

471
DYNAMIC AMPLIFIER USING YTTERBIUM-DOPED FIBER

Notes:

472
 GAIN FLATTENING FILTER OPTIMIZATION

Gain Flattening Filter Optimization

EDFAs have a wavelength-dependent gain; i.e., some wavelengths are amplified

more than others. A gain flattening filter restores all wavelengths to approximately the
same intensity. Project Gain Flattening Filter Optimization.osd demonstrates the
gain flattening filter optimization engine that is included with OptiSystem.
The system layout is presented in Figure 1. The WDM Transmitter generates eight
channels from 195 THz to 196.4 THz, with power of -20 dBm per channel. The Gain
Flattening Filter component is placed after the EDFA and it will equalize the amplifier
gain.

Figure 1 System layout of the fiber amplifier

Figure 2 shows the steps to access OptiSystem optimization dialog box.

473
GAIN FLATTENING FILTER OPTIMIZATION

Figure 2 Accessing the optimizations

Figure 3 shows the main parameters of the optimization. The optimization will run 10
times, targeting a gain ripple of 0.1 dB from 1500 to 1600 nm. The minimum
transmission value for the filter is -40 dB and the maximum value is -0.1 dB. The
transmission values of the filter will be optimized at the user defined frequencies
defined in the filter. The components tab should have the filter component selected
and the visualizers tab should have the dual port analyzer selected.
In order to run the optimization, simply calculate the project. Make sure the 'Run all
optimizations' check box is enabled. During the calculation the user can click on the
Optimization tab and visualize the progress of the optimizations (Figure 4).

474
 GAIN FLATTENING FILTER OPTIMIZATION

Figure 3 Optimization main parameters

475
GAIN FLATTENING FILTER OPTIMIZATION

Figure 4 Optimizing the filter

In this particular project we have the signal before and after the gain flattening filter
available in the report page (Figure 5). The shape of the filter is also available in the
report. The user can also visualize the overall gain of the amplifier using the WDM
analyzer and the value of the filter transmission values by looking at the filter
‘Transmission’ parameters.

476
 GAIN FLATTENING FILTER OPTIMIZATION

Figure 5 Signals before and after the filter:

477
GAIN FLATTENING FILTER OPTIMIZATION

Notes:

478
 AMPLIFICATION OF MULTIPLE MODES IN ER-DOPED MULTIMODE FIBERS

Amplification of multiple modes in Er-doped

multimode fibers

The Er-doped Multimode Fiber component was designed to allow the analysis of
multiple modes amplification in large mode area fibers. The components are very
similar to the original single-mode doped fiber amplifiers, however now the refractive
index information and ion distribution profile can be defined by the user (see main tab
parameters in the figure below).

Figure 1 Doped fiber parameters

Based on the parameters specified, the component's mode solver calculates the
modes supported for the signal, pump and ASE bands. The user can also define if all
supported modes will be equally excited ( for pump and/or signal channels) or the
power distribution ratio between the supported modes will be calculated based on
coupling coefficients, in this case an initial optical excitation field has to be defined for
pump and/or signal.

The following example shows an erbium-doped fiber (double clad fiber), where all
signal modes supported are equally excited (MM EDFA
ExampleAllModesEquallyExcited.osd).

479
AMPLIFICATION OF MULTIPLE MODES IN ER-DOPED MULTIMODE FIBERS

Figure 2 System layout.

The refractive index profile (step index) and Er ion distribution (uniform) used are
displayed in the figures below:

Figure 3 (a) Refractive index profile and (b) Ion distribution.

After running the simulation, the following results were obtained:

Total Signal, Pump, and ASE power evolution along the fiber length.

480
 AMPLIFICATION OF MULTIPLE MODES IN ER-DOPED MULTIMODE FIBERS

Figure 4 Total Signal, Pump, and ASE power evolution along the fiber length.

The fiber at 1550 nm (signal wavelength) supports 3 guided modes: LP01 and LP11
(two helical polarities).

The normalized radial intensity (azimuthal angle is 0) for LP01 (black line) and LP11
(red line) are displayed in the figure.

481
AMPLIFICATION OF MULTIPLE MODES IN ER-DOPED MULTIMODE FIBERS

Figure 5 Radial profiles.

The power evolution along the fiber length for LP01 (black line) and LP11 - sum of the
sin and cos components - (red line) were also calculated.

Figure 6 Power evolution.

Based on the modes power distribution at the fiber output, the profile of the output
beam is displayed:

482
AMPLIFICATION OF MULTIPLE MODES IN ER-DOPED MULTIMODE FIBERS

Figure 7 Output beam.

483
AMPLIFICATION OF MULTIPLE MODES IN ER-DOPED MULTIMODE FIBERS

Notes:

484
 S-BAND AMPLIFICATION USING TM-DFA

S-band amplification using Tm-DFA

The system designed amplifies a set of 16 channels in the S-band going from 1450
nm to 1525 nm. Three different amplifier lengths are simulated and the gain curve is
plotted in the report page.

Sample: MODB_ModulationSystem.osd

Figure 1 Thulium-doped fiber amplifier layout.

485
S-BAND AMPLIFICATION USING TM-DFA

Figure 2 Gain curves for 2.4 m (black), 7 m (red), and 8.1 m (green) of Tm-doped fibers.

486
 ULTRASHORT PULSE FIBER LASER

Ultrashort Pulse Fiber Laser

The system demonstrates a fiber laser, based on passive mode locking using a
saturable absorber, capable of producing ultrashort pulses

Sample: UltraShortFiberlaser.osd

The system layout presented in the figure below is a unidirectional ring cavity
composed of an optical fiber amplifier (here represented by a gain fiber numerical
modeling that is emulating an ytterbium-doped fiber), an undoped fiber and a DC
component that can be varied to change the net cavity dispersion. The saturable
absorber is used to achieve mode locking and generate the optical pulse. The white
light source provides the noise that initialize the pulse generation.

Figure 1 Ultrashort pulse fiber laser layout.

The system run multiple iterations and the convergence monitor component stops the
calculation once the output pulse reaches a stable condition.

487
ULTRASHORT PULSE FIBER LASER

Figure 2 Output (a) spectrum and (b) time-domain signal after first iteration

Figure 3 Output (a) spectrum and (b) time-domain signal after convergence in the results (1293 iterations).

488
Optiwave
7 Capella Court
Ottawa, Ontario, K2E 7X1, Canada

Tel.: 1.613.224.4700
Fax: 1.613.224.4706

E-mail: [email protected]
URL: www.optiwave.com