White Paper Introduction To EDFA Technology: 2.1 Physical Principle of EDF Amplification
White Paper Introduction To EDFA Technology: 2.1 Physical Principle of EDF Amplification
White Paper
2 EDFA Basics
2.1 Physical Principle of EDF Amplification
At the heart of EDFA technology is the Erbium Doped Fiber (EDF), which is a conventional Silica fiber doped
with Erbium ions. When the Erbium ions are illuminated with light energy at a suitable wavelength (either
980nm or 1480nm) they are excited to a quasi-stable long lifetime intermediate state (see Figure 1),
following which they decay back to the ground state by emitting a photon within the 1525-1565 nm band. If
photons already exist within this wavelength band, for example due to a signal channel passing through the
EDF, then these photons can stimulate the decay process (so called stimulated emission), resulting in
additional photons with the same wavelength and characteristics as the signal photons. Thus, if a pump
wavelength and a signal wavelength are simultaneously propagating through an EDF, energy transfer will
occur via the Erbium ions from the pump wavelength to the signal wavelength, resulting in signal
amplification.
RED-C Optical Networks, Ltd., Atidim Technology Park, Bldg. 3, P.O. Box 58101 Tel-Aviv 61580, Israel, sales@red-c.com
ww.red-c.com
Introduction to EDFA Technology
Unstable short-
lif etime state
980nm
pump
Quasi-stable
intermediate state
1480nm
pump
1525-1565 nm
Ground state
Figure 1: The energy levels of the Erbium ions that participate in the amplification process. The Erbium ions
can be either pumped by 980nm photons, in which case the ions pass through an unstable short lifetime
state before rapidly decaying to a quasi-stable state, or by 1480nm photons in which case they are directly
excited to the quasi-stable state. Once in the quasi-stable state, they decay to the ground state by emitting a
photon in the 1530-1560nm band. This decay process can be stimulated by pre-existing photons, thus
resulting in amplification.
EDF
Pump GFF
Combiner
980nm
Input Output
Pump
Detector Detector
Control Unit
Figure 2: Diagram of a typical single stage EDFA, showing the three basic components (EDF, pump and
WDM combiner), as well as additional optical and electronic components used to optimize performance and
control the amplifier.
Besides the three basic components described above, Figure 2 also shows additional optical and electronic
components used in a basic single stage EDFA. The signal enters the amplifier through the input port, and
then passes through a tap which is used to divert a small percentage of the signal power (typically 1-2%) to
an input detector. The signal then passes through an isolator, before being combined with pump energy
emitted by the 980nm pump laser diode. The combined signal and pump energy propagate along the EDF,
where signal amplification occurs, and then the amplified signal exits the EDF and passes through a second
isolator. The purpose of the two isolators, which allow light to pass only in a single direction, is to ensure that
lasing cannot take place within the EDF, which would then lead to instability of the output power.
Furthermore, the output isolator also acts as a filter for 980nm light propagating in the forward direction, thus
stopping the 980nm light from exiting the amplifier output port.
In a multi-channel WDM amplifier, a Gain Flattening Filter (GFF) is usually placed following the output
isolator in order to flatten the gain spectrum, as shown in Figure 3. The attenuation spectrum of the GFF is
designed to match the Gain spectrum of the EDF (operating at a given fixed gain), such that the combination
of the two produces a flat gain. Following the GFF the signal passes through an output tap used to divert a
small percentage of the output power (typically 1-2%) to the output detector. The output and input detectors
are used to monitor the input and output power respectively, and thus provide feed-back to the control unit,
which controls the amplifier by setting the pump laser current, and thus the amount of pump power injected
into the EDF. The control unit also provides external communication, for example via an RS232 interface.
n
r Output from EDF io
t GFF r Flat Gain
e a e
w u w
o n o
P e
tt P
A
Figure 3: Use of a Gain Flattened Filter to achieve a flat gain spectrum. The attenuation spectrum of the
filter matches the gain spectrum at the output of the EDF, such that the combination of the two provides a flat
gain spectrum.
Input Output
Detector Detector
Pump
FF loop
G = P_out/P_in
FB loop
Figure 4: Schematic diagram of Automatic Gain Control (AGC) in an EDFA. The control loop consists of a
Feed Back (FB) loop for maintaining stable gain, as well as a Feed-Forward (FF) loop for fast transient
suppression.
Figure 5: Output power of a single channel (blue curves) in response to a change in the total input power
(magenta curves). The maximum transient deviation of the output signal is <1dB, while the setting time
before stable operating is resumed is <100μs. Left – Transient response in the case of a sharp decrease in
input power (e.g. channel drop). Right - Transient response in the case of a sharp increase in input power
(e.g. channel add).
VOA GFF
Gain Gain
Stage I Stage II
Control Unit
Figure 6: Left: Basic design of a WDM Variable Gain EDFA, where the VOA is placed between two gain
sub-stages in order to improve NF. Right: Resulting improvement of NF compared to the case of a fixed gain
EDFA with a VOA placed beforehand. As can be seen, the variable gain EDFA provides much better NF
performance over the entire gain range.
Dual stage amplifiers are basically two amplifiers in one package, where there is access for an optical
component such as a DCM to be placed between them, as is illustrated in Figure 7. Most often the first
amplifier (the pre-amplifier) is variable gain, and the second amplifier (the booster) is fixed gain, such that the
amplifier as a whole provides variable gain operation. The control of both amplifiers is combined – in other
words the user sets the required net gain of the entire combination (including the DCF), and the control units
sets the gain of each of the two amplifier in order to achieve the net gain. The amplifiers are designed a-
priori to take into account the DCM loss. For example, the dynamic range of the input detectors of both
amplifiers is set accordingly, and the optical performance, such as NF, is specified already taking the DCM
loss into account. Since the DCM is often implemented using special Dispersion Compensating Fiber (DCF),
there can be a large optical delay between the first and the second stage of the amplifier. For this reason
transient suppression of each amplifier needs to be performed separately, and consequently each amplifier
has its own pump and own local control mechanism (in addition to the overall control used to set the net
gain).
Mid Stage
Access
Gain Gain
Line In Line Out
Stage I Stage I
Control
Figure 7: The basic scheme for a dual stage EDFA with mid stage access. The two amplifiers are packaged
in the same module and are controlled together with the mid-stage device (e.g. DCM) as a single unit.
Additionally, each amplifier also has its own local control loops.
4 Conclusion
Of the various technologies available for optical amplifiers, EDFA technology is by far the most advanced,
and consequently the vast majority of optical amplifiers deployed to date are based on EDFA technology.
Today it is possible to build broadband WDM EDFA’s which provide flat gain over a large dynamic gain
range, low noise, high saturation output power, and stable operation with excellent transient suppression.
These features are provided in small modules measuring <100x100 mm, and with power consumption of just
a few Watts. This combination of good reliable performance, together with relatively low cost, allows EDFAs
to address most application and functions in modern optical networks.