Dense Wavelength

Division Multiplexing
Basics (DWDM)
 ‫قطاع التدريب والتطوير‬
‫اإلدارة العامة لتطوير المهارات الفنية ونظم وتكنولوجيا المعلومات‬

‫‪Dense Wavelength‬‬
‫‪Division Multiplexing‬‬
‫)‪Basics (DWDM‬‬
‫‪Code: PD0104000020202‬‬
 1 WDM Introduction Pages (1-27)
Sub -
Sections

2 Linear Effects Pages (1-31)

DWDM
3 Non-Linear High-Power Effects Pages (1-15)

4 Optical Devices Pages (1-73)

5 Measurements Pages (1-10)

This document consists of 156 pages.

Chapter 1 : WDM Introduction

Chapter 1: WDM Introduction

Aim of study
This chapter introduces basis of DWDM.

Contents Pages

1.1 Why WDM 2

1.2 What is WDM 4
1.3 Varieties of WDM 5
1.4 Types of Multiplexing 7
1.5 Wave Length Plan 9
1.6 kinds of optical fiber 16
1.7 Light Propagation in Multimode Fiber 22

1
Chapter 1 : WDM Introduction

Chapter1

WDM Introduction

1.1 Why WDM

a- The difficulties of technology

As we know, signal transmission through fiber-optic systems becomes

increasingly difficult as the data rate on an optical channel increases. Dispersion
effects become more significant at higher speeds, and can limit transmission
distances, depending on the type of fiber. The faster the channel rate, the shorter
the distance signals can travel. Thus 2.5Gbps signal can go farther than 10Gbps
signals, and 40Gbps signals cannot go as far as 10Gbps. Thus one way to
achieve higher overall transmission rates over the same distance is to break the
signal into many parallel optical channels and transmit them at different
wavelengths through the same fiber.

Figure 1

2
Chapter 1 : WDM Introduction

b- Cost

Reduce the cost of network and save the fiber, WDM uses one fiber instead of
many fiber used by some single-channel systems.

WDM uses OA instead of REG. as we know, REG equipments are expensive,

different signal format and bit rate have to use different REG.

Figure 2

c- High Bandwidth Demand:

- Bandwidth are doubling every 3 months

- Internet traffic increases thousand-fold every 3 years

- upgrade of existing fiber networks (without adding fibers)

3
Chapter 1 : WDM Introduction

Figure 3

d- Transparency

Each optical channel can carry any transmission format (different asynchronous
bit rates, analog or digital)

e- Scalability

Buy and install equipment for additional demand as needed

1.2 What is WDM

WDM (Wavelength-division Multiplexing): WDM is a fundamental passive

optical component for optical system. Sending several signals through one
WDM with different wavelengths of light. By its special physical character, for
instance, a fused dual-channel WDM will be able to double the data transmission
bandwidth at very low cost.

4
Chapter 1 : WDM Introduction

Figure 4

1.3 Varieties of WDM

Early WDM systems transported two or four wavelengths that were widely
spaced.

WDM and the “follow-on” technologies of CWDM and DWDM have evolved
well beyond this early limitation.

a- WDM

Traditional, passive WDM systems are wide-spread with 2, 4, 8, 12, and 16

channel counts being the normal deployments. This technique usually has a
distance limitation of under 100 km.

b- CWDM

Today, coarse WDM (CWDM) typically uses 20-nm spacing (3000 GHz) of up
to 18 channels. The CWDM Recommendation ITU-T G.694.2 provides a grid of
wavelengths for target distances up to about 50 km on single mode fibers as
specified in ITU-T Recommendations G.652, G.653 and G.655. The CWDM
grid is made up of 18 wavelengths defined within the range 1270 nm to 1610 nm
spaced by 20 nm.

5
Chapter 1 : WDM Introduction

Figure 5

c- DWDM

Dense WDM common spacing may be 200, 100, 50, or 25 GHz with channel
count reaching up to 128 or more channels at distances of several thousand
kilometers with amplification and regeneration along such a route.

Figure 6

6
Chapter 1 : WDM Introduction

WDM Wavelengths features:

1. We used C_band (1530: 1565 nm) because it faced minimum attenuation.

2. We divided this band into 40 channels with separation 100G HZ for WDM.

3. We divided this band into 80 channel with separation 50G HZ for DWDM.

4. 50G HZ is the minimum separation without interference, and if we want to

increase channels we must use another band (L _band).

1.4 Types of Multiplexing

Multiplexing is sending multiple signals or streams of information through a

circuit at the same time in the form of a single, complex signal and then
recovering the separate signals at the receiving end. Basic types of multiplexing
include frequency division (FDM), time division (TDM), and wavelength
division (WDM), with TDM and WDM being widely utilized by telephone and
data service providers over optical circuits.

a- Time Division Multiplexing

Time-division multiplexing (TDM), as represented in Figure 7, is a method of

combining multiple independent data streams into a single data stream by
merging the signals according to a defined sequence. Each independent data
stream is reassembled at the receiving end based onthe sequence and timing.
Synchronous Optical Network (SONET), Asynchronous Transfer Mode (ATM)
and Internet Protocol (IP) utilize TDM techniques. In modern
telecommunications networks, TDM signals are converted fromelectrical to
optical signals by the SONET network element, for transport over optical fiber.

7
Chapter 1 : WDM Introduction

Figure 7

b- Wavelength Division Multiplexing

WDM combines multiple optical TDM data streams onto one fiber through the
use of multiple wavelengths of light. Each individual TDM data stream is sent
over an individual laser transmitting a unique wavelength of light.

Figure 8

8
Chapter 1 : WDM Introduction

c- FDM
Frequency Division Multiplexing (FDM) is a networking technique in which
multiple data signals are combined for simultaneous transmission via a shared
communication medium. FDM uses a carrier signal at a discrete frequency for
each data stream and then combines many modulated signals.
When FDM is used to allow multiple users to share a single physical
communications medium (i.e. not broadcast through the air), the technology is
called frequency-division multiple access (FDMA).

Figure 9

1.5 Wave Length Plan

a- Transmission Windows

If one looks at the possible wavelengthsfor the transmission of signals one has to
look at the fiber properties. Optical fibers are not suitable for transmission at all
wavelengths but only in certain windows. Today, usually the second
transmission window (around 1300nm) and the third and fourth transmission
windows from 1530 to 1565nm (also called Conventional Band) and from 1565
to1620nm (also called Long Band).are used. Technological reasons limit
DWDM applications at the moment to the third and fourth window. The losses
9
Chapter 1 : WDM Introduction

caused by the physical effects on the signal due by the type of materials used to
produce fibers limit the usable wavelengths to between 1280nm and 1650nm.
Within this usable range the techniques used to produce the fibers can cause
particular wavelengths to have more loss so we avoid the use of these
wavelengths as well

Figure 10

10
Chapter 1 : WDM Introduction

Figure 11

b- ITU Wavelength Plan:

Within those windows the ITU has defined in G.692 a wavelength plan for
DWDM systems to use. In fact, not the wavelengths have been defined, but the
frequencies. This does not matter though as the frequency f and the wavelength
λare connected by the relation:

C = f*λ

Where c is the speed of light.

These defined frequencies are given by the equation:

f=193.1 ±m*0.05THz

11
Chapter 1 : WDM Introduction

Which means that the ITU in G.692 initially uses a 50GHz grid. There are also
proposals for 100GHz and 200GHz spacing or even for unequal channel spacing
for specific applications. The wider spacing is easier to handle, but some of the
systems existing or planned are in fact already using 50GHz spacing or even
below. Further on, additionally to the C-Band the L-Band will be used in the
future.

c- Optical Supervisory Channel :

In addition to those "working" wavelengths another set of wavelengths (1510 or

alternatively 1480 and 1310nm) is reserved for use as the optical supervisory
channel, an additional optical channel connecting the DWDM network elements
together and fulfilling approximately the same purpose as the SDH overheads.

Figure 12

12
Chapter 1 : WDM Introduction

Figure 13

d- Bandwidth and wavelength

There is a correlation between the frequency f, the propagation velocity v

(„phase velocity“) and the wavelength λ. In this the frequency f is determined
bythe processes during thegeneration of radiation. The medium, in which the
wave is propagating, determines the phase velocity v. Consequently, the wave-
length λ is no independent quantity. It results from the frequencyand the phase
velocity .Thus the light has the same frequency but different wavelengths
indifferent substrates. For the propagation in a vacuum it is:

C * f= λ

In this, c is the vacuum velocity of light and λ the wavelength in the vacuum. All
correlations stated in the following between frequency andwavelengths refer to
the wavelengths in the vacuum.

13
Chapter 1 : WDM Introduction

In principle, with the standard singlemode fiber for telecommunicationa

wavelength range of approx. λ1=1280nm to λ2=1650nm can be utilized. In this,
the lower wavelength limitresults from the core diameter of the singlemode
fiber.The upper wavelength limitresults from the fact that above this limit the
attenuation coefficient rapidly increases and the fiber gets very sensitive
regarding macro bending. Corresponding to the pervious equation the resulting
usable wavelength rangeis from f1=235THz to f2=182THz. In this,
THz=Terahertz oscillations per second. Thus the intrinsic transmission capacity
of the singlemode fiber is:

Intrinsic f THz = 1253 transmission capacity

This transmission capacity is often called „bandwidth of the fiber“. From the
equation it is obvious that the transmission capacity of the singlemode fiber is
only used at a very small scale at present. A 2,5Gbit/s signal, for example, only
uses this bandwidth capacity with 0,005% and a 10Gbit/s signal with 0,02%! It
is obvious that the transmission capacity of a single mode fiber can be exploited
much better by a simultaneous transmission of several.

14
Chapter 1 : WDM Introduction

e- Wavelengths

Figure 14

Figure 15

15
Chapter 1 : WDM Introduction

Figure 16

1.6 kinds of optical fiber

Figure 17: Fiber Modes

16
Chapter 1 : WDM Introduction

Figure 17 illustrates the three different kinds of optical fiber:

 Multimode Step-Index
 Multimode Graded-Index
 Single-Mode (Step-Index)

The difference between them is in the way light travels along the fiber. The top
section of the figure shows the operation of “multimode” fiber. There are two
different parts to the fiber. In the figure, there is a core of 50 microns (µm) in
diameter and a cladding of 125 µm in diameter. The cladding surrounds the core.
The cladding glass has a different (lower) refractive index than that of the core,
and the boundary forms a mirror.

Light is transmitted (with very low loss) down the fiber by reflection from the
mirror boundary between the core and the cladding. This phenomenon is called
“total internal reflection”. Perhaps the most important characteristic is that the
fiber will bend around corners to a radius of only a few centimeters without any
loss of the light.

Multimode Step-Index Fiber

Figure 18: Multimode Step-Index Fiber

What happens is that there is only a finite number of possible paths for the light
to take. These paths are called “modes” and identify the general characteristic of
the light transmission system being used.

17
Chapter 1 : WDM Introduction

Fiber that has a core diameter large enough for the light used to find multiple
paths is called “multimode” fiber. For a fiber with a core diameter of 62.5
microns using light of wavelength 1300 nm, the number of modes is around 400
depending on the difference in refractive index between the core and the
cladding.
The problem with multimode operation is that some of the paths taken by
particular modes are longer than other paths. This means that light will arrive at
different times according to the path taken. Therefore the pulse tends to disperse
(spread out) as it travels through the fiber. This effect is one cause of
“intersymbol interference”. This restricts the distance that a pulse can be
usefully sent over multimode fiber.

Multimode Graded Index Fiber

Figure 19: Multimode Graded index Fiber

One way around the problem of (modal) dispersion in multimode fiber is to do

something to the glass such that the refractive index of the core changes
gradually from the centre to the edge. Light travelling down the center of the
fiber experiences a higher refractive index than light that travels further out
towards the cladding. Thus light on the physically shorter paths (modes) travels
more slowly than light on physically longer paths.
The light follows a curved trajectory within the fiber as illustrated in the figure.
The aim of this is to keep the speed of propagation of light on each path the

18
Chapter 1 : WDM Introduction

same with respect to the axis of the fiber. Thus a pulse of light composed of
many modes stays together as it travels through the fiber.
This allows transmission for longer distances than does regular multimode
transmission. This type of fiber is called “Graded Index” fiber. Within a GI fiber
light typically travels in around 400 modes (at a wavelength of 1300 nm) or 800
modes (in the 800 nm band).
Note that only the refractive index of the core is graded. There is still a cladding
of lower refractive index than the outer part of the core.

Single-Mode Fiber

Figure 20: Single-Mode Fiber

Note that this figure is not to scale. The core diameter is typically between 8 and
9 microns while the diameter of the cladding is 125 microns.

If the fiber core is very narrow compared to the wavelength of the light in use
then the light cannot travel in different modes and thus the fiber is called
“single-mode” or “monomode”. There is no longer any reflection from the core-
cladding boundary but rather the electromagnetic wave is tightly held to travel
down the axis of the fiber. It seems obvious that the longer the wavelength of
light in use, the larger the diameter of fiber we can use and still have light travel
in a single-mode. The core diameter used in a typical single-mode fiber is nine
microns.

19
Chapter 1 : WDM Introduction

A significant proportion (up to 20%) of the light in a single-mode fiber actually

travels in the cladding. For this reason the “apparent diameter” of the core (the
region in which most of the light travels) is somewhat wider than the core itself.
The region in which light travels in a single-mode fiber is often called the “mode
field” and the mode field diameter is quoted instead of the core diameter. The
mode field varies in diameter depending on the relative refractive indices of core
and cladding.

Core diameter is a compromise. We can't make the core too narrow because of
losses at bends in the fiber. As the core diameter decreases compared to the
wavelength (the core gets narrower or the wavelength gets longer), the minimum
radius that we can bend the fiber without loss increases. If a bend is too sharp,
the light just comes out of the core into the outer parts of the cladding and is lost.

You can make fiber single-mode by:

 Making the core thin enough.
 Making the refractive index difference between core and cladding small
enough.
 Using a longer wavelength.

Single-mode fiber usually has significantly lower attenuation than multimode

(about half). This has nothing to do with fiber geometry or manufacture.
Single-mode fibers have a significantly smaller difference in refractive index
between core and cladding. This means that less dopant is needed to modify the
refractive index as dopant is a major source of attenuation.
It's not strictly correct to talk about “single-mode fiber” and “multimode fiber”
without qualifying it - although we do this all the time. A fiber is single-moded
or multi-moded at a particular wavelength. If we use very long wave light (say
²
10.6 nm from a CO laser) then even most MM fiber would be single-moded for

20
Chapter 1 : WDM Introduction

that wavelength. If we use 600 nm light on standard single-mode fiber then we

do have a greater number of modes than just one (although typically only about
3 to 5).
There is a single-mode fiber characteristic called the “cutoff wavelength”. This
is typically around 1100 nm for single-mode fiber with a core diameter of 9
microns.
The cutoff wavelength is the shortest wavelength at which the fiber remains
single-moded. At wavelengths shorter than the cutoff the fiber is multimode.
When light is introduced to the end of a fiber there is a critical angle of
acceptance.
Light entering at a greater angle passes into the cladding and is lost. At a smaller
angle the light travels down the fiber. If this is considered in three dimensions, a
cone is formed around the end of the fiber within which all rays are contained.
The sine of this angle is called the “numerical aperture” and is one of the
important characteristics of a given fiber.
Single-mode fiber has a core diameter of 4 to 10 µm (9 µm is typical).

Fiber Refractive Index Profiles

Figure 21: Fiber Refractive Index Profiles

Figure 21 shows the refractive index profiles of some different types of fiber.

21
Chapter 1 : WDM Introduction

1.7 Light Propagation in Multimode Fiber

Figure 22: Light Propagation in Multimode Fiber

Light is bound within the fiber due to the phenomena of “total internal
reflection” which takes place at the interface between the core of the fiber and
the cladding.

The key feature of light propagation in a fiber is that the fiber may bend around
corners. Provided the bend radius is not too tight (2 cm is about the minimum for
most multimode fibers) the light will follow the fiber and will propagate without
loss due to the bends. This phenomena is called “total internal reflection”. A ray
of light entering the fiber is guided along the fiber because it bounces off the
interface between the core and the (lower refractive index) cladding. Light is
said to be “bound” within the fiber.

Figure 23: Reflection

22
Chapter 1 : WDM Introduction

If we consider the propagation of a “ray” in a multimode step index fibre we can

analyse the situation quite easily with the “laws of elementary physics”. “The
angle of incidence is equal to the angle of reflection.” This is illustrated in
Figure 23. This means that Ɵ 1= Ɵ 2.

Figure 24: Refraction

The important thing to realise about propagation along a fibre is that not all light
can propagate this way. The angle of incidence of the ray with the core-cladding
interface must be quite small or else the ray will pass through into the cladding
and (after a while) will leave the fiber.

Snell's Law

In order to understand ray propagation in a fiber we need one more law from
high school physics. This is Snell's law. Referring to Figure 24 and Figure 25:

n1 sin Ɵ 1 = n2 sin Ɵ 2
Where n denotes the refractive index of the material.

23
Chapter 1 : WDM Introduction

Figure 25: Refraction (2)

Notice here that:

1. The angle Ɵ is the angle between incident ray and an imaginary line normal to
the plane of the core-cladding boundary. This is counter to intuition but the
accepted convention.

2. When light passes from material of higher refractive index to a material of

lower index the (refracted) angle Ɵ gets larger.

3. When light passes from material of lower refractive index to a material of

higher index the (refracted) angle Ɵ becomes smaller.

Critical Angle

Figure 26: Critical Angle (1)

24
Chapter 1 : WDM Introduction

If we consider Figure 24 we notice that as the angle. Ɵ 1 becomes larger and

larger so does the angle Ɵ 2. Because of the refraction effect Ɵ 2 becomes larger
more quickly than Ɵ 1. At some point Ɵ 2 will reach 90° while Ɵ 1 is still well less
than that. This is called the “critical angle”. When Ɵ 1 is increased further then
refraction ceases and the light starts to be reflected rather than refracted.

Thus light is perfectly reflected at an interface between two materials of

different refractive index if:

1. The light is incident on the interface from the side of higher refractive index.

2. The angle Ɵ is greater than a specific value called the “critical angle”.

If we know the refractive indices of both materials then the critical angle can be
derived quite easily from Snell's law. At the critical angle we know that Ɵ 2 equal
90° and sin 90° = 1 and so:

Figure 27: Critical Angle (2)

25
Chapter 1 : WDM Introduction

In Figure 26 we see that for rays where Ɵ 1 is less than a critical value then the
ray will propagate along the fiber and will be “bound” within the fiber. In Figure
27 we see that where the angle Ɵ 1 is greater than the critical value the ray is
refracted into the cladding and will ultimately be lost outside the fiber.

Another aspect here is that when light meets an abrupt change in refractive index
(such as at the end of a fiber) not all of the light is refracted. Usually about 4%
of the light is reflected back along the path from which it came.

In Figure 27 there is a partial reflection present when most of the light is

refracted. These reflections are called “Fresnel Reflections” and occur in most
situations where there is an abrupt change in the refractive index at a material
interface. These reflections are an important (potential) source of disruption and
noise in an optical transmission system.

Numerical Aperture (NA)

Figure 28: Calculating the Numerical Aperture

26
Chapter 1 : WDM Introduction

One of the most often quoted characteristics of an optical fiber is its “Numerical
Aperture”. The NA is intended as a measure of the light capturing ability of the
fiber.

It is clear that there is a “cone” of acceptance (illustrated in Figure 28). If a ray

enters the fiber at an angle within the cone then it will be captured and propagate
as a bound mode. If a ray enters the fiber at an angle outside the cone then it will
leave the core and eventually leave the fiber itself.

The Numerical Aperture is the sin of the largest angle contained within the cone
of acceptance.

Or:

NA is related to a number of important fiber characteristics.

1. It is a measure of the ability of the fiber to gather light at the input end (as
discussed above).

2. Because it is a measure of the contrast in RI between the core and the cladding
it is a good measure of the light guiding properties of the fiber. The higher the
NA the tighter (smaller radius) we can have bends in the fiber before loss of
light becomes a problem.

27
Chapter 2 : Linear Effects

Chapter 2: Linear Effects

Aim of study
This chapter introduces attenuation, dispersion & polarization mode dispersion.

Contents Pages

2.1 Introduction 2
2.2 Attenuation 3
2.3 Dispersion 10
2.4 Polarization-Mode Dispersion 21
2.5 Review Questions 25
2.6 Answers to Review Questions 30

1
Chapter 2 : Linear Effects

Chapter2

Linear Effects

2.1 Introduction

Loss or attenuation and pulse dispersion represent the two characteristics of an

optical fiber most important in determining the information-carrying capacity of
a fiber optic communication system. In digital communication systems,
information to be sent is first coded in the form of pulses, and these pulses of
light are then transmitted from the transmitter to the receiver, where the
information is decoded. A typical fiber optic communication system consists of
a transmitter, which could be either a laser diode or a light-emitting diode,
whose light is modulated by the signal and coupled into an optical fiber. Along
the path of the optical fiber, there are splices, which are permanent joints
between sections of fibers, and repeaters, which boost the signal and correct any
distortion that may have accumulated along the path of the fiber. At the end of
the link, the light is detected by a photodetector, which converts the optical
signals to electrical signals, which are then processed electronically to retrieve
the signal. The greater the number of optical pulses that can be sent per unit time
and still be detectable and resolvable at the receiver end, the larger will be the
transmission capacity of the system. A pulse of light sent into a fiber gets
attenuated as it propagates through the fiber, and if the loss is large, there would
not be enough light for the detector to separate the signal from the noise, and
thus it cannot detect individual pulses. In addition to the attenuation, the pulse
broadens in time as it propagates through the fiber. This phenomenon, known as
pulse dispersion. Obviously, the lower the attenuation (and similarly, the lower
the dispersion), the greater will be the required repeater spacing and therefore
the lower will be the cost of the system.
2
Chapter 2 : Linear Effects

2.2 Attenuation

Attenuation in a fiber optic signal is the loss of optical power as the signal
travels through the fiber. Attenuation is caused by the fact that no manufacturing
process can produce a perfectly pure fiber. Either by accident or by design, the
fiber will always have some characteristic that attenuates the signal passing
through it.

The wavelength of the light passing through the fiber also affects attenuation. In
general, attenuation decreases as wavelength increases, but there are certain
wavelengths that are more easily absorbed in plastic and silica fibers than others.
One of the reasons for establishing standard operating wavelengths of 850 nm,
1300 nm, and 1550 nm in silica fiber and in the visible range of 650 nm for
plastic is because the wavelengths in between are considered high-loss regions.
Specifically, these wavelengths are in the ranges of 730, 950, 1250, and 1380
nm.

Attenuation must be measured differently from dispersion. As with dispersion,

the effect on a signal due to attenuation increases with the length of the fiber. If
the fiber’s makeup is consistent, the amount of attenuation can be predicted and
accounted for to some extent by adjusting the power of the source or by adding
repeaters, which collect a weak signal and amplify it.

3
Chapter 2 : Linear Effects

Note:

Attenuation provides a good example of the superiority of fiber over copper for
carrying signals. When an electrical signal is carried through copper wire,
attenuation increases with the data rate of the signal, requiring an increase in
transmission power or, more often, the use of repeaters. Attenuation per unit
length in an optical signal for a fiber of a given type is constant no matter what
the data rate, so repeaters can be farther apart, requiring fewer of them.

Attenuation behaves differently from dispersion, however, in the way that its
effects accumulate. As we have seen, dispersion is determined by factors within
the fiber and the signal’s wavelength and spectral width. None of these factors
changes as the signal passes through the fiber, so the amount of change caused
by dispersion can be calculated fairly simply.

Attenuation, however, is referenced to the signal’s reduction in power, and any

calculations must take into account the fact that as power is reduced, attenuation
will affect only the power that remains, thus altering the equation over the length
of the fiber.

For example, if attenuation reduces power by 1% over the distance of 1 km, then
only 99% of the original power will be left at the end of 1 km. At the end of
another kilometer, the remaining power is reduced by 1%, and so on. This
produces a more complex equation for determining attenuation, but it can still be
done.

Remember that attenuation is measured in decibels (dB). Decibels help us

account for the constant loss of power when we are measuring attenuation in a
fiber. In silica fiber, typical attenuation is about 2.1 dB/km at 1500 nm, and
attenuation in plastic fiber can be over 300 dB/km at 650 nm.

4
Chapter 2 : Linear Effects

While decibels are useful in measuring total attenuation, we can also divide
attenuation into two types: absorption and scattering.

Absorption

All materials, even the clearest glass, absorb some light. The amount of
absorption depends on the type of material and the wavelength of the light
passing through it. You can see absorption easily in sunglasses. Even on the
brightest days, only a fraction of the light energy passes through the tinted
lenses. The wavelengths that do not pass through are mostly absorbed by
impurities that have been placed in, or coated on, the lens material.

In an optical fiber, absorption occurs when impurities such as water or ions of

materials such as copper or chromium absorb certain wavelengths, as shown in
Figure 1. By keeping these impurities as low as possible, manufacturers can
produce fibers with a minimum of attenuation.

Figure 1: Absorption in optical fiber

5
Chapter 2 : Linear Effects

Rayleigh Scattering

Scattering is caused by atomic structures and particles in the fiber redirecting

light that hits them, as shown in Figure 2. The process is called Rayleigh
scattering, for Lord Rayleigh, a British physicist who first described the
phenomenon in the late nineteenth century.

Figure 2: Scattering in optical fiber

Rayleigh scattering is also the answer to the age-old question “Why is the sky
blue?” The blue that we see is actually the more prevalent blue wavelengths of
light from the sun being scattered by particles in the atmosphere. As the sun
moves toward the horizon and the light must pass through more of the
atmosphere, the scattering increases to the point where the blue light is almost
completely attenuated, leaving the red wavelengths, which are less affected by
the scattering for reasons that we’ll see shortly.

Rayleigh scattering depends on the relationship between wavelength and the size
of the structures in the fiber. Scattering increases as the wavelength of the light
approaches the size of the structures, which means that as the wavelength
decreases, it is more likely to be scattered. This is one of the main reasons that
infrared wavelengths are used in fiber optics. Their relatively long wavelengths

6
Chapter 2 : Linear Effects

are less subject to scattering than visible wavelengths. It also explains why the
sun turns red on the horizon. The shorter blue wavelengths are more likely to be
scattered by the similarly sized particles in the atmosphere than are the red
wavelengths.

Total Attenuation

Total attenuation is the combination of the effects of absorption and scattering in

a fiber. Figure 3 shows a typical attenuation curve for an optical fiber with the
effects of absorption and scattering combined. Note that the general curve is
caused by the effects of scattering, while the irregularities in the plot are caused
by specific impurities, such as hydroxyl molecules, absorbing light in those
wavelengths.

Note also the windows at the 850, 1300, and 1550 nm ranges. Remember that
while the 1300 nm range is better in terms of dispersion, it still has a higher
attenuation than the 1550 nm range, which is the reason for dispersion-shifted
fiber.

7
Chapter 2 : Linear Effects

Figure 3: An optical fiber’s attenuation curve

Bending Losses

In addition to characteristics within the fiber material, the actual condition of the
fiber can lead to losses as well. Because of bending, high-order mode light rays
can be lost in the cladding as the angle of the boundary layer changes in relation
to the light.

The types of bending we will look at are microbends and macrobends.

8
Chapter 2 : Linear Effects

Microbends

Microbends are small distortions of the boundary layer between the core and
cladding caused by crushing or pressure. Microbends change the angle of
incidence within the fiber, as shown in Figure 4. Changing the angle of
incidence forces high-order light rays to reflect at angles that prevent further
reflection, causing them to be lost in the cladding and absorbed.

Figure 4: Losses caused by microbending and macrobending

Macrobends

Macrobends occur when the fiber is bent around a radius that can be measured in
centimeters.

9
Chapter 2 : Linear Effects

As shown in Figure 4, these tight radii change the angle of incidence within the
fiber, causing some of the light rays to reflect outside of the fiber and, as with
microbending, be lost in the cladding and absorbed.

2.3 Dispersion

In general, dispersion is the spreading of light as it travels away from its source.
The light spreads because different components of it travel at slightly different
velocities, depending on the conditions in the medium through which it is
traveling and the wavelengths that make up the light. There are different kinds of
dispersion, however, and the kind that is taking place depends on several factors
in the fiber and in the light itself.

The greatest effect of dispersion is that as the light spreads, it can degrade or
destroy the distinct pulses of the digital signals in the light by making them
overlap each other, as shown in Figure 5, blurring and blending them to the point
that they are unusable. The effect grows more pronounced as the distance the
light travels increases.

10
Chapter 2 : Linear Effects

Figure 5: The effects of dispersion on a signal

The effect is similar to looking into a hallway through a frosted glass window. If
people are moving through the hallway close together, the glass spreads their
images so much that they merge with one another and look like a single mass
rather than individuals. If they spread out far enough from each other, however,
you can see each person moving past the window. The images are still spread
out, but the space between each person is great enough to see.

To prevent signal loss due to dispersion, it is necessary to keep the pulses far
enough apart to ensure that they do not overlap. This limits the signals to a bit
rate that is low enough to be only minimally affected.

Restricting the bit rate places a limit on the fiber’s bandwidth, or the amount of
information it can carry.

The types of dispersion that affect optical fiber are:

 Modal dispersion.

11
Chapter 2 : Linear Effects

 Material dispersion.
 Waveguide dispersion.
 Chromatic dispersion.
 Polarization-mode dispersion.

Let’s look at these types of dispersion more closely.

Modal Dispersion

We mentioned modal dispersion in the last chapter to explain why fibers are
classified as multimode and single-mode. It will help to review some of the
important points.

Modal dispersion results from light taking different paths, or modes, as it passes
through the fiber. The number of modes the light can take is determined by the
diameter of the fiber core, the refractive indices of the fiber core and cladding,
and the wavelength of the light.

A mode can be a straight line through the fiber, or it can follow an angular path,
resulting in reflections every time the light meets the interface between the core
and the cladding. The more reflections, the longer the path through the fiber, and
the longer the light takes to pass through it.

Depending on the mode, some parts of the light will pass through the fiber more
quickly than others. The difference in travel time can cause parts of the light
pulses to overlap each other, or in extreme cases to arrive in a different order
from the order they were transmitted. The signal is then no longer usable.

Methods for overcoming modal dispersion include:

12
Chapter 2 : Linear Effects

Lower bit rate Lowering the bit rate increases the gap between bits in the
signal. While dispersion will still affect them, they will not overlap one another,
and will still be usable. The drawback of this method is a reduction in
bandwidth, reducing the fiber’s ability to carry data.

Graded index fiber: Graded index fiber gradually reduces the refractive index
of the fiber core from the center toward the cladding, allowing the light that
follows a more angled path to speed up as it leaves the center and causing it to
slow down again as it reaches the center. This effect reduces the difference in
travel time between modes and allows wider bandwidths. Graded index fiber is a
moderately priced solution that allows wider bandwidths than multimode step
index fiber. In addition, the gradual change of indices as the light heads for the
cladding causes the light to curve back into the core of the fiber before it has a
chance to approach the cladding at a penetrating angle and be lost or reflected
with a destructive time delay.

Single mode fiber: Single-mode fiber has a core that is narrow enough for only
one mode to propagate, eliminating the problems caused by multiple modes.
This type of fiber requires more expensive connectors and equipment because of
the small core size and is typically used when very wide bandwidth requirements
justify the cost.

Remember that modal dispersion is measured in nanoseconds per kilometer

(ns/km). For example, if a fiber has a modal dispersion of 15 ns/km, the beam
that takes the longest path will fall behind the beam that takes the shortest path
by 15 ns for every kilometer of fiber length. This figure is used to determine the
maximum bit rate possible for a length of fiber.

13
Chapter 2 : Linear Effects

Material Dispersion

Material dispersion is the result of different wavelengths of light traveling at

different velocities in the fiber. Even the light from a laser is made up of several
wavelengths within a narrow range called the spectral width, which varies
depending on the source. A light-emitting diode (LED) has a spectral width of
20 nm to 170 nm, while a laser diode’s spectral width is much smaller, between
1 nm and 3 nm.

Recall the formula for determining the refractive index of a material,

n = c/v

Where n is the refractive index, c is the speed of light in a vacuum, and v is the
speed of the wavelength of light through the material. In this equation, n changes
with the wavelength of the light passing through the material. Remember that
this is the cause of white light breaking into its component colors in a prism.

When the different wavelengths travel at different velocities, the slower

wavelengths begin to lag behind as the light travels down the fiber core, causing
the light to spread as shown in Figure 6. If the light must travel a great distance,
the lag in the slower wavelengths can cause them to overlap the faster
wavelengths of the bits following them. As with modal dispersion, these
overlaps can degrade and ultimately destroy the signal.

Figure 6: Material dispersion in fiber causes some wavelengths to travel more slowly than others

14
Chapter 2 : Linear Effects

Because the wavelengths used in fiber optic transmissions have a narrow

spectral width, material dispersion takes place on a much smaller scale than
modal dispersion. Its effects in a fiber are measured in picoseconds per
nanometer of spectral width per kilometer (ps/nm/km), and are insignificant in a
multimode fiber when compared to the effects of modal dispersion.

Material dispersion only becomes a problem when modal dispersion is overcome

with single mode fiber and higher data rates are used over long distances.

Waveguide Dispersion

Waveguide dispersion occurs in single-mode fiber as the light passes through

not only the core, but also part of the cladding, as shown in Figure 7. Because,
by design, the core has a higher refractive index than the cladding, the light will
be traveling more slowly through the core than through the cladding.

Figure 7: Waveguide dispersion in optical fiber

While the difference in refractive indices of single-mode fiber core and cladding
are minuscule, they can still become a factor over great distances. In addition,
waveguide dispersion can combine with material dispersion to create another
problem for single-mode fiber: chromatic dispersion.

15
Chapter 2 : Linear Effects

Various tweaks in the design of single-mode fiber can be used to overcome

waveguide dispersion, and manufacturers are constantly refining their processes
to reduce its effects.

Chromatic Dispersion

Chromatic dispersion occurs in single-mode fiber, and results from the

combination of effects from material dispersion and waveguide dispersion.

When chromatic dispersion occurs, the effects of material dispersion, as shown

in Figure 8, compound the effects of waveguide dispersion. At lower data rates
and in multimode fiber, the effects of chromatic dispersion are so small as to be
unnoticed, especially when buried under modal dispersion. It is mostly a
problem in single-mode fiber carrying bit rates up to 10 Gbps over long
distances, where the detrimental effects build up.

Figure 8: Waveguide dispersion and material dispersion combine to create chromatic dispersion

One way to reduce chromatic dispersion is by taking advantage of the fact that
the relationship between wavelength, refractive index, and velocity is not linear.
In the infrared range of most fiber optic transmissions, the light’s velocity
through the medium drops as the wavelength increases until it reaches the range
between 1300 nm and 1550 nm. At wavelengths greater than 1550 nm, the

16
Chapter 2 : Linear Effects

longer wavelengths have a higher velocity. Somewhere in the 1300 nm to 1550

nm range there is a crossover where, depending on the specific composition of
the fiber, the refractive index is the same for the wavelengths within the narrow
spectral width of the light being transmitted. In other words, as shown in Figure
9, as the wavelength approaches this range, dispersion drops to zero. This zero-
dispersion point normally occurs at 1300 nm in a standard single-mode fiber.
Unfortunately, other characteristics of the fiber attenuate the signal at this
wavelength, making it unusable for long-distance runs.

Figure 9-a: Dispersion profile of a typical optical fiber

17
Chapter 2 : Linear Effects

Figure 9-b: Chromatic dispersion coefficients for various fiber types

There are two ways to reduce chromatic dispersion in fiber while maintaining
the energy of the signal: dispersion-shifted fiber and reduced spectral width.

Dispersion-Shifted Fiber

Dispersion-shifted fiber is a specially formulated single-mode fiber that shifts

the zero-dispersion point to 1550 nm, where the signal can travel a greater
distance through the fiber without significant attenuation. The refractive index
profile of dispersion-shifted fiber is shown in Figure 10. The peak in the center
of the profile reveals an inner core that has its highest refractive index at the
center. The refractive index gradually decreases toward a thin inner layer of
cladding. The smaller peaks represent a ring of silica with a higher refractive
index surrounding the inner cladding, slowing the light that would normally
increase its velocity in the cladding. This effect reduces waveguide dispersion,
18
Chapter 2 : Linear Effects

which in turn reduces chromatic dispersion. The use of dispersion-shifted fiber

allows much higher bit rates to travel over greater distances, and as a result, it
may also be used to carry several different channels at different wavelengths.
This can lead to four-wave mixing, which we will discuss shortly.

It seems that nobody can leave a good thing alone. Once engineers overcame the
problem of chromatic dispersion in single-mode fiber using dispersion shifting,
they decided to squeeze all the use they could out of it by piling on different
wavelengths to create multiple transmission channels. The idea behind this is
that different wavelengths can actually occupy the same space but remain
distinct from one another until they are sorted out at the other end of the fiber
link.

It makes very good sense, but then another problem cropped up. The
wavelengths used in the multiple channels must stay near the zero-dispersion
range of 1550 nm, so you end up with individual channels only 2 nm apart,
typically at 1546, 1548, 1550, and1552 nm, for example. It’s difficult for
anything to be only 2 nm apart and not interact, so interact they do. In fact, they
create new wavelengths that can interfere with the wavelengths that are part of
the transmission.

The problem gets exponentially worse with the number of wavelengths being
transmitted. The formula for predicting the number of new waves created is:

FWM = (n2 (n – 1))/2

Where FWM is the number of waves created and n is the number of wavelengths
being transmitted through the fiber. So if two wavelengths are being used, an
extra two wavelengths will appear. That’s not too bad in itself. But if you are
transmitting four original wavelengths:

19
Chapter 2 : Linear Effects

FWM = (16 × 3)/2 = 24

And eight wavelengths will produce:

FWM = (64 × 7)/2 = 224

The solution to four-wave mixing actually involves creating just enough

dispersion in the fiber to render the newly created wavelengths harmless to the
signals while leaving the original signal clear enough to use. The fiber created
for this purpose is nonzero-dispersion-shifted fiber (NZ-DSF).

Nonzero-dispersion shifting moves the zero-dispersion point slightly away from

the wavelength used for the transmission, usually about 10 nm above or below
the transmission wavelength, so there is sufficient dispersion to keep the effects
of four-wave mixing to minimum.

Figure 10: Refractive index profile of dispersion-shifted fiber

20
Chapter 2 : Linear Effects

Reduced Spectral Width

Because material dispersion is caused by an overabundance of wavelengths in

the optical signal, the simplest solution is to reduce the number of wavelengths
by reducing the spectral width. Recall that dispersion is expressed as
picoseconds per nanometer of spectral width per kilometer of fiber, so any
reduction in spectral width will have a significant effect on the amount of
material dispersion.

2.4 Polarization-Mode Dispersion

Polarization-mode dispersion (PMD) is masked by other forms of dispersion

until the bit rate exceeds 2.5 Gbps. In order to understand PMD, we must look at
an information pulse more closely.

Recall that light is an electromagnetic wave, consisting of an electrical and a

magnetic wave traveling at right angles to one another. The orientation of the
two waves along the path of propagation determines the light’s polarization
mode, or polarity.

As shown in Figure 11, it is possible to have different polarities of light traveling

through the fiber in a signal, occupying different parts of the fiber as they pass
through it. Because no fiber is perfect, there will be obstacles in one part of the
fiber that are not present in another. As a result, the light having one polarity
may pass an area without interference, while another polarity may pass through a
defective region, slowing it down.

Polarization-mode dispersion is not so much a function of the fiber’s overall

characteristics as it is a result of irregularities, damage, or environmental

21
Chapter 2 : Linear Effects

conditions such as temperature. Small areas of damage called microbends can

cause PMD, as can fiber that is not perfectly round or concentric.

Because PMD is caused by specific conditions within the fiber, and not the
fiber’s overall characteristics, it is difficult to assign a consistent PMD value to a
length of fiber. The exact amount of PMD changes with external conditions, the
physical condition of the fiber, and the polarization state of the light passing
through it at any given moment. For this reason, PMD is measured in terms of
the total difference in the travel time between the two polarization states,
referred to as the differential group delay (DGD) and measured in picoseconds.
The amount of PMD itself may be measured in ps/km 1/2.

Figure 11: Polarized light shown in a cross-section of optical fiber

22
Chapter 2 : Linear Effects

Figure 12: Polarization mode dispersion

Contributing factors are:

 Fiber core ellipticity.

 Transverse stress.
 Bending.
 Twisting.
 Isolators in EDFAs.
 Aging.

How Dispersion Affects Bandwidth

While different types of dispersion have different causes and are measured at
different rates (ns/km, ps/nm/km), all of them have one effect in common: they
place a limit on the bandwidth of optical fibers. As we saw in the previous
chapter, in fact, modal dispersion can cause the bandwidth of a fiber to narrow
dramatically as the distance increases.
23
Chapter 2 : Linear Effects

It seems that no matter what kind of dispersion a fiber is manufactured to

overcome, it is always going to be subject to another kind, at a higher
bandwidth. For this reason, one way of grading fibers is by the limit that
dispersion places on its ability to carry a signal. Depending on the type of fiber,
this limit can be expressed in one of two ways. In multimode fiber, where modal
dispersion is the overwhelming factor, it is expressed in megahertz of bandwidth
per kilo- meter of fiber (MHz/km), and in single-mode fiber it is expressed in
picoseconds per nanometer of source spectral width per kilometer of fiber
(ps/nm/km).

The MHz/km figure expresses how much bandwidth the fiber can carry per
kilometer of its length. The fiber’s designation must always be greater than or
equal to the product of the source bandwidth and the length of the cable.

When working with single-mode fiber, where chromatic dispersion is a limiting

factor, the more precise designation of ps/nm/km is necessary as a gauge of a
fiber’s ability to carry a signal. Instead of specifying the bandwidth that the fiber
can carry, the classification describes the dispersion that takes place within the
fiber. This is done because the bandwidth is no longer simply a factor of
distance. It can also be changed by narrowing the spectral width of the source.
As we’ve already seen, a small change in spectral width can significantly affect
the fiber’s bandwidth.

24
Chapter 2 : Linear Effects

2.5 Review Questions

1. What is dispersion?

A. The weakening of light

B. The spreading of light

C. The coloring of light

D. The diverting of light

2. What does dispersion affect most in a fiber optic signal?

A. Bandwidth

B. Signal strength

C. Wavelength

D. Color

3. What causes modal dispersion?

A. Light rays change wavelength as it passes through the fiber.

B. Light rays rotate as they pass through the fiber.

C. Light rays take different possible paths through the fiber.

D. Light rays become weaker as they travel through the fiber.

25
Chapter 2 : Linear Effects

4. Material dispersion is partly dependent on:

A. The power of the source

B. The bit rate of the source

C. The manufacturer of the source

D. The spectral width of the source

5. Where does waveguide dispersion take place?

A. Only in the core

B. Only in the cladding

C. In the core and the cladding

D. In fiber without cladding

6. Chromatic dispersion is:

A. Another name for modal dispersion

B. Another name for material dispersion

C. Another name for waveguide dispersion

D. A combination of waveguide dispersion and material dispersion

Review Questions 105

26
Chapter 2 : Linear Effects

7. How does dispersion-shifted fiber reduce chromatic dispersion?

A. It shifts the point of no dispersion to the wavelength of the light that travels
through the fiber best.

B. It adds enough modal dispersion to counteract the chromatic dispersion.

C. It shifts the wavelength of the light passing through it.

D. It shifts the light from the core to the cladding.

8. Another solution for chromatic dispersion is to:

A. Reduce the power of the light.

B. Reduce the wavelength of the light.

C. Reduce the spectral width of the light.

D. Reduce the length of the fiber.

9. Polarization-mode dispersion takes place when:

A. Light travels in opposite directions in the fiber.

B. Two polarities of light are affected differently by conditions in the fiber.

C. Light passes near a magnet, splitting it into north and south polarities.

D. Light passes through an electrical field, splitting it into positive and negative
components.

27
Chapter 2 : Linear Effects

10. For a fiber of a given bandwidth, as the length of the fiber increases, the
bandwidth:

A. Increases

B. Decreases

C. Does not change

D. Is independent of the length

11. What is attenuation?

A. The amount of light that the fiber can collect

B. The increase in power of a signal

C. The decrease in power of a signal

D. The change in wavelength of a signal

12. Windows are wavelength regions in fiber where attenuation:

A. Is lowest

B. Is highest

C. Drops to zero

D. Allows light to escape the fiber

28
Chapter 2 : Linear Effects

13. Attenuation is measured in:

A. MHz/km

B. ps/km/nm

C. dB/km

D. GHz

14. Microbends and macrobends can attenuate a signal by:

A. Changing the angle of incidence

B. Changing the cone of acceptance

C. Changing the numerical aperture

D. Changing the refractive index

Review Questions 107

15. Macrobends are caused by:

A. Twisting the fiber

B. Damage to the fiber

C. Curving the fiber

D. Pressurizing the fiber

2.6 Answers to Review Questions

29
Chapter 2 : Linear Effects

1. B. Dispersion is the spreading of light for various reasons as it moves away

from its source.

2. A. Dispersion affects bandwidth, since signal pulses must be kept apart to

keep from overlapping as they spread.

3. C. Modes are possible paths that light can travel through the fiber, and modal
dispersion is caused when some light takes paths that are subject to more
reflection off of the boundary between the core and the cladding, causing its path
to be longer.

4. D. The spectral width of the source is the range of wavelengths being sent
through the fiber.

Because material dispersion causes different wavelengths to travel at different

speeds, a large spectral width can increase the amount of dispersion taking place.

5. C. Waveguide dispersion takes place as the light passing through a single-

mode fiber passes through part of the cladding. Because the cladding has a lower
refractive index, the light speeds up, arriving ahead of the light in the core.

6. D. Chromatic dispersion takes place when the combination of waveguide

dispersion and material dispersion causes signals in a single-mode fiber to
overlap.

7. A. Dispersion-shifted fiber shifts the “zero-dispersion point,” or the point

where chromatic dispersion drops to zero, to the wavelength that allows light to
travel through the fiber with the least energy loss.

30
Chapter 2 : Linear Effects

8. C. Reducing the spectral width of the light cuts down the difference between
the slowest and the fastest wavelengths in the fiber, thus reducing the material
dispersion component of chromatic dispersion.

9. B. Light waves travel in different orientations, or polarities. When two

different polarities travel through a fiber, they can be affected differently by
conditions within the parts of the fiber in which they are traveling, causing one
polarity to travel more slowly than the other.

10. B. Because dispersion increases with the length of the fiber, the usable
bandwidth of the fiber decreases as the signal pulses must be kept farther apart
to avoid overlapping.

11. C. Attenuation is the loss of power in a signal as it travels through the fiber.

12. A. Windows are spectral regions determined by the composition of the fiber
where light suffers the lowest attenuation. Standard windows for fiber optic
signals are at 850 nm, 1300 nm, and 1550 nm.

13. C. Because attenuation takes a percentage of power and then takes the same
percentage of the remaining power, decibels (dB) are used to express the
constantly changing relationship between the signal level and the amount of loss.

14. A. Microbends and macrobends change the angle of incidence of the

core/cladding interface, causing light rays to reflect at angles that send them into
the cladding where they are absorbed.

15. C. Macrobends occur when the fiber is bent in a sharp radius.

31
Chapter 3 : Non-Linear High-Power Effects

Chapter 3: Non-Linear High-Power Effects

Aim of study
This chapter introduces nonlinear effects: Stimulated Raman Scattering (SRS), Stimulated
Brillouin Scattering (SBS), Four-Wave Mixing (FWM), Self-Phase Modulation (SPM) &
Cross-Phase Modulation (XPM).

Contents Pages

3.1 Introduction 2
3.2 Stimulated Raman Scattering (SRS) 3
3.3 Stimulated Brillouin Scattering (SBS) 5
3.4 Four-Wave Mixing (FWM) 8
3.5 Self-Phase Modulation (SPM) 10
3.6 Cross-Phase Modulation (XPM) 15

1
Chapter 3 : Non-Linear High-Power Effects

Chapter 3
Non-Linear High-Power Effects

3.1 Introduction

When light travels in a vacuum, individual waves from different sources do not
interact with one another. However, when light travels in a material, it can
interact with that material in various ways. This interaction can produce changes
in the light wave itself and cause interactions between different light waves with
the material acting as an intermediary.

The interaction of light with the material in optical fiber is typically very small
and thus interactions between different signals on the same fiber are also very
small.
However, since the signal travels long distances on fiber, very small effects have
the opportunity to build up into large ones. Non-linear effects are ones which
increase in significance exponentially as the level of optical power in the fiber is
increased. At low power levels there is little or no effect. As power is increased
the effects appear and can then become very significant. For example in a
particular context Stimulated Brillouin Scattering may have no measurable
effect on a signal of 3 mW but a significant effect if the power of the signal is
increased to 6 mW.
These effects can be grouped into two classes. “Elastic” effects where although
the optical wave interacts with and is affected by the presence of matter there is
no energy exchange between the two. The prime example of elastic scattering is
four-wave mixing. “Inelastic Scattering” is where there is an energy transfer
between the matter involved and the optical wave. Stimulated Brillouin

2
Chapter 3 : Non-Linear High-Power Effects

Scattering and Stimulated Raman Scattering are examples of this class. These
effects are discussed in the following sections.
As far as transmission on fiber is concerned the non-linear effects are nearly
always undesirable. After attenuation and dispersion they provide the next major
limitation on optical transmission. Indeed in some situations they are more
significant than either attenuation or dispersion. However, many optical devices
rely on just these same non-linear effects for their basic operation. A lot of
research goes into developing special fiber with increased levels of non-linearity
to build more effective devices.

3.2 Stimulated Raman Scattering (SRS)

Stimulated Raman Scattering is caused by a similar mechanism to the one which

produces SBS. However, the interactions involved are due to molecular
vibrations rather than acoustic ones. Scattered light can appear in both the
forward and backward directions. In a single-channel system the “Raman
Threshold” (the power level at which Raman Scattering begins to take effect) is
very high. Other effects (such as SBS) limit the signal power to much less than
the Raman Threshold in single-channel systems.

Figure 1-a: Stimulated Raman Scattering

3
Chapter 3 : Non-Linear High-Power Effects

Figure 1-b: Stimulated Raman Scattering

While Stimulated Raman Scattering is a not an issue in single-channel systems it

can be a significant problem in WDM systems. When multiple channels are
present, power is transferred from shorter wavelengths to longer ones. This can
be a useful effect in that it is possible to build an optical amplifier based on SRS.
But in the transmission system it is a source of noise.
Figure 1 shows the principle. Two wavelengths are shown before and after SRS.
Notice that power has been transferred from the shorter wavelength to the longer
one (from the higher energy wave to the lower energy one). This has resulted in
additive noise at the longer wavelength and subtractive noise at the shorter one.
This power transfer is caused by interactions of the light with vibrating
molecules.
Optical power so transferred is called the “Stokes Wave”.

4
Chapter 3 : Non-Linear High-Power Effects

Important characteristics of SRS are:

- The effect of SRS becomes greater as the signals are moved further and further
apart (within some limits). This is a problem as we would like to separate the
signals as much as we can to avoid four-wave mixing effects and when we do
we get SRS!
SRS can take affect over about 40 THz (a very wide range) below the higher
frequency (shorter wavelength) involved. That is, it can extend over a range of
wavelengths of about 300 nm longer than the shortest wavelength involved.
The effect is maximized when the two frequencies are 13.2 THz apart.
- SRS increases exponentially with increased power. At very high power it is
possible for all of the signal power to be transferred to the Stokes Wave.
One study concluded that in a 10-channel WDM system with 1 nm channel
spacing power levels need to be kept below 3 mw (per channel) if SRS is to be
avoided.

3.3 Stimulated Brillouin Scattering (SBS)

Stimulated Brillouin Scattering is a scattering of light backwards towards the

transmitter caused by mechanical (acoustic) vibrations in the transmission
medium (fiber). The reflected wave produced is called the “Stokes Wave”. The
effect is usually trivial but can be very important in situations where a high
quality, narrow line width laser is used at a relatively high power level.
Stimulated Brillouin Scattering is caused by the presence of the optical signal
itself.
Even though a signal level of a few milliwatts seems very small, in the tiny
cross-section of a single-mode fiber core the field can be very intense. An
optical signal is in reality a very strong electromagnetic field. This field causes

5
Chapter 3 : Non-Linear High-Power Effects

mechanical vibrations in the fiber which produce a regularly varying pattern of

very slight differences in the refractive index. The Brillouin Scattering effect is
caused by light being reflected by the diffraction grating created by the regular
pattern of RI changes. The reflected light is reflected backwards from a moving
grating! Hence its frequency is shifted by the Doppler Effect. The shift of the
reflected wave in standard single-mode fiber is downward in frequency by
around 11.1 GHz.

Figure 2: Movable grating

The effect, like Stimulated Raman Scattering (described in the following

section), is nonlinear and in practical systems requires a power level of
something above 3 mW for any serious effect to be observable (indeed usually
it's a lot higher than this). It also requires a long interaction length and a very
narrow line width (long coherence length) signal. In general the signal line
width must be less than about 100 MHz (around .1 nm) for SBS to become an
issue. The effect in the forward direction is experienced as an increase in
attenuation. This is rapid and adds noise to the signal. For narrow line width
signals SBS imposes an upper limit on the usable transmit power.

6
Chapter 3 : Non-Linear High-Power Effects

Figure 3: SBS Threshold Variation with Wavelength

The threshold value is the power level above which SBS causes a significant
effect.

In most current systems SBS has not been much of a problem for the following
reasons:
1. Direct modulation of the transmit laser's injection current produces a chirp
and broadens the signal. This significantly reduces the impact of SBS.
2. The effect is less in 1300 nm systems than in 1550 nm systems due to the
higher attenuation of the fiber.
3. Lasers capable of producing the necessary power level have only recently
become available and amplifiers are also a recent innovation.
4. At speeds of below 2.4 GHz it has not been necessary to use either very high
power or very narrow line width lasers.
5. SBS effects decrease with increase in speed because of the signal broadening
affect of the modulation.

7
Chapter 3 : Non-Linear High-Power Effects

In cases where SBS could be a problem the line width is often intentionally
broadened. This can be done by using an additional RF modulation on the laser
injection current, by using an external phase modulator or by using a “self
pulsating” laser. Of course increasing the line width mitigates against long
distance transmission because it increases the effect of chromatic dispersion.
However, SBS can be a major problem in three situations:
1. In long distance systems where the span between amplifiers is great and the
bit rate low (below about 2.5 Gbps).
2. In WDM systems (up to about 10 Gbps) where the spectral width of the signal
is very narrow.
3. In remote pumping of an erbium doped fiber amplifier (EDFA) through a
separate fiber. EDFA pumps typically put out about four lines of around only
80 MHz wide. Each of these lines is limited by SBS in the amount of power that
can be used. This can significantly limit the potential of remote pumping.

3.4 Four-Wave Mixing (FWM)

Figure 4: Four Wave Mixing Effects

One of the biggest problems in WDM systems is called “Four-Wave Mixing”

(FWM).
Illustrated in Figure 4, FWM occurs when two or more waves propagate in the
same direction in the same (single-mode) fiber. The signals mix to produce new

8
Chapter 3 : Non-Linear High-Power Effects

signals at wavelengths which are spaced at the same intervals as the mixing
signals. This is easier to understand if we use frequency instead of wavelength
for the description. A signal at frequency f1 mixes with a signal at frequency f2.
To produce two new signals one at frequency 2f1-f2 and the other at 2f2-f1.The
effect can also happen between three or more signals.

There are a number of significant points.

- The effect becomes greater as the channel spacing is reduced. The closer the
channels are together the greater the FWM effect.
- FWM is non-linear with signal power. As signal power increases the effect
increases exponentially.
- The effect is strongly influenced by chromatic dispersion. FWM is caused
when signals stay in phase with one another over a significant distance. The
lasers produce light with a large “coherence length” and so a number of signals
will stay in phase over a long distance if there is no chromatic dispersion. Here
chromatic dispersion is our friend. The greater the dispersion, the smaller the
effect of FWM - because chromatic dispersion ensures that different signals do
not stay in phase with one another for very long.
- If the WDM channels are evenly spaced then the new spurious signals will
appear in signal channels and cause noise. One method of reducing the effect
of FWM is to space the channels unevenly. This mitigates the problem of
added noise (crosstalk) in unrelated channels. However, it doesn't solve the
problem of the power that is removed from the signal channels in the process.

9
Chapter 3 : Non-Linear High-Power Effects

3.5 Self-Phase Modulation

When we refer to the refractive index of a medium, we never specify the

intensity levels since under normal circumstances the refractive index of a
medium is independent of the light intensity. On the other hand, at high optical
intensities, the refractive index of a medium is no longer a constant but depends
on the intensity of the light beam. If the refractive index of a medium for low
intensities is n0, at high intensities it can be written as

n = n0 + n2 I (3.1)

Where n2 is a constant that depends on the medium and I represents the intensity
of the beam. For example, for silica, n2 ≈ 3.2 x 10−20 m2 /W, and thus the
refractive index of silica increases with increased light intensity: for nominal
intensities the increase is, of course, extremely small. Thus, if we consider a
single-mode optical fiber with a mode area of 80 pm2 and couple light of power
100 mW, the intensity within the fiber would be about 1.25 x 10 9 W/m2 and the
corresponding increase in refractive index due to the presence of the light beam
is a tiny value of about 4 x 10−11 (i.e., 0.00000000004), so the refractive index
would increase from 1.44 to 1.44000000004. Although this refractive index
change is very small, when the light beam propagates in an optical fiber over
long distances (a few hundred to a few thousand kilometers), the accumulated
effects due to this increase can be significant. Since the phase of the light beam
depends on the refractive index of the medium, and it is the beam itself that is
changing the refractive index, which in turn changes its own phase, this effect is
referred to as self-phase modulation.
In a single-mode optical fiber, since the light beam propagates as a mode, we
can represent the intensity as a ratio of the power carried by the beam to the
mode area (Aeff).

10
Chapter 3 : Non-Linear High-Power Effects

Thus, in the case of an optical fiber, Eq. (3.1) is written as

Table 1 gives values of Aeff and v for some typical fibers.

2
Nonlinear Coefficient v at
Fiber Type Mode Area (pm ) −1 −1
1550 nm (W m )
−3
G.652 85 1.5 x 10
−3
G.653 46 2.8 x 10
G.655 52 (D > 0)
−3
56 (D < 0) 2.5 x 10
−3
NZ-DSF 73 1.8 x 10
−3
DCF 23 5.6 x 10
−3
PCF 3 43 x 10

Table 1: Mode Area and Nonlinear Coefficient for Some Common Fiber Types

Since nonlinearity-induced phase change depends on optical power, and since

the Optical power in the pulse decreases due to attenuation in the fiber, the effect
of nonlinearity would decrease continuously as light propagates through the
fiber. When light enters the fiber, it has maximum power and this result in a high
nonlinear phase shift. But as light propagates, its power decreases and the
corresponding nonlinear effect would decrease.

To find out about the effect of self-phase modulation on the propagation of an

optical pulse, let us consider the propagation of an optical pulse through an
optical fiber. Figure shows the electric field variation of an optical pulse: you
can imagine this to be like a snap shot of the pulse at any given time.

11
Chapter 3 : Non-Linear High-Power Effects

The oscillatory variation is due to the high frequency of the light signal, while
the envelope of the pulse (which defines the pulse shape) has a much slower
variation with time. Now, in the leading and trailing edges of the pulse, the
intensity is smaller than at the center of the pulse.

Figure 5: Optical pulse: the oscillatory portion is due to the high frequency of the pulse
and the envelope is the pulse shape.

(a)

(b)

Figure 6: (a) All cyclists travel at the same speed, and their separation remains the same
as they propagate. (b) If the cyclists close to the center start to travel slightly slowly, this
would result in crowding at the back and greater separation at the front.

12
Chapter 3 : Non-Linear High-Power Effects

Hence, due to the dependence of refractive index on intensity, the center of the
pulse would lead to a greater increase in the refractive index of the core of the
fiber than will the leading and trailing edges. Since the response of the medium
to changes in intensity are almost instantaneous (a short response time compared
to the time period of the optical wave, which is in femtoseconds) as the
intensity of the pulse changes within the pulse, the refractive index change
follows the change in intensity almost instantaneously. As the speed of
propagation of light depends on the refractive index, this would result in a slight
slowing down of the center of the pulse vis-a`-vis the leading and trailing edges.

To understand the implication of this, we consider an example in which seven

cyclists are traveling along a road at the same speed and spaced equally.
In this case the group of cyclists will retain their positions as they travel. Now, if
the cyclists closer to the center of the group start to travel a bit slower than the
rest, there would be a crowding of the cyclists toward the back end and greater
separation toward the front end.

In the case of light pulse propagation through an optical fiber, we can imagine a
similar situation when the higher-intensity portion of the pulse (around the
center of the pulse) travels more slowly than the ends. There would then be a
crowding of the waves toward the back end and greater separation toward the
front end. Since the period of oscillation decides the frequency of a wave, this
implies that the front end of the pulse would have a lower frequency and the
back end of the pulse would have a higher frequency (Fig.7). In such a pulse,
called a chirped pulse, the chirping is caused by nonlinear effects. Chirping
without a change in pulse shape leads to an increase in the frequency content of
the pulse, that is, to a broadening of the spectrum of the pulse.

13
Chapter 3 : Non-Linear High-Power Effects

Figure 7: When an optical pulse travels through an optical fiber in the presence of a
nonlinear effect, the frequency of the pulse varies with position within the pulse, leading
to a chirped pulse.

Figure 7 Shows simulation results of pulse propagation through an optical fiber

in the presence of nonlinearity only (neglecting dispersion), showing clearly that
the pulse shape remains the same while the frequency spectrum broadens.

The dispersion of a pulse depends on the spectral band- width of the pulse (i.e.,
on the frequency content of the pulse). Hence, in the presence of such a
nonlinear effect, the dispersive behavior of the pulse would get changed. It so
happens that when we operate a standard single-mode fiber (with zero dispersion
close to 1310 nm) at 1550 nm (i.e., a fiber operating with positive dispersion),
the presence of nonlinear effects indeed results in a reduction of the effective
dispersion. This implies that the dispersion caused in the pulse would
decrease as the input power is increased. Hence, the quantum of dispersion
compensation required in the presence of nonlinearity is in fact less than what is
predicted using linear effects only. This fact has to be taken into account when
designing a fiber optic system. For operation below the zero-dispersion
wavelength, wherein the dispersion is negative, the nonlinear effect indeed leads
to increased dispersion.

14
Chapter 3 : Non-Linear High-Power Effects

3.6 Cross-Phase Modulation

In WDM systems, within the fiber there are pulses propagating simultaneously
at different wavelengths. In the presence of nonlinear effects, each wavelength
would result in a change in refractive index of the fiber, depending on the power
carried by that wavelength. If we now consider light beams at two different
frequencies propagating simultaneously through the fiber, the change in
refractive index brought about by each of the beams will affect the propagation
of the other beam. This effect, termed cross-phase modulation (XPM), results in
crosstalk: the output from one channel now depending on the presence or
absence of the other channel. Since the signal pulses are random, sometimes
there would be overlap between the two signal pulses and sometimes there
would be no overlap. When they overlap there would be effects of cross-phase
modulation and when there is no overlap, there would be no cross-phase
modulation. This random nature results in a random noise of the channels,
resulting in a penalty in terms of increased bit error rates.

Figure 8: Cross-phase modulation takes place when pulses at different wavelengths

overlap

15
Chapter 4 : Optical Devices

Chapter 4: Optical Devices

Aim of study
This chapter introduces detailed structure of a DWDM line.

Contents Pages

4.1 Detailed Structure of a DWDM Line 2

4.2 Transponders 3
4.3 Filters and Gratings 6
4.4 Optical Multiplexer and Demultiplexer 12
4.5 Optical add / drop multiplexers 21
4.6 Optical Amplifiers 24
4.7 Optical Fibers 44
4.8 Dispersion Compensation Modules 46
4.9 PMD Compensation 48
4.10 Lasers and Modulators 49
4.11 Photodetectors 61
4.12 Isolators, Circulators and Connectors 65
4.13 Optical Switching Units 68

1
Chapter 4 : Optical Devices

Chapter 4
Optical Devices

4.1 Detailed Structure of a DWDM Line

When talking about the components of a DWDM line the question must be,
which components are actually special? The answer is:

 Lasers and Modulators.

 Optical Multiplexers / Demultiplexers.
 Optical Amplifiers.
 Dispersion Compensation Modules.
 PMD Compensators.
 Fiber.
 Photodiodes.
 Connectors and Isolators.

Most of these devices are of passive analog nature. This means that, in principle,
a DWDM line can be transparent to the carried signals. However, efforts are
being made to turn these passive devices into active components, controlled
either by electrical or optical interfaces, in order to achieve maximum
performance.

2
Chapter 4 : Optical Devices

Figure 1: Components in DWDM systems

4.2 Transponder
Convert from color to (black & white) optical signal Used transponder terminal
that convert from black & white to color signal from black & white equipments.

There are some equipments deal with color signal directly without transponder.

Figure 2

3
Chapter 4 : Optical Devices

Within the DWDM system a transponder converts the client optical signal from
back to an electrical signal and performs the 3R functions (see Figure). This
electrical signal is then used to drive the WDM laser. Each transponder within
the system converts its client's signal to a slightly different wavelength. The
wavelengths from all of the transponders in the system are then optically
multiplexed.
In the receive direction of the DWDM system, the reverse process takes place.
Individual wavelengths are filtered from the multiplexed fiber and fed to
individual transponders, which convert the signal to electrical and drive a
standard interface to the client.

Figure 3: Transponder Functions

Future designs include passive interfaces, which accept the ITU-compliant light
directly from an attached switch or router with an optical interface.
Operation of a Transponder Based DWDM System the next Figure shows the
end-to-end operation of a unidirectional DWDM system.

4
Chapter 4 : Optical Devices

Figure 4: Anatomy of a DWDM System

The following steps describe the system shown in Figure

1. The transponder accepts input in the form of standard single-mode or
multimode laser. The input can come from different physical media and
different protocols and traffic types.
2. The wavelength of each input signal is mapped to a DWDM wavelength.
3. DWDM wavelengths from the transponder are multiplexed into a single
optical signal and launched into the fiber. The system might also include the
ability to accept direct optical signals to the multiplexer; such signals could
come, for example, from a satellite node.
4. A post-amplifier boosts the strength of the optical signal as it leaves the
system (optional).
5. Optical amplifiers are used along the fiber span as needed (optional).
6. A pre-amplifier boosts the signal before it enters the end system (optional).
7. The incoming signal is demultiplexed into individual DWDM lambdas (or
wavelengths).

5
Chapter 4 : Optical Devices

8. The individual DWDM lambdas are mapped to the required output type (for
example, OC-48 single-mode fiber) and sent out through the transponder.

4.3 Filters and Gratings

4.3.1 General Properties

For DWDM it is essential to have the ability to filter out one particular
wavelength. Several filtering methods exist, most of these techniques are in one
way or another using interference. A good optical filter for DWDM is
characterized by the capacity of isolation (eliminate power from other channels)
and distortion (to minimize signal distortion due to filter response). If a filter
does not provide good isolation then there will be signal degradation due to
linear crosstalk. That means that power from other channels will reach the
receiver, interfering with the selected channel. But, in order to achieve good
isolation it is necessary to reduce the filter bandwidth, thus increasing filter
distortion. The system designer has to find a compromise between these factors.

Figure 5: Original DWDM signal spectrum

6
Chapter 4 : Optical Devices

Figure 6: Filtered signal spectrum

4.3.2 Interference

Light can be imagined as a wave. The interesting question now is what happens
if two of those waves meet?

The answer is interference. If the two waves have the same phase (that means
„mountain to mountain and valley to valley“, the two waves add up and create a
joint wave of higher amplitude. If the two waves have opposite phase though,
they cancel each other, the result is „nothing“.

7
Chapter 4 : Optical Devices

Figure 7: Interference

4.3.3 Fabry-Perot Interferometer

The easiest form of interferometer is the Fabry-Perot type. It consists of two

parallel plates that reflect light back and forth. By constructive and destructive
interference only a few wavelengths are able to pass, the others are reflected.
The criterion for constructive interference is that the differences in path length
of the multiple-reflected beams is equal to an integer multiple of the
wavelength. Thus, by varying the distance between the plates, certain
wavelengths can be selected.

For the other wavelengths the criterion is not fulfilled, therefore they are
reflected.

8
Chapter 4 : Optical Devices

Figure 8: Fabry-Perot interferometer

4.3.4 Dielectric Thin Film Filters

DTF Filters consist of alternate layers of high refractive index and low
refractive index, each layer being λ/4 thick.

Light reflected within layers of high refractive index does not shift its phase,
while light reflected in layers of low refractive index is shifted by 180°. The
condition for constructive interference once more causes one wavelength to pass
and the others to be reflected.
That means its function is similar to a Fabry Perot Filter but it is much more
"accurate", with narrow line width etc.

9
Chapter 4 : Optical Devices

Some features are:

 Low pass band loss: less than 0.3dB.
 Good channel spacing: better than 0.8nm.
 Low interchannel crosstalk: better than –28dB.

Figure 9: Dielectric thin film filter

4.3.5 Bragg Grating

A Bragg grating (or Bragg Reflector) consists of a number of parallel semi-

reflecting plates. Once more by using constructive and destructive interference
just one specific wavelength is completely reflected, if it satisfies the condition

d = n*λB/2
Where n is 1, 3, 5,...
Bragg reflectors have a very high reflectivity and are therefore employed as
mirrors for high power lasers.
10
Chapter 4 : Optical Devices

A variation of Bragg gratings is the so called fiber bragg grating: By varying the
index of refraction of a fiber core it is possible to achieve a kind of Bragg
grating, such that one wavelength is reflected, while the others pass through.

Figure 10: Bragg grating

Figure 11: Fiber Bragg grating

11
Chapter 4 : Optical Devices

4.3.6 Mach-Zender Filter

This method once more relies on interference.
A mix of two wavelengths arrives at the first coupler which distributes the
power equally on both lines. One of the lines is longer, thus introducing a
different optical path length and a phase shift.
Selecting that phase difference cleverly can mean that the first wavelength has
its interference maximum at the place of fiber one and the second wavelength at
fiber two, thus separating the two signals.
By introducing a heating device to regulate the difference in length, and thus the
phase shift, it is possible to tune a Mach-Zender filter.

Figure 12: Mach Zender filter

4.4 Optical Multiplexer and Demultiplexer

4.4.1 Introduction

An optical demultiplexer can be built as an association of optical filters or as a

single-stand device. The purpose is to extract the original channels from a
DWDM signal. The requested properties of this device are the same as for the
optical filter: isolation and signal distortion. However channel number and
spacing must be considered now because demultiplexers can impose limitations

12
Chapter 4 : Optical Devices

on the number of channels or the total available bandwidth. Most

demultiplexers are symmetrical devices and can also be used as multiplexers.

4.4.2 Prism

The easiest and best-known optical demultiplexer is the prism.

Using the effect of dispersion (different speed of light for different
wavelengths), light is split into its spectral components.

4.4.3 Diffraction Grating

The function of a diffraction grating is very similar to that of a prism, only here
interference is the important factor. A mixture of light is also split into its
contributing wavelengths.

With such a grating, sometimes also called a bulk grating, channel spacings of
down to 50GHz can be achieved.

Figure 13: Effect of a prism

13
Chapter 4 : Optical Devices

Figure 14: Effect of a grating

4.4.4 Dielectric Thin Film Filters

DTF filters are well suited to multiplex or demultiplex a small number of

channels. They are simply connected in sequence, each filter dropping one
specific wavelength.

Although this method is comparatively easy it has one drawback; each

reflection causes attenuation of approximately 0.1dB. That means the channel
demultiplexed last is attenuated to a much higher degree then the first one,
especially if we are talking about a higher number of channels. This property
restricts the use of DTFs quite severely, limiting the number of channels to
about 16.

The minimum spacing reachable with these devices is about 100GHz.

14
Chapter 4 : Optical Devices

Figure 15: Demultiplexing using narrowband DTF filters

4.4.5 Arrayed Waveguide Gratings

Waveguide Grating Routers (WGRs) are also called PHASARS or Arrayed

Waveguide Gratings (AWGs). These are one of the most important new devices
available for WDM systems.76 Built using planar waveguide technology the
WGR has similar functions to those of a Littrow grating. The basic function is
illustrated in Figure 16.
 It can take a multi-channel (multi-wavelength) input appearing on a single
input waveguide (port) and separate the channels onto different output ports.
 It can combine many inputs (of different wavelengths) from different input
ports (waveguides) onto the same output port.
 It can operate bi-directionally.
 It can be connected as an optical add-drop multiplexor.

15
Chapter 4 : Optical Devices

Figure 16: Functionality of an arrayed waveguide grating

It is bi-laterally symmetric and either side could be input or output (or both
could take place at the same time).

The input and output stages consist of star couplers called “Free Space
Couplers” (FSCs). The inside of an FSC is just a “free space”. The “grating
region” is just a set of parallel waveguides of different lengths. These
waveguides are far enough apart so that the evanescent field in one guide does
not extend into any other guide. Therefore there is no coupling of power
between the guides in the grating region.
On the input side, a single-mode input on one of the input waveguides will
couple to a very large number of modes in the free space region. These modes
then couple to the waveguides in the grating region. Because there are so many
modes involved the amount of power coupled from any particular input to each
of the waveguides in the grating region is equal. However the distance from any
particular input port to each of the grating waveguides is different. This means
that at the entrance to the grating region there are phase differences between
16
Chapter 4 : Optical Devices

modes originating at the same input port. Light from different input ports will
have different sets of phase relationships.

The central region of the coupler functions like a grating. However it is

important to note that it is not a grating (it just functions like one). The
waveguides in the grating region are sufficiently separated from each other that
power cannot couple from one to another. The length of each of the guides
differs from its neighbour by a fixed delta. This results in phase differences
between the signals when they reach the destination star coupler. Interference
effects cause the signal at a particular wavelength to be reinforced in one
particular output guide and to be extinguished in the other guides. These effects
depend on both the wavelength and the location of the input port.

The result is that there is destructive interference in most of the output

waveguides.
In just one (at each different wavelength) there will be constructive interference
(reinforcement). The particular output guide in which a given wavelength will
be output depends on both the location of the input waveguide AND the
particular wavelength. N wavelengths on a particular input are spatially
separated onto N separate output waveguides.

17
Chapter 4 : Optical Devices

Figure 17.a: Distribution of Optical Power in the Input FSC

There is no interference here from other input waveguides as no other inputs are
coherent with this one.

If we consider a single wavelength signal arriving on just one input waveguide.

 When the signal arrives at the first Free Space Coupler (FSC) it diffracts in
many directions and is distributed more or less evenly among the
waveguides of the central section. There is no interference here because no
other input signal is coherent with this one.
The signal would be reinforced in some of the output waveguides and
extinguished in others (depending on their locations).

18
Chapter 4 : Optical Devices

Figure 17.b: Operation of the Output FSC

 In the real device we have multiple waveguides in the free space region and
their lengths are different from one another. In addition the FSCs are shaped
in such a way as to influence the interference effects produced. (The
distances between different input and output waveguides of the FSC are
intentionally different.) As illustrated in the figure a particular wavelength
from a particular input waveguide is reinforced in one (and only one) output
waveguide and destructively interferes in all other output waveguides.
 In this example (with only one input port in use) multiple wavelengths
arriving on the single input port will be directed to different output ports.
 If we now disconnect the input from the port we were using and re-connect
it to a different input port we get the same effect as described above (the
signal is split out by wavelength). However, the output ports now used for
particular wavelengths will be different from the first example.
 A given wavelength input on one particular port will be directed to a specific
output port. The same wavelength input on a different port will be output on

19
Chapter 4 : Optical Devices

a different port! The output port selected depends both on the wavelength of
the input light and the input port it came from.
It is easy to see that if you structure such a device carefully you can arrange
different wavelengths to be directed to different output ports.

4.4.6 Mach Zender Interferometer

Using a cascaded set of Mach Zender filters it is possible to demultiplex (or

multiplex) a number of channels with different wavelength.

Those Mach Zender interferometers can be integrated on Silica substrates, using

conventional technology.

4.4.7 Fiber Bragg Gratings

To demultiplex a multi wavelength signal wavelength by wavelength it is

possible to use a combination of an optical circulator and a Fiber-Bragg grating
(FBG).

This method is particularly interesting for optical add-drop multiplexers as

single wavelengths can be easily dropped. Even more interesting is the
possibility to tune that device by changing the length constant of the Fiber-
Bragg grating using piezo technology.

One advantage is the comparatively low insertion loss of only 0.2dB per FBG.
A second one is the reachable channel spacing of only 25GHz.

20
Chapter 4 : Optical Devices

Figure 18: Functionality of a mach Zender interferometer

Figure 19: Fiber Bragg grating as a demultiplexer

4.5 Add Drop Multiplexers

Figure 20: Add-Drop Multiplexor Function

An add-drop multiplexor adds and/or removes a single channel from a
combined WDM signal without interfering with the other channels on the fiber.
21
Chapter 4 : Optical Devices

This function is illustrated in Figure 20. There are several devices which may
perform this function such as:
1. Array waveguide gratings
2. Circulators with FBGs

4.5.1 Array Waveguide Gratings

Figure 21: Array Waveguide Add-Drop Multiplexor

In this figure we see a WGR configured as an add-drop multiplexor.

Wavelength 1 is added to the multiplexed stream on the left of the picture and
dropped (demultiplexed) from the input stream on the right of the figure.
This is a very versatile function. For example multiple channels can be added or
dropped in the same operation. However, the signal loss is about 5 dB per pass
through the device. So channels that aren't added or dropped experience an
insertion loss of around 10 dB. In a system we need to equalize the signal power
between channels (so that the newly added channels are about the same strength
as the dropped ones) and probably to amplify the whole stream as well.

4.5.2 Circulators with In-Fiber Bragg Gratings

22
Chapter 4 : Optical Devices

Figure 22: Add-Drop Multiplexor Using FBG and Circulators

Here we are using an FBG with a pair of circulators to add and drop a single
channel. Operation is as follows:
 The signal enters at the left of the figure and is routed through the circulator
to the FBG.
 The non-selected wavelengths pass through the FBG to the next circulator.
 The selected wavelength is reflected by the FBG and then directed out of the
next circulator port.
 The wavelength to be added (which must be the same as the one just
dropped) enters through the “add-port” of the rightmost circulator.
 It travels around to the FBG and is reflected back to the circulator. This
process mixes the added channel with the multiplexed stream.
This configuration has a relatively low loss of 3 dB for the multiplexed stream.
It could be very suitable for operation in a looped metropolitan area network
(MAN) where a single fiber loop interconnects many locations within a city
area.

23
Chapter 4 : Optical Devices

4.6 Optical Amplifiers

Introduction

Fiber loss and dispersion limit the transmission distance of any fiber-optic
communication system. For long-haul WDM systems this limitation is
overcome by periodic regeneration of the optical signal at repeaters, where the
optical signal is converted into electric domain by using a receiver and then
regenerated by using a transmitter. Such regenerators become quite complex
and expensive for multichannel light wave systems. Although regeneration of
the optical signal is necessary for dispersion-limited systems, loss-limited
systems benefit considerably if electronic repeaters were replaced by much
simpler, and potentially less expensive, optical amplifiers which amplify the
optical signal directly. Several kinds of optical amplifiers were studied and
developed during the 1980s. The technology has matured enough that the use of
optical amplifiers in fiber-optic communication systems has now become
widespread.

Optical Amplifier Applications

 In-line amplifiers.
 Booster amplifiers.
 Pre-amplifiers.
In-line amplifiers are used to directly replace optical regenerators. Booster
amplifiers are used immediately after the transmitter or multiplexer to increase
the output power. Pre-amplifiers are used before the receiver or demultiplexer to
increase the received power and extend distance.
The use of each configuration as advantages and disadvantages that must be
considered by the systems designer.
24
Chapter 4 : Optical Devices

The problems come when considering non-linear effects in the transmission

fiber and also noise generated by the amplifiers.
Some of the requirements for optical amplifiers for DWDM purposes are:
 High gain.
 Low noise.
 Flat amplification profile.

Figure 23: Passage from optical/electrical regenerators to optical amplifiers

Figure 24: Applications for optical amplifiers

4.6.1 Erbium Doped Fiber Amplifier

25
Chapter 4 : Optical Devices

4.6.1.1 Introduction

In traditional long-distance optical fiber communication systems, compensation

of loss and dispersion is usually accomplished using electronic regenerators. In
an electronic regenerator, attenuated and temporally broadened optical pulses
are first detected by a photodetector (shown as “PD” in the figure), which
converts the optical signals to electrical signals. These are then processed,
retimed, cleaned of
noise, and amplified electronically, and the amplified electrical pulses drive a
laser diode (shown as “LD”) to produce optical pulses almost identical to the
original.
These regenerated pulses are then launched into the fiber link for transmission
to the next regenerator. Such regenerators are designed to operate at a specific
bit rate, and when there is a need to upgrade the system to higher bit rates, the
regenerator needs to be replaced. Since each regenerator operates at one optical
wavelength, when there is a requirement to increase the capacity by using
another signal at another wavelength through the same fiber, then at the
regenerator sites, the signals at different wavelengths first need to be
demultiplexed (separated) into separate paths, and each wavelength then needs
to be processed individually by separate regenerators.
After regeneration of the individual signals at different wavelengths, the signals
need to be multiplexed (combined) into a single output for further transmission
through a fiber. Such regenerators not only compensate for the loss and
dispersion suffered by the signal but also clean the signal of any accumulated
noise, and thus the reformed pulses after the regenerators are almost as good as
at the beginning of their journey. At the same time, for wavelength-division-
multiplexed transmission systems carrying multiple wavelength signals through
one fiber, a solution using electronic regeneration would be very expensive.
26
Chapter 4 : Optical Devices

Whenever the system limitation is due to insufficient optical power rather than
dispersion, what is needed is simply amplification of the signal, and optical
amplifiers can perform this job very well. Optical amplifiers are devices that
amplify the incoming optical signals in the optical domain itself without
conversion to the electrical domain, and have truly revolutionized long-distance
fiber optic communications.
Optical amplifiers have two advantages over electronic regenerators: They do
not need high-speed electronic circuitry, and, they are transparent to bit rate and
format and most important, can amplify multiple optical signals at different
wavelengths simultaneously. Their development has ushered in a tremendous
growth in communication capacity using wavelength-division multiplexing
(WDM), in which multiple wavelengths carrying independent signals are
propagated through the same single mode fiber, thus multiplying the capacity of
the link. Of course, compared to electronic regenerators, they also have
drawbacks: They do not compensate for dispersion accumulated in the link, and
they add noise to the optical signal. As we will see later, this noise leads to a
maximum number of amplifiers that can be cascaded so that the signal-to-noise
ratio is within the limits.
Optical amplifiers can be used at many points in a communication link.
Shows some typical examples. A booster amplifier is used to boost the power of
the transmitter before launching into the fiber link. The increased transmitter
power can be used to go farther in the link. The preamplifier placed just before
the receiver is used to increase the receiver sensitivity (the minimum power
required by the receiver to function properly). Inline amplifiers are used at
intermediate points in the link to overcome fiber transmission and other losses.
Optical amplifiers can also be used for overcoming splitter losses: for example,
for distribution of cable television.

27
Chapter 4 : Optical Devices

There are currently three principal types of optical amplifiers: the erbium-doped
fiber amplifier (EDFA), the Raman fiber amplifier (RFA), and the
semiconductor optical amplifier (SOA).
Today, most optical fiber communication systems use EDFAs, due to their
advantages in terms of bandwidth, high power output, and noise characteristics.

4.6.1.2 Principles of the Erbium-Doped Fiber Amplifier

Optical amplification by EDFA is based on the process of stimulated emission,

which is the basic principle behind laser operation. In fact, a laser without
optical feedback is just an optical amplifier. Atoms and molecules are
characterized by discrete energy levels and they make transitions between these
levels of energy whenever they absorb or emit electromagnetic radiation.
Figure 25 shows two lowest-lying energy levels of an atomic system: the
ground level with energy E1 and an excited level with energy E2. These are the
lowest-lying energy states that the atom or molecule can occupy. The atom or
molecule cannot have an intermediate energy value between the two energy
values shown. Thus, we can say that the energy of the atom or molecule is
quantized.
The atom described by different energy levels can interact with electromagnetic
radiation in three distinct ways: absorption, spontaneous emission, and
stimulated emission.

28
Chapter 4 : Optical Devices

Figure 25: Atoms can interact with electromagnetic radiation in three distinct ways: (a)
absorption; (b) Spontaneous emission; (c) stimulated emission

1. Absorption. In the case of absorption, an atom occupying a lower energy state

can absorb radiation of an appropriate wavelength and get excited to an upper
energy level (Fig. 25.a). The atom occupying energy level E1 can absorb
radiation at a frequency V0 given by the following equation and get excited to
the level with energy E2:

Where h, Planck’s constant, has a value of 6.634 × 10−34 J·s. Since the energy
values of the various levels are dependent on the atom, an atom will absorb light
of certain wavelengths only, which correspond to the various pairs of energy
levels.

29
Chapter 4 : Optical Devices

2. Spontaneous emission. An atom occupying an upper level can radiate

electromagnetic radiation spontaneously and de-excite itself to the lower level
(Fig. 25b), a phenomenon known as spontaneous emission. If the energy
levels are the same as considered above, the frequency of the radiation
emitted will again be V0. Spontaneous emissions are completely random and
appear in all directions. Light coming from most optical sources, including
the sun, is due primarily to spontaneous emission.

3. Stimulated emission. Apart from these two processes, an atom occupying the
upper energy level can also be stimulated to emit radiation at the frequency
V0 by an incident light wave at that frequency (Fig. 25c) in a process called
stimulated emission. The primary difference between spontaneous and
stimulated emission is that whereas the former emission is completely random
in direction, polarization, and so on, the latter is coherent with the incident
radiation.
This implies that the radiation emitted by the atom is identical in all respects to
the radiation that stimulates the atom, and in this process the incident radiation
gets coherently amplified by the stimulated emission process.
We may mention here that in an emission process the radiation is not
monochromatic but is spread over a certain frequency range. Thus, energy levels
have a certain width (usually referred to as line width), and atoms can interact
over a range of frequencies.
Now, when the atomic system is in thermal equilibrium (i.e., in equilibrium with
the surroundings), most of the atoms will be found in the ground level. Thus, if
light at a specific wavelength (corresponding to the atom) falls on this collection
of atoms, it will result in a greater number of absorptions (from ground level to
upper level) than stimulated emissions (from upper level to ground level), and
the light beam will suffer from attenuation. On the other hand, if the number of

30
Chapter 4 : Optical Devices

atoms in the upper level could be made greater than those in the lower level, an
incident light beam at the appropriate wavelength could induce more stimulated
emissions than absorptions, thus leading to optical amplification. Known as light
amplification by stimulated emission, this is the basic principle behind an
EDFA.

Figure 26: (a) Under normal equilibrium conditions, there are more atoms in the ground state
than in an excited state, and an incident light wave undergoes attenuation. (b) When there is
population inversion, the light beam gets amplified, due to the process of stimulated emission

31
Chapter 4 : Optical Devices

Figure 27: Three lowest-lying energy bands of erbium in silica matrix

The pump laser at 980 nm excites the erbium ions from the ground level E1 to
the level marked E3, from which they make a non radiative transition to level
E2. Level E2 is a metastable level, and population inversion between levels E2
and E1 is responsible for the amplification of signals in the 1550-nm band.

Upper level could be made greater than those in the lower level, an incident light
beam at the appropriate wavelength could induce more stimulated emissions
than absorptions, thus leading to optical amplification. Known as light
amplification by stimulated emission, this is the basic principle behind an
EDFA.

Figure 27 shows the three lowest-lying energy levels of erbium ion located
within silica glass. Light from a semiconductor laser at 980 nm (called a pump
laser) excites erbium ions from the ground state to the level marked E3 i.e.,
erbium atoms in the ground state absorb the 980-nm radiation and get excited to
the level marked E3. We may mention here that the photons corresponding to
the 980-nm wavelength have an energy of about 2 × 10−19 J, which represents
the energy difference E3 − E1. Level E3 is a short-lived energy level; after a few
microseconds, ions from this level jump down to level E2. The lifetime of level
32
Chapter 4 : Optical Devices

E2 is much longer, about 12 ms. Hence, ions brought to level E2 stay there for a
significantly longer time. Thus, by pumping hard enough, the population of ions
in level E2 can be made larger than the population of level E1 thereby achieving
population inversion between levels E1 and E2. In such a situation, if a light
beam at a wavelength corresponding to the energy difference (E2 − E1) falls on
the collection, it will get amplified by the process of stimulated emission. For
erbium ions, the energy difference E2 − E1 is approximately 1.28 × 10−19 J, the
corresponding wavelength falls in the 1550-nm band, and thus it is an ideal
amplifier for signals in the 1550-nm window. Now, in the case of erbium ions
located within silica glass, due to interactions between neighboring atoms, the
energy levels are not sharp levels but are broadened: that is, ions can have
energies over a range of values, which implies that as they jump from the higher
level to the lower level, their wavelengths can have a range of values.
Hence the system is capable of absorbing or emitting over a band of
wavelengths and consequently, of amplifying optical signals over a band of
wavelengths.
Figure 28: is a schematic of an EDFA that consists of a short piece (about 20 m
in length) of erbium-doped fiber (EDF), a single-mode fiber doped with erbium
(typically, with 100 to 500 parts per million) in the core, and which is pumped
by a 980-nm pump laser through a wavelength-division-multiplexing (WDM)
coupler.

33
Chapter 4 : Optical Devices

Figure 28: Schematic of an EDFA consisting of a 980-nm pump laser, WDM coupler, and short
piece of erbium-doped fiber

Tap couplers are used to monitor the input and output from the amplifier, and
the isolator prevents reflected light from entering the EDFA.

The WDM coupler multiplexes (combines) light of wavelengths 980 and 1550
nm from two different input fibers to a single output fiber. The 980-nm pump
light is absorbed by the erbium ions to create population inversion between
levels E2 and E1. Thus, incoming signals in the 1550-nm wavelength region get
amplified as they propagate through the population-inverted doped fiber. The
tap couplers are couplers that tap a very small fraction of the light from the input
and output to make it possible to measure the signal power entering and exiting
an amplifier. These values are used to control the amplifier for constant gain or
constant output power operation.
The isolator is a device that allows light to propagate along only one direction.
The isolator is placed to prevent any reflected light from entering the amplifier,
which otherwise can get destabilized and start to oscillate like a laser.

34
Chapter 4 : Optical Devices

4.6.1.3 GAIN

The gain depends on the doping concentration and doping profile of the erbium
doped fiber, the length of the fiber, and the pump power. Typical gain values of
an EDFA are about 20 to 30 dB i.e., the output power is about 100 or 1000 times
the input power. The gain provided by the amplifier depends on the erbium
doping in the doped fiber, the length of the fiber, and the pump power. These
parameters are usually optimized for achieving the desired gain characteristics.

Figure 29: Variation of gain with EDFA length for different values of pump powers

For a given pump power there is an optimum length for achieving maximum
gain.

35
Chapter 4 : Optical Devices

For a given input pump power, as we increase the length of the doped fiber, the
gain would first increase and then after reaching a maximum would start to
decrease. This happens because as the pump propagates through the doped fiber
it gets absorbed and thus its power reduces. After propagating a certain distance,
its power is too small to create population inversion, and thus after this length,
the doped fiber would start to absorb the signal rather than amplify it. Figure 29
shows the variation of gain with the length of the doped fiber for different pump
powers. Hence, for a given pump power there is an optimum length of the doped
fiber to achieve maximum gain.
For a given length of the doped fiber, as the pump power increases, we expect
the gain to increase. At the same time, as the pump power increases it creates
more and more population inversion, and once all erbium ions in the fiber are
excited, no more erbium ions are available and hence the gain would saturate.
Figure 30 shows a typical variation of gain with input pump power for different
lengths of the doped fiber, clearly showing gain saturation with increase in
pump power.

36
Chapter 4 : Optical Devices

Figure 30: Variation of gain with pump power for different lengths of erbium-doped fiber

4.6.1.4 Gain Flattening of EDFAs

There are basically two main techniques for gain flattening: One uses external
wavelength filters to flatten the gain while the other one relies on modifying the
amplifying fiber properties to flatten the gain. In gain flattening using external
filters, the output of the amplifier is passed through a special wavelength filter
whose transmission characteristic is exactly the inverse of the gain spectrum of
the amplifier. Thus, channels that have experienced greater gain in the amplifier
will suffer greater transmission loss as they propagate through the filter, while
channels that experience smaller gain will suffer a smaller loss. By tailoring the
filter transmission profile appropriately, it is possible to flatten the gain
spectrum of the amplifier.

37
Chapter 4 : Optical Devices

Figure 31: Principle of gain flattening in EDFAs

The filter transmission profile is exactly opposite the gain profile of the
amplifier, resulting in gain flattening.
Placing the gain-flattening filter after the amplifier will result in reduction of the
net gain of the amplifier. On the other hand, if the filter is placed prior to the
signal entering the amplifier, one finds that this results in increased amplifier
noise. Thus, in practical amplifiers, the gain-flattening filter is usually placed
within the amplifier (i.e., the filter is placed after a certain length of the doped
fiber, and the filter is followed by another piece of doped fiber). In this way, one
can optimize the amplifier for maximum gain and reduced noise while retaining
a flat gain spectrum.

38
Chapter 4 : Optical Devices

4.6.1.5 Noise in EDFA

If EDFAs can compensate for the loss suffered while propagating through a
fiber, the question that arises in one’s mind is whether it is possible to traverse
an arbitrarily long distance in the fiber by periodic amplification along the fiber
link provided that the dispersion effects do not limit the distance. This is, in fact,
not possible, due to the addition of noise by each amplifier, as discussed below.
In an EDFA, population inversion between two energy levels of erbium ion
leads to optical amplification by the process of stimulated emission. As
mentioned earlier, erbium ions occupying the upper energy level can also make
spontaneous transitions to the ground state and emit radiation. This radiation
appears over the entire fluorescent band of emission of erbium ions and travels
in both the forward and backward directions along the fiber. Just like the signal,
the spontaneous emission generated at any point along the fiber can be amplified
as it propagates through the population-inverted fiber. The resulting radiation is
called amplified spontaneous emission (ASE). This ASE, which has no
relationship with the signal propagating through the amplifier, is the basic
mechanism leading to noise in the optical amplifier.

39
Chapter 4 : Optical Devices

Figure 32

In a long-distance communication link with amplifiers spaced by spans of fiber

length L, although the signal levels are maintained at the same level after every
span, the noise power keeps increasing.
Each amplifier in a chain adds noise, and thus in a fiber optic communication
system consisting of multiple spans of optical fiber links with amplifiers, OSNR
will keep falling and at some point in the link when the OSNR falls below a
certain value, the signal would need to be regenerated Thus, as the signal passes
through multiple spans and amplifiers, there is a reduction in the optical signal-
to-noise ratio. Hence there is a maximum number of amplifiers that can be
placed in a link, beyond which the signal needs to be regenerated.

40
Chapter 4 : Optical Devices

4.6.1.6 Materials used for fiber amplifiers

Figure 33: Materials used for fiber amplifiers

4.6.2 Raman Amplifier

4.6.2.1 Principles of the Raman Fiber Amplifier

When we launch a high-power light beam into an optical fiber, we observe the
appearance of Raman-scattered light at the end of the fiber referred to as
spontaneous Raman scattering. If in addition to the strong pump light we launch
a weak light beam (referred to as a signal beam), with its wavelength lying
within the band of spontaneous Raman scattering, it leads to what is referred to
as stimulated Raman scattering (SRS). In this case, the pump and signal
wavelengths are coupled coherently by the Raman scattering process and the
scattered radiation, is coherent with the incident signal radiation, much like

41
Chapter 4 : Optical Devices

stimulated emission that occurs in the case of a laser. The coherent nature of the
process implies that the incident light gets coherently amplified by SRS. It is this
process that is used to build Raman fiber amplifiers.
The other interesting feature is that no matter what the wavelength of the pump
light is, the fiber can act like an amplifier in the wavelength range corresponding
to the spontaneous Raman scattering spectrum.
Hence, if we need to amplify signals in the 1310-nm (which corresponds to 229
THz) window, we need to choose a pump wavelength of about 1240 nm (which
corresponds to 242 THz), which will give a peak Raman scattering at a
wavelength of 1310 nm, and such a pump will lead to amplification of signals at
1310 nm. Similarly, if we need to amplify signals in the 1550-nm (which
corresponds to 194 THz) window, we need to choose a pump wavelength of
about 1450-nm (which corresponds to 207 THz); in each case the pump
frequency is about 13 THz more than the signal frequency. Notice that unlike
EDFA, which operated only in specific wavelength bands, Raman amplifier can
operate in any wavelength region.
In Raman amplifiers the pump beam can propagate in the same direction as the
signal, or in the reverse direction. The former case is referred to as co-
propagating (forward pumping) and the latter as contra-propagating (backward
pumping). The Raman scattering phenomenon is an extremely fast process with
time scales in the femtosecond (10−15 s) regime. This can lead to transfer of
power fluctuations from the pump to the signal. One way to avoid this is to have
backward pumping (Fig. 31), wherein the pump fluctuation–induced gain
fluctuations get averaged out, and thus the noise in the signal due to pump
fluctuations is much lower.

42
Chapter 4 : Optical Devices

Figure 34

Figure 34: Raman amplifiers usually operate with backward pumping, wherein
the signal and pump propagate in opposite directions. Using backward pumping
the noise present in the pump does not get transferred to the signal, and this is
the Preferred pumping configuration.(a) Using a 1240-nm pump, wavelength
signals at 1310 nm can be amplified. (b) Using the same fiber if the pump
wavelength is changed to 1450 nm, 1550-nm wavelength signals can be
amplified.

43
Chapter 4 : Optical Devices

4.7 Optical Fibers

Introduction

The optical fiber is the principal component of an optical communications

system. The design of optical fibers as improved dramatically since their first
use in the 1950s and the development of new and specialized fibers is
continuous.

An optical fiber consists of:

 The information-carrying glass (the core).

 With a slightly "better" glass (the cladding).
 A protective layer of plastic (the coating) is applied over the cladding.
This combination of core - cladding - coating is the fiber.
The core and cladding properties define the type of fiber. The principal property
considered is the refraction index of the different parts of the fiber.
By changing the refraction index it is possible to change the chromatic
dispersion and the importance of non-linear effects.

Some requirements are:

 Low chromatic dispersion.

 Low PMD.
 No FWM.

44
Chapter 4 : Optical Devices

Figure 35: Typical structure of an optical fiber with common dimensions

Chromatic Dispersion

The problem of the chromatic dispersion that is present on standard single mode
fiber (SSMF acc. G.652) can be solved by using dispersion shifted fiber (acc.
G.653), which has its zero-value not at 1300nm like usual standard single mode
fiber but at 1550nm. If this type of fiber is not installed or not available
dispersion compensation fiber or dispersion compensation filters can be used
together with standard single-mode fiber.

45
Chapter 4 : Optical Devices

Figure 36: Dispersion shifted fiber

Figure 37: Examples of doping profiles for dispersion shifted fiber

4.8 Dispersion Compensation Modules

To handle chromatic dispersion, dispersion compensation can be employed.

46
Chapter 4 : Optical Devices

4.8.1 Dispersion Compensation Fiber

A short length of fiber of a large dispersion coefficient opposite to the one of the
usual transmission fiber is introduced in the transmission path. This fiber is
usually coiled up and used as a module, the length of the compensation fiber
depends on how much normal fiber is being compensated for.

Figure 38: Dispersion compensation using compensation fiber

Chirp

Immediately after power is applied to a laser there is an abrupt change in the

carrier (electron and hole) flux density in the cavity caused by the lasing
operation itself. This density of charge carriers is one factor that affects the
refractive index. In addition, the temperature in the cavity increases quite
rapidly. This temperature increase is too localized to affect the length of the
cavity immediately but it does contribute to changing the refractive index of the
material in the active region (within the cavity).
These changes in the RI of the cavity produce a rapid change in the centre
wavelength of the signal produced. In the case of semiconductor lasers a
“downward” chirp is produced. The wavelength shifts to a longer wavelength
than it was immediately at the start of the pulse. It is not a large problem in short
47
Chapter 4 : Optical Devices

distance single-channel transmissions but in long distance applications and in

WDM systems chirp can be a very serious problem. This is due to the fact that it
broadens the spectral width of the signal and interacts with other aspects of the
transmission system to produce distortion. Indeed, the chirp problem is the main
reason that people use external modulators for transmission rates in excess of 1
Gbps.

4.8.2 Chirped Fiber Bragg Grating

The longer wavelengths are reflected earlier, shorter wavelengths later. This can
be used to compensate dispersion when combined with an optical circulator.

Figure 39: Effect of chirped fiber Bragg grating

4.9 PMD Compensation

As already mentioned, the problem of PMD surfaces with WDM systems,

especially, when high bitrates are involved.
Different techniques are presently in use, including optical, optoelectronic and
electronic.
Optically one could, e.g. use a chain of polarization controllers and fixed delays
or polarization controller/beam splitter and variable optical delay. The major
48
Chapter 4 : Optical Devices

problem is that PMD is a time dependent phenomenon, so the compensation has

to be done dynamically, not statically like with usual dispersion. The answering
time for such compensators should consequently be as short as possible, e.g. in
the range of 50ms.
In the example shown on the right an EDFA is used additionally because the
PMD compensator itself has a comparatively high insertion loss which has to be
catered for.

Figure 40: Layout of a PMD compensator

4.10 Lasers and Modulators

4.10.1 Introduction

Lasers have the function of providing optical power in order to generate optical
signals. DWDM systems uses semiconductor lasers which are composed by a
combination of n and p-type doped layers drived by a current.
The quality of the generated light depends strongly on the laser structure and
several types of devices were developed according to the required application.

49
Chapter 4 : Optical Devices

Lasers can be directly or externally modulated. In the first case the laser current
is the input signal turning on or off the light output signal. In the latter case the
laser current is kept constant and an external device modulates the light.

Some requirements for these transmission units are:

 Low wavelength tolerance.

 Narrowband.
 High wavelength stability.

4.10.2 Technical Parameters

1- Spectral Width
It is a fact that most simple semiconductor lasers do not produce a single
wavelength of light. They produce instead a range of wavelengths. This range of
wavelengths is called the “spectral width” of the laser.

2- Line width
Instead of producing a continuous range of wavelengths over their spectral
width, semiconductor lasers produce a series of “lines” at a number of discrete
wavelengths. Lines themselves vary in width (in different types of lasers) very
significantly. The line width is inversely proportional to the coherence length of
the laser.

3- Power

50
Chapter 4 : Optical Devices

The signal is attenuated as it travels on the fiber and thus the higher the signal
power you use the further you can go without needing to regenerate it. In
addition, theory tells us that in an optical receiver of a given type you need a
certain fixed minimum amount of power per bit transmitted. If you have a
working system and want to double the bit rate you must double the power (or
double the receiver sensitivity). But transmitters have limits to their power49
and receivers have limits to their sensitivity. Of course, you can get a higher bit
rate by reducing the attenuation (by shortening the distance between stations)
thereby increasing the signal power at the receiver. In some systems, signal
power, more than fiber capacity is the limiting factor.

Power control
One way of ensuring consistent operation over time (and perhaps saving the cost
of cooling) is to monitor the light level produced by the laser and to adjust bias
currents accordingly. This is often done by using a monitor diode at the back
facet of the laser. Provided the back facet lets some light out (it usually does)
you can measure the output power produced and control the laser accordingly.

4- Operating Wavelength (or Range)

Of course, lasers must be able to operate on wavelengths appropriate to the
system being designed. The operating wavelength of a laser depends on the
materials used for lasing (in exactly the same way as the wavelength of an LED
depends on the same materials) and on the geometry of the laser cavity.

5- Frequency (Wavelength) Stability

In a single-channel system using incoherent detection, a bit of instability
(wander) in the laser wavelength doesn't matter too much. However, if the

51
Chapter 4 : Optical Devices

system is using WDM techniques, each laser must keep within its allocated band
and wander matters a lot.
Fabry-Perot lasers vary an enormous .4 nm per degree Celsius of temperature
variation. Most of the single-mode lasers are significantly better than this, but
temperature control is critical.

Temperature Control
For most communications lasers temperature control is critical. Some of the
lower cost devices can be satisfactorily operated with just good heat sinking.
However, most lasers intended for long distance telecommunication applications
are packaged with thermoelectric coolers and thermostatic Control.

4.10.3 The wavelength of light emitted by Laser diode depend on

4.10.3.1 The distance between the mirrors by the following

formula

4.10.3.2 The wavelength of light is inversely proportional to the

band gap energy
The higher the energy the shorter the wavelength. The formula relating electron
energy to wavelength is given below:

52
Chapter 4 : Optical Devices

4.10.3.3 Band gap Energy and Possible Wavelength Range in

Various Materials

Table 1

Chirp or carrier frequency shift, are oscillations in the transmitted wavelength.

These oscillations are due to the dependence of the laser material on the applied
current. This effect occurs especially in directly modulated lasers and a practical
result is the increase of the signal bandwidth.
It should also be noted, that lasers also serve as pump-lasers for EDFAs. Those
lasers are of a rather high power and often are cascaded via different coupler
devices to give the required high pump power.

53
Chapter 4 : Optical Devices

Figure 41: Laser configuration with direct modulation

Figure 42: Laser configuration with external modulation

4.10.4 Lasers

4.10.4.1 Fabry-Perot-Laser

The easiest form of a semiconductor laser is the Fabry-Perot-Structure:

Basically, these lasers are planar light emitting diodes consisting of a n-type and
a p-type doped layer. Between these two layers, we've got the active region
where the lasing takes place.
To be able to select a lasing wavelength this region has to have the form of an
interferometer, which in this case can be done by cleaving. The length of the

54
Chapter 4 : Optical Devices

lasing cavity determines the lasing wavelength. Like a Fabry-Perot

Interferometer, the FP Laser is not very "good": In addition to our desired
wavelength we also get "side-wavelengths" which effectively give us a larger
linewidth.

Therefore the FP Laser is also called "Multi-Longitudinal-Mode" (MLM) Laser.

4.10.4.2 Distributed Feedback Lasers

DFB Lasers are commonly used for DWDM purposes. Here, we're applying
basically the same trick as in the Dielectric Thin Film filter: We don't use simply
mirrors at the end, but introduce a layer-structure in the lasing cavity, a bit like
in a Bragg Grating. This has a similar effect like a FP structure, just much better.
While having a slightly more complicated structure, DFB Lasers fit the
requirements very well:

 Narrow peaks (about 0.0001nm).

 Wavelength range 1520 to 1565nm and above (third and fourth window).
 Stable.

As they really only produce one lasing mode, the DFB lasers are also known as
SLM (Single Longitudinal Mode) Lasers.

55
Chapter 4 : Optical Devices

Figure 43: Fabry-Perot laser

Figure 44: DFB laser

4.10.4.3 Comparison

While FP Lasers are quite commonly used in usual single channel transmission
as sources of "grey" light, DFB lasers are the common laser source for DWDM
systems, due to there much smaller linewidth.

56
Chapter 4 : Optical Devices

Figure 45: Properties of different LED and laser types

4.10.4.4 Tunable Lasers

These devices pump out light at different wavelengths, and can switch from one
wavelength to the other very quickly. The newest development in laser
technology is tunable semiconductor lasers.

Those lasers are highly desirable as they have a lot of features which are
interesting for DWDM:

 One laser for several colors (spares!).

 Use in OADMs.
 Use in optical cross connects.
 Use for optical protection switching, etc.

From the technical point of view lasers can be tuned by varying the refractive
index of the lasing cavity. Changing that refractive index is equal to changing
the length of the cavity and thus the selected sending wavelength.

57
Chapter 4 : Optical Devices

4.10.4.5 Pump Lasers

For the pumping of EDFAs pump lasers are used, which usually put out rather
high powers, up to the watt range. The requirements are here more polarization
properties and power then linewidth.

4.10.4.6 Laser Classes

When talking about lasers, a few words about laser safety are in order, as with
DWDM products rather high laser powers are frequently used.

According to the safety provisions all laser products must be assigned to a Class
from 1 to 4 according to their hazard potential, labeled and fitted with the
required protective equipment.

An instruction program, to be repeated annually, is recommended for persons

who use Class 3 A to 4 laser products. Laser safety officers competent in the
field only need to be appointed for Class 3B and 4 laser products, however.
Unintentional radiation emission must be prevented for Class 2 to 4 laser
products. Class 2 or3A laser products in the service or work area must be clearly
recognizable and permanently labeled. Only qualified and trained employees
should be assigned to install, adjust and operate class 3A to 4 laser equipment.

Use of Class 2 and 3A lasers does not endanger the skin. An eye hazard may be
posed in the visible range of laser radiation if the blink reflex is suppressed and
Class 2 or Class 3A laser beams are shone directly into the pupil at close range.
In general there can be an eye hazard if beams emitted from Class 3A lasers are

58
Chapter 4 : Optical Devices

viewed with collecting optical instruments such as magnifying glasses,

telescopes etc.

Figure 46: Possible laser hazards

Figure 47: Table of laser classes

59
Chapter 4 : Optical Devices

4.10.5 Modulators

The modulation of lasers can be done by modulating the laser itself or using an
external modulator.
Direct modulation brings the problem of „chirping“, i.e. frequency oscillations
of the laser. The reason for this is that the index of refraction of the active region
changes slightly with the applied current, therefore the effective length of that
region changes and also the emitted wavelength.

As this is a problem in DWDM systems external modulators are often used.

They solve the problem of chirping, but they cannot avoid a certain broadening
of the line, which is physically inevitable when modulating a signal.

4.10.5.1 MQW Modulator

One type of modulator is the semiconductor MQW (Multiple Quantum Well)

modulator. Its functionality is simple:. when voltage is applied, light is
absorbed. It has the further advantage that it can be produced on the same
substrate as the laser itself.

4.10.5.2 Mach Zender Modulator

Its function is analogous to that of the Mach Zender filter. By varying the phase
of one arm of a M-Z filter the two parts of the signal either interfere
constructively ("on") or destructively ("off"). LiNbO3 can be used for this phase
control.

60
Chapter 4 : Optical Devices

Figure 48: Semiconductor MQW modulator

Figure 49: Layout of a Mach Zender modulator

4.11 Photodectectors

In all the optical transmission systems, usually two types of photodetectors are
used:

 PIN-Diodes.
 APD Diodes.

61
Chapter 4 : Optical Devices

4.11.1 PIN Diode

Figure 50: Layout of a PIN

The name comes from the structure of the diode: p-doped, intrinsic and n-doped
semiconductor material is used in a layer structure.
The diodes are reversely biased. Each incident photon causes an electron-hole
pair to be produced, the electron and hole drift towards the electrodes which in
turn causes a measurable current. This current is proportional to the number of
incident photons.
The wide intrinsic (i) layer has only a very small amount of dopant and acts as a
very wide depletion layer. There are a number of improvements here:
 It increases the chances of an entering photon being absorbed because the
volume of absorbent material is significantly increased.
 Because it makes the junction wider it reduces the capacitance across the
junction. The lower the capacitance of the junction the faster the device
response.

62
Chapter 4 : Optical Devices

4.11.2 APD Diode

APDs amplify the signal during the detection process. They use a similar
principle to that of “photomultiplier” tubes used in nuclear radiation detection.
In the photomultiplier tube:
1. A single photon acting on the device releases a single electron.
2. This electron is accelerated through an electric field until it strikes a target
material.
3. This collision with the target causes “impact ionization” which releases
multiple electrons.
4. These electrons are then themselves accelerated through the field until they
strike another target.
5. This releases more electrons and the process is repeated until the electrons
finally hit a collector element.
The result of the above process is that a single arriving photon can result in the
production of between 10 and 100 or so electron-hole pairs.

Figure 51: Layout of a photodiode

63
Chapter 4 : Optical Devices

Therefore APDs are especially well suited for applications where a very high
sensitivity is needed. It should be mentioned though that the avalanche process
is rather "noisy", causing a fluctuation of the gain factor.

Figure 52: Functionality of PIN and APD diode

Figure 53: Spectral response of different detector materials

64
Chapter 4 : Optical Devices

4.12 Isolators, Circulators and Connectors

The very-narrowband lasers used in DWDM are highly sensitive to reflected

power.
Such external reflections act as additional resonators and the result might be
wavelength instabilities, mode jumps, noise, etc.
Even a power 6 orders of magnitude lower than the transmitted power (about -
60dB) can significantly disturb the laser.

4.12.1 Isolators

Isolators are, to put it simply, devices that let light pass in one direction without
attenuation and do not allow light to flow in the reverse direction. In that respect
they are a kind of "optical diodes".

Thus, reflected power is highly attenuated (by about 30dB).

Technically speaking, isolators can be constructed as a combination of

polarization rotators and linear polarization filters.

Figure 54: Function of an optical isolator

65
Chapter 4 : Optical Devices

4.12.2 Circulators

A device of similar structure like the isolator is the circulator. It works as a kind
of multiport isolator, transmitting the input of port 1 to port 2, input of port 2 to
port three and so on.

Figure 55: Optical circulator

4.12.3 Connectors

Connectors have to be especially selected to guarantee low reflection. They

should satisfy two criteria:
 Physical contact.
 Angled contact.

66
Chapter 4 : Optical Devices

Types like E2000 or /APC or, generally speaking, high-return-loss connectors

are therefore good choices.
Additionally, all connectors have to be cleaned very carefully to avoid breaks in
the physical contact and also to avoid power being absorbed in those impurities.

Figure 56: Connector types

67
Chapter 4 : Optical Devices

Figure 57: Connector types

4.13 Optical Switching Units

Not really a part of DWDM lines, but of growing interest for the all-optical
network is the development of optical cross connects and of optical add-drop
multiplexers.
The latter could be implemented already with a number of techniques discussed
above, like tunable fiber-bragg-gratings, DTFs or tunable lasers. Newer
developments include more sophisticated and cheaper solutions.
The really big optical cross connect remains the challenge though. At the
moment optical traffic is connected with the help of distribution panels, within
the next years is will be necessary to switch at least part of the exploding data
traffic via quickly reconfigurable optical cross connects.

68
Chapter 4 : Optical Devices

Basically, it is possible to switch the traffic either electrically or optically. The

electrical switching fabric is only capable of handling comparatively low bit rate
traffic, therefore we will concentrate on the optical variants. Generally speaking,
larger switching fabrics can be created by cascading smaller units in one way or
another.
In addition to the switching fabric a number of additional modules can be used,
including wavelength transponders, regenerators, etc.

Figure 58: Functionality of an optical cross connect

At the moment quite a few different technologies are used on the way to the all
optical switch or the optical add-drop multiplexer. Differences are e.g. the
maximum size of the matrix or the switching time. Among those methods are:

69
Chapter 4 : Optical Devices

 Thermo-optical switches
Light is passed through glass that is heated up or cooled down with electrical
coils. The heat alters the refractive index of the glass, bending the light so
that it enters one fiber or another. The same can be done using polymer
technology.

Figure 59: Thermo optical switches

 MEMS (micro-electro-mechanical systems)

Micro-electro-mechanical systems (MEMS) are in widespread use in some
other industries, but their use for telecom applications is relatively recent.
In telecom, MEMS has become synonymous with the arrays of tiny tilting
mirrors used for optical switching fabric, although the same technology is
being used to make a wide range of other components as well.
Since MEMS creates so many mirrors on a single chip, the cost per switching
element is relatively low. However, since it involves moving parts, MEMS is
fairly slow to switch – requiring milliseconds to do so. This is fine for lambda
provisioning or restoration but is too slow for optical burst switching or optical
packet switching applications.

70
Chapter 4 : Optical Devices

Conventional MEMS works by reflecting the beam of light from the surface of a
tiny mirror. MEMS systems have moving parts, and the speed at which the
mirror moves is limited. By applying more current, the mirror can move faster,
but there's a limit to how much current can be sent into the array of mirrors. If
this weren't bad enough, it seems that the speed and angular displacement terms
in the calculation of the required current have integer powers of around 4 or 5,
and so the bottom line is that we have to put a lot of current into the array for a
small improvement in speed. By changing the mirror design so that the angle
through which light is bent is smaller, it's possible to achieve faster switching
speeds. This technique is known as "fast MEMS."
MEMS arrays can be built on a single-chip, single-plane approach. In other
words they are 2 dimensional (2D MEMS). In a simplistic approach it’s also
possible to stack a number of 2D MEMS arrays on top of each other to create a
3D MEMS array. In fact, real 3D MEMS systems are somewhat more complex
than this, but the general principle holds.
A huge drawback of 3D MEMS is the fact that the thousands of mirrors require
complex software to coordinate their operations. In particular, one vendor has
suggested that there are over a million lines of code in their implementation
(although the reference may be to the overall switch software and not just the
MEMS subsystem). While it’s possible to test software extensively, the
opportunity for bugs increases geometrically with the size of the code base.
Advantages are e.g. a large possible scale of the matrix (1000x1000 is under
discussion), low loss connectivity and compact design.
Some facts that have to be taken care of are mechanical stability and long term
reliability.

71
Chapter 4 : Optical Devices

Figure 60: 2-axis motion of MEMs OCX mirror

72
Chapter 4 : Optical Devices

Figure 61: 256x256 OXC switching array

73
 Chapter 5: Measurements

Chapter 5: Measurements

Aim of study
This chapter introduces measurement techniques.

Contents Pages

5.1 Introduction 2
5.2 Measurement Techniques 3

1
 Chapter 5: Measurements

Chapter 5
Measurements
5.1 Introduction

The required measurements when installing DWDM equipment are, in basis,

similar to the measurements performed with conventional optical systems.
Additional fiber parameters, like PMD, have to be taken into account and
spectrum analysis becomes indispensable.

The key points in field measurements for DWDM systems are:

 Optical Time-Domain Reflectometry.

 PMD measurements.
 Optical Power at certain reference points.
 Optical Spectrum Analysis.

Note that these points concern the evaluation and optimization of the link state
and not necessarily the DWDM equipment.

Some typically required optical measurement equipments are:

 OTDR meter.
 PMD meter.
 Optical Power meter.
 Optical Spectrum Analyzer.

The first two equipments are used to measure specific fiber characteristics and
the last concern the quality of the optical transmission.

2
 Chapter 5: Measurements

Even after a link status approval with these measurements, the quality and
lifetime of the connection cannot be assured due to additional effects that cannot
be easily measured in the field.

Additional measurements can be performed at the channel level by terminal

equipments. In the common case of using SDH terminal equipments,
performance measures at the SDH level can be made by evaluating the B1 and
B2 bytes. In this situation the entire system (SDH equipment, DWDM
equipment and link state) is being evaluated for that specific channel.

5.2 Measurement Techniques

5.2.1 Optical Time-Domain Reflectometry

OTDR measurements are used to evaluate the attenuation characteristics of a

fiber.
These measures concern the attenuation coefficient, splice loss or plug
connections allowing the installer to determine with considerable precision what
are the problems in a fiber and where they are located.

The OTDR technique consists of sending impulses to the fiber and measuring
the time delay and intensity of the backscattered signal. The backscatter effect
occurs because of the same reasons that we have attenuation on optical fiber,
scattering.
What happens is that some of the light gets reflected back due to changes in the
molecular density of the glass. Measuring this light is equivalent to measuring
fiber attenuation.

The structure of an OTDR is basically a light source to emit signal pulses and an
optical receiver connected to a data processing unit.
3
 Chapter 5: Measurements

The emitted signal is sent directly into the fiber and the incoming reflection
directed to the receiver by a beam splitter. The light source is synchronized with
the receiver so that time delay between outgoing and incoming signals can be
measured.

Much like the radar principle, the intensity of the reflected signal depends on the
fiber attenuation and occasional bends, twists or splices. The time delay of the
reflected signal is related to the position of the fault in the fiber.

The result of an OTDR measure is a curve of reflected signal intensity versus

time delay whereas the curve slope corresponds to the attenuation coefficient.
Discontinuities in the curve indicate additional attenuation due to splices and
peaks indicate large reflections, possibly due to reflections on connectors.

Figure 1: Basic structure of an OTDR analyzer

4
 Chapter 5: Measurements

Figure 2: Example of an OTDR output

5.2.2 PMD Measurements

As referred before, PMD measurements are indispensable before installing

WDM equipments.

The most basic technique to measure PMD is to transmit a signal in one end of
the fiber and measure the time delay between the received signal at different
polarizations. However, PMD measurement can become more complex when
considering wavelength dependence and mode coupling effects.

Even so, time must be taken into consideration because PMD depends strongly
on environmental conditions that change constantly. This means that regular
evaluations have to be performed in order to ensure the quality of the link.

5
 Chapter 5: Measurements

Because PMD measurements cannot be based on reflectometry like the OTDR

test equipment has to set in both sides of the transmission fiber. A transmitter
and receiver.

In some cases, mostly depending on the measurement equipment,

communication between test equipments as to be set up. However, actual
interferometer based equipments do not require this and both equipments can be
setup independently.

Figure 3: Configuration of a PMD measurement

5.2.3 Optical Power Measurements

After installing WDM equipment there are a certain number of test procedures
concerning power functionalities that have to be performed in order to check for
system or connection failures. Not only the output power of the equipment has
to be verified but also the power at the fiber terminal points and possible loss
due to dirty connectors or broken patch cords.

ITU-T as defined a set of reference points recommended for measure according

to a typical WDM connection. These points are:

6
 Chapter 5: Measurements

 S1...Sn – Reference points at the output connectors of the transmitter

equipment for channel 1 to n respectively.
 RM1...RMn – Reference points before the input connectors of the optical
multiplexer for channels 1 to n respectively.
 MPI-S – Reference point at the output connector of the optical booster.
 R' – Reference point before the input connector of an in-line amplifier.
 S' – Reference point at the output connector of an in-line amplifier.
 MPI-R – Reference point before the input of the preamplifier.
 SD1...SDn – Reference points at the output connectors of the demultiplexer
for channels 1 to n respectively.
 R1...Rn – Reference points before the input connectors of the receiver
equipment for channels 1 to n respectively.

Note that in some cases the transmitter equipment can be a composition of a

transmitter and a transponder. In this case the optical power between these two
elements must also be controlled.
The values obtained with these measurements must be in agreement with the
data resulting from planning in terms of power budget.

Figure 4: ITU-T Rec. G.692 Optical power measurement reference points

7
 Chapter 5: Measurements

5.2.4 Optical Spectrum Analysis

5.2.4.1 Measurement Principle

Figure 5: Spectrum Analyzer - Display Schematic

There are many occasions where we want to look at the wavelength spectrum of
the signal(s) on a fiber. One such occasion would be to examine the wavelength
spectrum of a WDM system to help understand system operation and to
diagnose faults. A spectrum analyzer scans across a range of wavelengths and
provides a display showing the signal power at each wavelength.

From this display we can calculate:

1. The power levels of each channel.
2. OSNR of each channel.
3. The spectral width of each channel.
4. Any interference between channels such as crosstalk possibilities.
8
 Chapter 5: Measurements

5. By connecting it in different places through the system we can track many

potential problems such as laser drift etc.

Figure 6: Spectrum Analyzer - Logical Structure

The logical structure of the device is shown in Figure 6.

 Light input from the fiber is passed through a tunable Fabry-Perot filter.
 The filter is scanned at quite a slow rate (perhaps 10 times per second)
through the range of wavelengths that we want to examine.
 Optical output of the FP filter is fed to an APD to convert it to electronic
form.
 The output of the APD will contain rapid variations due to modulation of
the signal in each channel etc. These modulations are averaged out
electronically so that the electrical signal level now represents the average
power level of the optical signal (average over a few milliseconds).
The electronic signal now needs to be scaled logarithmically as we need the
y-axis scale to be in dBm.
 The electronic signal is now fed to the y-axis control of an oscilloscope.
 The x-axis is swept across in synchronism with the wavelength setting of
the FP filter.
 This results in a display similar to that in the figure.

9
 Chapter 5: Measurements

Like OTDRs, spectrum analyzers vary widely in their capabilities and prices.
They range from large, very accurate and expensive laboratory instruments to
small, much less expensive devices about the size of a laptop computer. You can
even buy one that does not have a display and instead connects to your laptop
computer.

In using one you need to be aware of the resolution (minimum width) of each
wavelength measured and also of the accuracy of the instrument.

10
Telecom Egypt, Training and Development sector.
General department for Development of Technical, and Information systems & technology Skills
5 El Mokhaim El Daem Sreeet,
Nasr City, Cairo, Egypt.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, recording or otherwise.

Program name: Dense Wavelength Division Multiplexing (DWDM)

Program code: PD0104000020202