-
Free Space Optic (FSO) Link Design
Project Report
Submitted in Partial Fulfillment of the Requirements For the
Degree of
Bachelor OF TECHNOLOGY
IN
ELECTRONICS & COMMUNICATION ENGINEERING
By Falak Shah (09bec082)
Kavish Shah (09bec083)
Under the Guidance of Prof. Dhaval Shah
Department of Electrical Engineering Electronics &
Communication Engineering Program
Institute of Technology, Nirma University Ahmedabad-382481
May 2012
-
CERTIFICATE
This is to certify that the Minor Project Report entitled Free
Space Optic Link
Design submitted by Falak Shah (09bec082) & Kavish Shah
(09bec083) as the
partial fulfillment of the requirements for the award of the
degree of Bachelor of
Technology in Electronics & Communication Engineering,
Institute of Technology,
Nirma University is the record of work carried out by his/her
under my supervision
and guidance. The work submitted in our opinion has reached a
level required for
being accepted for the examination.
Date: 20/11/2012
Prof. Dhaval Shah
Project Guide
Prof. P.N.Tekvani
HOD (Electronics & Communication Engineering)
Nirma University, Ahmedabad
-
Acknowledgement
It gives us great pleasure in expressing thanks and profound
gratitude to Prof. Dhaval Shah,
Department of Electronics & communication Engineering,
Institute of Technology, Nirma University
for his valuable guidance and continual encouragement throughout
the Minor project. We are heartily
thankful to him for continuous suggestion and the clarity of the
concepts of the topic that helped us a
lot during the project.
We are also thankful to Prof. Yogesh Trivedi, Department of
Electronics & communication
Engineering, Institute of Technology, Nirma University for his
kind support in understanding the
fundamentals of wireless communication.
Lastly, we would like to thank our friends for providing us
constant inspiration and support
during various aspects of the project.
FALAK SHAH
[09BEC082] KAVISH SHAH
[09BEC083]
-
II
Abstract
Free Space Optics (FSO) is a communication technology uses that
light propagating in free space to
transmit data between two points. The technology is useful where
the physical connections by the
means of fiber optic cables are impractical due to high costs or
other considerations. Free-space-
optical links can be implemented using infrared laser light or
LEDs as a source and the receiver with
photodiode at the receiver end. This project aims at
understanding all that is needed in order to create
a transceiver for a FSO link. Beginning with a formal definition
and overview of the technology, it
goes on to explain the considerations for the transmitter and
receiver. Moving ahead, the channel
models for optical communications have been explained in the
final chapter. The practical design
issues for the transmitter as well as receiver have been
presented along with the theoretical
explanations. Lastly, the circuit designed for function as
transceiver and its working is covered.
Beaming light through the air offers the speed of optics without
the expense of fiber - IEEE Spectrum August 2001
-
III
Index
Chapter
No.
Title Page
No.
Acknowledgement I
Abstract II
Index III
List of Figures VI
List of Tables VIII
Nomenclature VIII
1 Introduction
1.1 Definition 1
1.2 Factors behind market growth 1
1.3 A Case Study 3
1.4 Advantages of FSO 4
1.5 Limitations of FSO 5
1.6 Applications of FSO 5
1.7 A typical system model 6
1.8 Objectives Of This Project 6
2 Transmitter for Free Space Optical Communication
2.1 Block diagram and practical circuit layout of FSO
transmitter
8
2.2 Qualities of the optical source 9
2.2.1 LED v/s LASER 9
2.2.2 A novel development in light sources-
VCSEL
10
2.2.3 Frequency(wavelength) of operation 11
2.3 Modulation Schemes in Optical Wireless
Communications
11
2.3.1 On-Off Keying 11
2.3.2 RZ OOK 12
2.3.3 Manchester Encoded Signal 12
2.3.4 Pulse Position Modulation 13
2.3.5 Comparison of Modulation Schemes 13
2.3.6 Conclusion 16
-
IV
2.4 All about LASERs 16
2.4.1 Unique characteristics 16
2.4.2 Types of LASERs available 16
2.4.3 Working of LASER Diode 17
2.4.4 Classes of LASERs based on eye safety
and power
17
2.4.5 Selection of LASER for FSO applications 17
2.5 All about LEDs 18
2.5.1 LED Operation and Characteristics 18
2.5.2 Types of LEDs and lifetimes 19
2.6 Driver Circuits 19
2.6.1 LED Driver 19
2.6.2 LASER Driver Circuit 21
2.7 Practical Design Steps 22
2.7.1 PC to Transceiver Interface 22
2.7.2 Using Hyper-terminal to send a file to a
remote computer
24
2.7.3 Selection of light source 25
2.7.4 Practical Driver Design models 27
2.7.5 Power Calculation 28
3 Receiver for Free Space Optical Communication
3.1 Block Diagram for Receiver of FSO 31
3.2 Photo Detector 31
3.2.1 Requirements of photo diode 31
3.2.2 Working principle 32
3.3 Different types of photo detector 32
3.3.1 PIN photo diode 33
3.3.2 Material selection for photo detector 33
3.3.3 Avalanche photo diode (APD) 34
3.3.4 PIN Photo Diode v/s APD 35
3.4 Noise in receiver 35
3.4.1 Dark current noise 36
3.4.2 Quantum noise 36
3.4.3 Thermal noise 36
3.5 Pre-amplifier 36
3.5.1 Low impedance pre-amplifier 37
-
V
3.5.2 High impedance pre-amplifier 37
3.5.3 Trans-impedance pre-amplifier 38
3.5.4 Selection of pre-amplifier 38
3.6 Decision Circuitry 38
4 Channel Models
4.1 Introduction to channel parameters 39
4.1.1 Atmospheric Turbulence 39
4.1.2 Scintillation Index 40
4.2 Various Channel models 41
4.2.1 Lognormal channel model with and
without perfect CSI
41
4.2.2 Gamma-Gamma Channel model 43
4.2.3 Negative Exponential Model 45
4.2.4 K channel model 45
4.2.5 I-K Channel model 47
4.3 Comparison of Channel Models 49
Conclusion 50
References 51
-
VI
LIST OF FIGURES
Fig. No.
Title Page
No.
1.1 10-Gbps FSO link, deployed by MRV Communications' Tele-Scope
1
1.2 Last-Mile Connectivity 3
1.3 Terabeam Transceiver 6
2.1 Block diagram of FSO Transmitter 8
2.2 Practical form of transmitter 8
2.3 Light power v/s current for LED and LASER 9
2.4 Small-signal frequency responses of an LED and an LD with
negligible parasitic
effects
10
2.5 BER performance for OOK (NRZ and RZ), from Eq. (2.38), and
L-PPM (L = 2, 4,
and 8)
14
2.6 Power spectrum of the transmitted signals for OOK (NRZ and
RZ), and L-PPM (L
= 2, 4, 8)
15
2.7 Example of (a) LED drivers, (b) shunt driver 20
2.8 Output power v/s current for LASER diode 21
2.9 LASER Driver Circuits 22
2.10 Ports DB-9 AND MAX232 23
2.11 RS232 to TTL interface 24
2.12 HyperTerminal screen 25
2.13 variety LEDs available 25
2.14 a typical LASER diode 26
-
VII
2.15 Collection of and gates as LASER Driver 27
2.16 LASER Driver using op-amp 28
2.17 Overall attenuation v/s distance plot for different
wavelengths 29
3.1 Block diagram of Simple Receiver 31
3.2 V-I characteristic of photo diode 32
3.3 energy band diagram of PIN photo diode 33
3.4 responsivity v/s wavelengths 34
3.5 sensitivity v/s Photodiode areas 35
3.6 various kinds of Noises 35
3.7 Low impedance circuits 37
3.8 Trans Impedance circuits 38
3.9 Decision Circuitry 38
4.1 HVB21 Models 40
4.2 Performance of perfect CSI at receiver for log-normal
channel model 42
4.3 Performance of imperfect CSI at receiver for log-normal
channel model 43
4.4 Performance of Gamma-Gamma channel model 45
4.5 Performance of K channel model 46
4.6 Performance of I-K channel model 48
-
VIII
LIST OF TABLES
1.1 Comparison of FSO with other technologies in terms of cost
2
2.1 Comparison of different baseband intensity modulation
techniques. 15
2.2 Relationship among Material, System Wavelength, and Band Gap
Energy for LED
Structures
19
2.3 Power Calculation 30
NOMENCLATURE
FSO Free Space Optics
LOS Line Of Sight
R.I. Refractive Index
S.I. Scintillation Index
CSI Channel State Information
BER Bit Error Rate
SNR Signal to Noise Ratio
IM/DD Intensity Modulation/Direct Detection
OOK On Off Keying
HVB Hufnagel Valley Boundary model
-
1
Chapter 1
Introduction
1.1 Definition
Free Space Optics, the industry term for Cable-free Optical
Communication Systems, is a
line-of-sight optical technology in which voice; video and data
are sent through the air (free
space) on low-power light beams at speeds of megabytes or even
gigabytes per second [1]. A
free-space optical link consists of 2 optical transceivers
accurately aligned to each other with a
clear line-of-sight. Typically, the optical transceivers are
mounted on building rooftops or behind
windows. These transceivers consist of a laser transmitter and a
detector to provide full duplex
capability. It works over distances of several hundred meters to
a few kilometres.
Figure 1.1 10-Gbps FSO link, deployed by MRV Communications'
Tele-Scope 10GE. Feb 12, 2010.
1.2 Factors behind market growth
Fibre optics provides an excellent solution for high bandwidth,
low error requirements and
can serve as the backbone for the internet infrastructure. Most
of the recent trenching to lay fibre
has been to improve the metro core (backbone). Carriers have
spent billions of dollars to increase
network capacity in the core, of their networks, but have
provided less lavishly at the network
edges. This imbalance has resulted in the "last mile
bottleneck." Service providers are faced with
-
2
the need to turn up services quickly and cost-effectively at a
time when capital expenditures are
constrained.
From a technology standpoint, there are several options to
address this "last mile connectivity
bottleneck" but most don't make economic sense.
Fibre - Optic Cable: Without a doubt, fibre is the most reliable
means of providing optical
communications. But the digging, delays and associated costs to
lay fibre often make it
economically prohibitive. Moreover, once fibre is deployed, it
becomes a "sunk" cost and cannot
be re-deployed if a customer relocates or switches to a
competing service provider, making it
extremely difficult to recover the investment in a reasonable
timeframe. Connecting with fibre
can cost US $100 000-$200 000/km in metropolitan areas, with 85
percent of the total figure tied
to trenching and installation [2].
Radio frequency (RF) Wireless: RF is a mature technology that
offers longer ranges distances
than FSO, but RF-based networks require immense capital
investments to acquire spectrum
license. Yet, RF technologies cannot scale to optical capacities
of several gigabits. The current RF
bandwidth ceiling is 622 megabits. When compared to FSO, RF does
not make economic sense
for service providers looking to extend optical networks
[3].
Wire & Copper-based technologies: (i.e. cable modem, T1s or
DSL): Although copper
infrastructure is available almost everywhere and the percentage
of buildings connected to copper
is much higher than fibre, it is still not a viable alternative
for solving the connectivity bottleneck.
The biggest hurdle is bandwidth scalability. Copper technologies
may ease some short-term pain,
but the bandwidth limitations of 2 megabits to 3 megabits make
them a marginal solution [3].
Table 1.1 Comparison of FSO with other technologies in terms of
cost [5].
-
3
The need for FSO is accelerated by several factors. First, more
and more bandwidth is needed
by the end user, which means that more data access must be
provided. As a fact, the number of
internet users will be increased to approximately 796 million by
the end of 2005 [4]. It has been
shown that the FSO implementation is not only cheaper compared
to the fibre optics, but also
compare to other popular technologies like the digital
subscriber line (DSL) or cable modem
services [5]. Providing last mile connectivity is extremely
difficult and expensive. In metropolitan
areas, an estimated 95 percent of buildings are within 1.5 km of
fibre-optic infrastructure. But at
present, they are unable to access it. Street trenching and
digging are expensive, cause traffic jams
and displace trees.
Figure 1.2 Last-Mile Connectivity
Working via a hub building, free-space optics can connect each
of the three buildings at the
left to a central office of competitive local exchange carrier
at 100-Mb/s. This office is a node on
a metropolitan-area ring, which is connected to a regional ring
by means of conventional fibre-
optics equipment [5].
1.3 A Case Study
In one free-space optics business case, a competitive local
exchange carrier (CLEC) has an
agreement with a large property management firm to provide
all-optical 100-Mb/s Internet access
capability to several buildings located in an office park. The
carrier is building its network by
-
4
leasing regional dark fibre rings and long-haul capacity from a
wholesale fibre provider. It has
identified a potential hub, or point-of-presence, less than a
kilo-meter from the office park and
within sight of one of its central offices. The CLEC currently
has no fibre deployed to target
customer buildings [see Figure 2].
When fibre was compared with free-space optics, deployment costs
for service to the three
buildings worked out to $396 500 versus $59 000, respectively.
The fibre cost was calculated on a
need for 1220 meters: 530 meters of trunk fibre from the CLECs
central office to its hub in the
office park plus an average of 230 meters of feeder fibre for
each of the runs from the hub to a
target building, all at $325 per meter. Free space optics is
calculated as $18 000 for free-space
optics equipment per building and $5000 for installation.
Supposing a 15 percent annual revenue
increase for future sales and customer acquisition, the internal
rate of return for fibre over five
years is 22 percent versus 196 percent for free-space
optics[2].
1.4 Advantage of FSO
FSO systems can carry full-duplex (simultaneous bi-directional)
data at gigabit-per-second
rates over metropolitan distances of a few city blocks to a few
kilo-metres [1].
Data is transmitted in the visible to infrared light spectrum
(terahertz spectrum range). Unlike
most of the lower-frequency portion of the electromagnetic
spectrum, this part above 300
GHz is unlicensed worldwide and does not require spectrum fees.
The only limitation on its
use is that the radiated power must not exceed the limits
established by the International
Electro technical Commission (Standard IEC60825-1).
Since data is beamed over the air and not via fibre-optic cable,
the carrier does not have to
lease or deploy wired infrastructure.
Cost Effectiveness: These free-space systems require less than a
fifth the capital outlay of
comparable ground-based fibre-optic technologies [5]. FSO thus
has compelling economic
advantages.
Rapid Deployment: Free-space optics enables very fast deployment
of broadband access
services to buildings. Installing an FSO system can be done in a
matter of days - even faster if
the gear can be placed in offices behind windows instead of on
rooftops.
-
5
The time-consuming and expensive process of getting permits and
trenching city roads is
completely avoided. Using FSO, a service provider can be
generating revenue while a fibre-
based competitor is still seeking municipal approval to dig up a
street to lay its cable.
1.5 Limitation of FSO
Here we are using air as a medium. So performance is highly
dependent on environment. So,
if the environment is not good our data rate is limited. We have
to design our model carefully
based on the environmental condition of the particular
place.
Line of sight is necessary. So, if there is an obstruction is
there between transmitter and
receiver this setup cannot be established. To avoid this, we
have to set this on the roof of the
tall buildings.
Comparing with optical fiber, its range is very much limited,
which also is dependent on
environmental condition. So we can use this only for LAN or MAN.
We cannot use this in
overseas condition.
As with any laser, eye safety is a concern. There are two
wavelengths of light, 850nm and
1550nm. The 1550nm units are, generally, safe due to the fact
that the human eye (aqueous
lens) absorbs the light energy and no damage will be sustained
to the retina. The 850nm
wavelength can cause damage to the retina. The person will not
be aware of the damage since
the retina has no pain receptors and invisible light does not
cause a blink reflex. Therefore
850nm lasers need to be installed carefully and ensure that
human eyes will receive the signal.
This is easily done by mounting the lasers on a wall.
1.6 Application of FSO
'Last-Mile' Network Solutions.
Temporary Network Provision.
CCTV Security Applications.
Industrial estates, Science parks and university campus where
number of separate buildings,
separated by roads or other obstacles, between which
communications links are frequently in
demand Backhaul for wireless cellular network.
Military Applications where more security is required.
Satellite Laser Communication.
LAN-to-LAN connections on campuses at Fast Ethernet or Gigabit
Ethernet speeds.
Speedy service delivery of high-bandwidth access to optical
fibre networks.
Re-establish high-speed connection quickly (disaster
recovery).
-
6
1.7 A typical system model
A typical free space optics communication system consists of: a
small laser source that can be
directly modulated in intensity at fairly high data rates; a
beam shaping and transmitting telescope
lens to transmit the laser beam through the atmosphere toward a
distant point; a receiving lens or
telescope to collect and focus the intercepted laser light onto
a photo detector; and a receiver
amplifier to amplify and condition the received communication
signal. The transmitted laser beam
passes through the atmosphere and can be absorbed, scattered or
displaced, depending on
atmospheric conditions and on the wavelength of the laser
source. In the case of high atmospheric
turbulence, an active tracking device may have to be used to
align the beam. Active tracking is
not necessary if sufficient laser power can be made available,
if the divergence of the beam can be
expanded and if the building and alignment are stable. Figure 3
is a photograph of a FSO unit that
operates at 1.55 m wavelength and can provide a data link at
speeds up to a Gbit/s [1]. The unit,
made by Terabeam, has a small single transmitted laser beam and
a larger receiver telescope lens.
It also has an optical video alignment TV that the installer
uses for initial alignment to the other
rooftop or window office unit.
Figure 1.3 Terabeam Transceiver
1.8 Objective of the project
1. Design a moderate speed FSO data link with transfer rates up
to 100Kbps.
2. Operating distances 200 to 300mts.
3. Much Cheaper as compared to the commercially available
equipment. The estimated basic
design cost was around Rs 3000/-. A commercial 850-nm
transceiver for a 10-100-Mb/s unit
spanning a few hundred meters can cost as much as $5000.
4. Design using readily available, cheap and indigenous
components instead of expensive,
specialized components.
5. Compact & Easy to install reliable Hardware.
-
7
6. Very less setup times.
7. Provide an excellent platform for design and testing of more
advanced FSO projects and
communication protocols.
For this, we first describe the components of the transmitter
and then receiver, both of which
are the elements of the link to be established. We aim at
designing the link for testing over small
distance under laboratory conditions and hence wont be including
any tracking mechanism.
-
8
Chapter 2
Transmitter for Free Space Optical Communication
2.1 Block diagram and practical circuit layout of FSO
transmitter
Figure 2.1 Block diagram of FSO Transmitter
Figure 2.2 Practical form of transmitter
The transmitter, which consists of two parts; an interface
circuit and a source drive circuit,
converts the input signal to an optical signal suitable for
transmission. The drive circuit of the
transmitter transforms the electrical signal to an optical
signal by varying the current flow through
the light source. This optical light source can be of two types:
(1) a light-emitting diode (LED) or
(2) a laser diode (LD). The information signal modulates the
field generated by the light source
and after passing through optics for concentrating the generated
beam moves to the channel. The
peltier element acts to cool the laser diode as it is very
sensitive to temperature.
Source Modulator Driver Circuit Light Sorce Beam
Concentrators
Cooling
Mechanism
-
9
2.2 Qualities of the optical source
It is important that the frequency response of the light source
exceeds the frequency of the
input signal as light is the carrier. It is this feature that
regulates the frequency of operation.
The light source should launch its energy at angles that maximum
portion is transmitted to
receiver end.
Faster speed of operation
long lifetime
high intensity
reasonably monochromatic (small spectral width)
temperature stability
2.2.1 LED v/s LASER
(i) LEDs do not produce so concentrated a beam as LASER and
hence are preferred for
indoor applications due to eye safety issues. In outdoor
environments, the properties of
LASER Diodes such as narrow spectra, high power launch
capability, and higher
access speed make these devices the favourite optical source for
long-distance and
outdoor directed-LOS links.
(ii) Light power v/s current as they differ considerably as
shown in below figure LEDs show linear characteristics near origin
whereas LASER above threshold. Also, a LASER at 30
C requires 70 mA to output 2 mW of optical power may require in
excess of 130 mA at
80 C). This implies that more current is required before
oscillation. So for lower current
supply LASERs are unsuitable.
Figure 2.3 Light power v/s current for LED and LASER showing
temperature dependency
of LASER [6].
(iii) Speed of operation. Laser diodes are much faster than LEDs
due to LED having
spontaneous recombination and LASER having simulated emission.
Modulation
bandwidth up to few MHz for LED as compared up to 10 GHz for
LASER.
-
10
Figure 2.4 Small-signal frequency responses of an LED and an LD
with negligible
parasitic effects.
(iv) Brightness of LASER as a light source is higher as it
combines the properties of an LED and a cavity reflector, producing
an external light radiation that is higher in power and has
a better focused beam as compared to LED.
2.2.2 A novel development in light sources-VCSEL
Vertical cavity surface emitting lasers (VCSEL), which offer a
safer peak wavelength
at 1.55 m [7], are becoming an increasingly attractive option
for outdoor and even indoor applications due to their
well-controlled, narrow beam properties, high modulation
bandwidth,
high-speed operation, excellent reliability, low power
consumption, and the possibility of
having array arrangements. It provides these advantages and is
cheaper in cost too. They
provide better carrier confinement for lesser heat dissipation
and better current flow. The
optical output power needs to be over 10 mW if the device is to
be used as a light source for
FSO outdoor applications. The optical output power of a
conventional VCSEL is not adequate
as a FSO light source, for conventional VCSEL devices to be used
as a light source for FSO
they are used as arrays to provide sufficient power [7].
2.2.3 A novel development in light sources-VCSEL
-
11
Free space optical communication typically operating in
unlicensed Tera-Hertz
spectrum bands (wavelength 8001700 nm) is used as it provides
improvement in signal bandwidth over operation in the RF
environment [9]. To achieve emission at a desired specific
wavelength, the material must allow a band gap variation, which
can be achieved through
different level of doping. Lasers in the 780925-nm and
15251580-nm spectral bands meet frequency requirements and are
available as off-the-shelf products. Most optical transmission
technology is designed to operate at a wavelength of 850 nm.
However, the latest technology
includes 1.55-m devices [8], such as above mentioned VCSELs
which are attractive due to the fact that, up to certain power
levels, they do not harm the human eye.
2.3 Modulation Schemes in Optical Wireless Communications
In optical wireless systems, the intensity of an optical source
is modulated to transmit signals.
This is because of the complexity and expensiveness of coherent
modulation techniques like
phase and frequency modulation [9]. A great number of
applications use Intensity
Modulation/Direct Detection (IM/DD) as the
transmission-reception technique due to its
simplicity of implementation [10].
Modulation schemes like QAM make more efficient use of the
bandwidth than schemes like
OOK. Researchers have found it difficult to apply advanced
modulation techniques like QAM on
lasers because of the way lasers are generated. If this were
achieved, lasers should be able to
attain greater QAM levels than microwaves because of their high
signal-to-noise ratio [12].
Applying more bandwidth-efficient techniques to lasers is not
necessary because of the wide
bandwidth available to lasers. Furthermore, lasers are unlikely
to interfere with other laser signals
because of their small beam spread. Therefore, there is not a
high motivation to research
bandwidth-efficient modulation for lasers.
For digital data transmission, there is no practical alternative
to digital modulation since it
provides source coding (data compression) as well as channel
coding (error detection/correction).
The transmission of the digital data can be done on a bit-by-bit
basis (binary encoding) or on a
bit-word basis (block encoding).
2.3.1 On-Off Keying
The simplest type of binary modulation scheme is OOK. In an
active high OOK
encoding, a one is coded as a pulse, while a zero is coded as no
pulse or off field. To
restrict the complexity of the modulator, the pulse shape is
chosen to be rectangular. The bit
rate is denoted as
Rb = 1/ Tb
Where Tb is the bit duration; and is directly related to the
rate at which the source can be
switched on and off. The normalized transmit pulse shape for OOK
is given by
-
12
In the demodulator, the received pulse is integrated over one
bit period, then sampled and
compared to a threshold to decide a one or zero bit. This is
called the maximum likelihood
receiver, which minimizes the bit error rate (BER).
Another important parameter that needs to be considered in any
modulation scheme is the
bandwidth requirement. The bandwidth is estimated by the first
zeros in the spectral density of
the signal. The spectral density is given by the Fourier
transform of the autocorrelation
function.
2.3.2 RZ OOK
There is a variation of OOK, in which the pulse shape is high
for only a fraction of bit
duration dTb with 0 < d
-
13
the pulse time is one half the bit times, and these results in
higher required bandwidths. The
BER is now the probability that the bit half interval containing
the pulse does not produce the
higher value. Since the Manchester signalling is identical to
2-PPM, all the results for PPM
can be applied directly in analyzing this scheme.
2.3.4 Pulse Position Modulation
In block encoding, bits are transmitted in blocks instead of one
at a time. Optical block
encoding is achieved by converting each word of l bits into one
of L = 2l optical fields for
transmission. One of the most commonly used optical block
encoding schemes is PPM, where
an input word is converted into the position of a rectangular
pulse within a frame. The frame
with duration f T is divided into L slots and only one of these
slots contains a pulse. This
scheme can also be denoted as LPPM, in order to emphasize the
choice of L. The transmit
pulse shape for L-PPM is given by
Since L possible pulse positions code for log2L bits of
information, the bit rate is
Rb = log2L/Tf .
The optimum L-PPM receiver consists of a filter bank, each
integrating the photocurrent in
one pulse interval. The demodulated pulse is taken to originate
from the slot in which the most
current level was found. If the demodulated pulse position is
the correct pulse position, log2L
bits are decoded correctly. Otherwise, we assume that all L -1
wrong position are equally
likely to occur. Therefore bit errors usually occur in
groups.
The BER for Manchester signals for L=2 is identical to the BER
of OOK modulation.
2.3.5 Comparison of Modulation Schemes
In order to compare different modulation schemes, the power and
bandwidth
efficiency, defined as the required power and bandwidth at a
desired transmission speed and
BER quality, are to be compared. Power efficiency can readily be
derived from the BER
expressions.
To achieve a given BER value, the power requirement in OOK and
L-PPM scheme can be
written as
-
14
It is fairly obvious that 8-PPM has the best BER performance,
and hence is the most power
efficient scheme. To achieve a given BER value, the comparison
of power requirement in
OOK and L-PPM scheme show that L-PPM requires a factor of ((L
/2) log2L)0.5
less power
than OOK to obtain a particular BER performance.
Figure 2.5 shows the BER performance of OOK, for both NRZ and
RZ, and L-PPM for L = 2,
4, and 8.
Figure 2.5 BER performance for OOK (NRZ and RZ), from Eq.
(2.38), and L-PPM (L = 2 , 4,
and 8)
-
15
Another important measure of performance is the bandwidth
efficiency. The bandwidth
required for modulation can be estimated from the first zero of
the transmitted signals power
spectrum. Fig.9 illustrates the spectral density envelope
(without the Dirac impulses) of the
transmitted signals for OOK and L-PPM. Note that only positive
frequency is shown and the
frequency is normalized to the bit rate Rb.
Figure 2.6 Power spectrum of the transmitted signals for OOK
(NRZ and RZ), and L-PPM (L
= 2, 4, 8).
The bandwidth efficiency is defined as the ratio between bit
rate and required bandwidth (in
bps/Hz). The required bandwidth is
B = Rb for OOK and B =LRb /log2L; for L-PPM.
Thus, the bandwidth efficiency of L-PPM can be shown to be at
least 1.9 times worse than
OOK. To conclude, the comparison results are also summarized
as
Table 2.1- Comparison of different baseband intensity modulation
techniques.
-
16
2.3.6 Conclusion
Signal transmission in optical wireless systems is generally
realized using an intensity
modulation technique. For FSO systems, although the power
efficiency is inferior to PPM,
OOK encoding is more commonly used due to its efficient
bandwidth usage and robustness to
timing errors [11]. Furthermore, the slot timing capability
places a lower limit on the slot
times that can be used in PPM systems, limiting their advantage
over OOK systems.
Therefore, in this research work, FSO systems are designed using
intensity modulation/direct
detection (IM/DD) with an OOK technique.
2.4 All about LASERs
2.4.1 Unique characteristics
Lasers have unique characteristics that set them apart from
other light sources.
Monochromatic: The output of a laser is light of a single colour
(the light is very nearly a
single wavelength). The difference between the output of a laser
and that of an incandescent
light bulb is analogous to the difference between a single tone
and white noise.
Coherence: All of the light waves start at the same instant in
time (all the waves are in step)
Directionality: The beam is either well collimated to start or
can easily be collimated or
otherwise manipulated. These special characteristics are very
important for laser
communication.
2.4.2 Types of LASERs available
Diode laser
Helium-Neon laser
Argon/Krypton ion laser
Carbon Dioxide laser
Helium-Cadmium (HeCd) laser
Of particular interest to FSO applications is the diode laser
source due to small size, ease of
handling, cost effectiveness, being electrically run and
functioning at the desired frequency
range. Most of these lasers are also used in fibre optics;
therefore, availability is not a
problem.
-
17
2.4.3 Working of LASER Diode
A 'laser diode', refers to the combination of the semiconductor
chip - driven by low
voltage power supply - that does the actual lasing, along with a
monitor photodiode chip (for
regulation of laser diode current using optical feedback
control) in the same package as the
laser diode. Because the band gap of a semiconductor depends on
the crystalline structure and
chemical deposition of the material, diode lasers can operate at
a specific wavelength by
changing the composition of the material system.
2.4.4 Classes of LASERs based on eye safety and power
Class 1- products are defined as inherently safe, which means
that they are safe even when
viewed with an optical instrument. They are not supposed to
present any hazard to the human
eye independently of their wavelength of operation and the
exposure time.
Class 2- applies to sources between 400 and 700 nm (visible
light), and it states that lasers in
this category are safe if the blink or aversion response of the
eye operates (the blink or
aversion response is the natural ability of the eye to protect
itself by blinking.
Class 3- laser with power range between 1 mW and 0.5 W [13]. The
energy emitted by this
type of source is dangerous not only if seeing a direct beam,
but also when seeing reflections.
Damage may occur in a period of time shorter than the blink
response of the eye.
The new regulation by IEC addresses the power density in front
of the transmit
aperture rather than the absolute power created by a laser diode
inside the equipment. For
example, the laser diode inside the FSO equipment can actually
be Class 3B even though the
system itself is considered to be a Class 1 or 1M laser product
if the light is launched from a
large-diameter lens that spreads out the radiation over a large
area before it enters the space in
front of the aperture. The new regulation also states that a
Class 1M laser system operating at
1550 nm is allowed to transmit approximately 55 times more power
than a system operating
in the shorter IR wavelength range, such as 850 nm, when both
have the same size aperture
lens.
2.4.5 Selection of LASER for FSO applications
We have used Class II LASERs with power up to 1 mill watt. These
lasers are not
considered an optically dangerous device as the eye reflex will
prevent any ocular damage.
(i.e. when the eye is hit with a bright light, the eye lid will
automatically blink or the person
will turn their head so as to remove the bright light. Class II
lasers won't cause eye damage in
this time period. No known skin exposure hazards exist and no
fire hazard exist. FSO uses
class II laser device.
-
18
Laser diodes with wavelengths around 635 nanometres are
available which is a red
beam. Deep Red (670 nm) and beyond, IR (780 nm, 800 nm, 900 nm,
1550 nm, etc.) up to
several micrometers are also available. Green and blue laser
diodes which have been produced
in various research labs, only operated at liquid nitrogen
temperatures, had very limited life
spans (~100 hours or worse), or both. Due to the sensitivity
curve of the human eye, a
wavelength of 635 nm appears at least 4 times brighter than an
equivalent power level at 670
nm. Thus, shorter wavelength laser diodes will be preferred
choice.
2.5 All about LEDs
Light-emitting diodes (LEDs) are semiconductor light-emitting
structures. Due to their
relatively low transmission power, LEDs are typically used in
applications over shorter
distances with moderate bandwidth requirements up to 155 Mbps.
Depending on the material
system, LEDs can operate in different wavelength ranges.
Advantages of LED sources include
their extremely long life and low cost.
2.5.1 LED Operation and Characteristics When an n- and a p-type
material are brought together, the electrons and the holes
recombine in the interface region. However, during this process,
a barrier (neutral region) is
generated and neither the electrons nor the holes have enough
energy to cross this barrier.
With zero bias voltage applied to the structure, the charge
movement stops and no further
recombination takes place. However, when a forward bias voltage
is applied, the barrier
decreases and the potential energy of the free electrons in the
n-type material increase. In
other words, the potential energy level of the n- side is
raised. The forward bias voltage
provides the electrons and holes with sufficient energy to move
into the barrier region. When
an electron meets a hole, the electron falls into the valence
band and recombines with a
hole. During this process, energy is released in the form of a
photon. The wavelength of the
light emitted during this process depends on the energy band gap
width Wg, as shown in the
following equation.
Wg= 1.24/
Table 2.2 shows a listing of semiconductor material systems and
the relationship between
band gap energy and emission wavelength. For free-space optical
applications, the Gallium
Arsenide (GaAs) and Aluminium Gallium Arsenide (AlGaAs) material
systems are of interest
because the emission wavelengths fall into the lower wavelength
atmospheric window around
850 nm.
-
19
Table 2.2 Relationship among Material, System Wavelength, and
Band Gap Energy for LED
Structures
2.5.2 Types of LEDs and lifetimes With respect to light
emission, LEDs are one of two types: surface-emitting LEDs or
edge emitting LEDs. Whereas surface-emitting diodes have a
symmetric Lambertian radiation
profile (a large beam divergence, and a radiation pattern that
approximates a sphere), edge
emitting diodes have an asymmetrical elliptical radiation
profile. LEDs are commercially
available in a variety of packages: TO-18 or TO-46. Some
packages include micro lenses to
improve the quality of the output beam.
LEDs typically operate at a modulation bandwidth between 1 MHz
and 100 MHz. LEDs that
can be used in applications that require a higher modulation
bandwidth are not capable of
emitting high optical power levels. A 1 mW LED is already
considered to be high power at
higher modulation speed. However, the lifetime of LEDs (the
length of time until the power is
reduced to half of the original value) can be as high as 105
hours. This corresponds to about
11 years.
2.6 Driver Circuits
High speed LED and LASER drivers are becoming more prominent in
the digital
industry due to increased speeds of data transfer. The term
"high speed" in the market sense
refers to data rates greater than one Mbps. There is a switching
speed and light intensity
trade-off that hinders the design for some applications.
2.6.1 Types of LEDs and lifetimes
a) Working
The LED driver controls the voltage across the diode and either
turns the diode "on" or
"off". The LED turns "on" when a forward bias greater than the
turn on voltage is applied,
and the diode begins to emit light. The driver must be designed
to produce a large enough
voltage so that the diode will give off the desired intensity.
When the driver turns the
LED "off" the voltage should adjust the diode to barely conduct.
This is necessary
because it takes far too long to turn on a diode once it has
been completely turned off.
-
20
The photon intensity from the diode when it is barely conducting
must be negligible when
designing a product.
b) Circuit Design
The basic construction for any LED driver, in core remains as
shown here. The LED are
operated with switching on and off of a current in the range of
a few tens to a few
hundreds mA. This current switching is performed in response to
input logic voltage
levels at the driving circuit. A common method of performing
this current switching
operation of the LED is shown in Figure 2.7(a). The common
emitter configuration is
adapted with a bipolar transistor providing current gain. In
this circuit, the output current
flowing through the LED is set by the value of R2. However, the
switching speed is
limited by the diffusion capacitance which means that the
bandwidth and current gain
have the trade-off relation. To increase the switching speed,
low impedance driver (shunt
driver) is developed as shown in Figure 2.7(b).
Figure 2.7 Example of (a) LED drivers, (b) shunt driver
c) Working of Shunt Driver Circuit
The shunt driver circuit simply puts the LED in parallel with
the driver output. This
circuit is patented because the old LED drivers had the diode in
series with the driver
output, and while in parallel the rise and fall times of signals
are much faster.
The output of the driver consists of a high speed switching
transistor. The carriers
built up in the junction of the diode are swept out quickly
through the shunt connection to
the transistor. When the transistor starts conducting it
reverses the direction of minority
carriers and recombines electron-hole pairs much quicker than
natural recombination.
Essentially the diode is "on" when the transistor stops
conducting and vice versa. This
circuit varies with different designs but is mainly used to
increase signal integrity, reduce
-
21
jitter, and decrease the extinction ratio. Maxim designed a high
speed LED driver circuit
with a data rate of 155 Mbps.
d) Higher Data rate implementation
Sumitomo Electric Industries has created a shunt LED driver
circuit using GaAs
semiconductors that is successfully tested at bit rates of 400
Mbps over a few centimetres
[14]. The shunt driver circuit is frequently used in current FSO
research. The bit rates
produced by high speed LED drivers would satisfy speed
requirements for smaller
communication networks (Mbps range), but not for the larger tier
networks (Gbps range).
2.6.2 LASER Driver Circuit
The laser transmitter circuitry is somewhat different from the
LED drivers since as
shown in the light-versus-current characteristics of the laser
(figure 10), the light output is
very small until the DC current reaches the threshold current.
After the threshold current, the
optical power is approximately linear with current. The problem
associated with typical lasers
is that the characteristic curve is not linear at high current
and tends to shift to the right as both
the temperature and device ages are increased. This results in
unwanted changes in output
power, extinction ratio, and turn-on delay in digital
transmission. Thus, the laser should be
biased near the threshold current when it is in off state to
reduce the turn-on delay and to
minimize any relaxation oscillations, and also to easily
compensate for variations in threshold
due to temperature and device ageing.
Figure 2.8 Output power v/s current for LASER diode
For biasing the laser, a bias control circuit is necessary in
designing laser driver
circuits. A simple laser driver circuit used to connect the
output of a current driver circuit
directly to the laser diode is shown in figure 12(a). The
threshold current for a laser is
provided by Vbias and modulation current is provided by source
resistor, Rmod, respectively.
-
22
This type of single-ended laser driver is typically used with
low operating speed due to the
unwanted parasitic inductance from the packages bonding wires.
When this parasitic
inductance is combined with the capacitance of the laser driver
circuits and lasers, it degrades
output of the lasers rise time and causes power supply current
ripple. Another example of the
laser driver circuit is shown in Figure 12(b) when the driver
circuit and the laser are placed in
different package. In this topology, a matching circuitry
between the driver and the laser is
necessitated to overcome the large impedance mismatch. In this
circuit, Ibias controls the DC
threshold current and Imod provides the modulation current for
the laser.
Figure 2.9 LASER Driver Circuits
2.7 Practical Design Steps
2.7.1 LASER Driver Circuit
While designing the transmitter, the first thing in transmitter
is the PC to transceiver interface.
Available options:
1) We can use MAX232 IC by MAXIM to convert RS232 signals from
PC to TTL and CMOS
logic levels and vice versa. MAX232 has now replaced the
previous 1488 and 1489
transmitter and receiver IC pair and is most commonly used in
any serial interfacing with
RS232. It is available at cost of approximately Rs 30 to 40 (at
Robokits).
-
23
Figure 2.10 Ports DB-9 AND MAX232
Working:
Serial RS-232 (V.24) communication works with voltages (between
-15V ... -3V are
used to transmit a binary '1' and +3V ... +15V to transmit a
binary '0') as per the electrical
specifications contained in the EIA (Electronics Industry
Association) for the RS232C
standard. Also, the region between +3 and -3 volts is undefined
and open circuit voltage
should never exceed 25 volts. On the other hand, classic TTL
computer logic operates
between 0V ... +5V (roughly 0V ... +0.8V referred to as low for
binary '0', +2V ... +5V for
high binary '1' ). Modern low-power logic operates in the range
of 0V ... +3.3V or even lower.
So, the maximum RS-232 signal levels are far too high for
today's computer logic electronics.
Therefore, to receive serial data from an RS-232 interface the
voltage has to be reduced, and
the 0 and 1 voltage levels inverted. In the other direction
(sending data from some logic over
RS-232) the low logic voltage has to be "bumped up", and a
negative voltage has to be
generated, too.
The MAX232 from Maxim just needs one voltage (+5V) and generates
the necessary
RS-232 voltage levels (approx. -10V and +10V) internally. The
MAX232 has a successor, the
MAX232A. The ICs are almost identical, however, the MAX232A is
much more often used
(and easier to get) than the original MAX232, and the MAX232A
only needs external
capacitors 1/10th the capacity of what the original MAX232
needs.
It should be noted that the MAX232 (A) is just a
driver/receiver. It does not generate the
necessary RS-232 sequence of marks and spaces with the right
timing; it does not provide a
serial/parallel conversion. All it does is to convert signal
voltage levels.
2) RS232 data cable which automatically converts the digital TTL
signal to RS232 and back
as mobiles need 5V or 3.3V supply can also be used.
3) MAX232N by Texas Instruments. It needs at least 1F capacitors
as compared to 0.1 F
capacitors in MAX2322A by MAXIM. It is also cheaper in
comparison- costs around Rs. 26.
4) Circuit for RS232 to TTL interface level converter can also
be used
-
24
Figure 2.11 RS232 to TTL interface
Complete explanation of this can be viewed at:
http://www.botkin.org/dale/rs232_interface.htm
2.7.2 Using HyperTerminal to send a file to a remote computer We
then make use of HyperTerminal available in Windows XP for
serial
communication testing using our PC. HyperTerminal is a program
that you can use to connect
to other computers, Telnet sites, and bulletin board systems
(BBSs), online services, and host
computers, using your modem, a null modem cable or Ethernet
connection.
1) Open HyperTerminal.
2) Open a saved connection file or create a new connection.
3) Connect to the remote computer.
4) On the Transfer menu, click Send File. In the Filename box,
type the path and name of the file
you want to send.
5) In the Protocol list, click the protocol your computer is
using to send the file. Click Send.
-
25
Figure 2.12 HyperTerminal screen
Setting up HyperTerminal
1) Performing this task does not require you to have
administrative credentials. Therefore, as a
security best practice, consider performing this task as a user
without administrative
credentials.
2) To open HyperTerminal, click Start, point to all programs,
point to Accessories, point to
Communications, and then click HyperTerminal.
3) You must have an active HyperTerminal session connected prior
to performing this
procedure. Both the sending computer and the receiving computer
must be using the same file
transfer protocol.
4) If you use the Modem protocol to transfer data, the remote
computer will receive the file
automatically and will not need to perform a manual receive
procedure.
2.7.3 Selection of light source
Figure 2.13 variety LEDs available
-
26
a) LEDs
Today, we have available in the market a large variety of LEDs
and laser diodes.
When looking for LED, we have IR LEDs at our desired range of
850 nm to 880 nm,
1200 to 1300 nm as well as 1550 nm available (although 850 nm
and 1550 nm remain the
most commonly available and used). They come in a large variety
of price ranges from Rs.
5 to as high as Rs. 500 (several manufacturers were contacted
and the prices are as quoted
by them for bulk purchase of IR LEDs) based on their spectral
width, power radiated, half
power angle and maximum operating frequency. A few of the IR LED
manufacturers are
listed below.
1) Hamamatsu
2) Dense light Semiconductors PVT. LTD.
3) Cree LED lights
4) Ray science Innovation Ltd.
5) Ad labs Pvt. LTD.
6) New age instruments and materials private Ltd.
For our experiment, we use typically,
850 nm LED (cheaper than 1550 nm although having slightly higher
atmospheric
attenuation index). 1550 nm would be preferred choice in longer
distance involving
designs.
20-50 nm full width half maximum spectral width
10 mW to 100 mW power are commonly used although up to 350 mW
available(as per
distance of transmission)
7 deg to 45 deg half power angle (as per cost consideration)
b) LASER Diode
Figure 2.14 a typical LASER diode
-
27
A wide variety of LASER sources are available but due to reasons
explained in the theory
portion, we make use of only LASER diodes. Laser diodes best
suited for FSO
applications are the above mentioned IR wavelength diodes. LASER
diodes are available
in IR range (up to 830 nm) but cost as high as Rs. 8000 for LD
of power 10 mW and 830
nm. . However a talk with the sales executives of New age
instruments and materials PVT.
Ltd and other laser diode suppliers and authorised agents for
Hamamatsu in India was
suggestive that laser diode with 1550 nm remains unavailable.
But for our low cost
experimental purposes, we can use the cheaply available red
lasers of 635 nm range. Here,
we look for
Power radiated (calculations shown below)
Beam spread angle in vertical and parallel directions.
Laser diodes are fairly monochromatic so, spectral width is not
so much of a concern.
Also, at all times in FSO, point lasers and not line lasers
should be considered as in line
lasers, divergence increases.
2.7.4 Practical Driver Design models
The HSDL4220 infrared LED is originally unsuitable for 10 Mbit/s
operations. It has a
bandwidth of 9 MHz, where 10 Mbit/s Manchester-modulated systems
need bandwidth of
around 16 MHz. Operation in a usual circuit with current drive
would lead to substantial
signal corruption and range reduction. Therefore Twibright Labs
developed a special
driving technique consisting of driving the LED directly with
15-fold 74AC04 gate output
in parallel without any current limitation. The same idea has
been put into action in the
circuit used in the project where a bunch of and gates have been
used for supplying higher
current when there is too low current for driving laser.
Figure 2.15 Collection of and gates as LASER Driver
-
28
Another method of design is to make use of the op-amp
The low speed transmitter mainly consists of an op amp, a BJT
(Bipolar Junction
Transistor) and a LASER. The main idea of the circuit is to
function as a constant current
source switched on and off by an external pulse generator. When
the supply (VCC) is high
enough, the current through the laser diode will be dependent on
the size of the resistor
(REmitter), the voltage applied to the positive port of the op
amp and the maximum output
swing of the op amp. An example circuit using this has been
shown here.
Figure 2.16 LASER Driver using op-amp
2.7.5 Power Calculation:
The FSO link model can be divided into three separate parts, the
optical transmitter,
the optical receiver and the transmission through the
atmosphere. For the calculation of
the link power budget the power equivalent Gaussian beam concept
is used [17].
1. OPTICAL TRANSMITTER SYSTEM
The attenuation of the transmitter system is given by the sum of
losses of its parts. The
attenuation of the cover window WT and attenuation due to the
Laser Diode to
Transmission medium coupling are given by its practical
measurement. The usual values
are WT = -1 dB and LD = -1 dB. The attenuation of the
transmitter system TS = -2.0 dB.
-
29
2. OPTICAL RECEIVER SYSTEM
The receiver system includes the receiver lens, the
concentrator, interference filter and the
detector. The attenuation of the receiver system RS is given by
a sum of looses of its parts.
The attenuation of cover window WR, receiver lens RXA, and the
attenuation due to the
transmission medium to PD coupling PD are given by its practical
measurement. The
practical measurement at the wavelength = 1550 nm gives us the
values: WR = -1 dB,
RXA = -0.3 dB and PD = -3 dB. For the wavelength = 1550 nm the
value of overall
receiver attenuation RS = -4.3 dB and for the 830 nm wavelength
the value RS = -6.8 dB.
3. NEP
Next, we have noise equivalent power calculations as (example of
Type : C30737E-500,
Perkin Elmer is considered) . For it, at 1 Mhz [15],
NEPDiode = Itot[A] / S [A/W]*(frequency)^0.5 = 1.14 x 10-12
W/Hz-1/2
B = 1 MHz. This gives NEP=-89.4 dBm.
4. Atmospheric attenuation
For the FSO-link the transmission through the atmosphere could
be described with
attenuation due to the particles influence and propagation
attenuation. The propagation
attenuation 12 is given by the link distance L12 and the full
transmitted angle represented
by the back distance L0. The attenuation due to the particles
influence part is for the clear
atmosphere and the wavelength of 1550 nm given by 1part = 0.48
dB/km [16]. The
overall attenuation of the atmosphere is given by a sum
This can be seen from the graph plotted below.
Figure 2.17 Overall attenuation v/s distance plot for different
wavelengths
-
30
5. LINK POWER BUDGET
For distance limit calculations it is necessary to calculate the
minimal value of the receiver
systems input power PRXA and the output power of the transmitter
system PTXA. The
minimum power PMIN to guarantee requested bit error rate BER =
10-6 is equal to the
photodiodes noise equivalent power NEP increased by the signal
to noise ratio SNR =
13.5 dB. The required minimum power at the photodiode PPD is
then PMIN increased by
the link power margin (20 dB reserve used). The minimal value of
the receiver systems
input power PRXA is then PMIN increased by the attenuation of
the receiver system
PRXA. Here for 50 mW system, PTX= 17 dBm is considered and range
calculated.
Table 2.3 Power Calculation
All units in dBm
NEP = -89.4 dBm
PMIN= NEP + 13.5 + 20 = -55.9 dBm
This means for the above mentioned laser diode of 50 mW power
and avalanche
photodiode, we get a theoretical maximum range of 10.5 km for
850 nm and 11.5 km for
1550 nm wavelengths used.
-
31
Chapter 3
Receiver for Free Space Optical Communication
3.1 Block Diagram for Receiver of FSO
The main purpose of the receiver is to detect the signal in form
of light, then to convert it into
electrical form, amplify it and detect the data that was
transmitted. The design of the receiver is
very complicated because of some reasons like; it must be able
to detect distorted or weak signals
and to make accurate decisions based on that distorted signal.
Optical receiver consists of mainly
3 parts called photo detector, preamplifier and signal
processing circuitry. Firstly, photo detector
converts optical signal into electrical signal that is current,
and this current changes with the light
level or intensity. Then, this electrical signal is very weak
due to distortion and it needs to be
amplified for further electronic processing. So, preamplifier is
used here. And finally for decision
making circuitry and some electronic circuit for further signal
processing is used [18].
Figure 3.1 Block diagram of Simple Receiver
3.2 Photo detector
The function of the receiver is to absorb photons and emit
electrons, means to produce the
electric current from the incident photons. Photo detector must
meet very high performance
requirements.
3.2.1 Requirements of photo diode
High response or sensitivity at the operating wavelength: high
current should be produced in
response to incident light.
Linearity: in order to minimize the distortion for analogue
transmission
Low internal noise: detector itself should produce low noise for
high performance
Sufficient bandwidth / fast response speed: helps at higher data
rate
-
32
Insensitivity to external conditions: it should not be affected
by external conditions like
temperature
Other requirements like cost effectiveness, long life,
reliability, high stability and small size.
3.2.2 Working principle Light is nothing but the bundle of
photons. When the light is incident on the material,
the photon is absorbed by an atom. If the energy level of the
photon is greater than the band
gap energy of the material, photon causes an electron emission
from conduction band to
valance band. So, free electron is generated, which is mobile
and it becomes an electric
current when potential difference is applied [19].
Where EP is energy of photon, Eg is energy of electron, h is
planks constant, f is
frequency.
Figure 3.2 V-I characteristic of photo diode
As shown in V-I characteristic of photo diode, the value of
reverse current increases
with increase in light intensity. And for particular value of
light intensity, current firstly
increases and then becomes saturated.
3.3 Different types of photo detector
There are several types of photo detectors like,
photo multiplier
pyroelectric detectors
semi-conductor based photo conductors
photo diode
photo transistor
Though photo multiplier is capable of low noise and very high
gain, it is not used in free space
optics because of its large size and high voltage
requirements.
-
33
Pyroelectric materials are suitable for detecting high speed
laser pulses using principle of
converting photons into heat, but it is not suitable for free
space optics due to its quite flat
response over broad spectral band [20].
In semi-conductor based photo conductor, photo diodes are mainly
used because of its small
size, fast response time and high sensitivity over photo
transistors.
There are 2 types of photo diodes those are mainly used,
o PIN photo diode
o Avalanche photo diode
Avalanche photo diode has its internal gain, while PIN photo
diode has not its internal gain,
which is well compensated by its larger bandwidth [18].
3.3.1 PIN photo diode PIN photo diode consists of P and N region
separated by a larger and very lightly
doped intrinsic region (i). When very high reverse bias voltage
is applied across this diode, the
intrinsic region is completely depleted. Now a photon is
incident on the diode, if the incident
photon has its energy greater than band gap energy of the
semiconductor, the photon gives its
energy to electron and electron gets excited from valance band
to conduction band. This
process free electron-hole pair, which is also known as photo
carriers, and when high electric
field is applied in the depletion region, it causes the photo
carriers to get separated and get
collected across reverse bias junction. Finally, this process
gives rise to flow the current in the
external circuit, known as photo current [21].
Figure 3.3 energy band diagram of PIN photo diode
The energy band diagram of PIN photo diode is shown here, when
photon has energy
greater than band-gap energy, it gives energy to electron as
shown in the figure.
3.3.2 Material selection for photo detector In selecting the
material for photo detector, there are mainly two parameters. One
is
responsivity and the other is the quantum efficiency.
Responsivity is defined as the photo
current generated by incident photon power. Responsivity R is
given as,
-
34
And another term is quantum efficiency, which is defined as the
ratio of number of electron
hole pairs generated and number of incident photons. Quantum
efficiency is given as,
Where is number of electron-hole pair generated and is the
number of incident
photons.
Figure 3.4 responsivity v/s wavelengths
As shown in the figure, the responsivity is the function of the
wavelength and it increases as
wavelength increases. But as wavelength increases beyond a
limit, then photon energy
becomes less than the band-gap energy of the material and
responsivity reduces suddenly.
Quantum efficiency is independent of wavelength.
The best material as photo detector is silicon for wavelengths
below 1 m, because in order to
produce very large current in photo diode, there must be very
large separation between hole
and electron, and somehow silicon gives the best separation
between hole-electron [21]. So, at
wavelengths below 1 m, silicon is used. And at higher
wavelengths between 1.1 m and 1.7
m, InGaAs is used, as its responsivity is more at these
wavelengths.
3.3.3 Material selection for photo detector
In avalanche photo diode, the principle of carrier
multiplication is used
in the diode. Here, the photo carriers travel in a region, where
very high electric field is
present, so receiver sensitivity is increased [22]. The most
important 2 terms here are electron
ionization rate and hole ionization rate. Electron ionization
rate is the number of separation of
electron-hole pair by an electron, and hole ionization rate is
same by hole. Now, if there is a
significant difference between these two numbers, then
multiplication factor increases.
Multiplication M, is multiplied current and is not multiplied
current
-
35
3.3.4 PIN Photo Diode v/s APD
Figure 3.5 sensitivity v/s Photodiode areas
As shown in figure, sensitivity decreases with decrement in
photodiode area. And the graph
shows the gain of average 10 dB in sensitivity using APD over
PIN photo diode.
3.4 Noise in receiver
Noise can be considered as an unwanted component that disturbs
or reduce the content
of the signal. Consideration of noise is important because it
helps us in finding the sensitivity
of the receiver and it puts the lower limit to the performance
of the receiver set by the signal
to noise ratio.
Figure 3.6 various kinds of Noises
As shown in figure, there are mainly three types of the
noises.
Dark current noise
Quantum noise
Thermal noise
-
36
3.4.1 Dark current noise Dark current noise is present in the
receiver as the continuous current flow even when there is
no any incident light. Dark current does not depend on the
optical signal. Dark current noise is
given as,
Where q is charge on electron, B is bandwidth and is dark
current. The value of dark
current noise is very less in silicon photo diodes.
3.4.2 Quantum noise
Quantum noise is produced by the random arrival rate of photons
known as quantum nature of
photons and this noise is signal dependant noise. The noise is
in directly proportion with the
amount of light incident. Quantum noise is given as,
Where is the average current of diode because of the average
incident optical power and B
is the noise bandwidth.
3.4.3 Thermal noise
Thermal noise is produced due to spontaneous fluctuation created
by collision between free
electrons and vibrating ions in conductor. It affects more in
resistors. Thermal noise is aroused
from photo detector as well as load resistors. Thermal noise is
given as,
Where k is Boltzmann constant, R is resistor, B is bandwidth and
T is the absolute
temperature.
As shown in the equation, the light incident on the detector
must be reduced for
more reduction in induced noise. Very narrow band pass filters
are used to select the
wavelength of a laser diode and then reduce the ambient light,
which is generated by the
fluorescent, incandescent lamps and sunlight. So, using this
filters noise can be effectively
reduced [23].
3.5 Pre-amplifier
The signal is received and converted into photo current by the
photo detector, but it
suffers from attenuation and its amplitude becomes very low. So,
some kind of amplification
is required there. Pre-amplifier is of 3 main types,
Low-impedance amplifier
High- impedance amplifier
Trans- impedance amplifier
-
37
While choosing which amplifier to use, there are mainly 3
parameters to be known are noise,
bandwidth and sensitivity. And load resistance plays an
important role in setting these 3
parameters [19].
Noise is receiver is inversely proportioned to the load
resistance (RL) of the circuit. Thermal
noise N,
Bandwidth of the receiver is also inversely proportioned to the
load resistance (RL).
Bandwidth B,
Sensitivity of the receiver circuit is directly proportioned to
the load resistance (RL).
Sensitivity S,
So, we can say that to keep thermal noise low, we must keep load
resistance high. But, with
high load, bandwidth decreases. In short, there is trade-off
between bandwidth and noise,
sensitivity [20].
3.5.1 Low impedance pre-amplifier As name suggests, this
amplifier has very low impedance.
Figure 3.7 Low impedance circuits
As there is low impedance, and bandwidth is inversely
proportional to load resistance, we can
get higher bandwidth at low impedance. But, this advantage is
hindered by the noise and
sensitivity of the circuit. Because at low impedance noise is
very high and sensitivity is low,
which is not tolerable [21]. So, there is trade-off between
sensitivity and bandwidth.
3.5.2 High impedance pre-amplifier
This amplifier is with very high impedance. This amplifier has
the same circuit diagram as of
low-impedance with one change of load impedance. In this case,
because of high impedance,
there is very low noise as well as good sensitivity [21]. But
bandwidth is low. So, this pre-
amplifier is used at narrow-band, not at wide-band.
-
38
3.5.3 Trans impedance pre-amplifier
This amplifier use feedback resistor as shown in figure.
Figure 3.8 Trans Impedance circuits
Trans-impedance amplifier is mostly used where more bandwidth as
well as more sensitivity
is required [21].
3.5.4 Selection of pre-amplifier
In conclusion, it can be said that low impedance amplifier is
not much used, because it causes
high noise and low sensitivity. Then high-impedance amplifier is
used for only narrow band
application, it cannot be used at wide band. Where, the most
widely used pre-amplifier is
trans-impedance amplifier, as it provides more sensitivity at
more bandwidth [22].
3.6 Decision circuitry
Figure 3.9 Decision Circuitry
In the receiver, after photo diode and pre-amplifier, there is
binary decision circuit.
This circuit is controlled by mainly a threshold value. This
decision circuit compares the
sample value with the threshold value, and accordingly, it
decides the perfect value, which
was transmitted [18]. The comparison is triggered using a clock
signal to synchronize. In
order to improve the performance of the receiver, some other
circuits like, forward error
correction, adaptive equalizers are also used. And after signal
is detected, further signal
processing circuitry is also connected to receiver.
-
39
Chapter 4
Channel Model
4.1 Introduction to channel parameters
4.1.1 Atmospheric Turbulence The main purpose In FSO channel
model, the most important factor is environment
and major impairments due to atmospheric effects. There are many
losses like free space loss
exponent, clear air absorption, scattering, refraction and
reflections considered as atmospheric
losses. Now, the refractive index at every different point in
environment will vary because of
temperature and pressure fluctuation will be different at
different points, this will result in
atmospheric turbulence. This atmospheric turbulence is
responsible for scintillation or signal
fading, which is irradiance fluctuation in received signal. The
effect of scintillation will
degrade the performance of overall established link, which will
finally increase bit error rate
for same signal to noise ratio over the optical link.
In order to understand the overall effect on BER due to
atmospheric turbulence, it is
important to describe the power spectrum of atmospheric
turbulence in its mathematical
model. That is derived using Kalmogorov theory as,
n =0.033Cn 2
113 , where
1
L0
1
l0
Where L0 and l0 are large and small eddy size of 10-100 m and 1
cm, respectively,
is the refractive index structure parameter that gives the
spatial frequency and it depends upon
the geographical location, altitude and time of day. Values of
for different turbulence
levels like weak turbulence, moderate turbulence and strong
turbulence:
= 10
-17 m
-2/3 for weak turbulence
= 10-15
m-2/3
for moderate turbulence
= 10-13
m-2/3
for strong turbulence
Refractive index structure parameter is almost constant for
horizontal path
propagation. But in vertical path propagation, temperature
gradient is different at different
altitude and thats why refractive index structure parameter
varies with altitude. Now, when
we want to measure Refractive index structure parameter for
slant edge, we have to consider
vertical propagation and that is why it is very difficult to
measure it for slant edge. There are
some models like SLC-Day model, clear 1 model, Hufnagel Valley
Boundary (HVB) model,
PAMELA model, Greenwood model, HV-Night model and Gurvich model,
which give
refractive index structure parameter for slant edges. But as a
special case of ground to satellite
communication for uplink the data, there are large variations in
the atmospheric conditions. In
-
40
these conditions, Hufnagel Valley Boundary (HVB) model gives the
best performance. So,
model should be chosen according to application.
The mathematical model of HVB is as shown below,
Where a1= 5.94 x 10-23
, a2= 2.7 x 10-16
, s1=1000 m, s2= 1500 m, s3= 100 m and h is
altitude (m), V is the root mean square wind speed in m/s which
controls high altitude
turbulence strength at ground level. The refractive index
structure parameter versus the
altitude, h has been shown in Fig. 1 for HVB-21 model with V= 21
m/s. For different values
of (0),
decreases with increasing height and is nearly independent of
(0) when
altitude is greater than 1 km.
Figure 4.1 HVB21 Models
4.1.2 Scintillation Index An optical wave that is propagating
through the atmosphere will experience irradiance
fluctuations, or scintillation. Scintillation is caused by small
temperature variations in the
atmosphere, which results in index of refraction fluctuations.
Theoretical and experimental
studies of irradiance fluctuations generally center on the
scintillation which is defined by, S
Here I denote irradiance that is the received intensity of the
optical field after passing
it through turbulent medium. Now from this value of S,
turbulence can be identified as strong
or weak. As shown in equation, S is basically ratio of standard
deviation to mean of
irradiance. Now if S is exactly 1, that means mean is equal to
standard deviation, in this case
the effect of turbulence is so high, so fluctuations are very
frequent such that deviation is
equal to the mean value of signal, so in this case there is
strong turbulence. On the other hand,
if S is less than 0.75, in that case deviation in signal or
fluctuations in the signal is less than
-
41
its average value, so the effect of turbulence is less in this
case than before one, so here is
weak turbulence.
4.2 Various Channel models
Various channel models are proposed for different conditions of
atmospheric
turbulence like strong and weak turbulence. There are basic four
models lognormal channel
model, gamma-gamma channel model, K distribution channel model
and I-K distribution channel model. For an example, Kiasaleh has
proposed the channel model with fading of
lognormal distribution and Al-Habash has given on gamma-gamma
distribution channel
model.
The statistical channel model is given by [24],
y = sx + n = Ix + n
Where s = I denotes the instantaneous intensity gain, x {0, 1}
the OOK modulated signal, n N (0,N0/2) the white Gaussian noise
with mean 0 and variance N0/2 because of random nature of electrons
at receiver electronic circuitry, the effective photo-current
conversion ratio of the receiver and I the irradiance. Where is
defined by,
Where is the quantum efficiency of the photo receiver, e the
electron charge, the signal wavelength, Planks constant and c is
the speed of light. And definition of I will change according to
models.
4.2.1 Lognormal channel model
A. With perfect channel state information (CSI) at receiver
As it is mentioned earlier that I depends on channel models, in
lognormal channel
model, I is
Where, Z is the Gaussian distribution with Mean 0 and variance
2. So, I will follow
log-normal distribution with mean and variance
[25].
Now finding Signal to Noise ratio (SNR) from all above equation,
it should be
2*E[I]2/N0, but using somewhat different definition of SNR, we
are taking formula as,
-
42
Now calculating error probability for this model using power
series approach,
Pe,L(g,x)=
Where Pe,L is bit error rate probability which is a function of
g (signal to noise ration)
and x (fading intensity).SNR can be calculated by
.Where, R is responsivity
of receiver, P is transmitted power and 1 and 0 are standard
deviation of noise currents for symbols 1 and 0.
As the channel coefficients h at different times are independent
identical variable, than
according to moment-generating function (MFG) the variance of h
can be calculated
as,
I2
= =
Where, and are mean and standard deviation of random variable x
at transmitter
and is mean of channel due to scintillation [25].
Now if y = hx+n, and power of x is Es, then signal power =
E[hx]2 = E[h]
2*Es, but
transmitted power is Es. And channel cant add or abstract power,
so transmitted power is equal to received power. So, E[h]
2 is equal to 1. Same will be the case here
with channel fading due to scintillation taking as h. So, 1 =
1.
. Finally, I2=
.
Below is the graph of BER v/s SNR for different value of .
Figure 4.2 Performance of perfect CSI at receiver for log-normal
channel model
-
43
Using all above equations [25], we can convert the value of SNR
in terms of h,
SNR =
B. With imperfect channel state information (CSI) at
receiver
Using the last equation of SNR from above study, we can find the
BER performance
using the imperfect channel knowledge at receiver [25].
Gauss-Markov Model is described as,
h1=
So, using h1 instead of h in BER equation, we have got
comparison of with CSI and
without CSI as below,
Figure 4.3 Performance of imperfect CSI at receiver for
log-normal channel model
4.2.2 Gamma-Gamma Channel model
For weak to strong turbulence channel, the Gamma-Gamma model is
used,
which is proposed by Andrews using modified Rytov theory and
gamma-gamma
power density function (pdf) as a useful mathematical model for
atmospheric
turbulence. And this pdf of irradiance is given by,
-
44
Where Ka(.) is the modified Bessel function of second kind of
order a. and
are the effective number of small scale and large scale eddies
of the scattering
environment. Modified Rytov theory defines the optical field as
a function of
perturbations which are due to large scale and small scale
atmospheric effects [24].
Now from this result, we can find the BER performance of scheme
as,
Pb=
Where D() is given by
D()=
K-
Where c1=
c2=
In above equation, and are the effective number of small scale
and large scale
eddies of the scattering environment and can be calculated as
[24],
And finally BER v/s SNR is plotted as,
-
45
Figure 4.4 Performance of Gamma-Gamma channel model
4.2.3 Negative Exponential model In case of strong turbulences,
there are more irradiance fluctuations. Where link length
spans several kilometers, number of independent scatter become
large [27]. Signal amplitude
follows a Rayleigh fading distribution which in turn leads to a
negative exponential statistics
for the signal intensity. Signal Intensity is given as,
Where is mean radiance of channel.
4.2.4 K Channel model For strong turbulence channels, where
Scintillation Index is nearly 1, that is standard
deviation is equal to average value of the signal and the value
of log intensity variance is
between 3 and 4, the intensity statistics are given by the K
distribution. The K turbulence
model can be considered as a combination of 2 different models
exponential distribution and
gamma distribution. We got excellent similarity between
theoretical and experimental values
using this model [30]. The K distribution channel model can be
derived from a modulation
process wherein the conditional probability density function of
irradiance, is governed by
the negative exponential distribution,
Here, is mean irradiance and it follows the gamma
distribution.
-
46
Where, is gamma function defined as,
, and is a channel
parameter related to effective number of discrete scatters. The
unconditional distribution for
irradiance is given as,
This integration results as,
=
Using a simple transformation, the pdf of instantaneous SNR can
be given as,
=
Where, K() is the modified Bessel function of the second kind of
order . is average
electrical SNR at the receiver. Which is given by = . As the
Bessel function is
denoted by K here, this channel model is known as K channel
model.
The BER v/s SNR plot is given as below,
Figure 4.5 Performance of K channel model
The limitation of the K channel model is that it lacks the
numerical computation in
much closed form, thats why I-K channel model is proposed
[29].
-
47
4.2.5 I-K Channel model
This channel model is working in both scenarios- weak turbulence
and strong turbulence.
Moreover it has less computation complexity than gamma-gamma
channel model. So, this
channel model is preferred over others [28].
The pdf of normalized signal irradiance is given as,
=
Where, is modified Bessel function of first kind of order , K()
is the modified
Bessel function of the second kind of order , and are channel
parameters related to
effective number of discrete scatters and coherence parameters,
respectively [27].
Again using a simple transformation, SNR is obtained as,
=
Now from the equation of channel capacity, , we have pdf of
capacity, C as following,
-
48
=
Now, outage probability, r is defined as,
.
So, pdf of outage can be written as,
=
The following shows the result of BER v/s SNR for I-K channel
model,
Figure 4.6 Performance of I-K channel model
-
49
4.3 Comparison of Channel models
Lognormal channel model is used in weak turbulence scenario and
key factor is Sigma(x).
Gamma-Gamma channel model is used in weak to strong turbulence
scenario and key
factor is Alpha and Beta.
K channel model is used in strong turbulence scenario and key
factor is Beta.
I-K channel model is used in strong turbulence scenario and key
factor is Raw.
-
50
Conclusion
In nearby future, FSO will become important and necessary medium
of information exchange
due to its advantages over fiber optics communication. Proper
low cost design of transmitters is a
viable and better option to prevent trenching and sunken cost of
fiber optics. For this project in
particular, the FSO transceiver was designed using red laser
diode and tested for 1 kbps data transfer
in laboratory conditions. The range extension can be done by the
use of higher power infrared laser
diodes. All the theoretical aspects for transmitter, receiver as
well as modulation techniques to be
used were studied and design issues arising were discussed. The
channel models for the free space
optic link were studied in detail and imperfect CSI model added.
The simulations for all present day
models was carried out using Matlab and the results presented.
Thus, a low cost prototype for free
space optical communication was designed.
-
51
References
1. Free Space Optics For Laser Communication Through the Air,
Dennis Killinger, Optics & Photonics News ,
October 2002, Optical Society of America
2. Fiber Optics Without Fiber, IEEE Spectrum, August,2001
3. Mobile computing and wireless communications, by Amjad
Umar
4. V. Ramasarma, Free Space Optics: A Viable Last-Mile Solution,
Bechtel Telecommunications Technical
Journal, pp. 22-30, December 2002
5. Free-Space Optics for Fixed Wire