Free Space Optic (FSO) Link Design

A Minor 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]

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-spaceoptical 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

II

Index
Chapter No. Acknowledgement Abstract Index List of Figures List of Tables Nomenclature 1 Introduction 1.1 Definition 1.2 1.3 1.4 1.5 1.6 1.7 1.8 2 Factors behind market growth A Case Study Advantages of FSO Limitations of FSO Applications of FSO A typical system model Objectives Of This Project Title Page No. I II III VI VIII VIII 1 1 3 4 5 5 6 6 8 9 9 10 11 11 11 12 12 13 13 16
III

Transmitter for Free Space Optical Communication 2.1 Block diagram and practical circuit layout of FSO transmitter 2.2 Qualities of the optical source 2.2.1 2.2.2 2.2.3 2.3 LED v/s LASER A novel development in light sourcesVCSEL Frequency(wavelength) of operation in Optical Wireless

Modulation Schemes Communications 2.3.1 On-Off Keying 2.3.2 2.3.3 2.3.4 2.3.5 2.3.6 RZ OOK

Manchester Encoded Signal Pulse Position Modulation Comparison of Modulation Schemes Conclusion

2.4

All about LASERs 2.4.1 2.4.2 2.4.3 2.4.4 2.4.5 Unique characteristics Types of LASERs available Working of LASER Diode Classes of LASERs based on eye safety and power Selection of LASER for FSO applications LED Operation and Characteristics Types of LEDs and lifetimes LED Driver LASER Driver Circuit

16 16 16 17 17 17 18 18 19 19 19 21 22 22 24 25 27 28 31 31 31 32 32 33 33 34 35 35 36 36 36 36 37
IV

2.5

All about LEDs 2.5.1 2.5.2

2.6

Driver Circuits 2.6.1 2.6.2

2.7

Practical Design Steps 2.7.1 PC to Transceiver Interface 2.7.2 2.7.3 2.7.4 Using Hyper-terminal to send a file to a remote computer Selection of light source Practical Driver Design models

2.7.5 Power Calculation Receiver for Free Space Optical Communication 3.1 3.2 Block Diagram for Receiver of FSO Photo Detector 3.2.1 3.2.2 3.3 3.3.1 3.3.2 3.3.3 3.3.4 3.4 3.4.1 3.4.2 3.4.3 3.5 3.5.1 Requirements of photo diode Working principle PIN photo diode Material selection for photo detector Avalanche photo diode (APD) PIN Photo Diode v/s APD Dark current noise Quantum noise Thermal noise Low impedance pre-amplifier

Different types of photo detector

Noise in receiver

Pre-amplifier

3.5.2 3.5.3 3.5.4 3.6 4 4.1

High impedance pre-amplifier Trans-impedance pre-amplifier Selection of pre-amplifier

37 38 38 38 39 39 40 41 41 43 45 45 47 49 50 50 51 53 53 54

Decision Circuitry Introduction to channel parameters 4.1.1 Atmospheric Turbulence 4.1.2 Scintillation Index Various Channel models 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 Lognormal channel model with and without perfect CSI Gamma-Gamma Channel model Negative Exponential Model K channel model I-K Channel model

Channel Models

4.2

4.3 5 5.1 5.2 5.3 5.4 5.5

Comparison of Channel Models Objectives of the Project Design specifications Circuit description Scope Result of the project

FSO Link Design

Conclusion

References

55

LIST OF FIGURES
Fig. No. 1.1 1.2 1.3 2.1 2.2 2.3 2.4 2.5 Title 10-Gbps FSO link, deployed by MRV Communications' Tele-Scope Last-Mile Connectivity Terabeam Transceiver Block diagram of FSO Transmitter Practical form of transmitter Light power v/s current for LED and LASER Small-signal frequency responses of an LED and an LD with negligible parasitic effects BER performance for OOK (NRZ and RZ), from Eq. (2.38), and L-PPM (L = 2, 4, and 8) Page No. 1 3 6 8 8 9 10 14

2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14

Power spectrum of the transmitted signals for OOK (NRZ and RZ), and L-PPM (L 15 = 2, 4, 8) Example of (a) LED drivers, (b) shunt driver Output power v/s current for LASER diode LASER Driver Circuits Ports DB-9 AND MAX232 RS232 to TTL interface HyperTerminal screen variety LEDs available a typical LASER diode 20 21 22 23 24 25 25 26

VI

2.15 2.16 2.17 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.1 4.2 4.3 4.4 4.5 4.6 5.1 5.2

Collection of and gates as LASER Driver LASER Driver using op-amp Overall attenuation v/s distance plot for different wavelengths Block diagram of Simple Receiver V-I characteristic of photo diode energy band diagram of PIN photo diode responsivity v/s wavelengths sensitivity v/s Photodiode areas various kinds of Noises Low impedance circuits Trans Impedance circuits Decision Circuitry HVB21 Models Performance of perfect CSI at receiver for log-normal channel model Performance of imperfect CSI at receiver for log-normal channel model Performance of Gamma-Gamma channel model Performance of K channel model Performance of I-K channel model Block Diagram of Transceiver Circuit Photograph of circuit board

27 28 29 31 32 33 34 35 35 37 38 38 40 42 43 45 46 48 51 52

VII

LIST OF TABLES
1.1 2.1 2.2 Comparison of FSO with other technologies in terms of cost Comparison of different baseband intensity modulation techniques. Relationship among Material, System Wavelength, and Band Gap Energy for LED Structures 2.3 5.1 5.2 Power Calculation project design specifications electrical characteristic of BPW-34 30 50 51 2 15 19

NOMENCLATURE
FSO LOS R.I. S.I. CSI BER SNR IM/DD OOK HVB Free Space Optics Line Of Sight Refractive Index Scintillation Index Channel State Information Bit Error Rate Signal to Noise Ratio Intensity Modulation/Direct Detection On Off Keying Hufnagel Valley Boundary model

VIII

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

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].
2

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 fibreoptics 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
3

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.
4

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 fibrebased 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).

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.
6

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.

Chapter 2

Transmitter for Free Space Optical Communication

2.1 Block diagram and practical circuit layout of FSO transmitter

Source

Modulator

Driver Circuit

Light Sorce

Beam Concentrators

Cooling Mechanism

Figure 2.1 Block diagra m of FSO Trans mitter

Figure 2.2 Practical for m of trans mitter 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.
8

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 maxi mum 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. 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.

(ii)

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.

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].

10

2.2.3 A novel development in light sources-VCSEL

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. T he 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
11

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 <1. The benefit obtained from this scheme is a reduction in transmitted power. However, as d decreases, the bandwidth requirement grows faster than the decrease in power requirement. Thus, this type of OOK is inferior to PPM, which offers less bandwidth to achieve a given reduction in power. For d = 0.5, this scheme is commonly called return-to-zero (RZ) OOK. In RZ-OOK, it is intuitive to show that the power requirement is reduced to half of the regular non-return-to-zero (NRZ) OOK discussed earlier, with the expense of doubling the bandwidth.

2.3.3 Manchester Encoded Signal

The basic disadvantage of OOK signalling is that key receiver parameter values, such as power levels, must be known to optimally set the threshold. A pulse format that avoids this difficulty uses pulse-to-pulse comparison for decoding. One way to do this in binary encoding is called Manchester coding, where a one is signified when the optical signal is on during the first half of the symbol interval, and a zero is signified when the optical signal is on during the second half of the symbol interval. The transmit pulse shape pm for m = {0, 1} can be written as

The demodulator separately integrates the detector output over the two half bit intervals and compares them for bit decoding. The bit is decoded according to which integration produces the higher value, and no threshold need be selected. The system still uses pulse signalling, but
12

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/T f . 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, log 2L 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

13

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)
14

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.

15

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.

16

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.
17

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 limi ted 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.

18

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 promine nt 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
19

because it takes far too long to turn on a diode once it has been completely turned off. 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
20

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 Vb ias and modulation current is provided by source resistor, R mod, respectively.
21

This type of single-ended laser driver is typically used with low operating speed due to the unwanted parasitic inductance from the packages bondi ng 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, I bias controls the DC threshold current and I mod 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 inter facing with RS232. It is available at cost of approximately Rs 30 to 40 (at Robokits).

22

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

23

Figure 2.11 RS232 to TTL interface Complete explanation of this http://www.botkin.org/dale/rs232_interface.htm can be viewed at:

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) 2) 3) 4) Open HyperTerminal. Open a saved connection file or create a new connection. Connect to the remote computer. 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.

24

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

25

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) 2) 3) 4) 5) 6) Hamamatsu Dense light Semiconductors PVT. LTD. Cree LED lights Ray science Innovation Ltd. Ad labs Pvt. LTD. 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

26

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

27

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.

28

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
29

5. LINK POWER BUDGET For distance limit calculations it is necessary to calculate the mini mal 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.

30

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 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.
31

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. 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.
32

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, 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, =
33

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 =

3.3.4 PIN Photo Diode v/s APD

34

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

3.4.1 Dark current noise

35

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, 2 = 2 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,
2 = 2 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,
2 = 4

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
36

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 (R L) of the circuit. Thermal noise N,

Bandwidth of the receiver is also inversely proportioned to the load resistance (R L). Bandwidth B,

Sensitivity of the receiver circuit is directly proportioned to the load resistance (R L). 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 preamplifier is used at narrow-band, not at wide-band.
37

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.
38

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, 1 1 L0 l0 2 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 2 the geographical location, altitude and time of day. Values of for different turbulence levels like weak turbulence, moderate turbulence and strong turbulence: n =0.033C2 - 3 , n
11

where

2 = 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 w ith 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
39

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,
2 = 1

27

10

exp

2 + 2 exp + 0 exp 1 2 3

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 2 2 2 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
40

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 = 2 Where, Z is the Gaussian distribution with Mean 0 and variance 2. So, I will follow 2 2 2 log-normal distribution with mean 2 and variance 4 ( 4 1) [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, = 2 E 2 E 2 N0
41

Now calculating error probability for this model using power series approach, Pe,L(g, x)= 2
1 1

(1) g (2k+1) 2 4+1 x 2 ( ) =0 2 2 +1 2 2+1 ! exp 2

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
4 2 2 (1+0) 2

.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, =
2 +2 2 4 +4 2

I2 = 2 []2 =

( 4 1)

= 1 1 2 Where, and are mean and standard deviation of random variable x at transmitter and 1 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. = 2 . Finally, I 2=( 4 1). Below is the graph of BER v/s SNR for different value of .
2

Figure 4.2 Performance of perfect CSI at receiver for log-normal channel model
42

Using all above equations [25], we can convert the value of SNR in terms of h, SNR = (1+0) 2
4 2 2 2

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= + 1 2 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, 2() (+ )/2 (+ ) 1 = 2 2 ; ()() > 0

43

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()=2
+4 4

2 6 () 0 (122 )2

K-(2

Where c1=

(
+1 ) 2

c2=( +
2

1 16

( )2 )
2

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], 0.49 2 = exp (1 + 0.18 2 + 0.56 12/5 )7/6 0.51 2 (1 + 0.69 12/5 )5/6 = exp (1 + 0.9 2 + 0.62 2 12/5 )5/6
1

1
1

And finally BER v/s SNR is plotted as,

44

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, 1 1 exp ; 0 0 Where 0 is mean radiance of channel. () = > 0

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,

1 1 exp ;

> 0

Here, is mean irradiance and it follows the gamma distribution.

45

1 exp ; ()

> 0
1 , 0

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, = 2

1 2

+1 +1 2 +1 ( ) 4

Using a simple transformation, the pdf of instantaneous SNR can be given as,
2 4
+1 3 +1 4

= ()

1 2

Where, K() is the modified Bessel function of the second kind of order . is average electrical SNR at the receiver. Which is given by =([])2 0 . 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].
46

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, 2 1 + =
1+
1 2

1 2
1+

1 2 1 + ; < 2 1 +
1+
1 2

1 2
1+

1 2 1 + ; >

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, 2 1 + 1 = 2 1 + 1
1+
1 3 2 4 +1 4

1 2 ; < (1+) 2 1 2 ; >

2 (1+) 2 2

2 1 +
1+

1 3 2 4 +1 4

2 1 +

Now from the equation of channel capacity, = log 2 (1 + ), we have pdf of capacity, C as following,

47

2 +1 (2) 1

1 +

1+

1 3 1) 4 2 (2 +1 4

1 2
1+ 2

1 =

2 1 + 1 +

2 1

< log 2 (1+ ) 2 1 2

1+ 2

2 +1 (2) 1

1+

1 3 1) 4 2 (2 +1 4

2 1 +

2 1 0

> 2 (1+ )2

Now, outage probability, r is defined as, = So, pdf of outage can be written as, 2 = 1 2 (1 + ) 2 1 +
1+

1+ 2 (2 1)4 2 1

1 2 ; < log 2 (1+) 2 1 2

1+ 2 1+ 2

2 1 +

1 1)4 2 (2

2 1

; > 2 (1+ ) 2

The following shows the result of BER v/s SNR for I-K channel model,

Figure 4.6 Performance of I-K channel model

48

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.

49

Chapter 5

FSO link design

5.1 Objectives of this Project
Design a moderate speed FSO data link with transfer rates up to 100 Kbps. Operating distances 200 to 300mts. 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. Design using readily available, cheap and indigenous components instead of expensive, specialized components. Excellent up-time and good reliability. Compact & Easy to install Hardware. The hardware was intended to be transparent to existing software platforms. Very less setup times. Provide an excellent platform for design and testing of more advanced FSO projects and communication protocols.

5.2 Design Specifications

The following table shows the specifications like maximum range, data rate, mode of communication, the interface with computer, and specification of lasers like wavelength, laser power, class of laser and power requirements. Maximum Range Data rate Mode Computer interface Laser beam Modulation Laser wavelength Laser power Laser Class Power requirements 10 centimeters (due to visible light laser used) 100 Kbps or higher Full Duplex COM Port RS232 (Serial port) On-Off Keying 670 nm < 5mW peak pulse power < 1mW average power Class II 9 to 12 V DC @ 60 mA Table 5.1 project design specifications
50

The following table shows the electrical characteristic of BPW-34 Min. Typ. Max. Reverse Breakdown Voltage V(BR) 60 Reverse Dark Current Iro 2 30 Rev. Light Current Ira 75 65 Half Angle Peak Wavelength 900 600-1050 Spectral Bandwidth Rise & Fall Time 100/100 Table 5.2 electrical characteristic of BPW-34 Unit V nA uA deg nm nm uS

5.3 Circuit description

Figure 5.1 Block Diagram of Transceiver Circuit Almost all digital devices which we are using today require either TTL or CMOS logic levels. Therefore the first step to connecting a device to the RS-232 port is to transform the RS-232 levels back into 0 and 5 Volts by RS-232 Level Converters. Two common RS-232 Level Converters are the 1488 RS-232 Driver and the 1489 RS232 Receiver and each package contains 4 inverters of the one type, either Drivers or Receivers. The driver requires two supply voltages, +7.5 to +15v and -7.5 to -15v. The transceiver is based on the MAX232A IC for transmitting and receiving RS-232 compatible voltage signals. The MAX232A generates +10V and -10V voltage swings using a dual charge-pump voltage converter from a single +5VDC voltage. This IC includes two receivers and two transmitters in the same package in order to make it full duplex. The
51

MAX232A version requires only 0.1uF capacitors for the charge-pump and inverter, whereas the MAX232 requires 1uF capacitors for the same. So, the advantage of the A version is that it has faster response times, and allows for faster data rates. The laser diode driver consists of a 7405 open-collector hex inverter IC to convert it to 5V DC and all the outputs of the inverters are coupled together to provide enough drive current for the laser diode which draws around 35mA @ 3V. The two 1N4001 diodes, in series with the laser diode, step down the voltage from +5VDC to around 3.6VDC which is close to the nominal voltage for the laser diode to drive it. The receiving sensor is a PIN Infrared photo-diode numbered BPW-34, which has a very wide spectral bandwidth from 600nm to 1050 nm. The signal from the photo-diode is amplified by the Non-inverting OPAMP amplifier 741. The gain of the amplifier is set to 10 by fixing the resistance ratio of 10:1. This amplifier serves 2 purposes: One it acts as a high impedance buffer to the output of the photodiode as this is needed since the photodiode cannot source more than a few microamperes of current and secondly the range of voltage variation between the light and dark condition at the output of the photo-diode was very small and had to be amplified to digital levels. The output of the OPAMP is then buffered via a pair of Schmitt trigger buffers to clean up and square the signal or sharp that signal. The output of the second buffer is then directly converted to a RS-232 standard signal using the MAX232A. The transceiver is designed in such a way that when no signal is present, at that time the laser is on. This helps to see where the laser is pointing during the laser-detector alignment or not. The transceiver is powered by a 9V DC battery and draws 80mA (laser on) and 40mA (laser off).

Figure 5.2 photograph of circuit board

52

5.4 Scope
Due to financial constraints we have implemented this technology at an elementary level. Availability of more funds will help us implement the circuit for higher speeds of the order higher Kbps to Mbps and also widen its communication range from meters to kilometers. The system implemented is basically a one to one bi-directional communication system. It can be modified into one to many multichannel system. One of the limitations of the above system is that it is not able to operate with efficiency in presence of fog, heavy rain or snow. This limitation can be overcome by implementing the several techniques mentioned previously for dealing with environmental factors. When line of sight communication is not possible then the system can be implemented by the use of reflection and deviation mirrors.

5.5 Result of the project

A circuit of laser transceiver has been successfully implemented. System has been tested to transfer rates up to 100 Kbps. The laser link works up to a distance of 10 centimeters (As laser costs near Rs. 3000, we have used a low cost laser for just testing purpose of the circuit. So, we are getting very less distance.) Character transfer has been successfully run using HyperTerminal.

53

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 were carried out using Matlab and the results presented. Thus, a low cost prototype for free space optical communication was designed.

54

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 Wireless Broadband (thesis) , Pramadityo Tjondronegoro, 10.12.2004 Munich University of Technology 6. Optical Wireless communications, IR for Wireless Connectivity,Roberto Ramirez-Iniguez, Sevia M. Idrus and Ziran Sun 7. D.C. OBrien, G. Faulkner, K. Jim, E. Zyambo, D.J. Edwards, M. Whitehead, P.N.Stavrinou, G. Parry, V.M. Joyner, R.J. Samsudin, D.M. Holburn, and R.J. Mears, High-Speed Integrated Transceivers for Optical Wireless, IEEE Communications Magazine, 41(3), 5862, 2003. 8. J.M. Senior, Optical Fiber Communications: Principles and Practice, 2nd ed. Prentice Hall International, 1992. 9. X. Zhu and J. Kahn, Free-space optical communication through atmospheric turbulence channels [A], IEEE Trans. on Communications, no. 2, pp. 12931300,2003. 10. Sabi S. and Vijayakumar N. Simulation of a Modem using Digital Pulse Interval Modulation for Wireless Optical Links. 10th National Conference on Technological Trends (NCTT09) 6 -7 Nov 2009. 11. M.Ijaz, O. Adebanjo, S. Ansari, Z. Ghassemlooy, S. Rajbhandari, H. Le Minh, A. Gholami and E. Leitgeb. Experimental Investigation of the Performance of OOK-NRZ and RZ Modulation Techniques under Controlled Turbulence Channel in FSO Systems. IEEE Trans. 2010. 12. THESIS, FREE SPACE OPTICS COMMUNICATION FOR MOBILE MILITARY PLATFORMS by Soo Sim Daniel Neo,December 2003. 13. C. Sinah, J. John, Y.N. Singh, and K.K. Tripathi, Design Aspects of High-Performance Indoor Optical Wireless Transceivers, presented at IEEE ICPWC2005, New Delhi, 2005, pp. 1418. 14. High-speed 1.3-m LED transmitter using GaAs driver IC , Suzuki, T. Sumitomo Electric Industries, Ltd., Shimaya, Konohana-ku, Osaka, Japan Ebata, T. ; Fukuda, K. ; Hirakata, N. ; Yoshida, K. ; Hayashi, S. ; Takada, H. ; Sugawa, T. 55

15. NEP-Menlo-E-SPC-0002 Issue 01 16. ANDREWS, L. C, PHILIPS. R. L., HOPEN, C. Y.: Laser beam scintillation with application, Washington: SPIE PRESS, 2001 17. KIVK, P. and WILFERT, O.: Power equivalent Gaussian beam for terrestrial optical wirele ss link power budget. In Proceedings of the Junior Scientist conference 2006. Vienna: Vienna University of Technology, 2006. 18. Nur Haedzerin Md. Noor, A.W. Naji, Wajdi Al-Khateeb, Theoretical Analysis of Multiple Transmitters/Receivers on the Performance of Free Space Optics (FSO) Link, 2011 IEEE International Conference on Space Science and Communication, 12-13 July 2011, Penang, Malaysia. 19. J. A. R. Pacheco de Carvalhol, H. Veigal, P. A. J. Gomes, Claudia F. F. P. Ribeiro Pacheco, A. D. Reis , Experimental Performance Study of a Very High Speed Free Space Optics Case Study, 2008 IEEE. 20. Jaime A. Anguita, Jaime E. Cisternas, Experimental evaluation of transmitter and receiver diversity in a terrestrial FSO link, IEEE Globecom 201 0 Workshop on Optical Wireless Communications. 21. Nur Haedzerin Md Noor, Wajdi Al-Khateeb, A.W. Naji, Experimental Evaluation of Multiple Link at the University of Beira Interior Campus: a

Transmitters/Receivers on Free Space Optics Link, IEEE Globecom 2010 Workshop on Optical Wireless Communications. 22. Shen Changyu, Feng Huajun, Xu Zhihai, Jin Shangzhongt, High Speed Free Space Optical Communication System for l km communication 23. Frank S. Marzano, Saverio Mori, Fabrizio Frezza, Paolo Nocito, Giorgio M. Tosi Beleffi, Gabriele Incerti, ElioRestuccia, Fernando Consalvi, EuCAP 2011 - Convened Papers. 24. Murat Uysal , Jing (Tiffany) Li, Error Rate Performance of Coded Free-Space Optical Links over Gamma-Gamma Turbulence Channels, Communications, 2004 IEEE International Conference, 20-24 June 2004, 3331 - 3335 Vol.6. 25. Hassan Moradi, Maryam Falahpour, Hazem H. Refai, Peter G. LoPresti, Mohammed Atiquzzaman, BER Analysis of Optical Wireless Signals through Lognormal Fading Channels with Perfect CSI, Telecommunications (ICT), 2010 IEEE 17th International Conference Publications, 493 497. 26. Hector E. NISTAZAKIS, Andreas D. TSIGOPOULOS, Michalis P. HANIAS, Christos. D. PSYCHOGIOS, Dimitris MARINOS, Costas AIDINIS, George S. TOMBRAS, Estimation of Outage Capacity for Free Space Optical Links over I-K and K Turbulent Channels, RADIOENGINEERING, VOL. 20, NO. 2, JUNE 2011. 27. LETZEPIS, N., FABREGAS, A. G. Outage probability of the free space optical channel with doubly stochastic scintillation. IEEE Transactions on Communications, 2009, vol. 57, no. 10. 56

28. ANDREWS, L. C., PHILIPS, R. L. Mathematical genesis of the IK distribution for random optical fields. Journal of the Optical Society of America A, 1986, vol. 3, no. 11, p. 1912 - 1919. 29. ANDREWS, L. C., PHILIPS, R. L. I-K distribution as a universal propagation model of laser beams in atmospheric turbulence. Journal of the Optical Society of America A, 1985, vol. 2, no. 2, p. 160 163. 30. ANDREWS, L. C., PHILIPS, R. L. Mathematical genesis of the IK distribution for random optical fields. Journal of the Optical Society of America A, 1986, vol. 3, no. 11, p. 1912 - 1919.

57