This chapter delves into detail about the various components used in the construction of the optical wireless transmitter and the principles and theories that make them work. The major component of the transmitter design was the optical front end, of which the optical source was the most important part. Optical sources include LASERS and LEDs and each of these sources are categorized under different types and sub classes. Due to its ready availability, cheap cost and taking into account the user's safety, an LED was preferred over a LASER source. The working principle, different types and structures are explained in this chapter.
Another important consideration of the transmitter design was the information source. The information which was to be sent was digital information in the TTL logic format. The information source used for this purpose was a laptop. The source was connected via a USB-TTL converter cable to the transmitter section. Testing of the transmitter was done using a signal generator.
The choices of the optical source as well as the transmitter design are included in this chapter.
4.2 Optical Source
4.2.1 Light Emitting Diodes (L.E.D)
LEDs are solid-state light sources that have a spectral range of emission from ultraviolet to the near infra-red. The wavelengths range from 370 nm to 16500 nm. As stated in (Jia-Min) Lui's book (1), commercial LEDs are made of III-IV compound semiconductors. The basic principle of operation of an LED is based on spontaneous emission of photos. L.E.Ds, unlike LASERS, emits incoherent spontaneous light. LEDs also do not require a threshold and emits light as soon as it is forward biased (1).
The figure 4.1 shows the common structure of a commercially available LED. An easy way to distinguish between the cathode and the anode is to check the length of the legs. The anode leg of the LED is longer than the cathode leg.
Fig 4.1: Structure of an LED
Source:http://upload.wikimedia.org/wikipedia/commons/thumb/f/f9/LED%2C_5mm%2C_green_%28en%29.svg/500px-LED%2C_5mm%2C_green_%28en%29.svg.png (20)
There are few advantages of using LEDS over LASERS. As T.D.C.Little.et.al (2) points out in their paper that, although lasers can allow for greater distance separation between transmitter and receiver and also greater data rates, lasers are harmful to the human eye and cannot be used alternatively as an illumination source. It is for these reasons that LEDs were opted for as an optical source for the purpose of indoor communication. J.M.Senior (3) also points out that, LEDs are cheaper and simpler to fabricate, are more reliable, require a simpler drive circuitry and most importantly have liner characteristics. The last advantage means that the LED has a linear light output against current characteristic as opposed to an injection LASER.
Fig 4.2: Light power v/s Current characteristics of an LED at different temperatures
Source: "Optical fiber communications: principles and practice" by John.M. Senior, Pg 424 (3)
4.2.2 Operating Principle
An LED operates on the principle of spontaneous emission. Spontaneous emission is defined as the process by which an atom in the excited state transitions to a lower energy state by giving off a photon. The photon is emitted and has energy defined by the equation below (3):
…………………….…… (4.1)
Where is the energy corresponding to the higher energy state, is the energy corresponding to the lower energy state, h is the Planck's constant and f is the frequency of the emitted photon.
An LED is basically a p-n junction diode that operates in the forward bias condition. When the p-n junction is forward biased current starts to flow from the anode to the cathode. As a result of the biasing the holes from the p-side and the electrons from the n-side now move towards the junction. When the electron-hole pairs recombine they move to a lower energy level and releases energy in the form of a photon. The wavelength of light emitted depends on the band gap energy of the p-n junction. In case of spontaneous emission atoms return to the lower energy level in a random manner (3).
Fig 4.3: Internal structure and working of an LED
Source: http://upload.wikimedia.org/wikipedia/commons/d/d7/PnJunction-LED-E.svg (21)
4.2.3 Types and Structures of Optical sources
Generally the optical sources used are either LEDs or LASER diodes. A few commonly used structures of light emitting diodes and laser diodes are explained in the following section.
Hetero-junction LEDs
Hetero-junction LEDs are of two types: single or double hetero-junction LEDs. Due to the poor optical confinement abilities of the single hetero-junction LEDs the double structure is adopted (1). The semiconductor structure of a typical LED is the Double Hetero-junction LED structure. The figure below shows the Burrus-type double hetero-structure surface-emitting LED:
Fig 4.4: Burrus type double Hetero-junction LED
Source: "Optical semiconductor devices" by Mitsuo Fukuda, Pg 94 (v)
According to John Gowar's (4), during the fabrication process of an LED, in case a visible light source is needed GaAs is doped with either N or ZnO. The different wavelengths of LEDs fabricated are show in the table below taken from Mitsuo Fukuda's book.IR LEDs are fabricated by using GaAs or AlGaAs.
Table 1: Different Active layer materials for different wavelengths
Source: "Optical semiconductor devices" by Mitsuo Fukuda, Pg 94 (v)
Edge Emitter LEDs
This is another high radiance structure of an LED which is used in optical communications. It has a very thin active layer placed in between two carrier confinement layers (3). There are also transparent guiding layer present between the two carrier confining regions, this helps reduce the amount of light that gets self-absorbed in the active region.
Fig 4.5: Edge Emitting LED
Source: "Optical fiber communications: principles and practice" by John.M. Senior, Pg 411 (3)
Semiconductor Laser Diodes
Based on M.J.N Sibley's book (5), LASER diodes work on the principle of stimulated emission. Gas LASERS also work on the similar principle. For stimulated emission to occur population inversion should exists (5).In the process of stimulated emission the incident light photon interacts with an already excited atom thereby bringing the atom from conduction band to valance band with the release of an addition photon with the energy similar to the band gap energy. The structure of a semiconductor laser is shown below:
Fig 4.6: Structure of a semiconductor laser diode
Source: http://www.explainthatstuff.com/semiconductorlaserdiodes.html (26)
4.2.4 LED selection
As explained at the start of the chapter, LEDs were preferred over LASER diodes. The choice of selection of the LED depended largely on 3 factors namely functionality, cost and safety. Initially the optical source to be used was an IR LED. And while on the one hand IR LEDS have very large modulation bandwidths making it possible to send information at high data rates. The main reason for the rejection of the IR LED as an optical source was due to safety concerns. On the basis of functionality and safety white LED was selected. This was because white LEDs could not only be used as a source of information but also as a form of illumination. Another reason for the selection of white LEDs was because white LEDs, which do not come under the high brightness category, are safe for viewing with the human eye.
White LEDs that are commercially used are classified into two types. This classification was based on how the LED generates "white light". The first type involves combining red, green and blue emitters in equal measure to form white. The second type involves a blue emitter and a yellow phosphor. The experiment uses the latter type as the optical source. In experiments conducted by H.L.Minh.et.al (6) with white LEDs it was found that white LEDs had a much lower modulation bandwidth as compared to blue LEDs. The modulation bandwidth of a white LED is in the range 2-3 MHz without any kind of post or pre equalisation circuitry being implemented (6). This bandwidth limitation was due to the slow temporal response of the phosphor present in the LED (6). The modulation bandwidth could be increased significantly by either applying equalization techniques, using multiple input sources or by using a filter at the receiver section (6). To keep the circuitry simple the first two options of using MIMO systems and using post or pre-equalization techniques to increase modulation bandwidth were not employed.
To increase the modulation bandwidth at the receiver section a blue filter was used. Although according to Lubin Zeng.et.al (7), the inclusion of a blue filter at the receiver section would block significant amounts of energy emitted by the LED. This implies that it was necessary to make sure that the LED was bright enough so as not to lose significant amount energy after light passed through the blue filter but not too bright to cause any kind of harm to the eyes of the viewer.
Based on the three factors mentioned above the LED selected was an InGaN high brightness white light emission LED from the AND series of LEDs. These were manufactured by Purdy Electronics. The LED used was the AND520HW, which was a 5mm high brightness and wide viewing angle LED.
LED Driver
After the selection of the light source, the next step was to determine how the light source was to be modulated based on the digital input being fed to it. LEDs are current controlled devices. The output characteristic of LED with respect to current varies linearly. According to Heather Brundage (8), a number of constant current LED drivers are available commercially but they are designed for low speed applications such as LED dimming. Because of this such drivers cannot be used for high speed optical communications.
However the most common method for controlling LEDs is by using a constant voltage and by limiting the current passing through the LEDs using resistors. In such cases transistors such as BJTs and FETs can be used to turn the LED either ON or OFF. For the experiment a high speed MOSFET driver IC was used. This was the TC4426A IC. Since the LED which was used was a low current, low voltage device, the push pull MOSFET set up present in the output section of the TC4426A IC was sufficient enough to drive the LED. As a result of this an additional MOSFET or BJT was not required to drive the LED. The functional block diagram of the TC4426A with push-pull setup is shown below:
Fig 4.7: Internal structure of TC4426A IC
Source: http://nd.edu/~lemmon/courses/ee40442/labs/docs/TC4428A.pdf (9)
The output current from the IC acts as a current source for the LED. A small portion of the output current was sunk through the load, which is a resistor connected to the ground. The IC is well suited for high speed applications as rise and fall times are both 35-40 nano-seconds (9). The IC was operated with a supply of 10V DC.
Input Data Signal
The scope of the project was to transmit any kind of information from the transmitter to the receiver using optical wireless principles. The information that was transmitted could be either in digital or analogue format. Sending of information in analog format would require additional components including an A/D converter as well as further circuitry for the implementation of various modulation schemes. As a result the receiver section would also consist of a demodulator and a D/A converter. This would complicate the overall system. As a result the experiment that was carried out used TTL digital data as input information.
The information to be transmitted was supplied by a laptop. The information from the laptop however, had to be converted into TTL logic. In TTL logic, logic "0" is represented by 0V and logic "1" is represented by 5V. One of the main concerns was to obtain the output from the laptop into TTL format.
The method of sending information to the transmitter section was to use one of the USB ports available on the laptop. Since the ports on the laptop were USB2.0, the data transmission rates, according to Jang-Jin Nam.et.al (10) were in excess of 480Mbps. Since such high data transmission rates made it problematic for transmission using a single LED source, the data rate was limited to 1Mbps in order for successful transmission. The relationship between modulation bandwidth and memory-less channels, as mentioned in Tommy Ó¦berg's (11) book, can be explained by using the Shannon-Hartley formula. This is given as:
Where C is defined as the symbol rate, B is the bandwidth and M is the number of levels for the particular coding used. TTL logic represents a 2 level logic level or a unipolar non return to zero form of coding. As explained in Bernard Skylar's (12) book, the waveform levels are either +V or 0. Unipolar non return to zero is shown in the fig 3.4.1:
And so by replacing M=2 in the above equation. This makes the signal rate defined as twice the modulation bandwidth.
Fig 4.8: TTL logic levels
Source: http://www.rigacci.org/docs/biblio/online/intro_to_networking/c2161.htm (25)
Other concerns such as the modulation schemes and setting up of any form of error correction was not required as the USB uses its own protocol to transmit data with specific synchronization sequences, packet sizes and end of packet indicators (8). The protocols for transmission of data are explained in the next section.
4.4.1 USB Protocols
FTDI chips' technical paper on USB (13) states that USB transmits data in packets which are sent LSB (Least Significant Bit) first. The main packet types are: Token, Data, Handshake and End of frame (13). Each of these packets is comprised up of different field types such as SYNC, PID, Address, Data, Endpoint, CRC and EOP (13).
Unlike TTL logic, USB follows NRZI coding (13). NRZI is shown in the fig below:
Fig 4.9: NRZI coding technique
Source: http://www.fiberoptics4sale.com/wordpress/line-coding-in-digital-communication/ (24)
USB ports use differential voltages to set the logic levels of "0" and "1" (8) (13). The differential serial line consists of two states namely: the J and K state. Logic "1" is received when the D+ line when it is 300mV greater than the D- line and logic "0" is received when the D+ line is 300mv less than the D- line (13).
The packet sizes of the USB protocol are given in the figures below:
Fig 4.10: USB Packet sizes and Structures
USB Token Packet
USB DATA Packet
USB Handshake Packet
USB Start of Frame Packet
Source:http://www.ftdichip.com/Support/Documents/TechnicalNotes/TN_116_USB%20Data%20Structure.pdf (13)
The token packet shown in Fig 4.10a is used to access the correct address and end point. Both the address and end point should be correctly decoded for normal operation (13). The data packet shown in Fig 4.10b can be of variable length depending upon the data to be transmitted. The handshake packet shown in Fig 4.10c is used to indicate whether the signal is either sent or not sent. And finally the start of frame packet shown in Fig 4.10d is used to indicate the start of a particular frame.
4.4.2 USB-TTL conversion
As established previously, USB operates on NRZI coding. This was not the preferred method of coding; since transmission of NRZI coding would be affect the performance of the LED (9). So this output needed to be converted into TTL logic. In order for this to be performed a USB-TTL converter was required.
For the conversion of USB to TTL, an FT232R USN-TTL serial cable was used. The FT232R cable houses an FT232R chip which converts USB data into asynchronous serial data at TTL levels (9). Some of the key features of this cable, as mentioned in the cable datasheet (14) include; low operating current which is nearly 70μA, Low USB bandwidth consumption, high output drive option, improved EMI performance to name a few. The figures below show the FT232R chip and its internal circuitry:
Fig 4.11: Internal structure of the USB-TTL converter along with the FT232RQ chip
Source: http://www.farnell.com/datasheets/81225.pdf (14)
The cable however uses the RS232 serial form of communication for its conversion into TTL logic. In RS232 communication logic "1" is represented by a negative voltage and logic "0" is represented by a positive voltage (9) (15). When this was converted into TTL logic levels by the on board chip, it represented TTL levels: logic "1" as 0V and logic "0" as 5V. In order to prevent the LED from overheating (9) this output signal was passed through a high speed hex inverter IC to get the required TTL levels for transmission. The inverter used for this purpose was the M74HC14 Hex-Schmitt Inverter. The MC74HC14 which was used had a very low propagation delay time of 14ns (16).
Fig 4.12: Schmitt Inverter with truth table
Source: http://www.electronics-tutorials.ws/logic/logic_4.html (22)
Optics
Optics was used in the experiment for the optical amplification of light from the LED. Generally for the amplification of optical signals, from the transmitter end, in wired optical communication optical amplifiers are used. The main reason as to why this was done was to improve the SNR at the receiver end and to overcome the losses sustained by the signal. Urquhart, P.et.al (17) stated that, the use of such amplifiers was to overcome losses incurred by the signal due to the fibre transmission medium, various other passive components and power division in multi-point systems. Without these amplifiers optical-to-electrical-to-optical regenerators were required (17). But in the case of wireless optical communication this method cannot be employed for the amplification of optical signals.
However alternate methods of optical amplification could be employed to ensure that the receiver section received the ample amount of light from the signal beam. The two methods which are used for amplification of wireless optical signals are:
Use of concentrators
Use of lenses
Ramirez-Iniguez, R.et.al (18) states that concentrators help improves the collection efficiency at the photo detector by focussing light beams incident over a large area onto the photo detector. This makes the use of smaller photodiodes, which have lower capacitance and greater sensitivity, possible for the reception of optical signals (18). Various different methods for the design of concentrator types have been proposed by Miñano, Juan C. et al (19). These included using the Winston-Welford method and the Simultaneous Multiple Surface method for the design of concentrators (19).
The fabrication of such concentrators made it much more difficult to commercially purchase than regular optical lenses. And since using a concentrator would involve use of a special adhesive gel, with the same refractive index as that of the concentrator, to attach it to the photo detector. Therefore the use of concentrators in the experiment was avoided. Instead the role of the concentrators was taken over by optical lenses. Since the communication system which was set up was a point-to-point system the main focus of the lenses were to make the light beam as directional as possible without much beam divergence.
In order to achieve a directional beam the best method of approach was to use two Plano-convex lenses at the transmitter and receiver sections. The non-curved surfaces of the lenses were made to face each other in a straight line. The aim was to place the convex section of the lenses in front of the LED at the transmitter section and the photo-detector at the receiver end. This would help collect the light emitted from the LED and focus it to the receiver lens which would then concentrate it to the photo detector. The block diagram in figure 3.5.1 illustrates this principle.
In the experiment carried out four lenses were used for the purpose of amplification. A pair of BK-7 Plano-convex lenses with focal lengths of 63mm was used initially. Then various distance measurements were taken by replacing the lens at the transmitter end with the 2 bigger lenses of focal lengths 25cm and 65 cm respectively.
Fig 4.13: Block diagram of the data-link using lenses
Transmitter Circuit Design
The various components making up the transmitter section have already been discussed in this chapter. The following section goes into detail about the transmitter circuit design and test circuit design. The basic functional block diagram of the transmitter section is shown below:
Fig 4.14: Block diagram of the transmitter section
The design of the transmitter section involves the connection of all of the components so far mentioned.
Circuit Diagram and Theory
Various soft-wares were used in the development of the circuit design as well as the simulation. These included Proteus 7.0, Multisim 11 and Multisim 12. Eventually Multisim 12 was chosen as the software on which all the circuits' were designed.
Fig 4.15: Transmitter circuit
The above figure 4.15 shows the circuit diagram of the transmitter section. Its working is as follows: the input data from the USB-TTL cable was fed to the hex-inverter where it was converted into TTL logic. This was then fed to the LED driver circuit. The LED driver circuit which was the TC4426A IC had the secondary input pin (pin 4) and the ground pin (pin 3) both grounded. The driver IC provided the necessary current to drive the LED i.e. to turn it ON and OFF according to the data being sent.
From the output pin of the driver IC a small resistor was connected to ground and the cathode of the LED was connected to the output pin of the IC. The value of the resistor R1 was calculated by using Ohms Law. The forward voltage of the LED was 3V and the typical forward current as given in the datasheet was 20mA. Based on these values and with the fact that the supply to the LED was 10V, the resistor value was designed as:
And so a standard value of 330Ω was used as the value of resistor R1. The reason why this resistor was connected was to ensure that some amount of current was sunk to the ground from the output of the TC4426A IC. This meant that the LED would not completely turn itself off when logic "0" was encountered. A small amount of current would flow through the LED even when it was supposed to be turned OFF. This had two advantages: firstly this reduced the propagation delay as the LED was not required to turn itself ON from an OFF state which causes a certain amount of delay and secondly it prevented the LED from overheating.
The LED was connected to the supply in series with a resistance R2 of value 470Ω. This series resistance was connected to ensure that the LED would not be damaged by the large amount of current flowing through it. This was because in the absence of a series resistor the LED would itself act as resistor to the input voltage source. Since the internal resistance of an LED is very small this meant that according to Ohms law large amount of current would pass through the LED thereby damaging it. This value was decided in a similar manner to resistor R1 using Ohm's Law as shown below.
The test circuitry, which was designed, however was slightly different from the circuit shown in fig 4.16. The circuitry for testing the transmitter section is shown below:
