CONTENTS
Chapters Page nos.
1. INTRODUCTION 1
1.1 Different types multiplexing techniques 2
1.1.1 Frequency Division Multiplexing 2
1.1.2 Time Division Multiplexing3
1.1.2.1 Synchronous Time Division Multiplexing 3
1.1.2.2 Statistical Time Division 3
1.1.3 Wavelength Division Multiplexing (WDM) 4
1.1.4 Code Division multiplexing(CDM) 4
2. PRINCIPLES OF OPTICAL TIME-DIVISION 5
MULTIPLEXING AND DEMULTIPLEXING
2.1Electrical and Optical Multiplexed Systems 5
2.2. Optical Time-Division Multiplexing 7
2.3. Optical Switching and Demultiplexing 12
2.4 Timing Recovery 19
3 .EXPERIMENTS 20
3.1Multiple-Laser Systems 21
3.2 Single-Laser System 31
4 CONCLUSIONS AND OUTLOOK 32
5.REFERENCES 35
Chapetr-1
1. Introduction
Progress in very high bit-rate light wave systems has been stimulated by consistent demands for expanded transmission capacity. These demands have led to increased interest in multi giga bit per-second pulse-code modulated (PCM) systems, and have put an emphasis on the need for high-speed and wide-band electronics in light wave transmitters and receivers. To date, it has generally been possible to meet these needs with high-speed Si and GaAs circuits, but at gigabit-per-second bit rates it becomes increasingly difficult to develop the necessary digital electronic circuits. One method to relieve this electronic speed bottleneck is to extend the well-known techniques of electrical multiplexing into the optical domain. The two main approaches to optical multiplexing are optical wavelength-division (or frequency-division) multiplexing and optical time-division multiplexing. This presentation concentrates on optical time division multiplexing. In optical time-division multiplexing (OTDM), a high bit-rate data stream is constructed directly by time-multiplexing several lower bit-rate optical streams. Similarly, at the receiver end of the system, the very high bit-rate optical signal is demultiplexed to several lower bit-rate optical signals before detection and conversion to the electrical domain. This approach to optical time-division multiplexing and demultiplexing moves the demand for high-speed performance away from electronic devices such as transistors, and places it on optical and optoelectronic devices such as pulsed semiconductor lasers and optical switches. The time-division multiplexing approach is a purely digital technique and is therefore compatible with the concept of an all-digital network that combines switching and transmission. In addition, optical time-division multiplexing offers system design flexibility, including the possibility of adjustable bandwidth allocation in different baseband channels and the possibility of simple system hardware in which only a single transmitter laser is required for all channels. The potential of optical time division multiplexing and demultiplexing for very high bit-rate PCM systems has been recognized for more than two decades but until recently there have been few system-level demonstrations of the technique at multi gigabit-per- second bit rates. The implementation of very high bit-rate. OTDM systems has been slow because electronic multiplexing has usually served adequately and because the necessary hardware, such as high-speed optical switches and compact pulsed semiconductor lasers, has only recently reached a sufficient state of refinement. This paper describes recent experiments in optical time-division multiplexing and demultiplexing that have been made possible by improvements in lasers and switch /modulators. Our emphasis is on very high bit-rate point-to-point transmission systems but many of the concepts are also relevant to multiuser systems and time-multiplexed photonic switching networks. We review system architectures, describe the requirements on individual system components, and give examples of transmission system experiments operating at bit rates up to 16 Gbit /s. One of the key factors affecting the performance of multiplexed systems is crosstalk between baseband channel . In presentation explore sources of crosstalk in optical time-division multiplexing and demultiplexing, and describe how the main system components affect the overall crosstalk performance. Under the simplest conditions, a medium can carry only one signal at any moment in time. For multiple signals to share one medium, the medium must somehow be divided, giving each signal a portion of the total bandwidth.The current techniques that can accomplish this include
1.1 Different types multiplexing techniques
1.frequency division multiplexing (FDM)
2.Time division multiplexing (FDM)
3.wavelength division multiplexing (WDM)
4.code division multiplexing (CDM)
1.1.1Frequency Division Multiplexing
Assignment of non-overlapping frequency ranges to each “user” or signal on a medium. Thus, all signals are transmitted at the same time, each using different frequencies.A multiplexor accepts inputs and assigns frequencies to each device. The multiplexor is attached to a high-speed communications line.A corresponding multiplexor, or demultiplexor, is on the end of the high-speed line and separates the multiplexed signals.
1.1.2Time Division Multiplexing:
Sharing of the signal is accomplished by dividing available transmission time on a medium among users.Digital signaling is used exclusively. Time division multiplexing comes in two basic forms:
1. Synchronous time division multiplexing, and
2. Statistical, or asynchronous time division multiplexing
1.1.2.1 Synchronous Time Division Multiplexing
The original time division multiplexing.The multiplexor accepts input from attached devices in a round-robin fashion and transmit the data in a never ending pattern.T-1 and ISDN telephone lines are common examples of synchronous time division multiplexing.
1.1.2.2 Statistical Time Division:
MultiplexingA statistical multiplexor transmits only the data from active workstations (or why work when you don’t have to).If a workstation is not active, no space is wasted on the multiplexed stream.A statistical multiplexor accepts the incoming data streams and creates a frame containing only the data to be transmitted. A statistical multiplexor does not require a line over as high a speed line as synchronous time division multiplexing since STDM does not assume all sources will transmit all of the time!Good for low bandwidth lines (used for LANs).
1.1.3Wavelength Division Multiplexing (WDM)
Give each message a different wavelength (frequency)Easy to do with fiber optics and optical sources, Dense wavelength division multiplexing is often called just wavelength division multiplexing .Dense wavelength division multiplexing multiplexes multiple data streams onto a single fiber optic line. Different wavelength lasers (called lambdas) transmit the multiple signals.Each signal carried on the fiber can be transmitted at a different rate from the othersignals.Dense wavelength division multiplexing combines many (30, 40, 50, 60, more?)onto one fiber.
1.1.4 Code Division Multiplexing (CDM)
Also known as code division multiple access (CDMA),An advanced technique that allows multiple devices to transmit on the same frequencies at the same time using different codes .Used for mobile communications. time. Each mobile device is assigned a unique 64-bit code (chip spreading code)
Chapter-2
2. PRINCIPLES OF OPTICAL TIME-DIVISION
M UL TIPLEXING AND DEMUL TIPLEXING
2.1Electrical and Optical Multiplexed Systems
The basic principle of time-division multiplexing and demuItiplexing is that each of the baseband data streams is allocated a series of time slots on the multiplexed channel.
A multiplexer (MUX) assembles the h higher bit-rate bit stream from the baseband streams and a demultiplexer(DEMUX) reconstructs bit streams at the original lower bit rate by separating bits in the multiplexed stream . The techniques for this process are well established for electrical time-division multiplexing and demultiplexing but are only now emerging in optical systems .Fig. I highlights the similarities and differences between electrically time-multiplexed and optically time multiplexed light wave systems . In this figure and in subsequent figures , thick lines are used for optical (fiber) signal paths and thin lines are used for electrical signal paths .In an electrically time-multiplexed system , Fig . lea),multiplexing is carried out in the electrical domain , before the electrical-to-optical (E/O) conversion . Demultiplexing is carried out after the optical-to-electrical (0 / E) conversion. For n baseband channels, each of bit rate B, the multiplexed bit rate is nB. Potential electronic bottle necks occur in the MUX and the E /0 converter, and in the 0/ E converter and the DEMUX , where the electronics must operate at the full multiplexed bit rate . These bottlenecks arise from a) speed limitations of digital integrated circuits, b) speed limitations of high-power and low-noise linear amplifiers used to drive the laser or modulator in the E/O converter and in the O/E converter , c) limited modulation bandwidths of lasers and modulators , and d)the fact that the receiver sensitivity offered by an avalanche photodiode degrades by more than 3 dB for every octave increase in receiver bandwidth. These problems have so far limited the maximum b it rate for electrically multiplexed systems to 10 Gbit/s. In the optically multiplexed system, Fig. l (b) , the electronic bottlenecks are removed by moving the E / 0 and o/ E converters (i.e, the transmitters and receivers) into the baseband channels. Multiplexing is carried out after the E /0 conversion and demultiplexing is carried out before the O/E conversion. All electronics associated with signal processing operate only at the baseband bit rate .Note that a control signal is needed to drive the demultiplexer. In general , this control signal could be either electrical or optical , depending on the demultiplexer technology. At p resent the most practical optical demultiplexers are based on electrooptic switches, which use electrical control signals . It will be shown later that the bandwidth of this electrical control signal need not be large for demultiplexing in an OTDM system. An important difference between electrically multiplexed and optically multiplexed systems is that in electrical systems the multiplexing and demultiplexing can be carried out at points in the system where the signal has been amplified to large levels . The signal-to-noise ratio is determined by the receiver and its associated low-noise front end,and is not affected by loss in the multiplexing or demultiplexing operations. In an OTDM system, on the other hand , the multiplex ing and demultiplexing is carried out on the optical signal, Thus optical losses reduce the signal level relative to the receiver noise, and losses must be kept small . It may be feasible, however, to place optical amplifers at selected points in the system to compensate for some of these losses .
2.2. Optical Time-Division Multiplexing
In this section we consider optical waveforms and timing requirements for optical multiplexing, and examine topologies for transmitters and multiplexers. The operation of time-multiplexing several lower bitrate baseband channels onto a higher bit-rate channel can be divided into three sub functions: sampling, timing, and combining. The sampling function takes samples of the incoming baseband data stream, thereby identifying the value of each incoming bit. The timing function ensures that the samples are available at the correct time slots on the multiplexed channel. The combining function assembles all the sampled baseband data streams to generate the higher bit-rate multiplexed data stream. In multi gigabit-per-second electrically multiplexed systems it is convenient to sample each of the input data streams using short sampling pulses that are timed to correspond to the appropriate time slots on the multiplexed bit stream. If the sampling pulse widths are less than one bit-period of the high bit-rate multiplexed signal, the combiner can be a simple summing circuit. A similar strategy is a preferred approach to optical time-division multiplexing because it can capitalize on mode-locked and gain switched semiconductor lasers, which are capable of generating pulses more than ten times shorter than electrical pulses. In this approach to multiplexing the sampling function is carried out in the E /0 converters (i.e., the system transmitters). Consequently, the mux in Fig. l(b) is required only to do the combining function. The following concentrates on this method of multiplexing. Fig. 2 shows the schematic of an E /O converter (transmitter) that can be used to sample the input data before optical combining.
Short optical pulses from a laser are incident on an optical modulator, which is driven by an input electrical data stream. The electrical data stream could be either in the return-to-zero (RZ) or the non-retum-to-zero (NRZ) format, but NRZ is usually preferable because it minimizes the bandwidth requirements of the baseband digital electronics, the modulator, and its drive amplifiers. The optical pulse train from the laser samples the electrical input data via the modulator, thereby converting it from NRZ in the electrical domain to RZ in the optical domain. An important feature of the laser-modulator combination in Fig. 2 is that when the laser and input data are correctly timed, the modulator is either fully "on" or fully "off" when the optical pulse passes through it. This means that the modulation process does not cause the optical signal to chirp. Sampling the baseband data in the E / 0 converter enables the optical combining function to be carried out using a passive device such as a fiber directional coupler power combiner. Another advantage of this approach is that independent of the type of combiner used in the system, the RZ output format from the E / 0 converters leads to low multiplexing crosstalk (see below). Furthermore, the optical power from the lasers in the E /0 converters is used efficiently since the signal in each baseband channel is zero during time slots to be occupied by other channels in the multiplexed data stream. This is a significant practical consideration because semiconductor lasers are average-power limited devices. The timing scheme for a general n-channel optical time division multiplexed system is shown in Fig. 3. The n optical signals incident on the combiner are RZ pulse trains with repetition rates B and with pulsewidths T(measured at the baseline). The incoming bit streams are temporally offset from one another by delays D. Data are encoded on each pulse train, before combining, so in general some of the individual pulses will have zero amplitude. However, in Fig. 3 all bits are shown as "ones" for clarity. When the pulse spacing in Fig. 3 is adjusted for maximum multiplexed hit rate, each pulse in the multiplexed bit stream just comes into contact with its nearest neighbors. Under these circumstances D = T, and the multiplexed bit rate is 1 / T. With laser pulses that are 10 ps wide at the baseline, for example, the multiplexed bit rate could be as high as 100 Gbit / s.
Fig. 3 shows that the RZ signal format provides low system crosstalk. Since each baseband signal is always nominally zero except in its allotted time slot on the multiplexed bit stream, it cannot interfere with other channels. In practice, the pulse stream will have a finite on / off ratio and the resulting baseline light signal between pulses will cause a component of crosstalk. Furthermore, leading and trailing tails on the pulses will also cause cross talk if the pulses are spaced such that tails overlap on adjacent pulses.