T305: Digital Communication

T305: Digital Communication

Dr. Farid Jradi : Tutorial 20

Introduction to the Tutorial:

This is the number twenty face-to-face tutorial of this course and its basic aim is to discuss the contents of: Systems and Processes Block 5: Part 2; Modulation.

Reference Material for this Tutorial:

This tutorial is based on the following references:

1-T305 Block 5: Systems and Processes: Part 2; Modulation

The various topics that will be covered in this tutorial are discussed below.

Topic 1: Multiplexing, Multiple Access and Duplexing
Topic 2: Modulation in Optical-Fibre Systems
Topic 3: Summary
Topic 4: Preparation for Next Tutorial

Topic 1: Multiplexing, Multiple Access and Duplexing

One of the purposes of modulation is to allow several signals to share the same transmission medium.

If a system is using multiplexing, it behaves as though it were several links in parallel as shown below. At each end there is a multiplexer which shares the single link between the multiple incoming and outgoing channels.

Fig. Multiplexing.

If a system is using multiple access, it will be used to provide channels to several different nodes on demand. Each node is directly connected to the medium and has to sort out the rules for getting access to the medium for itself. The nodes can be in different physical locations, as with terminals on Ethernet sharing a bus by CSMA/CD.

The idea of multiplexing was widely used in the development of the public switched telephone network (PSTN) and was a particular concern of the transmission engineers in that industry.

The idea of multiple access was used by the data communications industry for the way in which terminals share a transmission medium in a local area network, but also by the industry developing satellite communications.

There are broadly four ways of sharing a medium: space division, time division, frequency division and code division.

Space division is where the different signals use different parts of the physical medium, as happens in,for example, cellular mobile communications when the same frequency isre-used in geographically separate cells

Duplexing

Duplexingis the process of allowing two-way communication. Within cabled communication networks – between two telephone exchanges, for example – the two directions are usually provided over separate wires (or optical fibres), and so might be said to be using space division.

In wireless communications, this option isn't usually available, so one of the other schemes (time, frequency or code division) is used instead.

With optical fibre a device known as an optical circulatorcan be used, as shown below. The circulator uses three input/output fibres. Because of the internal construction of the circulator, light arriving at fibre 1 leaves from fibre 2 and light arriving at fibre 2 leaves from fibre 3.

Fig. Use of an optical circulator for duplexing.

Duplex operation over a single pair of electrical wires can be done, using a device known as an electrical hybrid, in a similar way.

There is a hybrid on the telephone, which directs the incoming signal to the earpiece of the handset and directs the signal from the mouthpiece out towards the local exchange. Some of the signal from the mouthpiece is intentionally sent directly to the earpiece.

The telephone would sound ‘dead’ to the user otherwise– but if too much went that way you’d be deafened by your own voice, and there wouldn’t be enough of the signal sent to the exchange! This fraction of the outgoing signal directed to the local earpiece is known as sidetone.

Time Division

Time division is also used for duplexing. Some systems for mobile communications, for example, have used time division duplex(TDD) so that the uplink (mobile terminal to base station) and downlink (base station to mobile terminal) communication can share the same channel.

The communication is structured into frames, and part of the frame is allocated to one direction of communication and the other part to the other direction. It may be that half the frame is used for each direction, which gives equal capacity to both directions, or it may be that more of the frame is allocated to one direction than the other.

As can be seen below, in TDD there is a period of time unused or ‘wasted’ while the direction of transmission swaps around. The wasted time is equal to twice the propagation time between the two terminals, and is an unavoidable overhead due to the use of TDD.

Fig. Time division duplex.

Frequency Division

Frequency division remains fundamental to the exploitation of many communication resources. For example, all radio transmissions use frequency division in the way that the available radio spectrum is shared between different services.

Within any one service, furthermore, frequency division multiplexing or multiple access is also still often used to share the capacity between different channels. Some mobile communication systems use either FDM or hybrid FDM/TDM schemes to allow multiple channels in one cell.

In GSM (shown below), for example, eight channels are multiplexed together by time division multiplexing on each carrier frequency, then carrier frequencies are spaced at 200 kHz intervals for frequency division multiplexing.

Frequency division is also used to allow duplex operation (frequencydivision duplex, FDD). In GSM, for example, the frequency band used for communication from the mobile to the base station is lower than from the base station to the mobile. Specifically, the carrier frequencies are in pairs for uplinks and downlinks, separated by 45 MHz.

Fig. Multiplexing and

duplexing in

GSM.

Code Division / Spread Spectrum

In time division multiplexing / multiple access the different channels are divided in time but use the same frequencies as shown below (a). In frequency division multiplexing / multiple access the signals are divided in frequency but are active simultaneously (b).

It is possible through the technique known as code division multiplexingto multiplex channels in such way that they are divided neither in time nor frequency, yet can still be separated or de-multiplexed (c).

Fig. Time division, frequency division and code division multiplexing.

Code division multiplexing can be explained using the concept of orthogonality. Orthogonality can in fact be used to explain all multiplexing – although it isn’t the usual explanation of FDM or TDM. Figure below shows the general idea.

Fig. Multiplexing

with

orthogonal

symbols.

Assume that each channel sends data using on–off keying with its own symbol. Channel 1 uses the symbol labelled S1, so that it sends S1 when it wants to send a data 1, it sends nothing when it wants to send a data 0. Channels 2 and 3 do the same using the symbols labelled S2 and S3, respectively. S1, S2 and S3 are chosen so that they are mutually orthogonal.

At the receiver, there are correlators for all three symbols. Assume that the channels are synchronized so that the symbols are sent by each channel simultaneously. Because the symbols are orthogonal, the presence or absence of S2 or S3 has no effect on the output of the S1 correlator, S1 and S3 do not affect the output of the S2 correlator, and S1 and S2 do not affect the S3 correlator.

If S1, S2, and S3 are segments of sinusoids of different frequencies, then we have an example of FDM.

Sets of code words having the orthogonal property are described as orthogonal codes, and from any set of orthogonal codes, orthogonal symbols can be generated.

There is a similarity between the coding for code division multiplexing and for direct sequence spread-spectrum. In each case, a single bit is converted to a code sequence- the spreading sequence in the case of spread spectrum and the orthogonal sequence in the case of code division multiplexing.

It is worth noting that the symbols used in code division multiplexing transmit simultaneously (they are not separated in the time domain) and their spectra occupy the same frequency range (they are not separated in the frequency domain). The 'division' in code division multiplexing is due to the orthogonality of the symbols.

Topic 2: Modulation in Optical-Fiber Systems

Light is an electromagnetic wave similar to a radio signal but at a very much higher frequency. Although modulation of an optical signal is in principle the same as modulation of a radio signal, in practice the approach is rather different.

The frequency of light used in optical fiber is greater than 1014 Hz.Radio frequencies used for communications have been graduallyincreasing over the years – with advancing technology and increasingdemand for bandwidth – but at the time of writing the highestfrequencies being exploited are under 100 GHz (i.e. 1011 Hz).

Thisdifference of more than three orders of magnitude in frequency makesa very significant difference to the way in which the signals can beprocessed. Nevertheless, advancing techniques for handling opticalsignals are making it increasingly possible to treat optical signals asvery-high-frequency versions of radio signals.

Modulation

The basic modulation scheme used for the vast majority of optical-fibre communication systems is binary amplitude shift keying. Nominally it is usually a form of on–off keying that is used, but, for practical reasons to do with the operation of laser diodes, sometimes the light is not quite switched fully ‘off’ for the low signalling state.

The transmitter is a semiconductor light source which is switched ‘on’ for a data 1 and ‘off’ for a data 0. There are two sorts of light sources used: light emitting diodes (LEDs) and laser diodes. The LEDs used are based upon the same technology as the LEDs widely used as indicator lights on electronic equipment – although, specially designed for communications applications.

Laser diodes are made of the same or similar material as LEDs but have a different physical construction, which makes them ‘lase’. It is generally possible to couple more light power from a laser diode into an optical fibre than from an LED, and the optical linewidthof a laser is narrower than that of an LED.

The linewidth is a measure of the range of optical frequencies contained in the optical output, which is not a pure, single-frequency, sinewave. Linewidths can be measured either in terms of wavelength or in terms of frequency. For example, LEDs typically have linewidths of the order of tens of nanometers (1 nm is 10−9 m).

Example:

If an LED has a linewidth of 10 nm centered on a wavelength of1300 nm, what is its linewidth measured in terms of frequency?

(Hint: the wavelength 1 and frequency ƒ of a sinusoid are related by1ƒ = c, where c is the speed of light, approximately 3 × 108 m s−1.)

Solution:

Optical fibre is made from either plastic or glass. Glass fibre can be made with attenuation much lower than plastic, and is invariably used for mid-to-long-distance communication. Plastic fibre is cheaper and easier to use, but is only suitable for short-distance communication, perhaps over a few meters or maybe up to a few tens of meters.

However, there is still a lot ofresearch effort being expended on developing lower-loss plastic, and it maybe that in the future plastic fibre could be suitable for use in LANs or eventhe ‘last mile’ of the telecommunications network

The wavelength dependence of the attenuation of telecommunication-grade glass fiber typically has the characteristics shown below. The shape of this graph is determined by the fundamental properties of the material from which the glass is made (silica) as well as the presence of impurities within the glass.

Fig. The attenuation oftelecommunication- grade singlemode optical fibre.

Notice in particular the rise at around 1400 nm. This is caused by the presence of a small amount of water in the glass. The lowest attenuation occurs between 1500 and 1600 nm, a range of wavelengths that is often referred to as the ‘1500 nm window’.

The other widely used ‘window’ is the region of low attenuation around 1300 nm. The 1300 nm window was used more widely in the telecommunications network for many years for two reasons: first that it was easier to manufacture 1300 nm optical light sources, and second, because of the effects of dispersion.

At the time of writing, 1500 nm is becoming increasingly important,especially for very long-distance communication (such as trans-Atlanticcables) both because of the lower attenuation of fiber at that wavelength,but also because of the development of optical amplifiers, as discussedlater.

Dispersion

When rectangular pulses are transmitted over optical fibre, as well as being smaller at the far end due to the attenuation, they are also spread out, resulting in inter-symbol interference. The spreading is caused by the effect known as dispersion.

Standard single-mode fibre, the type most widely installed in the past, has a dispersion minimum at around 1300 nm, and has significantly higher dispersion around 1550 nm. Dispersion in single-mode fibre depends upon the dispersion characteristics of the fibre and the spectral linewidthof the light source: the narrower the linewidth, the less will be the dispersion.

Repeaters: regeneration and amplification

To regenerate a digital optical signal it is, in practice, necessary to convert it to an electrical signal and do the regeneration electrically. Erbium-doped fibre amplifier(EDFA) amplifies the incoming optical signal with the addition of very little noise or distortion.

If the dispersion can be kept under control, the use of EDFAs is very attractive because of their simplicity and reliability and because of their transparency to the conveyed signal. Transparency means that it doesn’t matter what the signalling rate or modulation scheme is – or whether the signal is even analogue or digital.

Wavelength Division Multiplexing

Frequency division multiplexing (FDM) is used in optical fibre in order to increase the data-carrying capacity of a single fibre. One method of doing it is shown below, in general terms. Each of the channels originates from a separate laser diode operating at a different wavelength. The signals are brought together in a single optical fibre for transmission. At the receiver the signals at the different wavelengths are split to different photodiodes and separately detected.

Fig. Wavelength division multiplexing.

Because it is normal to refer to the wavelength of an optical-fibre system, this scheme is usually described as wavelength division multiplexing (WDM).

Topic 3: Summary

After Channel coding, data is formatted before putting it into transmission medium by coding followed by symbol generation. The coding might be line (commonly used for baseband systems) or a spreading code for spread-spectrum.

The symbols might be rectangular pulses (for a baseband system) or segments of sinusoids (for a modulated system). Pulse Shaping might be used with both baseband and modulated systems to restrict the bandwidth of the signal while also controlling inter-symbol interference.

The main functions of the line coding are usually to balance the signal (to prevent baseline wander) and/or to ensure frequent transitions (to simplify clock recovery at the receiver). The output from line coding is a sequence of signal states, which might be binary or multi-level. Each signaling state is then turned into a unique symbol by the symbol generation.

The symbols used to present different states can differ in magnitude or in shape (or both). Receivers can look for particular shapes within a received waveform by correlating against a reference copy of the shape being looked for. The correlation between orthogonal symbols is zero.

Systems using segments of sinusoids that differ only in amplitude are described as amplitude shift keying (ASK) systems. Systems whose symbols differ in phase are described as phase shift keying, and those whose symbols differ in frequency are described as frequency shift keying (FSK).

Special cases of binary modulation schemes include on-off keying (OOK) witch is ASK with one of the amplitudes 0, differential phase shift keying (DPSK) which is PSK combined with differential coding, and continuous-phase frequency shift keying (CPFSK) which is FSK without phase-steps between symbols.

Spread-spectrum modulation is designed deliberately to broaden the spectrum of the signal, and is done either through frequency hopping (which may be fast or slow hopping) or through direct sequence spreading.

If more than one channel is to share one transmission path as a multiplexed signal or for multiple access, then space division, time division, frequency division or code division may be used. The same approaches can also be used to allow duplex operation over a common transmission medium.

The most common modulation used with optical-fiber transmission is a form of ASK which is nominally OOK, but not quite switching off if the optical source is a laser diode.

Optical fiber is particularly important for long-distance transmission because fiber can be manufactured with very low attenuation. Multiple signals can share a single optical fiber through wavelength division multiplexing (WDM).

Topic 4: Preparation for Next Tutorial

1)Overview the contents of the Modeling Activities Block 6: Signal Impairments.