Biblio5
.pdf
250 CONCEPTS IN TRANSMISSION TRANSPORT
Figure 9.33 The WDM concept.
able, and in the 1550-nm band, with about 100 nm bandwidth available, which includes the operational bandwidth and guardbands, the capacity on a single fiber would be 40 WDM and 50 WDM channels, respectively, on each fiber.19 Then each fiber could transport an aggregate of 100 Gbps and 120 Gbps, respectively, for each band. At present (1998), 16-channel systems are available, off-the-shelf.
9.4.9 Fiber Optic Link Design
The design of a fiber optic communication link involves several steps. Certainly the first consideration is to determine the feasibility of such a transmission system for a desired application. There are two aspects of this decision: (1) economic and (2) technical. Can we get equal or better performance for less money using some other transmission medium such as wire pair, coaxial cable, LOS microwave, and so on?
Fiber optic communication links have wide application. Analog applications for cable television (CATV) trunks are showing particularly rapid growth. Fiber is also used for low-level signal transmission in radio systems, such as for long runs of IF, and even for RF. However, in this text we stress digital applications, some of which are listed below:
•On-premises data bus;
•LANs (e.g., fiber distributed data interface);
Figure 9.34 The use of regenerative repeaters on a WDM link.
19Guardbands are empty spaces to provide isolation between adjacent channels. This helps minimize interference from one channel to the next. Of course, that bandwidth allocated to guardbands is nonrevenuebearing, and thus must be minimized.
9.4 FIBER OPTIC COMMUNICATION LINKS |
251 |
•High-level PCM or CVSD configurations;20 SONET and SDH;
•Radar data links;
• Conventional data links where bit/ symbol rates exceed 19.2 kbps; and
• Digital video including cable television.
It seems that the present trend of cost erosion will continue for fiber cable and components. Fiber optic repeaters are considerably more expensive than their PCM metallic counterparts. The powering of the repeaters can be more involved, particularly if power is to be taken from the cable itself. This means that the cable must have a metallic element to supply power to downstream repeaters, thereby losing a fiber optic advantage. Metal in the cable, particularly for supplying power, can be a conductive path for ground loops. Another approach is to supply power locally to repeaters with a floating battery backup.
A key advantage to fiber over metallic cable is fewer repeaters per unit length. In Chapter 6 we showed that repeaters in tandem are the principal cause of jitter, a major impairment to a digital system such as PCM. Reducing the number of repeaters reduces jitter accordingly. In fact, fiber optic systems require a small fraction of the number of repeaters, compared with PCM for the same unit length, either on wire-pair or coaxial cable.
9.4.9.1 Design Approach. The first step in the design of a fiber optic communication system is to establish the basic system parameters. Among these we would wish to know:
• Signal to be transmitted: digital or analog, video/ CATV; bit rate and format;
•System length, fiber portion end-to-end;
•Growth requirements (additional circuits, increase of bit rates);
•Availability/ survivability requirements; and
•Tolerable signal impairment level, stated as signal-to-noise ratio on BER at the output of the terminal-end detector.
The link BER should be established based on the end-to-end BER. In the past we
had used 1 × 10−9 as a link BER. Bellcore now requires 2 × 10−10 end-to-end. One link value might be 1 × 10−12.
However, there is still another saving factor or two. If we have operation at the zerodispersion wavelength, about 1300 nm, dispersion may really not be a concern until about the 1-Gbps rate. Of course, at 1300 nm we have lost the use of the really low loss band at around 1550 nm. There is an answer to that, too. Use a fiber where the minimum dispersion window has been shifted to the 1550-nm region. Such fiber, of course, is more costly, but the cost may be worth it. It is another trade-off.
The designer must select the most economic alternatives among the following factors:
•Fiber parameters: single mode or multimode; if multimode, step index or graded index; number of fibers, cable makeup, strength;
•Transmission wavelength: 820 nm, 1330 nm, or 1550 nm;
•Source type: LED or semiconductor laser; there are subsets to each source type;
20CVSD stands for continuous variable slope delta (modulation), a form of digital modulation where the coding is 1 bit at a time. It is very popular with the armed forces.
252CONCEPTS IN TRANSMISSION TRANSPORT
•Detector type: PIN or APD;
•Use of EDFA (amplifiers);
•Repeaters, if required, and how they will be powered; and
•Modulation: probably intensity modulation (IM), but the electrical waveform entering the source is important; possibly consider Manchester coding.
9.4.9.2 Loss Design. As a first step, assume that the system is power limited. This means that are principal concern is loss. Probably a large number of systems being installed today can stay in the power-limited regime if monomode fiber is used with semiconductor lasers (i.e., LDs or laser diodes). When designing systems for bit rates in excess of 600 Mbps to 1000 Mbps, consider using semiconductor lasers with very narrow line widths (see Figure 9.29) and dispersion shifted monomode fiber. Remember that “dispersion shifting” moves the zero dispersion window from 1300 nm to 1550 nm.
For a system operating at these high bit rates, even with the attribute of monomode fiber, chromatic dispersion can become a problem, particularly at the desirable 1550nm band. Chromatic dispersion is really a form of material dispersion described earlier. It is the sum of two effects: “material dispersion” and waveguide dispersion. As one would expect, with material dispersion, different wavelengths travel at different velocities of propagation. This is true even with the narrow line width of semiconductor lasers. Waveguide dispersion is a result of light waves traveling through single-mode fibers that extend into the cladding. Its effect is more pronounced at the longer wavelengths because there is more penetration of the cladding and the “effective” refractive index is reduced. This causes another wavelength dependence on the velocity of light through the fiber, and therefore another form of dispersion. Thus the use of semiconductor lasers with very narrow line widths (e.g., <0.5 nm) helps mitigate chromatic dispersion (Ref. 11).
Link margin is another factor for trade-off. We set this dB value aside in reserve for the following contingencies:
•Cable reel loss variability;
•Future added splices (due to cable repair) and their insertion loss; and
•Component degradation over the life of the system. This is particularly pronounced for LED output.
CCITT recommends 3 dB for link margin; others (Ref. 9) recommend 6 dB. Ideally, for system reliability, a large margin is desirable. To optimize system first cost, as low a value as possible would be desired.
The system designer develops a power budget, similar in many respects to the path analysis or link budget of LOS microwave and satellite communication link design. However, there is little variability in a fiber optic link budget; for example, there is no fading.
For a first-cut design, there are two source types, LED and semiconductor laser. Expect a power output of an LED in the range of −10 dBm; and for the semiconductor laser budget 0 dBm, although up to nearly +10 dBm is possible. There are two types of detectors, PIN and APD. For long links with high bit rates, the APD may become the choice. We would expect that the longer wavelengths would be used, but 820/ 850-nm links are still being installed. We must not forget reliability in our equation for choices. For lower bit rates and shorter links, we would give LEDs a hard look. They are cheaper and are much more reliable (MTBF).
9.5 COAXIAL CABLE TRANSMISSION SYSTEMS |
253 |
Example Link Budget Exercise. The desired bit rate is 140 Mbps. What will be the maximum distance achievable without the use of repeaters? The detector is a PIN type. Turn to Table 9.6 and determine that the threshold (dBm) for a BER of 1 × 10−9 is
−46 dBm at 1.3 mm. One EDFA is used with a gain of 40 dB. This now becomes the
starting point for the link budget.
The light source is a laser diode with −0.3 dBm output. The receiver threshold is
−46 dBm, leaving 45.7 dB in the power budget. Add to this value the EDFA gain of 40 dB, bringing the power budget up to 85.7 dB. We allocate this value as follows:
•Fiber at 0.25 dB/ km;
•Two connectors at 0.5 dB each or a total of 1.0 dB;21
•Fusion splices every kilometer; allows 0.25 dB per splice; and
•A margin of 4 dB.
If we subtract the 1 dB for the connectors and the 4-dB margin from the 85.7 dB, we are left with 80.7 dB. Add the splice loss and the kilometer fiber loss for 1-km reels, the result is 0.5 dB. Divide this value into 80.7 dB, and the maximum length is 160 km between a terminal and first repeater or between repeaters. Of course, there is one less splice than 1-km lengths, plus the 0.45 dB left over from the 80.7 dB. This will be additional margin. Setting up these calculations in tabular form gives the following:
|
ITEM |
LOSS |
|
|
|
Connector loss @ 0.5 dB/ conn, 2 connectors |
1.0 dB |
|
Margin |
4.0 dB |
|
Splice losses @ 0.25 dB/ splice, 160 splices |
40.0 dB |
|
Fiber loss, 161-km fiber @ 0.25 dB/ km |
40.25 dB |
|
Total |
85.25 dB |
|
Additional margin |
0.45 dB |
|
|
|
|
For an analysis of dispersion and system bandwidth, consult Ref. 7.
9.5 COAXIAL CABLE TRANSMISSION SYSTEMS
9.5.1 Introduction
The employment of coaxial cable use in the telecommunication plant is now practically obsolete, with the following exceptions:
•The last mile or last 100 feet in the cable television (CATV) plant; and
•As an RF transmission for short distances.
It is being replaced in the enterprise network with high-quality twisted pair and fiber optic connectivities. Certainly in the long-distance network, the fiber optic solution is far superior in nearly every respect. In the next section we provide a brief review of coaxial cable systems.
21These connectors are used at the output pigtail of the source and at the input pigtail to the detector. Connectors are used for rapid and easy disconnect/ connect because, at times during the life of these active devices, they must be changed out for new ones, having either failed or reached the end of their useful life.
254 CONCEPTS IN TRANSMISSION TRANSPORT
9.5.2 Description
A coaxial cable is simply a transmission line consisting of an unbalanced pair made up of an inner conductor surrounded by a grounded outer conductor, which is held in a concentric configuration by a dielectric.22 The dielectric can be of many different types such as solid “poly” (polyethylene or polyvinyl chloride), foam, Spiralfil, air, or gas. In the case of an air/ gas dielectric, the center conductor is kept in place by spacers or disks.
Historically, coaxial cable systems carried large FDM configurations, over 10,000 voice channels per “tube.” CATV (community antenna television or cable television) systems use single cables for transmitted bandwidths in excess of 750 MHz. Coaxial cable systems competed with analog LOS microwave and often were favored because of reduced noise accumulation.
9.5.3 Cable Characteristics
When employed in the long-distance telecommunication plant, standard coaxial cable sizes are as follows:
DIMENSION (in) |
DIMENSION (mm) |
|
|
|
|
0.047/ 0.174 |
1.2/ 4.4 (small diameter) |
|
0.104/ 0.375 |
2.6/ 9.5 |
|
The fractions express the outside diameter of the inner conductor over the inside diameter of the outer conductor. For instance, for the large bore cable, the outside diameter of the inner conductor is 0.104 in. and the inside diameter of the outer conductor is 0.375 in. This is illustrated in Figure 9.35. As can be seen from Eq. (9.27) in Figure 9.35, the ratio of the diameters of the inner and outer conductors has an important bearing on attenuation (loss). If we can achieve a ratio of b/ a c 3.6, a minimum attenuation per unit length results.
For air dielectric cable pair, e c 1.0 |
|
|
Outside diameter of inner conductor, c |
2a |
|
Inside diameter of outer conductor, |
2b |
|
Attenuation constant (dB)/ mic |
|
|
|
|
|
|
5 f |
|
a1 + b1 |
|
|||
|
|
|
|
f |
|
|||||
a c 2.12 × 10− |
|
|
|
|
|
(9.26) |
||||
|
|
|
log b/ a |
|||||||
where a c radius of inner conductor and b c radius of |
||||||||||
outer conductor. |
|
|
|
|
|
|
|
|
||
Chararacteristic impedances (Ω) |
|
|||||||||
138 |
b |
|
|
b |
|
|||||
Z c f |
|
log |
a c |
138 log a in air |
(9.27) |
|||||
e |
||||||||||
Figure 9.35 The basic electrical characteristics of coaxial cable.
22Dielectric means an insulator.
9.6 TRANSMISSION MEDIA SUMMARY |
255 |
Figure 9.36 Attenuation-frequency response per kilometer of coaxial cable.
The characteristic impedance of coaxial cable is Z0 c 138 log (b/ a) for an air dielectric. If b/ a c 3.6, then Z0 c 77 Q . Using dielectric other than air reduces the characteristic impedance. If we use the disks previously mentioned to support the center conductor, the impedance lowers to 75 Q .
Figure 9.36 illustrates the attenuation-frequency characteristics of the coaxial cable discussed in the text. Attenuation increases rapidly as frequency is increased. It is a function of the square root of frequency, as shown in Figure 9.35. The telecommunication system designer is basically interested in how much bandwidth there is available to transmit a signal. For instance, the 0.375-in. cable has an attenuation of about 5.8 dB/ mi at 2.5 MHz and the 0.174-in. cable, 12.8 dB/ mi. At 5 MHz the 0.174-in. cable has about 19 dB/ mi and the 0.375-in. cable, 10 dB/ mi. Attenuation is specified for the highest frequency of interest.
Equalization (i.e., the use of equalizers) will tend to flatten the response curves in the figure at the expense of some added loss per unit length. Equalization is defined by the IEEE (Ref. 6) as “A technique used to modify the frequency response of an amplifier or network to compensate for variations in the frequency response across the network bandwidth.” The ideal result is a flat overall response. The CATV plant makes wide use of such equalizers.
9.6 TRANSMISSION MEDIA SUMMARY
Table 9.8 presents a summary of the performance characteristics of the five basic transmission media.
256
Table 9.8 Characteristics of Transmission Media
|
|
LOS |
Satellite |
|
Coaxial |
Item |
Wire Pair |
Microwave |
Communications |
Fiber Optics |
Cable |
|
|
|
|
|
|
Bandwidth |
2 MHz to 400 MHz |
500/ 2500 MHz |
500/ 2500 MHz |
120 GHz per band |
Up to 1 GHz |
Commone bit rates |
1.544/ 2.048 Mbps |
155 Mpbs |
2.048 Mbps |
2.4/ 10 Gbps |
100 Mbps |
Achievable bit rates |
100 Mbps |
622 Mpbs |
155 Mbps |
200 Gbps |
1 Gbps |
Limitations |
Length limited |
By statute |
By statute; delay |
Severing cable |
Serving cable |
Applications |
LANs, TelCo |
Long-distance/ short- |
VSAT networks, |
For every broadband |
CATV last mile/ last |
|
outside plant |
distance links, |
long-distance links, |
terrestrial |
100 feet; RF |
|
|
TelCo and CATV, |
video transport |
application |
transport short |
|
|
private networks |
|
|
distances; |
|
|
|
|
|
otherwise |
|
|
|
|
|
limited |
Notes: Wire pair is distance limited. The shorter the pair length, the higher the bit rate. Also, balance and accumulating capacitance with length affect bit rate. LOS microwave is limited by statute, meaning by the ITU Radio Regulations and the national regulatory authority.
Satellite communication faces the same legal limitations. Geostationary orbit (GEO) satellites have long delays, which could affect interactive data systems. Only one GEO satel lite relay allowed for a voice connectivity.
The limits of fiber optics are still being explored. All terrestrial buried and aerial cable systems are vulnerable to severing by natural disaster or by man. Coaxial cable is limited by amplitude-frequency response characteristics. In nearly every instance, fiber optic cable connectivity is preferred.
|
|
|
REVIEW EXERCISES |
257 |
REVIEW EXERCISES |
|
|
|
|
1. |
For very-high-capacity transmission systems (e.g., >20,000 equivalent voice chan- |
|||
|
nel), what transmission medium should be selected? |
|
||
2. |
What are the advantages of using the RF bands from 2 GHz to 10 GHz for trunk |
|||
|
telephony/ data? Name at least two. |
|
||
3. |
Discuss the problem of delay in speech telephone circuits traversing a geostation- |
|||
|
ary satellite. Will there be any problem with data and signaling circuits? |
|
||
4. |
Give four of the five basic procedure steps in designing an LOS microwave link. |
|||
5. |
Where is earth curvature maximum on an LOS microwave path? |
|
||
6. |
In a path profile, what are the three basic increment factors that are added to obsta- |
|||
|
cle height? |
|
|
|
7. |
When a K factor is 4/ 3, does the radio ray beam bend away or toward the earth? |
|||
8. |
Name at least four parameters that we will derive from a path analysis to design |
|||
|
an LOS microwave link. |
|
|
|
9. |
Calculate the free space loss of a radio link operating at 4100 MHz and 21 statute |
|||
|
miles long. |
|
|
|
10. |
What is the EIRP in dBW out of an LOS microwave antenna if the transmit power |
|||
|
is 2 W, the transmission line losses are 3.5 dB, and the antenna gain is 36 dB? |
|||
11. |
A receiving system operating at room temperature has a 5 dB noise figure, and its |
|||
|
bandwidth is 2000 kHz. What is its thermal noise threshold? |
|
||
12. |
A receiver operating at room temperature in a digital LOS microwave link displays |
|||
|
a noise figure of 2 dB. What is its N0? |
|
||
13. |
What theoretical bit packing can we achieve with QPSK? With 8-ary PSK? |
|
||
14. |
What efficiency can we expect from an LOS microwave parabolic antenna bought |
|||
|
off-the-shelf? |
|
|
|
15. |
What is the cause of the most common form of fading encountered on an LOS |
|||
|
microwave link? |
|
|
|
16. |
What theoretical bit packing can be achieved from 64-QAM? |
|
||
17. |
Why is line-of-sight something more than line-of-sight? Explain. |
|
||
18. |
Based on the Rayleigh fade margin criterion, what fade margin will we need for |
|||
|
a 99.975% time availability? For 99.9%? |
|
||
19. |
There are two kinds of diversity that can be used with conventional LOS micro- |
|||
|
wave. What are they? It is nearly impossible for one type to be licensed in the |
|||
|
United States. Which one and why? |
|
||
20. |
An LOS microwave link cannot meet performance requirements. What measures |
|||
|
can we take to remedy this situation? List in ascending order of cost. |
|
||
21. |
Satellite communications is just an extension of LOS microwave. Thus a satellite |
|||
|
earth station must be within |
|
of the satellite. |
|
|
|
|
||
258 CONCEPTS IN TRANSMISSION TRANSPORT
22. There are two basic generic methods of satellite access. What are they? List a third method which is a subset of one of them.
23. What are two major advantages of satellite TDMA?
24. Define G/ T mathematically.
25. Why is the geostationary satellite downlink limited? Give two reasons.
26. Receiving system noise temperature has two components. What are they?
27. Name three applications for VSAT networks.
28. What sets a VSAT aside from conventional geostationary (fixed) satellite earth stations?
29. What sort of digital format might we expect on an outbound link for VSAT operation?
30. What is the great, overriding advantage of fiber optic communication links?
31. What are the two basic impairments that limit the length of a fiber optic link?
32. How does dispersion manifest itself on a digital bit stream?
33. Name the three basic components of a fiber optic link (in the light domain).
34. There are three wavelength bands currently in use on optical fiber networks. Identify the bands and give data on loss per unit distance.
35. What does a glass fiber consist of (as used for telecommunications)?
36. What are the two generic types of optical fiber?
37. Identify the two basic light sources. Compare.
38. What is a pigtail?
39. What are the two generic types of optical detectors? Give some idea of gain that can be achieved by each.
40. Where do we place fiber amplifiers (in most situations)?
REFERENCES
1. Transport Systems Generic Requirements (TSGR): Common Requirements, Bellcore GR-499- CORE, Issue 1, Bellcore, Piscataway, NJ, Dec. 1995.
2. Allowable Bit Error Ratios at the Output of a Hypothetical Reference Digital Path for RadioRelay Systems Which May Form Part of an Integrated Services Digital Network, CCIR Rec. 594-3, 1994 F Series Volume, Part 1, ITU Geneva, 1994.
3. R. L. Freeman, Radio System Design for Telecommunications, 2nd ed., Wiley, New York, 1997.
4. Performance Characteristics for Frequency Division Multiplex/ Frequency Modulation (FDM/ FM) Telephony Carriers, INTELSAT IESS 301 (Rev. 3), INTELSAT, Washington, DC, May 1994.
5. Performance Characteristics for Intermediate Data Rate (IDR) Digital Carriers, INTELSAT IESS 308 with Rev. 7 and 7A, INTELSAT, Washington, DC, Aug. 1994.
REFERENCES 259
6. IEEE Standard Dictionary of Electrical and Electronic Terms, 6th ed., IEEE Std. 100-1996, IEEE, New York, 1996.
7. R. L. Freeman, Telecommunication Transmission Handbook, 4th ed., Wiley, New York, 1998. 8. Telecommunication Transmission Engineering, 3rd ed., Vol. 2, Bellcore, Piscataway, NJ,
1991.
9. 1993 Lightwave Symposium, Hewlett-Packard, Burlington, MA, Mar. 23, 1993.
10. S. Shimada and H. Ishio, Optical Amplifiers and Their Applications, Wiley, Chichester (UK), 1992.
11. G. P. Agrawal, Fiber-Optic Communication Systems, Wiley, New York, 1992.
12. J. Everett, VSATs: Very Small Aperture Terminals, IEE/ Peter Peregrinus, Stevenage, Herts, UK, 1992.