11.1 Introduction

Distribution system design is, ideally, the process of specifying the most economical network that will provide the required bidirectional bandwidth to the required number of terminal points and still meet the performance goals. The performance goals, in turn, may be derived from internal quality standards, governmental regulations, the perceived requirements for reliable signal transportation, or all of the above.

Typical system specifications will include noise, distortion, response variation, delivered signal levels (and their stability), and hum. Reliability, another important consideration, is discussed in Chapter 20, and the effects of imperfect impedance matching among components are discussed in Chapter 15. The distribution system performance must also be considered as part of a chain that includes the headend channel processing and terminal equipment effects on signals. These are considered in separate chapters as well. Finally, the coaxial distribution system may well be cascaded with a signal transportation link such as microwave or, more frequently, a fiber-optic circuit. These, also, are treated in different chapters.

The final leg in the broadband distribution system is the drop wiring leading from tap ports to and throughout homes. By tradition, broadband networks are designed to meet certain goals from headend output to tap port, with an additional allowance made for the effects of in-home wiring. Modern drop systems, however, may vary widely in their complexity and quality, and in fact, may constitute minidistribution networks by themselves. The final section of this chapter deals with drops.

We begin by discussing basic principles and applying them to the “hard cable” portion of the plant.

11.2 Carrier-to-Noise Ratio

Noise is defined by the IEEE as “unwanted disturbances superposed upon a useful signal that tend to obscure its information content.”¹ In many communications systems, that is a good working definition. In cable television, however, the term noise is generally applied to thermal noise, which is treated independently from disturbances caused, for instance, by intermodulation products, ingress, or other causes. One of the reasons for treating thermal noise differently from other disturbances is that the subjective effects of noise and various other types of interference are different for analog video signals. Another is that, unlike many communications systems, the sources and mechanisms by which interfering signals are generated are well understood, and a detailed independent analysis of each has proved useful. Therefore, in this book, unless otherwise stated, noise will be used to refer exclusively to thermal noise.

In general, thermal noise is uncorrelated, that is, unpredictable and nonrepetitive. Noise generated in one device or at one time is totally unrelated to noise in another device or in the same device at another time although the average level of noise power may be quite predictable. Since that is the case, when the noise from two independent sources is combined, the total noise power will be the sum of the original noise powers.

Carrier-to-noise ratio (C/N)is defined as follows:

(11.1)

where с and n are the scalar power levels of the carrier and noise, respectively. For analog television signals, such as NTSC, the carrier power is defined as the rms power that occurs during the sync tip (the highest power level in the modulated signal). Noise is defined as the rms noise power measured in a defined bandwidth (4 MHz for NTSC). Equation (10.19) gives the C/N for a channel transmitted through an individual amplifier, given the input signal level and noise figure.

In order to design a coaxial distribution system, it is necessary to compute the C/N of cascaded amplifiers, each followed by loss equal to the gain of the preceding amplifier. For two amplifiers, the general expression for composite C/N is approximated by the expression

(11.2)

where C/N₁ and C/N₂ are the C/N of each of the amplifiers calculated independently using Equation (10.19) (for example, when driven by a noise-free input signal). Although this general expression is useful when cascading two dissimilar network elements, a simpler approach is useful when cascading n identical amplifiers, separated by passive network losses equal to the amplifier gain:

(11.3)

where

These equations are quite useful, but their limitations must be emphasized.

The preceding conditions are required because the equations do not rigorously account for the effects of thermal noise though they are adequate for analyzing typical amplifier cascades encountered in cable television systems. For other applications, a more precise set of equations is recommended.

In a simple repeatered transmission line, the designer must “juggle” three parameters to reach a required end-of-line C/N: the noise figure of individual amplifiers, the input level to each amplifier, and the number of amplifiers to be cascaded. In more complex networks, the same variables apply, but the C/N “budget” may first have been allocated among several dissimilar network sections. As will be seen, noise considerations must be balanced against other requirements in reaching an optimal design.

11.3 Carrier to Distortion

Unlike the thermal noise, which is a function of only the transmission system characteristics and channel bandwidth, the distortion level depends on the signals carried. The distortion level of amplifiers is typically specified under one or two specific channel-loading conditions. Although it would be possible to simply specify distortion with just two or three carriers, the translation to expected IM products under the unique channel loading used in cable television is not straightforward, so vendors have resorted to testing under conditions that are easily related to actual field applications.

As discussed in Chapter 10, the channel loading under which amplifiers are typically tested and specified (for CTB and CSO) consists of independently generated, unmodulated carriers at each of the visual carrier frequencies specified in the ANSI/EIA-542 Standard frequency assignment list that falls within the bandwidth of the device. In addition, a portion of the bandwidth may be loaded with signals that simulate the effect of digitally modulated signals (often just band-limited white noise at a controlled level). Testing of XMOD is similar except that the carriers are simultaneously modulated.

The results of such testing do not match the distortion when loaded with analog television carriers (whose sync peak carrier levels match the unmodulated carriers), but the results are much more repeatable, and there are standard assumptions regarding the approximate improvement when loaded with modulated carriers: 6 dB in the case of CSO and 12 dB in the case of CTB and XMOD, though the improvement will be somewhat less with short coaxial cascades.

The buildup of distortion as amplifiers are cascaded is less straightforward than in the case of noise. If the deviations from linear response are not predictable from amplifier to amplifier, then the distortions can be considered uncorrelated, in which case they will add in the same way as noise; that is, the distortion products will increase by 10 log n, where n is the number of amplifiers. On the other hand, if every amplifier has the same general distortion shape (for instance, a flattening of the curve at voltage peaks), then the distortion will be highly correlated, and the magnitude of the products will increase at a 20 log n rate.

Since third-order distortion generally does result from symmetrical curve flattening and both CTB and XMOD arise from third-order distortion, most designs assume that these distortions increase at the 20 log n rate although that may be overly conservative when evaluating cascades of amplifiers with dissimilar characteristics or those that nominally cancel third-order distortion (for example, feed-forward amplifiers). Second-order distortions, on the other hand, are nominally canceled in the universally used push-pull gain modules. For that reason, the residual second-order distortions are not highly correlated, and many designs are based on a compromise 15 log n buildup.

As with cascades of identical amplifiers, the buildup of distortion when cascading different network sections depends on the degree to which the distortions are correlated.² In general,

(11.4)

where

Where identical amplifiers operating at the same output levels are cascaded, the expression for C/CTB can be reduced to

(11.5)

where

whereas the equivalent expression for CSO is

(11.6)

where

11.4 Noise-Distortion Trade-Off

The interplay between noise and distortion can be readily illustrated by considering a cascade of n identical amplifiers. It can readily be understood that the loss between each pair of amplifiers must be made identical to the gain of one amplifier for optimum performance. If the loss is less than that value, then each amplifier’s input level (and, hence, output level) will be greater than the previous amplifier, and the distortions will quickly build up to a high level. If the loss is greater than the gain, then the input of each amplifier will be less than the previous amplifier and thus contribute disproportionately to the overall C/N degradation.

Given equal gains and losses, let us assume a requirement that the cascade provide an analog television C/N of 50 dB and a C/CTB of 65 dB. (We could have considered any other distortion parameter as easily, but this will serve to illustrate the trade-offs involved.) Let us also assume that each amplifier has a noise figure of 8 dB, a gain of 22 dB, and a C/CTB of 80 dB at the required channel loading and at an output level per carrier of +40 dBmV. (Again, to make it simple, we will assume all carriers have the same level and the loss between the amplifiers is the same at all frequencies.)

If our cascade consists of just a single amplifier (n = 1), then we can calculate the minimum input level that will assure an adequate C/N. From Equation (10.19):

(11.7)

With a station gain of 22 dB, this corresponds to an output level of +20.8 dBmV.

Similarly, since CTB changes 2 dB for every 1-dB change in operating level, we can readily determine that we could increase the output operating level to +47.5 dBmV and stay within the distortion limit. Thus, the output level could vary over a 26.7-dB range while maintaining the required performance.

If additional amplifiers are cascaded, the effective noise figure increases by 3 dB for every doubling of the cascade (Equation (11.3)), so the minimum C_i (and thus output level) must increase by that amount to provide the same end-of-line C/N. Similarly, the distortion increases by 6 dB for every doubling of the cascade, so the maximum allowable output level must decrease by 3 dB to maintain the same end-of-line distortion levels. Thus the allowable range of signal levels decreases by 6 dB.

Table 11.1 shows the allowable output level range as a function of cascade for this example.

Table 11.1 Example of cascade effects on allowable amplifier operating levels

If we plot these general relationships as a function of cascade, we can see that there is a maximum attainable cascade and a unique operating level that allows that cascade to be realized. Figure 11.1 illustrates the usable operating range, along with the parameters that define the noise-distortion-cascade relationship. Complete system design requires that this calculation be performed for both second- and third-order distortion, as well as noise, and to pick operating levels that allow all specifications to be met. Generally, composite intermodulation noise (CIN) caused by digital signals is quantified as an equivalent increase in noise level that must be figured into the C/N calculation.

Figure 11.1 Relationship of cascade, noise, distortion, and levels.

Another way of looking at the interaction between parameters is to consider the case of a hypothetical tapped distribution line that just meets end-of-line noise and distortion specifications, as well as required signal levels at subscriber terminal equipment. The operator desires to increase the bandwidth and channel loading of this system without degrading performance.

At the increased upper frequency limit, the cable loss is higher. The designer can compensate for that loss by using higher-gain amplifiers, spacing amplifiers closer together, or reducing the per-foot loss of the interconnecting cable. Additionally, the increased number of carried signals will increase the distortion products in each channel if the per-carrier level is held constant. Finally, the loss of subscriber drop wiring will increase, requiring higher tap levels or lower-loss cables.

If higher-gain amplifiers are used, then either output levels must be raised or input levels lowered. Absent a change in technology, the former will further increase distortion, whereas the latter will cause a greater C/N degradation in each amplifier station. If amplifiers are spaced more closely together, then more will be required to reach the same physical distance, thereby increasing both noise and distortion. Even if amplifier spacing is retained, either amplifier output levels must be raised or drop cables replaced to maintain adequate levels at subscriber equipment at the highest channels. Absent changes in technology, the designer is faced with an insurmountable obstacle.

Several techniques can be used to get out of this “box”: (1) amplifiers with lower noise figures and/or distortions can be used; (2) both distribution and drop cables can be replaced with larger, lower-loss versions; and/or (3) methods can be used to transport signals to points closer to homes so that coaxial cascades can be shorter. Generally, upgrades to older all-coaxial systems employ elements of all three techniques. In particular, fiber optics or microwave is used as a supertrunk to transport signals throughout large service areas with less noise and distortion than reasonably possible with coaxial techniques.

11.5 System Powering

As discussed in Chapter 10, coaxial distribution systems are usually powered by multiplexing 50- or 60-Hz power with the signals and transporting both through a common cable. Through use of separate signal paths through amplifier stations and passive devices, power routing is configured independently from the signal paths.

Coaxial distribution systems may be powered from one end (centralized powering) or via power supplies located along the cable route (distributed powering). In the latter case, it is common to locate the power supply midway through the section of plant it supplies so that roughly half the current can flow in each direction from the insertion point, reducing both the current capacity requirement for components and resistive losses in the interconnecting cables.

The reliability of distribution system power has a major effect on overall reliability. Therefore, systems designed for high availability are generally built with coaxial distribution networks that are centrally powered with high-capacity standby supplies. Chapter 20 covers reliability and availability.

Figure 11.2 is a simplified schematic of a section of cable and several amplifiers extending one direction from a power supply. From a network designer’s point of view, the challenge is to minimize the number of power supplies required or, put another way, to determine the maximum number of amplifiers that can be powered from each supply.

Figure 11.2 Example of amplifier powering.

Given a decision on system supply voltages, several parameters cannot be exceeded in making this determination: (1) the current rating of the supply, (2) the current rating of any coaxial component along the route, and (3) the minimum usable voltage at any powered component (typically around 40 volts for 60- or 90-volt systems).

As previously discussed, most modern amplifiers use switching power supplies. These draw current in approximately inverse proportion to the supply voltage so that the total power drawn is a constant. This allows them to maintain high efficiency and low power dissipation. It also complicates the power calculation.

Figure 11.3 is a redrafting of the first seven amplifiers in Figure 11.2 to show the power component interaction. Though it would be possible to set up simultaneous equations for each powering situation and solve for all the critical voltages and currents, a successive approximation procedure is generally used for both manual and computer calculations. The process is more or less as follows:

Figure 11.3 Equivalent circuit of powered coaxial network.

The actual calculation is very easy on a computerized system. The “black magic” is in strategically placing power supplies and determining network power boundaries. Not only is it important to minimize the number of power supplies, but from a reliability standpoint, it is important to minimize the number of cascaded supplies that affect any given subscriber. Finally, it is important to locate power supplies where the commercial power is most reliable.³

A technique sometimes used to overcome both the maximum component current ratings and the voltage drop in the high-current sections near the power supplies is to add a second cable extending from power supply locations to additional strategically located power insertion points. Since these cables do not carry RF, they can be of lower characteristic impedance and, thus, use larger center conductors. For instance, the loop resistance of standard 0.625-inch-diameter P3 cable is 1.1 ohms/1,000 feet, whereas the same-size “power feeder” cable is 0.290 ohms/1,000 feet.

11.6 Signal Level Management

The requirements for level management are very different in the forward and reverse directions. In the downstream or forward direction, each amplifier receives signals from only a single source whose output is well defined. The design and operational requirements are constant amplifier output levels and adequate amplifier input levels. Additionally, to ensure that well-designed end-of-line receivers work satisfactorily, several parameters must be controlled: adequate signal strength to overcome receiver noise, avoidance of excess levels that could lead to overload, limits on adjacent channel level differences, limits on total instantaneous variation across the spectrum, and limits on changes with respect to time. These are more fully discussed in Chapter 24.

As discussed in Chapter 10, manufacturers typically specify performance with specific channel loadings and amplitudes. Since digital formats are generally more robust than analog television, it is common to specify performance under two conditions: a full spectrum of analog television signals and a defined mix of analog television and digital signals, where the digital signals are operated at a relatively lower level (as illustrated in Figure 10.19). Note that the net gain of the network, as measured from the output of one amplifier to the output of the next amplifier, is 0 dB in either case. The depressed levels of the digital signals relative to the analog signals are set in the headend where they are combined.

To the extent that the positive gain slope built into amplifiers does not exactly compensate for the losses between amplifiers, equalizers are used to eliminate the remaining variability. To the extent that interamplifier losses (including any equalizer losses) are less than amplifier gain, attenuators (also known as pads) are used to set the gain block input level. Pads may be placed before the first stage (generally, the only option with line extenders) and sometimes also between stages. If the latter option is available, then in cases where excess signal level is available, operators can independently set the operating levels of input and subsequent gain blocks to optimize noise-distortion tradeoffs. With amplifiers that offer multiple independent output stages, the interstage pads also allow setting of different output levels for different ports. Thus, a port that directly feeds a tapped line may be set to a higher level than one that will feed additional amplifiers and therefore needs to run at a lower distortion level.

In tapped lines, both distribution cable losses and drop cable losses must be considered. It is important that adequate and relatively uniform levels be delivered to terminal equipment (see Chapter 24). A 200-foot size 6 drop cable has about 8 dB greater loss at 750 MHz than at 5 4MHz. This differential loss tends to offset some of the intentional tilt in amplifier outputs if connected to a tap that is close to a station. On the other hand, if it is connected to the last tap in a string, it will aggravate the situation if the slope is already negative (high-end channels below the low-end channels in level). To prevent excessive level variation, in-line equalizers are sometimes inserted in tap strings. As a general rule, operators try to limit analog video channel level variation across the spectrum to about 10 dB. Typical design rules call for +15 to +20 dBmV levels at tap ports on the highest channels.

Signal level management in the upstream direction is much more complex and must be treated as an overall system problem rather than as just part of the design. Signals may originate from anywhere in the system and may be of various types so that the channel loading and levels may vary continuously at any given amplifier.

The preferred approach to upstream level and gain setting is to align the system from headend out so that the input level to each upstream amplifier is consistent; that is, if a level of x dBmV is inserted into the input of any upstream amplifier, the level received at the headend will be the same. Pads (and, if required, equalizers) are placed in the output of the gain modules to set overall system gains properly.

This does not, however, mean that the required upstream transmitter output levels for terminal equipment will be consistent. Consider Figure 11.4, which depicts a typical tapped line. Following normal practice, tap values are selected to produce approximately the same downstream level at the highest channel at each home. The circuit losses depicted are based on typical manufacturers’ specifications for four-port taps, the loss of 250 feet of .500-inch type P3 cable between the taps, and the loss of 100 feet of size 6 drop cable connected to each tap port. As can be seen, the downstream levels at the end of the drops are all within 1 dB.

Figure 11.4 Signal levels in tapped distribution line.

If we calculate the insertion levels required to reach the return amplifier, however, we see a nearly 20-dB difference due to the differential loss of the cable and the taps. Rather than manually set the level of each upstream transmitter, most manufacturers sense the level reaching the headend and remotely control each device independently to set the appropriate level. Another approach is to equalize each drop to approximately compensate for the differential cable loss and equalize required levels. As of the writing of this book, there is no uniformly accepted practice in this area.⁵ Chapter 16 deals with the numerous unique issues related to upstream transmission.

11.7 Signal Level Stability

As was discussed in Chapter 10, the attenuation of coaxial cable varies about 1% per 10°F. The losses of other components, as well as amplifier gains, also vary somewhat with time and temperature. Amplifier manufacturers offer two options in their equipment that are designed to compensate these variations: thermal compensators and automatic gain and/or slope control circuits (AGC/ASC).

Thermal gain compensation is open loop, meaning that the gain adjustment is only an approximation to the variation in interconnecting losses with temperature. AGC circuits work by sensing the level of one or more of the carried signals and controlling the gain of the station so as to maintain a constant output level for that signal. Automatic slope controls sense two signal amplitudes and control both the overall gain and the gain slope through the stage so that the levels of both signals are constant at the output port. Typical control ranges are 6–8 dB. Use of AGC and/or ASC circuits depends on the accuracy of headend level setting. Also, if one of the reference carriers fails, the entire distribution system can change gain sufficiently to cause serious distortion products.

The loss between amplifiers may be all cable or a combination of cable and passive devices. In the extreme case of an all-cable link whose loss is 35 dB, the loss will vary over a range of about 5.2 dB over a range from +110°F to −40°F(+ 43°C to − 40°C). Thus every amplifier will require some sort of compensation in a trunk network.

In a tapped distribution line or one that is split several times, half or more of the loss may take place in passive devices. Thus, thermal compensation may be required only in every second or third amplifier. The trade-off is that compensation networks may have loss that reduces available gain, whereas level variations will affect C/N, CTB, and CSO.

11.8 The Service Drop

So far we have dealt primarily with the portion of the distribution network that lies between the headend output and the tap ports that serve individual customers. Generally, network designers design to specifications for this portion of the network, assuming certain average performance in the headend and drop systems such that the end-to-end performance will meet regulatory requirements as well as subscriber expectations. Chapter 2 dealt with subscriber quality expectations, and Chapter 15 will cover end-to-end performance. This section will deal with the variation in drop subsystems and suggest the performance degradation that might be expected in a “standard” drop that must be taken into account in calculating end-to-end performance.

As a practical matter, in the United States, as in many other countries, all or most of the drop system may not be designed, built, or maintained by the network operator. Thus, although designers must be aware of the variation in drop performance, little can be done to control quality or design within homes.

11.8.1 Simple Passive Architectures

At its simplest, the drop will consist of a flexible (either overhead or underground) cable from the distribution tap port to the building entry point, a connection to the building ground system at that point, and another flexible cable within the building that leads to a user’s receiver, such as a television set or VCR. If there are multiple receiver locations, a signal splitter is used to divide the signal to feed more than one outlet.

If the loss in the interconnecting cables plus splitters is too great to assure adequate signal level at receiver input ports, then either (1) higher tap levels must be provided, (2) lower-loss drop cable must be used, or (3) amplification must be provided in the drop system. As discussed earlier in this chapter, tap signal levels generally are difficult to increase because of the interaction among levels, noise, and distortion in the distribution network. That leaves the builder of the drop with two practical choices: reduce loss or provide gain.

In the early history of cable systems, almost all drop wiring was done with size 59 cable (originally known as RG59). As bandwidths have increased beyond 300 MHz, common practice has been to upgrade the portion of the drop between tap and bonding (ground) block (the point where the cable enters the dwelling) to at least size 6 and, as required, size 7 and even size 11. The shift from 59 to 6 is easily justified by savings in the hard cable plant due to lower required tap levels. The larger sizes, however, are considerably more expensive and, in the case of size 11, require special connectors that add to both hardware cost and installation labor. Even in systems with size 6 outside drop cables, the inside wiring is generally size 59 since it is easier to handle and less obtrusive visually where exposed. Common practice is to design the hard cable plant to have drop levels that will accommodate most situations with size 6 exterior and size 59 interior cables, and to use larger cable sizes, both inside and outside, only where required because of exceptional drop lengths and/or a high number of splits.

A Reference Drop Structure

Although there are many single-receiver households, it is increasingly common to find two (or more) television outlets in cabled homes, and therefore a splitter that may be mounted at the entry point to the dwelling and does double duty as a bonding block. Given that, we can define a standard drop (Figure 11.5) that will encompass the majority of current home wiring. It will include 25–100 feet of size 6 cable from tap to house, perhaps a splitter, and size 59 connections to outlets, with lengths varying from 10 to 50 feet.

Figure 11.5 Typical drop configuration.

Passive Drop Transmission Characteristics

In order to assess the transmission characteristics of this standard drop, we need to first predict the quality of the components used, then use the tools developed in Chapter 10 to calculate their effect on the signals.

Component Quality. A prime determinant of the degree of degradation signals experience in the drop system is the quality of the components used. Unfortunately, just as there are no mandatory standards on drop wiring or design, so there are none for the components used. Historically, drop wiring was installed and maintained by cable operators (much of it still is), and they purchased reasonable quality components because it was less expensive than the trouble calls and customer dissatisfaction resulting from inferior materials.

As of the writing of this book, the Society of Cable Telecommunications Engineers (SCTE, an ANSI standards-making organization) is actively developing voluntary national standards covering virtually all drop components: cable, connectors, splitters, couplers, amplifiers, attenuators, filters, and so on.⁶ Covered performance parameters include return losses, transmission losses, environmental requirements, shielding, surge protection, labeling, and a host of others. The specifications for amplifiers additionally include gain, noise, and distortion. Of importance in estimating drop performance, most devices must exhibit return losses of at least 18 dB, and excess loss (that is, loss exceeding that expected from the power division ratio) in splitters and couplers may not exceed 1.0 to 1.2 dB (for instance, the loss through a two-way splitter must be no greater than 4.0 dB) though in some cases, they can rise above 2 dB. Port-to-port isolation in splitters must be at least 20 dB. Cable-shielding requirements are consistent with good-quality double-shielded cables. There is no assurance, of course, that existing products all meet the proposed standards or that even conforming products will continue to do so after years of exposure to weather.

Furthermore, network operators can hardly be assured that components installed by a homeowner will meet these voluntary, future standards. In particular, a conservative estimate is that return losses from components may be as poor at 10 dB, port-to-port splitter isolation may be as low as 10 dB, and cable losses may, after aging if not originally, exceed nominal values (Figure 10.5) by 10%.

Drop Cable. The characteristics of various types of drop cable were discussed in Chapter 10. Within the basic size options, however, are many quality and feature options. Chief among those are shielding integrity. Cables are available with a standard dual shield (foil and braid), triple shield (foil, braid, and foil) and quad shield (foil, braid, foil, and braid), with increasing effectiveness of shielding. Of serious concern to operators are dual shield (or worse) cables with sparse braid structures that are available in the retail market. Some of these poorly shielded cables do not adhere to SCTE tolerances for dimensions either, with the result that connectors are poorly attached, both electrically and mechanically, to the cable.

Aside from shielding options, cables are available with support messengers, with flooding compound (to prevent water from migrating through the cable structure), in a dual configuration (for dual cable distribution networks), and paired with one or more twisted copper pairs for carrying telephony or power along with RF in the same structure.

Finally, the nonmetallic portions of the cable are available with various degrees of flame resistance to meet electrical and fire codes when they are routed within some interior paths (such as plenums).

F Connectors. Whatever other components may be included in the drop, in North American cable systems, there will be at least four type F connectors — one at the tap, two at the bonding block, and one at the consumer’s receiver input. This connector, designed for low-cost mass manufacturing, is at the root of many drop problems. The female connector uses a nonprecious-metal, two-point spring clip contact for the center conductor; the male cable connector is a feed-through type that utilizes the center conductor of the cable itself for a pin; and the mated pair depends on a threaded joint plus a butt contact for an outer RF connection. The result is a connector family that is extremely craft sensitive in installation. Despite ANSI standardization,⁷ and detailed electrochemical studies,⁸ F connectors remain the leading cause of trouble calls and signal leakage in many cable systems.

In cases where power is multiplexed with RF signals on drop cables (typically to power point-of-entry telephone or data terminal equipment), the limitations of F connectors are exacerbated. A less-than-perfect contact that has a reasonable capacitance may pass RF signals but will not pass power. Furthermore, a loose contact, whether in the center conductor or shield, may generate electrical arcing when cables move with temperature changes or wind. These transients generally will have frequency components extending through the upstream band and into the lower downstream channels. Since they occur within the cable, they are tightly coupled to the transmission line in both directions and can cause severe transmission problems. The effects of electrical transients on upstream transmission are discussed in detail in Chapter 16.

In addition to inherent problems with the quality of the mating surfaces, male connectors vary in how they are attached to the flexible drop cable. Until about 1990, the most common construction used a sleeve that was pushed between the foil and braid of the cable. It was retained by use of a separate hexagonal crimp ring. Improved designs used a heavier, integral crimp ring. With both designs, however, field experience showed that the retention strength of the connector on the cable reduced over time. More recent designs have featured a full annular press-fit connection that has proved to retain its strength.

The use of a feed-through design (where the cable center conductor forms the connector pin) with more than one size of cable has created problems with the mating connectors. Even with highly compliant female contact structures, once a female connector has been mated with a large cable center conductor, its ability to form a reliable contact with a small cable is reduced. The center conductor of size 11 cable is beyond the capacity of standard female ports so that special pin-type male connectors are used with it. Pin connectors are also sometimes used with size 7 cable to protect mating ports.

Traps. In addition to the normal drop components, some combination of bandpass, band-stop, low-pass, or high-pass filters are often inserted into the drop circuit at the tap to control which signals reach subscribers’ receivers. The techniques used to restrict and tailor services are discussed in Chapter 18. However they are used, traps add loss. Generally, their out-of-band loss may be only a few tenths of a decibel, but their loss near cutoff frequencies can be as much as several decibels at the band edge of adjacent channels, coupled with a considerable slope through the channel. Where traps are used to control several channels individually, as many as four or five may be cascaded between tap and drop cable.

Transmission Loss. Given these variables, the total loss from tap to house will generally fall within the values given in Table 11.2, where the minimum losses result from short drop cables without a splitter, and the maximum losses are based on long drops, splitters, and two to three in-line traps.

Table 11.2 Tap to receiver loss ranges for typical passive drop systems

	Expected Loss Range
Frequency	Minimum	Maximum
5 MHz	0.2 dB	9.5 dB
40 MHz	0.5 dB	10.5 dB
54 MHz	0.6 dB	13 dB
550 MHz	1.8 dB	18.5 dB
750 MHz	2.1 dB	20 dB

System designers must take into account this loss range in setting signal levels, and allowable level variation, at system taps. For example, in the “maximum” loss configuration, note that the difference in loss is 7 dB over the downstream bandwidth of a 750-MHz system. If the flatness of the signal at the tap port has a downward slope with frequency, the combination could result in widely differing signal levels being presented to receiver input ports.

Microreflections, Group Delay, and Frequency Response Variation. Characteristics of terminal equipment are dealt with in depth in Chapters 23 and 24. For now, it is sufficient to state that cable operators are required to deliver signals directly to consumers’ television receivers under some circumstances (where analog channels are unscrambled and delivered on standard channels) and that the return loss of those receivers is essentially zero on untuned channels and may be as poor as 4–8 dB on the tuned channel. In any event, it is likely that an unused outlet will be left unterminated, so it should be assumed that 100% reflection occurs at each subscriber outlet. This will cause reflections whose delay and magnitude will depend on the isolation of the splitter ports and the length of the cables beyond the splitter. Using the methods discussed in Chapter 15, we can estimate the worst-case reflections. Figure 11.6 shows the results at various frequencies. Although operators should be prepared for occasional reflections of those magnitudes, a statistical study of actual cable systems showed 95% of all echoes, including those in the distribution plant, were more than 25 dB below carrier levels in magnitude.⁹

Figure 11.6 Worst-case microreflection magnitudes and delays in typical house wiring.

These reflections can lead to bit errors in digital transmissions since delayed signals cause successive transmitted symbols to overlap at the receiver and visible degradation of analog television pictures. As will be discussed in Chapter 15, the effects also include group delay and passband response variations. Amplitude response variations exceeding 1 dB in each channel are quite possible.

Signal Leakage and Ingress. Most difficult to assess is signal leakage. In practice, most leakage and ingress is not associated with components or cable but rather is due to inadequate shielding in customer-owned terminal equipment and/or improperly installed or tightened drop connectors. Poor-quality connectors and poorly shielded cables may also contribute problems.

Given the statistical nature of both egress and ingress, there is no way of predicting their effects. The effect of shielding of consumer equipment is discussed in Chapter 24. The effect of ingress on upstream communications is treated, along with other upstream transmission issues, in Chapter 16.

11.8.3 Integrated In-Home Wiring Systems

In response to a need to have defined performance levels and a desire to interconnect home video equipment and multiple program sources, several industry groups have designed integrated home wiring systems. Although a complete discussion of these is beyond the scope of this book, one example of such a system is the CEBus.

The Consumer Electronics Manufacturers Association (CEMA) of the Electronic Industries Association (EIA) has developed a complex home wiring standard under the umbrella title CEBus. CEBus supports message and information transmission over a variety of physical paths within homes, each with its own special label: twisted pair wiring (TPBus), subcarriers on the power wiring (PLBus), coaxial cable (CXBus), infrared through the air (IRBus), over-the-air transmitted low-power RF (RFBus). A separate, but related, specification covers direct interconnection of home audio and video equipment using specially designed multiple twisted pair cables (AVBus). These standards have been under development since the mid-1980s.¹⁰

A common language is specified for medium-independent signaling of messages and control information, whereas routers interconnect different types of media. The language used for signaling is known as the common application language (CAL) and is defined by ANSI/EIA Standard IS-60.

Of greatest interest to broadband network operators, the CXBus distributes both externally supplied signals (such as from a cable system) and internally generated RF signals over dual, size 6 drop cables through the residence. Down-stream signals from the external network are amplified, optionally combined with internal signals, then split to feed four or more “branch circuits” over one of the paired distribution cables. Each branch cable may feed a four-way splitter with shorter cables from there to individual termination points. Thus, one network may incorporate as many outlets as permitted by the cable losses and available RF power (see Figure 11.7). Note that the splitters may be internal to Node 0 and that this is, in fact, the preferred topology since all network elements and connectors are then accessible for maintenance or inspection, if required.

Figure 11.7 Basic CEBus coaxial topology.

The downstream bandwidth of the distribution system (the “external” cable in CEBus parlance) extends from 54 to more than 800 MHz, so as to allow direct transmission of UHF signals from local antennas, if desired (switching between the external network and local antennas is provided in the design). Upstream signals are carried in the 5–30-MHz band and are combined and amplified before insertion into the external network. The bandwidth from 450 to 546 MHz may optionally be used for downstream transmission of video from local sources to television receivers. The second (“internal”) cable has a more evenly split bandwidth and is used for routing of both messages and video among various consumer electronics devices. Data (information) channels are defined in 1.5-MHz bandwidth slices, with four contiguous channels required for internal transmission of normal analog video. Physically, the two cables are run together, with dual type F wall outlets provided for subscriber access.

All the signal processing is done at “Node 0.” Figure 11.8 shows the signal processing within the node. Locally generated signals that will stay within the dwelling may be inserted into any internal cable outlet. The block converter is used to convert the frequencies of the internally generated signals from lower upstream channels to downstream channels for redistribution. The output cable selector determines whether the converted channels are distributed over the internal or external cables.

Figure 11.8 CEBus Node 0 functionality.

Locally generated signals that are intended for transmission through the external cable system can be inserted into any external cable outlet. Those signals will be amplified by the return amplifier and transmitted into the cable drop.

The control channel regenerator is an optional feature that allows control signals to travel both directions between coaxial and twisted pair distribution networks.

The CXBus is a well-designed specification that should assure proper level control, minimal signal leakage or ingress, and well-controlled noise and distortion degradation. Nevertheless, there are important issues when external network signals are transmitted over such a network:

• The network has fixed bandwidth boundaries, both downstream and up-stream. If a network operator chooses to use frequencies outside those bandwidths, the signals will not reach subscriber outlets.

• As we discussed, the downstream and upstream amplifiers contribute both noise and distortion that must be accounted for in the end-to-end performance budget. Since only a few homes may be equipped with CEBus systems, it is unrealistic to upgrade the entire external network to accommodate them.

• Upstream levels may vary widely in power level (see Chapter 16). Therefore, the return amplifier must have a very wide dynamic range.

Despite these concerns, it is expected that CEBus systems will generally interface with most current cable television network with minimal problems.

11.9 Summary

Cable television coaxial distribution systems consist of a cascade of cable and amplifiers. The network typically branches in the downstream direction, and various passive RF devices are used both within and external to amplifiers to create the legs. The amplifiers are capable of amplifying signals flowing in both directions, with the downstream bandwidth extending from approximately the lowest broadcast television channel (54MHz in the United States) to an upper limit determined by the bandwidth needs of the network, but typically 400–870 MHz. The upstream bandwidth typically extends from 5 MHz to somewhere between 30 and 42 MHz.

In the downstream direction, distribution networks with more than a few amplifiers in cascade are usually designed and operated so that the gain, measured from amplifier output to amplifier output, is unity. When systems carry only one signal type (for example, analog video signals), their levels are often adjusted so that they increase linearly with frequency, as measured at amplifier output ports. This usually results in amplifier input levels that are approximately the same across the spectrum and creates the optimum balance between noise and distortion. When different signal formats are carried, levels may be adjusted to a different pattern of levels. A common example is digital signals carrying packetized data or video, which are operated 6–10 dB lower in level than equivalent analog video signals would be at the same frequencies.

Amplifier gains must be sufficient to offset the loss of the intervening cable and passive devices at the highest operating frequency. A combination of attenuators, positive gain slope with frequency, and special response-shaping networks are used to just offset interamplifier losses. The last few amplifiers may be operated at higher levels to allow feeding more subscriber taps between gain blocks.

A distribution network will exhibit an overall downstream carrier-to-noise ratio that is inversely proportional to the logarithm of the number of amplifiers and proportional to the signal levels at the inputs to each amplifier. It will generate distortion products that are primarily determined by the amplifier output levels, as well as the number of amplifiers. The bandwidth, operating levels, noise, and distortion are interdependent and, as a set, describe the essential features of the system. Other important characteristics include hum modulation, reflections, frequency response, and gain stability.

Power is carried to amplifiers over the same cables used for signals. Most commonly, 60–90 VAC, 60-Hz quasi sine waves are used in the United States. In the event of commercial power failure, standby power supplies replace the sine waves with trapezoidal waves. The saturation of magnetic circuits used to separate 60-Hz power from signals at each device is the largest contributor to hum modulation.

Cable television drop systems can be as simple as a single coaxial cable from tap to one television receiver, or as complex as a full, integrated distribution system with multiple video sources and receivers. Today, the vast majority of cable drop configurations consist of a cable from tap to home, an optional splitter, and additional cables to one to four individual outlets.

The loss (and its variation) over this network, along with reflections from the termination points and shielding integrity, will have an effect on the overall transmission quality to consumers’ receivers. In more complex networks that include amplification, the noise, distortion, gain stability, group delay, hum modulation, and so on must be included in end-to-end performance calculations. In the case of complex distribution systems, spectral usage by internal video sources may affect compatibility with the external network.

Endnotes