Broadband Cable Nets Serve Business Needs.

The Community Antenna Television (CATV) system had its inception in the early 1950s. As originally conceived, the CATV system simply provided basic limited channel TV service to restricted off-the-air reception areas. The complete CATV system commonly included three basic parts: a master antenna system (tower, preamps and appropriate antennas); a headend system (single-channel processors and modulators); and a broadband distribution system (trunk and feeder). Initially, the headend and broadband distribution systems provided for one-way only transmission; however, as operating techniques developed and equipment technology advanced, operators promoted and found interest in two-way transmission. The intent of this article is to familiarize the reader with the types of two-way systems in use and the system components used to operationally construct a two-way broadband RF distribution system.

Presently, there are three specific types of two-way broadband RF distribution systems in use. The generic name for a two-way RF distribution system is split-band system; however, specific names are more commonly used for the three system types noted above. These systems are called sub-split, mid-split and high-split systems, where the specific name refers to the frequency-band division for forward and reverse transmission. Although there are differences between the various equipment manufacturers as to forward and reverse frequency band edges, in general, the bands occupy approximately the same relative frequency blocks in the spectrum and loosely fall under the sub-split, mid-split and high-split classifications.

Of the three two-way systems noted above, the sub-split type is by far the most common in operation. The sub-split system was cable TV's first attempt at two-way operation and was developed initially because it easily fit into the frequency plan of the CATV system as it was originally conceived.

The original CATV subscriber system forward-frequency band was selected based on a need to accomodate standard off-the-air channel frequencies and the existing channel select capability of the television receiver. The original forward subscriber system, therefore, began at channel 2 (54 MHz) and extended through channel 13 (216 MHz) excluding the presently defined mid-band frequencies from 88 to 174 MHz. As knowledge of system operation developed and equipment technology advanced, operators began considering addtional channel capacity above the 12 channels initially offered. Equipment manufacturers responded to the demand for additional subscriber system forward-transmission bandwidth and extended the original upper 216-MHz limit to the present practical limit of 440 MHz.

Enterprising system operators looking for methods of increasing revenue began to consider the prospects of not only a link to the subscriber, but a link from the subscriber back to a central facility. Since the frequency spectrum below channel 2 was vacant, it seemed an ideal slot to provide the return link from the subscriber. With the standard forward system already constructed and the cable itself bidirectional and able to support information flow in either direction, it seemed logical that by adding a reverse signal path around each forward repeater amplifier, the existing system without major modification would support two-way transmission. CATV equipment manufacturers responded to this idea and engineered the proposed sub-band equipment and system; these became known as sub-split equipment and system, respectively. Immediately, sub-split reverse upgradable equipment became a necessity, and few systems built since the early 1970s are without sub-split reverse capability.

As applications for two-way transmission evolved, the need for a second type of two-way broadband RF distribution system developed. This system was to provide equal bandwidth in both forward and reverse directions and be used for interconnection between multiple common institutions. Because of the institutional application and the equal bandwidth requirement, the system was called interchangably an institutional or mid-split system. Equipment developed for institutional system application was and is most commonly called mid-split equipment.

As originally conceived, the institutional system was to have its own separate cable and was to shadow the main subscriber cable. In this dual cable configuration, the standard subscriber cable became known as the "A" cable system, and the institutional cable as the "B" cable system. Aslo, it was originally proposed that the sub-split reverse from the "A" cable system be added into the "B" cable system so that only a single common reverse path back to the headend would be required. This was accomplished through a "seventh" port connection on the "A" and "B" cable repeater amplifier station housings. However, in recent years this common A/B cable interconnection has been questioned due to the potential of "B" cable reverse contamination from stray signal ingress into the "A" cable sub-split subscriber system. Sub-split reverse ingress will be discussed in more detail later.

When mid-split equipment was originally developed, the upper frequency limit for forward amplifiers was approximately 270 MHz, and the lower frequency limit for reverse amplifiers was 5 MHz. Allowing for a guard band (diplex filter high-pass/low-pass crossover region) between forward and reverse, the total 5 to 270-MHz band was approximately equally divided for two-way operation. However, since the initial development of mid-split, the forward amplifier upper frequency limit has stepped upwards from 270 to 300, 330, 360, 400 and, finally, to the present limit of 450 MHz. As a result, the term mid-split no longer strictly implies equally divided frequency bands for forward and reverse. Although two-way mid-split and dual cable systems were much talked about during the early and mid-'70s, not much demand was created and not many systems of this type were constructed.

Interest in the third type of two-way system grew from the 1980 development of the 400-MHz expanded-bandwidth subscriber system. As 400-MHz technology advanced, system operators began looking for services that would interest franchising authorities and at the same time provide additional sources of revenue. As a result, renewed interest in the institutional system began, and at the request of system operators, equipment manufacturers began to dust off mid-split and at the same time examine a new equal bandwidth split-band system for 400 MHz. Since mid-split was still commonly used to describe split-band systems above 30 MHz nut below 300 MHz, the name given the 400-MHz equal-bandwidth split-band system was high-split. However, as upper-frequency expansion continued, the high-split term soon lost its original equal bandwidth connotation and, consequently, today no longer strictly implies equally divided frequency bands for forward and reverse. However, when readily available, it is expected that high-split will become the new standard for two-way institutional split-band RF distribution networks and will predominate in that role at least until further forward bandwidth extension occurs. Subscriber and Institutional Systems

CATV RF broadband distribution systems typically fall under one of two classifications. One is called a subscriber system and is used to provide services to the general public. The subscriber system may or may not include sub-split reverse. The other is called an institutional system and includes reverse and connects banks, schools, government facilities and the like. The application of the institutional system is now expanding in the area of supplying the data communications links (see article in March CN, page 64). The true institutional system always includes a reverse system with a frequency bandwidth greater than that of sub-split reverse. Occasionally, the individual functions are inter-mixed with the subscriber system providing a two-way link between institutions and the institutional system providing services to the general public. However, by historical definition, each still retains its identity in that the subscriber system forward-frequency band always begins at 54 MHz and may or may not include sub-split reverse, while the institutional system forward frequency begins above 100 MHz and always includes a broadband reverse system.

Both the subscriber system and the institutional system are constructed in a similar fashion and typically include a trunk system and feeder system. The trunk system is the primary transportation system and, as such, branches in tree fashion throughout the area to be cabled. The feeder system branches from each trunk station in a similar tree-like manner and serves as the connecting link between the trunk system and the end user (subscriber or institutional). In some instances, the institutional trunk is routed directly to the end user and the feeder system is not included.

Large-diameter coaxial cable, repeater amplifiers, three-port power dividers (splitters and directional couplers), power supplies and AC power inserters are the main component parts used to construct the trunk system. A single series trunk path called a cascade can be miles in length and can contain many trunk-class repeater amplifiers. Consequently, individual trunk amplifiers must be high-quality, high-reliability amplifiers that exhibit low levels of noise and non-linear distortion. Power dividers having equal and unequal loos characteristics are used to branch the trunk system. The splitter or equal-loss power divider is used at branch junctions where equal-length trunk branches are required; the directional coupler or unequal-loss power divider is used at branch junctions where unequal-length trunk branches are required. Careful system design and efficient use of the family of power dividers has a direct effect on the total number of cascaded trunk amplifiers and, consequently, on overall system performance. The trunk amplifier stations (also feeder amplifier stations) are AC-powered through the same coaxial cable that carries the RF signal. Pole-mounted 60-volt AC power supplies are located at intervals throughout the system and are connected to the coaxial cable through power inserts.

Small-diameter coaxial cable, repeater amplifiers, three-port tower dividers (splitters and directional couplers), directional taps and in-line equalizers are the main component parts used to construct the feeder system. Unlike the trunk, a single series feeder path or cascade is relatively short and typically contains not more than three line extender class repeater amplifiers. Consequently, individual feeder amplifiers do not require the same degree of sophisticated level control and frequency response circuitry characteristic of the individual trunk amplifier. Feeder amplifiers do, however, require the smae high quality, high reliability and low levels of noise and non-linear distortion characteristic of the trunk amplifier. The first feder amplifier is actually located within the trunk amplifier station housing and is called a bridging amplifier. The bridging amplifier receives signal from the trunk, raises the level and provides the source signal for up to four individual trunk station feeder cables. In addition, the bridging amplifier provides feeder-to-trunk isolation to minimize possible spurious signal leakage into the high-quality trunk transportation system. Feeder cables leaving the trunk are tapped with directional taps at regular intervals to provide connections for subscriber access to the system. Splitters and directional couplers are used to branch the feeder system just as described for the trunk. In-line equalizers are spliced into feeder cable spans to adjust for accumulated differences in signal level that occur due to the frequency-dependent loss variation of coaxial cable. Since each trunk station supports its own independent feeder system, there are generally as many independent feeder systems within a cable distribution system as there are trunk stations.

Subscriber and institutional trunk systems are constructed in exactly the same fashion and, except for trunk amplifier station diplex filters and reverse amplifiers, use exactly the same component parts. Subscriber and institutional feeder systems are, however, not always exactly similar in that the institutional feeder system does not typically provide connections for the general public. Consequently, the institutional feeder system may not include large quantities of directional taps and does not include sub-split-only, in-line equalizers. In fact, in many cases, the institutional feeder system simply consists of the bridging amplifier and direct cable connection to the user.

In recent years, all components used to construct both the subscriber and institutional trunk and feeder system have been broadband bidirectional. Consequently, for these systems to support reverse, it is only necessary to provide proper splitband diplex filters and reverse amplifier modules for each of the trunk and line extender stations. As it relates to components this is true; however, a proper working reverse system is guaranteed only if a separate reverse system design analysis is performed simultaneously with forward-system design. Some sub-split systems that were designed for forward-only operation have encountered problems when activating the sub-split reverse uprgadable option. Equipment Has Established Track Record

Equipment for two-way broadband RF distribution systems has been available to the cable TV industry for almost a decade. Moreover, many of the proposed products that would utilize two-way transmission are available and have been successfully operating in two-way cable systems for years. As a result, many of the initial problems have been eliminated, and much has been learned about the operational requirements and limitations associated with these types of systems. One problem long fought and still present is spurious signal ingress into the sub-split subscriber reverse system. Two factors have made this problem an especially difficult one to solve. One factor is the sub-split reverse-frequency band itself and the infinitely large numbers and high levels of non-cable-related 5 to 30-MHz signals always present in the environment. The other factor is related to subscriber system architecture and the potential problems due to loss of shielding integrity at the many subscriber connection points within the system.

Several articles that have been published recently refer to the "digital headend." The digital headend will process signals that are somewhat different from video signals and will use modulation methods that may be unfamiliar. Nevertheless, the analogy to the classical cable television headend is apparent.

On the customers' premises, racks of equipment consisting of digital multiplexers and standard-bandwidth modems are installed. The output of the modems are frequency-multiplexed on a two-way system with the output from other modems. At the earth station, there will be much larger "digital headends" receiving the modem RF signals and processing them to a format compatible with the satellite equipment. For installations where the digital earth station is not collocated with the cable headend, data translators may be required to allow full access to any location in the system. In many respects this "digital headend" is less complex than some of the very sophisticated cable headend systems being installed today.

The information that is to be sent by calbe can be multiplexed using time-division multiplexing (TDM) or frequency-division multiplexing (FDM). Some combination of time and frequency-division access will probably be used for most applications. Consider the data rates commonly used in satellite circuits (Figure 1). If many low-data-rate ports are available at one location, the most cost-effecive method of transmitting the signals is to time-division-multiplex before modulation. On the other hand, the higher-date-rate services are usually sent single-channel-per-carrier (SCPC) using frequency-division multiplexing on the coax or satellite.

Another item that needs to be mentioned is changing access according to the changing needs of different users (multiple access). Pure time-division systems may be made multiple-access by allowing different users to occupy different time slots on demand (TDMA). Likewise, frequency-division multiplex systems may be extended to demand access (FDMA or DAMA). Another method that is commonly used is called carrier-sense multiple-access/collision detection (CSMA/CD). Basically, a station wishing to transmit using CSMA/CD listens to the circuit. If the link is idle, it transmits. If two stations should transmit simultaneously (collide), each attempts to transmit again after a random delay. Two other widely accepted schemes for multi-access that are closely related are polling and token passing. In polling, a master controller polls each station on the network. If a station has a message to send, the controller gives the station the right to transmit. In token passing, the control is not centralized; the stations are given a sequence. After transmitting a message or no message, each station passes access rights (token) to the next station. This passing of the token continues in a cyclical fashion.

In addition to the multiplexing and accessing method, one also has to consider the modulation that is to be used. The basic possibilities are amplitude-shift keying (ASKe, phase-shift keying (PSK) and frequency-shift keying (FSK). An attractive method for transmitting on non-linear systems (satellites) is PSK. PSK transmission can take place using any number of phases; for example, two phases (biphase, BPSK) and four phases (quadra-phase, QPSK). BPSK and QPSK are widely used in satellite circuits. When transmitting at high data rates on cables, the frequency spectrum must be conserved. Since the cable is relatively linear and the signal-to-noise ratios are high, an attractive modulation method is a combination of amplitude and phase-shift keying. The measure of transmission efficiency normally used is called bits per Hertz. This is the number of bits per second that can be transmitted in a 1-Hz bandwidth. The theoretical number of bits per Hertz for BPSK is one, for QPSK it is two, and for a combination of amplitude and phase-shift keying, it can be three or more. For data rates of T1 or larger, a combination of amplitude and phase-shift keying can conserve spectrum on the cable.

On the other hand, consider the case of a large number of users at different locations, each user having a relatively low data rate, and each user transmitting occasionally. Here a modulation method can be employed that uses spectrum less efficiently, but yields less costly hardware, such as FSK. In addition, the system could employ CSMA/CD or token passing, giving all users access to the same channel. In most cases, it is evident which type of modulation method should be employed.

Let's consider an example of two current business communication needs. Business A wanted to establish a dedicated communication link to another city with eight voice channels, one 9.6-kb/s circuit, one 56-kb/s circuit and one 224-kb/s circuit.

Business B wished to communicate with 40 voice channels through the same satellite communications earth station. Twenty-four voice channels can be digitized on a commercially available channel bank and sent on a T1 (1.544-Mb/s) circuit. Data multiplexers can be added to one channel bank in order to provide the eight voice channels and the various data circuits. The T1 output from the channel bank is then fed into a T1 modem and converted to the appropriate frequencies for transmission on the cable (Figure 2). The 40-voice-channel requirement for Business B can be satisfied using two channel banks and two T1 modems. Both customer requirements are satisfied economically using the configuration shown in Figure 2.

Before exploring the question of capacity of a broadband cable system, it would be beneficial to look at a couple of examples of data circuits on cable. Figure 3 is a schematic representation of a mid-split broadband distribution system.

Signal flow in the conventional (downstream) direction takes place in the band 174 to 440 MHz, while reverse (upstream) signals are carried in the band 5 to 108 MHz. This two-way capability is achieved by the use of diplexing filters and dual amplifiers at each signal amplification station.

Suppose that at point "A" there is a bank branch office equipped with a data terminal that must communicate with the master computer located in the head office at point "B." The data terminal at "A" is connected to a modem that can transmit onto and receive from the distribution system, and the master computer at "B" is similarly connected. It is clear that there is no way in which a signal can travel directly from "A" to "B" using the route LE-2, TA-2, TA-3, LE-1 or vice versa. Any signal in the range of 5 to 108 MHz will travel only toward the headend, and any signal in the range of 174 to 440 MHz will travel only away from the headend. However, the addition of one simple device at the headend will solve this problem, and allow great flexibility in connecting additional customers to the system. The device is a frequency translator, which receives signals at the headend in the "reverse" frequency range and "translates" them to the "forward" frequency range. It can be designed to translate a single 6-MHz channel or several channels, depending on the amount of traffic and the availability of unoccupied channels.

Therefore, when "A" wishes to send a message to "B," "A's" modem transmits at a frequency in the range of 5 to 108 MHz. The signal returns to the headend, where the translater shifts it to an available channel in the "forward" spectrum. "B" can then receive the transmission.

Figure 4 is a schematic representation of one version of a midsplit private broadband network. The difference between this distribution network and the network of Figure 3 is that all communication is between the headend and the remote sites; that is, no direct communication is allowed between remote sites. This agreement could arise with a company that would have a headquarters with multiple remote buildings or if a network has built solely to distribute communiations to and from a satellite site. In this case, the modems at the remote sites, A and B, would transmit in the 5 to 108-MHz band and receive in the 174 to 440-MHz band. The modems located at the headend, C, can use an inverted frequency plan; that is, transmit in the 174 to 440-MHz band and receive in the 5 to 108-MHz band.

The data transmission capacity of a broadband system depends on the following factors:

* Available bandwidth (expressed as the number of unoccupied 6-MHz channels). In an institutional system, the whole cable spectrum may be given over to data transmission.

* Number of data circuits.

* Modem spectral occupancy, which is a function of both the data rate and the bandwidth efficiency of the modem. Hpothetical Intracity System

To help visualized the potential capacity, we will examine first the case of the intracity, point-to-point connection. Modems for broadband systems are available from a number of sources (including Scientific-Atlanta); therefore, for the purpose of this discussion, we will assume a hypothetical, medium-speed modem that can handle data at rates up to 19.2 kb/s. The figures we shall use are reasonable and quite representative of state-of-the-art devices that meet the requirements of high reliability and low cost.

Using Bi-Phase-Shift Keying (BPSK) as our modulation scheme, we may expect to transmit a 19.2-kb/s data stream in a bandwidth considerably less than 100 kHz. We wish to allow a certain safety margin, however, so we will suppose that 100 kHz represents the "channel" separation required by the modems. A quick calculation shows that a single 6-MHz TV channel could support transmissions from 60 such modems.

To determine the actual capacity of a typical system, refer to Figure 5. This diagram is an extension of Figure 3, since it illustrates the two-way communication between points "A" and "B." The cable system has a reverse channel, designated T7, available and a forward channel, designated H, currently unused. Each 6-MHz channel has been divided into 60 subchannels," numbered 1 through 60, so that we can readily described the "slots" that the modems require for transmission and reception.

When modem "A" transmits a signal, it uses subchannel 1 of channel T7. This can be written as T7(1). The signal travels in the reverse direction to the headend, where it is intercepted by the frequency translator and converted to the channel H, subchannel 1, or H(1). The transmission can now travel in the forward direction throughout the distribution system, and be received by modem "B," which is tuned to H(1).

Transmission in the opposite direction, from "B" to "A," must make use of a different subchannel. This is because both modems may be required to transmit and receive simultaneously (full-duplex operation). Therefore, modem "B" transmits in T7(2), and modem "A" is tuned to receive H(2).

To summarize, a pair of modems operating in full-duplex mode requires the allocation of two subchannels in the reverse-frequency range, and two corresponding subchannels in the forward-frequency range.

Broadband technology offers the immediate incentive of reduced cost and improved quality of transmission, and the confidence that increasingly sophisticated networking capability and faster data transportation, necessitated by the growing demand for information transfer and the concomitant growth of data technology, are tasks that fall naturally with the scope of coaxial cable systems.