Hybrid Fiber Coaxial Systems

Coaxial cable systems provide great improvements in television reception in many parts of the country. Even in areas where reception was good, CATV was popular because of increased programming options. Many new channels were premium or pay-per-view channels, thereby affording broadcasters new sources of revenue and giving content creators new outlets for their work.

The Introduction of Fiber Trunks

Despite the success of coaxial cable networks, pure coaxial cable systems are insufficient in several ways for high-speed residential broadband service. First, channel capacity is still insufficient to compete with Direct Broadcast Satellite (DBS). Coaxial systems can provide roughly 40 channels, but DBS subscribers can receive more than twice as many channels, affording them more innovative programming options. DirecTV in particular has done well with enhanced sports packages. Cable networks require additional channel capacity to be competitive.

Second, coaxial systems lack robustness. If an amplifier malfunctions near the head end (for example, by losing power), all subscribers downstream from the bad amplifier lose service. Referring back to Figure 3-1, if amplifier 4 breaks, then Jimmy, Rosie, Mom, and Junior will lose service.

Third, signal quality is insufficient for large numbers of users. In Figure 3-1, Junior or Leonardo might be 40 to 50 miles from the head end. The number of amplifiers between the head end and the last subscriber is called the cascade depth. Junior is six amplifiers from the head end; that is, his cascade depth is 6. Cascade depths of 40 to 50 are possible in some systems in the United States. Amplifying a signal 40 times is like copying a tape 40 times: The last signal is never quite true to the original.

Finally, coaxial cable systems are very complicated to design and operate. Referring back to Figure 3-1, if many new subscribers living near Jimmy begin service, it is possible that their taps would cause enough attenuation to require a new amplifier for serving Mom's neighbor-hood. But even if there were no changes in subscribership, there are dozens of amplifiers, splitters, pads, taps, cables of various impedences, thermal changes (which lengthen or shorten cables as ambient temperature changes), and pirates (doing their dastardly deeds between the head end and subscribers) spread over a 40- to 50-mile radius. Keeping power equalized for all subscribers is a difficult problem.

Ongoing operational network design is also complicated by rusting components, leakage due to perforation of insulation, temperature gradients, loosened fittings, malice, and other annoying causes. These imperfections are difficult to find and, when found, typically require onsite maintenance. Each instance of onsite maintenance costs about $50 to $100 per incident for an onsite visit or truck roll, which compares unfavorably to the $20 to $40 monthly charges paid by subscribers. Software programs exist to assist technicians, but for the most part, they use manual processes as well as a lot of intuition and experience.

To combat these problems, cable operators came up with the idea to use fiber optic cable in place of coaxial cable trunks. The total system would have both fiber and coaxial cables, hence the term hybrid fiber coaxial (HFC) networks. With the invention of the linear light source and analog fiber systems, it became practical to introduce fiber into the trunks. The requirement for analog fiber systems was to maintain compatibility with the existing analog metallic plant. Also, the use of analog fiber was cheaper and more reliable than conversion to and from digital fiber.

By using fiber, the cascade depth can be reduced to a handful of amplifiers, perhaps four or five. Reliability and equalization problems associated with deep cascades are reduced. The systems offer more bandwidth and are easier to design. Figure 3-2 shows how the system in Figure 3-1 would evolve with the introduction of fiber.

The diagram shows that trunk coaxial cables have been replaced by multiple fiber cables. If a fiber link breaks, fewer homes lose service. The fiber cables are terminated at media translators called fiber nodes that convert optics into electronics. Therefore, the cable system is segmented into smaller clusters, each of which is defined by those homes served by a single fiber node. In this case, Picasso, Einstein, and Leonardo are passed by a single-fiber node and therefore are in the same cluster. It is the intent of MSOs to limit clusters to 500 to 2000 homes passed.

One to six coaxial cables are attached to the fiber node. In the forward direction, the fiber node splits the analog signal so that the same signal is sent on each coaxial cable. Traffic flowing in the opposite direction—that is, from the home to the head end—is referred to as upstream, reverse, or return path. In the return path, traffic is received from the coaxial cables and is multiplexed in the frequency domain onto the fiber. The definition and purposes of the return path are discussed in more detail in the section "Upstream Transmission" later in this chapter.

Between the fiber node and the subscriber, the assembly of coaxial cables, amplifiers, splitters, and taps is the same as the pure coaxial cable scenario. HFC systems reduce the number of amplifiers, thereby increasing available bandwidth and reducing maintenance.

A key innovation underpinning HFC deployment was the development of amplitude modulated fiber-optic transmission. Until the invention of the linear light source in the 1980s, it was not possible to have amplitude modulated (AM) fiber; without AM fiber, it was not possible to have analog transmission over fiber. Without analog transmission over fiber, cable would have had to perform analog-to-digital and digital-to-analog conversions (as discussed in Chapter 2) to use fiber, which is a costly proposition. AM fiber was central to the evolution of HFC cable systems. Figure 3-3 shows how a cable HFC network maps to the DAVIC reference model.

HFC Upgrade Costs

The cost of upgrading to fiber varies by geography, labor rates, and to some extent on what dollars are allocated to upgrade costs and what dollars are allocated to operational costs. Because metallic cables need periodic replacement anyway, it makes sense to replace metallic cables with fiber cables during normal maintenance. Coaxial cable can be reused to feed electric power to the fiber node, and thus find a second life.

The cost of upgrades includes the cost of optoelectronics (fiber and laser transmitters), changes to the cable plant (connectors, power supplies, passives, and occasionally new cable), and labor. An important cost factor is whether the new cable is overhead or buried (aerial or burial). Numbers to upgrade systems are typically proprietary. The S1 prospectus of Ciena Corporation (Nasdaq-CIEN) indicates a total cost of $70,000 per route mile.