Cables

Fiber optic cables have different types of mechanical connections. The type of connection determines the quality of the fiber optic lightwave transmission. The different types we’ll discuss here are the flat-surface, Physical Contact (PC), Ultra Physical Contact (UPC), and Angled Physical Contact (APC).

The original fiber connector is a flat-surface connection, or a flat connector. When mated, an air gap naturally forms between the two surfaces from small imperfections in the flat surfaces. The back reflection in flat connectors is about -14 dB or roughly 4%.

As technology progresses, connections improve. The most common connection now is the PC connector. Physical Contact connectors are just that—the end faces and fibers of two cables actually touch each other when mated.

In the PC connector, the two fibers meet, as they do with the flat connector, but the end faces are polished to be slightly curved or spherical. This eliminates the air gap and forces the fibers into contact. The back reflection is about -40 dB. This connector is used in most applications.

An improvement to the PC is the UPC connector. The end faces are given an extended polishing for a better surface finish. The back reflection is reduced even more to about -55 dB. These connectors are often used in digital, CATV, and telephony systems.

The latest technology is the APC connector. The end faces are still curved but are angled at an industry-standard eight degrees. This maintains a tight connection, and it reduces back reflection to about -70 dB. These connectors are preferred for CATV and analog systems.

PC and UPC connectors have reliable, low insertion losses. But their back reflection depends on the surface finish of the fiber. The finer the fiber grain structure, the lower the back reflection. And when PC and UPC connectors are continually mated and remated, back reflection degrades at a rate of about 4 to 6 dB every 100 matings for a PC connector. APC connector back reflection does not degrade with repeated matings.

What is Fibre Channel?
Fibre Channel is a set of communication standards designed to provide high-speed data transfer over a duplex, serial interface. It’s an open standard that supports multiple protocols including higher-level protocols, such as FDDI, SCSI, HIPPI, and IPI, to manage data transfer.

Although it operates at a range of 133 Mbps to 4 Gbps, Fibre Channel is most commonly used at speeds of 1 or 2 Gbps. A working standards group recently announced that 10-Gbps speeds are expected in soon.

Why is it called Fibre Channel?
Originally, Fibre Channel was designed to support only fiber. When copper was added, the International Standards Organization (ISO) task force changed the spelling of fiber to fibre instead of renaming the technology.

Fibre Channel history.
Fibre Channel was first developed in 1988, and the American National Standards Institute (ANSI) formed a committee in 1989. To ensure interoperability, IBM®, Hewlett-Packard®, and Sun Microsystems® formed the FCSI (Fibre Channel Systems Initiative), a temporary organization, in 1992. FCSI later dissolved, and development was handed over to the FCA (Fibre Channel Association) in 1994. ANSI accepted Fibre Channel as a standard in 1994.

The best of both worlds.
This hardware-based standard combines the best of both channel and network communication methods into one I/O interface. It takes advantage of hardware-intensive, quicker point-to-point channel links that offer low overhead, such as SCSI bus technology, as well as the broad connectivity and long-distance benefits of software-intensive network technology.

Where Fibre Channel is used.
Fibre Channel is used to transfer large amounts of data quickly between supercomputers, mainframes, workstations, desktop computers, storage devices, displays, and other peripherals.

Fibre Channel offers reliability, scalability, congestion-free data flow, Gigabit bandwidth, compatibility with multiple topologies and protocols, flow control, self management, hot pluggability, speed, cost efficiency, loop resiliency, and distance. This makes it ideal for large data operations such as Internet/intranets, data warehousing, networked storage, integrated audio/video, real-time computing, on-line services, and imaging.

The most popular application for this technology right now is Storage Area Networks (SANs). Independent methods of centralized storage management within a SAN (e.g., RAID, tape backup or library, CD-ROM library) run more efficiently with a Fibre Channel backbone.

Fibre Channel topologies.
Fibre Channel can be connected by three methods. In all cases, the topology of the network is transparent to the attached devices.

Point to point is the simplest topology, which uses simple bidirectional links between two connected devices.

Arbitrated loop is the most common topology and the most complex. It is distributed, connecting up to 126 devices across shared media, and it offers shared bandwidth. Two ports on the loop establish a point-to-point, full-duplex connection through arbitration among all ports.

The cross-point or fabric-switched topology uses 24-bit addressing to connect up to 2 (to the 24th) devices in a cross-point switched configuration. This enables many devices to communicate at the same time and does not require shared media.

Fibre Channel layers.
Fibre Channel protocol is divided into five hierarchical layers: The three bottom layers, FC-0–FC-2, define the physical transmission standard. Layers FC-3 and FC-4 address interfaces with other network protocols.

FC-0: Media and interface layer that defines the physical link.

FC-1: Transmission encode/decode layer. Information is encoded 8 bits at a time into a 10-bit transmission character (8B/10B from IBM).

FC-2: Signaling protocol layer that serves as the transport mechanism performing basic signaling and framing. FC-2 includes the following classes of service:
• Class 1 provides dedicated connections. Intermix is an optional type of Class 1 service in which Class 1 frames are guaranteed a special amount of bandwidth.
• Class 2 is a frame-switched, connectionless service, also known as multiplex. It guarantees delivery and confirms receipt of traffic.
• Class 3 is a one-to-many, connectionless, frame-switched service. It’s similar to Class 2 except it uses buffer-to-buffer flow control and does not confirm frame delivery.

FC-3: Common-services layer that provides common services required for advanced features such as striping, hunt groups, and multicast.

FC-4: Upper layer for protocol mapping of network and channel data transmitting concurrently over the same physical interface.

Fibre Channel media.
Fibre Channel runs at up to 1 Gbps over copper or fiber, but for higher speeds, fiber is required. Copper-wire cable can be video coax, miniature coax, or, most commonly, shielded twisted pair with a DB9 or HSSDC connector. Fiber choices include 62.5- or 50-µm multimode and 7- or 9-µm single-mode fiber, all with an SC connector.

Other Fibre Channel equipment includes disk enclosures, drivers, extenders, hubs, interface converters, host bus adapters, routers, switches, and SCSI bridges.

The Digital Visual Interface (DVI) standard is based on transition-minimized differential signaling (TMDS). In a typical single-line digital signal, voltage is raised to a high level and decreased to a low level to create transitions that convey data. To minimize the number of transitions needed to transfer data, TMDS uses a pair of signal wires. When one wire goes to a high-voltage state, the other goes to a low-voltage state. This balance increases the data-transfer rate and improves accuracy.

There are different types of DVI connectors: DVI-D, DVI-I, DVI-A, DFP, and EVC. DVI-D is a digital-only connector.

DVI-D is a digital-only connector. DVI-I supports both digital and analog RGB connections. Some manufacturers are offering the DVI-I connector type on their products instead of separate analog and digital connectors. DVI-A is used to carry an analog DVI signal to a VGA device, such as a display. DFP, like DVI-D, was an early digital-only connector used on some displays; it’s being phased out. EVC (also known as P&D) is similar to DVI-I only it’s slightly larger in size. It also handles digital and analog connections, and it’s used primarily on projectors.

One of the most important cable measurements is Near-End Crosstalk (NEXT). It’s signal interference from one pair that adversely affects another pair on the same end.

Not only can crosstalk occur between adjacent wire pairs (“pair-to-pair NEXT“), but all other pairs in a UTP cable can also contribute their own levels of both near-end and far-end crosstalk, multiplying the adverse effects of this interference onto a transmitting or receiving wire pair.

Because such compounded levels of interference can prove crippling in high-speed networks, some cable manufacturers have begun listing Power Sum NEXT (PS-NEXT), FEXT, ELFEXT, and PS-ELFEXT ratings for their CAT5e and CAT6 cables. Here are explanations of the different types of measurements:

NEXT measures an unwanted signal transmitted from one pair to another on the near end.

PS-NEXT (Power Sum crosstalk) is a more rigorous crosstalk measurement that includes the total sum of all interference that can possibly occur between one pair and all the adjacent pairs in the same cable sheath. It measures the unwanted signals from multiple pairs at the near end onto another pair at the near end.

FEXT (Far-End crosstalk) measures an unwanted signal from a pair transmitting on the near end onto a pair at the far end. This measurement takes full-duplex operation into account where signals are generated simultaneously on both ends.

ELFEXT (Equal-Level Far-End Crosstalk) measures the FEXT in relation to the received signal level measured on that same pair. It basically measures interference without the effects of attenuation—the equal level.

PS-ELFEXT (Power Sum Equal-Level Far-End Crosstalk), an increasingly common measurement, measures the total sum of all intereference from pairs on the far end to a pair on the near end without the effects of attenuation.

The background
As today’s networks expand, the demand for more bandwidth and greater distances increases. Gigabit Ethernet and the emerging 10 Gigabit Ethernet are becoming the applications of choice for current and future networking needs. Thus, there is a renewed interest in 50-micron fiber optic cable.

First used in 1976, 50-micron cable has not experienced the widespread use in North America that 62.5-micron cable has.

To support campus backbones and horizontal runs over 10-Mbps Ethernet, 62.5 fiber, introduced in 1986, was and still is the predominant fiber optic cable because it offers high bandwidth and long distance.

One reason 50-micron cable did not gain widespread use was because of the light source. Both 62.5 and 50-micron fiber cable can use either LED or laser light sources. But in the 1980s and 1990s, LED light sources were common. Since 50-micron cable has a smaller aperture, the lower power of the LED light source caused a reduction in the power budget compared to 62.5-micron cable—thus, the migration to 62.5-micron cable. At that time, laser light sources were not highly developed and were rarely used with 50-micron cable—mostly in research and technological applications.

Common ground
The cables share many characteristics. Although 50-micron fiber cable features a smaller core, which is the light-carrying portion of the fiber, both 50- and 62.5-micron cable use the same glass cladding diameter of 125 microns. Because they have the same outer diameter, they’re equally strong and are handled in the same way. In addition, both types of cable are included in the TIA/EIA 568-B.3 standards for structured cabling and connectivity.

As with 62.5-micron cable, you can use 50-micron fiber in all types of applications: Ethernet, FDDI, 155-Mbps ATM, Token Ring, Fast Ethernet, and Gigabit Ethernet. It is recommended for all premise applications: backbone, horizontal, and intrabuilding connections, and it should be considered especially for any new construction and installations. IT managers looking at the possibility of 10 Gigabit Ethernet and future scalability will get what they need with 50-micron cable.

Gaining ground
The big difference between 50-micron and 62.5-micron cable is in bandwidth. The smaller 50-micron core provides a higher 850-nm bandwidth, making it ideal for inter/intrabuilding connections. 50-micron cable features three times the bandwidth of standard 62.5-micron cable. At 850-nm, 50-micron cable is rated at 500 MHz/km over 500 meters versus 160 MHz/km for 62.5-micron cable over 220 meters.

Fiber Type: 62.5/125 µm
Minimum Bandwidth (MHz-km): 160/500
Distance at 850 nm: 220 m
Distance at 1310 nm: 500 m

Fiber Type: 50/125 µm
Minimum Bandwidth (MHz-km): 500/500
Distance at 850 nm: 500 m
Distance at 1310 nm: 500 m

As we move towards Gigabit Ethernet, the 850-nm wavelength is gaining importance along with the development of improved laser technology. Today, a lower-cost 850-nm laser, the Vertical-Cavity Surface-Emitting Laser (VCSEL), is becoming more available for networking. This is particularly important because Gigabit Ethernet specifies a laser light source.

Other differences between the two types of cable include distance and speed. The bandwidth an application needs depends on the data transmission rate. Usually, data rates are inversely proportional to distance. As the data rate (MHz) goes up, the distance that rate can be sustained goes down. So a higher fiber bandwidth enables you to transmit at a faster rate or for longer distances. In short, 50-micron cable provides longer link lengths and/or higher speeds in the 850-nm wavelength. For example, the proposed link length for 50-micron cable is 500 meters in contrast with 220 meters for 62.5-micron cable.

Migration
Standards now exist that cover the migration of 10-Mbps to 100-Mbps or 1 Gigabit Ethernet at the 850-nm wavelength. The most logical solution for upgrades lies in the connectivity hardware. The easiest way to connect the two types of fiber in a network is through a switch or other networking “box.“ It is not recommended to connect the two types of fiber directly.