INSTALLATION: Surge protection prevents network disasters
Electrical power problems can be extremely detrimental to data-communications networks. A power surge that lasts just one-millionth of a second can render a local area network (LAN) inoperative. Estimates of the annual cost of damage from power-related problems range from millions to billions of dollars
A zoned approach to surge protection can stop power surges before they hurt your network.
Anthony O. Bird and John C. Roth / Atlantic Scientific Corp.
Electrical power problems can be extremely detrimental to data-communications networks. A power surge that lasts just one-millionth of a second can render a local area network (LAN) inoperative. Estimates of the annual cost of damage from power-related problems range from millions to billions of dollars. Regardless of the cost, a power problem that damages your database server or the LAN that it is on can rob you of the ability to serve your clients, and the toll from that is incalculable.
Most people are familiar with the debilitating blackouts caused by lightning, falling trees, and the occasional squirrel. But these events rarely cause power outages or equipment failure. More often, power problems originate from your electric utility or from the everyday operation of equipment within your building, which means that computer systems and networks are threatened by the very environments in which they function. In fact, almost half of the electrical power problems recorded are generated internally.
To set up our LANs, we have run hundreds or thousands of feet of cable throughout our buildings using a variety of cable types, each capable of carrying enough of an electrical surge to damage the communications port on your computer or server. Indeed, the incidence of equipment input/output port failures has skyrocketed in the past few years.
Prevention, not recovery
It costs much less to avoid a disaster than to recover from one. One way to avoid network disruption is to create "zones" that lower transients in steps. Such a strategy provides the safest electrical environment for your sensitive electronic equipment.
Electrical power problems occur in various forms. Each has its own characteristic that enables it to damage electronic equipment. These problems include the following:
Swell and sag: This term describes a temporary raising (swell) and lowering (sag) of voltage. These power anomalies are usually not harmful if the resulting voltage is within 10% of the nominal voltage recommended for your equipment. Swells and sags are usually caused by the utility company when it is trying to adjust for heavy demand during extreme hot or cold weather. A sag of long duration is commonly called a brownout. Together, these problems account for less than 15% of all power problems.
Outage: This is commonly referred to as a blackout. Blackouts are generally not harmful to equipment but can bring business operations to a standstill. Fortunately, total outages account for less than 1% of all power problems.
Surge: Surges are often called transient voltages or spikes. These sharp rises in voltages can, in a microsecond, far exceed safe voltage levels. As they pass through personal computers, modems, fax machines, and other delicate electronic systems, they cause system downtime, data loss, degradation, or immediate damage.
Lightning can cause a surge when it directly hits power or telephone lines. Even a nearby hit can cause a surge from either inductive coupling within your building or ground potential differences from one side of your building to the other. Surges can also be caused by grid switching by the utility company or by equipment operating within your building, such as elevators, copiers, and air conditioners.
Most surges go unnoticed day after day. If no protection is installed against them, surges can slowly degrade the sensitive electronic components inside your equipment. Then, with little or no warning, the equipment fails. It is estimated that more than 95% of equipment failure is from component degradation over time.
There are many ways to defend against unwanted electrical power surges. Used together, they can provide the disaster avoidance necessary to protect your network and your business:
Structural lightning protection: Installed outdoors, this protection provides a controlled low-impedance path for lightning current to flow to ground. The required ground resistance and number of paths to ground are specifically chosen to limit the voltage rise on the structure, reducing the chance of flashover. Without a lightning-protection system in place, lightning can penetrate the building, resulting in a flashover to inner building wiring. Flashover means that lightning current flows through the wiring, with potentially disastrous consequences. If your building requires structural lightning protection, follow codes such as NFPA 78-1983 of the National Fire Protection Association (Quincy, MA) and the American National Standards Institute (New York City). Structural lightning protection by itself, however, cannot prevent damage to sensitive electronics.
Grounding system: The building grounding and bonding system should be designed in accordance with the NFPA's National Electrical Code (NEC). Any departure from this code may pose a safety risk and compromise the effectiveness of other protection systems (see "Earth grounding basics," above).
Surge-protection devices (SPDs): These devices come in all types and capacities. But just as important as the capacity of each surge-protection device is its placement. To protect your entire facility and be cost-effective, SPDs should be located where each device will be most effective. SPDs can be installed at the demarcation point to protect against surges coming in on the dial-up or ISDN line or leased lines such as T1s and digital data service. Internally, SPDs can protect LAN hubs, servers, and routers from induced surges coming across twisted-pair and coaxial cables.
Uninterruptible power supplies (UPSs) and generators: Where continuous power is required to keep critical operations functioning, UPSs and alternative power sources such as diesel generators should be an integral part of an overall protection system design. However, a UPS by itself is not a solution for all problems; it is a component in a complete defensive strategy.
Zones are areas delineated and defended by the application of surge protection. A number assigned to each zone indicates the severity of the surge voltage that can occur in that environment. The higher the zone number, the lower the surge voltage. Each higher-numbered zone number "nests" within a lower-numbered zone.
Zone 0: The most hostile zone is the unprotected environment outdoors. The magnitude of surge currents and correspondingly high electromagnetic fields can be extreme, as in the case of lightning. In this zone, avoidance is the best solution but is not always practical. Getting off the golf course is prudent for people, but in the case of a building, the best solution for Zone 0 is structural lightning protection. Electric utilities can also add to the hostility of Zone 0 by grid switching, power factor corrections, and short circuits.
Zone 1: Ideally, the boundary of Zone 1 is the facility itself, which usually will provide a significant zone of protection. However, any conductive path penetrating the facility boundary-such as cables, telecommunications lines, and unused metallic conduits-can pose a threat. During a lightning storm, surge currents of 10,000 A to more than 100,000 A can flow for a short time. If these surge currents are allowed to flow uncontrolled over inner building wiring, high differential voltage differences will occur throughout the building due to different wiring impedances. For electronic systems interconnected by communications cables, there could be significant damage even if each system is protected locally by an AC surge protector.
To create Zone 1 and make it a safer and more controlled surge environment, surge-protection devices should be installed on all cables crossing from Zone 0 to Zone l. Such main panel SPDs should have the capacity to withstand a 90,000- to 300,000-A surge. Any other metallic conductors that cross the boundary should be bonded together and directly connected to ground. The NEC requires primary surge-protection devices to be installed on all incoming telephone lines. These devices are usually sufficient to create the Zone 1 environment for incoming telephone lines. All SPDs installed at the Zone 1 boundary should be grounded to the same electrical ground. Special care is required when AC power leaves Zone 1 and reenters Zone 0 to supply external equipment such as outside lighting, security cameras, or rooftop equipment. Surge protection should be installed at the panel to which this external equipment is connected.
Zone 2: In buildings, electrical distribution costs are significantly reduced by using subpanels throughout the building or on different floors, as in the case of a high-rise office building. Ideally, both AC and data-communications lines should enter Zone 2 at the same location. Surge protection should be used on the AC subpanel. Here the surge protection can be very effective with SPD ratings of only 40,000 to 90,000 A.
The data-communications cabling should also be protected at the demarcation point installed by the telephone company. The decision to protect data cabling crossing into Zone 2 will depend on the cable exposure and run length. When data cabling exceeds 100 meters from the demarc (50 meters in electrically noisy environments) or is routed through large loop areas where induced surges could occur, protection should be installed at the Zone 2 boundary. In a facility where data-communications networks join equipment connected to many different subpanels, it is very important to install surge protection at each subpanel. Likewise, the hub in the telecommunications closet must be protected both on the AC and the data side of cabling.
Zone 3: This is the zone in which we work, commonly known as the desktop, although the discussion that follows applies to mission-critical servers as well. Unfortunately, it is here that the Band-Aid approach to surge protection can make matters worse.
Today, it is common to have a data-communications and a dial-up telephone line coming to the desktop along with the AC power. Placing a UPS on the AC power going to the desktop or server is fine, but if the same is not done on the telephone and data-communications lines, serious damage can result because during a surge, the voltage potential difference between a component protected by a UPS and an unprotected LAN is greater than before. Such a major difference in voltage potential can wreak havoc on the LAN and telephone system.
If AC surge protection has been applied to Zones 1 and 2, then a lower protection capacity of 30,000 to 60,000 A will usually be adequate for creating Zone 3. The let-through voltage of the surge protector is the key specification here: The lower the let-through voltage, the better. Forget joule ratings; a good surge protector grounds a surge fast enough that joules of energy won't build up.
LAN and telephone lines should be protected at the same point as the desktop or server AC protection, which will provide protection against surges coming into Zone 3 from the telephone or LAN cabling.
Anthony O. Bird is president and John C. Roth is national sales manager at Atlantic Scientific Corp. (West Melbourne, FL).
Earth grounding basics
Stephanie Woodbury, ATI Tectoniks NA Inc.
An effective earth grounding system is the basis of any electrical site protection strategy. Earth grounding is integral to personnel safety and the successful long-term operation of equipment.
Grounding provides a low-impedance path to earth for high currents such as lightning strikes or power surges. For the safety of those employed in a building, it is necessary to maintain good earth grounding to ensure all noncurrent-carrying metal parts are at ground or zero potential. Grounding also provides a reliable path to earth for low currents such as radio-frequency interference or static that builds up on system enclosures and signal return paths. Standards for equipment performance mandate installation of a reliable, stable, low-resistance grounding system.
The most important element of any grounding system design is soil evaluation. A soil resistivity test is the most accurate way to analyze the soil. This test should be the first step in designing a grounding system because the information it yields is the key to ensuring predictable ground grid performance. A soil-resistivity profile is also important because resistivity varies widely at different depths and in different soil types. Soil resistivity readings at 5 and 10 ft may be very different from readings at 20 and 40 ft. Typically, resistivity decreases with depth; when the reverse happens, you have the information you need to design the best and most cost-effective solution for those conditions.
Electrolytic grounding systems are an excellent alternative to copper-clad driven rods, which are subject to damage during installation (the act of driving them into the ground may damage the thin copper coating), difficult soil conditions (these rods depend on the environment for their stability), and corrosion. Electrolytic rods are factory-filled copper tubes that weep moisture when exposed to air. The rod body is pure copper, which prevents corrosion due to dissimilar metals, and the electrode continuously provides its own moisture in any conditions, wet or dry. The rod is stable and low in resistance year-round. The backfill material used is dense, adhering to the electrode's surface and creating an excellent conductive medium.
Businesses make large capital investments in their networking equipment, and they expect uninterrupted, reliable services from their networks. Including standards-compliant grounding systems to protect their investment makes sense from both an operational and cost perspective.
Stephanie Woodbury is sales engineer at ATI Tectoniks NA Inc. (Torrance, CA).