Lightning poses a risk that can both directly and indirectly affect electronic systems in buildings, homes and cities. Direct lightning strikes can be devastating to human life, and the electromagnetic fields produced by the instantaneous surge of current can permanently damage equipment. The average lightning bolt almost instantaneously transfers up to 500,000 Joules of energy to the surface of the earth and, with over 1 billion strikes hitting the earth annually, lightning protection is highly valuable. Lightning protection applications and subsequent surge protective technology can vary greatly from base stations to data centers. The following information mostly focuses mostly on surge protective devices for systems with coaxial and Ethernet interconnect.
Lightning Protection Standards
The lightning protection standard IEC/BS EN 62305 breaks down the risks associated with a direct lightning strike, an indirect lightning strike and the resulting electromagnetic energy. As shown in Table 1, it is important for lightning protection planning to break down the areas at risk, or lightning protection zones (LPZs), and potential damage caused from a strike. The standard specifies three types of damage a lightning strike can cause:1) injury to living beings, 2) physical damage to structures and 3) failure of internal systems. Naturally, anything that can potentially harm a human will require layers of precaution. A hospital would then require significant buffers both on the external structure and on the internal equipment to protect the life within.
Underwriter’s Laboratory lists several standards for lightning protection that are geared more toward the equipment implemented for surge protection (see Table 2). The standard for surge protective devices, UL 1449, describes the modes of protection necessary in electrical paths for transient voltages based on the way in which power is supplied to a system.
End-devices are powered with either single-phase or polyphase AC power. When a device is connected between one powered line and neutral (local ground), power is supplied as single-phase AC. A three-phase system would therefore connect to three
powered lines and neutral. As shown in Table 3, surge protection is necessary between all combinations of conductive paths where differences in potential can occur. Datasheets will often list protection modes of a surge protective device with descriptors such as line-to-neutral (L-N), line-to-ground (L-G), and line-to-line (L-L) protection. Protection modes may also be described as common mode or differential mode surge protection.
A common-mode surge occurs when a surge affects all conductors in a local area equally, such as at the N-G node. Also known as a normal surge, a differential-mode surge occurs between any two conductors at a given location ― or, a surge between the line and neutral wires. Most modern equipment is inherently immune to a common-mode surge.1 According to American National Standard (ANSI) C62.41, most common-mode surges are diverted from the building where the worst case surge within a building is at 0.17 Joules, whereas differential-mode surges are many orders of magnitude higher.
Types of Surge Protectors
Coaxial Surge Protectors: Coaxial interconnect used in outdoor applications is subject to direct and indirect lightning strikes that can adversely affect the transmission line and connected circuitry. The switch from vacuum tubes to solid-state devices, particularly on the transmitter end, has elevated the risks associated with lightning strikes ― the sensitive integrated circuits are easily damaged by remnant transient voltage surges.
Coaxial cables are designed to minimize the ohmic losses of a desired high-frequency signal because of the skin effect while protecting the internal signal from outside interference with the outer conductor (or shield). The skin effect is a phenomenon that occurs when most of a signal is pushed to the outer edges of a conductor as frequency increases. Therefore, a radio-frequency (RF) signal generally does not penetrate below several one-thousandths of an inch into the conductor, as opposed to electrons moving freely all over the conductor in most AC applications. The large-diameter ground cable typically used at 60 Hz is then flattened to surround the center conductor that is transmitting a high-frequency signal. This allows the shield of the cable to act as a Faraday cage and prevents external interference from reaching the internal signal wire. An ungrounded shield will induce voltages and act as an antenna, radiating signals. A lightning strike could potentially destroy this delicate balance as it causes a difference in potential in both the inner and outer conductors of a coax. Moreover, the majority of a lightning strike’s energy falls from DC to 1 MHz, and a coaxial cable can easily carry these surges, resulting in destructive changes in the operating frequency of the connected electronic equipment.
Because of the relatively low-frequency nature of lightning, coaxial lightning protectors will often function as a DC block or a high-pass filter, diverting DC and low-frequency 50/60 Hz voltages to shield ground. In the case of gas tube surge protection, if the surge is too large to divert, the current causes to the fuse to open, eliminating the path between the SPD and shield ground (see Figure 1).
Surge protective devices used in distributed antenna system applications have the additional consideration of leverage low passive intermodulation (PIM) distortion components. PIM is a major source of interference that is particularly difficult to troubleshoot. It occurs when the passive hardware of a highly sensitive RF system under high power generates frequency products that can mix and interfere with a desired signal. The difficulty in solving the problem caused by PIM stems from the fact that it is difficult to filter out without affecting the original carrier signal because the passive components are abundantly used. With no inexpensive and quick way of eliminating PIM, system installations will require the use of low-PIM components. These components are manufactured with a careful selection of connector materials that generate inconsequential PIM levels when mated, and their manufacture minimizes the use of ferromagnetic materials (a common source of PIM).
Ethernet Surge Protectors: Surge protection for the prolifically used Ethernet backbone in data centers is essential in preventing downtime that can cost millions of dollars. According to Emerson and Ponemon’s downtime assessment study, the average cost of an unplanned data center outage was almost $9,000 per minute as of 2016 — a number that grows larger every year. Although human error and uninterruptible power supplies account for the bulk of data center outages (nearly 50 percent)2, it is critical to have a line of defense against unexpected power surges caused by direct and indirect lightning strikes because they can cause unforeseen damages to equipment.
Data Center Equipment
Remote lightning strikes, power surges from motors and generators and even natural geomagnetic disturbances such as solar flares can cause voltage transients in the feed line between equipment and the power supply. Transient voltage surge suppression for data center protection can come in many forms. Switchgear is often the first line of defense to divert power surges away from the facility. The uninterruptible power supply is another major form of transient voltage surge suppression in which instantaneous power is provided to the computers and data storage equipment via a flywheel or battery (depending upon the type of uninterruptible power supply) in case of an emergency when the power source fails. The redirection of massive amounts of energy requires major forms of heat dissipation. Surge protective devices are necessarily installed in power distribution units to safely break up massive surges without smoke or explosions.
Surge Protection for Data Center Interconnect (DCI)
Suppressing minor voltage surges has become increasingly critical with the advent of faster transmission speeds of Ethernet communication lines. Massive data centers are facing an increase in density requirements leading to tighter and tighter space and specification constraints — the 3.5 watts of power used to power the QSFP+ transceiver for a 40 Gbps throughput can power a QSFP28 transceiver for up to 100 Gbps. A surge that an older 10Base-T interface could handle would destroy a 100Base-T and beyond interface. The need to keep cost low has meant the Ethernet interface is now integrated onto the main printed circuit board assembly, increasing the risk of destroying the sensitive internal circuitry. Microelectronic integrated circuits can be affected on the transistor level, where the input lines are susceptible to damage from an electrostatic discharge (e.g., a static discharge).3 For this reason, fiber-optic cables are almost exclusively used inside buildings and inside campus installations despite the cost. Many data centers have even switched from direct-attach cables to active optical cable for the short link distances because of the direct-attach cables’ susceptibility to electromagnetic interference, representing another attempt to minimize noise sources for data center equipment.
Surge protection for Ethernet data lines is therefore essential to the success of a data center. Still, an Ethernet surge protector can, in and of itself, cause distortions that disrupt data and that cause costly latencies. Protective components have an inherent capacitance that can introduce distortions. This leads to a balancing act between making use of low-capacitive components while increasing the ability of the circuit to handle high energy surges — a 3.3-volt supply line for a microelectronic integrated circuit requires signal lines to have less than 5 picofarads of capacitances and is required to handle 1.2 Joules of energy3. Moreover, limitations in the printed circuit board layout can contribute to the degradation. Minor discontinuities can cause signal attenuation, phase distortion and a degraded common-mode rejection. In order to adequately protect the cable and connected equipment, two protectors can be attached on each end of the line. Naturally, this increases the insertion loss of the signal transmission and increases the possibility of introducing a ground loop into the line. Multi-stage surge protectors offer much higher levels of protection with the trade-off of minor signal loss (see Figure 2).4
Protection for PoE-based Systems
Power over Ethernet (PoE), which passes electric power along with data on twisted-pair Ethernet cabling, is now often used in smart security camera systems. It is especially critical to keep PoE protected from power surges that pose safety and security risks in the case of a failure. Smart cameras offer much higher resolutions than their coaxial-based counterparts in closed-circuit television applications. Because these cameras are far more modular, they can be used as stand-alone entities. Analog cameras must to be connected to a digital video recorder. The applications for these cameras range from industrial process monitoring to in-vehicle image processing.
PoE-based systems consolidate both the power and data communications over one line in which four of the eight lanes are often used to power a device and the other four transmit data. As defined in the IEEE standard 802.3, power can be delivered to a powered device one of three ways: endpoint power sourcing equipment (PSE), midspan PSE or extender PSE. Endpoint PSE sends power and data directly to a powered device from a power over Ethernet-based switch or hub, whereas midspan PSE, or a PoE injector, provides power and data from a non-PoE-enabled switch. Extenders simply increase the range of the PoE endpoint.
Both the PoE interconnect and the powered device must survive electrostatic discharge events and power transients that rise above Underwriter’s Laboratory (UL) standards. For instance, the UL standard for “Access Control System Units” lists that components (i.e., the power supply, door controller and card reader.) must be capable of suppressing a surge of 2,400 volts/12 amps on the output circuitry ― a difficult specification to accomplish with low-capacitance surge protection.5 It is therefore important that transient voltage protection be provided for both the PoE switch or PoE injector as well as the subsequent connected powered devices, such as the camera, card reader or door sensor (see Photo 1).
This is often accomplished with circuitry consisting of transient voltage suppressor diodes that can experience avalanche breakdown ― a phenomenon that allows for extremely large currents within materials. Thyristor surge protective devices are also avalanche-triggered diodes that can often feature better power handling capabilities because they are crowbar devices with a low on-state resistance that can carry large currents while maintaining a capacitance. Thyristor surge protective devices require a power reset after a surge, so they are generally not used in higher-voltage, low-speed signal lines, whereas transient voltage suppressor diodes may be the best option in a higher-powered DC system.
PoE Installation Considerations
Even with the surge protection equipment, the powered device’s voltage rating must be well above the clamping voltage of the transient voltage suppressor diode or thyristor surge protective devices. A transient voltage suppressor diode that can withstand up to thousands of volts, for instance, may not go into avalanche breakdown until almost 100 volts. A high enough voltage rating ensures that the powered device is not damaged by smaller and more frequent transient voltage spikes.6
Power-sourcing equipment for powered devices often comes with isolated DC-DC converters to minimize the risk of harming a user with exposed powered conductors as well as avoiding a ground loop. The surge protector itself would also probably need to isolate the cable shield that is at a different potential from the safety ground to prevent a ground loop. It is also important to note that PoE surge protection devices may have either Mode A (Power Pins: 1/2 and 3/6) or Mode B (Power Pins: 4/5 and 7/8) compatibility. Generally, most PoE-enabled switches are Mode A devices, and injectors are Mode B-compatible. Some surge protectors do offer both mode A and B compatibility for power.7
Surge protective devices are an indispensable tool to buffer the ill-effects of a direct and indirect lightning strike in a variety of applications. Although much thought and effort must go into preventing damage to life, this unpredictable surge of high energy can now cause a high degree of damage to increasingly sensitive solid-state devices. These solid-state devices not only require additional protection from induced voltage and current transients, but they also need surge protective devices with low capacitance to maintain their high data rates. It is therefore critical not only to redirect the bulk of the lightning surge from a building, but it also is increasingly important to protect internal equipment from lightning remnants. Legacy technology that had the benefit of withstanding minor surges has, for the most part, been replaced with much faster circuitry and tighter tolerances.
To read more of the April issue of AGL Magazine, Click Here
Manuel Martinez is product manager at L-Comm.