Published In October 2000
Fire damaged a three-story, single-family frame residence in New Jersey in 1985. Certain physical and circumstantial evidence suggested the possibility that lightning may have caused the fire. Lightning may serve as an ignition source of a fire by directly striking an object; by an arc discharge between two conductive objects at different induced potentials; by a current surge in circuitry and equipment resulting from an induced voltage; by the flow of substantial electrical current, which causes overheating, melting, or vaporizing of metal; or by arcing of lightning current from conductors at high-resistance grounds or grounding conductor bonding points. Lightning strikes, or discharges of massive charges of static electricity, cause about 2 percent of all fires. Lightning and consequent power system surges can result in voltage levels sufficient to cause failures of electrical equipment. The photograph depicts inside telephone wiring that exhibits a typical lightning-damage pattern. A lightning discharge to the ground has an overall potential difference of 100 to 1,000 kV, with an average peak current of 20 kA. Lightning strike currents reach their peaks in 1 to 10 microseconds, and decay exponentially to one-half their peak values in 10 to 100 microseconds. Lightning protection includes devices for conducting lightning energy into the earth without damage to structures, apparatus, or personnel. Each piece of electrical equipment has an inherent basic impulse level (BIL), above which it will be damaged by a lightning surge. If equipment provided with surge protection was nevertheless damaged, its BIL protection margin may have been inadequate because the rise time to crest of the lightning surge was much shorter than the nominal assumed standard. Such decreased rise times can increase surge arrester discharge voltages by 10 to 30 percent. In this case, the fire originated in the extreme southwest corner of a rear cellar extension, in the ceiling below the first level flooring. A low burn was found in the southwest corner of the rear first floor den. The amount of damage to the beams at the origin of the fire corresponded to a smoldering time beginning when an earlier fire alarm for a nearby residence was received after a lightning strike. At that time, a circuit breaker for a branch circuit in that residence had tripped open. That circuit breaker fed an outdoor light, which had been damaged by the lightning strike. Fire damage was concentrated where the telephone drop entered the burned residence. This area was located at least 30 feet from the electrical power service entrance panel. Four residences near the burned house had inoperative telephone lines and “blown” carbons in their telephone protectors. The cover of a telephone terminal box for the service drop supplying the burned house, mounted on a pole, had been blown off; the faceplate under the cover was blackened. The telephone drop cable for the burned residence was found lying on the ground. No evidence of a telephone protector or grounding conductor was found after the fire either outside, or in the basement where the telephone service drop entered through the basement wall. My analysis included the initial assumption that a protector had been present indoors, and that its remains could not be found due to fire damage. I then concluded that the cause of the fire, and departures from good installation practice and applicable industry standards that contributed thereto, were the following: In the fire-damaged residence, lightning energy that entered the basement on the telephone service drop ignited flammable structural material and caused the fire there. Assuming that a telephone protector and grounding conductor were in the residence during the lightning storm, and that because of post-fire activities they were not recovered, it was concluded that this equipment and/or installation was defective, since the lightning stroke did not contain an unforeseeably great amount of energy. This was established by the lack of significant lightning damage to any other nearby residence fed off the same telephone feeder. In the telephone company’s “Standard Practices” documents, the fusing characteristics of telephone distribution drop wire, such as 19-gauge copper, were compared with the safe time-current capabilities of the types of protectors in use in that neighborhood. Approximately twice as much current was required for equal exposure times in order to exceed a protector’s safe time-current capability. It was concluded that the residence lacked a surge protector and grounding conductor required by NEC Rule 800-2. Thus lightning energy started the fire immediately inside the basement wall. It is standard practice among telephone utilities that cable and communication wire pairs be coordinated with the selection of surge protectors so that the wires will fuse open before the time-overcurrent damage limit of the protector is exceeded, thereby eliminating a possible fire hazard in a customer’s premises. The alternative would be for the protector to blow up with consequential premises damage. Telephone protectors in nearby residences that were subjected to similar and probably greater levels of lightning energy exhibited no evidence of damage. A telephone protector, if indoors, would have been mounted directly on or near combustible materials in the basement, and would have been the source of ignition during the lightning strike. Many standards required that protectors not be installed near easily ignitable materials. These included Rule 800-2b of the 1975 NEC, and the “Standard Practices” of various telephone companies. The telephone drop cable leading to the residence was affixed to trees, which does not comply with Rule 280 of the 1984 National Electrical Safety Code (NESC). This method of drop cable attachment made it more likely that lightning would enter the house. The telephone company’s “Standard Practice” stated that tree attachments are a potential source of trouble, and are therefore not desirable. The telephone strand (messenger cable) was not bonded to the power company’s multi-grounded neutral. NEC Rule 800-31(h)(7) and NESC Rule 99C require bonding. The hazards of separately grounded, unbonded telephone and power grounding systems, because of potential differences between them, are well known. This bond was installed after the fire. Recent repairs to the bare conductors of a nearby overhead power distribution line were near the telephone cables. Assuming a lightning strike on the power lines, which then flashed over to the messenger-supported telephone lines, caused this damage, proper bonding to the power company multi-grounded neutral and appropriate installation and placement of telephone protective devices could have prevented the fire. No fuses, as required by telephone company installation standards, were found in the telephone lines between where they were exposed to power voltages and the burned residence. Conforming closely to guidance provided by recognized standards, such as the NEC, is a major step for avoiding catastrophic electrical failures. Only statistically infrequent, substantial lightning surge energies can’t be discharged harmlessly by well-established protection methods. MAZER is a consulting electrical engineer who currently specializes in electrical safety issues. His telephone number is 202-338-0669, and his e-mail address is email@example.com.