From measurements made on communication lines and on utility power lines, both outside and within structures, ‘‘standard’’ waveforms have been derived for designing protective systems. Nevertheless, any given lightning strike, particularly if it is very close, may produce a ‘‘non-standard’’ transient waveform. Communication lines generally have a different exposure to lightning than power lines. Communication lines are often mounted beneath distribution power lines, and communication lines are usually shielded and bundled. Based on the measurements of Bodle and Gresh (1961), reasonable worst-case over voltages to which telephone lines are exposed can be expected have a rise time to peak value of 10 ms and a time to decrease to half of peak value of 1000 ms. Such a transient is called a 10/1000 ms waveform. This voltage waveform does not represent a typical transient, but rather a composite worst case specified for testing the lightning immunity of communication systems. Figure below shows a plot of the first 60 ms of the 10/1000 ms waveform. The 10/1000 ms waveform is found in a number of standards, as are similar waveforms such as the 10/700 ms. One standard for gas tube arresters calls for the 10/1000 ms waveform with a peak current of 500 amperes. The peak voltage/peak current found in other representative standards is 1500 volts/40 amperes, 800 volts/100 amperes, and 1500 volts/200 amperes.
While studying the failure of household electric clock motors, Martzloff and Hahn (1970) found that for a clock motor insulation level of 6000 volts, the motor failure rate was 1 percent of the failure rate that occurred with a 2000 volt insulation level. The implication of this observation is that there are few transients on the 60 Hz 120 volt service with a maximum value greater than 6000 volts, but there are a considerable number of transients greater than 2000 volts. Because of the spacing between wires and other metal components, the insulation level in household electrical wall outlets and circuit breaker boxes is generally between 5000 and 10 000 volts. Thus one would not expect to observe higher level transients inside a structure because of prior electrical breakdown at these weak points. In other studies of transients on power lines inside structures, it was found that most transients recorded on 120 volt lines were in the range of several hundreds of volts with the upper 1 percent level in different studies being between 300 volts and 2000 volts. Lightning transients in the thousand-volt range on a structure’s 120 volt lines are probably due to lightning within a few hundred meters of the structure, an event that occurs several times a year to most structures in the United States.
The most common waveforms specified in standards for designing lightning protection of low-voltage power lines within structures (as well as for designing protection for high-voltage equipment on distribution and transmission lines) are the 8/20 ms current waveform and the 1.2/50 ms voltage waveform. The 1.2/50 ms voltage waveform is shown in Figure. The 8/20 ms current waveform and the 1.2/50 ms voltage waveforms are intended to approximate the direct effects of a lightning first return stroke which is typically responsible for the largest magnitude transient during a lightning discharge. Note, however, that the rise time of the current in subsequent return strokes (return strokes following the first stroke) is generally less than 1 ms, and thus this rapidly changing transient signal is not accounted for in these standard waveforms. Within structures, the maximum voltage specified for testing is generally about 6000 volts (because, as noted above, other elements of the power system will suffer electrical breakdown first, preventing higher voltages), and the maximum current is about 3000 amperes. In addition to the unipolar voltage waveforms shown in Figure, a so-called ring waveform, based on the measurement of Martzloff and Hahn (1970), is specified in some standards. The ring waveform of Martzloff and Hahn (there are other ring waveforms) is also shown in Figure. Martzloff and Hahn observed that transient overvoltages on low-voltage systems often have a rise time to peak that is less than 1 ms, followed by a decaying oscillation at a frequency of about 100 kHz. One standard sets the peak current for the ring waveform at 500 amperes and the peak voltage at 6000 volts. Individual SPDs are rated by the maximum continuous operating voltage (rated voltage) they can withstand without failure, the total transient power and energy they can withstand without failure, the peak current they can withstand without failure, and the peak voltage that will appear across the arrester terminals when a current of a certain magnitude and waveform is passed through the arrester.