Surge measurement errors introduced by coaxial cables

Nearly all surge or impulse voltage measurements involve the use of a cathode-ray oscillograph (CRO) and measuring cable. If an ideal cable is assumed, i.e., one with no losses and a surge impedance which does not vary with frequency, conditions giving zero measurement error can readily be set up. However, all cables have some attenuation and their surge impedance is never the same for all frequencies.1,2,3 In low-voltage impulse work (where very short cables can be used) such factors may be negligible under certain conditions, but for measuring high-voltage surges a fairly long cable (50 feet or longer) is usually required and possible errors caused by the cable should be considered, especially for measuring steeply rising surges. The purpose of this paper is to discuss such errors and devise methods for measuring and minimizing them.