An Investigation of Overvoltage Effects on a Gas-Blown Spark Gap

This experimental investigation addresses the anomalous overvoltage breakdown and subsequent recovery characteristics observed in gas-blown spark gaps. Previous investigations have reported a two-stage recovery process in self-break, pulse-charged, gas-blown spark gape in which the spark gap recovers first to an intermediate breakdown strength before recovering to the undisturbed breakdown voltage. The recovery process has been explained in terms of two different characteristic times. The first is that required to obtain initial recovery in which the spark gap regains its intrinsic breakdown strength. The second time is that required to obtain final recovery in which the pulse-charged spark gap regains the overvoltage breakdown strength associated with the voltage increase during the statistical delay. In this interval, the spark gap operates on the "plateau" (a region of constant, highly stable breakdown strength) which has been correlated with the intrinsic breakdown strength of the pulse-charged spark gap when suitably preionized. Both the time to initial recovery and final recovery have been found to be functions of gas velocity. Initial recovery is controlled by the gas density in the interelectrode volume. After an initial breakdown the resulting hot arc debris precludes the spark gap from exhibiting its full intrinsic strength. By convecting this debris downstream, the electrodes are cleared of the arc debris much faster than the gas naturally recovers As a result, this investigation has demonstrated a 10 kHz repetition rate (two-pulse) at 100X of the spark gap s intrinsic breakdown strength with a gas velocity of 360 m/s (Mach 1). The velocity dependence of final recovery has also been studied . It is clear that final recovery occurs when the statistical delay before breakdown is no longer precluded by the effects of the first pulse. Two mechanisms have been proposed to explain the reduction of the statistical delay on the plateau. The first is thermionic emission of electrons from the hot electrodes. The second is boundary layer entrapment of easily ionized impurities. The consequences of gas flow on each mechanism have been estimated and compared to experiment. At this time, only the boundary layer impurity hypothesis correctly predicts the observed inverse relationship between the time to final recovery and the gas velocity.