Vacuum Breakdown Voltage Distributions Between Different Contact Arrangements and Materials

Abstract

Although electrical breakdowns are often reported as a single value, in general breakdowns for both vacuum and gases are in fact a distribution of values. This work measures the breakdown voltage distribution for two general contact arrangements (floating and fixed shield) using Cu-Cr contact material. Tests were performed using the lightning impulse voltage pulse, with a 1.2/50 ps rise/decay time. The breakdown distributions were acquired by the always breakdown method. In the always breakdown method, the voltage is set to a high enough level that the contact gap always breaks down, providing a datapoint for each voltage application. Breakdown rates were measured from a few percent up to 100%. The data was compared to an empirical model based on the Weibull distribution and a model taken from work on linear accelerators. The Weibull distribution with a high shape parameter gave good agreement with the data across a wide range of measurements. A model taken from linear accelerator data at very low breakdown rates also successfully matched the data and the calculated Weibull distribution for breakdown rates <10%, using similar parameters in both models. The extremely steep dependence of the breakdown rate on the voltage necessitates a vacuum breakdown model that can explain both the occurrence of the breakdown and the dramatic change in the frequency of the breakdowns. A comparison with a broad range of other published data demonstrates that detailed comparisons benefit from determining the breakdown distribution and/or a clear choice of the breakdown rate being compared. Comparisons focusing on general trends and broad ranges of parameters can be more flexible.