Backfires in Mercury Arc Rectifiers

Mercury arc rectifier design of today was arrived at in an endeavor to reduce the frequency of backfire. Efficiency and cost are sacrificed to obtain sufficient backfire immunity. Study of backfires in actual rectifiers and laboratory tubes indicates that this occurrence is a random one. Backfires differ in kind from other breakdown phenomena because they do not occur with certainty after some definite limitation has been reached, and in fact no limit can be chosen above which or below which backfires will or will not take place. The state of a rectifier in relation to backfiring must be described in terms of a backfire probability, or average rate of backfire. There are no commercial tank rectifiers for which the mean frequency of backfire is absolutely zero, but some probably reach the low value of one per several years. The significance of the random nature of backfire on the rating and testing of rectifiers is discussed, and suggestions are made as to means for testing which lead to the determination of the true quality of a rectifier. Data are given showing the variation of the backfire rate with current and voltage. The application of these ideas to research and development is illustrated by several examples. The theory of the formation of a backfire is described, and backfires are attributed to ``causes'' which lead to the formation of a cathode spot on an anode, but which do not correspond to any fundamental limitation of the rectifying process.