Theoretical analysis of longevity testing on bubble memory devices

This paper reports theoretical study results about magnetic bubble device long-term reliability testing. The bubble during propagation along permalloy tracks will be represented by a simple, one dimentional stochastic model. An equation to describe fluctuation in cylindrical bubble radius is approximated in the Langevin type stochastic differential equation, in which a set of small effects, such as interaction among bubbles and crystal nonuniformity, are considered as a white noise forcing term. Estimating the average time to bubble annihilation or run-out (bubble memory mean time to failure) is reduced to a level-crossing problem for a random process. Calculated bias field margin degradation shows good agreement with experimental results for an actual bubble device. Bubble material parameters for obtaining maximum operation time can be suggested. Furthermore, for problems where analytical solutions for the proposed model are difficult to get, Stochastic Hybrid Computer application is described with a practical example.