Storage Reliability Of Electronic Components

For military electric equipments that "long-term storage, once use", the non-reliability of electric components play an important part in them. With carrying out the long-term storage experiments, a mainly purpose is to evaluate the component non-operating reliability, as well as to provide design conferences and guidelines and basic data for the equipment reliability.First, environmental effects on electric components are introduced. The thermal effect, chemical effect and radiative effect are discussed in detail. The inner material thermal mismatch of components is an inherent reason to cause thermal effect;furthermore, the storage ambient temperature change is the external cause. Chemical effect may erode the pins, packages, internal metal system, Al-Au contacts and inner bonding of the device.Second, Two kind of methods for electric components non-reliability evaluation, storeroom storage evaluation and accelerated stress method are discussed. Meanwhile, the experiment technique in this study is illustrated in detail. To sum up the current experiment results, the component failure rate is obtained. Beside, whether the functional parameters of the device during storage period are diversified distinctly was analyzed, as well as whether there exist significant differences among four storage regions.Third, the failure modes and failure mechanisms of electric components under the non-operating condition is investigated. Experiment results indicated that the corrosion, bonding failure, fixture and package failure are the mainly causes to induce components failures. According this failure modes and failure mechanisms exposed during experiment, many corresponding improvements in design and manufacture stage were put forward to enhance the inherent reliability of the devices.Finally, since the storage data processing methods and storage life predicting techniques have been the barrier of the study, there are no perfect techniques forreference. Considering the advantage of other methods applied in prediction, the gray system theory and BP neural network were introduced in this study to predict the storage life. The prediction results indicated that it has certain precision and it may explore a new technique on the storage life prediction. But those two methods are based and localized on the data statistical processing. It should consider more the combination with failure physics model.