Maturity Factors in Predicting Failure Rate for Linear Integrated Circuits

Abstract

There are differences between failure rates obtained from test results, and failure rates obtained from the failure rate models. In some cases, failure rates obtained from failure rate models are 10 to 50 times larger (worse) than failure rates obtained through testing of the same devices. This indicates that the available failure rate models need some modification. It is not the intent of this paper to disqualify the available failure rate models. This paper suggests only that the available failure rate models be modified by including two maturity factors: ? date code factor (technology maturity in general) ? manufacturer's factor (maturity of specific manufacturer) This paper is based on tests performed on more than 18000 linear ICs (integrated circuits) manufactured by National Semiconductor and on comparative tests performed on more than 2500 linear parts from other IC manufacturers. This paper offers date code factors which could be used in prediction of failure rates for linear IC's. It also offers manufacturer's factor for devices manufactured by National Semiconductor. However, offered date code factors and manufacturer's factor are only indicative, and given values should be confirmed or improved by further investigations. This paper deals mostly with prediction of failure rates for linear ICs. But the same approach for developing date code factors and manufacturer's factors could be used for developing similar factors for other ICs (besides linear), and for other manufacturers.