RELIABILITY ASSESSMENT BASED ON ACCELERATED DEGRADATION - A CASE-STUDY

被引:102
作者
CAREY, MB [1 ]
KOENIG, RH [1 ]
机构
[1] AT&T BELL LABS,READING,PA 19604
关键词
QUALIFICATION; DEGRADATION DATA; PARAMETRIC DATA; NONLINEAR ESTIMATION;
D O I
10.1109/24.106763
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
This paper describes an analysis strategy (experimental and analytic) to extract reliability information from the measured degradation of devices submitted to elevated stress. The strategy is applied to the estimation of the reliability of an integrated logic family (ILF). The ILF is an element of the Supervisory Logic Circuit, a component of a new generation of submarine cables. The experiment consisted of monitoring several electrical parameters while aging the devices under elevated temperature. We model changes observed in one of these electrical parameters (propagation delay) and identify and quantify various sources of variability observed in the data. We use these models to predict (with 95% confidence) that a randomly selected device operating at 40-degrees-C will see a change of no more than 2 nsec in propagation delay in 25 years, the lifetime of a submarine cable. An analysis of the degradation rate leads to the conjecture that the observed change in propagation delay is due to the diffusion of impurities (Na) through the bulk of the oxide layer (SiO2).
引用
收藏
页码:499 / 506
页数:8
相关论文
共 6 条
  • [1] [Anonymous], 1988, NONLINEAR REGRESSION
  • [2] Beadle W. E., 1985, QUICK REFERENCE MANU, P1
  • [3] NONLINEAR MIXED EFFECTS MODELS FOR REPEATED MEASURES DATA
    LINDSTROM, MJ
    BATES, DM
    [J]. BIOMETRICS, 1990, 46 (03) : 673 - 687
  • [4] ONEILL R, 1971, J ROY STAT SOC C-APP, V20, P338
  • [5] Seber G.A.F., 1989, NONLINEAR REGRESSION, DOI DOI 10.1002/0471725315
  • [6] Shewmon P., 2016, DIFFUSION SOLIDS