Abstract: A probabilistic method is utilized to predict solder joint reliability of surface mounting devices. Chip component 0603 is taken as an example. Based on the simplified joint geometry model and Coffin-Manson relation, equations and solutions of thermal fatigue life are established and validated to determine the probability of failure for lead-free solder joints. Uncertainties in the solder’s geometry are often related to manufacturing processes, and the ambient temperature varies due to the diversity of end-users. Using the approach, the influence of each uncertainty on the thermal fatigue life is quantified. This information can be used to optimize the product design, establish acceptance criteria for quality control, and predict after-sales product quality. The result shows that the thermal fatigue life of the solder joints decreases significantly with the increase of random variables. The key aspects of the normally distributed random variables, solder joint geometry, H, L and thermal range Δ are involved to predict the low cycle thermal fatigue life according to the experimental data of small sample with measured standard deviation and mean value.

Key words: Lognormal distribution, Probabilistic analysis, Random variable, Solder joint, Thermal fatigue