Solder joint failure history has generally been assumed to follow a straight line when plotted as a lognormal or 2‐parameter Weibull distribution. Test results presented here show that a deviation from straight‐line behaviour occurs at low percentage failure probabilities. This indicates that solder joint failure history is more correctly characterised as a 3‐parameter Weibull distribution with a failure‐free period of life for true wearout failures. The solder joint failure distribution characteristic is also affected by applied strain. Lower strain, in addition to increasing median life, also improves the distribution such that the number of cycles‐to‐first‐failure is increased compared with the median cycles‐to‐failure. The ratio of cycles‐to‐first‐failure/median cycles‐to‐failure and apparent Weibull slope increases as strain decreases in a predictable manner. The effects of part elevation, part size, solder joint volume and shape, conformal coating, temperature differential, and alternative board materials are also presented with test data showing the effect of variation of these parameters.
[1]
R B Abernethy,et al.
Weibull Analysis Handbook
,
1983
.
[2]
Keith C. Norris,et al.
Reliability of controlled collapse interconnections
,
1969
.
[3]
Donald S. Stone,et al.
Deformation and Fracture of Pb-Sn-Eutectic Under Tensile and Fatigue Loading
,
1992
.
[4]
Z. Mei,et al.
Fatigue Lives on 60Sn/40Pb Solder Joints Made With Different Cooling Rates
,
1992
.
[5]
H. D. Solomon,et al.
Predictions of solder joint fatigue life
,
1990,
40th Conference Proceedings on Electronic Components and Technology.
[6]
D. S. Steinberg,et al.
Vibration analysis for electronic equipment
,
1973
.