|
|
|
Board, Package and Die Thickness Effects Under Thermal Cycling ConditionsA model that enables the scaling of solder joint failure cycles for board and die or substrate thickness effects under thermal cycling conditions is discussed. Analysis Lab |
|
Authored By:Jean-Paul Clech EPSI Inc. Montclair, NJ, USA SummaryThis paper presents a first-order model that enables the scaling of solder joint failure cycles for board and/or die or substrate thickness effects under thermal cycling conditions. The model also allows for the prediction of failure cycles for solder joints of double-sided, mirrored assemblies based on failure data for single-sided assemblies. ConclusionsA first-order design-for-reliability tool / model has been developed that captures the impact of board and/or component thickness on solder joint reliability under thermal cycling conditions. The tool allows for scaling of failure cycles from one board thickness to another, or from one assembly configuration to another (single-sided or mirrored). Solder joint life under thermal cycling conditions goes as the inverse of the assembly stiffness. The model has been validated against thermal cycling results for 1 to 3.3 mm thick test boards and 0.2 to 3.7 mm thick components. The model also accounts for the global CTE mismatch between board and component, as well as the component size (DNP) effect. Other important parameters, such as stand-off height, assembly pitch and pad sizes are being added to the model (to be presented in a future publication). The model is algebraic and is easily implemented in a spreadsheet. This allows for quick assessments of the relative impact of changes in design parameters and material properties on solder joint reliability, independent of soft solder compositions. The board thickness effect is confounded with concurrent changes in in-plane moduli and CTEs. The interpretation of thermal cycling test results and the extrapolation of test data to field conditions rely heavily on realistic estimates, or better, measurements of board properties. In our updated survey of test board properties from multiple sources (Figure 19), board modulus is anywhere in the range 10 to 30 GPa and inplane CTEs are in the range 12 to 21 ppm/degree C. Effective board properties vary widely and cannot be assigned randomly when entered into reliability models. The stiffness or strength-of-materials approach, which is the foundation of the board and component thickness model, provides qualitative and quantitative insight into controlling parameters and competing factors: a) the board and component stiffness in flexure and under axial loading; b) the distance H that separates the neutral planes of the board and of the component. The latter increases the compliance of thicker board assemblies, a potential factor in some thick boards having longer solder joint lives than thin board assemblies as observed in accelerated thermal cycling and finite element simulations (Teng et al., 2002; Shih et al., 2004; Wu et al., 2014). These situations were not modeled in this paper because relevant board and package material properties were not available. Initially Published in the SMTA Proceedings |
|
Comments
|
|
|
|
|
|
|