One of the key challenges in achieving integration between design and manufacturing activities for cost-effective and timely product development is simulation of the science of key manufacturing processes. A virtual-factory environment which allows modeling and proactive optimization of these processes is the focus of this paper. Such a modeling environment will allow designers and manufacturing engineers to integrate design and manufacturing considerations to assess reliability and supportability early in the development cycle. Within such an integrated development environment, different design options can be evaluated in a timely manner, and candidate process windows can be identified and optimized up-front before the first prototype is built. The impact of design changes on manufacturability and the impact of manufacturing processes on product attributes can now be quantitatively evaluated. Specifically, this paper identifies modeling strategies available for several key manufacturing processes in the surface mount assembly technique. Models like these form the backbone of any mechanistic process simulation tool. Potential links between these individual process models are discussed, in order to allow proper integration of the entire process.

Modeling strategies currently being labeled by industry as “virtual-factory” tools, are mostly empirical models based on statistical, neural-net and fuzzy-logic techniques. The limitations of these approaches are related to the cost and time required to “train” or “teach” the model (which is essentially a model calibration process), and to the inability to extend a model trained on one technology to a vastly different technology without extensive retraining. These limitations can be partially alleviated by developing more mechanistic process models based on the first principles of the fundamental science of the processes. The basic approach is to understand and model the structure-processing-properties relationships for each process in the manufacturing cycle. The model becomes more and more generic as it focuses increasingly on the fundamental principles of the manufacturing process.

While the basic principles of virtual-factory environments are quite generic, we focus on surface-mount technology (SMT) for CCA/MCM design and manufacturing processes, as an example case study. Integrated modeling approaches and results of sample parametric sensitivity studies are presented for key SMT processes such as lead forming, solder paste evaluation, PWB warpage evaluation, solder paste printing, solder reflow, and solder joint geometry evaluation.

The lead forming model predicts the elastic springback at the end of the forming process, and predicts the state of residual stress which governs fatigue reliability. The PWB warpage model predicts the maximum warpage likely to occur during the reflow cycle, since this can affect the uniformity of solder joint geometry. The solder paste rheology assessment model is based on the rheology of the binder and the metal content. The solder printing model predicts the amount of solder paste that is likely to get deposited if the stencil opening geometry and the paste rhcology are known. The solder joint geometry prediction is based on the lead and bond pad geometry, and solder paste volume. The reflow model predicts the temperature attained at the joint, for a given oven temperature history and a given assembly. The goal in all of these models is to be able to track the effect of any of the associated key manufacturing parameters on the quality of the interconnect formed.

This content is only available via PDF.
You do not currently have access to this content.