This research focuses on developing a model to predict the yield of a printed circuit board manufactured on a given assembly line. Based on an extensive literature review as well as discussion with industrial partners, it was determined that there is no tool available for assisting engineers in determining reliable estimates of their production capabilities as they introduce new board designs onto their current production lines. Motivated by this need, a more in-depth study of manufacturing yield as well as the electronic assembly process was undertaken. The relevant literature research was divided into three main fields: process modeling, board design, and PCB testing. The model presented in this research combines elements from process modeling and board design into a single yield model. An optimization model was formulated to determine the fault probabilities that minimize the difference between actual yield values and predicted yield values. This model determines fault probabilities (per component type) based on past production yields for the different board designs assembled. These probabilities are then used to estimate the yields of future board designs. Two different yield models were tested and their assumptions regarding the nature of the faults were validated. The model that assumes independence between faults provided better yield predictions. A preliminary case study was performed to compare the performance of the presented model with that of previous models using data available from the literature. The proposed yield model predicts yield within 3% of the actual yield value, outperforming previous regression models that predicted yield within 10%, and artificial neural network models that predicted yield within 5%. A second case study was performed using data gathered from actual production lines. The proposed yield model continued to provide very good yield predictions. The average difference with respect to the actual yields from this case study ranged between 1.25% and 2.27% for the lines studied. Through sensitivity analysis, it was determined that certain component types have a considerably higher effect on yield than others. Once the proposed yield model is implemented, design suggestions can be made to account for manufacturability issues during the design process. / Master of Science

Manufacture testing of digital integrated circuits is essential for high quality. However, exhaustive testing is impractical, and only a small subset of all possible test patterns (or test pattern pairs) may be applied. Thus, it is crucial to choose a subset that detects a high percentage of the defective parts and produces a low defective part level. Historically, test pattern generation has often been seen as a deterministic endeavor. Test sets are generated to deterministically ensure that a large percentage of the targeted faults are detected. However, many real defects do not behave like these faults, and a test set that detects them all may still miss many defects. Unfortunately, modeling all possible defects as faults is impractical. Thus, it is important to fortuitously detect unmodeled defects using high quality test sets. To maximize fortuitous detection, we do not assume a high correlation between faults and actual defects. Instead, we look at the common requirements for all defect detection. We deterministically maximize the observations of the leastobserved sites while randomly exciting the defects that may be present. The resulting decrease in defective part level is estimated using the MPGD model. This dissertation describes the MPGD defective part level model and shows how it can be used to predict defective part levels resulting from static defect detection. Unlike many other predictors, its predictions are a function of site observations, not fault coverage, and thus it is generally more accurate at high fault coverages. Furthermore, its components model the physical realities of site observation and defect excitation, and thus it can be used to give insight into better test generation strategies. Next, we investigate the effect of additional constraints on the fortuitous detection of defects-specifically, as we focus on detecting dynamic defects instead of static ones. We show that the quality of the randomness of excitation becomes increasingly important as defect complexity increases. We introduce a new metric, called excitation balance, to estimate the quality of the excitation, and we show how excitation balance relates to the constant τ in the MPGD model.

Manufacture testing of digital integrated circuits is essential for high quality. However, exhaustive testing is impractical, and only a small subset of all possible test patterns (or test pattern pairs) may be applied. Thus, it is crucial to choose a subset that detects a high percentage of the defective parts and produces a low defective part level. Historically, test pattern generation has often been seen as a deterministic endeavor. Test sets are generated to deterministically ensure that a large percentage of the targeted faults are detected. However, many real defects do not behave like these faults, and a test set that detects them all may still miss many defects. Unfortunately, modeling all possible defects as faults is impractical. Thus, it is important to fortuitously detect unmodeled defects using high quality test sets. To maximize fortuitous detection, we do not assume a high correlation between faults and actual defects. Instead, we look at the common requirements for all defect detection. We deterministically maximize the observations of the leastobserved sites while randomly exciting the defects that may be present. The resulting decrease in defective part level is estimated using the MPGD model. This dissertation describes the MPGD defective part level model and shows how it can be used to predict defective part levels resulting from static defect detection. Unlike many other predictors, its predictions are a function of site observations, not fault coverage, and thus it is generally more accurate at high fault coverages. Furthermore, its components model the physical realities of site observation and defect excitation, and thus it can be used to give insight into better test generation strategies. Next, we investigate the effect of additional constraints on the fortuitous detection of defects-specifically, as we focus on detecting dynamic defects instead of static ones. We show that the quality of the randomness of excitation becomes increasingly important as defect complexity increases. We introduce a new metric, called excitation balance, to estimate the quality of the excitation, and we show how excitation balance relates to the constant τ in the MPGD model.