PROJECT TITLE :

Yield Prediction Through the Event Sequence Analysis of the Die Attach Process

ABSTRACT:

Die attach is the process of mounting a plurality of dice to a printed circuit board (PCB) or substrate. Die attach is crucial to the thermal and electrical performance of semiconductor merchandise, considerably affecting the final yield of PCBs. In general, the die attacher records alarm events, modification events, and maintenance events in a very log. Alarm events occur when dice are not aligned well to the mounting positions on a PCB. Modification events are recorded when product sorts are changed or raw materials of different suppliers are introduced. Maintenance events are recorded whenever the staff conduct corrective actions because of alarm events. We tend to empirically observed that completely different sequences of events have totally different effects on the final yield. In this paper, we propose a data mining approach that predicts the ultimate yield of a PCB using the event sequences recorded in the log of the die attacher. We have a tendency to propose a predictive association rule considering the event sequence (PARCOS) algorithm that creates a group of rules, in which every rule estimates the yield for a sequence of events. An experiment with a piece-web site dataset demonstrated that the PARCOS algorithm had a yield prediction accuracy that was a minimum of nine% on top of those of the regression models that didn't contemplate the event sequences.