Trusted Data Acquisition in Microelectronics Manufacturing as a Basis for ML-Optimized Processing

Abstract

In the digitization of manufacturing processes, one of the major goals is the connection of production facilities and the use of data to digitize business processes. In order to optimize manufacturing processes and maximize the quality of the resulting product, further process information and data directly from the work piece and from the manufacturing environment are required to achieve a holistic system view in addition to the selected data that the manufacturing systems already provide. Within the SiEvEI 4.0 project, a research consortium from industry and academia works on process digitization for a manufacturing scenario where high value electronic goods are built in a distributed manufacturing environment. The key research topics addressed are the implementation of a Chain of Trust [CoT] for trustworthy distributed manufacturing and the application of artificial intelligence/machine learning to analyze and eventually optimize manufacturing processes. The basic concepts of this approach have been presented at IMAPS 2021. As an update, this paper reports on the actual experimental evaluation of these concepts in two different assembly lines, including data acquisition, data handling and AI processing with the goal to optimize processing targeting higher production yield and product quality. Specific for this work is the focus on high mix/low volume SMD assembly using fully automated equipment and Solder Ball Application. As a result, the paper presents the experimental validation of the manufacturing process digitization and the use this digital description of a process combination to make a distributed manufacturing flow safe and increase product/process quality.