Colin Dankwart is the team leader for performance services at ficonTEC, specifically concentrating on edge computing, machine learning, and other techniques for optimizing machine performance. He has a background in optics and optimization algorithms, having worked with them for nearly a decade prior to joining ficonTEC in 2017. Having gained extensive experience in applying this knowledge to optical test and assembly and extending his expertise to motion systems, he focuses on improving the performance of ficonTEC machines using data-driven approaches.
High volume manufacturing of photonic systems requires assembly machines with high throughput, yield and availability. To meet this demand, the motion systems involved must offer high accuracy combined with high motion speed. At the boundary of performance, these factors often oppose each other, making it necessary to find the optimum trade off. Under changing conditions in the field, finding & maintaining this optimum can impose significant challenges to local operators & engineers alike. We present an approach that automatically adapts the motion system to changing conditions without human intervention. Thereby, best possible yield can be achieved without sacrificing throughput.