No abstract

This contribution describes the development of a fuzzy‐based process control system for the soft process drying of natural material foils. The system should maintain constant product moisture to facilitate automatic treatment of the foils in subsequent processing steps. A pilot plant was constructed to support the development of the control system. The core of the experimental setup is an optical measuring device, that provides for the simultaneous in‐line determination of drying foil water content and thickness at different places in the drying plant. The sensor possesses a measuring range of 10 %–80 % for water content (X*) and 50 μm–500 μm for thickness. Compared to a reference method, the system shows an accuracy of ± 3 % for the water concentration and 3.5 % for the foil thickness. No significant drift in the output signal could be detected for 5 weeks. These quantities, together with other process information, were used as inputs for a multistage fuzzy control system, which adjusts the outputs temperature and velocity of drying air and belt velocity for optimal soft process drying. In normal operation without the controller, foil moisture variations of 7 % over a period of 2 hours were obtained, while applying the fuzzy logic controller decreased the variations to 4 %. After disturbances of up to 17.5 % in initial viscosity, 6.8 % in concentration and 20 % in foil thickness variations, outlet moisture content variations of only 6.2 % were attained.

Printed circuit board (PCB) surface finish and its reactive wetting process were identified as a main cause of voiding. During reflow soldering liquid flux residues adhered to the interface PCB / liquid solder alloy form voids while the solder alloy is in a liquid state. In contrast to that no voids are found when no reactive wetting occurred. Furthermore, we observed that the number of detected voids after the reflow process strongly depends on the tin layer thickness: Many voids are formed in the case of a thin tin layer, no voids are found in the case of a very thick tin layer. A model is established to explain this phenomenon.

No abstract