Automatic quality inspection of printed circuit boards using a cobot

Automatic quality inspection of printed circuit boards using a cobot
Mirmex Motor, a Belgian company specialising in the production of customised DC voltage motors, knocked on our door to automate the manual quality control steps in their production process. The inspection process involves very repetitive and time-consuming tasks such as electrical tests and visual inspections by operators. Therefore, Mirmex partnered with Flanders Make to find automated systems to improve efficiency.
The challenge lay mainly in the unique characteristics of the motors: thin, flexible and with different dimensions. Moreover, the visual defects that needed to be detected were often very small. Mirmex envisioned an automated solution that could perform both visual and electrical inspections of the motors, with the ability to sort them based on the results.
‘The workload is lower, while our quality control became a lot more effective.’
Mirmex Motor
What did Flanders Make do for Mirmex Motor?
To meet these expectations, we conducted a feasibility study. We selected a cobot with a two-finger gripper for the final electrical inspections, while we developed a specially designed work cell for visual inspection at the beginning of the production line. This work cell contained a stack of motors that we placed in front of the cobot, which picked them up individually using a gripper with suction cups. We modified the gripper using off-the-shelf pneumatic elements to reduce costs and increase flexibility. After picking up, each motor was placed on a conveyor belt, flattened with a Plexiglas cover and inspected with a linescan camera system and a modified computer vision algorithm.
Our feasibility study showed that the proposed system successfully picked up the motors. The system operated within Mirmex's speed requirements, confirming its effectiveness. From the employees' point of view, the implementation of the automated system significantly reduced the workload. Previously, inspection tasks were performed manually, but with the cobot and computer vision system taking over these tasks, we reduced the workers' involvement to verifying uncertain inspection results, saving them a lot of time.