Board-to-Board Connector Inspection
Ensure BTB connections between PCBs are functional
Graphical programming environment for deep learning-based industrial image analysis
Board-to-Board (BTB) connectors provide signal connections between two printed circuit boards (PCBs) without the use of a cable, which saves space in tight configurations. BTB connectors have pins on one side and contacts on the other that must match the corresponding terminals on the PCB. Each consists of a molded plastic base that holds many metal contacts.
BTB connectors with damaged elements or contaminants can pass electrical testing and be sent out as acceptable. Such parts are often unreliable in use and cause intermittent faults that are hard to debug in the field. Visual inspection is more reliable and flags such minor defects.
The molded base can suffer from a variety of defects, including burn, short shot, dust, scratches, deformation, and foreign inclusions, as well as misplaced, bent, or missing pins or contacts. Many of these defects are hard for the human eye to identify at the high speeds and volumes necessary for inspection.
Manual inspections are good at detecting cracks or molding defects in these connectors but can only check a sample of parts at the required speeds. The typical inspection procedure is for automated optical inspection (AOI), using traditional rule-based vision tools, to inspect each connector, followed by human inspection of a sampling of the connectors.
AOI machines can have a high false positive, or overkill, rate, while manual inspection has a low throughput rate, even on the relatively small sample inspected.
Cognex Deep Learning increases both volume and accuracy, which helps meet market demands. Cognex Deep Learning’s defect detection tool trains on a set of labeled images of both good and bad BTB connections. It then reliably detects and marks anomalies anywhere on the connector, ensuring that only connectors without defects move on to board assembly.
BTB connector inspections require handling unpredictable variations, and Cognex Deep Learning within the AOI machine can more quickly and accurately identify these variations. Compared to traditional rule-based machine vision, deep learning can maintain high-speed inspections even under rigorous throughput requirements.
This means that all products can be passed through the AOI machine for inspection, without requiring a follow-up statistical sample by human inspectors. AOI-based inspection speed is twice that of a human inspector and the accuracy rates eliminate the need for most human inspectors.