Void Inspection of Complex Assemblies using Computed Laminography
By Jeff Urbanski, Yxlon Sales Account Manager and Peter Koch, Electronics Project Management While speaking with and visiting many electronics manufacturing plants in the United States, we have heard dissatisfaction with their ability to measure porosity accurately. They stated the inability to detect flaws in high-density, double-sided boards because of more fine pitch, complex components and assemblies. Despite the general trend of microchipminiaturization, many manufacturers we work with continue to build very large PCBs. As their products continue to be optimized for the highest performance, many EMS providers are left with dated equipment that provide vague low-quality images of their products. For this reason, products are no longer passing audit, making them unsuitable for their end customerswho are incorporating their products into critical end-products, often in the medical device, communications, aerospace, or automotive segments. Both small and large electronics manufacturing companies are at-risk to produce low performing products if they are not able to verify their product quality.
Several types of components exist in electronics that have “hidden” or “bottom-terminated” solder joints, including ball grid arrays (BGA), land grid arrays (LGA), quad flat no leads (QFN), and insulated-gate bipolar transistors (IGBT). Regardless of their different shapes, all can be prone to solder voids and other conditions such as non-wet, head-in-pillow, opens, and shorts.
For SMT components like resistors, capacitors, and quad flat packages (QFP), solder joints can be inspected with AOI. However, for the “hidden” components this is not possible because the solder joints are directly between the component and the board and often are only microns thick. When these solder joints overlay other pads on the back side of a board, this occludes the image and does not allow the reader to interpret solder joints clearly.
Due to this market need, we have described a few improvedmethods and analysis tools that work together to provide more accurate void measurement at critical interfaces than previously existed. First, computed laminography (CL) is an X-ray technique thateffectively provides a virtual cross section for single or multiple layers in a component, usually at board-component interface. In comparison to computed tomography (CT), CL allows data to be collected by rotating the object around the same axis the beam is shooting. This enables the X-ray source to get very close to the PCB. Depending on the PCB size, a user can magnify a small area to a much higher resolution than they could using CT to scan the same object without sectioning it from the PCB. This flexibility allows users to zoom in or out to an appropriate resolution for the features that require inspection.
The resulting volume can be analyzed as individual slices, sections, or as the whole volume. Slices can be analyzed quickly and accurately because of the reduced background “noise” in this VoidInspect workflow software, by Yxlon powered by ORS(Object Research Systems), walks the user through the process of defining the measurement area at the actual cross section of the solder joint, not only the CAD estimated position of the pad. Then they can adjust the threshold to define the voids. Once this process is set up on one or more images, it can be saved as a process that can be recalled automatically. Once defined, this process then runs itself and the data can be reviewed for acceptable and failed pins. Or the data can be exported for more detailed review by the process engineering team in their quality system.
For these LGA pins and other solder joints, it has been verified that the combination of the high-quality X-ray image, and theseinspection algorithms result in extremely repeatable and accurate void measurement results. We look forward to future learning and sharing of theseadvanced inspection tools being developed in partnership between Yxlon and ORS, both are companies of the Comet Group. For additional information regarding VoidInspect, follow this link.