First inspection workflow for the semiconductor industry.

X-ray data analysis to reduce waste and drive 80% gain in inspection speed

Launched in 2019, the VoidInspect workflow project between ORS and IXS had a tangible customer problem to solve from the start: Enabling the rapid non-destructive analysis of voids (enclosed cavities) inside the solder joints of board components, especially for ball grid arrays, to ensure the proper thermal stability.

The choice of laminography, as opposed to conventional computed tomography (CT), seemed the obvious solution to this challenge. This is because CT requires a small sample size to create a high-quality 3D model. In surface mount technology, this means destroying the printed circuit board, as the area of interest must be separated out. But computed laminography, which cuts a 3D volume into virtual slices for analysis, is non-destructive and can be performed over the entire inspection area while achieving high resolution. This technique, combined with new intelligent algorithms, reduces waste and inspection time by automatically localizing the slice with the best resolution to highlight and evaluate the problematic defects. As Peter Koch reports, “We can gain up to 80% in speed compared to conventional methods”.

In order to complete the evaluation even faster, its execution can be performed even by an untrained operator – with a minimum number of steps and no physical interaction. A goal that was clear from the beginning was that VoidInspect’s user interface needed to be user-friendly.

The rapid, high-quality results are achieved through a combination of several ingredients: the Yxlon image processing (the best on the market), the workflow itself, the high-end hardware, and the FF CT software ensuring quality in the reconstruction of the laminography data. It all adds up to a perfect recipe for precision. Even now, the VoidInspect workflow can be added to any FF CT software from Yxlon for off-line applications.

It is not just about collaborating, but co-creating

Completed in only nine months, the workflow based on the AI capabilities of Dragonfly, ORS’s flagship software, is an example of effective project management across teams. Mathieu Gendron, Lead Architect at ORS, recalls that the project’s success was closely related to the feedback rounds initiated with IXS at a very early stage, allowing for the quick creation of mockups that could be critiqued and precisely adjusted before programming. For the two teams, accustomed to agile project management, it was not only about collaboration, but co-creation. “We saved a lot of time by communicating closely, and thus knowing exactly what to do. We were able to split up the tasks efficiently among the developers,” Mathieu Gendron explains.

A game changer for semiconductor customers?

Until now, void measurement in complex boards was very time-consuming and only possible to a limited extent. It is therefore no surprise that the first semi-automatic high quality void analysis done via an algorithm in laminography may well be a game-changer. For the semiconductor industry, which must ensure flawless quality for complex advanced chip and wafer-level packaging, the solution created by the ORS and IXS teams is a very welcome one. Peter Koch confirms the interest aroused: “Customers started to send us their samples to try the workflow – and they are impressed. We are seeing the sales numbers increasing.”

VoidInspect is just the beginning

Motivated by the outcome of this cross-team collaboration, the teams are already planning how they can advance x-ray data processing further to solve customers’ pain points. One thing is certain: By releasing this first workflow, they have unleashed enormous potential for the future of inspection.