Limtronik advances AI-driven quality control into an intelligent control loop

Detecting deviations before they propagate through production – and controlling processes so that defects are prevented from occurring in the first place: the capabilities of automatic optical inspection (AOI) are advancing significantly with AI. Limtronik laid the technological groundwork early on. Since 2011, the electronics manufacturer has relied on 3D SPI systems from Koh Young, and since 2013 it has used the data gathered from them for statistical process analysis. With the later integration of 3D AOI systems, a foundation of real 3D measurement and process data built up over years has emerged – one that today enables new AI applications and closed control loops.

"We invested early in 3D measurement technology and the systematic evaluation of the data it generates. At the time, the focus was mainly on quality assurance and process optimization. Today it's clear that, in doing so, we also laid an important foundation for the use of AI. The more reliable and meaningful the data, the more tasks intelligent systems can take on – all the way to the automated optimization of manufacturing processes," explains Gerd Ohl, Managing Director of Limtronik GmbH.

Measuring instead of interpreting

Koh Young takes a clear technological approach with its inspection systems: measure rather than interpret. The 3D inspection systems capture precise measurements of the height, volume, area, and shape of solder joints and components. This produces objective, reproducible data that can be statistically evaluated and used to optimize inspection and manufacturing processes.

SmartRep has been a sales and service partner of Koh Young since 2009 and found early receptiveness at Limtronik for the concept of 3D measurement data. Limtronik invested in 3D SPI systems back in 2011. Since 2013, Limtronik has relied on SPC analyses to statistically evaluate and systematically optimize manufacturing processes. In the 2020s, 3D AOI systems were integrated into Limtronik's smart manufacturing operations. In switching from 2D to 3D AOI technology, Limtronik was able to reduce false-call rates by 70 to 80 percent.

With the use of AI, this data foundation is gaining even greater significance. AI applications can draw on concrete geometric measurements and statistical process data, using them not only for defect detection but also for automated decision-making.

AI takes on tasks at the review station

One example of this is Koh Young's Smart Review. Smart Review automates the previously manual evaluation of AOI defect reports. The AI uses the 3D measurement data to filter out false calls and can feed its findings back into the inspection process. This reduces the workload at the review station while making inspection continuously more stable.

Koh Young relies on incremental learning for this. The rationale is that an incorrect decision by an operator should not enter the AI model as an incorrect training label. New assignments are therefore not adopted without review. The reliability of a decision is taken into account before a new association is incorporated into the autonomous evaluation model.

From detecting to preventing defects

The Koh Young Process Optimizer (KPO) goes a step further: the 3D measurement data generated by SPI and AOI is used to actively optimize the manufacturing process. KPO analyzes correlations between individual process steps and can derive targeted corrections for printers and pick-and-place machines.

For example, if the SPI detects systematic deviations in the volume, height, or position of solder deposits, KPO can analyze the measurements and feed corresponding corrections back into the printing process. AOI measurement data, in turn, can be used to detect positional deviations of components and draw conclusions about the placement process.

This creates a closed control loop: measure, analyze, correct, and measure again.

"Our goal is to prevent defects from occurring in the first place – and thereby sustainably increase process stability and first pass yield," explains Gerd Ohl.

Limtronik at EFX 2026

Limtronik will also demonstrate how digitalization, automation, and AI interact in electronics manufacturing at the EFX trade fair, taking place October 6–8, 2026, in Stuttgart. As a member of SEF Smart Electronic Factory e. V., the company will participate in Hall 9, Booth 9C60, in the live use case "From Order to Delivery." Together with other members, the association will showcase an end-to-end production process from order placement through to delivery.