Knocking Down the Bone Pile: Hot Air vs. Infrared PCB Rework
Nolan’s Notes: Why Mexico’s Time Has Come
Standard of Excellence: ‘Quality’ Is Your Company's Greatest Competitive Advantage
Applying test strategies simulation to the test and inspection workflow should be top of mind in the digital age, yet we notice that many engineers are still not applying this critical piece to get the test and inspection workflow optimized.
In the past, test engineers tended to focus on one piece of the test and inspection workflow: AOI, boundary scan, ICT, flying probe, functional test, and the like. In their focus on one strategy, they generally ignored the other test and inspection tests, and had little idea of what was and was not covered by these other strategies. Essentially, applying the maximum coverage from their area of focus or assuming what pieces may be covered from other strategies is like taking a shot in the dark because you don’t really know the facts.
The key to designing a good test and inspection process is to model that process before starting to produce test programs and then see the coverage results reflected. The concept is to build a digital twin of the test and inspection process to get the visibility that’s needed to make important decisions.
Digital twin relates to the test and inspection workflow by using data analytics or AI-driven simulation technology to create predictive models which have been tuned and can continuously be tuned based on real test results. Simulation models allow the user to determine the effectiveness of the test and inspection steps. Where are coverage duplications that are not needed? What components still have missing coverage or low coverage from the process that need supplementation or more creative solutions to achieve the desired test coverage?
A good predictive model of a test strategy will allow a comprehensive report to be generated about the coverage as it applies to the board being examined. The simulation model for each test strategy needs to know the strategy’s capabilities. For example, if we’re looking at AOI, what type of machine will be used—2D or 3D? Will it have good camera resolution, or laser capability for planarity and tombstoned components? With what effectiveness can solder joints be examined? The model can then be tuned to reflect the machine to be used and predict the test coverage offered by the machine on the board. The same idea can be applied with other test strategies, such as X-ray, ICT, flying probe, etc. Depending on the machine to be used, the model needs to understand the capabilities, accessibility of the board, and the way some programming features will be used to, again, produce a report of comprehensive test coverage.
To read this entire article, which appeared in the October 2023 issue of SMT007 Magazine, click here.
