Greatly Improving the Speed and Accuracy of Conformal Coating Inspection for PCBs

Manufacturing

DarwinAI

Conformal coatings protect electronic components—typically printed circuit boards (PCBs). Inspecting such coatings is notoriously difficult, and usually requires manual examination to detect defects that may result in rework or scrap. A leading global manufacturer of aerospace components approached DarwinAI, looking for an alternative to an inefficient and expensive manual inspection process.

The Problem

Due to the difficulty of inspecting conformal coatings and the mix of PCB models involved, the existing process was slow, with high escape rates and false positives.

DarwinAI developed and deployed a vision-based AI that assists human inspectors by automatically detecting and distinguishing between a range of defects.

The Results

fewer hours inspecting
60%
less false positives/escape
3x
solution payback
1 month

Difficulties with the existing inspection process

  • Complexity
  • Slow, Manual Inspections
  • High False Positive and Escape Rates

Automating a complex task

Conformal coatings can experience a range of defects, including: cracking; delamination; bubbles, pinholes, and foam; and uneven or incorrect thickness.

Detecting and differentiating between these faults has historically required human expertise. Even today, with advanced AI solutions available, automation is extremely rare due to the complexity and lack of training data.

DarwinAI’s patented technology achieves higher accuracy using less data. The AI models at the heart of DarwinAI’s automated solutions deliver rapid decisions that speed up inspection processes.

DarwinAI’s Solution

In only one month, DarwinAI developed an automated, AI-driven vision inspection solution that:

  • Uses multiple views of the board under inspection—provided by two mounted cameras and a rotary
  • Rapidly identifies and distinguishes between defects (e.g., too much coating, too little coating, etc.)
  • Shows operators the reason for an inspection outcome and allows them to validate the decision
  • Provides detailed logging that enables rigorous auditing and increases organizational confidence 

Importantly, the flexibility of DarwinAI’s technology ensured straightforward deployment and operation by requiring few defect samples to get started—a huge benefit in high-mix/low-volume production and leveraging human-in-the-loop operation to quickly achieve production-level performance accuracy.

Difference-making results

Ultimately, DarwinAI’s automated solution:

  • Slashed average inspection time to only 20 seconds
  • Cut false positive and escape rates to less than 5%
  • Reduced labor costs by 60%
  • Delivered a one-month return on investment

Contact us to learn more about this case study.