Chapter 11: Inspection Standards and Techniques
After components are placed and soldered, the PCBA must undergo rigorous inspection to
confirm that all operations have been performed correctly and that the final product meets the
required quality standards (e.g., IPC-A-610). Manual inspection is giving way to advanced
automated techniques that provide greater speed, consistency, and capability for complex,
miniature assemblies.
11.1 Manual Visual Inspection (MVI)
While automated inspection dominates, Manual Visual Inspection (MVI) still plays a role,
especially for low-volume production, complex rework verification, or final quality checks.
●Tools: Uses stereo microscopes, magnifiers, and dedicated lighting.
●Technique: Highly skilled inspectors search for defects based on IPC-A-610 criteria
(10.1.1).
●Limitations:
○Subjectivity: Highly dependent on the inspector's skill and fatigue.
○Speed: Too slow for high-volume SMT production.
○Hidden Defects: Cannot see under components (e.g., BGAs, QFNs).
11.2 Automated Optical Inspection (AOI)
Automated Optical Inspection (AOI) is the primary inspection tool on most SMT lines,
providing fast, automated visual checks for a wide range of defects.
11.2.1 2D vs. 3D AOI
●2D AOI: Uses multiple high-resolution cameras with different lighting angles (e.g., red,
green, blue, white LEDs) to capture 2D images of the board. It detects defects based on
deviations from the expected image (e.g., missing part, wrong polarity, shorted leads).
It's very fast and effective for many defects. ●3D AOI: The current standard for high-reliability SMT. In addition to 2D cameras, it uses
structured light (e.g., laser projectors or digital fringe projection) to create a
topographical, 3D image of the PCB. This allows it to:
○Measure Solder Fillets: Critically, it can measure the height and shape of solder
joints, identifying insufficient or excessive solder, which 2D AOI struggles with.
○Improved False Call Reduction: By understanding the 3D shape, it can
differentiate between acceptable variations and true defects more reliably,
reducing false calls (the machine flagging a good joint as bad).