AOI · 0201/0402 · Sub-10 µm

0201/0402 SMD Inspection — Automated Microscopic Quality Control with AI

Replace manual microscope checks with a high-resolution imaging system and AI-powered defect detection that reliably catches sub-10 µm issues — at production speed.

Defect Coverage System Overview
LIVE · TELECENTRIC
0402 · R143 PASS
SMD 0201/0402 inspection
Resolution2 µm
Offset3 µm
Bridges0
VerdictPASS

Automate what you currently check under a microscope.

Intelgic replaces manual microscopic inspection of ultra-miniature SMDs (0201: ~0.6×0.3 mm, 0402: ~1.0×0.5 mm) with a high-resolution imaging system and AI-powered defect detection that reliably catches sub-10 µm issues — at production speed.

Trusted on the shop floor of
Gentex RxSight clariant 1800Accountant Aptive Save A Lot TCS Transportation Solid Dareesoft Petromin logo CostaRicaABC logo SunEdison logo Linde logo Cummins logo Sammons Industrial Mobile Maintenance Link X

Micron-level Accuracy

Consistent Pass/Fail

Full Traceability (DQC)

Manual microscope inspection
001 · The Problem

Manual inspection is slow and inconsistent.

Manual microscope inspection is slow, subjective, and hard to scale. Tiny variations in lighting, fatigue, or interpretation lead to misses and false rejects — especially on dense PCBs packed with 0201/0402 parts. Intelgic's automated AOI station delivers repeatable, objective results, integrates with your line, and generates digital records for audits and warranties.

002 · Defect Coverage

What Intelgic automates.

Typical defect categories captured by our AOI system across placement, solder, joint integrity, components, foreign matter, and PCB features.

DEFECT · 01

Placement

Offset/shift, rotation, skew, tombstoning, tilt.

DEFECT · 02

Solder

Insufficient/excess solder, bridging, opens, voids, wetting issues.

DEFECT · 03

Joint Integrity

Micro-cracks, cold joints, fillet shape anomalies.

DEFECT · 04

Component

Missing/polarity wrong, counterfeit/marking mismatch (OCR), package damage.

DEFECT · 05

Foreign Matter

Fibers, dust, flux residues, solder balls.

DEFECT · 06

PCB Features

Pad damage, solder mask delamination, mask encroachment.

003 · System

System overview.

Imaging hardware, Certainty AI, and throughput integration designed for ultra-miniature SMD inspection.

Imaging Hardware

  • Industrial cameras: 20–65 MP with 2–3 µm pixels for micron-scale resolution.
  • Telecentric lenses: distortion-free metrology on tiny leads and fillets.
  • Lighting geometry (software-controlled): coaxial brightfield, low-angle darkfield, polarized ring, multi-wavelength (UV/Vis/IR) to reveal glare-prone solder details.
  • Precision motion: X-Y stage or gantry; optional Z-axis autofocus for uneven boards; fiducial-based alignment.

Certainty AI (Intelgic)

  • Deep learning + anomaly detection: learns "good" joints and flags outliers — ideal when labeled defect data is scarce.
  • Pixel-level segmentation: measures defect area/length; supports thresholds in µm or pixels.
  • Multi-class classification: defect type, severity grading, component presence/polarity.
  • Adaptive recipes: switch models/parameters per product or panel step; A/B changeover in minutes.
  • Analytics & DQC: per-board/per-joint records, trends, heat-maps, Digital Quality Certificates for every PASS unit.

Throughput & Integration

  • Inline or offline: integrates after SPI/placement/reflow, or as a bench-top cell.
  • Comms: MES/ERP/PLC via OPC-UA, REST, SQL; barcode/QR linking.
  • User roles: operator, QA engineer, admin; audit trail & versioned recipes.
004 · Workflow

How it works.

A five-step inspection cycle from load-and-align through traceable DQC issuance.

STEP · 01

Load & Align

Panel enters station; fiducials auto-align the field of view.

STEP · 02

Programmed Imaging

The system executes a scan path; lighting modes switch to expose different defect signatures.

STEP · 03

AI Inspection

Models classify/segment defects, compute metrics, and apply acceptance criteria.

STEP · 04

Decision & Actions

PASS/FAIL per component, with optional mark/reject/stop-line outputs.

STEP · 05

Record & Report

Results stored with timestamps, images, measurements, lot/serial; DQC issued for PASS boards.

005 · Performance

Performance at a glance.

Calibrated, application-tuned specifications for ultra-miniature SMD inspection.

METRIC · 01

Defect Sensitivity

< 10 µm (application-dependent; optics & lighting optimized per board).

METRIC · 02

Placement Accuracy

±5–10 µm typical with telecentric optics.

METRIC · 03

False Rejects

Reduced vs. rules-based AOI through learned normal-appearance modeling.

METRIC · 04

Changeover

New product recipe loading in minutes; model fine-tuning supported.

006 · Engineering

Engineering deep-dive.

Why telecentric + AI matters, the acceptance criteria we use, and how traceability is delivered.

PANEL · 01

Why Telecentric + AI Matters for 0201/0402

  • Telecentric optics remove perspective error → trustworthy micrometer measurements on pads, leads, and fillets.
  • AI anomaly detection doesn't need every defect label → rapid deployment even when true defects are rare.
  • Lighting orchestration reveals solder topography and micro-cracks that single-mode illumination can hide.
PANEL · 02

Acceptance Criteria (Examples)

Defined with your QA team; below are common defaults.

  • 0201/0402 placement offset ≤ 25 % pad width; rotation ≤ 5°.
  • Solder bridge gap ≥ 20 µm; solder ball diameter ≤ 50 µm and count per zone thresholds.
  • Fillet coverage ≥ pad rule; void fraction ≤ x % (per IPC class).
  • Crack length < 10 µm or no-crack rule for safety-critical lines.
PANEL · 03

Data, Traceability & DQC

  • Per-component logs: images before/after decisions, defect masks, metrics in µm.
  • Lot/serial binding: scan board ID; auto-associate to work order and recipe version.
  • Dashboards: FP/FN trend, top defect Pareto, station OEE, heat-maps across the panel.
  • DQC: digitally signed certificate for every PASS board; cloud or on-prem storage.
007 · Comparison

Manual microscope vs Intelgic AOI.

Side-by-side comparison across the metrics that matter for production-grade QA.

Aspect
Manual Microscope
Intelgic AOI
Consistency
Operator-dependent
AI criteria, repeatable
Speed
Slow; non-scalable
Inline, panel-level fast scans
Sensitivity
Limited by fatigue
Sub-10 µm with high-mag optics
Records
Sparse notes/photos
Full image & metric archive + DQC
Changeover
Retrain humans
Load recipe; optional quick re-train
Cost of Quality
High rework/escape risk
Lower false rejects & escapes
008 · FAQ

Frequently asked questions.

Yes. Most customers migrate to automated AOI for primary decisions and keep minimal spot-checks for audit.

No. We bootstrap with good-only anomaly models and incrementally add labeled defect data to refine classes.

We mitigate with coaxial + polarized + darkfield lighting and train models on varied exposures for robustness.

Yes. We design the scan path, FOV tiling, and inference pipeline to your cycle time; parallel cameras are available for high-throughput lines.

Telecentric optics, stage calibration, and per-recipe scale validation (grid targets) ensure µm-accurate metrology.

Solution

Industrial AI

LEGAL

Head Quarter

17352 Murphy Ave Suite 101
Irvine, CA 92614
Phone: (949) 317-2420

Branch Office

5ES7-H, 5Th Floor, East Tower, Mani Casadona, Plot no -IIF/04, Action Area -IIF, Newtown, Kolkata-700160

Certificate

intelgic ISO 27001 certification

Social links

©2026 Intelgic Inc. All Rights Reserved.