FAQ

Iris Ant provides four AI AOI models across two major series:

Selection Recommendation: Choose T7/T7D for 100% inline production line inspection. Select T3 Pro or T5 Pro for sampling or low-to-medium volume production. T5 Pro features higher accuracy, making it suitable for micro-packages and fine-pitch IC inspection.

Inline AOI (T7 / T7D): Integrates directly into SMT assembly lines, operating synchronously with screen printers and pick-and-place machines for 24/7 uninterrupted real-time inspection. Ideal for high-volume, short-takt-time production lines.

Offline AOI (T3 Pro / T5 Pro): Deployed independently without occupying main line footprint. Perfect for small-to-medium batch sizes, multi-product line changes, and new product introduction (NPI) verification.

Selection Criteria: Choose Inline if your priority is throughput maximization and full automation. Choose Offline if you require flexible deployment and frequent product switching.

Yes. The Iris Ant SF-2 series 3D SPI is designed for SMT assembly lines to perform 100% 3D volumetric measurement on printed solder paste, accurately measuring volume, area, height, and offset.

Core Technology: Utilizes Programmable Structured Light Module (PSLM) technology with zero mechanical wear. Achieves a repeatability precision of <10% (6 Sigma), supports 0201 micro-component inspection, features built-in SPC analytical tools, enables 5-minute programming and 3-minute line changeover, and includes lifetime free software updates.

Traditional AOI relies on rigid RGB thresholds and fixed rules, leading to heavy false alarms and escapes when handling identical background colors, marking interference, or complex DIP solder joint classification.

Iris Ant AI AOI leverages multi-dimensional feature extraction and deep learning models to resolve:

• Same-Color Interference: Accurately identifies components even when their color closely matches the substrate.
• Black Inductor & Crystal Markings: Intelligently ignores surface markings to focus strictly on structural defect identification.
• DIP Solder Joint Classification: Uses locally trained AI classification models to adapt to varying process standards.
• Component Alignment: Maintains high alignment positioning precision despite surface contamination or dirt.

Not at all. Iris Ant AI AOI operates on a 100% local deployment model. All model training and real-time inspection take place on your factory's local server or industrial PC. Core process intelligence and proprietary product designs never leave your facility or get uploaded to vendor cloud servers.

Yes. The 80% false alarm rate reduction is validated by real-world SMT production line data through three technical mechanisms:

1. AI Automatic Programming: Generates precise parameters automatically based on a golden reference board, eliminating false alarms caused by improper manual threshold settings.
2. Continuous Learning Mechanism: Operator feedback on inspection results continually refines the AI model, improving accuracy over time.
3. Multi-Dimensional Feature Extraction: Replaces single greyscale thresholds with holistic multi-feature analysis to minimize false judgments.

Operation is straightforward and completed in 4 simple steps within 5 minutes:

1. Capture Reference Board: Take an image of a good board to set up the baseline sample.
2. AI Automatic Modeling: AI generates inspection windows, sets parameters, and detects array layouts automatically.
3. Confirm Results: Review and verify the evaluated inspection criteria.
4. Feedback & Optimization: Run inspection and provide feedback on false alarms; the AI automatically updates parameters.

No specialized AOI programming engineers are required. Operators can run the system proficiently after 30 minutes of basic training.

The system supports flexible programming methodologies:

• AI Smart Programming: Requires no CAD or Gerber files—only a single good reference board.
• Gerber Import: Supports standard RS-274X and RS-274D formats.
• CAD Coordinate Import: Automatically indexes and maps component libraries.
• Manual Teaching: Allows manual creation of inspection windows when Gerber files are absent.

Equipment state directly impacts detection accuracy. On-site data indicates that over 70% of misjudgments stem from hardware condition rather than algorithms.

Recommended Maintenance Schedule:

• Daily: Clean equipment outer casing, optical lenses, and glass stage.
• Monthly: Inspect light source attenuation and calibrate camera parameters.
• Quarterly: Refresh golden board verification programs to validate inspection efficacy.
• Annually: Conduct comprehensive system calibration and program re-verification.

Replacement Thresholds for Key Consumables: LED Light Source: ~12,000 hours | PU Squeegee: ~600k strokes (Steel: ~1.2M strokes) | Glass Stage: ~18 months.

When a board is flagged as NG, operators must manually verify each red-highlighted region on screen:

• False Alarm: Mark as OK and send feedback to the technician to optimize the inspection program.
• Real Defect: Mark defect location, attach a tracking card, and route to the repair station.
• 3 Consecutive Identical Defects: Halt line immediately and notify process engineers to investigate root cause.

Using the industry-standard Golden Board Verification Method:

1. Create NG Samples: Intentionally introduce typical defects (insufficient paste, bridging, missing parts, misalignments) onto a sample board.
2. Daily Verification: Prior to shift startup, IPQC passes the NG sample board to verify program response and logs the results.
3. Program Fine-Tuning: If abnormalities occur, technicians adjust program settings promptly.

Note: Solder paste on SPI NG samples must remain in production-equivalent condition to ensure meaningful test reference data.

1. Confirm PCB Specifications:
   • Inline T7 Series: PCB Size ≤ 500 × 325 mm (Customizable).
   • Offline T3/T5 Pro: PCB Size ≤ 470 × 330 mm.
   • Confirm minimum component size (Standard: 0201; Optional: 01005).
2. Verify Workshop Infrastructure:
   • Power Supply: 220V (Inline: 1.2 kW / Offline: 1.0 kW).
   • Pneumatic Supply: 4 to 6 Bar required for Inline models.
   • Dimensions & Weight: T7 Series (1000×1500×1600 mm / 1000 kg); T3/T5 Pro (860×1040×1318 mm / 345 kg).
3. Define Optional Features in Advance: Barcode scanning/MES connectivity, 3D height measurement, offline programming, and remote operation.
4. Incorporate Service Terms into Contract: Lifetime free software upgrades and free replacement guarantee for 3D projection heads (for first-time buyers).

Traditional AOI program setup takes 3 to 5 hours during product changeover, whereas AI AOI reduces it to 5 minutes, minimizing line downtime drastically.

Quantifiable Financial Benefits:

• Eliminates dedicated AOI programmer positions, saving $15,000–$40,000 per engineer annually.
• Cuts false alarms by 80%, substantially reducing manual review labor costs.
• Reduces costly rework losses caused by escaped defects.

Complementary deployment is highly recommended. SPI and AOI address distinct inspection stages across SMT lines:

Data Correlation Value: Correlating SPI data with AOI defect results validates whether SPI thresholds are set correctly and reveals whether SPI alerts represent true process risks.

Step 1: Check Hardware First Before Adjusting Algorithms. Over 70% of sudden misjudgments originate from hardware condition anomalies.

Troubleshooting Checklist:

• Is the light source attenuated or contaminated?
• Is the camera lens clean and clear?
• Is the conveyor belt vibrating or tracking off-center?
• Are PCB clamps or support pins worn out?
• Are there scratches on the glass stage?

Resolution Steps:

1. Confirm no PCB remains inside the conveyor track.
2. Click the "Reset" button on the UI.
3. Click "Start" to resume testing.

Preventive Action: Regularly clean track rails and inspect conveyor belt tension and alignment.

Troubleshooting Steps:

1. Inspect whether the fiducial mark is obstructed by foreign objects or dirt.
2. If contaminated, eject board and clean surface.
3. If clean, adjust fiducial lighting brightness under the Tools / Fiducial Tool menu.

Common Causes & Solutions:

• Emergency Stop Pressed: Inspect and release the E-Stop button.
• Main Power Switch Off: Switch on the main power breaker.
• Software Conflict: Log out and restart the application, or reboot the controller PC.
• Motion Card Loose Connection: Check card interface cables for bent pins or loose connections.

SPC data generated by SPI/AOI systems serves as a critical asset for SMT process control, enabling manufacturers to:

• Predict stencil lifespan by monitoring aperture wear.
• Track solder paste freshness via volumetric consistency trends.
• Correlate ambient humidity with printing behavior.
• Evaluate operator operational consistency across shifts.

Core Value: Transforms quality control from reactive post-inspection to proactive drift prevention before physical defects occur.

• MES Integration: Real-time inspection data transfer for end-to-end traceability.
• SPC Reporting: Automated generation of Cp/Cpk quality trend reports.
• Remote Operation: Status checking and report retrieval via web browsers.
• Offline Programming: Program setup on standalone workstations without interrupting active production.
• Open API: Flexible integration capabilities with mainstream MES and ERP systems.