Process Optimization

1. Why AOI Alone Is Not Enough

AOI detects assembly results, but it does not understand:

• Printing process stability

• Normal solder volume variation

• Acceptable paste behavior

Without upstream context, AOI often flags normal variation as defects.

SPI fills this gap by providing quantitative, process-level insight.

2. What SPI Data Is Useful for AOI Adjustment

Not all SPI data is equally valuable. Focus on:

• Solder volume distribution

• Height and area trend

• Cp / Cpk values

• Pad-level consistency

• Trend drift over time

These metrics indicate whether a solder joint appearance is normal or risky.

3. Establish a Stable SPI Baseline

Before linking SPI to AOI:

1. Run SPI under normal production conditions

2. Collect data across multiple boards and shifts

3. Confirm process stability (Cp/Cpk ≥ target)

4. Eliminate known printing issues

AOI adjustment based on unstable SPI data will amplify problems.

4. Classify AOI False Calls Using SPI Context

For each AOI false call, ask:

• Did SPI show abnormal paste behavior?

• Is this variation within SPI control limits?

• Is the defect cosmetic or functional?

This classification separates real risks from acceptable variation.

5. Map SPI Metrics to AOI Parameters

Create logical relationships such as:

AOI rules should respond to SPI reality, not static assumptions.

6. Implement Tiered AOI Thresholds

Instead of one fixed limit:

• Tier 1: Normal SPI → Relaxed AOI

• Tier 2: Borderline SPI → Standard AOI

• Tier 3: SPI drift → Tight AOI + alert

This dynamic approach dramatically reduces false calls.

7. Adjust AOI by Component and Pad Type

SPI data varies by:

• Component size

• Pad geometry

• Solder paste aperture

Use SPI segmentation to:

• Tune AOI rules per component family

• Avoid over-penalizing small passives

• Focus inspection on high-risk joints

8. Integrate SPI Trends into AOI Review Logic

When SPI trend alarms appear:

• Prioritize AOI review

• Trigger engineering review

• Increase inspection sensitivity temporarily

When SPI is stable:

• Minimize manual AOI review

• Maintain smooth line flow

9. Manage Change Control Carefully

SPI-driven AOI changes should:

• Be documented

• Require engineering approval

• Be reversible

Uncontrolled parameter drift creates long-term risk.

10. KPIs to Track After Implementation

Measure success using:

• AOI false call rate

• AOI review time per board

• Line blocking incidents

• First-pass yield (FPY)

• Inspection takt time

Data should confirm improvement—not just perception.

11. Common Implementation Mistakes

Avoid:

• Blindly loosening AOI rules

• Ignoring SPI data quality

• Over-reacting to short-term spikes

• Adjusting AOI without process ownership

SPI-driven adjustment is engineering discipline, not trial and error.

Conclusion

SPI data-driven AOI adjustment transforms inspection from a reactive filter into a process-aware quality system.

When SPI and AOI work together:

• False calls decrease

• Inspection flow stabilizes

• Quality risk becomes visible early

This approach represents mature SMT process control.

Linking Suggested Reading

• SPI vs AOI Optimization – Joint Strategy

• AOI False Call Optimization

• Inspection and Buffering Strategy

• SMT Line Blocking and Bottlenecks

• SMT Knowledge Center
  

Thomao Engineering Insight

AOI sees the result.
SPI understands the process.
Data connects both.