Skip to main content

🇻🇳 178 profiles · Engineering review

Vietnam Manufacturing

How to Verify a Stamping Parts Supplier's AOI and X-Ray Inspection Capabilities?

July 17, 2026

High-tonnage mechanical and hydraulic stamping press at AN THANH TECHNOLOGY COMPANY LIMITED, a DEWIN-audited manufacturing facility

Verifying a stamping parts supplier's AOI and X-Ray inspection capabilities

Last year, a client’s stamping supplier claimed “100% inspection” on every shipment — until a batch of mounting brackets arrived with hidden micro-cracks that shut down an assembly line for three days.

To verify a stamping parts supplier’s AOI and X-Ray inspection capabilities, request equipment calibration records, Gage R&R studies, and defect detection logs. Then conduct an on-site audit to confirm that inspection programs are actively maintained, properly integrated into the production flow, and supported by a mature Quality Management System.

This guide walks through the specific documents, questions, and audit tactics that separate a supplier who truly operates advanced inspection from one who simply owns the machines. Let’s break it down section by section.

What Documentation Should I Request to Confirm AOI and X-Ray Equipment Ownership?

One lesson we learned early in our supplier audit work across Vietnam and China is that a brochure photo of a machine tells you nothing about whether it actually runs during your production window.

Request equipment asset lists with serial numbers, original purchase invoices or lease agreements, annual calibration certificates traceable to national standards, and maintenance logs showing service frequency. Also ask for the equipment's OEM specification sheet to confirm resolution, inspection speed, and defect-detection thresholds.

Equipment documentation checklist confirming AOI and X-Ray ownership and calibration records

Why an Asset List Alone Is Not Enough

Many stamping suppliers will hand you a generic equipment list during the RFQ phase. It might say "AOI System — 1 unit" or "X-Ray — available." That tells you almost nothing. You need to dig deeper.

Start by requesting the documents below. Each one serves a different verification purpose.

Document What It Proves Red Flag If Missing
Equipment purchase invoice or lease contract Supplier actually owns or leases the unit May be borrowing or sharing equipment
OEM specification sheet Camera resolution, field of view, throughput speed Cannot confirm the machine suits your part geometry
Annual calibration certificate Measurement accuracy is traceable and current Results from the machine cannot be trusted
Maintenance and service log Equipment is kept in working condition Risk of drift, downtime, or unreliable readings
Gage R&R study 1 report Repeatability and reproducibility across operators/shifts Inspection data may vary widely between runs

Calibration Traceability Matters

A calibration certificate should reference a recognized standard body 2 — for example, A2LA or a national metrology institute. When our team audits facilities in Southeast Asia, we check that the calibration date 3 is within the last 12 months and that the certificate covers the specific measurement parameters relevant to your part. For dimensional metrology on stamped brackets, the AOI system's pixel resolution and lighting configuration must match the tolerances on your drawing.

Gage R&R: The Hidden Indicator of Process Maturity

A Gage R&R study reveals whether the inspection system produces consistent results regardless of which operator runs it or which shift it operates on. If a supplier cannot produce a Gage R&R report for their AOI or X-Ray system, their inspection data is unreliable. In our experience coordinating PPAP packages for U.S. buyers, we treat a missing Gage R&R as a deal-breaker for any part with critical tolerances.

Lease vs. Ownership

Some suppliers lease equipment or share it across multiple facilities. This is not inherently bad, but you need to know. A shared X-Ray unit might only be available one day per week. That means your parts could sit in queue, delaying shipments — a pain point we see frequently.

✔ Calibration certificates must be traceable to a recognized national or international standard True
Without traceable calibration, measurement results have no verified accuracy, making inspection data unreliable for quality decisions.
✘ Owning an AOI machine automatically means a supplier has a validated inspection process False
Equipment ownership says nothing about recipe programming, operator training, calibration status, or integration into the production flow. A machine sitting idle provides zero quality assurance.
Video: Sample metal parts from audited Vietnamese factories in the DEWIN network. (@dewintech)

How Do I Evaluate a Stamping Supplier's AOI Programming and Defect Detection Accuracy?

A few months ago, we reviewed a supplier's AOI setup for a set of Z-shaped spring clamps similar to the ones we regularly source. Their machine was top-tier, but the detection recipe had not been updated since the original tool was revised — meaning it was inspecting against an outdated reference.

Evaluate AOI programming by requesting the defect library, reviewing recipe revision control logs, and asking for documented false-call rates. A mature supplier maintains a "Golden Part" reference, tunes recipes after every engineering change, and tracks defect Pareto data weekly to improve surface defect detection accuracy.

Evaluating AOI programming and defect detection accuracy using golden part references

Understanding AOI in Metal Stamping Context

AOI for stamped parts is different from AOI in PCB assembly. In precision metal stamping, AOI cameras and structured lighting 4 detect scratches, burrs, missing punch-outs, dimensional deviations, and edge quality issues. The system compares each part image against a digital reference — the "Golden Part" — and flags anything outside the programmed tolerance window.

Here is a quick comparison so you understand the distinction:

Feature AOI for PCB Assembly AOI for Metal Stamping
Primary defects detected Solder joints, missing components Burrs, scratches, dimensional deviation, edge cracks
Reference standard Gerber file overlay Golden Part image or CAD-to-part comparison
Lighting technique Diffuse, multi-angle Structured, directional to reveal surface texture
Typical placement Post-reflow Inline after stamping press or at final QA station
Key accuracy metric Component presence/absence Dimensional accuracy within ±0.05 mm or tighter

Recipe Revision Control

Every time a tool is modified, a material lot changes, or an engineering revision updates the drawing, the AOI recipe must be updated. Ask the supplier: "How do you manage recipe revisions?" A strong supplier will have a documented change-control process 5 tied to their Quality Management System. A weak supplier will say "we update it when needed" — with no log to prove it.

False-Call Rate and Escape Rate

Two numbers tell you everything about AOI effectiveness:

  • False-call rate: How often the system flags a good part as defective. A well-tuned system keeps this below 10%. Above 15%, operators start ignoring alerts — which defeats the purpose of automation.
  • Escape rate: How often a defective part passes through undetected. This is harder to measure but can be inferred from downstream rejection data.

When we manage final inspection for our clients, we cross-reference the supplier's AOI data against our own outgoing quality checks. Discrepancies reveal recipe gaps immediately.

The Role of AI and Deep Learning

Some advanced AOI systems now use AI-driven deep learning algorithms to distinguish between harmless surface marks — like residual oil patterns — and true functional defects like micro-cracks. If your supplier claims AI-enhanced inspection, ask them to demonstrate the training dataset and show you how the model was validated. A vague claim of "AI inspection" without evidence is a marketing statement, not a quality capability.

In-Die Vision Sensors

For high-volume stamping, some suppliers deploy in-die vision sensors that inspect parts during the press stroke itself. These sensors detect tool wear, slug pulling, and misfeed conditions in real time. This is a strong indicator of process maturity. Ask whether their system triggers an automatic press stop when a defect threshold is reached — this is the "closed-loop" quality system concept that separates top-tier suppliers from average ones.

✔ A Golden Part reference must be re-validated after any tooling or material change True
Tooling wear and material lot variations alter part geometry and surface characteristics, making the previous Golden Part reference inaccurate for defect comparison.
✘ A low false-call rate means the AOI system catches all defects False
False-call rate measures over-detection (flagging good parts). A system can have a low false-call rate while still missing real defects if the recipe’s sensitivity thresholds are set too loosely.

What Questions Should I Ask During a Factory Audit to Verify X-Ray Inspection Processes?

During a recent supplier audit in southern China, I watched a factory tour guide walk us past an X-Ray machine that was clearly powered off and covered in dust. When I asked the last time it ran a production lot, the answer was "maybe two months ago." That single observation changed our entire assessment.

During a factory audit, ask where X-Ray sits in the process flow, which part families require it, what the escalation criteria from AOI to X-Ray are, how often the tube and detector are calibrated, and whether operators hold NDT Level II certification. Request sample inspection reports with actual defect images, not just pass/fail summaries.

Factory audit questions to verify X-Ray inspection process and operator certification

When X-Ray Is Actually Necessary

Non-destructive testing via X-Ray is not standard for every stamped part. It is critical when:

  • Parts have internal features invisible to surface inspection (e.g., welded sub-assemblies)
  • The application is safety-critical — automotive braking components, aerospace structural brackets
  • Parts are thick-gauge and could harbor internal porosity or inclusions
  • Stamped-and-overmolded assemblies need internal geometry verification

For general-purpose mounting brackets and clips, X-Ray may not be required. But if your parts go into an automotive or aerospace assembly, your buyer will likely require it — and so will IATF 16949 or AS9100 certification standards.

The 10 Audit Questions

Here is a structured list of questions to ask, along with what each answer reveals:

Question What a Good Answer Looks Like Red Flag Response
Where is X-Ray positioned in your process flow? "After AOI flags a suspect part, or 100% for safety-critical families" "At the end, if the customer asks for it"
Which part families go through X-Ray? Specific list tied to risk classification "Everything" or "whatever the customer wants"
What triggers X-Ray from AOI? Defined escalation rules in the control plan "Operator judgment"
How often is the X-Ray tube calibrated? Per OEM schedule, with records available "When something looks off"
Do your operators hold NDT Level II or equivalent? Yes, with certifiable training records Hesitation or no certification
Can you show a recent X-Ray report with defect images? Detailed report with annotated images Pass/fail summary only
What is your reject disposition process? Quarantine, root-cause analysis, corrective action "We set them aside"
Do you use 2D X-Ray or 3D Computed Tomography 6? Depends on application; CT for complex assemblies Claims CT but cannot demonstrate
How is X-Ray data stored and traced to lot numbers? Digital archive linked to lot/batch traceability system Paper records or no traceability
Is X-Ray data included in your PPAP submission? Yes, as part of the measurement results section "Only if requested"

3D Computed Tomography vs. Standard X-Ray

Standard 2D X-Ray gives you a flat projection — good for detecting voids and large inclusions. 3D Computed Tomography (CT) creates a full volumetric model of the part, allowing engineers to slice through it digitally and inspect internal geometry at high resolution. CT is more expensive and slower, but it is the gold standard for complex stamped-and-overmolded assemblies or parts with tight internal tolerances.

Statistical Process Control Integration

A mature supplier feeds X-Ray data into their Statistical Process Control 7 system. This means they track defect trends over time, not just pass/fail per part. Ask to see their SPC charts for X-Ray-inspected part families. If they cannot produce them, their X-Ray process is reactive, not preventive.

How Can I Confirm My Supplier's Inspection Reports Match Actual Production Quality?

A trade-off we constantly weigh in our supply chain management work is trust versus verification. We want to trust our suppliers' reports, but experience has taught us that the only reliable confirmation comes from independent cross-checks.

Confirm report accuracy by conducting First Article Inspection on initial samples, running periodic third-party dimensional audits, comparing supplier AQL data against your incoming inspection results, and starting with a trial order before committing to volume. Any persistent gap between supplier reports and your findings signals a systemic quality problem.

Confirming inspection reports match actual production quality through first article inspection

The Trial Order Strategy

Before committing to a large production contract, place a small trial order — typically 200 to 500 pieces. Inspect these parts thoroughly at your facility or through a third-party lab. Compare your results against the supplier's inspection report for the same lot. Look for:

  • Dimensional measurements that fall outside the supplier's reported range
  • Surface defects that were not flagged in their AOI report
  • Inconsistencies in material test certificates versus actual material hardness or thickness

This approach costs very little relative to the risk it mitigates. When we coordinate trial orders for our U.S. clients, we perform final inspection at the supplier's facility and then arrange for the buyer's incoming inspection team to verify independently. Discrepancies are documented and discussed before any volume order is placed.

AQL Alignment

Make sure you and your supplier agree on Acceptable Quality Levels 8 before production starts. Here is a standard framework:

Characteristic Type Recommended AQL Inspection Level
Safety-critical dimensions AQL 0.65 or 100% inspection Tightened (Level III)
Major functional dimensions AQL 1.0 Normal (Level II)
Minor cosmetic attributes AQL 2.5 Normal (Level II)
New supplier (first 5–10 lots) AQL 0.65 across all categories Tightened until consistency proven

If your supplier's reports consistently show zero defects at AQL 2.5 but your incoming inspection finds issues at AQL 1.0, something is wrong with their sampling plan, their measurement method, or both.

First Article Inspection as the Baseline

First Article Inspection validates the initial setup — tool alignment, material conformance, dimensional accuracy. It is the foundation upon which all subsequent inspection builds. But FAI alone is not enough. It catches systematic errors (wrong tool, wrong material). It does not catch random isolated defects that occur during a production run. That is what AOI is for.

The best approach is a layered system: FAI validates the setup, inline AOI catches random defects in real time, and X-Ray provides non-destructive testing for high-risk parts. This is the tiered inspection strategy that IATF 16949 and AS9100 effectively require for certified suppliers.

Cross-Referencing Data Over Time

Do not just check one shipment. Track supplier quality data over five to ten consecutive lots. Plot the defect rate trend. A stable, low defect rate indicates a mature process. A fluctuating rate — even if individual lots pass — suggests the process is not under Statistical Process Control and will eventually produce a bad shipment.

Our team maintains a supplier scorecard for every active source. We track on-time delivery, first-pass yield 9, and defect PPM. When a supplier's AOI reports show consistently perfect results but our incoming data shows occasional failures, we escalate immediately. That gap is where quality escapes live.

✔ A trial order is the most cost-effective way to verify a new supplier’s actual inspection capabilities True
A small trial batch allows independent measurement comparison against supplier reports before committing to volume, exposing gaps in their quality system with minimal financial risk.
✘ A supplier reporting zero defects consistently means they have excellent quality False
Consistently zero-defect reports often indicate under-inspection, overly loose acceptance criteria, or data manipulation rather than genuinely perfect production. Some level of detected-and-contained defects is normal and healthy.

Conclusion

Verifying a stamping supplier's AOI and X-Ray capabilities is about confirming process maturity — not just checking for equipment on the floor. Demand documentation, ask precise audit questions, and validate reports against independent inspection data before scaling production.

Frequently Asked Questions

What Documentation Should I Request to Confirm AOI and X-Ray Equipment Ownership?

One lesson we learned early in our supplier audit work across Vietnam and China is that a brochure photo of a machine tells you nothing about whether it actually runs during your production window.

How Do I Evaluate a Stamping Supplier's AOI Programming and Defect Detection Accuracy?

A few months ago, we reviewed a supplier's AOI setup for a set of Z-shaped spring clamps similar to the ones we regularly source. Their machine was top-tier, but the detection recipe had not been updated since the original tool was revised — meaning it was inspecting against an outdated reference.

What Questions Should I Ask During a Factory Audit to Verify X-Ray Inspection Processes?

During a recent supplier audit in southern China, I watched a factory tour guide walk us past an X-Ray machine that was clearly powered off and covered in dust. When I asked the last time it ran a production lot, the answer was "maybe two months ago." That single observation changed our entire assessment.

Footnotes


1. Statistical method for analyzing measurement system variation in quality control. ↩︎


2. NIST guidelines on metrological traceability for calibration certificates. ↩︎


3. Authoritative accreditation body for calibration and testing laboratories. ↩︎


4. Technical overview of structured light used in automated optical inspection systems. ↩︎


5. ISO 9001 requirements for quality management and documented change control. ↩︎


6. Authoritative NIH guide explaining how CT creates 3D volumetric images. ↩︎


7. NIST Engineering Statistics Handbook section on statistical process control methods. ↩︎


8. Definition and application of acceptable quality limits in manufacturing sampling. ↩︎


9. Standard definition of the FPY quality metric. ↩︎

Manufacturing RFQ

Have a STEP Model or 2D Drawing for This Project?

Share your drawing, material, quantity and finishing requirements to begin a manufacturing quote. Confirm pricing, delivery and inspection requirements during the review.

Upload CAD & Get Quote →

This article is part of DEWIN's technical manufacturing engineering knowledge base. Factory-specific details come from each factory's profile page — please verify current information there or with our team.

Quick Sourcing Inquiry