IDB-QCI-004
Production · QC · inspection
Quality control and inspection
Reference for AQL sampling, defect classification, inspection-point design, process capability, and the golden-sample workflow that catches defects before they ship.
Abstract
Quality control is the operational expression of the specification. A spec sheet without a matching QC plan creates ambiguity: each side judges quality against an unstated standard. The QC plan converts the spec into measurable inspection criteria, sampling sizes, disposition rules, and the records that survive an audit or a recall.
Section 1 covers the four production-line inspection points. Section 2 covers AQL sampling math (ISO 2859-1 / ANSI Z1.4). Section 3 covers defect classification. Section 4 covers process capability (Cp, Cpk). Section 5 catalogues process-specific defects. Section 6 provides the QC plan template structure.
1.Inspection points
Four checkpoints, each with a defined scope, sample basis, and disposition authority.
| Point | When | Sample basis | Decision | Records |
|---|---|---|---|---|
| IQC (incoming) | Materials arrive | Per shipment, per AQL | Accept / reject / quarantine | C of C, lot, inspection report |
| IPQC (in-process) | During assembly | First article + per AQL | Continue / rework / stop | First-article report, control charts |
| FQC (final) | Finished units | Per AQL on packed boxes | Pass / hold / scrap | Per-lot disposition log |
| PSI (pre-shipment) | Before goods ship | 3rd-party per AQL | Pass / hold / fail batch | PSI report + photographs |
1.1Golden sample workflow
- Physical sample signed and dated by both buyer and supplierthe acceptable production unit.
- Production QC measures against the golden sample, not against the prototype, render, or buyer's expectations.
- Both parties keep an identical signed sampleyours in the office, theirs on the production floor.
- Photograph and documentHigh-res images, dimensional report, tag attached to the physical sample.
- Disputes resolve by reference to the golden sampleNot by argument; by inspection of the signed reference.
2.AQL sampling
ISO 2859-1 (international) and ANSI/ASQ Z1.4 (US-equivalent) define the sampling math. Same tables; different cover. Level II / Normal inspection is the default.
2.1Standard AQL levels by defect class
- **Critical defects0.0 % AQL.** Any critical defect rejects the batch.
- **Major defects2.5 % AQL.** Significantly reduces usability.
- **Minor defects4.0 % AQL.** Cosmetic, does not affect function.
Tighter AQLs (1.0 % major, 2.5 % minor) for medical, automotive, mil-spec. Looser AQLs (4.0 % major, 6.5 % minor) for promotional or one-shot goods.
2.2Sample size table (Level II, Normal inspection)
| Lot size | Code | Sample | Acc/Rej @ 2.5 % | Acc/Rej @ 4.0 % | Acc/Rej @ 1.0 % |
|---|---|---|---|---|---|
| 91–150 | F | 20 | 1 / 2 | 2 / 3 | 0 / 1 |
| 151–280 | G | 32 | 2 / 3 | 3 / 4 | 1 / 2 |
| 281–500 | H | 50 | 3 / 4 | 5 / 6 | 1 / 2 |
| 501–1200 | J | 80 | 5 / 6 | 7 / 8 | 2 / 3 |
| 1201–3200 | K | 125 | 7 / 8 | 10 / 11 | 3 / 4 |
| 3201–10000 | L | 200 | 10 / 11 | 14 / 15 | 5 / 6 |
| 10001–35000 | M | 315 | 14 / 15 | 21 / 22 | 7 / 8 |
| 35001–150000 | N | 500 | 21 / 22 | 30 / 31 | 10 / 11 |
2.3Switching rules
Quality drifts over time; the inspection level adapts.
| Direction | Trigger | New level |
|---|---|---|
| Normal → Tightened | 2 of 5 lots rejected | Tightened (tighter sample) |
| Tightened → Suspend | 5 lots rejected | Stop production |
| Tightened → Normal | 5 consecutive lots pass | Normal |
| Normal → Reduced | 10 consecutive lots pass + low defect rate | Reduced sample |
| Reduced → Normal | 1 lot fails | Normal |
3.Defect classification
3.1Critical defects (0.0 % AQL)
- Safety hazardsSharp edges (per IEC 62368-1 test fingers), exposed conductors, battery shorts, choking risk on toys (<31 mm part per ASTM F963).
- Failure to perform primary functionDoes not power on, sensor doesn't read, transmitter doesn't transmit, app cannot connect.
- Regulatory non-complianceMissing required marks (CE, FCC ID, country of origin), wrong importer details, RoHS-restricted substances above threshold.
- Counterfeit or substituted componentsAny deviation from the approved BoM, even when functional.
3.2Major defects (2.5 % AQL)
- Reduced performanceBattery life ≤80 % of spec, sensor accuracy outside tolerance, charging slower than spec, RF range below qualified.
- Visible cosmetic flawsScratches >0.5 mm at arm's length (~50 cm), gap >0.3 mm at mating, mis-aligned parts, distorted logos.
- Wrong configurationWrong firmware version, wrong colour, missing accessory, wrong packaging insert.
- Packaging damageCrushed boxes (>10 % surface dented), missing inserts, deformed inner trays.
3.3Minor defects (4.0 % AQL)
- Subtle cosmeticMarks <0.5 mm not visible from 50 cm, minor finish irregularity, gloss inconsistency.
- DocumentationManual misprint not affecting safety, label slightly misaligned (<1 mm), language tag in wrong position.
- PackagingBarcode placement off by a few mm, slight crease in retail box.
3.4Per-product defect catalogue (template)
Each defect needs an entry. Photo each defect type at the actual scale and lighting. The catalogue is the contractual reference for accept/reject decisions.
`` Defect: Scratch on cosmetic A-surface Class: Major Limit: Visible from 50 cm under 500 lux office lighting Photo: CAT-MAJ-001.jpg (with ruler at 1:1 scale) Notes: Below 0.5 mm: minor. Above 1 mm: critical. ``
4.Process capability (Cp, Cpk)
Process capability indices quantify how well a process holds tolerance. Required for medical/automotive; useful for any tooled production.
4.1Cp — process potential
``` Cp = (USL − LSL) / (6σ)
USL = Upper Specification Limit LSL = Lower Specification Limit σ = Process standard deviation ```
- Cp ≥ 1.0Process barely fits within tolerance. Marginal.
- Cp ≥ 1.33Industry-typical target for production.
- Cp ≥ 1.67Six Sigma target. Critical features.
- Cp ≥ 2.0Ultra-stable process. Rare in consumer hardware.
4.2Cpk — process performance (centered)
`` Cpk = min[(USL − μ) / (3σ), (μ − LSL) / (3σ)] μ = Process mean ``
Cpk accounts for process centering. A process can have high Cp but low Cpk if it's not centered. Cpk ≤ Cp always.
4.3Worked example
Critical dimension: 25.00 ± 0.20 mm. Production sampled 50 units.
- Sample mean μ = 24.92 mm
- Sample std dev σ = 0.04 mm
- USL = 25.20 mm, LSL = 24.80 mm
- Cp = (25.20 − 24.80) / (6 × 0.04) = 0.40 / 0.24 = 1.67 (excellent spread)
- Cpk = min[(25.20 − 24.92) / 0.12, (24.92 − 24.80) / 0.12] = min[2.33, 1.00] = 1.00 (off-centre)
- Action: Adjust tooling to centre the mean at 25.00 mm. Cpk will rise to ~1.67.
5.Process-specific defects
Each manufacturing process has characteristic failure modes. Include the relevant set in the QC plan.
5.1Injection molding defects
| Defect | Cause | Severity | Visual signal |
|---|---|---|---|
| Sink mark | Uneven cooling, thick walls | Major | Surface depression over thick section |
| Flash | Worn mold, high pressure, low clamp | Major | Thin film at parting line |
| Short shot | Low pressure, cold material | Critical | Incomplete fill, visible holes |
| Weld line | Flow fronts meeting | Minor / structural | Faint line on surface |
| Burn mark | Air entrapment | Major | Dark spot, often near vent |
| Warp | Uneven wall thickness | Critical | Part doesn't sit flat |
| Splay | Moisture in material | Minor | Silvery streaks |
| Drag mark | Poor draft, scuff during ejection | Major | Linear scratch in ejection direction |
| Jetting | Gate flow turbulence | Minor / cosmetic | Snake-like pattern from gate |
5.2PCBA defects
| Defect | Cause | Severity | Detection |
|---|---|---|---|
| Tombstoning | Reflow profile, pad imbalance | Critical | AOI / visual |
| Solder bridge | Stencil aperture, paste volume | Major | AOI / electrical test |
| Cold joint | Reflow profile (insufficient peak) | Major | Visual / X-ray |
| Missing component | Pick-and-place feeder issue | Critical | AOI / electrical test |
| Wrong polarity | Operator error, programming | Critical | AOI / function test |
| Lifted pad | Rework damage | Major | Visual after rework |
| Insufficient solder | Stencil aperture too small | Major | Visual / X-ray |
| Component shift | Reflow drift, no fiducial | Minor / major | AOI |
| BGA voids | Reflow profile / flux | Major | X-ray (mandatory for BGA) |
5.3Sheet metal defects
- BurrsSharp edges from cutting; functional + safety risk. Deburr per ISO 13715 (broken edge).
- Spring-backBent angle deviates from intent. Material- and tool-dependent; compensate in tool design.
- ScratchesTooling marks or post-process handling.
- Mis-aligned holesPunch wear or stamp misalignment. Punch lifetime: typically 100k–500k strokes.
5.4Final assembly defects
- Mis-aligned screwsStripped threads (over-torque), backed-out screws (under-torque). Spec torque per fastener.
- Mis-mating connectorsForced or partial seating. Specify keying or asymmetric profiles to design out.
- Wrong accessory bundledMixed SKU at packaging station. Bar-code scan per unit at packaging eliminates.
- Surface contaminationFingerprints, dust, oil from line. Gloves and dust extraction at cosmetic stations.
6.QC plan template
The deliverable. One document, version-locked with the spec sheet.
6.1Document structure
6.2Standard inspection equipment
| Tool | Use | Calibration interval |
|---|---|---|
| Caliper (digital, 0.01 mm) | General dim | 6 months |
| Micrometer (0.001 mm) | Critical dim | 6 months |
| Comparator microscope | Optical comparison | 12 months |
| Hi-pot tester | Electrical safety | 12 months |
| Multimeter (DMM) | Continuity, V, I | 12 months |
| Drop tester | Drop test per IEC 60068-2-32 | 12 months |
| Salt-spray chamber | Corrosion test ASTM B117 | 12 months |
| Environmental chamber | T/RH cycling per IEC 60068 | 12 months |
| Force gauge | Latch/snap force | 6 months |
| Light meter | Cosmetic inspection lighting | 12 months |
6.3Standard inspection lighting
- Office lighting300–500 lux. Acceptable for routine cosmetic inspection at arm's length.
- Inspection station750–1 000 lux, daylight 5500 K. Catches subtle cosmetic defects.
- Critical surface inspection2 000+ lux with directional + diffuse mix. Catches micro-scratches.