# CNC Inspection Methods Compared: CMM, NDT and PMI

> Seventeen measurement and NDT methods compared by geometric capability, accuracy, cost tier, cycle time and qualification, with CMM and UT worked examples.

https://www.sinbo-machining.com/wiki/reference/inspection-methods/

---

## On This Page

- Inspection methods compared (17 methods; cost tier is editorial)
- What CMM accuracy do I need to verify a ±0.02 mm position callout?
- Can a caliper check GD&T true position?
- How long does a full-FAI CMM inspection take per part?
- Can blue-light scanning replace my CMM for GD&T sign-off?
- How accurate is a shop-floor Rockwell test?
- How deep can ultrasonic testing find a crack?
- Why did my buyer reject magnetic particle on the aluminum bracket?
- What can XRF verify on incoming material?
- How sensitive is a helium leak test?
- Do salt-spray hours prove outdoor life?
- What reports prove my supplier's inspection capability?
- When does a part need a full FAI again?
- Is a 0.092 mm bore-position reading really out of tolerance at Ø0.05?
- How do I prove a Ø120 mm 4140 Q&T shaft has no internal cracks?
- How to Use These Numbers
- Frequently Asked Questions
- Standards & Sources
- Related Pages
- Download This Table

---

# CNC Inspection Methods Compared

Seventeen measurement and NDT methods compared by geometric capability, accuracy, cost tier, cycle time and qualification, with CMM and UT worked examples.

## Inspection methods compared (17 methods; cost tier is editorial)

| Method | Measures (GD&T scope) | Accuracy / range | Cost tier | Cycle per part | Destructive? | Operator qualification | Detects | Standard | Source |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| CMM, bridge (touch-trigger / scanning) | size, position, flatness, roundness, perpendicularity, parallelism, concentricity, profile — full 3D GD&T | MPE ±(1.7–2.5 + L/250–300) µm; high-end ±(0.6 + L/1000) µm; to ~1 m class | 5 | 3–15 min | no | ISO 10360 calibration report + GD&T-literate measurer | any geometric deviation; probing cannot find internal voids | ISO 10360-2; ASME B89.4.1 | S1–S5 |
| Blue-light / laser 3D scanning | surface profile, 3D best-fit, full-field deviation map | point 0.017–0.05 mm; volumetric 0.02–0.1 mm | 4 | 1–4 min (scan + align) | no | moderate; mesh/alignment skill | form deviation over the whole surface; line-of-sight limits undercuts; internal cavities unreachable | VDI/VDE 2634 Part 3 | S9, S10 |
| Profile projector (optical comparator) | 2D profile, radius, angle, pitch, 2D true position, edges | ±(1.5–2.5 + L/100) µm; magnification ±0.1%; stage 203 × 102 mm typ. | 2 | 15 s–2 min | no | low; parallax/edge training | 2D contour deviation, burrs, thread profile; no depth or roughness | Mitutoyo PJ/PH specs | S6–S8 |
| Vision measuring machine (VMM) | 2D size, position, small features, high feature count | ±(0.8–1.5) µm | 3 | 10–60 s auto | no | low–moderate; program once, run operator | 2D geometry on small/flat parts; edge-defined features | Mitutoyo QV specs | S6 |
| Calipers | linear size (OD/ID/depth/step) quick check | ±0.02–0.04 mm | 1 | seconds | no | none (shop-floor universal) | coarse size error; not for GD&T | ISO 13385-1 class | S6 |
| Outside micrometer | diameter, thickness | ±0.001–0.003 mm | 1 | seconds | no | low | diameter/thickness deviation; lapped faces needed | ISO 3611 | S6 |
| Height gauge on surface plate | height steps, 1D comparative, scribing | ±0.003–0.008 mm | 1 | < 1 min | no | low | height/step error; runout with lever attachment | Mitutoyo indicator class | S6 |
| Surface roughness tester (Ra/Rz stylus) | Ra, Rz, Rt per ISO 4287 | ±0.5–2% of reading; resolution 0.001 µm; a 2 µm tip cannot resolve features below about 4 µm | 1 | 30 s per trace | contact but non-damaging | low; cutoff/filter training | surface texture deviation; not form | ISO 4287; JIS B 0601 | S6 |
| Hardness — Rockwell (HRB/HRC) | bulk hardness (strength / heat-treat proxy) | ±0.5 HRC bench class; loads 60/100/150 kgf | 1 | < 1 min per indent | small indent remains (semi-destructive to cosmetic surface) | low; ASTM E18 procedure training | wrong heat-treat condition; soft spots; missing case | ASTM E18 / ISO 6508 | S15, S16 |
| Hardness — Brinell (HB) | bulk hardness on rough/large parts | 10 mm WC ball, 500/3000 kgf, 15 s; best on castings/forgings | 1 | about 1 min | visible indent (larger than HRC) | low | grade mix-up vs expected HB; Q&T condition | ASTM E10 / ISO 6506 | S15, S16 |
| Hardness — Vickers/Knoop micro (HV/HK) | case depth (CHD), thin sections, coating cross-section | loads 10 gf–10 kgf; lab method | 2 | sample-prep dependent (minutes) | cut + polish section — destructive to the part | moderate (metallurgical prep) | case depth out of spec; decarburization | ASTM E92/E384; ISO 6507 | S15, S16 |
| Ultrasonic testing (UT, incl. PAUT/TOFD) | internal flaws, wall thickness, lack of fusion | 1–5 MHz; penetration 100–300 mm; TOFD depth ±0.5 mm on 25 mm wall; finds tight cracks < 0.1 mm | 3 | minutes per weld/joint | no | ASNT Level II per SNT-TC-1A / CP-189 / ISO 9712 | internal cracks, voids, LOF, inclusions, thickness loss | ASTM E428/E317; EN 1714; ASME Sec V; AMS 2632 | S11, S12 |
| Magnetic particle testing (MT) | surface and near-surface cracks in ferromagnetic parts | cracks below the 1.6 mm class band (A903 MT column); dry/wet, visible/fluorescent | 1 | minutes | no (may need demagnetize/clean) | ASNT Level I/II; NAS 410 for aerospace | grinding, fatigue and quench cracks; seams | ASTM E709; E1444 (aero); ISO 9934 | S13, S14, S24 |
| Liquid penetrant testing (PT) | surface-opening cracks, any material incl. non-ferrous | pairs with MT on castings (MT below 1.6 mm, PT complementary); fluorescent more sensitive than visible | 1 | 15–60 min (dwell incl.) | no (residue cleanable) | ASNT Level I/II | surface cracks, porosity openings, laps — incl. aluminium and austenitic stainless | ASTM E165/E1417 | S14 |
| Radiographic testing (RT / X-ray) | internal voids, porosity, inclusions, weld penetration | volume defects by density contrast; thickness limit set by energy (typ. ≤ 100 mm steel; Co-60/linac beyond) | 4 | 10–30 min per exposure | no (radiation hazard to operator, not part) | ASNT Level II; radiation-safety license (site dependent) | porosity, shrinkage, slag, LOF (planar defects angle-sensitive) | ASTM E94; EN 17635; ASME Sec V | S11 |
| Salt-spray cabinet (NSS) | corrosion resistance of coatings | 5% NaCl, 35 °C per ASTM B117 / ISO 9227; screening only | 2 | 24–1000+ h cabinet (parallel batch) | yes for the tested coupon (exposure degrades the finish) | low; chamber operation | coating failures: white/red rust onset, blister, creep | ASTM B117; ISO 9227 | batch1 T3 table (caliber cross-reference) |
| PMI — OES spark / XRF / helium leak test | chemical composition (OES all elements incl. C; XRF no C); hermeticity | OES 0.01% class (carbon weak — LECO combustion better); XRF portable non-destructive; helium MSD 5×10⁻¹²–5×10⁻¹³ Pa·m³/s, response < 1 s | OES/XRF 2; helium 3 | OES/XRF seconds; helium minutes | OES leaves a spark mark (slight); XRF and helium non-destructive | OES/XRF instrument training + PMI procedure; helium vacuum training | grade mix-up, off-spec incoming chemistry, leak rate over limit | ASTM E415/E1086/E572; API RP 578; helium: vendor nameplate | S20–S23 |

> **How to read this table.** Cost tier 1 (lowest) to 5 (highest) is an editorial rating compiled from published equipment cost anchors (CMM around $215K class, scanner $142K class, projector $48.5K class) — not a standard value. GD&T column lists the ASME Y14.5 / ISO 1101 characteristics the method can evaluate. Salt-spray hours are a screening comparison, not outdoor life (same caliber as the coating-thickness table). Helium leak rates are instrument nameplate values with a conservative method-level caliber logged in the conflicts.

## What CMM accuracy do I need to verify a ±0.02 mm position callout?

A bridge CMM at ±(1.7–1.9 + L/300) µm resolves it with roughly a 7:1 margin — at 300 mm the MPE is about ±2.7–2.9 µm against a 20 µm tolerance zone, inside the 4–10:1 rule-of-thumb band. Route each purpose with the matrix below before booking machine time.

| Inspection purpose | Primary method | Quick screen | Disabled / ineffective | Accuracy anchor |
| --- | --- | --- | --- | --- |
| Size + full GD&T (position/profile/flatness) | CMM | blue-light scan (full profile fields) | calipers cannot evaluate GD&T | MPE ±(1.7 + L/300) µm |
| Small 2D profile / thin deformable parts | projector / VMM | — | contact probes deform thin walls | ±(1.5 + L/100) µm |
| Surface roughness Ra/Rz | stylus roughness tester | — | CMM and projector cannot measure roughness | ±0.5–2% of reading |
| Heat-treat condition check | Rockwell HRC (batch sampling) | Brinell (rough large parts) | hardness alone cannot identify the grade | ±0.5 HRC |
| Case depth (carburized/nitrided) | Vickers micro section | — | macro hardness (HRC/HB) cannot see the layer | CHD per ISO 2639 |
| Internal flaws (cracks/voids/LOF) | UT incl. PAUT/TOFD | RT (best on volume defects) | MT/PT only see surface openings | TOFD depth ±0.5 mm per 25 mm |
| Surface cracks (ferromagnetic) | MT | PT (only option for non-ferromagnetic) | UT near-field masks the surface | MT below the 1.6 mm class |
| Coating corrosion comparison | NSS salt spray | — | cabinet hours ≠ outdoor life (ISO 10683) | endpoint must state white vs red rust |
| Grade / mix-up verification (incoming) | OES (carbon matters) or XRF (no carbon) | spark test (shop qualitative) | XRF cannot see carbon — cannot separate 304 from 304L | OES 0.01% class |
| Hermeticity | helium mass spectrometry | pressure decay (coarse leaks) | bubble method only locates gross leaks | 5×10⁻¹² Pa·m³/s class |

## Can a caliper check GD&T true position?

No — a caliper's ±0.02–0.04 mm accuracy only supports linear size; position needs a CMM (±(1.7 + L/300) µm class). Hand tools stay valuable for what they actually resolve:

| Hand tool | Accuracy | What it honestly resolves |
| --- | --- | --- |
| Calipers | ±0.02–0.04 mm | linear OD/ID/depth/step quick check — no GD&T |
| Outside micrometer | ±0.001–0.003 mm | diameter and thickness |
| Height gauge | ±0.003–0.008 mm | height steps; runout with a lever attachment |
| Stylus roughness tester | ±0.5–2% of reading | Ra/Rz only — not form, not size |

## How long does a full-FAI CMM inspection take per part?

Plan 3–15 min per part on a bridge CMM — 14 min is typical for a part with 7 radii R0.15 ± 0.02 and a stepped bore. A profile projector pre-screen (under 90 s per part) shortens the CMM queue; how much shorter depends on your scheduling — treat that as a shop editorial estimate, not a source figure.

| Timing / cost anchor | Value | Status |
| --- | --- | --- |
| Bridge CMM cycle | 3–15 min per part | published range |
| Landing-gear bushing case: 7 radii R0.15 ± 0.02 + stepped bore | 14 min per part | source case |
| Projector pre-screen | under 90 s per part | published |
| CMM queue-time reduction from pre-screening | scheduling-dependent (editorial estimate) | no source figure — deliberate replacement of an unsourced claim |
| CMM equipment anchor | $215K class | TCO table |
| Scanner equipment anchor | $142K class | TCO table |
| Projector equipment anchor | $48.5K class | TCO table |

## Can blue-light scanning replace my CMM for GD&T sign-off?

Not for position of internal features — scanners reach 0.017–0.05 mm point accuracy and give a full-field profile map, but line-of-sight limits blind holes. Use them for profile of surface; keep the CMM for datums and true position.

| Scanner class | Stated accuracy | Note |
| --- | --- | --- |
| GOM Scan 1 | 0.017 mm single point | VDI/VDE 2634 Part 3 metering standard |
| ATOS Q | 0.02 mm class | current-generation rated value |
| Handheld laser | 0.02 mm class | imported industrial tier |
| ATOS (2010 lab data) | 0.002–0.02 mm, single frame 1 s | older wide band kept for reference |

## How accurate is a shop-floor Rockwell test?

±0.5 HRC on a bench tester at 60/100/150 kgf — enough to verify a 28–34 HRC pre-hard 4140 band, but hardness cannot identify the grade itself. The three methods divide the work:

| Method | Division of work | Standard |
| --- | --- | --- |
| Rockwell (HRB/HRC) | batch hardness verification of heat-treated parts | ASTM E18 / ISO 6508 |
| Brinell (HB) | rough, large parts: castings and forgings | ASTM E10 / ISO 6506 |
| Vickers/Knoop micro (HV/HK) | case depth (CHD), thin sections, coatings — destructive section | ASTM E92/E384 / ISO 6507 |

## How deep can ultrasonic testing find a crack?

Conventional UT penetrates 100–300 mm of steel at 1–5 MHz, and TOFD sizes crack depth to ±0.5 mm on a 25 mm wall — MT and PT stay surface-only. Wall-thickness check: a 3.4 µs back-wall echo at 5,850 m/s reads 9.94 mm.

| Parameter | Value | Source |
| --- | --- | --- |
| Frequency band | 1–5 MHz conventional | S11 |
| Penetration in steel | 100–300 mm | S11 |
| TOFD depth sizing | ±0.5 mm on a 25 mm wall | S11 |
| Tight crack detection | below 0.1 mm | S11 |
| Thin-material spec | 0.5–12.7 mm at 2.0–10.0 MHz | S12 |
| Wall-thickness formula | 5,850 m/s × 3.4 µs ÷ 2 = 9.94 mm | S11 |
| Qualification | ASNT Level II (SNT-TC-1A / CP-189 / ISO 9712) | S24 |

## Why did my buyer reject magnetic particle on the aluminum bracket?

MT only works on ferromagnetic material (ASTM E709 scope) — for aluminium and austenitic stainless the surface-crack method is penetrant testing (ASTM E165). The casting acceptance split that confuses buyers:

| Method | Covers | Source |
| --- | --- | --- |
| MT | ferromagnetic parts; detects cracks below the 1.6 mm class band | S13, S14 |
| PT | any material; complementary above 1.6 mm on castings — a division of labor, not a conflict | S14 |

## What can XRF verify on incoming material?

Full alloy chemistry except carbon — so XRF cannot separate 304 from 304L. Add spark OES (0.01% class carbon) or require a mill cert when carbon matters; helium leak sits in the same incoming-verification toolbox:

| PMI tool | Strength | Limit |
| --- | --- | --- |
| XRF | portable, non-destructive, seconds per reading | cannot detect carbon |
| Spark OES | all elements incl. carbon at 0.01% class | leaves a small spark mark; bench instrument |
| LIBS | handheld; fast screening | lower precision than OES for certification |
| Carbon (LEEC combustion) | accurate carbon when OES is marginal | sample destruction |
| Helium MSD leak test | 5×10⁻¹²–5×10⁻¹³ Pa·m³/s, response under 1 s | nameplate sensitivity, not a method guarantee |

## How sensitive is a helium leak test?

A mass-spectrometer helium leak detector resolves 5×10⁻¹²–5×10⁻¹³ Pa·m³/s within seconds — pressure-decay test floors are orders of magnitude coarser. The 10⁻⁹ caliber seen elsewhere is a conservative method-level figure; this page quotes instrument nameplate and logs the difference.

| Caliber | Sensitivity | How to read it |
| --- | --- | --- |
| Instrument nameplate (MSD) | 5×10⁻¹²–5×10⁻¹³ Pa·m³/s, response < 1 s | vendor high-end class (S22, UGC-sourced; logged as nameplate value) |
| Conservative method-level | 10⁻⁹ Pa·m³/s order | S23; may describe the whole rig rather than the detector |
| Pressure decay / bubble | orders of magnitude coarser | volume- and temperature-dependent; no fixed floor (declared gap) |

## Do salt-spray hours prove outdoor life?

No — ASTM B117 hours are a screening comparison; ISO 10683 states accelerated results cannot be directly related to service. Same caliber as the coating-thickness table, cross-referenced:

| Parameter | Value |
| --- | --- |
| Electrolyte | 5% NaCl solution |
| Cabinet temperature | 35 °C |
| Solution pH | 6.5–7.2 |
| Collection rate | 1–2 mL per 80 cm² per hour |
| Meaning | screening comparison — not outdoor service life |

## What reports prove my supplier's inspection capability?

An annual ISO 10360 CMM calibration certificate, a gauge list with ISO/IEC 17025-traceable certificates, and NDT operator ASNT Level II certificates cover most buyer audits. Assemble them per scenario:

| Scenario | Dimensional evidence | Material / process evidence | Deliverable |
| --- | --- | --- | --- |
| AS9102 full FAI (new part number) | CMM report on all characteristics + ballooned drawing (Form 3) | MTR/CoC + HT/plating/NDT certificates (Form 2) | Form 1 + 2 + 3 + ballooned drawing |
| AS9102 delta FAI (revision / machine / supplier change) | re-test affected characteristics only | certificates for the changed processes only | delta Form 3 referencing the original FAIR |
| PPAP Level 3 (automotive) | dimensional report + MSA / gauge R&R | material / performance test reports | PSW + 18-element package, Cpk ≥ 1.67 on key characteristics |
| Buyer capability audit | annual ISO 10360 CMM calibration certificate + 17025-traceable gauge list | ASNT Level II certificates + approved special-process vendors | one package covers most audits |

## When does a part need a full FAI again?

AS9102 Rev C triggers a full FAI on a new part number, a fit/form/function design change, a process or supplier change, or a production lapse of more than two years — and after a two-year lapse the standard allows a full or a delta FAI, so pin down which one your customer requires.

| Trigger | FAI scope |
| --- | --- |
| New part number | full FAI |
| Design change affecting fit, form or function | full FAI (or delta per agreement) |
| Process or supplier change | delta FAI on affected characteristics |
| Production lapse over two years | full or delta FAI per AS9102 Rev C — confirm which with the customer |
| Revision | AS9102 Rev C, published 2023-06 |

## Is a 0.092 mm bore-position reading really out of tolerance at Ø0.05?

Yes — and the budget math is what settles the argument. The combined standard uncertainty is √(6² + 4² + 5² + 3²) = √86 ≈ 9.3 µm (k = 1); at the source's expanded caliber U(k=2) ≈ ±0.018 mm. Either way 0.092 − 0.018 = 0.074 mm still exceeds the 0.025 mm radial boundary — the deviation stands, it is not measurement error.

| Step | Item | Value | Note |
| --- | --- | --- | --- |
| 1 | CMM MPE at L = 200 mm | ±(1.9 + 200/300) ≈ ±2.57 µm | Zeiss Contura-class specification (S4/S5) |
| 2 | Combined standard uncertainty u_c (k=1) | √(6² + 4² + 5² + 3²) = √86 ≈ 9.3 µm | probe 6 + scale 4 + thermal 5 + setup 3, root-sum-of-squares (S3 budget, values in µm) |
| 3 | Expanded uncertainty U (k=2) | ≈ ±0.018 mm | 2 × u_c, per the source's own U(k=2) = ±0.018 mm caliber |
| 4 | Verdict | 0.092 − 0.018 = 0.074 mm > 0.025 mm | out of tolerance stands under both calibers — argument only matters near the boundary (ISO 14253-1) |
| 5 | Fitting requirement | ≥ 16 points, least-squares cylinder; 3-point method amplifies 4.3× | S3 verification practice |

## How do I prove a Ø120 mm 4140 Q&T shaft has no internal cracks?

UT, not RT or MT: MT/PT see only the surface, and a solid bar needs high-energy RT that misses planar flaws anyway. Run straight-beam longitudinal plus angle-beam shear at 2–5 MHz, evaluate to ASME Sec V with ASNT Level II operators.

| Step | Item | Value |
| --- | --- | --- |
| 1 | Method elimination | MT/PT surface-only; RT needs high energy, insensitive to planar flaws |
| 2 | Frequency selection | 2–5 MHz (fine-grained forging); 0.5–12.7 mm thin stock per AMS 2632 at 2.0–10.0 MHz |
| 3 | Wall-thickness check | 5,850 m/s × 3.4 µs ÷ 2 = 9.94 mm |
| 4 | Acceptance caliber | crack-like (planar) indications evaluated per ASME Sec V; pressure-boundary forgings typically zero tolerance |
| 5 | Personnel | ASNT Level II; report bound to the day's calibration status |

## How to Use These Numbers

- Cost tier 1–5 is an editorial rating from the published TCO anchors (CMM $215K class, scanner $142K class, projector $48.5K class) — not a standard value and not a quote.
- The CMM-queue-time reduction from projector pre-screening is deliberately stated as a scheduling-dependent editorial estimate: no source figure exists for it, and an unsourced percentage claim was removed from this page on verification.
- Uncertainty arithmetic: combine component budgets as a root-sum-of-squares at k = 1, then multiply by 2 for the ~95% expanded U(k=2) — quoting a k=1 result as U is a classic report error.
- GD&T evaluation: only the CMM covers all 3D characteristics; projectors and VMMs are 2D instruments, calipers and micrometers cannot evaluate GD&T, and roughness testers cannot see form.
- NDT division of labor: MT ferromagnetic surfaces only (aluminium and austenitic grades go to PT), UT for internal and depth sizing, RT for volume defects but angle-sensitive on planar flaws, XRF blind to carbon.
- Salt-spray hours are screening numbers, not outdoor life — the endpoint (white vs red rust) must be stated, same caliber as the coating-thickness table.
- The GB/T 16857.2-2017 acceptance protocol is cited by standard number only (ISO 10360-2 equivalent); the ISO 10360-2 acceptance test itself is 5 lengths × 7 directions × 3 repetitions = 105 points.
- Pressure-decay and bubble-method leak sensitivities have no citable fixed floor (volume/temperature dependent) — declared gap, ranked qualitatively only.

## Frequently Asked Questions

### What CMM accuracy do I need to verify a ±0.02 mm position callout?

A bridge CMM at ±(1.7–1.9 + L/300) µm resolves it with roughly a 7:1 margin — MPE about ±2.7–2.9 µm at 300 mm against a 20 µm zone.

### Can a caliper check GD&T true position?

No — a caliper's ±0.02–0.04 mm accuracy only supports linear size; position needs a CMM.

### How long does a full-FAI CMM inspection take per part?

Plan 3–15 min per part on a bridge CMM — 14 min is typical for a 7-radius, stepped-bore part; a sub-90 s projector pre-screen shortens the queue by a scheduling-dependent amount (editorial estimate).

### Can blue-light scanning replace my CMM for GD&T sign-off?

Not for internal-feature position — scanners reach 0.017–0.05 mm but are line-of-sight; keep the CMM for datums and true position.

### How accurate is a shop-floor Rockwell test?

±0.5 HRC on a bench tester at 60/100/150 kgf — fine for a 28–34 HRC band, but hardness cannot identify the grade.

### How deep can ultrasonic testing find a crack?

Conventional UT penetrates 100–300 mm at 1–5 MHz; TOFD sizes depth to ±0.5 mm on a 25 mm wall.

### Why did my buyer reject magnetic particle on the aluminum bracket?

MT only works on ferromagnetic material (ASTM E709) — aluminium needs penetrant testing (ASTM E165).

### What can XRF verify on incoming material?

Full alloy chemistry except carbon — it cannot separate 304 from 304L; add spark OES (0.01% class) or a mill cert when carbon matters.

### How sensitive is a helium leak test?

A helium MSD resolves 5×10⁻¹²–5×10⁻¹³ Pa·m³/s within seconds; pressure-decay floors are orders of magnitude coarser.

### Do salt-spray hours prove outdoor life?

No — ASTM B117 hours are a screening comparison; ISO 10683 states accelerated results cannot be directly related to service.

### What reports prove my supplier's inspection capability?

An annual ISO 10360 CMM calibration certificate, a 17025-traceable gauge list, and ASNT Level II operator certificates cover most buyer audits.

### When does a part need a full FAI again?

AS9102 Rev C: new part number, fit/form/function change, process or supplier change, or a production lapse over two years — lapse cases may be full or delta FAI per customer requirement.

## Standards & Sources

- S1 — CNC57, ISO 10360 / JIS B7440 CMM acceptance indicators (7 directions × 5 lengths × 3 repetitions = 105 points)
- S2 — GB/T 16857.2-2017 / ISO 10360-2 (equivalent acceptance-test standards, cited by number only — the original listing's URL was truncated, downgraded per verification)
- S3 — TestGageInsight, How does a coordinate measuring machine work (bridge ±(1.7+L/300), gantry ±(2.5+L/250), Leitz ±(0.6+L/1000); uncertainty budget 6/4/5/3; U(k=2) ±0.018 mm; TCO anchors; 16 points / 4.3×)
- S4 — JPT Online, Zeiss CMM specifications (MPE 1.9 + L/300 µm)
- S5 — Zeiss Contura-class MPE-E ±(1.9 + L/300) µm listings (UGC reprints, corroborated by the official specification sheets)
- S6 — Japan Monozukuri Lab, Mitutoyo instrument accuracy tables (micrometers, calipers, SJ roughness, CRYSTA-Apex, projectors, QUICK VISION)
- S7 — Mitutoyo PH-A14 profile projector official specifications (0.001 mm resolution, ±0.1%/±0.15%, 203 × 102 mm stage)
- S8 — TestGageInsight, profile projector applications and limits (±(1.5 + L/100) class, 2D-only, 22–48 s per part)
- S9 — Industrial scanner accuracy tiers (GOM Scan 1 0.017 mm, ATOS Q 0.02 mm; VDI/VDE 2634 Part 3) — industry Q&A reprint
- S10 — CAS Institute of Advanced Manufacturing, ATOS scanner parameters (0.002–0.02 mm, 1 s per frame, 2010 unit)
- S11 — Atlantis NDT, Ultrasonic Testing ultimate guide (1–5 MHz, 100–300 mm, TOFD ±0.5 mm/25 mm, tight cracks < 0.1 mm, ASNT Level II)
- S12 — SAE AMS2632C, ultrasonic inspection of thin materials 0.5–12.7 mm at 2.0–10.0 MHz
- S13 — ASTM E709-21 official page, magnetic particle testing guide
- S14 — ASTM A903/A903M-99(2024), steel castings surface acceptance (MT below 1.6 mm, PT complementary, 5 acceptance levels)
- S15 — ZwickRoell, hardness testing standards map (E18/6508, E10/6506, E92-384/6507, E140)
- S16 — TestMetals, hardness testing practice (E18 loads, E10 3000/500 kgf, micro 10 gf–10 kgf)
- S17 — SteelCalculator, steel hardness chart (220↔19 to 660↔62; Fu ≈ 3.45 × HB)
- S18 — BalloonPlot, AS9102 Rev C FAI complete guide (Form 1/2/3, ballooning, triggers)
- S19 — New Tech Metals, AS9102 Rev C (effective 2023-06-28), FAI vs PPAP
- S20 — UTEC, unknown-material identification (hardness → spark → OES; OES 0.01% class)
- S21 — LGC Standards, PMI guide (XRF/OES/LIBS; ASTM E415/E572/E1086/E1999/E2465, API RP 578)
- S22 — Helium mass-spectrometer leak detector nameplate values (5×10⁻¹²–5×10⁻¹³ Pa·m³/s, response < 1 s) — encyclopedia entry, logged as nameplate
- S23 — Shangling, gas leak test methods (bubble / pressure decay / helium MSD, 10⁻⁹ conservative caliber)
- S24 — Atlantis NDT, ASTM E709 application page (ASNT Level I/II/III, E1444 aerospace)
- S25 — TestGageInsight, optical comparator vs CMM vs laser scanning comparison

## Related Pages

- [/wiki/quality/measurement/](/wiki/quality/measurement/) — Long-form measurement capability guide behind this quick table
- [/wiki/quality/parts-rejected-by-cmm/](/wiki/quality/parts-rejected-by-cmm/) — Why CMM reports reject parts — and how to argue with the evidence
- [/wiki/reference/hardness-conversion/](/wiki/reference/hardness-conversion/) — Sibling quick table: hardness scale conversion for the Rockwell/Brinell rows

## Download This Table

- CSV — /wiki/reference/download?slug=inspection-methods&format=csv&lang=en
- JSON — /wiki/reference/download?slug=inspection-methods&format=json&lang=en
- Markdown — /wiki/reference/inspection-methods/index.md