KEY TAKEAWAYS

✓
Always verify the datum reference frame first — misaligned datums invalidate every measurement that follows.

✓
A dimension shown as ‘out of tolerance’ in red does not automatically mean a reject; understand bonus tolerance and MMC/LMC modifiers before escalating.

✓
GD&T callouts like true position and flatness are measured in 3D space — a simple caliper cannot replicate them.

✓
Measurement uncertainty (typically reported as U= values) must be factored into borderline pass/fail decisions.

✓
Request the native CMM report file, not just a PDF summary — summary reports can omit critical deviation data.

A Coordinate Measuring Machine captures discrete X, Y, and Z point coordinates on a part surface using a calibrated probe. From those points, the software constructs geometric elements — planes, cylinders, spheres, lines — and compares them against the nominal values in your CAD model or drawing. This is fundamentally different from manual inspection. A micrometer measures a linear distance between two contact points. A CMM reconstructs geometry and evaluates it against toleranced feature definitions, including complex GD&T callouts that cannot be evaluated any other way.
What CMM does not measure: surface finish (Ra/Rz), material hardness, coating adhesion, or internal defects. Those require separate instruments — profilometers, hardness testers, and NDT methods respectively. Engineers sometimes assume a clean CMM report means the part is fully conformant. It means the dimensionally inspected features are conformant. Confirm with your supplier which features were inspected and which were not. A good inspection plan covers all drawing callouts, not just the easy ones. At Nimble’s certified partner network, CMM inspection is included with orders — and the report covers all critical-to-function features identified during DFM review.

A CMM report is only as good as the inspection plan behind it. If the plan skips features, the report gives false confidence. Always ask: ‘What features were excluded and why?’

The header is not boilerplate — it contains information you need to validate before reading a single measurement. Look for:

• Part number and revision level — confirm it matches your current drawing. A report run against Rev B when you released Rev C is meaningless.
• Date of inspection and operator ID — relevant for traceability if a field issue surfaces later.
• CMM machine ID and calibration date — the machine must be within its calibration interval. Expired calibration = suspect data.
• Measurement software and version — PC-DMIS, Calypso, Metrosoft, and others calculate certain GD&T callouts slightly differently. Knowing the software helps resolve discrepancies.
• Temperature at time of inspection — CMM labs are ideally maintained at 68°F (20°C). Deviations cause thermal expansion errors, especially on aluminum and long features.

Experienced engineers scan this section in 30 seconds. Procurement teams often skip it entirely. Don’t. A report with an expired CMM calibration date is not a valid quality record and should be rejected before you read a single dimension.

The datum reference frame (DRF) is the coordinate system to which every measurement on the report is anchored. If the DRF is set up incorrectly, every positional and orientation tolerance on the report is wrong — even if the CMM ran perfectly. The DRF is defined by your drawing’s datum feature symbols (A, B, C) and their precedence order in feature control frames. The CMM operator constructs this coordinate system by probing the datum features in the correct sequence. A common error: probing a datum feature with too few points. A plane requires a minimum of three points, but three points define any plane — including a warped one. Best practice is six or more points for a primary datum plane to average out surface variation.
On the report, look for a section labeled ‘Alignment’ or ‘Coordinate System Setup.’ It will list the features used to establish each axis and the residual error of the alignment fit. A large residual error (sometimes shown as RMS deviation) is a red flag — it means the datum features themselves may be out of spec, or the operator used an incorrect probing strategy. Any downstream positional callout using that DRF is suspect until the datum alignment is verified. This is one of the most common sources of false failures and false passes in machined part inspection.

If Datum A is a face with flatness issues, the entire DRF rocks on that surface. Always check that datum features themselves are within their own tolerance before trusting positional data built on them.

The measurement results table is the core of any CMM report. Each row typically contains:

• Feature name or ID — correlates to a balloon number on the drawing or a labeled feature in the CAD model.
• Nominal — the target value from the drawing.
• Actual (measured) — what the CMM found.
• Deviation — the algebraic difference between actual and nominal (actual minus nominal). A positive deviation means oversize or shifted in the positive axis direction; negative means the opposite.
• Tolerance — the allowable deviation, either as a bilateral range (+/- value) or a unilateral limit.
• Status — pass (often green) or fail (often red).

The deviation column is where engineers should focus, not just the pass/fail color. A dimension that passes by 0.0001 inch on a 0.005 inch tolerance is statistically close to the edge — worth flagging for process capability review. A dimension that shows consistent deviation in the same direction across multiple parts indicates a systematic offset, not random variation. That is a process problem, not an outlier, and it warrants a corrective action conversation with the supplier.

GD&T callouts are where most non-metrology engineers get lost. Each control frame result in a CMM report follows the same structure as the drawing callout: geometric characteristic, tolerance value, and applicable material condition modifier. Here is how to read the most common ones:
True Position: Reported as a diameter zone (e.g., 0.008 inch diameter zone). The actual value shown is the calculated positional error in diametral terms — it must be less than the tolerance zone diameter to pass. If the drawing uses the MMC modifier (circled M), the report will also show bonus tolerance — additional tolerance earned when the feature departs from maximum material condition. Always check whether bonus tolerance was applied; it often turns apparent failures into passes.
Flatness and Straightness: Reported as a single value representing the total band within which all measured points must fall. Unlike position, these have no datum reference — they are self-contained measurements of form only.
Cylindricity and Circularity: These evaluate roundness and cylindrical form. High values on bore features can indicate chatter, tool deflection, or worn tooling. A cylindricity failure on a bearing bore is a functional failure, not just a paperwork issue. Profile of a Surface: Increasingly common in aerospace work. Reports the maximum deviation from the nominal surface across all probed points, evaluated against the specified bilateral or unilateral tolerance.

Bonus tolerance under MMC can be significant. A 0.010 inch true position tolerance at MMC on a 0.250 inch hole could grow to 0.020 inch or more if the hole is produced at LMC. Always look for the bonus column before calling a positional failure.

Every CMM measurement has associated uncertainty — a quantified doubt about the true value of the measurement. Uncertainty arises from probe calibration limits, thermal effects, surface finish of the measured feature, number of sample points, and the mathematical fitting algorithms used. On well-run reports, uncertainty is expressed as a U= value at a stated confidence level (typically 95%, or k=2). ASME B89.7.3.1 provides guidance on how to handle uncertainty in conformance decisions.
For borderline results — where the measured deviation is close to the tolerance limit — uncertainty matters enormously. If a dimension measures 0.0048 inch deviation against a 0.005 inch tolerance, and measurement uncertainty is +/- 0.0005 inch, the true value could actually be 0.0053 inch. The part may be nonconforming. Rigorous aerospace and defense suppliers operating within an AS9100-registered quality system will flag this explicitly. They apply a guard band — accepting only results that fall inside the tolerance by at least the uncertainty value. This is more conservative than simple pass/fail, but it is the correct approach for safety-critical hardware. If a supplier’s CMM report shows no uncertainty values, ask for them. A report without uncertainty disclosure is incomplete.

The rule of thumb: CMM measurement uncertainty should be no greater than 10% of the feature tolerance being measured. On a 0.002 inch tolerance, your CMM system uncertainty must be 0.0002 inch or better. Tighter tolerances require better equipment and more controlled environments.

Knowing what a good report looks like also means knowing when something is wrong. Here are the most common issues engineers encounter:

• Missing features: The inspection report covers 12 dimensions, but your drawing has 28 callouts. Ask what the inspection plan was and why features were excluded.
• Rounded actuals: If every actual value ends in a round number (e.g., 1.0000, 0.5000), the data may have been manually entered rather than machine-generated. Legitimate CMM output almost always shows more decimal places with minor variation.
• Identical actual values across multiple parts: If 10 parts all show exactly the same measured value for a given feature, something is wrong. Real manufacturing produces variation. Identical values suggest copy-paste, not measurement.
• No probe qualification record: The probe must be calibrated against a reference sphere before each inspection session. If this is absent, probe compensation errors may corrupt the data.
• Datum alignment RMS above 0.001 inch: On tight-tolerance work, a high-residual datum alignment should prompt re-inspection after corrective action on the datum features.

Nimble’s certified partner network operates under AS9100 and ISO 9001 quality systems, which require documented inspection plans, traceable calibration records, and controlled measurement environments — the infrastructure that prevents these red flags from appearing in the first place.

A dimensional nonconformance is not automatically a scrapped part. Most quality systems provide a formal path to use-as-is (UAI) or repair dispositions through a Material Review Board (MRB) process. The key question is functional impact: does the deviation affect fit, form, or function? A positional deviation of 0.007 inch on a clearance hole in a non-critical bracket is very different from the same deviation on a locating hole for a precision assembly.
When a part fails, the first step is to verify the failure is real — not a measurement artifact from a datum setup error, probe calibration issue, or thermal error. Request a reinspection if the margin is small. If the failure is confirmed, the supplier should issue a nonconformance report (NCR) documenting the deviation, root cause, and corrective action. For aerospace and defense work under AS9100, this is mandatory, not optional.
For use-as-is dispositions on non-flight-critical parts, an engineering disposition signed by the responsible engineer is typically sufficient. For flight hardware or safety-critical applications, the customer and often the end-use OEM must approve the disposition. Never accept a verbal ‘it’s fine’ — get the disposition in writing with an engineering signature. The CMM report, NCR, and disposition together form the traceability package that protects you if the part ever comes back as a field issue.

Always request the NCR and engineering disposition in writing before accepting a nonconforming part. A verbal approval is not a quality record — and it will not protect you during a customer audit or failure investigation.

READY TO SOURCE?

Table of Contents

Related Services

Get a Quote in 24h

Upload your drawing and get a detailed quote with free DFM review.