A CMM best-fit alignment can make a CNC inspection report look better without changing the part. That does not automatically make the analysis wrong: the question is whether the permitted movement matches the drawing requirement. For acceptance, use the specified datum reference frame and evaluation rules. Keep an unconstrained best-fit study clearly identified as diagnostic evidence unless the requirement actually permits that freedom.
This FUMING guide is for engineers and sourcing teams reviewing a supplier’s coordinate measuring machine report. It focuses on one decision: whether an alignment change has legitimately evaluated the requirement or merely redistributed the error. It complements our broader CNC machining services information without claiming that a particular machine, tolerance or certification is included in every project.
Why the same point data can produce different results
An alignment locates the measured part relative to the coordinate system used for evaluation. Translation and rotation change the relationship between measured locations and nominal locations. They do not remove material, straighten a wall or move a drilled hole. Consequently, two reports from the same data can disagree when their reference systems differ.
Hexagon’s PC-DMIS documentation distinguishes two-dimensional, three-dimensional and user-defined best-fit alignments. The allowed degrees of freedom and starting reference matter. Software offering an option does not establish that the option is suitable for a drawing’s acceptance requirement.
A useful supplier discussion therefore begins with the controlled drawing, not the green or red report cells. Ask which datum features establish orientation and location, which characteristic is being evaluated, and what movement the evaluation permits. Record any interpretation agreed with the design authority before comparing pass/fail outcomes.
Separate machine setup, acceptance and diagnosis
| Alignment purpose | Useful evidence | Buyer question |
|---|---|---|
| Initial machine setup | Safe part location and measurement access | Does the final evaluation use the required references? |
| Drawing acceptance | Controlled datum and characteristic definitions | Are constraints consistent with the drawing? |
| Diagnostic best fit | Offsets, rotations and residual patterns | What changed, and what manufacturing action does it suggest? |
A rough setup alignment can be adequate for finding features and still be inadequate for reporting their final relationship. Conversely, a diagnostic best fit can help locate a setup shift even when it must not determine acceptance. Request distinct names for these operations so the report does not silently inherit a convenient coordinate system.
ZEISS’s discussion of CMM inspection errors identifies datum selection, alignment, sampling and the wider measurement setup as contributors to dependable results. A reference chosen because it is easy to reach may not represent the intended assembly interface. The remedy is to resolve the measurement plan, rather than simply collect more points around the same unsuitable reference.
An example of an offset being hidden
Consider an illustrative mounting plate whose hole pattern is positioned relative to specified locating surfaces. Suppose all holes share a similar sideways displacement relative to those surfaces. A free translation fitted to the hole pattern can reduce the reported hole deviations. The mating assembly, however, may still locate the plate from the original surfaces.
The best-fit study is valuable because the common movement may point to a machining setup offset. It is not by itself evidence that the datum-referenced requirement passes. The reviewer should retain the original evaluation and attach the diagnostic comparison with the translation explicitly reported.
This example is conceptual, not a customer result. Real position or profile evaluation can involve material-boundary conditions, simultaneous requirements and permitted datum mobility. Those rules must be interpreted for the actual drawing. Avoid replacing a valid standards-based evaluation with an oversimplified rule that every coordinate must always be fixed.
Request the settings that can change the answer
The phrase “best fit applied” is too vague for a reproducible report. Hexagon’s best-fit setup instructions expose feature selection, fit method, weights and alignment freedom as configurable inputs. Ask the supplier to preserve those inputs alongside the result rather than sending only a screenshot of the final deviations.
- Drawing, CAD and inspection-program revision, with the inspected part identifier.
- Datum features and their precedence for each acceptance characteristic.
- Included and excluded measured features, with reasons for exclusions.
- Translation and rotation freedoms, fit method and any nonuniform weighting.
- Point distribution, feature construction and relevant filtering choices.
- Before-and-after deviations and the transformation used for the diagnostic study.
- Named reviewer and the agreed rule for acceptance near a limit.
Keep these records with the first-article inspection package. A report becomes easier to compare across facilities when both teams can recreate its reference system without guessing from a colored model.
Check acquisition before changing the alignment
If the same part gives unstable results after repeated placement, an alignment debate may be premature. Review seating, accessible datum areas, burrs, cleanliness, stylus qualification and part condition. A fixture must support the intended measurement state; clamping a flexible wall flat can change the geometry being evaluated.
Repeat a limited set of important features after unloading and reloading the part. Keep the inspection program unchanged during this comparison. If results shift with placement, document that variability before editing fit settings. Our datum-transfer review addresses the related risk of changing references between operations.
Also distinguish equipment traceability from suitability for a particular measurement. NIST’s explanation of metrological traceability connects a result to a reference through a documented calibration chain, including uncertainty. Traceability alone does not prove that a particular alignment or sampling strategy evaluates your drawing correctly.
Use a controlled comparison to resolve disagreements
- Freeze the drawing revision, measured dataset and original report.
- List the disputed characteristics and the datum frame each requires.
- Reproduce the original evaluation before adding a diagnostic alignment.
- Change one documented analysis choice at a time.
- Compare transformation values and residual patterns, not only overall pass counts.
- Agree whether the cause is acquisition, evaluation, machining or an unresolved drawing interpretation.
- Release a corrected report only after approval, retaining the superseded record.
If two laboratories still disagree, exchange an identified part and an agreed measurement plan. Avoid treating a second machine’s result as automatically superior. Differences in sampling, feature association, temperature condition or acceptance rules can survive a change of equipment. The practical objective is a reproducible decision about the part’s intended function.
What to include in a purchase-order inspection note
A useful note might request: “Report specified characteristics in the drawing-required datum reference frame. Identify diagnostic best-fit results separately, preserve transformation settings, and obtain approval before substituting a different acceptance evaluation.” Adapt that wording to the project and applicable drawing standard; it is an example, not a replacement for engineering review.
State which dimensions require individual results and which can be summarized. Define the sample selection, file format and treatment of disputed results. Where uncertainty affects acceptance, address it explicitly using the approach discussed in our measurement uncertainty and acceptance guide. Do not ask a supplier to hide a borderline result by increasing rounding or removing unfavorable points.
Frequently asked questions
Is best fit always unsuitable for acceptance?
No. It can be appropriate where the actual requirement and evaluation method permit the selected freedom. The problem is an undocumented or incompatible substitution. Confirm the requirement first and preserve enough settings for another reviewer to reproduce the result.
Can a color map replace the dimensional report?
A color map helps communicate deviation patterns, but it needs a stated alignment, scale and scope. It should not replace characteristic-level results when those are required for acceptance. A visually uniform map may omit precisely the interface that controls assembly.
What should I send FUMING for a report review?
Send the controlled drawing and model revisions, marked critical interfaces, current report and disputed characteristics through the FUMING contact page. Include the intended assembly references and any approved inspection instructions. The review scope and measurement deliverables should be agreed before production or reinspection begins.
