Common Causes of Poor CMM Repeatability

Common Causes of Poor CMM Repeatability

Poor CMM repeatability is one of the most frustrating dimensional inspection problems because it creates doubt everywhere.

The part measures good, then it measures bad. One operator gets one result, another operator gets something different. A setup looks fine during prove-out, then becomes unstable during production.

The CMM often becomes the center of the argument, but the machine is not always the root cause.

Repeatability problems usually come from the measurement process around the CMM. The issue may be the fixture, part condition, probe/stylus system, alignment strategy, feature evaluation method, environment, operator setup and run practices, or a combination of several small problems adding together.

The goal is not to guess. The goal is to isolate which part of the measurement process is changing.

What This Checklist Covers

This checklist helps separate common sources of CMM repeatability problems before the team starts changing programs, rerunning parts, or blaming the machine.

Common review areas include:

1. Fixture and setup stability — whether the part is located, supported, restrained, and loaded consistently.
2. Probe/stylus system condition — whether the probe/stylus system is set up and qualified correctly, built rigidly, and appropriate for the feature being measured.
3. Alignment and measuring strategy — whether the program establishes the part coordinate system using stable, repeatable, and functionally appropriate features.
4. Feature evaluation and reporting logic — whether the measured result reflects the inspection requirement and is not being distorted by point strategy, fitting method, filtering, or report setup.
5. Part condition and handling — whether temperature, burrs, chips, oil, surface condition, or part movement is affecting the result.
6. Operator setup and run practices — whether operators are loading, cleaning, clamping, qualifying, running, reviewing, and escalating the process consistently.
7. Environment and timing — whether temperature, machine warm-up, part stabilization, vibration, airflow, or inspection timing is influencing the result.

The goal is to identify the unstable variable instead of treating every poor repeatability result as a CMM problem.

Signs Your CMM Repeatability Problem Needs Review

These are common indicators that repeatability problems may be coming from the measurement process, not just the part or the CMM.

1. The same part measures differently after being removed and reloaded.
2. One operator gets a different result than another operator using the same program.
3. The result changes after probe/stylus system qualification.
4. The part passes and fails without a clear production pattern.
5. A feature is unstable even though the machine appears to be running normally.
6. The setup looks acceptable, but the part does not seat the same way every time.
7. GR&R results are poor, but no one can clearly identify whether the problem is the part, fixture, program, operator practice, or measurement method.
8. The CMM program has been adjusted more than once without solving the repeatability issue.
9. The report shows variation, but the team cannot explain where the variation is entering the measurement process.
10. Production, Quality, and Engineering are debating the result instead of isolating the source of the change.

If any of these sound familiar, the issue usually is not solved by simply rerunning the part. The measurement chain needs to be reviewed.

Fixture and Setup Instability

A CMM program can only be as stable as the setup supporting it.

If the part is not located, supported, and restrained consistently, the measurement result can shift even when the part itself has not changed. This does not always mean the fixture is poorly designed. Sometimes the fixture is acceptable for holding the part but not stable enough for the measurement strategy being used.

Common fixture and setup issues include:

1. Part not seated consistently.
2. Clamp force changing from operator to operator.
3. Burrs, chips, oil, or debris affecting contact.
4. Part resting on inconsistent surfaces.
5. Flexible parts moving during loading or measurement.
6. Fixture wear or loose components.
7. Setup method depending too much on operator judgment.

A repeatability problem that changes between setups is often a setup control problem before it is a CMM problem.

Probe/Stylus System Issues

The probe/stylus system has to be appropriate for the feature, access condition, and measurement strategy.

A stylus system that is too long, too flexible, poorly qualified, damaged, or not suitable for the feature geometry can create repeatability problems. Small deflections or qualification issues may not be obvious until the same feature is measured multiple times or compared across operators and setups.

Common probe/stylus system issues include:

1. Stylus length or extension that is not rigid enough.
2. Tip size that is not appropriate for the feature being measured.
3. Qualification not performed correctly.
4. Probe/stylus system changed without confirming the impact.
5. Poor access angle causing inconsistent contact.
6. Dirt, wear, or damage on the stylus tip.
7. Using one stylus setup for features that need different access or contact behavior.

A strong measurement strategy starts with a probe/stylus system that is qualified correctly, built rigidly, and suitable for the features being measured.

Weak Alignment or Measuring Strategy

Alignment strategy is one of the most common sources of unstable CMM results.

If the program aligns the part using features that are unstable, poorly formed, too small, flexible, dirty, or not functionally appropriate, the entire measurement result can shift. The measured feature may not be changing. The coordinate system used to evaluate it may be changing.

Common alignment and measuring strategy issues include:

1. Aligning to features with poor form.
2. Using too few points or weak feature coverage.
3. Measuring datums in a way that does not support the print requirement.
4. Building an alignment that is sensitive to small setup changes.
5. Using a theoretical alignment that does not match the physical setup.
6. Measuring features with a method that does not match the tolerance requirement.
7. Evaluating the wrong feature relationship for the decision being made.

Poor repeatability often appears in the reported characteristic, but the cause may be upstream in the alignment or measuring strategy.

Feature Evaluation and Reporting Logic

The number on the report depends on how the feature was measured and evaluated.

Different fitting methods, filtering choices, point distribution, outlier handling, and characteristic setup can change the reported result. This is especially important on features with form error, rough surfaces, interrupted geometry, thin walls, or features that are difficult to access.

Common evaluation issues include:

1. Inconsistent point strategy.
2. Weak coverage of the actual functional surface.
3. Filtering or outlier settings that are not understood.
4. Fitting method that does not match the inspection intent.
5. Characteristic setup that does not match the drawing requirement.
6. Reporting output that hides the source of variation.
7. Comparing results from different measurement methods as if they are equivalent.

Before blaming the part or the CMM, the evaluation method needs to be reviewed.

Part Condition and Handling

Sometimes the part really is changing, or the part condition is changing enough to affect the result.

Temperature, burrs, chips, oil, surface condition, handling, part relaxation, or manufacturing variation can all affect repeatability. This is especially true when tolerances are tight, materials are sensitive, or parts are measured before they have stabilized.

Common part-condition issues include:

1. Parts measured at different temperatures.
2. Parts not thermally stabilized before inspection.
3. Burrs or chips affecting fixture contact.
4. Flexible parts changing shape after unclamping.
5. Surface finish or form variation affecting measurement points.
6. Parts handled differently between operators.
7. Mixed production conditions being compared as one population.

A stable measurement process requires a stable part condition. If the part condition changes, the measurement result may change with it.

Operator Setup and Run Practices

Operator variation does not always mean the operator is doing something wrong.

Often, it means the process gives operators too much room to make different decisions. If the setup method, cleaning method, probe/stylus system qualification method, run sequence, or escalation path is not clearly defined, different operators will create different conditions.

Common operator setup and run practice issues include:

1. Different cleaning or deburring habits.
2. Different loading pressure or clamp sequence.
3. Different interpretation of setup instructions.
4. Different reaction when a result fails.
5. Different decisions about rerunning a part.
6. Different probe/stylus system qualification habits.
7. Different understanding of what the report means.

Good training and clear work instructions reduce variation because they reduce the number of judgment calls required during routine inspection.

Environment and Timing

Environmental effects are easy to underestimate.

Temperature change, machine warm-up, part warm-up, air flow, vibration, and time between production and inspection can all influence measurement results. These effects are not always dramatic, but they can matter when tolerances are tight or parts are sensitive.

Common environment and timing issues include:

1. Measuring parts before they reach inspection-area temperature.
2. Comparing first-shift and second-shift results without controlling conditions.
3. Running measurements before the machine is stable.
4. Measuring near doors, air flow, or heat sources.
5. Inspecting parts immediately after machining or washing.
6. Ignoring fixture temperature.
7. Comparing results collected under different shop conditions.

Environment does not need to be perfect, but it does need to be understood and controlled well enough for the tolerance being measured.

Poor CMM repeatability is rarely solved by simply rerunning the part.

The better approach is to isolate the measurement chain: fixture, part condition, probe/stylus system, alignment strategy, feature evaluation, operator setup and run practices, and environment.

Once the unstable variable is identified, the team can make a targeted correction instead of continuing to debate whether the CMM, the part, or the operator is the problem.

Related Resources

For a deeper review of unstable inspection results, see Measurement Reliability Troubleshooting.

If repeatability issues are creating pass/fail uncertainty, review the False Reject Root-Cause Checklist.