A bracket passes inspection, is lifted out of its nest, and fails after being loaded again. The immediate temptation is to tighten the clamp or average the two numbers. Neither action tells the team whether the part changed, the setup changed or the measurement process is too variable for the decision.
For precision OEM sheet-metal assemblies, fixture repeatability and part-to-part variation must be investigated separately. Otherwise a supplier may adjust a capable fabrication process to compensate for an unstable inspection method.
Define the variation you are trying to explain
Repeated measurements without moving a part answer a narrower question than repeated measurements after unloading and reloading it. The first may largely reflect instrument and measurement-cycle variation. The second also includes locating, support contact and loading technique.
A study involving different operators adds another source of variation. Measuring different parts introduces actual product differences. Combining all of these into one unlabeled range prevents useful diagnosis.
NIST’s gauge R&R guidance treats repeatability, reproducibility, stability and study design as distinct issues. A short fixture check is useful, but it should not be presented as a complete measurement-system evaluation.
| Check | What stays controlled | What it can reveal |
|---|---|---|
| Repeat without unloading | Same part, position and operator | Short-term instrument or cycle variation |
| Unload and reload | Same part and instruction | Locating and loading sensitivity |
| Change operators | Common parts and method | Differences in setup execution |
| Measure a representative part set | Controlled measurement process | Product spread plus remaining measurement variation |
Make the nest’s intended contacts visible
Begin with a physical check. Look for burrs, debris, coating buildup, worn pins, loose stops or a contact surface that was never intended to locate the part. A folded return touching the fixture wall can replace the drawing datum without anyone changing the program.
Separate locating from supporting and holding. Record which surfaces establish the datum frame, which pads prevent sagging and which clamps retain the part. If clamp force or sequence matters, it belongs in the method.
Do not quietly add a shim to make one part sit correctly. A shim changes the measurement system and may conceal a product nonconformance. Investigate the cause and obtain approval for any revised setup.
Run the reload check before comparing suppliers
Select a stable part and use a written sequence. Measure it, remove it fully, reload it and repeat the measurement. Keep the instrument, program, environment and part condition unchanged as far as practical. Record every result rather than retaining only the preferred reading.
If the readings move with reloads, inspect the seating and restraint before judging process capability. If they remain consistent for one operator but shift between operators, compare the loading sequence and interpretation of the instruction.
A thin sheet-metal part can also be altered by repeated clamping. Check whether the study itself is bending, marking or otherwise changing the sample. Repetition is not valid evidence if the measured item is no longer stable.
Design the wider study around the production decision
Use parts that represent the range and geometries the method will encounter, including relevant finishes and difficult interfaces. A study of nearly identical easy-to-seat parts may miss the mechanism causing real production disagreements.
Decide with the quality team how operators, repeats, order and analysis will be handled. Randomizing the measurement order can reduce memory and sequence effects. Agree the evaluation criteria with the customer; a single percentage threshold should not be copied into every application.
A gage R&R study also does not resolve every metrology issue. Bias, stability and the chosen acceptance rule remain relevant. Good short-term repeatability cannot rescue a consistently incorrect datum interpretation.
Keep process adjustment separate from measurement repair
If the fixture is unstable, correct and re-evaluate it before changing bend compensation or hole coordinates. Otherwise the shop may introduce real product variation while trying to chase measurement noise.
After a method change, identify which earlier decisions may be affected. Reinspection or customer review should follow the agreed nonconformance process. Do not overwrite the original measurements with new values and erase the reason for the change.
Request evidence that supports release
An RFQ inspection package should name the fixture revision, loading instruction, relevant measurement study and response to a suspect gauge. Connect these records to the production control plan and the product configuration.
The useful outcome is a defensible distinction: what variation belongs to the parts, what belongs to measurement, and what action each one requires. That is more valuable to an OEM than a large spreadsheet of unexplained readings.
Need a Custom Sheet Metal Quote?
Send us your CAD drawings — DRAmetal's initial response target is within 24 working hours. Quotation, DFM and lead-time timing depend on file completeness and engineering scope.
GET A FREE QUOTE →
