A paper cup destructive test gage study is needed when the test changes or destroys the specimen, so the same cup cannot be measured repeatedly. Leak-to-failure, peel or compression tests may show variation from both the cups and the method. Treating different cups as if they were one identical part can overstate measurement error or hide product variation.
This is a study-design guide, not a prescribed statistical model. A competent quality engineer must select the study structure and confirm whether samples within a production group are homogeneous enough for the intended assumptions.
Diagnostic decision table
| Study condition | Possible structure | Critical check |
|---|---|---|
| Every operator tests subsamples from each group | Crossed or expanded design may be possible | Within-group specimens are sufficiently homogeneous |
| Each operator receives unique sample groups | Nested or expanded design may apply | Part identity is nested under operator |
| Extra fixed factors must be studied | Expanded design may be needed | Model matches the actual allocation |
| Only one specimen exists per condition | Gage variation cannot be isolated easily | More justified samples or another method are required |
Define what the destructive test measures
Write the response, unit, endpoint and specimen preparation. A time-to-leak test differs from a pass/fail hold test, even when both use liquid. Record how the cup is filled, held and observed, and which event ends the test. Separate operator judgement from instrument readings where possible.
Identify all steps that can vary between laboratories or operators: conditioning, fixture, loading rate, liquid, timing and data processing. Stabilize the method before designing the formal study. A statistical analysis cannot rescue results collected under changing definitions or undocumented specimen preparation.
Create defensible homogeneous sample groups
Select production groups expected to provide similar cups within each group while spanning the relevant process range between groups. Preserve forming position, material lot, time and handling identity. The homogeneity assumption should come from process knowledge and supporting evidence, not simply from consecutive sample numbers.
Minitab guidance explains that destructive gage studies rely on similar specimens being treated as the same part for analysis. If within-group product variation is large, it can mask method variation. Plan enough independent groups and specimens with the analyst rather than taking all test pieces from one unusually stable period.
Match allocation to crossed or nested structure
In a crossed arrangement, every operator evaluates material representing every selected group. This requires defensible comparable specimens within each group. If each operator tests unique specimens or groups that cannot be shared across operators, a nested or expanded design may better describe the actual experiment.
Randomize test order where safe and practical, while blocking known constraints such as limited conditioning batches. Record the allocation before testing. Do not relabel unique specimens with the same identifier after the fact merely to make software accept a preferred model. The worksheet must represent the physical sample relationships.
Preserve invalid tests and method events
Record every planned specimen, including fixture slips, spills, sensor errors and departures from the endpoint rule. Mark invalid tests using predefined criteria and retain the reason. Silent deletion can make the method appear more repeatable and prevents investigation of failure-prone steps.
Have the analyst separate estimated operator, repeatability and specimen-group components as supported by the design. Review plots and raw values, not only a single percentage. The interpretation must reflect assumptions, balance and uncertainty. Avoid applying a crossed result when the actual samples were nested or confounded with operator.
Improve the method and confirm the change
Use the evidence to standardize preparation, fixtures, training or endpoint detection. Control the revised method and run a new planned study with representative groups. Do not pool old and new method data into one variability estimate unless the analyst has designed that comparison explicitly.
Define how ongoing checks will detect drift in the destructive method. Send HANNAI the response definition, sample-group logic, allocation table and failure observations when the results relate to machine quality. This keeps true cup variation visible while preventing unstable testing from driving unnecessary process adjustments.
Related equipment and next checks
Review the paper cup production line and the related guide to dimension measurement repeatability study. The Engineering Notes archive connects these checks with wider machine planning.
Technical reference
- Minitab: study choices for destructive measurements — general reference; confirm applicability to the installed equipment.
Frequently asked questions
Can the same cup be tested twice after destruction?
No. The study must use justified comparable specimens and an appropriate design.
Is a crossed gage study always required?
No. Allocation and homogeneity determine whether crossed, nested or expanded analysis fits.
Should invalid test attempts be deleted?
Retain them with documented reasons and apply predefined validity rules.
Can one stable production group represent the full method?
Use groups that support the intended range and analysis.
Review the evidence with HANNAI
Send your machine model, component identification and the observations described in this guide. Include the drawing or material reference and any before-and-after measurements so the required next check can be identified.