Paper cup process capability is meaningful only when the dataset describes a defined, stable process and a trustworthy measurement method. A high Cpk value calculated from a few selected good cups does not demonstrate that a line will reliably meet the buyer specification during normal production.
This guide prepares the evidence for a capability study. It does not set a universal acceptance index, sample size or cup tolerance. A competent quality specialist should choose the analysis, uncertainty treatment and decision rule appropriate to the characteristic and contract.
Diagnostic decision table
| Before calculating | Required evidence | Common error |
|---|---|---|
| Define the characteristic | Drawing feature and limits | Mixing rim and body measurements |
| Validate measurement | Repeatability and suitable resolution | Assuming calibration is sufficient |
| Establish stability | Time-ordered production observations | Removing inconvenient drift |
| Choose the analysis | Distribution and sampling assumptions | Using one formula for every dataset |
Choose one characteristic and one specification
Identify the exact feature, drawing revision, unit and applicable specification limits. Define the measurement plane, orientation and conditioning state for flexible cups. If only one limit applies, do not invent a second limit to fit a convenient spreadsheet. Keep the engineering target separate from the observed process average.
Agree why the capability result is needed: a commissioning review, a production improvement or a contractual acceptance. The purpose affects the observation period and required evidence. A short trial on one cup size should not be presented as a capability certificate for every product that the machine can theoretically form.
Make the production data representative
Collect measurements in production order and retain material lot, forming position, recipe and relevant stop events. Include the operating conditions covered by the claim. Do not inspect a large group and calculate the index only from the cups that already passed, because that truncates the observed variation.
Separate materially different products or process regimes before analysis. Mixing cold-start output, stable running and a tooling change into one unlabeled column hides the structure of the data. Excluding a condition may be appropriate for a narrowly defined study, but the exclusion and its effect on the scope must be stated.
Check stability before predicting performance
NIST describes capability as a comparison between an in-control process and specification limits. Review the time-ordered data for changes in level, spread and other relevant patterns before making that comparison. Choose suitable control-chart or time-series methods with the quality specialist rather than relying on a histogram alone.
If a drift is associated with a material change or restart, investigate it and define the process state being studied. Do not delete the affected readings simply to improve the index. Preserve the original record and document any new trial after corrective action. Stable does not necessarily mean acceptable; a stable process can consistently produce the wrong dimension.
Use an index with stated assumptions
For the conventional two-sided normal-process model, Cp compares specification width with process spread, while Cpk also reflects how close the mean is to the nearer limit. The selected estimate of spread, distribution assumptions and sample support matter. State the method rather than reporting only an unexplained number.
Ask the analyst to assess non-normality, dependence among observations and the effect of measurement variation. Report uncertainty or confidence information where required. Avoid silently switching between within-subgroup and overall spread estimates; those choices answer different questions about the period and process variation represented.
Make the conclusion narrower than the evidence
A useful capability report identifies the feature, limits, sample source, observation period, measurement method, stability assessment and calculation approach. State the agreed acceptance criterion and whether the evidence supports it. Keep the actual observed nonconforming count available alongside any model-based estimate of future performance.
Send HANNAI the time-ordered readings and process context if a forming adjustment is being considered. A low index can reflect centering, variation, measurement or mixed conditions; these require different responses. Confirm the relevant cause and repeat an appropriately scoped study after the authorized change instead of using the index itself as a machine-setting instruction.
Related equipment and next checks
Review the paper cup production line and the related guide to measurement repeatability checks. The Engineering Notes archive connects these checks with wider machine planning.
Technical reference
- NIST process capability and its assumptions — general reference; confirm applicability to the installed equipment.
- NIST assessment of process stability — general reference; confirm applicability to the installed equipment.
Frequently asked questions
Does a high Cpk from selected good cups prove capability?
No. The sampling and process assumptions must support the claim.
Is a stable process automatically within specification?
No. Stability and conformity are different questions.
Can one study approve every cup size?
Only the documented scope is supported; other configurations need appropriate evidence.
What Cpk target should be used?
Use the agreed quality or contractual criterion, not a generic online threshold.
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.