An approved sample shows what “good” looked like at one point in time. Repeatable production needs a document that carries the approved configuration into every relevant process stage.
A control plan provides that operating link. It defines what must be controlled, where the control occurs, which method governs the result, how often the check runs, which record preserves the evidence, and what happens when the result is abnormal.
For a buyer, the useful question is whether each critical requirement can be traced to an executable control and a defined reaction.
What a control plan actually controls
A specification states the required result. A work instruction explains how to perform a task. An inspection report records observations. A control plan connects these elements across the manufacturing flow.
Each controlled characteristic needs six fields: requirement or risk; process point; method and reference; sample size and frequency; record; and reaction.
A row that says only “check resistance” or “inspect appearance” leaves essential decisions open. The operator still needs to know when the check occurs, which controlled method applies, how the result is recorded, and which event triggers adjustment, segregation, escalation, or a process hold.
Start from the approved configuration and the failure path
Build the plan from the approved bill of materials, drawings, approved sample, functional requirements, packaging requirements, process flow, and known failure modes. The control plan should reference those controlled sources instead of recreating them.
A typical empty disposable AIO path may cross incoming components, subassembly, electrical connection, reservoir and airway assembly, closure, functional verification, cosmetic inspection, packaging, and lot release.
Place a characteristic at the earliest practical point where variation can be prevented or detected. Final inspection may reveal a defect after its containment window has already grown and its original cause has become harder to reconstruct.
Separate product characteristics from process characteristics
Product characteristics describe the hardware output: a dimension, resistance result, airflow response, fit, seal condition, function, appearance, or identification mark. Process characteristics describe the controlled inputs used to create that output: fixture condition, joining parameter, component orientation, tooling state, or another production variable.
Both layers matter. A finished device may pass one functional check while the process producing it is drifting. A recorded process setting also cannot prove every output characteristic is acceptable. A strong row connects the process control to the product result it protects.
Choose CTQs by consequence and detection opportunity
Critical-to-quality characteristics should follow the approved product definition and the consequences of failure. Review whether the feature affects basic function or compatibility, becomes difficult to detect after later conversion, contributes to leakage or airway restriction, is sensitive to tooling or material variation, or would require buyer approval if it deviated.
The drawing remains the complete requirement set. The control plan concentrates recurring attention on characteristics where loss of control creates meaningful risk. It should also show why the selected method and frequency can detect a relevant change soon enough to limit exposure.
Build every row as a closed control loop
1. Identify the characteristic precisely
Use the characteristic name and reference from the approved drawing or specification. “Airflow” is incomplete if the decision concerns pressure response under a defined fixture condition. “Appearance” needs named defect families or a controlled visual standard.
2. Name the process point
State whether the control occurs at incoming receipt, setup approval, first-off, in-process, after a specific assembly step, final functional verification, packaging, or release.
3. Lock the method and reference
Name the gauge, fixture, test program, visual standard, approved sample, or controlled instruction. Record the revision whenever a method change could alter the result. Our guide to vape hardware test-method validation explains why comparable measurements require agreement on the measurement system before acceptance decisions are made.
4. Set sample size and frequency from risk
Frequency should reflect detection opportunity, process stability, change points, the potential containment window, and the cost of delayed detection. Applicable triggers may include startup, first-off, a project-defined interval, tool or material change, maintenance, or restart after downtime.
Sample size should reflect measurement variability, available process knowledge, the chance and consequence of missing relevant variation, and the project’s agreed decision rule; it should be justified separately from how often the check runs.
The project must define its own frequency. A universal number would ignore the actual process, risk, capability, and buyer agreement.
5. Specify the record and traceability keys
The record should preserve the result, production sequence or time reference, line or station when relevant, equipment or fixture identity when needed, operator or system identity, and the related component or finished-goods lot. Those keys determine how accurately the team can define the affected window later.
6. Define reaction authority before production
A usable reaction states both the immediate controls and the decision authority. The plan should identify who may place material on hold, who decides disposition, which evidence is required after correction, and who may authorize restart. Those roles must come from the project’s approved responsibility matrix; the article does not assign universal factory titles.
The reaction sequence should cover:
- Stop or hold where appropriate.
- Segregate output since the last verified acceptable check.
- Identify the affected component, work-in-process, and finished-goods scope.
- Confirm the measurement system.
- Investigate the assignable cause.
- Correct the process and document disposition.
- Verify effectiveness.
- Resume under the authorized restart rule.
A repeated check that produces one passing unit does not establish containment. Adjustment without preserving the original result can also erase evidence needed to find the cause.
Three example rows, with project-specific limits left open
The examples below demonstrate document logic. They are editorial examples rather than ILEVA production records, specifications, test results, certifications, or universal acceptance plans.
| Characteristic | Process point | Method/reference | Frequency | Record | Reaction |
|---|---|---|---|---|---|
| Coil-resistance result | After electrical connection and before enclosure completion | Controlled resistance fixture and approved instruction | Project-defined first-off, interval, and restart triggers | Actual result, lot, station, fixture identity, and sequence reference | Hold the defined window; confirm the measurement system; investigate connection and component causes; verify correction; obtain authorized restart |
| Draw-activation response | Completed hardware functional verification | Approved airflow or pressure-response setup and controlled procedure | Project-defined functional-screening plan | Result, method revision, station, and lot | Segregate the defined window; check fixture and airway condition; investigate common causes; verify correction and disposition |
| Housing/reservoir interface condition | After assembly and before final enclosure or packaging | Specified dimensional or fixture check plus controlled visual criteria | Startup, relevant change events, and project-defined intervals | Results tied to component and assembly lots | Hold further conversion where appropriate; contain from the last verified point; evaluate tooling and component contribution; verify before restart |
The approved specification supplies the actual limits. The agreed method makes the resulting evidence comparable.
Keep process control and lot acceptance separate
A control plan manages production over time. An AQL sampling plan supports a defined lot-acceptance decision. A stable in-process trend may improve confidence, yet final lot inspection can still be required. A passing final sample does not erase a known abnormal production window.
First Article Inspection confirms the initial production output against approved requirements; the control plan then carries selected characteristics into recurring execution.
Treat the empty-hardware boundary explicitly
The supplier’s empty-hardware plan should preserve the interfaces the buyer or contract filler will later depend on, such as closure condition, airway integrity, traceable configuration, and agreed compatibility evidence. The buyer’s filling parameters, oil-specific settings, post-fill closure window, and filled-product transport controls belong to a separately approved filling and finished-product plan.
This boundary prevents an empty-hardware control plan from making unsupported claims about the buyer’s oil, process, or final filled system while still protecting the handoff between the two operations.
Revise the plan when the controlled state changes
Review the plan after an approved design or material change, new tooling or fixture, process relocation, repeated defect, complaint, measurement-method revision, corrective action, or evidence that the current frequency is not detecting drift early enough.
The change-control record authorizes the new state. The control-plan revision shows how that state will be controlled and when the revision becomes effective.
Buyer review questions before a PO or pilot run
Ask the supplier to provide the actual controlled plan or an agreed buyer-facing extract, then walk it against the real process flow:
- Which approved requirement or risk does each CTQ control?
- Where is the earliest practical prevention or detection point?
- Which method revision governs the result?
- How were sample size and frequency chosen, and which events reset or increase the frequency?
- Can each record be tied to the affected lot and production window?
- Who can hold material, decide disposition, and authorize restart?
- What evidence is required before the process resumes?
- How do FAI, deviations, AQL release, traceability, and change control connect to the plan?
The buyer does not need unrelated proprietary process detail. The objective is enough controlled information to verify that the approved configuration has an executable control path.
Where the control plan fits
Sample testing and First Article Inspection establish approved references. Method agreement supports comparable measurements. The control plan turns selected requirements into recurring action. Component traceability defines the affected scope. AQL sampling supports lot acceptance. Deviation and change-control processes govern authorized exceptions and revisions.
When sourcing Empty Disposable Pens for an OEM, white-label, or sample-stock program, align the control-plan fields before treating a successful sample as evidence of repeatable production.
Final takeaway
The strongest control plan is the one another trained person can execute without guessing. Every important row connects an approved requirement to a process point, method, frequency, record, reaction, and named decision authority.
That structure turns a general promise of quality control into an auditable path from approval through production evidence and release.
