Manufacturing Control Plan: From Process Risk to Response

Start with the process step and the failure consequence

A process flow identifies where work happens. A risk review such as FMEA identifies what can fail and why the consequences matter. The control plan specifies how the process will be controlled and what the team does when the check fails.

AIAG publishes a dedicated Control Plan manual alongside its APQP guidance. Use the current customer and sector requirements where they apply. This article supplies an original teaching record, not the licensed manual, an official template or proof that a process meets an automotive requirement.

Make the characteristic measurable

In the illustrative assembly, a fastener joint has an assumed specified installation torque of 8.0±0.5 N m. This is fictional, not a torque recommendation. The critical issue is to connect the stated requirement to an approved method and a meaningful reaction.

A programmed torque setting alone does not establish the torque actually achieved or joint performance. Check the approved tool method, calibration, measurement uncertainty, joint behavior and applicable validation. Gage R&R addresses a separate measurement-study question; it cannot qualify every torque method automatically.

Choose the frequency from risk and detection needs

A convenient hourly sample may leave a long interval of suspect production after a tool failure. A continuous check also has failure modes: bypass, missing records, incorrect part linkage or a detector that misses the actual problem. Choose the rule using consequence, process behavior and demonstrated detection.

If a hypothetical line makes 120 units per hour and a failure could have occurred anytime since a last accepted check 2 hours ago, up to 240 units lie inside that time window before other evidence narrows it. That is a containment scope, not 240 proven defects. Missing timestamps or rework routes can widen uncertainty.

Write the reaction before the first failed result

The reaction needs to protect product and establish who investigates, verifies the correction and authorizes restart. Safe stopping and work on equipment remain governed by site procedures. An article must not supply an improvised machine-isolation instruction.

Define what happens when the measurement itself is invalid or missing. Do not classify missing evidence as a passing result. Link the suspect scope to lot traceability and retain the original failed value even after adjustment.

1. Map the operation

   Identify the step, characteristic, requirement and relevant risk.

2. Verify the check

   Confirm that the method detects the intended condition under representative work.

3. Define the response

   Name the product hold, responsible role, investigation and authorized restart.

4. Control the revision

   Change the plan, instruction, measurement and training together at an explicit effective point.

Keep specifications separate from control limits

Specification limits state the applicable product requirement. Control limits describe process variation estimated using a suitable chart and baseline. A result can lie within specifications while the process shows a signal that deserves investigation.

Use the control-chart guide for that interpretation. Do not paste specification limits into a chart and label them calculated control limits. Keep unit, sampling rule, chart method and response consistent with the actual data.

Audit whether the plan can be used on the shift

Ask a qualified user to follow one check and one failure response. Can they find the right revision, identify the affected product and contact the authorized owner? Observe the work rather than judging the neatness of the spreadsheet.

The download contains the fictional joint example and blank fields for your own checks and responses. Review any real control plan against the site's approved process and applicable customer requirements.