How Should One Porosity Finding Trigger Containment Across a Production Lot?
One porosity finding should trigger immediate hold of the affected part and a risk-based containment population defined by casting cavity, production lot or shot range, machining fixture, tool and operation, inspection history and functional severity. It should not automatically condemn every historical lot or release adjacent parts without evidence.
Preserve the as-found indication, map its location and determine whether it crosses a seal, pressure, thread, bearing or structural zone. Traceability then narrows the most credible population for expanded inspection.
What Information Must Be Frozen Before Rework
Photograph under defined scale and lighting, record coordinates, opening dimensions, cluster and machined depth, then retain part, chips or tool evidence where useful. Record casting cavity/lot, date or shot range, machine, fixture pocket, program revision, cutter and tool life. Do not blend, impregnate or cut deeper until engineering accepts the evidence plan.
Identity | Why It Narrows Containment |
|---|---|
Casting cavity/tool | Tests whether internal distribution follows one source |
Shot/lot/time range | Connects process-window or tool-condition changes |
Machining fixture/pocket | Separates cut depth and location effects |
Tool and life | Checks tearing or smearing mechanism |
Functional test result | Sets severity beyond appearance |
How to Set the Initial Hold Population
Begin with material since the last accepted inspection or process verification relevant to the zone. Include unmachined castings that will receive the same cut when the mechanism may be upstream. If traceability cannot separate cavities or time, the hold must expand accordingly. Customer or shipped-stock notification depends on demonstrated risk and contractual rules.
A pressure-boundary leak or crack-like indication justifies faster and broader escalation than a permitted isolated pit on a hidden dry face. Severity, not production convenience, determines the first boundary.
How Expanded Inspection Should Be Designed
Inspect the same mapped zone using a method capable of finding the observed condition. Add functional leak or load tests when the finding affects performance. Review matched parts from different cavities, times and fixture positions so the sample can test patterns. Random sampling without strata can miss a cavity-specific issue.
When screening unmachined parts by radiography or CT, prove that the method can detect the relevant feature before using negative results to release. Machining selected controls may be necessary to expose the exact plane.
Sampling Stratum | Question | Possible Escalation |
|---|---|---|
Same cavity, same time | Is the condition clustered locally? | Expand time window |
Other cavities, same time | Is the issue tool-wide or cavity-specific? | Review common process controls |
Same fixture/tool path | Could machining create/expose the pattern? | Tool and stock-depth study |
Earlier accepted lots | When did the signal begin? | Customer/shipped stock review |
When the Lot Can Be Released or Must Expand
Release requires a documented acceptance basis, completed expanded inspection, disposition of every held subgroup and evidence that the inspection could detect the relevant condition. If additional findings appear, expand by the pattern they reveal. Repeated location in one cavity directs upstream investigation; findings only after one machining tool or fixture position direct machining review.
Do not reduce containment because a quick rework makes parts pass appearance. The original functional risk and repaired condition still need validation.
How Containment Feeds Corrective Action
Send a common defect map to casting engineering and post-machining. Review gate/overflow/vent or vacuum/cooling and process records alongside stock depth, fixture location, cutter and tool load. Select controlled trials from the evidence instead of changing many variables.
After correction, verify more than the single failed zone: confirm casting distribution, machining cleanup, dimensions and functional integrity across representative cavities and lots. Update the control plan and reaction threshold.
What Objective Evidence Closes Containment?
Containment can close when the affected population is defined, every held subgroup has a disposition, the screening method has demonstrated capability for the observed condition, no unacceptable expansion pattern remains, correction is implemented where needed and a planned number of representative subsequent lots meets both defect and functional requirements. The quality authority should record the effective date and any temporary elevated sampling.
A root-cause statement without verification does not close containment. Neither does screening only easy-to-inspect surfaces when the original finding was on a hidden machined plane.
Closure Gate | Required Record |
|---|---|
Population complete | Cavity, lot, time, fixture and shipped status |
Screening capable | Detectability and inspector/equipment setup |
Held stock dispositioned | Use, rework, repair, scrap and retest identity |
Corrective action verified | Representative cavity/lot results after change |
Control plan updated | New monitor, limit and reaction owner |
If the mechanism remains unresolved but screening reliably protects function, engineering can approve a temporary controlled plan with explicit review date. It should not quietly become the permanent process without capability, cost and recurrence review.
Keep rejected and accepted comparison specimens through the corrective-action verification period. They provide a stable reference if another lot produces a borderline indication or an inspection method changes.
One finding should create a disciplined evidence boundary, not panic or denial. Good traceability narrows the population; qualified inspection and functional severity determine whether containment can close or must expand.
