How Should Leak, Proof-Pressure and Dimensional Evidence Be Released Together?
Leak, proof-pressure and dimensional evidence should be released through one serialized part and process record so the buyer can see that the correct geometry was tested under the correct method. Dimensions verify interfaces and remaining material, a leak test measures containment to a defined sensitivity, and a proof test checks survival at a specified load. One result cannot replace the others.
What Each Type of Evidence Actually Proves
Dimensional inspection verifies the drawing characteristics measured by the selected method. It does not demonstrate a continuous pressure boundary. A leak test detects pressure or flow behavior under defined conditions, but a pass does not prove that every wall has the intended thickness. A proof test shows the part withstood its prescribed condition; it may not have the sensitivity to detect the allowable service leak.
Evidence | Primary Question | Cannot Prove Alone |
|---|---|---|
Dimensional report | Are critical features geometrically compliant? | Leak tightness or structural survival |
Leak test | Does measured leakage meet the defined limit? | Minimum wall or life |
Proof test | Does the part withstand specified test load? | Fine leak rate or burst margin |
X-ray/CT as specified | Are reportable internal indications present? | A connected leak path in every case |
Cleanliness report | Does residual contamination meet the method? | Pressure performance |
How the Test Method Should Be Defined
State test medium, fill and vent method, pressure or vacuum, ramp, stabilization, hold, temperature, fixture restraint, allowable leakage, measurement resolution, calibration and retest rule. Product engineering sets values from system risk. The same nominal pressure-decay test can give different readings if part volume, trapped air, temperature or stabilization changes.
Define whether the test is performed on the bare machined housing, with temporary seals, or in an assembled condition. The fixture must seal outside the product boundary without covering the suspected path. Track fixture leakage separately and challenge the setup with known references where the method requires it.
How to Link Results to the Same Part
Use a serial or durable lot identity that connects casting cavity, die and machine, melt or material lot where required, heat/finish state, machining machine and fixture, program, tools, inspection, washing and pressure tests. First-article and validation samples should normally have individual identity because averaging results across different parts breaks causal analysis.
Routine production can use the approved traceability level from the risk plan, but a failed part must be traceable to the affected population. Do not overwrite a failed leak value after rework. Preserve original, repair action, repeat result and disposition.
How Internal Imaging Supports but Does Not Replace Functional Testing
X-ray inspection or CT can help characterize indications in pressure walls, port bosses or machined regions. Technique, orientation, resolution and acceptance must match the feature. Small connected paths can be difficult to infer from a two-dimensional projection, while a visible indication may not connect to a fluid surface.
Correlate images with machining depth, leak location, sections or other qualified evidence during development. Use the correlation to improve process controls and focused sampling. Do not use an image pass to waive a required functional test unless the product authority formally approves that validation strategy.
How to Set First-Article, Pilot and Routine Release
First article should establish complete dimensions and test-method feasibility on representative cavities. Pilot production should show variation over machine warm-up, tool life, fixture positions and casting sources. Routine control then combines capable production checks with scheduled audits and reaction rules.
Stage | Evidence Package | Decision |
|---|---|---|
Development | Zone map, stock, imaging/test correlation | Choose process and methods |
First article | Full dimensional, cleanliness and functional data | Confirm drawing/test compliance |
Pilot | Cavity, tool-life and fixture variation | Confirm short-run capability |
Routine production | Defined inspection/test plus audits | Maintain released state |
How to Investigate a Leak Failure
Preserve the part without impregnation, sealant or grinding. Confirm the tester with its reference, repeat only according to the approved rule and locate whether leakage comes from casting material, a machined face, thread, plug, seal or fixture. Review test temperature and stabilization before declaring a casting defect.
Map the path to cavity, gate/vent region, local section and machining depth. If destructive analysis is needed, document location and orientation before cutting. A leak near a port may arise from connected porosity, thread damage, seal-face geometry or an incorrect temporary plug; each cause needs a different correction.
How Proof-Test Failures Should Be Contained
Stop release and contain the population back to the last verified condition. Preserve pressure trace, fixture state, part temperature, deformation and failure location. Confirm that the method did not apply an unintended external load or hydraulic shock. Engineering decides whether the failure represents material, geometry, machining or test-fixture behavior.
Do not lower the test condition or increase wall thickness without root-cause evidence. A heavier wall can move thermal concentration and create another casting risk. Revisions require renewed structural, casting, machining and functional review.
What a Combined Release Report Should Contain
The report should include part and drawing revision, pressure-zone ID, material and cavity traceability, machining route, critical dimensions, test equipment and calibration, method parameters, raw values, pass/fail, inspection images where required, rework status and approval. Retain units and unrounded raw data sufficient to understand boundary results.
The production quality system should trend dimensions and functional results together. A changing leak-rate baseline with stable dimensions may indicate casting or test drift; a dimensional shift without leakage still requires correction before the interface reaches its failure limit. Combined evidence makes that relationship visible.
Which Changes Require Revalidation?
Review changes to alloy condition, gate, vent, cooling, die insert, port core, machining stock, datum, fixture, seal-land cutter, washing, impregnation, plug, test fixture, test software and acceptance method. Link each change to affected characteristics and repeat the evidence needed to restore confidence.
A release is credible when dimensions, internal-risk controls, cleanliness, leak behavior and proof evidence refer to the same approved design and production route. Keeping those records separate without common identity can create five passing documents that do not describe the same part.
