Locate the first separation before enlarging the corner
A crack appears near the corner of a drawn EMI shield, and the first proposed change is a larger radius. That may eventually be a useful design option, but the visible final crack does not establish where the damage began. A separation can reach the rim after starting in the wall, or extend inward from an edge that was already present during forming.
Start by tracing the earliest supported location and manufacturing stage of the separation. Combine that sequence with the condition of the adjacent metal. The purpose is to distinguish plausible failure paths before changing a controlled corner, replacing the material or adjusting a forming operation on the basis of one photograph.
Preserve the fracture and the material around it
Retain the failed piece with its identity, orientation and process stage. Photograph the complete corner and both accessible faces before cleaning, opening the crack farther or cutting a section. Those interventions can alter the surface and remove information about the original condition.
Keep adjacent material rather than saving only a small broken fragment. The surrounding wall can show whether the separation lies within a broader deformed region, whether it meets an original cut edge and how it relates to the top-to-wall transition. A detached fragment without its location can be difficult to interpret even under a powerful microscope.
Identify whether the visible line is an actual separation in the substrate, a coating feature or an observation still awaiting confirmation. Do not infer depth from darkness. If examination requires a destructive preparation, choose and document the location so that the procedure answers the unresolved question rather than simply producing an attractive cross-section.
Reconstruct which edges existed when the damage appeared
Map the corner through the actual sequence: incoming blank, relevant forming stages, trimming, local cutting and subsequent handling. Show the original perimeter and any opening present before the draw. Distinguish those boundaries from edges created only after the corner had already been formed.
This chronology can disprove an attractive explanation. If an identified intermediate part already contains a separation before final trimming, that later trim cannot have initiated the earlier observed damage. The trim may expose or enlarge it, but those are different contributions. Conversely, a defect first found after trimming leaves the earlier condition uncertain unless the pre-trim examination had suitable coverage.
Where representative stage samples are available through a controlled process trial, inspect the same material region at each stage. Preserve how that region moves from the blank into the wall. Final-part coordinates alone may not identify which original edge or blank region supplied the cracked corner.
Compare an edge-origin path with a wall-localization path
An edge-origin hypothesis should identify the particular boundary from which the separation is thought to have developed. Examine its condition before the relevant forming step where evidence is available, and compare neighboring edge regions. A rough edge beside the final crack is a lead, not automatic proof that edge quality caused the failure.
A wall-localization hypothesis asks whether deformation became concentrated within the wall or transition before separation. Look for the distribution of local thickness and shape around the suspected region, using a suitable method and a consistent section direction. A narrow reduced section can be relevant, but its interpretation must account for the material, geometry and stage represented.
Steel-industry guidance distinguishes necking-related failures from local fractures that can occur without an observable neck. Its specific examples include advanced high-strength steels. The practical lesson is limited but useful: absence of visible thinning does not prove that a feature is sound, and does not by itself prove an edge-origin mechanism in a different shield material.
Allow an unresolved branch. A severely opened final crack may have destroyed the earliest evidence, or available sections may miss the initiation region. Reporting that limitation is more useful than forcing every corner failure into one of two labels.
Measure neighboring metal without confusing section geometry
A thickness comparison should relate to the local material section, not an oblique projection through a curved corner. Record the section plane and the surfaces used to define the measurement. If the prepared plane cuts the wall obliquely, the apparent distance across the section can differ from thickness measured normal to the wall.
Compare several identified positions around the suspected origin and a relevant intact region. A single minimum value from the broken lip does not describe the deformation field that preceded fracture. Likewise, a sound remote roof measurement cannot establish that the corner retained the same thickness.
Preserve edges during preparation. Mounting and polishing can affect the appearance of thin boundaries through rounding, smearing or loss of support. The laboratory should select preparation appropriate to the material system and retain images of the as-received condition. A prepared image is evidence from that selected location and preparation history, not a complete reconstruction of the whole crack.
Do not use this investigation to invent a new minimum wall requirement. A measured local change can help test a failure explanation while the acceptance decision still depends on the controlled product definition and the actual condition found.
Use comparisons that test the suspected origin
If the evidence points toward a pre-existing cut edge, a useful controlled comparison holds the material and forming target comparable while changing the specific edge condition under investigation. The process engineer must define how that comparison is made. Changing the corner radius, stock condition and edge preparation together would not isolate the edge contribution.
If the evidence points toward deformation concentrated within the wall, compare the observations through the relevant forming sequence and review the actual material state, tool contact and material-flow conditions. Producer forming guidance treats surface condition and lubrication as meaningful process inputs. Their relevance does not establish which one caused the failure in the present tool.
Use specimens that represent the intended route. A hand-formed corner with a different constraint cannot be treated as a complete reproduction of the production draw. It can still answer a narrower question if the difference and purpose are explicit. Keep diagnostic specimens separate from parts submitted as evidence that the final route is acceptable.
For each trial, decide in advance what observation would support or contradict the proposed explanation. Finding fewer visible cracks may be useful, but an apparent improvement based on selected favorable pieces is not the same as demonstrating the intended population under the relevant conditions.
Two crack histories require different next steps
In a hypothetical first case, an identified intermediate sample has a small separation connected to an opening that existed before drawing. Later samples show that separation extending toward the corner. The investigation should preserve the original opening edge and its forming history. Enlarging a remote corner radius without examining that path could miss the relevant input.
In a second hypothetical case, the earliest supported separation lies in a wall transition before the final rim is trimmed. Nearby intact material shows a concentrated change in section, and the crack reaches the rim only later. That evidence makes the wall-forming condition important and rules out final trimming as the origin of the earlier separation. It still does not identify a unique material or tool cause.
These examples are reasoning patterns, not descriptions of customer failures. They illustrate why the final crack location and the initiating location may differ, and why the available chronology should constrain the explanation.
Confirm the remedy without replacing one failure mode with another
A proposed change should address the supported failure path while preserving the required corner envelope and neighboring features. If a radius change is justified, assess its effect on the actual internal and external boundaries. If a process input changes, evaluate the resulting part through the relevant later operations instead of stopping at a favorable intermediate appearance.
Use the manufacturing review to present the origin map and controlled trial evidence. Resolve any proposed geometry change through the drawing review, and preserve the failure observations and specimen history in the quality record. The investigation is stronger when it explains the path that produced the separation and the evidence that the remedy addresses it, rather than merely showing a corner that no longer looks cracked in one selected view.