Find the first condition that changes a passing part into a failing assembly
A dimensional report can be correct while a shield still refuses to seat on its PCB. The report describes particular features, references and measurement conditions. The assembly introduces other parts, different locating surfaces and an installation sequence. The useful question is which of those differences explains the observed interference.
Do not begin by enlarging the cover, relaxing every tolerance or asking for a tool correction. First separate three possibilities: the supplied part does not meet an applicable requirement; the assembled combination differs from the approved combination; or the released requirements do not fully describe the interface that must work. Each possibility leads to a different corrective action.
Stop a fit trial when continuing would bend the shield or damage the board. A forced assembly may erase the evidence needed to understand the original problem. Keep an untouched sample and a failing combination available while the investigation proceeds.
Confirm the combination before comparing dimensions
Identify the actual shield, PCB and any mating lid, clip, connector or enclosure involved. Check their revisions against the assembly definition. A cover can conform to its own current drawing while being tried on an earlier board. Likewise, the board may have the correct outline but contain a substituted component whose package differs from the released model.
Compare the files used for inspection with those used for PCB layout and mechanical assembly. Pay particular attention to locally saved models, footprint libraries and drawings exported before a change. Updating a model in a design system does not necessarily update every footprint or assembly that references it.
This check should produce a specific answer: these identified parts are intended to assemble together under this released configuration. If that relationship is uncertain, resolve it before treating the fit result as a supplier dimensional failure. Preserve the uncertain combination as evidence; do not silently replace its identity with the revision someone expected to receive.
Connect assembly location with the inspection setup
A datum establishes a reference for evaluating geometry. The physical surfaces used to locate the part during inspection therefore matter. Ask how the inspection setup established orientation and position, then compare that with how the board or mating frame locates the shield.
For example, an outside width might pass while the inside wall at the component height interferes. Both observations can be true. The measured width may not control wall angle, a local formed feature or the position of the wall relative to the mounting feet. Repeating the outside measurement more accurately would not resolve that mismatch.
Mark the first apparent interference on a section view. Trace its position to the drawing references and to the board references separately. If those chains do not connect, the immediate issue may be an incomplete interface definition. Use the drawing and interface review to decide which functional relationship needs clearer control, rather than adding an arbitrary tighter number to the nearest dimension.
Keep free, supported and installed conditions separate
A thin formed part may respond to the way it is supported. Inspection fixturing can establish repeatable location, but excessive restraint can also change the geometry being observed. Ask whether the reported values describe the free part, a specified restrained condition or a part already attached to a board.
Then reproduce the relevant condition deliberately. Inspecting a cover while someone squeezes its sides is not equivalent to placing it freely on the mating frame. A frame checked before soldering is not automatically represented by the same frame after attachment. Neither observation should replace the other when the project requires both.
The drawing owner must determine the applicable conditions and acceptance criteria. Do not declare every thin shield a formally non-rigid part or add a free-state symbol without reviewing the governing drawing practice. The practical aim is simpler: make sure the supplier and assembler are evaluating the same physical state before debating the result.
Use the failure pattern to choose the next check
Observe where the part stops and what was already engaged at that moment. A uniform lateral offset suggests a different investigation from one corner lifting after the opposite side seats. These patterns generate hypotheses; they do not establish root cause on their own.
- The whole outline appears shifted. Check orientation, board origins, footprint placement and the features that locate the shield before remeasuring every wall.
- One local region blocks insertion. Compare that region with the actual component, connector, solder deposit or formed detail. Inspect the relevant section rather than only the top view.
- The part rocks or stops above its intended seat. Identify the first supporting contact. Check the board condition, mounting features and possible internal obstruction before assuming the roof is too low.
- Fit changes after a later operation. Compare the same identified assembly immediately before and after that step. Investigate what the operation changes, including support, attachment and adjacent hardware.
A photograph is useful for locating the symptom, especially with orientation identified. It rarely establishes the hidden contact or the force that produced it. Combine images with direct inspection and a controlled comparison that can distinguish competing explanations.
Check the model, then check what the model represents
A three-dimensional interference check can reveal a collision missed by a plan view. Its usefulness depends on the model geometry, placement and the objects included in the check. Confirm component heights, board side, shield orientation and the selected model reference surface. A simplified box may omit a protrusion that matters, while an incorrectly positioned model can invent an interference.
Also review the checking scope. Software rules may compare selected component bodies without checking every interaction with the PCB itself. A clean report is meaningful only for the objects and relationships actually tested.
Use the digital model to target physical checks, not to overrule an observed obstruction. If the model shows clearance but the hardware does not fit, compare the two at the interference location. The discrepancy may be a model omission, a changed part, a placement error or a physical departure from requirements. That comparison is more informative than repeatedly rerunning an unchanged collision check.
Two hypothetical investigations with different outcomes
Consider a hypothetical cover that stops against one connector while its length, width and height pass inspection. An older assembly model shows the connector below the wall opening. The actual board uses a different connector variant with a raised housing feature. The next useful test is a verified component-envelope comparison. Changing the cover tool before resolving the approved connector configuration could solve the wrong problem.
In a second hypothetical case, a loose frame accepts its lid, but the lid will not engage after the frame is mounted. The investigation compares the same frame before and after attachment, using the intended support and orientation. If the mounted geometry changes, the team examines the attachment process and board condition. If it does not, the next check may be the installation path or an internal component contacting the roof.
Neither example proves a universal cause. Their value is the order of evidence: isolate the changed condition, measure the relevant interface, and choose an action that addresses the demonstrated difference.
Change only what the evidence supports
Where a part is outside an applicable requirement, the response may involve correction of the part or process. Where the wrong revisions were combined, restore the intended configuration and assess whether the attempted assembly damaged either item. Where compliant parts can still interfere, the design team needs to review the combined tolerance and interface definition.
That last case deserves particular care. A single nominal fit does not demonstrate compatibility across permitted variation. Review the dimensions that act together and the extremes relevant to the interference. Tightening one feature can move cost or inspection effort without resolving an uncontrolled relationship elsewhere.
Retest the proposed correction with representative mating parts and the intended assembly sequence. Keep the original symptom and the changed variable visible in the comparison. The quality and inspection scope helps separate dimensional evidence, assembly confirmation and the system tests that remain. A shield that now seats correctly has passed a mechanical milestone; its electrical performance still requires the project's own validation.