A reworked first article may prove that the corrected sample conforms while failing to represent the planned production route. The IAQG 9102 page describes an aerospace first-article standard and applies only when the governing contract or requirement invokes it.
A reworked first article can meet product acceptance criteria without representing the normal PCBA production route. If technicians add operations, replace components or adjust the sample until it passes, the final result describes that sample’s actual history. It does not automatically show that the intended repeat process will produce the same result.
For an OEM, two questions therefore need separate answers: is this physical article acceptable, and what does its evidence establish about routine production? Treating them as one decision can hide a process problem behind a successful sample.
This guide explains how to review a reworked first article without discarding useful evidence or overstating it. The examples are hypothetical engineering scenarios, not GNSEMS customer results. Actual rework methods, product acceptance and first article requirements must follow the project controls and relevant technical authority.
Separate Product Conformity from Route Representativeness
Product conformity concerns whether the article meets the applicable requirements after its actual processing. Route representativeness concerns whether that article and its evidence support the production route that the team intends to repeat. A positive answer to one question does not supply the other.
Suppose a first article has a placement problem that a technician corrects manually. Suitable inspection may establish the corrected placement on that board. However, the corrected board alone cannot establish that the unchanged placement program will place the next board correctly.
Similarly, replacing a suspect component may produce a functioning sample while leaving material selection or process handling unresolved. The replacement can be part of an appropriate disposition, but the normal route still needs evidence addressing the cause of the original concern.
The existing SMT first article inspection guide describes the wider review sequence. This article concentrates on the evidence boundary when the reviewed sample no longer follows that sequence exactly as planned.
Preserve the sample’s first observed condition
A passing result adds to the history; it does not erase the initial failure. Record the original observation before intervention. Include the unit identity, configuration, operation, inspection or test result and relevant conditions. Photographs can help with visible features, while electronic results may be necessary for programming or functional concerns.
A repeat pass should add to the history rather than replace the initial failure. Otherwise, a reviewer sees a clean result without knowing that the sample required extra work. Preserve the relationship between the observation, intervention and subsequent verification.
Use neutral wording while the cause remains uncertain. A failed test can arise from the product, the setup or the method. Record what occurred and what changed instead of assigning a cause simply because one intervention was followed by a pass.
The solder-joint defect escalation guide provides context for preserving evidence during an investigation. Apply the same principle to the first article record so the final approval does not lose the event that required review.
Record Every Intervention on the Reworked First Article
Record the extra operations when assessing what the sample can prove. List the operations performed beyond the intended normal route. Depending on the case, these may include component removal, replacement, local heating, cleaning, programming, mechanical adjustment or repeated inspection. This list is a record of actual work, not a recommendation to perform those operations.
Identify the instruction and authorization under which each intervention occurred. Record the affected positions or features, the relevant materials and equipment, and the person or role completing the work. The required detail depends on the product and process controls.
Also capture additional exposure where it matters to the technical review. Extra heat, handling or processing can affect the interpretation of later evidence. This article does not set universal rework-cycle limits or operating parameters; those require the applicable product, material and approved process requirements.
Keep the original production configuration distinct from the rework configuration. If an engineer changes a program or fixture during the intervention, identify which results belong to each condition. A single unqualified “passed” entry is insufficient when the conditions changed.
Compare the planned route with the sample’s actual route
Compare the reworked first article route with the released production route, including material, sequence, equipment, manual intervention, inspection, and test.
Separate evidence for rework acceptance from evidence for normal production. Create a concise comparison before deciding what the article represents. The following table illustrates the questions using generic operations. It is not a completed inspection record or a prescribed PCBA process.
Comparing a reworked first article with the intended route
Route element
Question about the actual sample
Effect on the evidence review
Material selection
Did the sample receive a replacement or special selection?
Check whether normal material control is represented
Assembly operation
Was a defect corrected through an extra manual step?
Separate corrected product evidence from normal-process evidence
Programming
Did firmware or configuration change after a failed result?
Associate each result with the correct software baseline
Inspection
Were settings or acceptance criteria changed during review?
Establish the validity and applicability of the final method
Handling and exposure
Did the article receive additional relevant processing?
Assess any effect on acceptance and representativeness
Final status
What exact decision does the passing result support?
Avoid converting sample acceptance into unrestricted route approval
The comparison should explain the differences that matter. Do not fill the record with irrelevant detail while omitting the intervention that changed the outcome. A reviewer should be able to identify where the actual sample departed from the intended repeat process.
Decide Which Results from a Reworked First Article Remain Useful
Rework does not automatically invalidate every result associated with the sample. Some earlier measurements may remain applicable if the intervention did not affect the characteristic or the conditions supporting that measurement. Other results may need repetition because the intervention changed their basis.
The technical review should establish that boundary. For each affected characteristic, identify whether the earlier evidence remains valid, needs new verification or cannot support the proposed conclusion. Keep the rationale linked to the intervention rather than using a blanket statement that all prior evidence is either good or bad.
For example, replacing a component can require renewed evidence for related identity, assembly and functional requirements. Whether it affects a separate measurement elsewhere on the board depends on the actual process and relevant characteristics. Do not infer independence solely from physical distance.
The PCBA validation checklist can help organize the applicable requirements. First article evidence should remain specific about what was verified and under which conditions; it should not claim broader reliability or process capability without supporting evidence.
Use corrected normal-route evidence when the question requires it
If the first article exposes a problem in the intended production route, correct that route through the applicable controls. Then determine what evidence is needed to demonstrate the corrected condition. A new article processed through the corrected normal route may be appropriate, depending on the requirements and the issue.
The important distinction is the verification objective. Rechecking the reworked sample can establish its final condition. Processing a representative article through the corrected route can address whether that route produces the required outcome without the same special intervention.
A new sample is not automatically sufficient either. It must use the intended baseline and relevant production conditions. If an expert again performs unplanned adjustments or specially selects material, the new sample may repeat the representativeness problem in a different form.
Connect any route or configuration correction to PCBA engineering change control . Record the effective point and treatment of old work in process. A successful new article does not automatically settle the disposition of earlier units.
Separate a normal operation from an exceptional intervention
A reworked first article can retain evidence only where the intervention did not alter the characteristic, process question, or acceptance condition under review.
Not every manual operation is rework. Many legitimate production routes include controlled manual tasks. The review question is whether the sample followed the approved route and whether the evidence represents the route intended for routine use.
An extra intervention introduced because the sample failed deserves its own history even if the same tool is used elsewhere in normal production. Conversely, an approved planned operation should not be described as an exception merely because it is manual. Use the actual instructions and processing history to distinguish them.
If the team proposes adding a formerly exceptional step to the normal route, treat that as a route change. Review its technical basis, execution controls and verification needs. Renaming the intervention as a standard step does not demonstrate that it is suitable or consistently executable.
Also consider whether the planned first article process intentionally includes a defined adjustment or calibration. Where it does, record that controlled activity and assess representativeness against the intended route. The presence of an adjustment alone is not enough to determine whether the article is representative.
Review repeated tests without losing the original evidence
A sample that passes after several attempts needs an understandable result history. Record the sequence and any changed conditions. If the test method itself was corrected, distinguish the invalid method’s results from results obtained under the approved method, using the responsible technical review.
If nothing relevant changed, an intermittent failure remains a question to investigate. Do not present the final pass as proof that the earlier failure was unimportant. The project should decide what additional evidence is needed to interpret the observation.
Link further investigation to the PCBA failure analysis process . Keep unit identity and configuration consistent across the investigation so results do not become detached from the actual article.
The first article approval should state any limits that remain. For instance, product acceptance after a reviewed intervention may coexist with an open question about the routine process. A concise decision record can preserve that distinction without requiring operators to interpret every test log.
Use an outcome matrix instead of one blanket pass
The scenarios below illustrate why two approval questions are useful. They are not universal release rules. Apply the product requirements, technical findings and designated approval authority to the actual case.
What different first article outcomes can support
Observed outcome
Product question
Normal-route question
Reworked sample meets reviewed criteria
Acceptance may be supportable within the approved disposition
Extra intervention may leave routine production unrepresented
Corrected route produces a representative conforming article
Review acceptance against applicable requirements
Review whether the evidence covers the intended route and scope
Final pass follows unexplained repeated failures
The result history needs technical interpretation
The unresolved behavior may limit route approval
Sample uses specially selected replacement material
Assess the actual material and resulting product
Normal material selection may need separate evidence
Earlier unaffected measurements remain applicable
Retain them with a justified scope
Do not extend that retention to changed process conditions
A single status label can still be used in an operational system if the underlying permission is clear. The risk arises when “first article passed” is treated as unrestricted approval despite a record that supports only a narrower conclusion.
Work through a hypothetical placement correction
Imagine a first article with a connector in the wrong orientation. A technician corrects the connector under an approved instruction, and suitable inspection confirms the final condition. This hypothetical example does not prescribe how to rework a connector or establish that rework is acceptable for a particular product.
The corrected sample can provide evidence about its final orientation and other verified requirements. However, if the placement data remain unchanged, the next board may repeat the original problem. The sample’s pass therefore cannot stand alone as evidence that the intended route is ready.
The team reviews the cause and corrects the relevant controlled input. It then defines the evidence needed from the corrected normal route. The reworked sample retains its own history and disposition, while the new route evidence supports a separate production decision.
Suppose the original connector also served as the test interface. The intervention may affect how earlier functional evidence should be interpreted. The reviewer considers that connection rather than assuming that a visual correction leaves every earlier result unchanged.
Keep sample identity stable through the investigation
Track the physical article through removal, replacement, testing and storage. If a board changes its configuration, update the relevant record without losing its earlier identity and history. A new label should not make a reworked article appear to be a different, untouched sample. Likewise, a replacement article needs its own identity so evidence from two boards does not become merged into one apparently continuous sequence.
Record retained parts or other investigation objects when their identity matters. A removed component may provide information about the original concern, while the replacement component supports the final configuration. Confusing the two can undermine both root-cause analysis and the first article review. Use the project’s approved evidence-handling controls rather than relying on an informal photograph with no case reference.
Explain the limits of a successful demonstration
A first article demonstration can cover only the conditions it represents. If the sample ran on one station, used one fixture and received one firmware target, the reviewer should not automatically extend the result to unverified alternatives. Rework adds another condition to that scope. State which aspects of the evidence remain common and which depend on the sample’s special history.
For the final handoff, include an explicit list of unresolved route questions rather than embedding them in a long narrative. Assign the next verification or decision to an owner and identify the affected work. This keeps a useful product acceptance decision from being misread as a claim that every manufacturing question has been settled. The resulting record can support progress while preserving the exact limits of the available evidence.
What Should the Final Reworked First Article Review Conclude?
Product acceptance and production-route representativeness are separate judgments. The review package should state the article’s final product status and the scope of route evidence separately. Include the original observation, intervention history, retained and repeated results, corrected baseline where relevant, and remaining questions.
Make the permitted next action clear. Acceptance of one reworked sample may authorize its defined disposition without authorizing the rest of the order. Approval of a corrected route may apply to future work while earlier suspect units remain under separate control.
For a project conversation, prepare the sample history, approved intervention, original and final results, and the intended repeat route. Discuss first article evidence with GNSEMS using those records. The objective is a defensible explanation of what the sample proves and what still needs verification.
Questions about a reworked first article
Does rework automatically make a first article unusable?
No. It changes the evidence review. Some results may remain useful and the final article may be acceptable under the applicable disposition. The team must still assess which characteristics and production conditions the article represents.
Can a reworked sample authorize normal production?
Only if the applicable requirements and reviewed evidence support that conclusion. Extra intervention can leave the intended route unrepresented. Separate the sample’s acceptance from evidence that the corrected normal process is suitable to repeat.
Should the original failed result stay in the record?
Yes. Preserve it with its conditions and subsequent technical interpretation. A later pass adds evidence but does not erase the earlier observation or the intervention that changed the sample.
Is a manually assembled first article necessarily unrepresentative?
No. Manual work may be part of the approved normal route. Compare the actual sample history with the intended production process. The relevant issue is a meaningful difference in conditions, not whether an operation uses manual labor.
