Do not assume a reported missing SMT component is a defect until the board configuration and the indication are verified. Match the physical board and reference designator to the released assembly variant, check whether the position is DNP/DNI, and have a qualified person review the image or inspect the board.
If the component was required and is absent, preserve the board and available records. Build a timeline for that same unit and position: what is the last stage at which a component was visible at the position, and what is the first stage at which its absence was confirmed? Visibility alone does not prove its identity or continued presence. Missing records leave a gap; they do not establish a cause.
1. Identify the board and the intended configuration
Start with one board whenever possible. Record the assembly number, board revision, serial or other unit identifier, panel position if relevant, side of the board, and reported reference designator. A photograph should show enough context to locate the position without relying solely on a handwritten label.
Then reconcile that identity against the released documents for the build:
- The applicable BOM and assembly drawing or other approved configuration information.
- The selected product variant and any approved changes effective for the unit.
- The placement data and program revision, as supporting records rather than automatic proof of design intent.
- DNP, DNI or equivalent instructions for the specific position and variant.
A position can be deliberately unpopulated on one variant and required on another. A component that appears in a generic BOM is not necessarily required on every board. Conversely, a missing placement-file entry does not by itself authorize leaving a required part off. Resolve conflicting documents through the responsible engineering and configuration-approval process before deciding what the board should contain.
For related file checks, see the pick-and-place file checklist and what a BOM contains.
2. Review the indication before classifying the board
An automated optical inspection (AOI) alarm, a customer photograph or an operator observation is a reported indication, not yet a confirmed defect. A qualified reviewer should check the relevant image against the correct board side, position, orientation and inspection program. If the image is unclear, arrange a controlled physical examination and document what is visible before disturbing the site. For the roles and limitations of visual and automated methods, see PCB inspection.
A component may be obscured, misidentified or outside the useful image area. An image might also belong to a different unit or panel location. If identity or visibility cannot be established, record the result as inconclusive, not present or absent.
If the position is confirmed DNP/DNI for the released variant and there is no conflicting requirement, document the configuration review. If it was required and its absence is confirmed, hold the affected unit while engineering and quality determine the investigation and disposition. Consider whether related units need containment based on traceability and risk; do not assume the reported unit represents the whole build.
3. Build a same-unit, same-position evidence ledger
A useful ledger separates an observation from what someone inferred from it. For every entry, capture the unit identifier, reference designator, stage, record source, timestamp or production sequence, reviewer, and any limits on visibility or traceability. Check that clocks and sequence references are consistent before treating timestamps as exact.
The key stages are:
- Configuration before production. Was the part required in the released variant? This establishes intent, not physical presence.
- Paste inspection. A solder paste inspection (SPI) result can document paste before placement. It does not show that a component was subsequently placed or retained.
- Placement. A recorded placement event may show what the equipment recorded for the programmed operation. It does not prove the part remained on the board afterward; an alarm or lack of one is not a complete physical history.
- Post-placement inspection. A usable image or observation of the correct unit and position may establish visible presence before reflow. If no such record is available, that state is unknown.
- Post-reflow and later inspection. A human-reviewed image or physical examination may establish visible absence. Note any intervening handling, inspection or transport that the records identify.
- Authorized corrective work and retest. Record what was authorized, what was performed, by whom and when, the part and material traceability where applicable, and the subsequent inspection and test results.
The last known visible presence at the position and first known absent entries define an evidence interval, not necessarily the exact moment of loss. An image alone may not identify the visible component as the specified part or establish that it remained there afterward. If the first trustworthy observation is an absent position and no earlier presence record exists, the last known visible presence is unknown. If records do not resolve the interval, the stage or root cause remains unknown.
An image showing a component at a location is evidence of visible presence. It does not automatically establish correct part identity, solder quality or electrical function. Similarly, a record that no rework was found is not proof that no intervention occurred unless the completeness and scope of that record system are established.
4. Fictional example: one position across eight related boards
This hypothetical example illustrates the ledger; its numbers are not inspection rates or acceptance thresholds.
A customer reports that U7 appears missing from board S04. The related build contains eight identified boards, S01 through S08. The released variant review says U7 is required on all eight. The report concerns S04 only; the other seven have not yet been assessed for U7.
Simple accounting:
- Related boards: 8
- Board with U7 confirmed absent at the reported review: 1
- Related boards not yet assessed at U7: 7
- Check: 1 + 7 = 8
The counts describe the investigation status, not a defect rate for the build.
| Stage for S04, U7 | Available evidence | What it supports |
|---|---|---|
| Released configuration | Approved variant documents reviewed | U7 was required; this is not evidence of physical presence |
| Before placement | SPI record shows paste at the U7 site | Paste was observed before placement, not the component |
| Placement | Unit-linked record reports a placed event for U7 | The equipment recorded the event; later retention is unproven |
| After placement, before reflow | Image linked to S04 and U7; a reviewer can see a component at the site | Last known visibly present |
| After reflow | Image linked to S04 and U7; a reviewer confirms the site is empty | First known visibly absent |
| Intervention records | No rework record before the post-reflow image is found in the available system; coverage of other interventions and system completeness are not established | Earlier intervention cannot be ruled out solely from this search |
| Later authorized corrective work | An approved instruction and signed record state U7 was installed; a later image shows a component at the site | Visible presence after the work, not proof of identity or the original cause |
| Retest and disposition | Required retest results are not yet available; quality disposition is pending | The board is not ready for release on this evidence alone |
In this sequence, the visible change occurred between the two usable images. That interval includes reflow and any other handling that occurred between them. It does not prove that reflow caused the absence, nor does the placement event prove that U7 remained secure until reflow. The later corrective work does not erase the earlier finding or close the investigation.
If the pre-reflow image could not be tied reliably to S04, the last known present state would instead be unknown. If the post-reflow image were obscured, absence would remain unconfirmed until another valid observation became available.
5. Compare hypotheses without turning a clue into a verdict
Once the evidence interval is defined, engineering can develop hypotheses consistent with it. Depending on the records, examples include:
- The part was not successfully placed, despite a recorded operation.
- It was placed but moved or detached during a later operation or handling.
- The apparent absence is an inspection or identification error.
- An undocumented or poorly linked intervention occurred.
- A configuration or programming discrepancy affected the position.
Not every hypothesis fits every ledger. In this example, a reliable pre-reflow image of the U7 position on S04 narrows the question: a simple claim that no component was ever visible at the site would conflict with that image. It still leaves several possible explanations for what happened afterward.
Use comparisons that preserve unit and position traceability. For example, compare the affected position with the same position on related units whose configurations match, or compare available records from before and after a documented change. Look for differences in timing, material, program, handling and inspection evidence—not merely a pattern that appears convenient.
Where a controlled trial is justified, engineering should define the question, approve the setup, change one factor where practical, and record both the change and the observations. Repeated, reversible comparisons may strengthen or weaken a hypothesis. A single feeder swap followed by a good board does not establish root cause: material, timing, handling or other conditions may also have changed. A machine alarm is likewise a lead to investigate, not a complete explanation.
Feeder and cover-tape checks can be relevant: verify the documented part and setup, and look for evidence of feeding or tape presentation problems. Keep those checks linked to the affected unit and time rather than treating them as a substitute for the board ledger.
6. Treat equipment records as configuration-dependent evidence
Ask what records actually exist for the line, program and unit, and what identifiers and time range they cover. An unavailable image or log should be marked unavailable; it should not be reconstructed as fact.
Some equipment configurations can record pickup and placement images or link production events to board identifiers. Availability, enabled options, coverage and retention depend on the actual system and its settings. Confirm those details for the specific build before relying on a record.
7. Assign responsibility through disposition
The customer provides the reported unit identity, symptoms, photographs and applicable requirements, and participates in decisions requiring customer authorization. The customer should not have to decide root cause from an image alone.
Engineering reconciles configuration, evaluates the evidence interval and hypotheses, and defines approved corrective work or further investigation when appropriate. Engineering also identifies suitable inspection and functional retest requirements for the specific product.
Production preserves the unit and records, carries out containment and corrective work only under the authorized instruction, and records the work performed. An operator should not install a part simply because an empty footprint looks suspicious.
Quality reviews the evidence, traceability, nonconformance handling, completed corrective work and required retest results. It coordinates disposition under the applicable agreements and approval process. Depending on the finding, the outcome might be confirmation of an intended DNP position, approved corrective work, rejection or scrap, or another explicitly authorized disposition. Classify the work as rework, repair or another disposition under the applicable requirements. A use-as-is decision requires the relevant approval; it is not implied by an inconclusive investigation.
After corrective work, inspect the affected position and complete the applicable retests under the approved plan. Record results against the same board identifier and designator. The PCB assembly inspection report checklist covers the associated inspection, exception and release records. Release only after the required reviews and disposition are complete. If the cause cannot be established, say so in the record; an unknown cause and a completed, approved board disposition are separate facts.
To discuss an assembly question or submit build information, request a quote.
FAQ
Does a missing-component alarm prove a board is defective?
No. Confirm the board and position, review the released variant and DNP/DNI instructions, and have a qualified person assess the indication. An unclear or incorrectly linked image remains inconclusive.
Can SPI prove that a component was installed?
No. SPI documents paste before placement. A later component-presence claim needs evidence from a stage at which the component could actually be observed.
Does a recorded placement event prove the part stayed on the board?
No. It is evidence of what the equipment recorded for the operation. Later inspection or other reliable evidence is needed to establish visible presence or absence after that point.
What if there is no image between placement and the reported absence?
Mark the relevant state as unknown and use the other available, traceable records to bound the investigation. Do not assign an exact loss stage or root cause merely to close the ledger.
Who can authorize rework and release?
Authority depends on the product requirements and agreements. Engineering defines the technical instruction, production performs authorized work, and quality reviews the required evidence and disposition; customer approval is involved where required.
