What Validation Should Buyers Require for an Expedited Replacement Lot After a Dispensing Supplier Shipment Is Rejected?
Direct answer: Treat an expedited replacement lot as a recovery lot, not as routine replenishment. Before release, verify that the supplier bounded the original failure, implemented a relevant correction, disclosed every process change made during the rush build, and produced traceable evidence for the failed characteristic. Then apply tightened receiving inspection, application-specific functional or destructive tests, and a controlled line trial. Release only a limited quantity first; normal volume follows after stable recovery lots.
Who this is for: supplier quality engineers, process engineers, purchasing managers and factory leaders recovering from a rejected dispensing, potting or meter-mix shipment. Buyer readiness: L5 deployment and restart.
Why an expedited remake carries different risk
A normal lot is produced inside an established schedule. A recovery lot is often built with overtime, reserved material, a different shift, recently serviced equipment or an improvised inspection window. Those changes may be necessary, but each can introduce variation. The supplier may also focus so heavily on the original defect that a new defect escapes elsewhere.
The goal is not to repeat every qualification test automatically. It is to prove three things: the original mechanism is controlled, rush-build changes did not invalidate approval, and the lot works under the buyer’s real application conditions.
| Validation layer | Main question | Evidence | Weak substitute |
|---|---|---|---|
| Failure containment | Was the rejected population correctly bounded? | Lot, machine, shift, material and time-window genealogy | “Only one sample failed” |
| Correction relevance | Does the action address the observed mechanism? | Before/after process evidence | Operator retraining with no cause |
| Change review | Did the rush build alter the approved process? | Documented comparison and authorization | Supplier says recipe is “basically the same” |
| Lot verification | Does this specific lot conform? | Raw measurements and release records | Generic certificate |
| Application proof | Will it run and cure correctly at the buyer? | Controlled trial and downstream results | Visual appearance only |
Gate 1: reconstruct the original failure before accepting the fix
Start with the failed requirement, actual measurement, test method and sample identity. Confirm the measurement system was valid. If the original result cannot be reproduced or explained, the supplier may be correcting the wrong mechanism. “No defect found” is not proof that the replacement lot is safe.
- Identify all potentially affected supplier and buyer inventory.
- Map resin, hardener, valve, pump, mixer, software, fixture and operator genealogy as applicable.
- Separate occurrence cause from escape cause: why it happened and why final inspection missed it.
- Record what remains unknown and add controls for that uncertainty.
- Determine whether adjacent lots share the same inputs or production window.
For a visible label defect, bounding may be straightforward. For intermittent ratio drift or internal potting voids, the affected window may extend back to the last verified good process check.
Gate 2: review every rush-build change
| Possible change | Risk introduced | Buyer evidence |
|---|---|---|
| Different material lot | Viscosity, filler settling, cure or adhesion variation | COA, shelf life, conditioning and comparison to approved range |
| Alternate pump, valve or mixer | Ratio, flow, pressure and dead-volume change | Equipment identity, calibration and output study |
| Different shift/operator | Setup, purge, cleaning or inspection inconsistency | Qualification and witnessed first-piece record |
| Overtime or shortened cure | Handling before sufficient cure | Time/temperature record and cured-property check |
| Expedited transport | Temperature excursion or packaging damage | Packaging, logger and receiving-condition evidence |
| Temporary software/recipe edit | Path, volume or alarm-limit drift | Revision, checksum, approval and rollback control |
If corrective action changes a validated input or process window, decide explicitly whether partial or full revalidation is needed. A replacement lot is not the place for an undocumented experiment.
Gate 3: target inspection at the original defect and its neighbors
Use tightened inspection after a rejection, but do not simply increase every sample count. Put most effort into characteristics linked to the failure mechanism and potential side effects of the correction. ISO 2859-1:2026 describes AQL-indexed lot-by-lot attribute sampling and switching toward tightened inspection when quality deteriorates. The selected plan still needs to match the contract, defect class and lot context.
| Original failure | Recovery-lot check | Neighboring risk | Release trigger |
|---|---|---|---|
| Mix-ratio drift | Independent A:B output by weight across start, middle and end | Pressure, flow and cure variation | All results inside approved window; stable trend |
| Bubbles/voids | Witness samples, cross-section or validated NDT | Fill level, vacuum cycle and cure | Defined void criteria met |
| Bead position/size | First-piece image and measured geometry | Overflow, stringing and cycle time | Capability or agreed lot evidence accepted |
| Poor adhesion | Representative substrate coupon and approved test | Surface preparation and open time | Strength/failure mode meets specification |
| Incomplete cure | Mixed coupon under recorded time/temperature | Ratio, contamination and material age | Approved cure endpoint achieved |
| Traceability gap | 100% identity and record reconciliation | Unapproved source or revision | No unresolved genealogy gaps |
Accept-zero approaches can be useful when contractually defined, but ISO 28594 notes limitations for destructive testing or where screening is impractical. Destructive tests need a justified specimen plan; they cannot be copied blindly from an attribute-sampling table.
Gate 4: test the application, not just the delivered item
Incoming checks may show that a metering component is dimensionally correct while missing its effect on dispense stability. Run a controlled line trial using the actual adhesive, substrate, fixture, recipe and environment. Start with a limited quantity that can remain traceable through downstream checks.
| Application | Minimum recovery trial | Critical evidence | Do not rely on |
|---|---|---|---|
| EV battery potting | Representative fill under normal volume and vacuum conditions | Ratio, void, cure and thermal-path evidence as specified | Surface appearance |
| PCB dispensing | First article plus representative board/run | Deposit geometry, keep-out compliance and functional result | Robot position accuracy alone |
| LED driver potting | Controlled enclosure fill and cure | Fill level, connector cleanliness, cure and electrical checks | One hand-mixed coupon |
| Automotive sensor sealing | Representative bead and assembled seal | Continuity, adhesion and leak/seal result | Bead photograph only |
| Power electronics TIM | Representative dispense and assembly compression | Mass, bondline/coverage and thermal requirement | Nominal programmed volume |
Gate 5: use staged release instead of all-or-nothing approval
- Quarantine received: records and identity are checked; no production use.
- Approved for engineering trial: a named quantity enters a controlled test order.
- Limited production release: trial passes; tightened checks and traceability remain.
- Lot release: required lot evidence passes and authorized personnel approve use.
- Normal supplier monitoring: only after consecutive recovery lots and CAPA effectiveness meet exit rules.
For many nonregulated industrial projects, three consecutive accepted lots across representative production conditions can be a useful starting point before reducing controls. It is not a universal rule. Customer-specific and regulated requirements take precedence.
What the supplier’s expedited release package should contain
| Record | Required detail | Why it matters |
|---|---|---|
| Lot genealogy | Material, equipment, shift, operator, time and quantity | Bounds later containment |
| Change comparison | Approved condition versus recovery-build condition | Reveals hidden rush changes |
| First-piece result | Actual values, method, date and authorization | Shows the corrected process started correctly |
| In-process trend | Values across beginning, middle and end | Detects drift masked by one result |
| Final inspection | Raw data by defect class and sample identity | Supports lot-specific acceptance |
| CAPA status | Containment, root cause, action and effectiveness plan | Separates emergency correction from permanent prevention |
ISO guidance on documented information identifies records for change review, authorized release, traceability and nonconformity disposition. Preserve the authorizing person and acceptance criteria, not merely a date stamp.
Reject or repeat the recovery lot when these signals appear
- The original failure reappears in any critical characteristic.
- The cause remains unknown and the defect cannot be reliably screened.
- The supplier made an unapproved material, equipment or recipe change.
- Raw results conflict with the certificate or omit failed samples.
- The process needs repeated manual adjustment to remain in tolerance.
- Traceability cannot separate trial output from normal production.
- Destructive results fail even when visual checks pass.
- The supplier asks to ship before authorized release is complete.
A second failed recovery lot is not just another lot rejection. It should trigger escalation of supplier status, fallback allocation and possible requalification or source-removal decisions.
Buyer checklist before authorizing the shipment
- Original defect and affected window are documented.
- Measurement validity is confirmed.
- Correction addresses both occurrence and escape causes.
- Every rush-build change has been disclosed and reviewed.
- Sampling and destructive-test plans are approved before results arrive.
- Application trial quantity, owner and stop criteria are defined.
- Customer approval requirements are checked.
- Recovery units remain identifiable downstream.
- Next-lot tightened controls and exit rules are written.
- Supplier CAPA continues after the emergency shipment.
Evidence sources
The ISO 2859-1:2026 overview describes lot-by-lot attribute sampling and switching rules. ISO 28594:2017 covers combined accept-zero sampling and process-control procedures while noting limitations for destructive tests. ISO’s documented information guidance identifies change-review, traceability and authorized-release records. Use the purchased standards, contract and customer-specific requirements for the actual plan.
Frequently asked questions
Can a supplier certificate alone release the replacement lot?
No. It supports release, but buyers should independently verify selected critical outputs and the original failure mechanism.
Must the buyer repeat full qualification?
Not always. Revalidate the scope affected by the failure, correction and rush-build changes. Major changes may justify broader revalidation.
Should every unit receive 100% inspection?
Identity or screenable critical checks may need it, but 100% visual inspection cannot detect hidden cure, adhesion or void defects.
Can production start before CAPA is fully closed?
A controlled recovery lot may run after credible containment and correction, provided residual risk is approved and CAPA deadlines remain active.
How many clean lots are enough?
Three consecutive accepted lots can be a starting point for some industrial projects, but risk, customer and regulatory requirements control the actual number.
What should buyers send OBO Precision?
Send the failed result, drawings, material data, machine recipe, corrective change, replacement-lot records and required production rate.
Validate the recovery lot before the line depends on it
OBO Precision can review the original failure, material, equipment changes and production target, then recommend a controlled dispensing trial and release evidence.
Finished-goods follow-up: After a recovery lot enters production, use a controlled finished-goods release plan before authorizing customer shipment.
Post-shipment follow-up: After releasing recovery-lot finished goods, use a defined post-shipment monitoring and customer containment plan to detect recurrence early.
Complaint-response follow-up: If the first customer issue appears, use a structured recovery-lot complaint and containment process before calling it isolated.
Customer containment: When affected stock remains at the customer, use a validated customer stock screening plan rather than relying on an unproven visual sort.
