The mold returns from maintenance, production restarts, and an edge now has flash. “It started after maintenance” is a useful lead. It is not yet a finding about what the maintenance work caused.
The first task is to reconstruct the transition: the last documented condition, what work was actually done, how the tool was returned to service, and which samples first showed the change. That timeline lets tooling, process, and quality staff investigate the same event instead of working from three different accounts.
What did “good before maintenance” mean?
Find the last accepted inspection record and any retained reference samples. Check whether the same edge, cavity, and inspection criterion were examined. A note saying “parts okay” may refer to dimensions while saying nothing about a fine edge feature.
Locate the flash on an approved drawing or permitted photograph. Mark the cavity, part orientation, extent, and sample ID. Use the job's definition and measurement method. “More flash” is hard to compare when one person looks at a trimmed part and another examines it before trimming.
Autodesk's definition of flash identifies excess polymer escaping where the mold separates. The definition names the observed defect; it does not identify which component or process condition allowed it.
Do not write “no flash before maintenance” merely because no one logged it. If the old inspection does not cover the present question, state that limitation. The best available baseline may be incomplete, and it should remain visibly incomplete.
“Maintenance completed” is a status, not a change list
Ask for the work-order reference, work scope, actual findings, changed components, and return-to-service record. Distinguish planned work from completed work. If a task was deferred or a component could not be checked, that belongs in the history.
Keep observations neutral. “Insert I-2 was replaced under work order W-42” is a traceable event. “The new insert is wrong” is a hypothesis unless inspection has established the discrepancy. Component drawings and approved specifications belong with the tooling owner; an operator's observation record should point to them.
BASF's troubleshooting guide, printed page 28, includes mold fit, sealing-surface damage, deformation under pressure, and cavity pressure among flash-related considerations. This supports keeping both tooling and process evidence in view. It does not justify assigning the defect to the most recently replaced component.
The American Society for Quality's discussion of change analysis describes investigating changes associated with a shift in performance. For this task, the timeline is a way to organize candidate explanations. A chronological association still needs evidence before it becomes a cause.
Record the restart as well as the maintenance
A restart may involve a recipe load, a different material container, a reconnecting step, a waiting period, or another approved setup action. List what the records show, even when it happened outside the maintenance team's work. Otherwise “after maintenance” can become a catch-all description for several changes.
Retain the first process record available after restart and the record associated with the affected samples. A later screenshot may reflect intervening adjustments. Identify those adjustments by time and owner; do not let the investigation silently compare the old process with a third, later condition.
The first detected defect is not necessarily the first defective shot
Constructed teaching example. The times, IDs, and counts below are invented. The table is a record-review exercise, not a containment boundary or a recommended inspection schedule.
| Time / event | Available record | Interpretation |
|---|---|---|
| 08:40 / before stop | Inspection Q-18 records no flash at E-3 on the retained samples from all four cavities. | A documented comparison exists for those samples. |
| 09:00 / maintenance begins | Work order W-42 lists cleaning and replacement of insert I-2. | Confirm completed work and return-to-service evidence. |
| 10:15 / restart | Recipe revision D is loaded. Pre-stop reference was revision C; the change reason is not attached. | A second change needs reconciliation. |
| 10:20 / first retained restart samples | Dimensional checks recorded; E-3 flash assessment not recorded. | These records do not establish the E-3 condition. |
| 10:45 / defect recorded | Cavity 2 has flash at E-3 on 6 of 10 inspected parts. Other cavities: 0 of 10 each. | A cavity-specific pattern is observed in this sample window. |
The insert replacement and recipe revision are both relevant. The table does not rank them as causes. It also does not prove that the first flashed part occurred at 10:45. That is the time the defect was documented; the earlier interval has a gap.
A useful handover note reads: “Flash recorded at E-3 in cavity 2 after restart. Pre-stop samples and W-42 retained. Recipe C-to-D change needs explanation; E-3 assessment missing for the first restart samples. Quality owns product disposition; tooling and process owners are reviewing their records.”
The product boundary must be determined under the actual quality procedure. The example cannot establish that all parts before 10:45 were acceptable or that every part after maintenance is defective. Keeping the gap explicit helps the responsible owner make that decision.
Use the pattern to ask a narrower question
| Observation | Useful next request | Conclusion to avoid |
|---|---|---|
| One cavity, one repeatable edge | Compare the marked location with the actual work scope and tooling inspection findings. | “The replacement insert must be the cause.” |
| Several cavities changed | Review shared process and restart records alongside the tool's findings. | “Several cavities means the tool is fine.” |
| Only later samples show the defect in the records | Locate intermediate samples and establish what was inspected when. | “The tool was good until the first complaint.” |
| An adjustment reduced visible flash | Review the exact change and all required part checks; retain the original tooling question. | “A process response proves there was no tooling issue.” |
A cavity pattern helps focus the review without making it conclusive. If the tooling owner finds a documented assembly or condition discrepancy, preserve the inspection result and the authorized correction. If no discrepancy is found, record what was actually checked, not “tooling ruled out” without scope.
Keep the evidence request separate from the work instruction. “Verify this interface against the controlled drawing” is a review assignment. It is not a direction to alter a support, change a fit, add a shim, or adjust clamp force. Those decisions require the relevant qualified owner and approved procedure.
What a useful closure record contains
If corrective work or a process trial follows, give it an ID linked to the timeline. Record the resulting authorized state, collected samples, same-feature inspection, other required checks, and product disposition. An entry that says only “maintenance fixed it” leaves the next investigation without a reproducible comparison.
If the cause remains unresolved, close the shift record with an open action rather than a guessed diagnosis. The maintenance-change worksheet includes space for evidence gaps and owners. Carry those entries into the shift handover.
When short shot and flash appear together during the investigation, the paired-defect guide helps keep the response counts separate. If a request to engineering keeps producing more questions than answers, use the technical-help request template to assemble the next review.
Sources and scope
Autodesk supplies the short definition of flash; BASF supplies the brief list of relevant tooling and process considerations. ASQ supports change analysis as a problem-investigation approach. The timeline, cavity example, review questions, and worksheet are original MTF material. No maintenance event, verified root cause, repair outcome, or equipment dimension is presented as a real factory case.