“Changed the settings. Parts look better.” That note leaves the next technician with two unanswered questions: what changed, and what does “better” mean?
A useful injection molding troubleshooting log connects a specific observation to a recorded change, identifiable samples, and a decision. Someone who missed the trial should be able to reconstruct it without asking the person who ran it to remember the details.
This guide is for operators learning to document problems and technicians recording an approved process trial. It covers the record, not the choice of machine settings. Your plant’s process limits, sampling plan, and approval rules still govern the work.
Write the question before the change
Start with the part and the observation. “Short shot” is a useful defect name, but it is not enough to locate the problem. A more useful entry is: “The end of the small rib is incomplete on cavity 2. Cavity 1 has a complete rib in the samples checked.” Add the sample identifiers and the time window.
Keep the proposed explanation in a separate field. “Air trapped at the rib” is a hypothesis unless evidence has established it. When a log puts that phrase in the observation box, later readers can mistake a suggestion for a finding.
Then state the trial question narrowly: “Does the approved change reduce incomplete ribs in cavity 2 under the comparison conditions?” That is a question a record can address. “Find the best process” is too broad for a single log entry.
Build a record you can compare
Give each trial an ID. Use it on the log, sample containers, permitted photographs, and exported machine records. Keep those references in the plant’s approved record system. A file named “latest settings” becomes ambiguous as soon as another file replaces it.
| Record | What belongs in the entry |
|---|---|
| Job identity | Date and shift; part and revision; press and mold IDs; active cavities; resin grade and lot; relevant additive or regrind identification. |
| Reference condition | Approved process-sheet or recipe revision, its location, and the last accepted inspection record. Record any known departure from that reference. |
| Observation | Feature and defect location; cavity; time window; inspected and affected counts; sample or permitted photo IDs. |
| Trial plan | Question, proposed explanation, authorized change, permitted bounds, stop criteria, and approver or approved instruction reference. |
| Before and after | Exact control label, original and trial values, units, change time, and the corresponding measured or reported process results. |
| Sampling | Stabilization criterion and whether it was met; collection period; cavity labels; sample count; inspection method, instrument ID, and relevant measurement conditions. |
| Outcome and next action | Results, missing evidence, whether the comparison was usable, final process state, sample/product status, next owner, and review time. |
Separate a setting from a result
A setting tells you what was requested. A measurement or machine result tells you something about what happened. Label them separately. A temperature setpoint should not be recorded as a measured melt temperature; a machine-side pressure value should not be relabeled as cavity pressure.
Use the equipment’s full label and units. “Pressure: 80” is incomplete. If a display offers several pressure values, record which one you used and where it came from. Write “not measured” when no relevant measurement exists. Do not fill the space with a nearby number just to complete the form.
RJG’s discussion of cavity-pressure data and process transfer explains why matching machine settings alone does not establish that the plastic experienced matching conditions. For this log, the practical consequence is simple: preserve the distinction between the control setting and the evidence collected.
Make the sample groups comparable
Decide how samples will be collected before looking at the trial result. Record the agreed stabilization criterion and collection window. There is no universal number of cycles to wait that this worksheet can supply for every material, mold, and change.
Keep cavity identity when it matters to the question. If one cavity is the concern, a combined mold-wide count can hide what happened there. Use the same inspection definition for both groups, including the same feature and acceptance reference. Record relevant differences in measurement timing or conditioning rather than assuming they were identical.
A trial can become hard to interpret even when nobody makes another intentional adjustment. A resin lot changes, production stops, a water-supply condition changes, or the inspection method changes. Give such events their own timestamped entry. They may explain why the before-and-after groups cannot support the intended comparison.
Use a simple comparison for a simple question
For an approved single-change comparison, record the planned change and track other relevant conditions. Do not assume that changing one machine control affects only one aspect of the process.
When the question involves several settings acting together, the trial needs a suitable experimental plan. NIST’s explanation of one-factor-at-a-time experiments shows their limitation: they do not reveal interactions between factors. A tidy log cannot compensate for a design that cannot answer the question.
A better-looking sample, but an unfinished decision
Constructed teaching example. The job, document IDs, counts, and times below are invented to demonstrate recordkeeping. They are not a factory case, recommended sampling plan, or machine recipe.
A two-cavity part has an incomplete rib on cavity 2. The technician records a trial under instruction TP-07. The actual setting values would be in that plant’s controlled trial instruction; this example intentionally supplies no transferable process values.
| Entry | Recorded result |
|---|---|
| Trial / reference | EX-014; part Demo Rib rev B; press P-04; mold M-12; process sheet PS-04 rev C; resin and lot unchanged in the example. |
| Baseline, group A | 10:10–10:15. Cavity 2: 6 incomplete ribs in 20 inspected parts. Cavity 1: 0 in 20. Visual reference VR-02 used for both cavities. |
| Change | 10:18. Technician carried out the change recorded in TP-07. Before/after control values and units attached to EX-014. |
| Trial, group B | 10:30–10:35, after the example’s agreed stabilization check. Cavity 2: 1 incomplete rib in 20 inspected parts. Cavity 1: 0 in 20. Same visual reference. |
| Missing evidence | The required dimensional results for group B are pending. No repeat trial has been completed. |
| Permitted conclusion | Fewer incomplete ribs were observed in the cavity-2 trial sample. Overall acceptance and repeatability have not been established. |
| Handover | Groups A and B remain identified as trial samples. Quality owns the pending measurements. Final press status and authorized disposition must be entered before handover; this example does not prescribe either. |
The visual count changed from 6/20 to 1/20. That is worth recording exactly. “Defect fixed” would overstate it: one defect is still present, another required characteristic has not been checked, and the result has not been repeated. Calling it an “83% improvement” would make a small teaching sample sound like a demonstrated production gain.
Now imagine group B had been inspected immediately after molding while group A was inspected later, with no agreed measurement timing. The log should disclose that difference. The next action may be to establish a valid comparison, rather than to make another process change.
Give the result an honest status
Use a short conclusion followed by its evidence. These four statuses are useful starting points:
- Supports further evaluation: the planned response moved in the expected direction; identify outstanding checks.
- No observed improvement: the measured response did not improve in this comparison. Keep the record so the next shift does not repeat the same unexplained trial.
- Inconclusive: missing measurements, mixed samples, changing background conditions, or an unsuitable design prevent the intended conclusion.
- Stopped: record the reason, time, product status, and responsible person. Follow the applicable response procedure.
Keep a separate field for product release and final machine state. A promising trial result is not release authorization. Likewise, writing “returned to baseline” does not prove that the current output is acceptable. Record the actual instruction followed and the required confirmation.
Use the worksheet without creating a second record system
Open and print the blank trial log, or use its fields to discuss a change to your existing form. It is an original MTF teaching aid, not a validated production form. Ask the process and quality owners which fields they already capture and which references the log should point to.
Before handing the record to someone else, check whether they can locate the samples, identify the actual change, distinguish missing evidence from a negative result, and find the next owner. If they cannot, improve that entry while the details are still available.
Carry unfinished work into the shift handover record. If you are collecting examples for a development discussion, the operator-to-technician skills map shows how a well-documented observation or trial can demonstrate a specific skill.
Apply the record to a specific problem
- Short shot and flash together: identify which parts have each defect and whether the trial traded one problem for another.
- Sink marks after increasing hold pressure: separate the setting, measured response, and unconfirmed feeding-path explanation.
- Same weight, different dimensions: compare sample history and measurement methods before attributing the difference to the process.
Sources and scope
- NIST/SEMATECH e-Handbook: One variable at a time — the limitation of one-factor-at-a-time experiments when factors interact.
- RJG: cavity-pressure data in injection molding — the distinction between machine settings and conditions experienced by the plastic. RJG sells training and process-monitoring systems; the link supports this distinction, not a purchase recommendation.
The worksheet, field explanations, and teaching example are MTF editorial work. The sources support the specific principles identified above; they have not endorsed this form. This guide does not establish operating limits, sample sizes, acceptance criteria, or a process-validation method.