The hold-pressure setting went up. The depression over the boss barely changed. Before increasing it again, check what the trial actually established: did more material reach the part, did the affected area respond, and were the samples compared the same way?
A higher machine setting is not proof of more effective packing at a particular feature. Persistent sink marks can involve the feeding path, local section geometry, cooling, or an incomplete comparison. The useful next step is to locate the missing evidence, then give the process owner a question that a controlled trial can answer.
This guide is an evidence-review exercise for operators and developing technicians. It provides no universal pressure, time, or temperature values. Any live trial follows the job’s approved limits and sampling procedure.
Make “the sink is unchanged” a usable observation
Mark the exact feature on an approved drawing or permitted photograph. Is the depression opposite a rib, near a boss, or over another thick region? Record cavity identity, sample IDs, and when the parts were inspected after molding. Keep the viewing direction and lighting consistent when the acceptance method is visual.
If the requirement is dimensional, use the specified measurement method. A photograph may document location without measuring depth. Do not replace a required depth or surface-profile result with “looks okay,” and do not assign a numerical depth from an uncalibrated image.
Autodesk’s sink-mark and void guidance describes contraction in thicker sections: the surface may pull inward, or an internal void may form when the skin resists that movement. An acceptable-looking surface alone does not establish internal soundness. Required internal or functional checks remain part of the job’s acceptance plan.
BASF’s troubleshooting brochure, “Sink marks,” printed page 14, notes that depressions can appear after ejection. This makes the inspection timestamp relevant. A freshly ejected trial sample and an older reference sample are not a well-defined visual comparison unless their different histories are accounted for.
“We added hold” leaves several questions unanswered
Retrieve the before-and-after process record. Did hold pressure change, hold time change, or both? Was it one step of a profile? Which control label and units were used? Were transfer, cooling, material preparation, or other conditions changed in the same window? Start with the exact record, not a reconstructed shorthand.
Then separate the requested setting from available results. Identify machine-side pressure data as machine-side data. If cavity-pressure data exist, record sensor location and the trace reference; a sensor at one location is not a measurement everywhere in the part. If no relevant result was captured, “delivery response unknown” is an honest entry.
BASF also lists pressure loss in the feeding system and an absent melt cushion among packing-related sink causes. Bring the available pressure and cushion records to the process review, using the machine’s exact labels. A cushion value alone does not demonstrate effective packing at the boss.
The first review can be organized around three questions:
- Was the planned change carried out? Match the trial ID and time to the actual recipe and available process history.
- Did a measured response change? Compare identified part weights, the defect measurement, and other required characteristics. Preserve individual results instead of only a rounded average.
- Was the comparison usable? Check sample timing, cavity labels, instrument/method references, and other events during collection.
These questions are intentionally separate. An unchanged sink measurement does not mean the trial did nothing. Conversely, an increased part weight does not prove that a local depression improved or that the part meets its other requirements.
A weight plateau does not identify the frozen location
A hold-time study and a hold-pressure comparison ask different questions. In a hold-time study, weight is compared across planned holding durations under controlled conditions. Do not combine results from several pressure changes into that curve without an experimental plan that accounts for them.
Ask how the trial plan handles total time in the mold and any resulting cycle changes. Record the plan and actual timing alongside the samples. A hold-time label by itself does not describe all the conditions under which the comparison was made.
Beaumont’s “Gate Freeze—Is It Always the Gate?” explains a critical limitation: a sprue, runner, or thin part section can interrupt feeding before the gate itself freezes. Its examples include a hot-gated part whose thin section isolates a thicker region. A weight plateau therefore does not, by itself, locate the interruption at the gate.
- Sprue / runnerIdentify the feeding system.
- GateMark the entry to the part.
- Thin part sectionCheck the route beyond the gate.
- Thick featureLocate the observed sink.
For a cold-runner tool, separate runner weight from individual cavity-part weights when reviewing such a study. Beaumont discusses comparing those responses to investigate the feeding path. The pattern can guide further investigation; it is not a substitute for confirming where feeding was lost.
The sketch is deliberately incomplete for many real tools. A hot-runner system, multiple gates, or a different part layout needs its own map. Mark what is known from the tool documentation and leave uncertain paths unconfirmed. Do not copy this sequence onto a job whose geometry does not match it.
Check what “flat” means in the underlying data
Before interpreting a plateau, ask for individual measurements, the scale’s displayed increment, and the approved method. “No change” may mean no reported change at the precision retained in a spreadsheet. It is not automatically evidence of exactly equal mass.
Check whether every entry represents the same object: a trimmed part, an untrimmed part, or a complete shot including the runner. Confirm that cavity identity survived collection. A whole-shot average answers a different question from a particular cavity’s response.
A simple review note can read: “Reported part weights show no further increase over the tested hold-time steps at the recorded measurement precision. Individual results and runner data are attached. Freeze location has not been confirmed.” That preserves the observation without converting it into a tooling conclusion.
More weight, but no evidence yet about the sink
Constructed teaching example. The following records are invented to demonstrate interpretation. They are not production data, a sampling recommendation, or a pressure-setting guide.
A cosmetic cover has a depression opposite a boss. The team makes one authorized hold-pressure change and labels the resulting samples B. The exact values remain in the fictional controlled trial record; the example does not supply transferable machine settings.
| Item | Before / group A | Trial / group B |
|---|---|---|
| Sample identity | A-01 through A-05; cavity 2 | B-01 through B-05; cavity 2 |
| Individual part weights | Available in attachment W-A; same trimming and weighing method | Available in W-B; all five reported weights exceed the A-group range |
| Sink assessment | Visual classification recorded against the job reference at the required inspection time | “Looks the same” written during collection; inspection time and reference not entered |
| Other requirements | Reference inspection attached | Dimensional results pending |
| Available conclusion | The trial group has higher reported part weights. The sink comparison is incomplete, and overall acceptance has not been established. | |
The next useful action is to complete the agreed inspection of the identified samples, if they remain suitable for that method. If not, arrange a properly controlled replacement comparison. Another pressure increase would not repair the missing inspection record.
Suppose the completed assessment then shows no improvement at the boss. That would support a more specific statement: “Under this comparison, additional part mass was not accompanied by an observed improvement at the measured sink location.” It still would not prove where the added material went, establish a frozen gate, or establish that more pressure is the right next trial.
Keep the sample group’s limitations visible. Five samples in a teaching table do not establish process capability. No release decision follows from this example.
Turn the result into one evidence request
| What is missing? | A useful request to the responsible owner |
|---|---|
| The actual trial is unclear | Reconcile the recipe revision, control changes, timestamps, and collected sample IDs before comparing results. |
| The sink comparison is inconsistent | Confirm the inspection reference, viewing or measurement method, and time after molding; repeat the comparison if needed. |
| Only whole-shot weight is available | Plan identified cavity-part and, where applicable, runner measurements under the approved sampling method. |
| A feeding interruption is suspected | Ask the process/tooling owner what evidence would distinguish the possible locations on this tool’s feeding path. |
| The local feature remains the concern | Bring the marked drawing, section information, retained samples, and trial results to a process/tooling/design review. |
Use the packing review worksheet as an attachment to the trial log. Finish with the evidence-supported conclusion, open measurement, next owner, and actual product status. Keeping an unresolved question explicit is more useful than recording “gate frozen” because the chart appeared flat.
If the next trial creates flash while a feature remains incomplete, use the paired-defect guide. If weight stabilizes but dimensions do not match, continue with the dimensional comparison record.
Sources and scope
Autodesk supports the brief sink/void mechanism; BASF supports the observation that a sink can appear after ejection. Beaumont’s technical article, credited to Jennifer Schmidt, supports the distinction between a feeding interruption and proven gate freeze. Beaumont supplies molding engineering and training services; this is not a service recommendation. The record sequence, diagram, worksheet, and sample review are original MTF educational material. No verified factory case or supplier endorsement is claimed.