Last updated: 2026-09-05 · 5 min read

Mold acceptance is where projects either bank their guarantees or discover they never had any. The difference between the two outcomes is not the buyer’s presence at a final trial — it is the evidence pack: the written, verifiable record that the tool fills, cycles, ejects and holds dimensions as agreed. A mold accepted on documents can be defended in any conversation that follows; a mold accepted on a good feeling cannot be defended anywhere. This article lays out the acceptance as a checklist discipline: what evidence to demand, what each document proves, what to inspect physically before crating, and what to do when a line fails.

Key facts

<>

  • Acceptance is an evidence review, not a ceremony: the decision rests on documents that can be re-read months later, not on the mood in the trial room.
  • The core evidence pack: trial report with settings, dimensional report against the drawing, CMM data files, trial video, spares list, and as-built drawings.
  • Functional criteria are process-anchored: the mold cycles cleanly at documented settings on a stated press class — not at whatever settings make the samples look good.
  • The dimensional report is judged against the drawing’s GD&T, feature by feature — a report without datums and tolerances is a photo, not a measurement.
  • Payment release is the enforcement mechanism: the final milestone buys the evidence pack, which is why it comes last.

    • Shrinkage — The dimensional reduction from mold cavity to cooled part. Amorphous resins: 0.4–0.7% (uniform). Semi-crystalline: 1.0–2.5% (directional). Mold makers apply shrinkage factor when machining cavity. Actual shrinkage varies with wall thickness, gate location and packing pressure. The related guide covers adjacent topics.

    Acceptance happens twice: at the press and at the desk

    The cost asymmetry explains why the discipline exists at all. A gap caught at acceptance costs a correction on the factory floor, where the people, machines and context are all standing by; the same gap caught in production costs freight, downtime, arguments over whose fault, and a correction executed without any of that context. Acceptance is the last moment the evidence is inexpensive to demand — which is why the demand is made in writing, before the crate closes.

    Two verifications make up a complete acceptance, and they answer different questions. The press verification — the final trial — answers whether the tool behaves: does it fill, eject, cycle safely and repeat? It is witnessed directly or through a timestamped video, and it produces the raw material of acceptance: samples and settings. The desk verification answers whether the record exists: every claim the trial made reduced to documents that survive — the report, the measurements, the data files, the drawings as actually built. Buyers who attend the trial but skip the desk verification accept a memory. Buyers who do both accept an asset with a paper trail — the paper trail being exactly what makes the mold transferable, insurable and defensible for the rest of its working life.

    The evidence pack, document by document

    Document What it proves What to check before signing
    Trial report The mold ran on a real press with recorded settings and produced samples Machine ID and tonnage, material and drying state, full settings sheet, findings list with dispositions
    Dimensional report Samples meet the drawing’s dimensions and geometric callouts English GD&T with datums, feature-by-feature measured vs. nominal, instrument identified, sample shot numbers cited
    CMM data files The dimensional report can be re-verified independently Raw data included, not just a PDF summary; file format you can open; measurement programs referenced
    Trial video Mechanical behavior no photo carries — cycling, ejection, action movement Timestamped, mold number visible, full cycles shown, close-ups of actions and ejection
    As-built drawings What was actually made, including every correction made during the build Corrections incorporated, water and electrical layouts as connected, spare insert locations marked
    Spares list and BOM The tool can be maintained without archaeology Every component with specification; included spares shipped against the list; supplier contacts for long-lead items

    The table’s quiet test: for every row, ask “could I hand this to a different shop in two years and have them run this tool?” A complete evidence pack answers yes — it is the same package a mold transfer would demand later, produced now, while every fact is fresh and every fact-holder is still on the payroll.

    Functional acceptance criteria

  • Cycles at documented settings. The mold runs a stable, repeated cycle at the settings written in the report, on the stated press class — the settings are part of the acceptance, because a mold that only behaves at mystery parameters cannot be handed to a production floor.
  • No flash at the documented clamp. Parting surfaces close under the tonnage stated; the flash behavior — or its absence — is judged against the material and pack window of record, as the flash diagnosis discipline prescribes.
  • Ejection clean and complete. Parts release without drag marks beyond agreed limits, without manual assist, and without witness that will grow into wear.
  • Actions operate fully and repeatably. Sliders, lifters and any mechanisms complete their travel on consecutive cycles; nothing binds, nothing needs a nudge.
  • Weight and cycle stability. Consecutive shots show stable part weight and cycle time — the signature of a process that will hold in production, not just produce a fortunate sample.
  • Cooling as connected. Circuits flow as the as-built drawing says, at the temperatures of record; a mold accepted with undocumented cooling will rediscover its temperature map in production, expensively.

    • Dimensional acceptance: reading the report that matters

    The dimensional report is the document buyers most often accept too quickly. A report that proves something has four properties: it references the drawing revision it was measured against; it carries the GD&T framework — datums, tolerances, callouts — rather than a bare dimension list; it shows measured values feature by feature, not pass/fail stamps alone; and it identifies the instrument and the sample it measured. With those four, you can audit the report yourself: scan for features sitting near their limits, check that critical callouts were measured at all, and compare the measured set against the drawing’s controlled-characteristic list. The tolerance conversation settled at DFM becomes here a checklist executed, not a feeling revisited. And the raw CMM data matters as much as the summary: a summary can be regenerated to please, but raw data is evidence — it is what lets your own quality team re-derive the verdict without borrowing the factory’s.

    Physical inspection before crating

  • Identity and traceability. Mold number, cavity identification, date stamps and version marks present and matching the documents — the marks that keep a ten-year-old spare insert from becoming a mystery.
  • Parting surfaces and cavity condition. Photographed as-found, clean, with no unreported weld marks, dents or damage; the photographs join the evidence pack as the condition baseline.
  • Water and electrical as connected. Circuit layout matches the as-built drawing; connectors labeled; no jury-rigged wiring that the BOM cannot explain.
  • Spares against the list. Counted physically, not against an email attachment — included spares packed with the tool, missing items named before the crate closes.
  • Preservation applied. Rust preventive on machined surfaces, parting faces protected, plates secured for transit — the ship-readiness the transfer packing sequence details.

    • The acceptance sequence, in order

    1. Set the criteria before the trial. The acceptance checklist is agreed at DFM or RFQ stage, in writing — criteria invented at the trial table are emotions wearing a checklist costume.
    2. Witness or video-verify the final trial. The mold cycles at documented settings; samples are pulled consecutively, labeled with shot numbers.
    3. Review the evidence pack against the checklist. Every document in the table above, checked line by line; gaps named in writing while they are still quick to close.
    4. Physically inspect the tool. The crating checklist above, executed with the factory, photographs filed.
    5. Resolve findings or record them. Everything fails-fast items go to correction with owners and dates; cosmetic observations that do not block acceptance go on record with the factory’s disposition.
    6. Release final payment against the pack. The milestone structure holds the last payment until the evidence is complete — acceptance and payment are the same act, deliberately.
    7. File everything. The evidence pack joins the program’s permanent record — the version a transfer, a dispute, or a year-five service will ask for.

    When a line fails

    Something usually does, and a failed line is not a failed acceptance — it is an acceptance doing its job. The discipline is in the disposition: dimensional misses on controlled features go back for steel or process correction before shipment, with the correction owned and dated; documentation gaps get filled before the crate closes, because the price of filling them only climbs after the tool leaves; and genuine judgment calls — a cosmetic zone slightly outside its agreed limit, a cycle-time estimate revised — are decided explicitly, in writing, with the buyer’s signature on the decision rather than absorbed into silence. What a structured acceptance prevents is the specific failure where a small miss ships unexamined and reappears as a production problem with no evidence trail to reopen. A miss recorded is a fixable miss; a miss shipped is a change order wearing a disguise.

    Gate vestige — The visible mark left on the part surface where the gate was trimmed. Edge gate: 0.5–2.0mm vestige. Pin-point gate: 0.2–0.5mm. Valve gate: 0.05–0.1mm (flush). Vestige size affects cosmetic appearance and may require secondary trimming. The related guide covers adjacent topics.

    I have shipped molds with a handshake and I have shipped molds with a binder, and I will tell you which buyers called me back with problems. The binder buyers call to order more tools. Not because their molds were better — some of my handshake molds were perfectly good tools — but because when their production team asked “prove it,” they could. Acceptance is not distrust of the factory; it is respect for the version of this project that will exist in five years, run by people we have not met yet.

    — Arthur, Tooling Engineer at aCatMould

    Frequently asked questions

    Do I need to attend the final trial in person?

    Attendance helps judgment calls — cosmetic zones, trade-offs — but a structured acceptance does not require it: the timestamped video plus the evidence pack carry the same verification remotely. Attend for the conversation, not for the evidence; the evidence should stand without you.

    Who pays for corrections found during acceptance?

    The fault line agreed at DFM decides: errors against the agreed design and specifications are the factory’s to correct; new requirements the buyer introduces at acceptance are new scope, priced as such. A written DFM is what keeps this conversation short — which is exactly why it exists.

    Should the final payment wait for the complete evidence pack?

    Yes — the milestone-for-deliverable structure exists precisely so the last payment buys the last deliverable. Payment released on enthusiasm has no enforcement left; the evidence pack is what the final milestone purchases.

    What if the factory’s dimensional report and my own measurement disagree?

    The raw CMM data settles it: re-derive the measurement from the files, check datum setup and instrument calibration first, then compare feature by feature. Disagreements without raw data on both sides are opinions — and opinions do not close acceptance.

    Is the mold “accepted” if it runs well but the drawings are missing?

    Functionally yes, administratively no — and the gap surfaces at the worst times: first maintenance, first spare order, first transfer. As-built drawings are not paperwork garnish; they are the tool’s memory. Accept when the pack is complete, including them.

    How long does a proper acceptance take?

    The trial day is one day; the desk review of the evidence pack is one to three days for a buyer’s quality team, longer for a first program with a new factory. Compared with the cost of a missed acceptance surfacing in production, the review is the highest-leverage engineering time in the whole project.

    Key takeaways

    1. Acceptance is an evidence review: trial report, dimensional report, CMM data, video, as-built drawings and spares list — each checked, not assumed.
    2. Functional criteria are process-anchored: the mold cycles cleanly at documented settings on a stated press class.
    3. The dimensional report is judged against the drawing’s GD&T feature by feature, with raw CMM data included.
    4. Physical inspection before crating — identity marks, surfaces, circuits, spares, preservation — closes what documents cannot see.
    5. Final payment releases against the complete evidence pack; a recorded miss is fixable, a shipped miss becomes a change order.

    If a tool of yours is approaching shipment, send us the drawing set and the draft acceptance criteria. You will get a complete acceptance checklist — evidence pack, functional criteria, physical inspection — structured for your program, from a named engineer, before any money moves.

    Sources
    ISO 20457:2018 — Plastics moulded parts: Tolerances and acceptance conditions
    ASTM D638 — Standard Test Method for Tensile Properties of Plastics, ASTM International
    Plastics Industry Association — SPI mold standards and surface finish sets

    Sources: Plastics Technology · Plastics Technology Online · Society of Plastics Engineers (SPE) · MoldMaking Technology