Last updated: 2026-09-01 · 4 min read

Why Do Parts From the Same Mold Vary Between Cavities — and What Fixes It?

Cavity-to-cavity variation means that parts ejected from the same mold, in the same shot, measure differently from one cavity position to the next. The root causes are almost always in the runner balance, cooling layout, or clamp-plate deflection — not in the resin or the press itself.

Runner system — The channel network that delivers melt from the machine nozzle to each cavity. Cold runner: solidifies between shots, produces runner waste. Hot runner: stays molten, eliminates runner waste but adds $15k–$40k to mold cost. See also related guide.

Key Facts

  • In a naturally balanced 4-cavity mold with a symmetrical runner, cavity-to-cavity dimensional spread is typically 10–30 % of the total tolerance band.
  • In an artificially balanced (H-pattern) runner or a family mold, the spread can 50–80 % of tolerance if the runner is not corrected after T1.
  • The most common symptom is one cavity consistently producing parts at the high end of the tolerance while another sits at the low end — not random scatter.

    • Runner system — The channel network that delivers melt from the machine nozzle to each cavity. Cold runner: solidifies between shots, produces runner waste. Hot runner: stays molten, eliminates runner waste but adds $15k–$40k to mold cost. The related guide covers adjacent topics.

    Where the Variation Comes From

    Three mechanisms dominate, and they compound rather than cancel.

    Runner imbalance. In an H-pattern or X-pattern runner, the flow path from the sprue to each cavity is not identical. Even a few millimeters of extra runner length shifts the pressure drop and the melt temperature that arrives at the gate. The cavity with the shortest path sees higher packing pressure and packs out larger; the cavity with the longest path packs out smaller.

    Cooling asymmetry. Drilled cooling channels rarely every cavity equally. The cavities nearest the water inlet run cooler than those at the far end of the circuit, which means they freeze earlier and shrink less. A 5 °C mold-temperature difference between two cavities produces roughly 0.05–0.10 % difference in shrinkage — which matters on a 0.1 mm tolerance.

    Platen deflection. On a large mold in a small press, the clamp plates deflect under injection pressure. The cavity in the center of the mold sees slightly more clamp force than the cavities at the corners. That difference manifests as flash on the center cavity or short shots on the corner cavity — or as a dimensional shift of a few microns on critical surfaces.

    How to Diagnose Which Mechanism Is at Fault

  • Run a short-shot study: if the fill pattern shows one cavity filling first and another last, the runner is imbalanced. The fix is runner re-sizing or a mold-flow re-simulation.
  • Measure mold surface temperature at each cavity with a contact pyrometer after a stable run.
    A spread above 5 °C between cavities points to cooling asymmetry — add baffles or re-route the circuit.
  • Check the parting-line witness on the ejected parts.
    A heavier witness on center cavities indicates platen deflection; the fix is stiffer support pillars under the B plate or a larger press.
  • If the dimensional pattern is consistent across 20 consecutive shots (same cavity always high, same cavity always low), the cause is tooling — not process drift.

    • Mechanism Symptom Diagnostic Fix
    Runner imbalance one cavity consistently larger short-shot fill study re-size runner legs
    Cooling asymmetry cycle-to-cycle variation in specific cavities mold surface temp map baffles, re-route circuit
    Platen deflection center vs corner cavity difference parting-line witness check support pillars, larger press
    Cavity insert tolerance random scatter across cavities CMM each insert re-machine or swap inserts

    A practical rule: on a multi-cavity mold with tolerances tighter than 0.1 mm, the mold-flow simulation should include a runner balance check before steel is cut — and the T1 report should include cavity-to-cavity dimensional data across at least 10 consecutive shots, not just a single sample from one cavity.

    Frequently Asked Questions

    Is cavity-to-cavity variation worse in family molds?

    Yes, because the cavities are different shapes and volumes, so each fills and packs differently. The runner balance for a family mold is a compromise by design — which is why family molds are acceptable only when the tolerance requirement on each cavity is relaxed enough to absorb the inherent imbalance.

    Can valve gating fix cavity-to-cavity variation?

    Partially. A valve-gated hot runner with independent timing per nozzle lets the process engineer tune fill speed and hold pressure for each cavity. That addresses runner imbalance but does not fix cooling asymmetry or platen deflection.

    How much cavity-to-cavity spread is acceptable?

    For a dedicated multi-cavity tool on a non-cosmetic engineering part, the spread should consume no more than 30 % of the tolerance band. If it uses more than 50 %, the tool needs runner correction before production starts — otherwise every cavity has to be held to a tighter nominal window, which raises scrap rate.

    Key Takeaways

    1. Cavity-to-cavity variation is a tooling problem, not a process problem — adjusting the machine settings cannot fix a runner that is physically unbalanced.
    2. The diagnosis is straightforward: short-shot study for runner, temperature map for cooling, parting-line witness for deflection.
    3. The fix belongs in the mold-flow simulation and the T1 report, not in production troubleshooting after the tool is approved.

    If the DFM review covers multi-cavity layouts, it should specify the acceptable cavity-to-cavity spread as a fraction of the tolerance band — and the T1 acceptance criteria should include a statistical sample from every cavity, not a convenience sample from one.

    Sources
    Beaumont Technologies — mold filling analysis
    Autodesk Moldflow — simulation
    MatWeb — material property data

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