Why Warpage in Injection Molding Is Differential Shrinkage, Not Just Cooling

Buyers usually blame a hot spot in the mold. Hot spots matter, but they are one trigger among several.

The actual mechanism is differential shrinkage. Plastic contracts as it cools from melt to solid. Where two regions contract by different amounts or in different directions, internal stress builds.

The mold holds the part flat while the stress builds. At ejection the constraint disappears, and the stress releases by bending the part.

A local shrinkage difference of just 0.1–0.3% can distort a part by several millimeters, per ZetarMold’s warpage guide. That is why a part can look perfect in the mold and fail inspection the next morning.

Thick Meets Thin: The Most Common Design Cause

Thick sections cool slower and shrink more than the thin walls beside them. The thin wall freezes first. Then the thick zone keeps contracting and drags the thin wall with it.

You see this opposite ribs and bosses, or any lump of plastic the designer forgot to core out.

Keep wall thickness within ±10% of nominal, or the thick zone will pull the thin one out of flat. Core out heavy sections and replace them with ribbed structures.

Hold ribs to 50–60% of the adjacent wall, and bosses to about 60%. Where thickness must change, taper the transition at 3:1 or gentler. The wall thickness masterclass covers these rules in full, with the per-material ranges.

Gate Location Sets the Stress Pattern

Melt near the gate packs under high pressure; melt at the far end of flow packs under low pressure. The high-pressure region shrinks less, the low-pressure region shrinks more, and the difference freezes in as orientation stress.

The gate decides the flow direction, and the flow direction decides where the frozen-in stress points.

Long flat parts gated from one end are the classic twist case. Fix it with a central gate, multiple balanced gates, or a fan gate that spreads the flow.

A ScienceDirect warpage study measured the same effect, with polymer near the gate shrinking less than polymer at the end of flow. Gate sizing and placement rules live in the runners, gates, and vents overview.

Over-Packing Warps Parts Too

Packing feeds shrinkage, so the instinct is to raise it when parts warp. That instinct backfires past a point.

Extra pressure packs the gate region solid while the far end has already frozen. The pressure gradient becomes a stress gradient, and the part bends after ejection.

Over-packing near the gate while the far end stays under-packed builds the stress gradient that bends the part after ejection. The fix is to back packing off in 5–10% steps until just before gate freeze-off, per ZetarMold’s process protocol.

A decayed packing profile, high early and tapering off, evens out volumetric shrinkage along the flow path. Packing and cooling belong to the process window, covered in the scientific molding guide.

Material Choice Sets Your Shrinkage Baseline

Some resins simply move more than others. Semi-crystalline plastics like PP and nylon shrink 1.5–3%, while amorphous ABS and PC stay near 0.4–0.7%.

Worse, semi-crystalline resins shrink more across the flow direction than along it. That directionality never shows in CAD.

PP and PA shrink 1.5–3% while ABS and PC stay near 0.4–0.7% — that gap is the warpage budget.

Changing resin is rarely free, since cost and material approvals both shift. Still, quoting the low-shrink option shows the buyer what flatness costs.

Glass Fiber Cuts Shrinkage but Adds Direction

Glass fiber looks like a cure. PP with 20% glass drops to roughly 0.3–1.0% shrinkage.

The catch is that fibers restrict shrinkage along their own direction. Where flow is unbalanced, the part shrinks differently along and across the fibers, and twists.

Glass fiber cuts total shrinkage but makes the part shrink differently along the fibers than across them. Fictiv’s warping guide documents a 30% glass-filled nylon bracket that warped one way until its gate moved to the center of mass.

For filled grades, move the gate or the flow path to balance fiber alignment, and slow the injection speed.

Ejecting Too Hot Bends the Part

Some warpage is mechanical, not thermal. A part ejected before its core cools below the heat deflection temperature is still soft inside.

Ejector pins dent it, and it finishes cooling with nothing holding it flat.

A part ejected before its center drops below the heat deflection temperature has no stiffness left to fight internal stress. Extend cooling until the center is at or below the HDT, and check it with an infrared camera at ejection.

Put pins on ribs and thick sections, use sleeve ejectors for cylindrical features, and keep the ejection force even. Pin and sleeve choices are covered in the ejector pins and sleeves guide.

Warpage That Arrives Late: Stress and Moisture

Some parts pass inspection at the press and bow days later. Frozen-in stress relaxes over time or under heat, and the part moves when it does.

Nylon is the classic late warper. It absorbs moisture from the air, and the dimensions drift as it does.

Nylon parts that measure flat on Monday can bow by Friday in a humid warehouse. Annealing relieves residual stress in semi-crystalline parts like PA and PBT.

Dry moisture-sensitive parts before use, and hold tight parts in cooling jigs until they stabilize.

Diagnose Warpage in One Hour

Guessing burns whole shifts. Run a short-shot study and point an infrared camera at the part as it ejects.

Short shots reveal fill balance and the thermal image reveals cooling uniformity, separating flow-driven warpage from cooling-driven warpage within an hour. That one-hour diagnostic comes from ZetarMold’s shop protocol.

Match the symptom to the cause before changing anything:

This symptom-cause-fix table follows the diagnostic in Fictiv’s warping guide. Warpage in injection molding rewards diagnosis over guesswork, and this table is the fastest starting point. Fix the cheapest variable first: mold temperature balance costs nothing, and it resolves more warpage cases than any other single adjustment.

Try the Interactive Defect Troubleshooter

Common Questions