A part warps when different areas shrink by different amounts as the plastic cools, pulling it out of flat. It's mostly set by design and tooling — uniform walls, ribs instead of thick sections, even cooling, and a sensible gate — so the cheapest time to prevent it is before the tool is cut.
All plastic shrinks as it cools. If the whole part shrank by the same amount it would just come out slightly smaller — which is why the tool is cut oversize to compensate. Warp happens when shrinkage is uneven: one area cools or shrinks faster than another, and the difference bends or twists the part. The biggest drivers are:
How a part warps hints at the cause — and the cause points to a fix (these often combine, so confirm with your molder or a mold-flow check rather than treating one as certain):
Resin matters too: semi-crystalline resins (such as PP and nylon) generally warp more than amorphous ones (such as ABS and PC), because their larger shrinkage is also less uniform, varying with cooling rate and flow orientation — worth weighing when the part has large flats.
A DFM review flags the warp-prone features before any steel is cut — far cheaper than correcting a finished tool.
Every warped part comes from the same root cause — one part of the wall shrinks more, or at a different time, than the part next to it, and the mismatch bends the part as it cools. Nothing warps if it all shrinks the same amount at the same moment. So the whole job of preventing warp is finding where the shrinkage is uneven and evening it out. There are four places that unevenness comes from, and reading which one you have tells you the fix.
A thick section stays molten longer and shrinks more than the thin wall beside it; when it finally shrinks it pulls the part toward the thick side. This is the most common warp and the easiest to design out: keep the wall uniform, core out thick bosses and ribs, and blend thickness changes with a taper rather than a step. A part that is already uniform rarely warps from thickness.
Even a uniform wall warps if one face of the tool runs hotter than the other, because the hot side freezes last and shrinks more. A classic case is a tall rib or a boss sitting over a core pin that cannot shed heat — the steel around it runs hot and that region shrinks differently. The fix lives in the tool: balanced cooling lines on both halves, extra cooling (baffles, bubblers or beryllium-copper inserts) in the hot spots, and a mold-temperature check that both halves actually run where the setpoint says. This is why cooling-channel layout is part of warp control, not a separate topic.
Glass-filled resins are the warp trap people miss. Glass fibers line up with the flow, and the part shrinks far less along the fibers than across them — sometimes several times less. That built-in anisotropy bows flat panels and pulls boxes out of square even when the wall is perfectly uniform. Unfilled semi-crystalline resins (PP, PE, nylon, acetal) have a milder version of the same problem because the polymer chains orient in flow. The levers here are gate location and gate count (they set the flow direction), and sometimes a resin or filler change; a fill simulation is the honest way to predict it before steel is cut.
Too little holding pressure leaves the far end of the part under-packed and higher-shrinking; too much over-packs the gate region and locks in stress that releases as warp once the part is out of the tool. Ejecting a part before it is cool enough, or pushing on it unevenly with too few ejector pins, bends it while it is still soft. These are process and tool issues: balance the pack, hold until the gate freezes, give the part enough cooling time, and support it on ejection.
| Cause | Why it warps the part | Fix |
|---|---|---|
| Uneven wall thickness | A thick section stays molten longer and shrinks more than the thin wall beside it, pulling the part toward the thick side | Keep the wall uniform; core out thick bosses and ribs; blend thickness changes with a taper, not a step |
| Non-uniform cooling | If one mold face runs hotter than the other, that side freezes last and shrinks more, bowing even a uniform wall | Balance the cooling lines on both halves; add cooling at hot spots; confirm both mold halves run at the setpoint |
| Fiber orientation (filled resin) | Glass fibers line up with the flow and the part shrinks far less along the fibers than across them, bowing flat panels | Set flow direction with gate location and gate count; verify with a fill simulation; consider a resin or filler change |
| Residual stress from packing | Under-packing leaves the far end higher-shrinking; over-packing locks in stress at the gate that releases as warp after ejection | Balance the pack; hold pressure until the gate freezes so shrinkage is fed evenly end to end |
| Ejected too early / uneven ejection | Pushing the part out while it is still soft, or on too few ejector pins, bends it before it has set | Give the part full cooling time; support it on enough ejector pins so it lifts out evenly |
A part can measure fine at the press and warp hours or days later as residual stress relaxes and the resin finishes shrinking — semi-crystalline resins in particular keep shrinking a little after molding. That is why a warp problem is checked on parts that have sat and stabilized, not on a hot shot straight off the tool, and why the first-article inspection waits for the parts to settle. If the part is flat at the press and bowed the next morning, suspect orientation, locked-in stress or uneven post-molding shrinkage before blaming the wall alone.
Design first, because it is free: uniform walls, cored thick sections, ribs at half to six-tenths of the wall (a standard design-for-manufacture ratio; see Protolabs' injection-molding design guidelines), symmetry where possible. Then the tool: balanced cooling and gate placement that controls the flow direction. Then the process: balanced packing, full cooling time, even ejection. We run the cheap design fixes before anyone pays to re-cut steel, and where glass-fill or a flat cosmetic panel makes warp likely, a fill-and-warp simulation up front is far cheaper than discovering it on a T1 sample. The entered answer is general guidance; your part is confirmed against its drawing, resin and a mold review.
Why do molded parts warp? Uneven shrinkage as the part cools pulls it out of flat — from uneven walls, gate location, or fiber orientation (large flats amplify it).
How do you prevent it? Uniform wall, ribs over thick sections, even cooling, a sensible gate; catch it in DFM before tooling.
Warp vs. sink vs. shrinkage? Shrinkage is overall size reduction; a sink is a local dimple; warp is the whole part bending from uneven shrinkage.
Do glass-filled resins warp more? They warp differently — fibers shrink less along the flow than across it, so a flat part can bow predictably; plan for it in the gate and design.
Can it be fixed after tooling? Partly, via cooling/mold-temp/process tuning or a holding fixture, but a geometry that wants to warp needs a design change — cheaper before the tool is cut.
JCH Design & Manufacturing can review your part for warp-prone features — uneven walls, large unsupported flats, gate placement — and propose changes before any steel is cut. Send your part and target quantity; NDA first.
Book a 30-minute call