Parts Customization / Processing Service Expert

From concept to product
From prototyping to production, our production sites can handle all your manufacturing projects.
Home / All / Injection Molding / How can I reduce common defects like warping and sink marks in injection molded parts?

How can I reduce common defects like warping and sink marks in injection molded parts?

Update Time:2026/10/12

The short answer

Warping and sink marks are volumetric problems. Amorphous resins like ABS and polycarbonate shrink 0.3 to 0.8 percent; semi-crystalline PP, POM and nylon shrink 1.0 to 3.0 percent and sink and twist more. Uniform walls, ribs at 40 to 60 percent of the wall, cooling balanced within 5 degrees Celsius, and correct hold pressure fix most of it.

Two defects, one mechanism

A sink mark is a shallow depression on the surface directly above a thicker internal feature. The mechanism is simple: plastic shrinks as it cools, a thick section stays molten longer than the skin around it, and when the core finally contracts it pulls the still-soft surface inward. The thicker the local section relative to the nominal wall, the deeper the dimple. A boss or a rib root is the usual location.

Warping is the same shrinkage acting unevenly. When one region of a part shrinks more than another, or one face cools faster than the other, internal stress builds and the part distorts to relieve it. Flat panels bow, round parts go oval, and mating surfaces stop aligning. It is the more expensive of the two defects because it usually breaks assembly fit rather than appearance, and it is often discovered at the assembly line rather than at the press. For scale, molded shrinkage measured between processing temperature and ambient can reach 20 percent by volume; what matters is not the total but the difference between one point of the part and another.

Matrix of injection molding defects and their fixes across four columns showing the defect, its mechanism, the design fix and the process lever: sink marks where the core shrinks after the skin with a rib thickness of 40 to 60 percent of the wall fixed by raising hold pressure and holding longer, cooling-driven warpage where one face cools first fixed by balancing both tool halves and keeping faces within 5 degrees Celsius, fibre-driven warpage from anisotropic shrinkage fixed by moving or adding a gate and lowering fill speed, and dimensional drift from unstable shrinkage fixed by drying the resin to a minus 40 degree dew point
Read it left to right and the cost rises. The design fix is free at the drawing stage and expensive after the steel is cut; the process lever is cheap today and only partly effective on an existing tool.

The resin family sets the size of the problem

Shrinkage is a property of molecular structure, not a setting. Amorphous resins such as ABS, polycarbonate, polystyrene and acrylic keep a disordered structure as they cool and shrink roughly 0.3 to 0.8 percent with good dimensional stability. Semi-crystalline resins such as polypropylene, polyethylene, nylon 6, nylon 6/6 and POM arrange into crystallites as they cool, which changes specific volume much more, and they shrink roughly 1.0 to 3.0 percent. That larger total shrinkage is what makes them sink and warp more readily for exactly the same geometry and the same process window.

Fillers change the picture in a way that surprises people. A 30 percent glass-filled PBT can shrink about 0.2 percent along the flow direction and 0.8 percent across it, so total shrinkage falls but anisotropy rises fourfold, and the part twists even though it shrinks less. That is why a glass-filled grade solves a sink-mark problem and creates a warpage problem in the same revision.

Range band chart of mold shrinkage in percent of nominal dimension by resin family on a shared axis: amorphous resins such as ABS, PC, PS and PMMA shrink 0.3 to 0.8 percent, semi-crystalline resins such as PP, PE, nylon 6, nylon 6/6 and POM shrink 1.0 to 3.0 percent, and a 30 percent glass-filled PBT shrinks 0.2 percent along the flow direction and 0.8 percent across it
Semi-crystalline resins occupy three times the band of amorphous ones. The design rules are identical; the tolerance budget is not.

Design fixes that remove the defect instead of hiding it

The first correction is the wall, and it is the only one that costs nothing at the drawing stage. Keep the nominal wall uniform, and where a thickness change is unavoidable, transition gradually over a distance of at least three times the thickness change rather than stepping it. A stepped wall concentrates shrinkage exactly where the step is.

Ribs and bosses come next. A rib should be 40 to 60 percent of the adjacent nominal wall in thickness, 2.5 to 3 times the wall in height, and spaced at least twice the wall apart. A boss should follow the same thickness rule, with a radius where it meets the base and gussets running to the nearest wall rather than a larger diameter. These proportions matter because a rib that is as thick as the wall just reproduces the sink mark it was meant to prevent. Where the requirement is stiffness rather than thickness, coring a thick section out and ribbing it removes the problem entirely. Radii help too: an internal corner should carry a radius of at least half the local wall thickness, because a sharp corner is simultaneously a stress concentration and a flow restriction.

Gate, runner and pack: the process half

Once the tool exists, the press has three effective levers. Gate size and position is the first: a gate that freezes before the thick section has been packed cuts off the feed, and a gate that fires straight at a thick wall leaves a sink mark on the far side. A working rule is a gate depth of at least 70 percent of the wall thickness, so a polycarbonate part with a 2.5 mm wall wants a gate around 1.8 mm deep, and a fan or tab gate spreads the flow better than a pinpoint gate on a large flat surface. Sink marks opposite the gate are the classic signature of a gate that is too small or a hold pressure that is too low.

Packing is the second lever. Hold pressure is commonly set at 50 to 70 percent of injection pressure, and hold time at a minimum of two seconds per millimetre of wall, so a 2 mm section wants at least four seconds. Short hold time lets material backflow out of the cavity into the runner, which increases shrinkage and brings the sink mark back. Runner balance is the third, and it is invisible until it is not: in a multi-cavity tool, a fill-time difference above about 5 percent between cavities produces a different shrinkage rate in each, so measure fill time per cavity and use a naturally balanced runner rather than a ladder layout.

Cooling is the single biggest lever on warpage

Cooling accounts for roughly 50 to 70 percent of the injection molding cycle, and uneven cooling is the most frequent cause of warpage. If the core side runs at 60 degrees Celsius and the cavity side at 80, the part bows toward the hotter side, because the cooler face sets first and the hotter face shrinks afterwards. The working targets are a temperature difference below 5 degrees Celsius between any two points on the mold surface, and cooling channels placed within one channel diameter of the cavity wall, so a 10 mm channel sits no more than 10 mm from the steel face. Cores that cannot take straight-through lines get baffles or bubblers.

Where a geometry has hot spots a drilled circuit cannot reach, conformal cooling is the upgrade. Channels printed to follow the cavity shape remove heat 15 to 30 percent more efficiently than straight drilled lines, and published studies report cycle-time reductions in the 20 to 40 percent range along with a clear reduction in warpage. Pulsed or rapid cooling is a cheaper intervention: alternating high-flow coolant cycles have been reported to narrow mold-surface variation from about plus or minus 8 degrees Celsius to plus or minus 2, and to cut scrap on thin-wall parts by 12 to 18 percent. Both are worth costing against the scrap and rework they remove rather than against the tooling line alone.

Four-step diagram ordering corrections for warping and sink marks in injection molded parts from cheapest to most expensive: fix wall thickness and rib proportions on the drawing, then set gate size and position and packing pressure and time, then balance the cooling in both halves of the tool to within 5 degrees Celsius, and finally change the resin grade or dry it properly
The order is not arbitrary. Each step costs more to change than the one before it, and each is less effective once the tool is cut.

Material and drying: the part nobody checks

Moisture is an underrated cause of both defects, and it produces exactly the symptom people blame on the tool. Nylon 6/6 with a moisture content above about 0.15 percent splays and develops unpredictable shrinkage; ABS needs two to four hours at 80 degrees Celsius, nylon 6/6 four to six hours at 80 degrees Celsius to a dew point near minus 40 degrees Celsius. Skipping the dry has been measured at around plus or minus 0.3 percent shot-to-shot shrinkage variation, which is the entire tolerance band on many molded parts. Regrind with unstable molecular weight does the same thing, which is why any regrind policy belongs in the process sheet rather than in an informal practice at the press.

Material choice is the last lever and the bluntest. Moving from a semi-crystalline resin to an amorphous one, or to a glass-filled grade, changes the base shrinkage and often removes the defect — but it also changes impact strength, chemical resistance, temperature rating and cost, so the substitution is an engineering decision and not a shortcut.

When the defect is a design decision rather than a fault

Some parts will always move, and the honest answer is to specify the movement rather than chase it. A large flat panel in a semi-crystalline resin will bow; either accept a stated flatness limit, with 0.5 mm per 200 mm a common commercial figure, or add ribs and warpage compensation into the tool rather than adding packing pressure. A thick section required to carry a thread will keep a sink mark unless it is cored out or the thread is moved to a molded-in insert. And a drift of 0.1 mm over a 100 mm span is normal for a molded part in a semi-crystalline resin, so a drawing that expects machined-metal flatness on a molded panel is specifying the wrong process rather than the wrong settings.

Three limits belong on any troubleshooting brief. Warpage on an existing tool can be reduced by process change but rarely eliminated, because the imbalance is built into the cooling layout. Sink marks can be hidden by texture and by a matte finish, which is a legitimate cosmetic solution but not a dimensional one. And every defect fix has a cost on another line: extra hold time lengthens the cycle, conformal cooling adds tooling, and a glass-filled substitute raises material cost and abrasion on the tool. The cheapest geometry that meets the requirement is the target, not the flattest part that can be molded at any price.

See injection molding for the process these defect rules belong to, plastic injection molding for the service that runs the tool, and our partner network if you need a second opinion on an existing tool.

Scope and sources. The gate depth rule of at least 70 percent of wall thickness with the polycarbonate example of a 1.8 mm gate on a 2.5 mm wall, the hold pressure band of 50 to 70 percent of injection pressure, the hold time rule of two seconds per millimetre, the 5 percent cavity-to-cavity fill time limit and the naturally balanced runner advice, the cooling channel placement within one channel diameter of the cavity wall, the 5 degrees Celsius mold surface limit, the 60 against 80 degrees Celsius core and cavity example, the drying schedules of 0.15 percent moisture and 4 to 6 hours at 80 degrees Celsius for nylon 6/6 and 2 to 4 hours for ABS, the plus or minus 0.3 percent shrinkage variation from skipping the dry, the 0.2 against 0.8 percent flow and transverse shrinkage of a 30 percent glass-filled PBT, and the flatness symptoms of 0.1 mm over 100 mm and 0.5 mm per 200 mm come from a shrinkage and warpage root-cause study. The rib thickness of 40 to 60 percent of the nominal wall, the boss wall rule, the note that amorphous resins shrink 0.4 to 0.7 percent against 1.0 to 2.5 percent for semi-crystalline resins, the description of sink marks as a volumetric problem opposite thick sections, and the three drivers of warpage — non-uniform cooling, non-uniform wall thickness and fibre orientation — come from a molding defect prevention guide. The industrial shrinkage range of 0.1 to 3.0 percent, the high-shrinkage family list of PP, PE, PA6, PA66 and POM and the low-shrinkage list of ABS, PC, PS and PMMA come from a plastic shrinkage reference. The figure that molded shrinkage can reach 20 percent by volume between processing temperature and ambient, and the list of causes — low injection pressure, short pack and hold, high melt and mold temperature — come from a shrinkage and warpage chapter. The finding that cooling is 50 to 70 percent of cycle time, that conformal cooling removes heat 15 to 30 percent more efficiently, and that pulsed cooling narrows mold temperature variation from about plus or minus 8 to plus or minus 2 degrees Celsius for a 12 to 18 percent scrap reduction, come from an efficiency study of molding cost per part. This version replaces the unsourced scrap-rate example and the survey figures that appeared in an earlier version of this answer. All values are working ranges for review against a specific part, resin and tool, not guarantees.