Plastic Extrusion Defects: 12 Problems, Causes and Fixes

1. Diagnose before you adjust

Extrusion is a continuous process, so a defect that appears at one point on the line will repeat on every metre after it. That makes diagnosis easier than in injection moulding — you can see exactly where in the line the problem starts, and you can watch it change as you adjust one variable.

Three questions identify most extrusion defects before any setting is touched:

• Does it appear before or after the calibration sleeve? A defect visible at the die but gone after calibration is a die or melt problem. A defect that appears only after calibration is a cooling, vacuum or haul-off problem. This single observation halves the search.
• Does it change with line speed? If the defect worsens when you speed up, it is shear or cooling related. If it is unaffected by speed, it is usually tooling or material.
• Is it continuous, periodic or random? Continuous means a steady-state cause — die condition, temperature, formulation. Periodic points at something rotating: a screw, a haul-off belt, a worn idler, a gear pump. Random points at contamination or feed variation.

2. Out-of-tolerance cross-section

The profile is the right shape but the wrong size — a wall too thick, an outside dimension drifting, a chamber collapsing slightly. This is the most common extrusion complaint and the most expensive, because it usually means scrap that has already been coiled or bundled.

Causes. Die wear in the land area, which increases flow and thickens the wall. A worn or scored calibration sleeve, which no longer holds the outside dimension. Melt temperature drift, which changes both viscosity and drawdown. Haul-off speed drifting relative to extruder output, which changes the drawdown ratio. Vacuum loss in the calibration tank. A formulation change — a different filler loading or a different regrind ratio changes shrinkage and drawdown.

Fixes. First check the calibration vacuum and the water temperature, because they are the cheapest to verify. Then measure the haul-off speed and the screw speed together and confirm the drawdown ratio is what it was at first-article approval. Then check the die and sleeve for wear with a pin gauge — a sleeve worn by 0.1 mm shows up as a 0.1 mm dimension shift, which is enough to fail a tight tolerance.

The pattern to watch. A dimension that drifts slowly over days is wear. A dimension that jumps between shifts is a setting or a different operator. A dimension that changes with the season is usually cooling water temperature.

3. Melt fracture and sharkskin

A fine, regular, matte ripple across the surface — sometimes called sharkskin — or in severe cases a rough, torn, spiral surface. It appears when the surface shear stress at the die exit exceeds what the melt can withstand.

Causes. Output too high for the die land length. Melt temperature too low, raising viscosity. A die land that is too short, so the melt does not have time to relax before exiting. A die lip that is damaged or poorly polished. A material with a very high molecular weight or a high filler loading, which fractures at lower shear rates.

Fixes. Raise the melt temperature in 5 °C steps — this is the fastest fix and usually enough. Reduce output. If the problem returns at the required output, the die needs a longer land or a larger die opening, which is a tool modification. A fluoropolymer processing aid can raise the critical shear rate significantly and is a legitimate formulation fix for a difficult section.

Do not try to polish out sharkskin. It is not a surface defect; it is a flow instability that happens after the melt leaves the die. Polishing the lip will not remove it.

4. Die lines and drag lines

Continuous longitudinal lines running the length of the profile, usually straight and often on one face. They are the easiest defect to diagnose because their position on the section points directly at the location on the die.

Causes. A scratch or nick on the die land or the die lip. Carbonised material or additive build-up (plate-out) on the die. A burr on the calibration sleeve entry. Poor die polishing in a corner or a narrow channel. In PVC, a build-up of degraded material at a flow stagnation point.

Fixes. Locate the line on the section, then inspect the die at the corresponding angular position. Clean and polish. If the line returns within hours, the cause is build-up rather than damage, and the fix is a formulation or die-temperature adjustment — a slightly hotter die land reduces adhesion of the deposit.

Why it matters commercially. A die line on a hidden face is cosmetic and may be acceptable. The same line on a visible face fails first-article approval and, if it reaches the customer, generates a return. Agree which faces are cosmetic before the tool is cut.

5. Bubbles and voids

Bubbles visible through a translucent profile, or blisters on the surface. In an opaque profile they may only show up when the part is cut or when a section is x-rayed.

Causes. Moisture in the resin, which turns to steam in the barrel. Air entrained at the feed throat because the feed is too hot or the screw is starving. Volatiles from an overheated melt or from a high regrind content. A vent that is blocked, on a vented extruder. Decomposition of a blowing agent where one is not wanted — a risk when cellular and rigid PVC run on the same line.

Fixes. Dry the material and verify the dryer rather than trusting it. Check the feed throat temperature — a hot feed throat causes the pellets to stick and the screw to starve, drawing air. Reduce melt temperature. Check the vent. Reduce the regrind ratio and confirm the regrind is dry.

The distinguishing test. Cut the profile and look at the bubbles. Moisture bubbles are round, evenly distributed and often concentrated toward the centre. Air entrainment gives elongated voids aligned with the flow. Volatile bubbles are irregular and usually accompanied by a smell.

6. Black specks and carbon

Dark particles embedded in an otherwise good profile, ranging from a few microns to several millimetres. In a white profile they are immediately visible and are an automatic reject.

Causes. Degraded material held in a dead spot — behind the screw flight, in a breaker plate, in a screen pack, in a die channel that is not swept by flow. Overheated zones producing local degradation. A worn screw or barrel with excessive clearance. Contaminated regrind or a dirty feed system. In PVC, any period of stopped flow.

Fixes. Purge with a proper purging compound rather than with virgin resin, and change the screen pack. Check every zone against the temperature profile. If the specks return within a shift, the problem is a dead spot in the flow path and the screw or die has to come out.

The preventive discipline. Black specks are almost always a shutdown-and-startup problem, not a running problem. A documented purge procedure at every stop, and a barrel kept at temperature with the screw forward during short stops, prevents most of them.

7. Rough surface, orange peel, poor gloss

A general loss of gloss, a stippled or pebbled surface, or a mottled appearance. Different from melt fracture, which is a regular high-frequency ripple.

Causes. Melt temperature too low, so the compound does not fuse properly against the die and the calibrator. Moisture. A die land that is too short or too cold. Insufficient vacuum in the calibration sleeve, so the surface is not pressed against the polished sleeve. A calibration sleeve that is contaminated or has lost its polish. Too much filler or an incompatible additive in the formulation.

Fixes. Raise the melt and die temperature. Check the vacuum level at the sleeve — a leaking seal or a blocked vacuum slot reduces it. Clean and polish the sleeve. If the surface is mottled rather than uniformly rough, the cause is mixing and the answer is a screw or formulation change.

8. Warpage and bend

The profile is dimensionally correct on the line and bends, twists or cups after cooling, or after a day in the warehouse.

Causes. Differential cooling — one side of the profile cooling faster than the other, usually because the water is directed at one face. Uneven wall thickness, so thick sections are still shrinking when thin ones have frozen. Asymmetric section design, where the mass is not balanced about the centreline. Too-rapid cooling overall, which locks in stress. Haul-off or take-up tension applied while the profile is still soft. Improper stacking or coiling while warm.

Fixes. Balance the cooling first — check that water is distributed around the whole section and that both sides of a flat profile see the same temperature. Slow the cooling slightly and lengthen the tank. Reduce haul-off tension. Allow the profile to cool before it is bundled or coiled.

The design point. An asymmetric section will always want to bend, and no amount of process tuning fully removes it. A cooling fixture or a short annealing pass after the tank is often the practical answer, and a balanced section is the permanent one.

9. Sink marks in thick walls

A depression on the surface opposite a thick section or a heavy rib. Less common in extrusion than in injection moulding, because the melt is under continuous pressure, but it appears wherever the wall is thick enough to shrink after the skin has frozen.

Causes. A wall section that is too thick for the cooling available. Insufficient melt pressure at the die, so the section is not packed. Vacuum pulling a soft wall inward. A hollow chamber with too little internal support pressure. Too-fast cooling on the outside relative to the inside.

Fixes. Reduce the wall thickness — this is the real answer and it is a design change. Raise the internal pressure on a hollow profile. Adjust the vacuum so it holds the outer surface without collapsing the wall. Slow the cooling on the thick side.

10. Colour streaks and shade variation

Swirls, dark streaks, light patches, or a shade that drifts over the length of a run.

Causes. Masterbatch or pigment not dispersed, because the melt temperature is low or the screw has insufficient mixing. A let-down ratio that is wrong or inconsistent. Screw wear producing unmelted particles. Dead spots in the die holding degraded material. Regrind ratio varying through the run. For a pre-compounded colour, a batch-to-batch difference from the compounder.

Fixes. Raise back pressure and melt temperature to improve dispersion. Verify the let-down ratio by weighing. Fix the regrind ratio and hold it constant rather than topping up with whatever is available. If the shade drifts gradually over hours, the cause is usually thermal — the die is slowly building up or the compound is slowly changing.

The commercial detail. Colour is judged against a standard under a specified light source. Agree the standard, the light source and the acceptable colour difference before the first run, otherwise every shipment becomes a negotiation.

11. Delamination

The outer layer separates from the core — in a co-extruded profile, a laminated profile, or a profile made from mixed regrind. It shows as a peeling edge, a blister along the length, or a layer that can be lifted with a fingernail.

Causes. Incompatible materials at the interface. Interface temperature too low, so the two melts do not entangle before freezing. Contamination — moisture, oil or dust — at the interface. A co-extrusion die that is not bringing the streams together at the right point. In a laminated profile, an adhesive that is under-cured or applied too thin. In mixed regrind, incompatible polymer contamination.

Fixes. Raise the interface temperature by adjusting the two melt temperatures toward each other. Confirm the material pairing against a compatibility chart rather than assuming. Clean the die and the feed system. For laminated profiles, check the adhesive and the laminating roll temperature.

The one that cannot be fixed on the line. If the two polymers are chemically incompatible, no setting will produce a bond. That is a design decision that has to be reversed, not a process problem.

12. Cracks after cutting

The profile is sound on the line, then cracks at the cut end, along a corner, or between chambers — sometimes minutes later, sometimes days later.

Causes. Internal stress from too-rapid cooling, released when the profile is cut. A dull or misaligned saw blade, which introduces a notch. Cutting the profile while it is still warm. A sharp internal corner acting as a stress concentrator. In PVC, a formulation with insufficient impact modifier. Solvent contact after extrusion — some cleaning agents will stress-craze PVC and polycarbonate.

Fixes. Slow the cooling so the profile leaves the tank closer to ambient. Sharpen and align the blade, and check the cut-off saw's clamp pressure. Cut cool rather than warm. Increase the corner radii on the section.

Why it is worth chasing. A crack that appears after delivery looks like a material failure and destroys confidence in the whole shipment, even when the root cause is a saw blade. If cracks appear after a delay, an annealing pass is the usual remedy while the process is corrected.

13. Cut length and cut angle

Lengths outside tolerance, ends not square, or a length that varies within a bundle.

Causes. Haul-off speed varying, or the encoder slipping on the profile. A cut-off saw that is not synchronized with the haul-off. Blade deflection on a large section. The profile moving during the cut because the clamp is worn. Thermal expansion — a profile cut warm and measured cold will be short.

Fixes. Check the encoder wheel for wear and slip, and confirm it is running on a clean, dry surface. Verify the saw is synchronized rather than free-running. Replace worn clamp pads. For tight length tolerances, measure at a defined temperature and state it in the specification.

14. The troubleshooting matrix

Read across, not down. A defect with a plausible cause in only one column is usually that cause.

Cooling, calibration and haul-off carry more defects in extrusion than in injection moulding, because a continuous process has no natural end point where the part is "done".

15. When it is the die, not the process

Some defects cannot be processed out. Recognising them early saves weeks of parameter chasing.

• The defect is in the same angular position on the section every time. That is a specific location on the die or the sleeve.
• The defect survives a full die clean and a new screen pack. It is damage or a design problem, not contamination.
• The defect appears at every output and every temperature. The die geometry is wrong for the section.
• One face is rough and the opposite face is fine. Asymmetric flow in the die.
• The profile is out of tolerance in a way that no drawdown adjustment fixes. The die was cut to the wrong allowance.
• The defect only appears above a certain line speed and never goes away. The die land is too short for the output the commercial case requires.

The commercial point: a defect that needs a die modification should be found during first-article approval, on a few hundred metres, not after twenty thousand metres have been coiled. That is what a proper first-article dimensional report and a retained sample are for.

16. Frequently asked questions