The extrusion process melts thermoplastic in a heated barrel with a rotating screw, then forces the melt continuously through a shaped die; the emerging profile is calibrated, cooled, pulled by a haul-off, and cut or wound. It is a continuous process, so the whole job is holding dimensions steady over hours of running, not shot to shot.

Extrusion is how pipe, tubing, window profiles, sheet, film, and wire insulation get made, anything with a constant cross-section. This is the process view: what the screw and barrel zones do, what the die and haul-off control, and why melt temperature and pressure are the two numbers that predict a good run. Extrusion is one of the core plastics manufacturing processes and it behaves very differently from the discrete cycle of injection molding.

What is the extrusion process?

Extrusion is continuous conversion of plastic pellets into a constant cross-section shape. Pellets drop from a hopper into a heated barrel, a rotating screw conveys and melts them, the melt is filtered and pushed through a die that gives it its shape, and the still-soft extrudate is then cooled and sized, pulled at a steady speed, and cut to length or wound onto a roll. Nothing about it starts and stops each cycle, once it is running, it just runs, which is exactly why a small drift anywhere shows up as a length of off-spec product before anyone reacts.

Because it is continuous, extrusion quality is a rate-balancing problem. The screw pushes melt out at some rate; the haul-off pulls product away at some rate; the cooling has to remove heat at the rate the line is moving. When those three agree, dimensions hold. When one drifts, a screw running hotter, a puller creeping faster, the product gets thinner or thicker a hundred feet down the line, far from the cause.

What do the barrel zones and screw do?

The barrel is divided into independently heated zones, typically four to eight on a production extruder, plus one or more die zones, each held by its own PID temperature controller. The screw running down the middle is not a simple auger; it is built in three functional sections. The feed zone conveys solid pellets from the hopper. The compression (or transition) zone has shrinking flight depth that squeezes and shears the plastic, and this is where most of the melting actually happens, from shear and pressure as much as from the barrel heaters. The metering zone then builds steady pressure to push a uniform melt through the screens and die.

Extruder screw zones and the temperature profile Feed, compress, meter, shear does much of the melting HOPPER DIE FEEDCOMPRESSIONMETERING convey pelletsshear + meltbuild pressure cooler at feedhotter toward die
The screw is a machine in itself: deep flights convey, shrinking flights in the compression zone melt by shear, and the metering zone builds the steady pressure the die needs. Barrel heaters ramp up from feed to die.

One consequence surprises people: on a fast-running line, so much heat comes from shear and friction that the barrel heaters may actually shut off and the melt holds temperature on its own. That is why melt temperature is not the same as heater setpoint, and why you measure the melt, not just the barrel.

What does the die do, and what happens after it?

The die gives the melt its cross-section, but the die alone does not set the final size, what happens right after it does. As soft extrudate leaves the die it is calibrated (sized) and cooled, and the cooling method depends on the product: pipes and profiles run through a water bath or vacuum sizing tank, while sheet and film are cooled on chill rolls or by air. Then a haul-off pulls the product at a controlled speed, and here is the key point: the haul-off speed relative to the output rate sets the final wall thickness. Pull faster and the profile thins and stretches; pull slower and it thickens. This is why an experienced operator adjusts the puller, not the die, to bring a wall thickness back on target, and why the haul-off is treated as a primary quality control rather than a conveyor.

After the die: calibrate, cool, pull, cut or wind The die gives the shape; the haul-off sets the size DIE WATER BATH / SIZINGpipe & profile CHILL ROLLS / AIRsheet & film HAUL-OFFspeed = thickness CUT to length WIND on roll measure thickness inline, drift shows up far downstream of its cause
Whether the die makes pipe, sheet, or film, the cooling method and the haul-off speed finish the job. Because the line is long, thickness drift is best caught inline, not at the end of the roll.

What can you make by extrusion?

Any constant cross-section. Change the die and, to a degree, the downstream cooling, and the same extruder makes very different products.

ProductDie / methodCooling & take-off
Pipe & tubingAnnular dieVacuum sizing tank, water bath, cut to length
Window & trim profilesProfile dieCalibration tooling, water bath, cut to length
SheetFlat (slot) dieChill-roll stack, cut or roll
Cast filmFlat dieChill roll, wound on a roll
Blown filmAnnular die, inflated bubbleAir cooling, collapsed and wound
Wire & cableCrosshead dieWater trough, wound on a reel
One machine, many products. The die sets the shape and the downstream train sets the size and finish, which is why an extrusion changeover is really a die-and-train changeover.

What are the common extrusion defects?

Extrusion defects are mostly steady-state problems, they persist and produce long runs of scrap until corrected, which is what makes them expensive.

DefectWhat it looks likeUsual cause
Melt fracture / sharkskinRough, matte, or rippled surfaceToo high a shear rate at the die; melt too cold
Die linesStreaks running with the flowNicks or contamination on the die land
Gels / fisheyesHard specks or lumps in filmDegraded or unmelted material, contamination
Thickness variationDimensions drift out of toleranceRate/haul-off imbalance, temperature drift
Voids / bubblesInternal holes, foamy spotsMoisture in resin, entrapped air, overheating
Warpage / bowProduct curves or twistsUneven cooling, uneven wall thickness
Most extrusion defects trace to melt condition, die condition, or a rate imbalance. Because the line is continuous, each one keeps producing scrap until the process is corrected, not just the part.

Why monitor melt temperature and pressure?

Because they are the two numbers that tell you what the plastic is actually doing, and neither is the same as a heater setpoint. Melt temperature measured in the stream near the die, reflects the real thermal state after shear heating, run too hot and you degrade the polymer and get gels; too cold and you get melt fracture and poor fusion. Melt pressure at the screen pack and die reflects flow resistance, a rising pressure often means a clogging screen pack, and a swinging pressure means unstable output that will print as thickness variation.

These two signals are the earliest warning an extrusion line gives. A melt pressure creeping up over an hour is a screen change you can plan; the same clog discovered as a pile of thin, out-of-tolerance pipe is scrap you cannot undo. That is why serious lines trend melt temperature and pressure continuously rather than glancing at a gauge, and treat a drift as a signal to act, not a curiosity.

How do you run an extrusion line well?

The lines that hold tolerance and yield treat the process as one balanced system. A practical order of operations:

  1. Dry the resin when it needs it. Moisture-sensitive materials bubble and gel if run wet; drying is a prerequisite, not an option.
  2. Set melt temperature by measurement. Dial in barrel zones to hit a target melt temperature at the die, not a set of heater numbers copied from memory.
  3. Balance output and haul-off. Lock screw speed and puller speed together so wall thickness holds; treat a drift in either as a defect in the making.
  4. Watch melt pressure for the screen pack. Trend it so a clogging screen becomes a planned change instead of a scrap event.
  5. Measure dimensions inline. Catch thickness drift at the line, not at the end of the roll or the cut table.
  6. Manage startup and changeover scrap. Purging and die swaps burn material and time; this is quick-changeover territory for extrusion.

Where does data tie yield and OEE together?

Extrusion punishes late detection more than most processes because it never stops. A discrete press makes one bad part per bad cycle; an extruder makes bad product continuously until someone intervenes, so the minutes between a drift starting and someone seeing it convert directly into scrap length. The signals that predict trouble, melt temperature, melt pressure, screw speed, haul-off speed, inline thickness, already exist on the line. The question is whether they are trended together and tied to the run.

When they are, an extrusion line becomes measurable the way OEE for plastic extrusion intends: speed loss, scrap, and downtime visible in one place, trended the way OEE calculation lays out. A supervisor sees the melt pressure climbing or the thickness trending toward the tolerance edge and acts before the roll is ruined. That connected layer is what Harmony runs on the plant floor on top of the extruders and gauges already installed, no rip-and-replace, and the CLS case study shows the same idea running production and quality on one record.

By the numbers

Where extrusion sits, from primary sources: