Plastic Sheet Extrusion: Roll Stack, Polishing and Thickness Control

What a sheet line actually consists of

People describe sheet extrusion as if it were profile extrusion with a wider die. It is not. A profile line pulls a shape through a calibrator and the shape is essentially fixed once it leaves the die. A sheet line has a second forming step — the roll stack — and that is where most of the dimensional and surface quality is won or lost.

A typical line runs in this order:

Extrusora (single screw, 30–37 L/D, vented if the resin needs it) → screen changer → melt pump (on anything that needs gauge held to ±2%) → static mixer → flat die (T-die or coat-hanger) → roll stack (two, three or four rolls) → cooling rollers or a cooling conveyor → thickness gauge → edge trim → haul-off → winder or shear.

The die sets the width and the initial thickness. The roll stack sets the final thickness, the surface finish on both faces, and the flatness. Between them there is a gap of perhaps 200 mm, and everything in this article happens in that gap.

The roll stack: geometry decides gauge

The roll stack is three (sometimes two or four) temperature-controlled chrome-plated steel rolls arranged vertically or at an angle. The sheet passes through the first nip, wraps part of the way around the middle roll, and exits through the second nip. Wrap length on the middle roll is the single most-used lever for surface quality, because that is where the sheet is held flat against a polished surface while it cools below its glass transition or softening point.

Table 1 — Roll stack configurations. The third roll is not there to cool faster; it is there to keep the sheet flat while it does.

Roll diameter is usually 300–450 mm for sheet under 6 mm and larger for heavy gauge. Smaller rolls give a shorter contact arc, which means less cooling per pass and a longer line, but they also reduce the bending stiffness of the stack, which matters when you are running 2 m wide sheet at 30 m/min.

The first nip

The first nip gap is the sheet thickness plus a small allowance for the elastic recovery of the polymer after it leaves the nip. For amorphous resins like PS, ABS, PC and PMMA, recovery is 2–5%. For semi-crystalline resins like PP and PE, it is larger and less predictable because crystallisation is still happening as the sheet leaves the roll.

This is why PP sheet is so much harder to hold to gauge than PS. The sheet is not just springing back — it is shrinking as it crystallises, and the amount depends on how fast it cooled, which depends on roll temperature, which is exactly the thing you are trying to control.

Table 2 — Nip allowance by resin. The semi-crystalline resins are the reason two lines running the same nominal gauge will quote different tolerances.

Crown, bow and the fight against the middle

A roll is a beam. Under nip pressure it deflects, and the deflection is greatest at the centre. If you set a perfectly parallel nip with cold rolls, the sheet will be thicker in the middle than at the edges — often by 3–6% on a 1,500 mm wide stack.

Three things are used to fight this, and they are usually combined:

Roll crown. The roll is ground with a slight convex profile so that when it deflects under load it becomes straight. Crown amount is calculated from the nip load and the roll's bending stiffness, and it is fixed at manufacture — you cannot adjust it. A roll crowned for 2 mm sheet at 80 N/mm will be wrong at 6 mm sheet at 150 N/mm.

Cross-axis (roll bending). Hydraulic rams apply a bending moment to the roll ends, counteracting the nip deflection. This is adjustable during production, which is why it is the preferred method on modern lines. Most lines have two or four rams and a control loop driven by the thickness gauge.

Roll skew. Rotating the roll slightly in the horizontal plane so the nip is not parallel. This changes the gauge profile to a wedge — thicker on one edge. Skew is a crude tool but it is the fastest way to fix a one-sided gauge problem, and it is what an operator reaches for at 2 a.m.

Cooling: the part nobody budgets for

The roll stack does two jobs at once — it forms and it cools. How much cooling it delivers is set by the contact time and the roll temperature, and both are constrained.

Contact time is the wrap length divided by line speed. On a 400 mm roll with 180° wrap, the contact length is about 1.26 m. At 10 m/min that is 7.5 seconds of contact; at 30 m/min it is 2.5 seconds. That is not much for a 3 mm sheet, which is why most lines have additional cooling rollers or a cooling conveyor after the stack.

Table 3 — Roll temperatures. Note that the amorphous engineering resins want warm rolls and the polyolefins want cold ones. Getting this backwards is one of the most common commissioning errors.

Roll temperature control is usually by internal spiral baffle with water, sometimes oil for PC and PMMA. The temperature uniformity across the roll face matters as much as the set point — a 5 °C gradient from centre to edge will show up as a gloss difference you can see in reflected light.

Thickness control: manual, auto die, and the gauge

Three levels of thickness control exist, and the price difference between them is large.

Table 4 — Thickness control options. The step from manual to auto die is worth it above about 3,000 tonnes a year; below that, the scrap saving does not repay the capital.

The gauge itself is usually a beta or X-ray transmission sensor on a scanning frame, or increasingly a fixed array of sensors across the width. It measures total mass per unit area, which for a known density is thickness. It cannot tell you about a void or a gel, so it is always paired with visual inspection.

Where the gauge sits matters. Placed immediately after the roll stack it measures the sheet before it has fully relaxed, so it reads slightly thick. Placed after the cooling section it reads the true final thickness but the correction arrives late. Most lines put it after the first cooling roller and apply a dead-time compensation in the controller.

What each resin punishes you for

Sheet extrusion is a process where the resin dictates the line settings, and getting them wrong produces a specific, recognisable defect each time.

Table 5 — Resin settings and the failure each one is known for. The drying column is not optional for the hygroscopic resins; hydrolysis shows up as a drop in impact strength long before it shows as a visible defect.

Troubleshooting: gauge, surface and flatness

Table 6 — The nine faults that account for most sheet line downtime. Note that only two of them are roll stack problems; the rest are upstream.

The one that catches people out is curl after cutting. A sheet that is flat on the line and curls into a tube the next day is showing you that the two faces cooled at different rates. The causes are usually a 10 °C difference between top and bottom roll, or a sheet that wrapped the middle roll on one face only. It is worth checking before you blame the resin.

Specifying sheet: what to put on the drawing

If you are buying extruded sheet, the drawing should state: nominal thickness and tolerance band, width and edge condition (trimmed, square, or as-extruded), surface finish on each face (polished, matt, textured, with gloss value), flatness (typically 2 mm per metre for general purpose), and colour with a reference standard.

The tolerance band is where most of the cost sits. Moving from ±10% to ±5% is usually free — it just requires the supplier to have a gauge and to use it. Moving from ±5% to ±2% requires auto die control, and that is a different class of supplier. Ask which one you are actually being quoted.