What Thickness Tolerance Can You Achieve With a Modern Film Extrusion Line?

Every buyer of a film extrusion line eventually asks the same question, and it is the question that decides whether a project is profitable or permanently loss-making: what thickness tolerance can you actually achieve? The honest answer is uncomfortable, because the single printed number that sales brochures love to quote is almost always meaningless until you attach four qualifiers to it. You must know whether the figure is a percentage of nominal or an absolute micrometer value. You must know whether it refers to the transverse direction (TD, across the web) or the machine direction (MD, down the web). You must know the statistical basis, because a 2-sigma spread is not a 3-sigma spread, and neither is the full range between the thickest and thinnest point. And you must know the operating condition, because a tolerance measured on one perfect roll under factory acceptance test is not the tolerance your customer sees across a month of multi-shift production.

YuanSu, a Wanplas factory, has spent decades building plastic film, sheet and board extrusion lines where gauge control is the central engineering problem, not an afterthought. Our core leadership carries decades of experience in extrusion molding and polymer processing equipment, and our modern machining base and assembly workshop are organized around one goal: make the same thickness, roll after roll, shift after shift. As part of the Wanplas group, we draw on a network that exports to more than 100 regions and shares the group’s quality and service standards, while we focus the factory entirely on film, sheet and board extrusion technology. This article is written to settle the tolerance question with data rather than slogans. We explain what tolerance numbers actually mean, what a modern line can really deliver by process route, how the variation is measured, how the control loop closes it, and most importantly, why the tolerance a machine is capable of is rarely the tolerance a process delivers.

Defining Thickness Tolerance: Why the Number Means Nothing Without Context

Before any supplier can promise a tolerance, both sides have to agree on what “tolerance” is being measured. In film and sheet extrusion the term is used loosely, and that looseness is where most disputes begin. A number that looks identical on two quotations can mean two completely different things once you unpack the basis.

Percentage of nominal vs absolute micrometer value

A tolerance expressed as plus or minus a percentage of nominal thickness scales with the product. For a 20 micrometer stretch film, a plus-or-minus 2 percent tolerance equals plus-or-minus 0.4 micrometer. For a 2 millimeter sheet, the same plus-or-minus 2 percent equals plus-or-minus 40 micrometers. Both are “2 percent,” but the absolute burden on the die, the rolls and the gauge is an order of magnitude apart. Thin films are far harder to hold in absolute terms, so when you compare two lines always convert the claim to absolute micrometers at your real production thickness. A supplier who only speaks in percentages is hiding the harder half of the problem.

Two-sigma, three-sigma and the full range

Statistical process control describes spread with sigma bands. A 2-sigma band covers about 95 percent of readings; a 3-sigma band covers about 99.7 percent. The “range” is simply the maximum reading minus the minimum reading over the sampling window. These three numbers diverge sharply as variation becomes non-uniform. A web with a slowly drifting mean can show a tiny 2-sigma but a huge range, because the extremes sit far outside the central band. When a brochure states “tolerance plus-or-minus 3 percent” you must ask: is that plus-or-minus 2-sigma, plus-or-minus 3-sigma, or the observed range? Only plus-or-minus 3-sigma against a stable mean is a defensible production guarantee. Anything else is a best-case snapshot.

Transverse direction vs machine direction

TD variation is the profile across the width of the web at a single instant. MD variation is the change of average thickness along the length of the web as time passes. They have completely different root causes and different fixes. TD error comes from die lip geometry, the air ring or chill roll, and local temperature differences. MD error comes from extruder output drift, melt pressure pulsation, draw ratio change and line speed wobble. A line can be excellent in MD and poor in TD, or the reverse. Any tolerance claim that does not separate TD from MD is incomplete, and in practice the TD number is the one that decides whether your roll is usable edge to edge.

Edge zones and sampling discipline

The measured width matters. Real webs are thicker near the edges because of end effects in the die and the air ring, and because the gauge scanner often cannot reach the last 50 to 100 millimeters on each side. If a supplier computes tolerance only over the central usable width and discards the edge band, the reported number looks better than what your slitter actually yields. You should agree up front how many millimeters of edge are excluded, how many measurement points are taken across the width, and what the scan period is. A scanner that sweeps the width every 4 seconds catches a slow TD wander; one that sweeps every 20 seconds may average it into invisibility.

Rule of thumb: a tolerance claim is only valid if it states the percentage basis, the absolute micrometer value at your thickness, the sigma level, the TD and MD split, the excluded edge width, the sampling duration and the operating condition. Anything shorter is marketing, not engineering.

The Common Ways Tolerance Claims Get Inflated

Nowhere in plastic machinery is the gap between brochure and reality wider than in gauge tolerance. Knowing the standard inflation tricks is how a buyer protects a project. These are the patterns we repeatedly see, and each one quietly doubles or triples the apparent capability.

Reporting MD but hiding TD

Machine direction control is comparatively easy: stabilize extruder output, lock line speed, close the weight loop, and MD variation falls to 1 to 2 percent. Transverse control is the hard part. A supplier who quotes “plus-or-minus 2 percent” while only having validated the MD direction is offering you half a line. Always demand the TD figure as a separate, named number.

Quoting 2-sigma instead of range or 3-sigma

A 2-sigma spread of plus-or-minus 1.5 percent can balloon to a 3-sigma of plus-or-minus 2.5 percent and a full range of plus-or-minus 5 percent or more once you include the tails. If the contract is written against 2-sigma, the supplier can pass while your customer still rejects rolls for the occasional thick or thin spot. Insist the acceptance basis be 3-sigma or full range.

Excluding the edge 100 millimeters

As noted, discarding edge band before computing the profile removes the worst TD points. If your finished product uses the full width, the exclusion is dishonest. Agree the excluded edge width in writing and keep it small, typically 20 to 30 millimeters per side to clear the die lip land, not 100.

One perfect roll vs long-term multi-roll variation

A factory acceptance test on a single stable material, run by the builder’s best operator, in a climate-controlled plant, at one setpoint, is the machine capability. It is not your process capability. Real production means material changeovers, grade switches, multi-shift crews, ambient swings and wear. Ask for data over at least 8 to 24 hours and across more than one roll before you treat a number as real.

The “best roll” trap

Some quotations quietly quote the single best roll from a trial as the line’s capability. Capability is a statistical property of a population, never the best sample in it. The only honest capability number is computed from a continuous production window with enough points to be meaningful, typically thousands of gauge readings.

Achievable Tolerance Capability Matrix by Process Route

This is the core of the article. Below is the capability you can expect from a modern line, split by process route and by control configuration. Every figure is a 3-sigma basis unless stated, expressed as plus-or-minus percentage of nominal, with the absolute value implied at your thickness. These are realistic machine-capability targets on well-commissioned lines; your process capability will sit wider, which the later section explains.

Process route	Configuration basis	TD tolerance (3-sigma)	MD tolerance (3-sigma)	Typical Cpk target	Measurement basis
Blown film, single layer	Manual die gap, fixed air ring, no automatic control	plus-or-minus 4 to 8 percent	plus-or-minus 2 to 3 percent	1.0 to 1.33	Beta-ray or capacitive scan, manual correction
Blown film, single layer	Automatic air ring (12 to 72 zones) plus closed loop	plus-or-minus 2 to 4 percent	plus-or-minus 1.5 to 3 percent	1.33 to 1.67	Beta-ray or capacitive scan, automatic zone loop
Cast film, single layer	Standard die, manual lip adjust bolts	plus-or-minus 1.5 to 3 percent	plus-or-minus 1 to 2 percent	1.33	Beta-ray or X-ray scan, manual correction
Cast film, single layer	Auto die lip (thermal or servo bolts) plus beta or X-ray closed loop	plus-or-minus 0.8 to 1.5 percent	plus-or-minus 1 to 2 percent	1.33 to 2.0	Beta-ray or X-ray scan, automatic lip loop
Sheet, 0.2 to 3 mm	Standard die, manual adjust, precision calender rolls	plus-or-minus 1.5 to 3 percent	plus-or-minus 1.5 to 3 percent	1.33	Beta-ray, X-ray or laser, manual correction
Sheet, 0.2 to 3 mm	Auto die lip plus closed loop plus precision calender rolls	plus-or-minus 0.8 to 1.5 percent	plus-or-minus 0.8 to 1.5 percent	1.33 to 2.0	Beta-ray, X-ray or laser, automatic lip loop
Multi-layer co-extrusion (total thickness)	Co-extrusion die, per-layer melt pumps, closed loop on total	plus-or-minus 1.0 to 3 percent	plus-or-minus 1 to 2.5 percent	1.33	Beta-ray, X-ray or infrared, total gauge loop
Multi-layer co-extrusion (individual layer)	Per-layer melt pump plus flow-balanced manifold	plus-or-minus 5 to 15 percent per layer	plus-or-minus 5 to 15 percent per layer	1.0 to 1.33	Microtome cross-section or multi-wavelength IR (calibrated)

Reading the matrix correctly

Blown film is the hardest to hold in TD because the bubble is a free, oscillating film cooled by an air ring; even an automatic air ring with 12 to 72 zones rarely beats plus-or-minus 2 percent, and simple lines sit at plus-or-minus 4 to 8 percent. Cast film is easier because the melt is laid onto a temperature-controlled chill roll and a die lip can be adjusted locally; with an automatic lip and beta or X-ray loop you reach plus-or-minus 0.8 to 1.5 percent. Sheet behaves like a thick cast film and reaches similar numbers when the calender roll is precise. The critical row is the multi-layer one: holding total thickness tight does not mean each layer is tight. A five-layer barrier film can be plus-or-minus 1.5 percent on total gauge while the tie or barrier layer drifts plus-or-minus 10 percent, which is enough to fail a laminate or lose barrier function. Total thickness passing is not proof of layer thickness passing.

Rotation and lay-flat accumulation

In blown film, rotating the die or oscillating the haul-off averages TD error into a more uniform roll, but it does not remove the error; it spreads it circumferentially. On a lay-flat roll this shows as a slow helical thick-thin pattern. The matrix figures assume proper rotation; without it the TD number on any single point worsens even if the average looks acceptable. This is why “rotating die” must be a named line feature, not a silent assumption.

What Drives Variation: Six Layers, Quantified

Tolerance is the sum of every upstream disturbance. To tighten it you have to attack each layer. We quantify the typical contribution of each so you can prioritize spend.

Layer 1: The die

The die sets the baseline geometry of the web. Die lip gap uniformity, finished to micrometer grade, decides the minimum achievable TD spread. A lip finished to plus-or-minus 2 micrometers gives a cleaner profile than one finished to plus-or-minus 8 micrometers. The internal flow channel must distribute melt with equal residence time across the width; a poor manifold forces the operator to pre-bend the lip to compensate, which steals adjustment range. Die temperature zoning matters: a plus-or-minus 0.5 to 1 degree Celsius variation along the lip changes local viscosity and therefore local flow by a measurable few percent. Automatic die lips, using thermal-expansion bolts or servo actuators, give resolution of roughly 1 to 5 micrometers of gap change and respond in seconds, which is what makes plus-or-minus 1 percent TD possible on cast lines.

Layer 2: Melt stability

The extruder must deliver a steady, pulsation-free melt. Screw metering zone design and consistent back pressure keep output steady. Without a melt pump, pressure at the die swings with screen pack clogging and screw beat; typical pressure pulsation is plus-or-minus 3 percent, which translates directly into MD thickness variation. A gear-type melt pump removes that pulsation, cutting pressure variation to about plus-or-minus 0.3 percent and greatly tightening MD. Melt temperature uniformity and the changing pressure drop across the screen changer are the other two contributors; a clogged filter that the operator has not swapped is a classic, slow MD drift source.

Layer 3: Material

Resin is never perfectly constant. Melt flow rate batch-to-batch differences change output per revolution. Regrind blend ratio variation changes both viscosity and shrinkage. Moisture causes foaming and local thinning; masterbatch dispersion unevenness causes local structure changes that some gauges misread as thickness. Melt strength and bubble stability dominate blown film: a resin with poor bubble stability oscillates, widening TD beyond anything the air ring can correct. Specifying a tight resin spec and holding the regrind ratio constant is cheaper than buying a better gauge.

Layer 4: Draw and cooling

For blown film, air ring airflow and air temperature uniformity set the freeze line. An automatic air ring with 12 to 72 independently controlled zones can reshape the bubble profile; more zones means finer correction, but only if matched to die lip segment count. Frost line height stability keeps MD consistent. For cast film, the air knife, vacuum box and chill roll do the work: chill roll temperature must hold within plus-or-minus 0.5 degree Celsius across its face, and the roll itself must be ground true to micrometer-class roundness and run-out, because any roll eccentricity prints directly onto the sheet as a periodic MD thickness wave.

Layer 5: Haul-off and winding

Line speed must be steady to within about plus-or-minus 0.1 percent, because any speed wobble is an instant MD thickness wave through the draw ratio. Tension closed-loop control prevents the web from stretching or relaxing as roll diameter grows; without roll-diameter compensation the winding tension drifts and the MD gauge follows. The draw ratio itself, the ratio of line speed to extrusion speed, must be held constant, which ties back to both extruder output and haul-off speed being stable together.

Layer 6: Environment

Blown film is sensitive to plant air currents; a draft across the bubble bends it and shifts TD. Day-to-night ambient temperature changes alter resin viscosity and roll dimensions. Floor vibration couples into thin-gauge lines. These are why a quoted factory acceptance tolerance, measured in a still, climate-controlled bay, rarely survives a draughty production hall without attention to plant layout.

Disturbance layer	Typical raw effect if uncontrolled	Primary control
Die lip and manifold	TD plus-or-minus 3 to 8 percent	Micrometer-finished lip, auto lip actuators, zoned heating
Melt pulsation	MD plus-or-minus 3 percent pressure swing	Gear melt pump, stable back pressure
Resin and regrind	Plus-or-minus 1 to 3 percent output shift	Tight MFR spec, fixed regrind ratio
Cooling and draw	TD plus-or-minus 2 to 5 percent (blown)	Automatic air ring, true chill roll, stable frost line
Haul-off and wind	MD plus-or-minus 1 to 2 percent	Speed loop plus-or-minus 0.1 percent, tension and diameter comp
Environment	Slow drift, hard to number	Draft control, climate stability, isolated foundation

Measurement Methodology: How You Actually Know the Thickness

You cannot control what you cannot measure, and gauge measurement is its own discipline. The sensor type decides accuracy, material sensitivity and cost. Below is the honest comparison.

Sensor principle	Accuracy basis	Material and thickness fit	Strengths	Weaknesses
Beta-ray gauge	Radiation absorption, density-compensated	Most films and sheets, wide thickness range	Non-contact, stable, industry standard for closed loop	Needs radioactive source license and shielding, insensitive to thin layers
X-ray gauge	X-ray absorption, density-compensated	Most films and sheets, wide range	Higher signal than beta, good for thicker sheet	Needs source license, more costly, still layer-blind
Infrared absorption gauge	Selective absorption by polymer and additives	Single-polymer films; multi-wavelength for co-extrusion	Can resolve layer structure with multi-wavelength	Strongly material and additive sensitive, needs calibration per recipe
Capacitive sensor	Dielectric thickness, non-contact	Thin films, clean polymers	Low cost, fast, simple	Sensitive to moisture and additive change, not for all polymers
Laser triangulation	Distance to surface, geometric	Sheet and thick film, opaque webs	Direct, layer-independent, good on sheet	Needs two-sided access or reference roll, surface finish dependent

Scanning vs fixed-array

A scanning gauge sweeps across the width on a rail, giving a full TD profile at the cost of a scan period, typically a few seconds. A fixed-array places many sensors across the width and reads all points simultaneously, catching fast TD events the scanner would miss, but at far higher sensor cost and with more calibration points. For most film lines a single scanning beta or X-ray head is the right balance; for high-speed thin film where TD events are fast, a fixed array earns its cost.

Offline and arbitration methods

Online gauges must be checked against truth. A contact thickness gauge with 0.1 micrometer resolution is the basic offline check. The most reliable arbitration is the weigh method: cut a known area, weigh it, divide by area and density to get average thickness. This is the reference most disputes are settled by, because it is independent of sensor physics. For multi-layer films, a microtome cross-section under a microscope is the only way to arbitrate individual layer thickness, since no online sensor cleanly separates layers without careful calibration.

Calibration and traceability

Every gauge drifts with temperature and ages with use. A standard shim calibration cycle, typically weekly to monthly depending on duty, keeps it honest. Two silent error sources deserve care: temperature drift of the sensor, and density input error. If you change resin grade and forget to update the density the gauge uses, every reading shifts. For infrared, an uncalibrated recipe change can move the number by several percent while the real thickness has not moved at all.

The Closed-Loop Control Chain

A gauge alone does nothing. Control happens when the measured profile is compared with the target and the difference is pushed back into actuators. The chain is: measure thickness across the width, build the actual profile, subtract the target profile, compute the correction per zone, drive the die lip bolts or air ring zones, and iterate until the error converges.

TD loop details

The number of control zones must match the die lip segment count. A 40-zone die driven by a 12-zone controller leaves most of the width uncorrected; the unused actuators are dead money. The dead time, the transport delay from die to gauge, sets how often the loop can safely act. Act too fast against a delayed signal and the loop oscillates, overcorrecting thick to thin and back. The control period is therefore tuned to the dead time, and the gain is set conservatively to avoid hunting. A well-tuned loop reaches target tolerance in roughly 15 to 45 minutes after a setpoint or material change, which is also where start-up scrap comes from.

MD loop details

Machine direction is closed on weight, the grams per square meter, using the average gauge plus line speed. When average thickness drifts, the loop trims extruder output or line speed to bring grams per square meter back. This is why a melt pump plus a good weight loop is the cheapest route to tight MD; the pump removes the pulsation and the loop removes the slow drift.

Convergence and start-up scrap

Because the TD loop must wait out the dead time and integrate the error, the first rolls after a change are off-spec. Budget for this: a line changing grade several times a day spends a real fraction of its output in convergence. Faster, better-tuned loops shrink this window, which is a direct yield gain, not just a quality gain.

Machine Capability (Cpk) vs Process Capability

This is the most valuable distinction in the whole article, and the one most often ignored. A line has two different tolerances, and conflating them is the source of most disappointed buyers.

Machine capability

Machine capability is what the hardware delivers under ideal, repeatable conditions: one material, one setpoint, stable environment, a skilled operator, short window. It is usually expressed as Cpk, the centered process capability index. A line demonstrated at Cpk 2.0 under factory acceptance test is genuinely excellent hardware. The supplier is not lying when they quote it.

Process capability

Process capability is what you get in production: material changes, grade switches, multi-shift crews with different habits, ambient day-night swings, screen pack changes, wear, and the occasional rushed changeover. The same line that showed Cpk 2.0 at acceptance may run at Cpk 1.1 in your plant. That is not fraud; it is the difference between a machine and a process.

Cause of the gap	Typical Cpk drop	What closes it
Material batch and regrind variation	2.0 down to 1.3	Tight resin spec, fixed regrind ratio, incoming check
Multi-shift operator difference	1.3 down to 1.2	Recipe management, SPC dashboard, training
Ambient and draft swing	1.3 down to 1.15	Plant climate and draft control
Screen pack and wear neglect	1.3 down to 1.1	Planned maintenance, pump protection
Poor changeover discipline	1.3 down to 1.1	Standard changeover, fast loop tuning

How to write the acceptance clause

Protect yourself in the contract. State the resin grade and batch window, the nominal thickness and width, the statistical basis (3-sigma or range, not 2-sigma), the sampling duration (minimum 8 hours, preferably 24), the excluded edge width, and the达标 window as a Cpk floor such as 1.33. Require the demonstration on your actual material if possible, not the builder’s favorite grade. Require the FAT report to include the raw gauge log, not just a summary number, so you can recompute the basis yourself.

The Economic Value of Tightening Tolerance

The reason to spend on gauge control is money, and the money is in resin, not in the hardware. We express this only in percentages and indices, never in absolute currency.

Downgauging from a tighter tolerance

If your process runs at plus-or-minus 6 percent, you must set the nominal thickness high enough that even the thin tail stays above the customer’s floor, so you carry extra resin across the whole roll. Pull the tolerance to plus-or-minus 3 percent and you can lower the nominal thickness by a corresponding percentage while still clearing the floor. That nominal reduction flows straight into lower resin consumption and lower unit energy per square meter, because you are melting and chilling less polymer. The material cost index, set at 100 points baseline for the loose-tolerance case, drops toward roughly 94 to 97 points when the tolerance is halved, a recurring saving on every meter produced.

Complaints, returns and yield

Tight, symmetric gauge cuts customer complaints and returns, because the off-spec tail that triggers rejection shrinks. Roll utilization rises because less edge and less off-weight length is scrapped. Downstream thermoforming and lamination yield improves, because a uniform web forms and bonds uniformly instead of splitting or blistering at thin spots. Each of these moves the effective cost index down further without touching the machine price.

Where the spend pays back

The payback order is usually: stabilize the melt with a pump (low cost, high MD gain), add a real closed loop on a good gauge (medium cost, high TD gain), then tighten the die and rolls (higher cost, marginal further gain). Buying the gauge without closing the loop, or closing the loop on a poorly finished die, wastes the spend. The earlier sections explain why.

Standards and Reference Methods

Several published standards define how thickness is measured and how capability is judged. Cite them in your specification so both parties share a method. ASTM D6988 covers guidance for expressing film thickness measurement data. ISO 4593 specifies the mechanical scanning method for plastics film and sheeting thickness. ASTM D374 covers thickness of solid electrical insulation and related plastic sheet by mechanical means. ISO 2286 relates to rubber- or plastics-coated fabrics and the determination of roll characteristics including mass per unit area, useful alongside the weigh method. GB/T 6672 is the Chinese national standard for plastic film and sheet thickness by mechanical scanning, widely used in Asian supply chains. For capability and control charts, the concepts in ISO 7870 (control charts) and ISO 22514 (statistical methods in process management) give the framework for Cpk and SPC discipline. CE marking applies to the machinery’s electrical and safety design and is a plain compliance point, not a gauge guarantee.

Common Mistakes That Waste the Investment

We close the technical part with the errors that quietly destroy gauge projects, drawn from lines we have been asked to fix.

Buying the gauge but not the loop

A beta or X-ray scanner that only displays a profile, with no actuator link, tells you you are wrong but does nothing about it. The operator still walks to the die. You paid for visibility and got no control. The loop and the actuators are the value, not the sensor.

Control zones far fewer than actuators

A 60-bolt automatic die driven by a 16-zone controller cannot correct most of the width. Match zone count to actuator count, or the unused bolts are decoration.

Ignoring the dead time

Tuning the loop faster than the die-to-gauge transport delay causes oscillation. The line hunts between thick and thin and the average looks fine while the range is terrible. Respect the dead time in the tuning.

Using infrared on multi-layer without calibration

An infrared gauge reading a co-extruded film with a recipe it was not calibrated for reports a thickness that drifts with layer ratio, not reality. Calibrate per recipe, or use beta or X-ray for total gauge and microtome for layer truth.

Quoting the best roll as capability

We said it earlier and repeat it: capability is a population statistic. One excellent roll is a sample, not a specification. Demand the continuous log.

Tolerance Requirements by Application

Different end products tolerate different gauge error. Over-specifying wastes money; under-specifying loses the account. The table maps typical needs to a sensible target so you can size the line honestly.

Application	Why gauge matters	Typical TD target	Recommended control tier
Food packaging film	Seal uniformity, downgauge cost	plus-or-minus 2 to 4 percent	Auto die lip or air ring, beta or X-ray loop
Protective and surface film	Even adhesion, no thin tear	plus-or-minus 3 to 5 percent	Standard die, manual or basic loop
Agricultural film	Strength at thin spots, cost	plus-or-minus 4 to 8 percent	Automatic air ring on blown film
Thermoforming sheet	Uniform wall, no thin-out	plus-or-minus 1.5 to 3 percent	Auto die lip, beta or laser loop
Optical and electronic film	Clarity, coating uniformity	plus-or-minus 0.8 to 1.5 percent	Auto die lip, X-ray or laser, tight loop
Lithium-battery related film	Layer uniformity, safety	plus-or-minus 1 to 3 percent total, tight per layer	Per-layer pump, IR or microtome layer check

YuanSu Production Lines Built for Tight Gauge

YuanSu builds the full range of film, sheet and board extrusion lines, and gauge control is designed in from the die upward. The two lines below are representative of our film and sheet programs; both are available as multi-layer co-extrusion configurations with online measurement and closed-loop control as options.

CPP / CPE / EVA Casting Film Extrusion Line

Our cast film line uses a co-extrusion-capable die with micrometer-finished lips and zoned temperature control. With the optional automatic lip and a beta-ray or X-ray scanner closed loop, the line reaches the plus-or-minus 0.8 to 1.5 percent TD band in the matrix above. Winding speed reaches up to 600 meters per minute on suitable grades. Multi-layer co-extrusion and online measurement are standard options.

Parameter	Representative configuration
Web width	1000 to 3000 mm
Thickness range	15 to 200 micrometers
Achievable tolerance	TD plus-or-minus 0.8 to 1.5 percent (auto lip plus closed loop); MD plus-or-minus 1 to 2 percent
Output	200 to 800 kg/h, by width and layer count
Line speed	up to 600 m/min on thin grades
Screw diameter and L/D	65 to 150 mm, L/D 30 to 33
Installed power	150 to 400 kW, by configuration
Gauge and closed loop	Beta-ray or X-ray scanner, scanning or fixed array; automatic die lip, 40 to 120 zones; grams-per-square-meter MD loop

PET / PE / PP / PS Stretch Film Extrusion Line

This line targets stretch and surface film where MD stability and symmetric gauge decide whether the roll performs. A gear melt pump removes extruder pulsation, and the optional automatic lip with beta-ray or capacitive scanning holds TD in the plus-or-minus 1.0 to 2.0 percent band. Multi-layer co-extrusion is supported for cling and core layers. The profile’s stated film thickness tolerance of plus-or-minus 2 percent reflects this family’s capability.

Parameter	Representative configuration
Web width	500 to 2500 mm
Thickness range	8 to 50 micrometers
Achievable tolerance	TD plus-or-minus 1.0 to 2.0 percent; MD plus-or-minus 1.5 to 2.5 percent
Output	150 to 600 kg/h
Line speed	up to 600 m/min
Screw diameter and L/D	55 to 120 mm, L/D 30 to 33
Installed power	120 to 350 kW
Gauge and closed loop	Capacitive or beta-ray scanner; automatic die lip; gear melt pump for MD stability

PET / GAG / PLA Sheet Extrusion Line

For thermoforming and packaging sheet, flatness and symmetric gauge decide forming yield. This line pairs the co-extrusion die with precision calender rolls ground to micrometer roundness, and the optional beta-ray, X-ray or laser gauge with an automatic lip loop holds plus-or-minus 0.8 to 1.5 percent with flatness at or below 0.1 mm per meter, matching the sheet technical claim in our program.

Parameter	Representative configuration
Web width	600 to 2000 mm
Thickness range	0.25 to 2 mm
Achievable tolerance	plus-or-minus 0.8 to 1.5 percent (auto lip plus closed loop plus precision roll)
Flatness	at or below 0.1 mm per meter
Output	300 to 1000 kg/h
Line speed	5 to 60 m/min
Screw diameter and L/D	75 to 150 mm, L/D 30 to 35
Installed power	200 to 600 kW
Gauge and closed loop	Beta-ray, X-ray or laser; automatic die lip; roll-gap trim; online measurement standard

Selection Guide: Requirement to Model

Matching the product to the line and the control tier prevents both overspend and rejects. Use the table below as a starting point; final configuration is confirmed against your resin and target.

Product type	Thickness band	Tolerance need	Width	Recommended model	Gauge and loop tier
Cast packaging film	15 to 80 micrometers	plus-or-minus 1 to 2 percent TD	1000 to 2500 mm	CPP / CPE / EVA Casting Film Line	Beta or X-ray, auto die lip, MD weight loop
Stretch and cling film	8 to 30 micrometers	plus-or-minus 1.5 to 2.5 percent	500 to 2000 mm	PET / PE / PP / PS Stretch Film Line	Capacitive or beta, auto lip, gear pump
Functional film (TPU/PVB/POE/EVA)	20 to 250 micrometers	plus-or-minus 1 to 2 percent	1000 to 2400 mm	TPU / PVB / POE / EVA Film Line	Beta or X-ray, auto lip, multi-layer pump
Thermoforming sheet	0.3 to 2 mm	plus-or-minus 1 to 1.5 percent	600 to 1500 mm	PET / GAG / PLA Sheet Line	Beta or laser, auto lip, precision roll
Barrier co-extruded film	20 to 120 micrometers	tight total, watch layers	1000 to 2200 mm	Multi-layer cast or stretch configuration	Per-layer pump, IR or microtome layer check
Lithium-battery related film	10 to 60 micrometers	plus-or-minus 1 to 3 percent total	800 to 1600 mm	Functional film co-extrusion line	Per-layer pump, X-ray total, layer calibration

Service, FAT and Lifecycle Support

YuanSu delivers more than hardware. Every line is run at our plant before shipment with a factory acceptance test that includes a live gauge capability measurement, and you receive the FAT report with the raw thickness log, not just a summary number, so the tolerance is provable rather than promised. Our engineers handle on-site installation and commissioning, and we provide the formula and process expertise, including raw material formulation and parameter optimization, that decides whether machine capability becomes process capability on your floor.

As part of the Wanplas group, the line carries the shared service promise of USD 500 free parts per year, warranty replacement of damaged parts, and an open-factory policy: you are welcome to visit, watch a line run, and see gauge tolerance measured on real material before you commit. We train your operators in process and statistical process control discipline, because the Cpk 2.0 machine only stays near Cpk 1.33 in your plant if your team runs the loop and reads the charts. Remote monitoring and lifetime consultation keep the line tuned after handover, and our turnkey scope covers factory planning, utility design, commissioning and mass-production support.

Frequently Asked Questions

Conclusion

Thickness tolerance on a modern film extrusion line is not a single number; it is a profile defined by process route, control configuration, statistical basis and operating condition. A modern cast film line with an automatic die lip and a beta-ray or X-ray closed loop reaches about plus-or-minus 0.8 to 1.5 percent transverse on a 3-sigma basis, a blown film line with an automatic air ring reaches about plus-or-minus 2 to 4 percent, and a precision sheet line reaches a similar plus-or-minus 0.8 to 1.5 percent with flatness at or below 0.1 mm per meter. The harder truth is that machine capability, often Cpk around 2.0 at acceptance, is not process capability, which in daily production commonly settles near Cpk 1.1 to 1.3; closing that gap is mostly about material discipline, loop tuning and operator training, not more steel.

YuanSu, a Wanplas factory, designs gauge control in from the die upward across our cast film, stretch film, functional film and sheet programs, with multi-layer co-extrusion, online measurement and closed-loop options available on each. If you are specifying a line and want the tolerance written as a provable process capability rather than a brochure claim, send us your resin, target thickness, width and required Cpk, and we will propose a configuration, run it at our plant with a live gauge measurement you can witness, and deliver it with the FAT report and the lifecycle support that keeps the tolerance real on your floor. We welcome you to our factory to see the numbers measured, not just quoted.