Nonwoven Defect Attribution: Roller vs Process Guide

+86-19322109200

team3@huataogroup.com

language

English

Website Group

HUATAO GROUP

News & Events

We Help Customers To Meet The Great Challenges Of Our Time.

VIEW PRODUCTS

scroll down

Home

Back to Overview

Roller or Process? The Complete Engineering Guide to Nonwoven Defect Attribution

Nov 11,2026

Every nonwoven plant eventually faces the same question in a quality meeting: the fabric is out of specification, the recipe has not changed, and the calender roll was inspected last shutdown. Is the roll guilty, or is the process?

The answer is rarely a matter of opinion, and it almost never requires a shutdown to establish. It requires three measurements, one list, and the discipline to change a single variable at a time. This guide sets out the full method, the physics behind it, the specification data that makes it possible, and the maintenance thresholds that keep it working.

HUATAO has manufactured industrial rollers since 2008 for nonwoven, paper, film and composite producers, and the diagnostic patterns below are drawn from the engineering support work that accompanies those installations. The methods are process-neutral: they apply whether the roll in question came from us or from another supplier.

Quick Answer

Roller faults are geometric or thermal and repeat at a fixed machine-direction interval equal to a roll circumference or a simple fraction of it. Process faults vary in spacing, follow forming or pumping cycles, and respond to setpoint changes. Measure defect spacing, compare it with every roll circumference in the line, confirm with hot runout and cross-face thermal data, then settle the question with a controlled swap or a single-variable change. If spacing matches nothing, the cause is upstream of the nip.

10 Core Keywords

nonwoven defect attribution
calender roll runout
defect periodicity analysis
thermal bonding window
basis weight coefficient of variation
nip impression and crown
induction heated calender roll
spunbond line troubleshooting
spunlace drum perforation
roller reconditioning threshold

Two Failure Families, One Decision Point

Nonwoven defects sort into two families that behave differently under measurement.

Family one: geometry and surface. A shell that is not round, a crown that does not match the load, an engraving that has lost land width, a chrome surface with pits, a bearing with radial play. These produce defects that are stable in position, stable in period, and indifferent to setpoint changes. They are properties of hardware.

Family two: material and energy. Polymer lot variation, fibre blend drift, spin finish, quench air imbalance, melt pressure fluctuation, jet pressure instability, tension variation, dwell time changes from speed. These produce defects that move, vary in spacing, and respond when a parameter changes. They are properties of the process.

The decision point is the nip, where both families express themselves on the same fabric. That is why the debate is so persistent: the evidence appears in one place while the causes originate in two different places.

The Seven-Step Attribution Protocol

The protocol below fits inside one shift and can be run without stopping production. It assumes a vision system or at least a winding-station inspector, a steel rule, an infrared thermometer or camera, and access to roll drawings.

Step 1: Characterise the defect

Pull three metres of defective web and lay it flat on an inspection table under even lighting. Identify the smallest repeating unit and mark ten consecutive repeats from a fixed reference. Measure the total distance and divide by ten.

Record three attributes: mean spacing in millimetres, cross-direction position (fixed lane, walking lane, full width), and whether the defect is visible on both surfaces or only one. One-sided defects usually indicate contact with a single roll; two-sided defects suggest a through-thickness effect such as bonding or a nip.

Step 2: Build the roll list

List every rotating element in the line with diameter and circumference. Include calender rolls, anvil rolls, guide rolls, spreader rolls, suction drums, doffers, cooling rolls, and pull rolls. A 300 mm roll is 942 mm around; 450 mm is 1,414 mm; 600 mm is 1,885 mm.

Divide the measured spacing by each circumference. A quotient of 1.00 is conclusive. A quotient of 0.50 indicates a two-lobe error. A quotient of 0.33 indicates a three-lobe error or a three-start engraving. A quotient that is close but not exact suggests slip, and should be re-measured across twenty repeats.

Step 3: Read the periodicity data

If an inline inspection system is installed, filter the defect log by type and cross-direction position, then plot a histogram of machine-direction intervals. A sharp peak indicates periodicity; a flat distribution indicates randomness. Most systems will also flag "periodic defects" directly, which is the single most useful feature for this purpose.

A vision report does something else that manual inspection cannot: it records defects that the operator missed, including low-contrast ones. Those extra data points frequently reveal a second, lower-amplitude periodicity hiding beneath the obvious defect — a common situation when a surface fault and a thermal fault coexist.

Step 4: Measure runout at operating temperature

Mount a dial indicator on a rigid stand and measure total indicator reading at four axial stations, with the roll at production temperature and steady state. For a large shell, steady state means at least 45 minutes after reaching setpoint, because the thermal time constant of a 400 mm shell is typically 15 to 25 minutes.

Acceptance depends on the application. For a bonding calender, 0.02 mm is a common working limit; for a chill roll in film service, tighter values apply. What matters is consistency: the same measurement method, the same stations, the same temperature, recorded against the roll serial number every time.

Step 5: Map thermal uniformity

Scan the roll face with an infrared camera at low rotational speed. Record maximum cross-face deviation, axial gradient, and the location and width of any cold or hot sector.

On an Induction Calender Roll, a deviation above 2 °C usually has one of three causes: rotary joint or siphon condition limiting flow, scale or fouling in internal channels, or an imbalance between induction zones. The first two produce a stationary cold sector; the third produces a gradient that follows the coil layout. Distinguishing them takes one additional data point — the flow rate through the roll.

Step 6: Run a controlled change or a swap

If a spare roll is available, swap it and run the identical recipe for twenty minutes. This is the highest-value diagnostic action available and the one most often postponed.

If no spare exists, change one variable by a known amount and observe. Nip load by 10% is the safest first move, because it can be reversed without quality risk. Temperature by 5 °C is the second, but the thermal time constant demands a twenty-minute wait before evaluating. Roll faults are indifferent to both; process faults are not.

Step 7: Document the verdict

Write down the spacing, the matched circumference, the runout values, the thermal map result, the change made, and the outcome. Attach the fabric sample with the spacing marked. This record is the difference between an organisation that learns and one that repeats the same investigation every quarter.

ALT: Induction heated calender roll prepared for a runout and thermal uniformity check

Defect Signature Reference

The table below maps common observations to their most probable causes. It is a starting point, not a verdict; each row assumes the previous steps have been completed.

Signature	Period behaviour	Cross-direction behaviour	Probable cause	Confirmatory test
Bands at one circumference	Exact, stable	Fixed lane or full width	Runout, local surface damage	TIR at temperature
Bands at half circumference	Exact, stable	Fixed lane	Oval bore, two-start engraving	Roundness measurement
Bands at third circumference	Exact, stable	Fixed lane	Three-lobe error, three-start pattern	Engraving pitch check
Diagonal streaks	Scales with speed	Walks across width	Bearing play, orbital error	Vibration spectrum
Edge-only weak bonding	Not periodic	Symmetric at both edges	Crown or parallelism error	Impression film
Centre-only weak bonding	Not periodic	Symmetric at centre	Over-crown, excessive load	Impression film at reduced load
Patchy weak spots	Random	Random	Basis weight variation, forming	CV at five positions
Gradual strength loss	Not periodic	Full width	Roll face temperature drift	Infrared face scan
Repeating creases	Exact, matches guide roll	Fixed offset	Guide roll drag, tension, spreader	Bearing drag check
Isolated gloss marks	Random or semi-periodic	Fixed lane	Polymer buildup or chrome pit	Surface inspection at angle

Process Parameters and Their Defect Signatures

Understanding what each process parameter does to the fabric is what allows a process fault to be recognised without a swap test.

Melt pressure. Variation beyond roughly ±0.5 MPa widens filament diameter distribution. Coarse filaments bond poorly at unchanged setpoint, producing weak spots that appear across the full width simultaneously. Signature: simultaneous onset across the width, no periodicity, recovery when pressure is stabilised.

Quench or drawing air temperature. A drift of more than about 1 °C changes filament cooling and therefore crystallinity, which changes bonding behaviour. Signature: gradual onset, full width, often accompanied by a change in hand-feel before any visible defect appears.

Jet pressure in spunlace. Fluctuation of ±2 bar changes entanglement energy. Signature: stripes of low tensile strength, often aligned with drum rotation visually and therefore misattributed to the Spunlace Drum. A one-second-interval pressure log resolves the question in twenty minutes.

Nip load. Increasing load widens contact and improves bonding until the point where the shell deflects enough to reduce edge pressure. Signature: improvement in the centre with deterioration at the edges, which is often misread as a crown problem that "suddenly appeared".

Line speed. Raising speed reduces dwell time in the nip. A 20% speed increase cuts bonding energy proportionally unless temperature is adjusted. Signature: uniform loss of bond strength at higher speed, recovering at lower speed — a process behaviour that is sometimes mistaken for a roll that "only misbehaves at speed".

Tension. Tension imbalance across the width causes tracking drift and diagonal wrinkles that press into the web at the calender. Signature: creases that repeat at a guide roll circumference, not the bonding roll. Checking bearing drag on guide rolls is a five-minute job that resolves a surprising number of these cases; the Chrome Drafting Roll specification notes cover the surface and tolerance requirements that prevent drag-induced marking in the first place.

Thermal Dynamics and the Attribution Error They Cause

Heated rolls are first-order thermal systems. A step change in heating medium temperature produces roughly 63% of the final surface change within one time constant and 95% within three. For a 400 mm shell with a 25 mm wall, that constant is typically 12 to 20 minutes.

Operators adjust on a much faster cycle. The result is a systematic attribution error: three changes are made within ten minutes, evaluation happens at twenty minutes, and the credit goes to whichever change the observer expected to work. This is not a discipline failure unique to any plant; it is a consequence of a thermal system whose response is slower than human patience.

Two procedural fixes cost nothing. First, impose a minimum dwell of twenty minutes after any temperature change before drawing a conclusion. Second, log the timestamp of each change next to the timestamp of each observation, so the sequence is visible in retrospect.

A third fix costs a little: install a surface temperature sensor at the exit of the nip. It closes the loop faster than a wall-mounted thermocouple and gives the operator something to watch during the lag.

Instrumentation Worth Having

A plant does not need a metrology lab to run this protocol. The minimum viable set is short.

A steel rule with 1 mm graduations and a flat inspection table.
A dial indicator with 0.001 mm resolution and a magnetic base.
An infrared thermometer with a known emissivity setting, or preferably a thermal camera.
Impression film or lead foil in 0.05 mm and 0.1 mm thickness.
A vibration meter or accelerometer with spectrum output.
Access to roll drawings and a per-roll record file.

The thermal camera is the largest single improvement in diagnostic speed, because it converts a two-hour thermocouple survey into a one-minute scan. The per-roll record file is the cheapest, because it converts every future argument into a comparison.

Specification Data That Makes Rolls Diagnosable

A roll that cannot be measured cannot be diagnosed, and a roll that arrives without records cannot be measured against a baseline. Five data items should accompany every roll for a nonwoven line.

Runout at operating temperature. Not only cold. State the measurement stations and the temperature at which the value applies.

Crown value with method. Crown without a measurement method is an assertion. Specify whether it is verified by impression film, by a crown gauge, or by diameter measurement at defined stations.

Thermal uniformity for heated rolls. Maximum cross-face deviation at a defined heat-up condition, plus the axial profile at three positions.

Surface specification. For bonding rolls: roughness average plus peak density, and coating thickness with hardness. For engraved rolls: pattern depth, land width, and pitch with tolerances.

Reconditioning limits. The manufacturer's stated limits for regrinding, replating, and replacement, expressed as measurable values rather than as years of service.

Buyers who request these five items get a different class of hardware, and more importantly they get the ability to attribute a future defect in a shift instead of a month. The specification depth described for Spunbond Nonwoven Rolls and Spunlace Nonwoven Rolls is a reasonable benchmark for comparing quotations from any source.

Maintenance Programme and Threshold Logic

Thresholds should come from the process window, not from habit. A defensible method: determine the maximum runout that still bonds acceptably at the lowest load in the product range, then set the maintenance limit at 70% of that value to leave margin for thermal growth. On many spunbond bonding rolls this lands between 0.015 and 0.020 mm.

Crown limits derive from contact width uniformity: if impression film shows contact width varying by more than 15% across the face at production load, crown or parallelism is out of specification for that product mix.

Coating limits are best expressed as defect frequency: when pits or release marks appear more than once per 1,000 m, replating is due. Thickness measurement alone misses localised failure and is the reason many plants discover coating breakdown from fabric rejects rather than from inspection.

A workable routine for critical bonding rolls:

Interval	Activity	Record
Every shutdown over 4 h	Surface inspection under oblique light, cleaning verification	Photo, roughness check
Monthly	TIR at operating temperature, four stations	Values per station, angular position
Quarterly	Infrared face scan, axial profile	Thermal map file
After bearing or frame work	Nip impression at production load	Contact width trace
Annual	Full dimensional review against drawing	Report with recommendations

Industry Context: Why This Matters More Than It Did

Three current developments make attribution discipline more valuable than it was five years ago.

Line speeds are rising. ITMA Asia + CITME 2026 opens in Shanghai on 20–24 November with 180,000 m² of exhibit space and more than 1,900 exhibitors, including a nonwoven process zone and new digitalisation and composites sections. Higher speed shortens nip dwell time, which narrows the tolerance band for runout, crown and thermal uniformity. A roll that performed acceptably at 250 m/min may not at 320 m/min. Source

Quality expectations are rising with the market. IndexBox's spunlace bonded nonwoven forecast to 2035 projects roughly 6.8% baseline CAGR, with electronic-grade and cleanroom grades at 7–9% and qualification cycles of 12–18 months. In that environment a defect that reaches a customer costs far more than the diagnosis would have. Source

Component quality is being recognised as an uptime variable. A recent technical note from a hydroentanglement screen supplier describes the "small component paradox": the cheapest component in the line can cause the most expensive downtime, and operators compensate for poor component performance by raising pump pressure, which accelerates wear and creates irregular density lines. The same compensation loop applies to calender rolls — temperature and load are raised to force bonding, and the defect changes shape rather than disappearing. Source

Case Study: Two Faults at One Frequency

A spunbond producer running 3.2 m wide fabric reported bond bands every 1,410 mm with reduced tensile strength in the banded lanes. The assumption was a worn calender surface, and a regrind was scheduled for the next shutdown.

Before removal, the maintenance team completed the protocol. The roll list showed a 450 mm bonding roll with a 1,414 mm circumference, matching the observed spacing within measurement error. Runout at 190 °C measured 0.031 mm, well above the 0.02 mm limit, confirming the surface as a genuine contributor.

The roll was reground to a target of 0.015 mm TIR at temperature, with crown verified by impression film. After reinstallation, band amplitude fell by roughly 80% — but did not vanish. A second spacing measurement showed the same period at much lower amplitude, and an infrared scan found a 3.1 °C cold sector across a 300 mm band of the face.

The remaining cause was scale in the internal channels, reducing flow in one axial region. Descaling and a flow test removed the residual band entirely. Two independent faults had been producing defects at the same frequency, which is why the first fix appeared partially effective and why a second measurement was essential.

The transferable lesson is that a confirmed match between spacing and circumference identifies the roll, but does not identify which property of the roll is at fault. Surface, geometry and thermal condition each need their own measurement.

ALT: Nonwoven fabric surface inspected for periodic defects after thermal bonding

Frequently Asked Questions

Can a process fault produce a spacing that exactly matches a roll circumference?

Yes, if a driven element upstream has the same circumference as the roll being blamed, or if the process cycle happens to coincide. This is why the roll list must include every rotating element, and why cross-direction position is used as a second discriminator.

How do I handle a defect that only appears on wide products?

Width-dependent defects usually indicate crown or deflection behaviour, because load per unit width changes with width. Check impression film at the specific product width rather than at a standard condition.

Is a cold runout measurement ever sufficient?

For incoming inspection of a cold roll, it is the only option, and it should be recorded as a baseline. For diagnosis on a running line, hot measurement is required, because thermal distortion can add 0.02 mm or more on a large shell.

What if the plant has no spare roll and no vision system?

The single-variable change becomes the primary tool. Nip load first, then temperature with a twenty-minute dwell. The absence of instrumentation slows diagnosis but does not prevent it; what it does prevent is confident attribution of small effects.

How often should rolls be reground?

By measurement, not by calendar. When TIR trend approaches the maintenance limit, schedule the regrind during a planned shutdown. A roll that has moved from 0.012 mm to 0.018 mm over six months will cross a 0.02 mm limit within two quarters.

Do coating and surface finish affect defect frequency?

Yes, indirectly. A rough or contaminated surface changes release behaviour, which changes local heat transfer and can generate banding that mimics a mechanical fault. Surface condition should be measured, not judged by eye.

Where does an embossing calender roll fit in this method?

Engraved rolls such as the Embossing Calender Roll add a second periodicity — the pattern pitch — on top of the circumference. Both must be checked: pattern pitch explains fine repeating marks, circumference explains broader bands. Pattern depth loss is progressive and shows as a gradual softening of hand-feel.

What records should be kept per roll?

Serial number, drawing revision, material and heat treatment certificate, hot and cold runout, crown method and value, coating thickness and hardness, thermal maps, and every reconditioning event with resulting measurements.

Conclusion

Attribution is a measurement discipline. The three measurements — spacing, hot runout, thermal uniformity — plus one list of roll circumferences will resolve the majority of nonwoven defects within a shift. The remaining cases need a controlled change or a swap, and they need the discipline to change one thing at a time and write it down.

The plants that do this well are not the ones with the most instruments. They are the ones with a per-roll record, a trained inspector, and a habit of demanding records with new hardware. That habit is available to any plant, and it is the cheapest reliability improvement in the building.

For engineering support on a specific defect, dimensional records for an existing roll, or specification review before ordering, our team works directly from line data and roll drawings. Background on the full nonwoven roll range is available under Nonwoven Fabric Industry, and field notes are published in the Industry Blog. Project examples are collected in Case Studies.