Fabric Knowledge
Horizontal Bands and Crease Marks: What a Delivered Roll Is Telling You
A band across the width is usually a record of a machine stop. What causes bands and creases, how to tell the two kinds of crease apart, and a ten-minute bench check.
A roll is opened on the bench, three metres are pulled out, and there is a band across the width that was not on the sample. Twenty minutes later someone finds a crease near the core. Both get reported as "defective fabric", and both are usually the visible record of something specific that happened in the plant or in the container.
The two complaints arrive together often enough that they get treated as one quality problem. They are not. A transverse band is almost always a thermal or tension event during coating. A crease near the core is almost always a pressure event after coating. The first is decided by how the line was run; the second by how the roll was wound, packed and stacked.
This article separates them: what produces each, how the spacing of a band narrows down the cause, what a buyer can check in ten minutes without instruments, and what a supplier can and cannot control once the roll leaves the loading bay. It is written from our own production records and our packing specification, and it applies to any coated woven fabric, not only ours.
Key takeaways
- A transverse band is a record, not a random flaw. It marks a moment when the fabric's heat history or tension changed relative to the metres either side of it.
- The spacing narrows the cause. Irregular intervals point to line stops; a band repeating at a fixed pitch points to a roller or a guide, because the pitch matches that roller's circumference.
- There are two kinds of crease and they have different owners. A tension crease runs along the edge and mostly recovers when hung; a pressure crease sits near the core, runs across the width, and does not recover.
- Start and stop effects live at the ends. The first and last three metres of a roll carry most of the risk, so inspect there first.
- Defect count is graded, not argued. In our packing specification a 30 m roll with three marked defects or fewer is Grade A, four is Grade B, five is Grade C.
Where these numbers come from
Two first-party sources, both internal documents rather than published research:
- Production records. Our coating line is run as a continuous process. A stop is treated as an event: the fabric that was inside the oven or the coating bath when the line stopped has a different residence time from the metres before and after it, and that difference is what shows up as a band.
- The packing specification (
Packing Details, Orenda). This document defines the roll build, the defect-marking procedure and the A/B/C grading used at final inspection.
Nothing in this article is a laboratory result, and there is no third-party study behind it. Where a statement is our practice rather than a general rule, it says so. Where a cause is a hypothesis that fits the symptom, it is described as a first check rather than a diagnosis.
A band across the width is a record of a moment
Coating a woven fabric means passing it through a bath or a die, then through an oven, under tension, at a constant speed. The coating pick-up, the cure and the tension are all functions of time in the line. Anything that interrupts time — a stop, a speed change, a tension adjustment — leaves a horizontal trace, because the fabric is moving across the width while the process changes along the length.
Three events produce most of the bands we see:
Event | What changes in the fabric |
|---|---|
Line stop with fabric in the oven | Longer residence time in that section: over-cure, slight colour shift, sometimes a stiffness difference |
Line stop with fabric in the bath | Different pick-up on that section: a band that is marginally heavier or lighter |
Restart | Tension step: the first metres after restart are pulled differently from steady-state running |
The practical consequence is uncomfortable but useful: a band is evidence of a process event, and a process event is recordable. A supplier who runs the line continuously and records stops can tell you where a band came from. A supplier who cannot describe the cause is describing a process they do not monitor.
This is also why the sample card usually looks perfect. Samples are cut from the steadiest part of a run. The band is a property of the length, not of the fabric construction, so it cannot appear on a 50 mm swatch.
The spacing narrows the cause
A buyer cannot see the line, but the band pattern carries information about it. Two patterns cover most cases.
Irregular spacing. Bands at random intervals, often with slightly different intensity, usually correspond to stops: a splice, a roll change, a power interruption, a shift handover. This is the pattern our own technicians describe when they say the line must not be allowed to stop.
Fixed pitch. Bands repeating at a constant distance from each other are a different problem: something rotating is touching the fabric once per revolution. The giveaway is that the pitch stays the same when the line speed changes, because it is set by a circumference, not by time. On a coating line the usual suspects are a guide roller with a build-up on one point, a dancer roller with a flat spot, or a wind-up roller that is out of round.
Pattern | Likely cause, and what it implies |
|---|---|
Irregular intervals, varying intensity | Line stops, roll changes, restarts. A process monitoring problem |
Constant pitch, same intensity | A rotating component with a local defect. A machine maintenance problem |
One band at a regular position in every roll | The same point in the wind-up cycle. A winding or core problem |
Band only at the first or last few metres | Start-up or run-out. A normal process boundary; should be trimmed or declared |
The last row matters commercially. Every continuous line has a start and an end, so the first and last metres of a roll are systematically more exposed. That is not a defect in the fabric; it is a boundary condition. It becomes a complaint only when those metres are sold as if they were mid-run fabric.
Two kinds of crease that get blamed on each other
Creases are where the conversation usually breaks down, because two different faults produce a similar-looking line and they belong to different parties.
Tension creases form during post-treatment, when the web enters a machine with uneven tension across its width. One edge runs slightly faster than the other, the fabric puckers, and the pucker is pressed flat by a nip roller. The signature: the crease runs roughly along the length, near one edge, and often recovers partly when the fabric is hung under its own weight, because nothing was permanently deformed — it was only folded.
Pressure creases form after coating, when the roll is wound or stored under load. The signature: the crease runs across the width, sits near the core, and does not recover. It is a physical impression, not a tension artifact. The innermost layer of a roll takes the highest pressure per square metre, because the whole roll's weight is concentrated there.
Feature | Tension crease → Pressure crease |
|---|---|
Direction | Along the length → across the width |
Position | One edge, anywhere along the roll → near the core, or where the roll was stacked |
Recovers when hung | Partly or fully → no |
Created by | Uneven infeed tension at post-treatment → roll build, stacking, transport |
Owner | Coating plant → winding, packing and logistics |
Our packing specification addresses the pressure side explicitly, and the measures are mechanical rather than procedural:
- a rigid inner paper tube (76 mm bore, 6 mm wall) keeps the roll cylindrical and stops it collapsing onto itself;
- a layer of EPE foam between the tube and the first wrap of fabric removes the hard-core impression — the specification's stated purpose is to prevent permanent pressure marks on the innermost layer;
- machine tightening after winding removes the air gaps that let a roll shift and rub against itself in the container;
- the outer cardboard tube (240 mm bore, 5 mm wall) takes stacking load that would otherwise reach the fabric;
- hollow end caps locate the roll axially, so it cannot slide inside the tube and abrade its own face.
None of that fixes a tension crease. It is worth saying plainly: a crease that runs along the edge is not a packing problem, and no amount of foam will remove it.
What to check in the first ten minutes
This is the procedure we would use on our own incoming material, and it needs a bench, a light source and a tape measure.
- Unroll the first three metres and the last three metres. These carry the start and stop exposure. Look across the width at a grazing angle with a light behind the fabric, not straight on.
- Mark the intervals between bands. If three or more bands appear, measure the distance between them. Equal distances mean a rotating component; unequal means stops.
- Find the creases and classify them. Run a finger along the line, note the direction relative to the roll axis, then hang a one-metre length for an hour and see whether the line relaxes.
- Check the core end. Open the end cap and look at the innermost two turns. This is where pressure creases are born, and it is the only place where the cause is unambiguous.
- Count defects against length. Our grading unit is 30 m with marked defects counted; the grade follows from the count, not from an opinion.
Ten minutes of that produces a description a supplier can act on: "three bands, irregular, first two metres, one edge crease that relaxes when hung, one pressure crease 400 mm from the core." That is a technical claim. "The fabric is bad" is not, and it usually gets answered with a replacement roll rather than a fix.
What we control, and what we do not
Being specific about the boundary is more useful than a promise.
Inside our plant: whether the line runs continuously; how a stop is handled and whether the affected section is identified; infeed tension at post-treatment; winding tension; core specification; foam layer; outer tube; end caps; the defect count and grade on the label.
Outside our plant: how the roll is stored after delivery, whether it is stacked on its side under other rolls, how long it sits in a hot warehouse, and how it is handled on site. A pressure crease can be created by a forklift and a pallet just as easily as by a winder.
Nobody's control: the first 300 mm of any roll, which is where the line starts. Every mill trims some of it. Ask how much.
That last question is a good one to ask any supplier, including us: how many metres at the head and tail of a roll do you treat as unsellable? A supplier with a number has a trimming rule. A supplier without one is selling the start of the run.
What A, B and C actually mean
Defect grading turns a subjective argument into a count. Our specification, applied at final inspection, counts marked defects per 30 m:
Grade | Marked defects per 30 m |
|---|---|
A | 3 or fewer |
B | 4 |
C | 5 |
Every defect is marked in red during production, at the point it is found, for three reasons: it cannot be missed later; downstream cutting can avoid it; and the count becomes data instead of a memory. The marking is also what makes the grade auditable — you can unroll a roll and count the same marks the inspector counted.
Where defects cluster within one metre, the normal practice is to cut that section out rather than leave it in the roll, which changes the delivered length. This is worth asking about before an order, because it explains a length that is slightly short of nominal.
Our take
If a roll arrives with a band, the useful question is not "is this acceptable" but "what produced it". The band pattern usually answers that within a minute of measuring, and the answer decides who pays: a stop is a process issue at the mill, a fixed-pitch band is maintenance, a pressure crease is packing or logistics, and an edge crease is infeed tension. Four different owners, four different fixes.
The part buyers under-use is the grading number. A grade is a count, and a count can be verified independently. That is a much stronger position than a visual judgement, in either direction.
Three things to do on the next delivery
- Open the first and last three metres first. That is where start and stop effects are, and finding them early keeps them out of a claim about the whole roll.
- Measure band spacing before calling it a coating fault. Equal spacing means a roller; unequal means a stop. The two need different responses.
- Ask for the defect grade and the count, not just "first quality". A number can be checked; an adjective cannot.
Sources
- [S1] Orenda packing specification, Packing Details — roll build, core dimensions, EPE layer, defect marking, A/B/C grading, outer tube, end caps. Internal document, 2026.
- [S2] Orenda production records — continuous-run requirement, stop handling, infeed tension at post-treatment. Internal, reported 2026-11-09.
No third-party data is used in this article.
What we do not claim
- We do not claim that our rolls are free of bands or creases. We describe what produces them and how they are graded.
- We do not publish a defect rate, a rejection rate or an on-time figure. Those are not measured in a form we can support.
- The band-spacing table is a diagnostic heuristic from production experience, not a controlled study. A fixed pitch strongly suggests a rotating component; it does not identify which one.
- We do not state a warranty term or a service life.
- We do not name or compare any other manufacturer's process.
Data notes
Figure | Value · source · type |
|---|---|
Inner paper tube | 76 mm bore, 6 mm wall — [S1], stated |
Inner tube weight | ~2.9 kg at 2.5 m, ~3.7 kg at 3.1 m — [S1], stated |
Outer paper tube | 240 mm bore, 5 mm wall — [S1], stated |
Outer tube weight | ~9 kg at 3.2 m, ~7.6 kg at 2.6 m — [S1], stated |
End cap weight | ~0.2 kg each — [S1], stated |
Grade A / B / C | 3 or fewer / 4 / 5 defects per 30 m — [S1], stated |
Inspection zone | first and last 3 m — [S2], practice |
Band spacing test | equal pitch = rotating component; unequal = stops — [S2], practice, not a study |
