October 22, 2026 | Production Notes
A coil can be dead on gauge at every point we measure and still ruin a stamping line. Thickness says nothing about internal stress. When the strip reaches a press, is slit into narrow widths, or is simply cut to length, that stress is released and the sheet moves — edges ripple, the centre lifts, the whole panel takes on a twist. This is the flatness problem, and it is why a levelling and skin-pass step exists between the rolling mill and the coater.
| Defect | What you see | Where it comes from |
|---|---|---|
| Edge wave | Both edges ripple like a curtain; the middle is taut | Edges elongated more than the centre during cold rolling |
| Centre buckle | Middle lifts off the table, edges sit down | Centre elongated more than the edges |
| Oil-can (crossbow) | Sheet rocks on a flat surface, or pops when you press a corner | Uneven extension across the width plus coil-set memory |
| Coil set / L-bend | Sheet keeps a curl after uncoiling | Winding tension stored in the coil |
| Sickle (camber) | Cut sheet curves to one side | Residual stress released by slitting |
The first two are visible before cutting. The last three usually are not — they appear only after the customer slits, shears or stamps, which is exactly when they hurt.
Cold rolling squeezes the strip between work rolls that deflect elastically under load. Cooling across the width is never perfectly even either. Both make some fibres of the steel longer than others. If the difference is large, you see a wave. If it is small, the sheet looks flat but the fibres are held in a locked, unequal state — a spring under load. Coiling adds its own memory on top. Slitting and cross-cutting then break the balance and the shape defect becomes visible.
A tension leveller pulls the strip through a small cluster of work rolls under high longitudinal tension, typically 60–80% of yield, and bends it alternately up and down. Bending under tension lowers the effective yield point, so the strip reaches plastic deformation with far less force than pure stretching would need. The controlled result is a small permanent elongation — in industry practice 0.5% to 2.0%, with 1.0–2.0% common on low-carbon strip of 0.4–2.0 mm.
The point is not to "press the wave flat". It is to elongate every fibre across the width by the same amount, so they all end up the same length. Do that and the internal stress is equalised: the sheet stops remembering that it used to be wavy. That is why levelled strip stays flat after slitting, while rolled-only strip springs back into a sickle.
The same pass usually carries a textured (matte) work roll. Two things happen at once, and they are often confused:
That roughness is not cosmetic. It gives the paint film a mechanical key to grip, it stops turns of coil from sticking and sliding against each other during winding and transport, it holds a thin oil film so the sheet does not gall the die, and it hides the very light waviness that no leveller removes completely. Skin-passing also breaks the yield-point elongation plateau, which is what prevents stretcher-strain (Lüders) marks appearing as bands across a formed panel.
| Problem | Fixed by levelling? | What actually fixes it |
|---|---|---|
| Thickness out of tolerance | No | Rolling mill gauge control / buy the tighter class |
| Wrong mechanical properties | No | Grade and annealing cycle |
| Zinc or paint weight short | No | Coating line set-up |
| Scratches, roll marks, dents | No | Handling and packing discipline |
| Waves, buckle, coil set, crossbow | Yes, within limits | Levelling elongation matched to the defect |
Severe edge wave beyond roughly 2% equivalent strain cannot be recovered by levelling alone; the strip is already stretched past what the machine can equalise, and pushing more elongation just thins the edges. At that point the coil should have been downgraded at the mill, not rescued at the coater.
Flatness is a measurable property, so specify it instead of asking for "flat". EN 10131 gives normal and special (FS) classes as the maximum gap between the sheet and a flat surface, mm, for low-carbon sheet:
| Nominal width | t < 0.70 mm | 0.70–1.20 mm | ≥ 1.20 mm |
|---|---|---|---|
| 600–1200 mm | 12 / 5 | 10 / 4 | 8 / 3 |
| 1200–1500 mm | 15 / 6 | 12 / 5 | 10 / 4 |
| ≥ 1500 mm | 19 / 8 | 17 / 7 | 15 / 6 |
Values are normal class / special class (FS). Editions and product grades differ slightly — check the current standard text before you sign. GB/T 708 covers the same property for cold-rolled sheet, and coil flatness is normally checked with the sheet on a flat table using a straight edge and feeler gauge, or on a levelled inspection table.
For appliance panels we would add one sentence to the PO: "Flatness to EN 10131 special class (FS), measured on cut sheet on a flat table; no visible edge wave or centre buckle after cut-to-length." That sentence is worth more than a page of adjectives, because it is the only version of "flat" a supplier can be held to.
Thickness is what you measure. Flatness is what the press feels. If a supplier quotes gauge but cannot tell you the levelling elongation or the flatness class, the gauge number is only half the story.
