A coil can pass every gauge and flatness check on a temper mill and still come back from a customer's paint line with adhesion complaints, or fail a deep-draw trial because the surface texture that looked fine on a profilometer readout wasn't distributed the way the forming die needed it to be. Skin pass rolling is one of the last chances a mill has to correct or ruin a coil, and the two variables that decide the outcome — roughness transfer from the work roll and elongation percentage — interact in ways that are hard to hold steady across a campaign using manual roll-change schedules and fixed elongation targets. This piece looks at where temper mill control typically drifts, what AI-based roughness and elongation management actually changes on the floor, and how a short demo can walk through roughness transfer modeling on your own product mix.
Why Skin Pass Is Harder to Control Than It Looks
Skin pass rolling applies a small, deliberate reduction — often well under two percent — to a coil that has already been cold rolled and annealed. The purpose isn't gauge reduction at all; it's suppressing the yield point elongation that causes stretcher strains during forming, flattening the strip, and transferring a controlled surface texture from the work roll onto the product. Because the reduction is so small, the process is extremely sensitive to roll condition, tension, and speed, and small drifts that would be irrelevant on a heavy reduction pass become visible defects here.
The roughness transfer itself is a function of the work roll's own texture, which degrades steadily from the moment a freshly ground or shot-blasted roll goes into service. A roll that starts a campaign transferring the exact roughness a customer specified for paint adhesion or deep-draw lubricant retention can, by the end of that same campaign, be transferring a texture thirty or forty percent smoother — and the mill's control system, tracking only elongation percentage against a fixed target, has no visibility into that drift unless someone is pulling and measuring samples by hand.
Roughness Targets by Product Application
Different downstream processes need different Ra ranges, and getting this wrong in either direction creates a real customer problem — too smooth and paint or lubricant won't key properly, too rough and the surface shows through a painted finish or scores a forming die prematurely. Holding the right band consistently, coil after coil, is the actual job of temper mill control, not just hitting an average across a shift.
| End Use | Typical Ra Target | Consequence of Miss |
|---|---|---|
| Exposed automotive panel, paint-grade | 0.8 – 1.4 µm | Paint adhesion failure or visible orange-peel finish |
| Deep-draw appliance panel | 1.0 – 1.8 µm | Lubricant retention loss, galling during forming |
| Electrical / lamination steel | 0.3 – 0.7 µm | Coating adhesion or stacking factor issues |
| General commercial sheet | 1.2 – 2.2 µm | Wider tolerance, but still fails if roll runs bright |
Elongation Control Under Real Operating Conditions
Elongation percentage looks like a simple setpoint on a mill HMI, but the force needed to hit a given elongation shifts constantly with incoming hardness variation from the annealing line, strip width, temperature, and the roll's own condition. A fixed force curve tuned for a mid-campaign roll produces under-elongated strip early in the campaign, when the roll is sharper than average, and can over-elongate toward the end as the roll smooths. Model-based control adjusts the force curve coil by coil against a live estimate of roll condition rather than a static schedule, which keeps both elongation and roughness inside spec through the full roll life instead of only near the middle of it.
Strip Shape and Roughness Are Not Separate Problems
A tension or shape correction made without accounting for its effect on roughness transfer can solve one problem while quietly creating another. Bending force and tension distribution across the strip width change the actual contact pressure between work roll and strip at any given point, and contact pressure is one of the direct drivers of how efficiently texture transfers. A shape correction applied through edge tension adjustment, for instance, can locally increase or decrease roughness transfer near the strip edges even though the elongation reading at center width looks unchanged — a mismatch that only shows up later as edge-to-center roughness variation on a customer's incoming inspection report.
What Changes for the Process Engineer Day to Day
In practice, this shifts roll change decisions away from a fixed tonnage or calendar schedule and toward a condition-based trigger tied to the actual roughness requirements of upcoming orders. A process engineer running a mixed order book — some coils needing tight paint-grade roughness, others on looser commercial tolerance — can sequence the schedule so the sharpest part of a roll's life is used against the tightest specs, and the tail end of the campaign is reserved for product where the wider tolerance still holds. That sequencing alone, done manually, is difficult to sustain across shifts and operators; done against a live roughness model, it becomes a scheduling input rather than a judgment call.
It also changes what a rejected coil investigation looks like. Instead of starting from a customer complaint and working backward through paper roll-change logs, the roughness and elongation trend for that specific coil, tied to the roll's wear state at the time it was rolled, is already available — which shortens root cause investigations from days to a single review session.
