How Much Zinc Does a Coil Line Rub Off Galvanized Strip?
The brackets passed every dimensional check for nine months. Then a customer ran a 480-hour salt spray test on three samples and all three showed red rust at the cut edges and along one face.
The steel was correct, the coating was correct on the incoming coil, and the coil line was rubbing zinc off one face of the strip.
The Complaint Arrived Nine Months After the Run
This was an appliance bracket, 1.5 mm hot-dip galvanised, Z275, run on a 600 kN line at 45 SPM. Roughly 90,000 parts had shipped before anyone tested corrosion.
Dimensionally the parts were fine. Flatness was inside 1 mm over a metre, burr height under 0.05 mm, and the die had run three months without a sharpening.
The failure was cosmetic until it was structural. Rust at a cut edge on an indoor bracket is a complaint. The same rust on a bracket that sees condensation becomes a warranty claim.
What made the case interesting was the pattern. Corrosion appeared on the upper face of the strip as it ran through the line, and almost never on the lower face.
That asymmetry is the first clue. A coating defect from the mill does not know which face is up. A machine that touches one face does.
What the Coating Mass Test Showed
We took three samples: incoming coil, strip after the straightener section, and finished part. Coating mass was measured by the strip method the lab uses, stripped and weighed per unit area.
Incoming strip read 134 to 136 g/m² per face. That is where Z275 should sit, allowing for the normal mill tolerance on a 137.5 g/m² nominal.
Strip after the straightener read 118 to 125 g/m² on the upper face. The lower face had barely moved.
So the line was removing 12 to 17 g/m², roughly 9% to 12% of one face, before the strip reached the die. Nothing in the die could account for that.
The finished part read lower again, because blanking and forming add edge exposure. The line's share was the middle number, and it was the one nobody had measured before.
The finished part read lower again, because blanking and forming add edge exposure. The line's share was the middle number, and it was the one nobody had measured before.
Sampling between sections is what made the number usable. Strip taken after the entry pinch rolls showed a 4 g/m² loss; strip after the straightener showed 12 to 17.
That gradient localised the damage to two metres of machine. Without it, the whole line from decoiler to die is a suspect and the investigation takes weeks.
Why Zinc Moves to the Rolls and Stays There
Zinc is soft. On the hardness scale it sits near 40 to 60 HV, against 120 to 180 HV for the cold-rolled substrate underneath.
Any roll that presses on that surface with slip between roll and strip will smear the coating rather than roll over it. The zinc transfers to the roll and stays there as a thin film.
Once the film forms, the roll is no longer steel on zinc. It is zinc on zinc, which has a higher friction coefficient and picks up more material with every metre.
On this line the upper work rolls of the straightener section had a visible grey bloom. The lower rolls were clean, which matched the coating data exactly.
Roll pressure was the trigger. The head had been set for 2.0 mm mild steel and never reset when the job moved to 1.5 mm galvanised.
Penetration on the last pair was deep enough to mark the strip, and on a coated surface a mark is removed material, not just a visual defect.
Penetration on the last pair was deep enough to mark the strip, and on a coated surface a mark is removed material, not just a visual defect.
Roll diameter changes the picture as well. A 60 mm roll on the same penetration wraps the strip through a tighter arc than a 90 mm roll, and the contact pressure at the tangent point goes up with it.
That is why a head with more, smaller rolls can be gentler on flatness and harsher on coating at the same time. The two goals pull in opposite directions on thin coated strip.
Three Fixes and What Each One Costs
We tried the cheap fix first, because it is also the fastest to reverse. Setting the head correctly recovered most of the loss on its own.
| Fix | What it recovers | Where it costs you |
|---|---|---|
| Reset penetration and gap for 1.5 mm | 8 to 11 g/m² back on the upper face | Flatness margin shrinks, so a coil with bad coil set may need a second pass |
| Change upper work roll finish to a finer Ra | Less mechanical keying of zinc to the roll | New roll set, and a finer finish grips less, which can reduce feed authority |
| Shorten the roll cleaning interval | Stops build-up before it becomes a smear | Two hours of downtime a week, and a cleaning routine someone has to own |
We ended up doing all three, in that order, over five weeks. Coating loss after the change measured 3 to 5 g/m², which is inside normal handling variation.
The roll finish change cost the most and delivered the least on its own. If the budget is tight, the setting and the cleaning schedule carry most of the result.
Blind Spots in a Coating Mass Test
Coating mass loss is not the same thing as corrosion failure. A part can lose 15 g/m² and still pass 480 hours if the remaining coating is uniform and the edges are clean.
Salt spray is also a poor proxy for a real service environment. It is a constant, aggressive, standardised attack. A bracket inside a dry appliance sees none of that.
The test cannot separate line damage from coil variation. If the incoming coil was already at the low end of its tolerance, a 12 g/m² loss looks worse than it is.
Nor can it tell you where in the line the zinc went. We inferred the straightener section from the roll bloom, and confirmed it by sampling between sections.
Without that sampling, the die was an equally plausible suspect.
Finally, the measurement has a floor. Below about 2 g/m², the difference between two labs is larger than the effect you are chasing.
Finally, the measurement has a floor. Below about 2 g/m², the difference between two labs is larger than the effect you are chasing.
Nor does the case prove that 5 g/m² is the right target for every plant. It is the figure that restored this customer's salt spray result, on this coating weight, with this die.
A plant running Z120 or a heavier passivation film would need its own baseline before writing a number into a purchase order.
How to Set the Line Before the Next Coil
Six checks turn this from an argument into a routine. All of them are cheap and none of them need a lab on site.
- Record the incoming coating. One sample per coil lot, upper face and lower face, written into the job sheet.
- Set penetration by gauge, not by feel. Coated strip at 1.5 mm is not the same setup as 2.0 mm bare steel.
- Inspect the upper rolls weekly. A grey bloom is the earliest visible sign and it precedes the coating data by weeks.
- Compare faces after the line. A one-sided loss points at contact; a two-sided loss points at the coil.
- Log the cleaning date. Build-up returns on a schedule, and the schedule is set by metres run, not by calendar.
- Check the entry pinch rolls too. They touch the same face and are usually overlooked because they are outside the head.
The Number to Put in the Purchase Spec
If you buy coated coil for a line, write a coating retention figure into the spec. Ours now reads 5 g/m² maximum loss per face across the line, measured on strip after the straightener section.
That single line changed how the operators set the head. It also gave the coil supplier a fair target, because the measurement point is defined.
Roll geometry and pressure are the two variables that decide whether that number is achievable. Both are worked through in the coil line guide library.
Does losing 12 g/m² of a Z275 coating matter?
Yes at the margins. Z275 is 137.5 g/m² per face, so 12 g/m² is roughly 9%. Parts that passed 480 hours before the change typically drop to 350 to 400 hours after it.
How often should rolls be cleaned on galvanised coil?
On this line, every 40,000 metres or two weeks, whichever comes first. Beyond that the film turns into pickup and the marks become visible.
Can 0.3 MPa of extra roll pressure remove zinc?
Pressure alone does not remove it. Pressure plus slip does, and slip appears when the head is set for a heavier gauge than the strip being run.
Is a 1.5 mm galvanised coil harder to straighten than 1.5 mm mild steel?
The yield strength is similar, but the surface is softer. You need the same flatness with less penetration, which usually means one more roll pair rather than more pressure.
What coating loss should a buyer accept on delivery?
Under 5 g/m² per face is normal handling. Between 5 and 10 g/m² is a setup problem worth investigating. Above 10 g/m² is damage.
Where to go next
Coating retention is one of several specs that a coil line can quietly break, and all of them trace back to how the rolls are set. The roll geometry side of that question is covered in our coil line knowledge base.
The Coil Line Guide Library
