Why Does 304 Stainless Work-Harden Inside a Coil Line?
304 stainless work-hardens on every pass through the line, and the first sign is usually not a crack. It is a part that measured correctly at the start of the coil and drifted by 0.3 mm at the end.
We have watched this play out on automotive trim, appliance panels and food equipment, always with the same opening complaint: the line is fine, the material changed.
The material did not change. It arrived with one temper and reached the die with another, because the line itself did the cold working.
What the Part Shows Before Anything Cracks
Annealed 304 arrives at roughly 210 to 240 MPa yield. Ten percent cold reduction takes it past 400 MPa, and the line can deliver that reduction without anyone noticing.
The visible sequence usually runs in this order, and each step is easy to explain away on its own.
Length drift. Parts grow along the feed direction as the coil runs down. The first fifty pieces sit inside tolerance and the last fifty do not.
Springback. A 90-degree bend comes out at 93 degrees and needs a bigger over-bend than the setting sheet allows.
Feed resistance. The strip pulls harder through the rolls, so the feed length shortens and the pilot pins start marking the holes.
Edge cracking. Small transverse cracks appear on the outside of a bend, on the slit edge first, at the tightest radius in the tool.
By the time a crack shows, the strip has usually absorbed 15 to 25 percent strain between the straightener, the feed rolls and three or four drawing stations.
Four Places the Strip Gains Strength
Work hardening in 304 is not one mechanism. It is austenite transforming to martensite under strain, and it happens wherever the strip is bent past its elastic limit.
The straightener is the first suspect because it bends the strip more times than anything else on the line.
A nine-roll stack on 1.5 mm material reverses the strip through five or six alternating bends before it reaches the feed rolls. Each bend is real strain.
The feed rolls contribute through slip. Every time the roll spins on the strip instead of gripping it, the surface is burnished and locally hardened.
The die is the largest contributor and the least examined, because it is expected to form the part. Draw stations add strain by design.
The fourth place is the coil itself. Two coils of 304 with the same certificate can differ in skin-pass reduction, and one will form while the other splits.
Roll counts and diameters are published per model in the machinery specification index, which is the first place to look when a stack is suspected of over-working the strip.

| What changes in the strip | How it shows at the press | What it costs if you ignore it |
|---|---|---|
| Yield strength rises 150 to 200 MPa | Longer parts, tighter pilots, higher tonnage per station | Die shimming and a re-cut blank, usually a full day of tool time |
| Springback increases with strength | Angles open up, flanges twist, assembly gaps appear | Over-bend changes on every station, then a weld fixture that no longer fits |
| Ductility falls as martensite forms | Edge cracks at the tightest radius, usually on one side | Scrap rate climbing from 1 percent to 6 percent before anyone measures it |
| Magnetic permeability rises | No effect on the press, but a clean field test for the cause | Nothing, which is why it is the cheapest diagnostic on this list |
How to Prove Which Place Is Doing It
Before changing a single setting, establish whether the hardening is happening before the die or inside it. The two answers lead to different fixes.
- Hold a magnet against the coil before the straightener, then against a coupon cut after it. Annealed 304 barely attracts; transformed 304 grips the magnet hard
- Measure a fixed feature on the first, middle and last part of one coil, and write the three numbers down
- File a corner of strip from both ends and compare how the file bites. A noticeable difference confirms a hardness gradient along the coil
- Run one coil with the straightener rolls opened to a light pass and compare the same three measurements
- Log the coil number against the scrap count for two weeks. If scrap tracks coil identity rather than shift, the mill temper is the variable
The magnet test takes thirty seconds and settles the argument in most cases. It works because the martensite that forms under strain is ferromagnetic while annealed austenite is not.
It is not a measurement. It is a yes-or-no answer to whether the line is transforming the material before the die sees it.
Setting a Line for a Material That Keeps Changing
The fix is not a single number. It is a shift in how the line is set, from removing curvature to removing as little strength as possible.
Reduce penetration before anything else. On 304, set the straightener for flatness and no more. Extra penetration buys a flatter strip and pays for it in yield strength.
A workable starting point on 1.5 mm 304 is a gap at thickness minus 0.03 mm on the entry rolls and near zero on the exit pair, adjusted until the strip lies flat on a plate.
Use back tension instead of pressure. A modest drag from the uncoiler keeps the strip taut against the rolls and lets a lighter pass do the same flattening work.
Cut the die clearance argument short. 304 wants 8 to 10 percent clearance per side. Running a CRS tool at 5 percent on stainless is a punch-snap waiting to happen.
Slow the start of the coil. The first 30 metres off a new coil carry the mill's own variation, and that is where most cracks begin. Many plants run the first 50 strokes at 60 percent speed.
Sort coils by certificate. Ask the mill for skin-pass reduction and tensile direction. Two coils that differ by 40 MPa will not run on one setting sheet.
- Set the entry gap first, then flatten with the exit pair, and stop as soon as the strip lies flat
- Bring the uncoiler brake up until the strip stops lifting off the entry roll, then back off one increment
- Record the setting on the sheet with the coil number, not just the date
- Re-check part length after the first 50 strokes of every new coil, not once per shift
Our coil line knowledge base treats springback and slip as two lines in one error budget rather than separate faults, which is the right frame here.
The Point Where 304 Stops Rewarding You
There is a boundary to all of this, and it sits closer than most tooling engineers expect.
If the part needs more than 12 percent elongation at a bend, no straightener setting will save it. The material has to arrive softer, which means a different temper or a different grade.
If the drawing is deep, a coil line is the wrong entry point. Blank-fed presses with interleaved annealing exist for exactly this reason.
Forcing stainless through a coil line to reach the same shape costs more in scrap than it saves in handling.
If your supplier cannot state temper, the data misleads you. Part-length drift will look like a line problem when it is a purchasing problem.
And 304L behaves differently from 304. The lower carbon grade transforms less readily, so a setting sheet proven on 304L will over-work 304 at the same penetration.
FANTY has run more than 200 lines into production across 12 years, and the stainless jobs that succeed are the ones where the material spec was settled before the machine spec.
How much yield strength does 304 gain per pass?
Roughly 25 to 45 MPa per significant cold pass on 1.5 mm strip. Three passes through a nine-roll stack plus one draw station commonly adds 150 MPa, which is enough to move a bend angle by 3 degrees.
Does a 5-roll stack work-hard 304 less than a 9-roll stack?
Yes, on the same penetration setting, because there are fewer bend reversals. The trade is flatness: a 5-roll stack on 1.5 mm 304 usually leaves 4 to 6 I-units of residual curvature.
Can the magnet test replace a hardness check?
No. It confirms transformation, not magnitude. Use it to locate the cause, then confirm with a portable hardness tester reading in HRB or HV on samples cut from both ends of the coil.
At what thickness does 304 stop being a coil line material?
Above 4 mm the roll load and the work hardening both climb steeply, and most plants move to cut-to-length blanks. Below 0.5 mm the handling problem shifts to buckling instead.
Why does scrap track coil number rather than shift?
Because temper variation between coils is larger than the variation between operators. Two coils from the same mill order can differ by 40 MPa yield, which is more than any setting change on the floor.
Where to go next
Material behaviour is one input to the larger question of how much error the line can absorb before the part fails.
The accuracy pillar works from the decoiler to the die and assigns each source a share of the tolerance, which is where stainless jobs usually discover they had no margin.
See the error budget pillar



