Common Stamping Defects and Their Causes

Most common stamping defects are not random; they are a variable that moved. A defect that appears on one part in fifty is not a defect, it is a distribution, and the job is to find which variable moved. A defect that appears on every part since a coil change is a material fault. A defect that appears every 40,000 strokes is tool wear. Telling those three apart before anyone touches the die is the difference between a two-hour fix and a fortnight of polishing.

The five variables that produce almost every stamping defect are material, clearance, force, surface and sequence. Each one leaves a signature: material faults track the coil, clearance faults track the stroke count since the last regrind, force faults track the press and the cushion setting, surface faults track the lubricant and the die finish, and sequence faults track the station order. This page works through the symptoms in that order.

How do you tell a material fault from a tool fault?

Use the boundary test. Find the last good part and the first bad part, then look at what changed between them. If the boundary is a coil change, a new pallet or a shift with a different storage position, suspect the material. If the boundary is a shutdown, a regrind or a tool repair, suspect the tool. If the boundary is nothing at all, the tool is wearing gradually and the defect is on a trend, not a step.

The second test is the gauge. Measure thickness and elongation on a sample from the failing end of the coil, not from the top of the stack. Cold-rolled strip is rolled and coiled under tension, and the last 10–15 m of a coil carries different residual stress from the body. A drop in elongation from 36% to 33% is enough to turn a draw that has run for a month into a splitter.

Storage matters as much as the mill. CRCA that has sat for six months in a warm shop strain-ages: carbon and nitrogen diffuse to the dislocations, the yield point returns, and the material that was drawing-quality at delivery behaves like a stiffer grade. The visible symptom is a Lüders band, a band of coarse stretcher strains across the panel, appearing at 1–3% strain. The fix is a first-in, first-out store and a mill certificate that states the age, not a die change.

What causes a split at the draw radius?

Strain above the forming limit. The forming-limit curve, constructed to ISO 12004-2 on specimens between 0.3 mm and 4 mm, tells you the major strain a grade tolerates at a given minor strain. A draw that sits within 10% of that curve will split on the first coil with slightly lower elongation, which is why a die that has run for months can start splitting without a single tool change.

Measure the wall. A cup drawn from 1.5 mm sheet that thins to 1.25 mm at the punch nose has lost 17%; at 1.20 mm it has lost 20% and has almost no reserve. The thinning is concentrated opposite the blank-holder gap, so measure at four points around the circumference rather than at one.

Three corrections, in order of cost:

• Enlarge the punch nose radius. Doubling it from 3t to 6t spreads the strain over a longer arc and is often enough on its own.
• Reduce the first-draw depth. A shallower first draw and a second draw is more tool but a stable process.
• Change the grade. Moving from 5052-H32 to 5052-O, or from a D grade to a DD grade of CRCA, buys elongation at a small material premium.

Why do flanges wrinkle?

Because the flange has to compress as it flows inward, and compression buckles anything thin that is not restrained. The blank holder restrains it by friction and by pressure, but only where the holder actually touches. On a part with a cut-out, a step or an asymmetric blank, there are regions the holder cannot reach, and that is where the wave starts.

On CRCA, blank-holder pressure of 1.5–2.5 MPa controls a flange in the 1–3 mm range. Above 2.5 MPa the friction stops the flange flowing, the wall thins instead, and the split arrives where the wrinkle used to be. On aluminium the working band is lower, 0.7–1.5 MPa, because the material flows at lower stress.

When pressure alone will not do it, add a draw bead. A bead of 3–5 mm width and 1–2 mm depth adds a controlled restraining force at a specific location, which is more useful than raising the pressure everywhere. Balance the bead around the part: an unbalanced bead twists the flange, and twist is harder to correct than a wrinkle.

Why does the burr grow across a production run?

Because the clearance grows. Every stroke wears the punch edge and the die edge, and the clearance measured at the start of the run is not the clearance at 50,000 strokes. Blanking clearance on CRCA runs about 5–8% of thickness per side. Once wear pushes it past that band, the material no longer shears cleanly; it bends over the die edge and tears, and the burr height climbs.

Set a burr limit and a regrind trigger. A burr of 5–10% of thickness is the usual acceptance band, so 0.10–0.20 mm on 2 mm sheet. Measure it at the point where the punch exits, on the same feature every time, with a profile projector or a burr gauge. A clearance gain of 0.03 mm per side is a sensible regrind trigger for a production tool, because the burr curve steepens after that.

Burr direction matters to the next operation. The burr sits on the side the punch exits, so flipping the strip moves it to the other face. If the part is powder coated, a burr above 0.2 mm shows as a thin edge where the coating pulls back during cure, and it fails a salt spray test before it fails a dimensional check. State the burr side on the drawing; a burr limit with no side is half a specification.

Why does a punch drag the slug back up?

Because the slug has nothing holding it down once the punch starts to withdraw. The cut slug sits in the die bore, the punch face is wet with lubricant, and the pressure difference and the oil film can lift it. On the return stroke it comes back into the strip, gets hit again on the next stroke, and leaves a torn, double-struck hole with a plug of metal welded to the punch face.

Clearance and punch geometry are the controls. Too little clearance makes the slug grip the punch; a punch with a slight back taper, or with a shear angle that leaves the slug tilted, releases it instead. A vacuum slug extractor or an air blast through the die bore fixes the rest. In a progressive tool the cost of a slug pull is not the one bad part, it is the station that takes the hit on the next stroke.

Galling follows a different route. Aluminium and stainless both cold-weld to tool steel under pressure, so the die radius picks up metal, the pickup work-hardens, and the next part is scored along the draw direction. The tell is a scored wall with a matt texture and a visible build-up on the die. Polish the radius, coat it, and check the lubricant before touching the clearance.

Why does a panel look wavy after coating?

Because the flatness was marginal before coating and the coating made it visible. A powder coat at 60–80 µm dry film thickness does not hide a 0.3 mm oil-can, it outlines it. The cure cycle also heats the panel to 180–200 °C, which can relax residual stress and move a panel that measured flat on the bench.

The physical cause of oil canning is a flat region with no curvature, held in a state of residual stress by the surrounding form. Push it and it snaps through to the other stable shape. Three fixes:

• Add crown. 1–2 mm of crown across a 500 mm panel is usually enough to make the panel stiff and stable.
• Add a bead or a swage. A single stiffening bead turns a flat field into two curved ones.
• Balance the blank. An asymmetric blank leaves one side of the panel more stretched than the other, and the difference shows up as a diagonal wave.

Check the panel on a surface plate with a 0.5 mm feeler before coating, not after. A panel that will not pass a feeler check in the white will not pass it after cure.

What does a defect report need to contain?

A defect that is described as "cracked" cannot be fixed. A defect report that says where, when and how much can. Record the following on every reject:

• The feature and the position on the part, with a photograph and a scale.
• The stroke count since the last tool regrind.
• The coil or heat number and the mill certificate for that lot.
• The press, the cushion pressure and the lubricant batch.
• The measured value against the drawing tolerance, not a description.
• Whether the part was measured free or clamped.

That record is what turns a monthly argument into a trend line. It is also the input for a first article inspection or a PPAP submission: tools and dies are only as good as the feedback loop that corrects them, and most corrections are cheaper when they are made from data rather than from a reject bin.

How Ganesh Multitech closes the loop on a stamping run

The tool room builds and maintains the press tools, progressive dies, welding fixtures and checking fixtures in-house, so a regrind, a radius change or a bead adjustment is a shop-floor action rather than an external purchase. Checking fixtures are built with the tool, which means the first article is measured in the fixture the production part will be measured in.

Press work runs on mechanical presses from 20 T to 300 T with 250 T punching, across five units in Mumbai (Vasai) and Nagpur, under ISO 9001:2015 and ISO 14001:2015. First article inspection and PPAP documentation are available, formed tolerance is ±0.5 mm typical with tighter features by agreement, and the custom-part range is 0.8–40 mm. Coating defects are closed out with dry film thickness and cross-hatch adhesion records rather than a visual opinion.

Parts that are pressed here go into gensets, elevators, racks and railway equipment for customers including Mahindra, Mahindra Powerol, Godrej & Boyce, TKE, Eros Elevators, Minda, Gunnebo and Indian Railways, and the tooling has run since 1992.

Which defects are a design problem, not a process problem?

Some defects cannot be pressed out. A hole 1 mm from a bend will distort on every tool. A flange shorter than 1.5t plus the radius will not form a clean bend. A draw deeper than the alloy's limiting draw ratio will split, whatever the blank-holder pressure. A part with a machined pad or a ground face is not a stamping.

The test is simple: if the correction requires the metal to do something it cannot do, the answer is a drawing change, not a die change. The inspection guide for repeat production sets out how to catch those cases at the first article, when a drawing change is still cheap. The springback guide covers the one defect that most often gets blamed on the tool and belongs in the tool design.

We do not fine-blank, we do not run presses above 300 T, and we do not repair a design defect by sorting. If the capability does not exist, the quotation will say so.

What to put on the drawing

• Material grade and thickness with the standard reference, so the forming limit can be checked against the right curve.
• Burr side and burr height limit, measured at a named feature.
• The features that must not carry a die mark, and the faces that will be visible after coating.
• Bend angles with a tolerance and a note on whether the part is measured free or clamped.
• Flatness or oil-can limits on any large flat panel, with the measuring method.
• The general tolerance class from ISO 2768-1, plus any tighter callout on pitch and profile.
• The annual volume, so the regrind and maintenance plan can be priced into the tool.

Send the drawing, a photograph of the defect and the stroke count to request a quote. The stamping and press tools hub carries the companion guides on tonnage, tooling and station sequence.