The paint is already on when the coil arrives

An appliance cabinet is not painted. It arrives painted, on a coil, and everything afterwards — blank, draw, pierce, fold, foam — happens to a finished surface. The mill primes and topcoats it and cures both at around 200 °C, faster than any assembly-line paint shop, and the part never sees that oven again. What it sees instead is a press.

That order of operations is the whole specification. A pre-painted coil is bought by what it survives, not by what it looks like on the sample card.

Half a millimetre, held flat by foam

A refrigerator door outer panel is usually 0.5 mm. That is 3.93 kg per square metre, so a door 700 mm by 1,300 mm carries about 3.6 kg of steel. The sides are often 0.45 mm and the back 0.3 mm of SGCC, because none does structural work: the polyurethane injected behind the panel bonds to it and to the inner liner, and the finished door is a sandwich. The skin is a facing, not a beam.

That is why the door is flat at half a millimetre, and why the bare panel feels like a drum: the stiffness arrives later, at around 40 kg per cubic metre. Design it as though the steel carried the door and it will be specified at 0.8 mm, 60 per cent more material for work the foam already does.

The hem is a bend with no radius

The visible edge of that door is a hem: the outer panel is folded 180° around the edge of the inner liner, and the outside of a 180° fold is a zero-radius bend. A 20 µm polyester film on 0.5 mm steel will crack at zero radius and usually survives one thickness of radius, 0.5 mm, so the tool is given that radius and the coating is bought on a bend result, not a gloss number.

The T-bend test folds a coated strip flat on itself, tapes the fold and pulls; the tape lifts cracked coating, and the result is graded in T. Most appliance polyesters are sold at 1T to 2T. A coil that passes 2T but not 1T should not be sent to a line that hems, however good its colour match.

Why the visible panel is electro-galvanized

Hot-dip zinc freezes into crystals a few millimetres across, and under 20 µm of paint the spangle reads as a faint orange peel no primer hides. That is why visible panels are electro-galvanized. SECC deposits zinc from a bath, grain by grain, so the surface is uniform by construction; that page lists zero spangle beside regular spangle because on an electro line there is no crystal to control.

Its appliance range runs 0.3 to 3.5 mm thick and 800 to 1,830 mm wide. Coating weights go from 3 to 90 g per square metre per side, and the page gives the conversion: 50 g/m² of pure zinc is about 7.1 µm. A 40/40 coating is therefore 5.7 µm a side, and the electro-galvanized coil is ordered by that number, not by a Z-grade label borrowed from roofing.

Six microns of zinc behind twenty of paint

Zinc is 7.14 g per cubic centimetre, so every 7 g/m² is a micron. The paint is 20 µm and the zinc is 6: paint is the barrier, zinc is the insurance behind it, protecting steel only where the film is broken. The exposure indoors is condensation and kitchen steam, not chloride, and the usual acceptance, 72 to 120 hours of neutral salt spray, mostly measures the paint's adhesion and its resistance to undercut at the scribe. Doubling to 80/80 buys little on a panel that will never see salt.

Behind the compressor the back panel runs wet, and there the argument changes: a hot-dip galvanized back, or its zinc-aluminium alternative, is chosen on how it behaves at a cut edge, not on coating mass.

The edges that never get coated

Every hole is a bare edge: hinge positions, handle cut-outs, the latch notch. The area is small — a 0.5 mm face around a 100 mm hole is 157 mm² — but it is bare steel in a wet room, and it is where a cabinet rusts first. The plate behind the hinge is a DC01 pressing, and it pulls out before the panel corrodes.

Three ways out. Fold the edge into a flange so the bare face points inward; fit a plastic insert; or touch in the edge at the end of the line, which works for two holes and not twenty. The burr matters for the same reason: 0.05 mm on the visible face prints through 20 µm of paint as a line, and burr height follows die clearance, 5 to 8 per cent of thickness a side. A coil at the bottom of the tolerance and a die set for nominal give a different pierced edge on the last 200 m than the first, and the blanking line says nothing until the panels are painted.

Stainless where the water sits

The drum, the door ring and the dishwasher interior are a different problem: wet, hot and full of detergent. 430 stainless covers most of it at 16 to 18 per cent chromium and no nickel, with 483 MPa minimum tensile and 22 per cent elongation, and it is magnetic. Its expansion is 10.4 µm/m·°C against 17.2 for a 304 austenitic, so a drum grows 40 per cent less over the same temperature rise, which makes the running clearance to the outer tub easier to hold.

The limit is the same thing that makes it cheap. No nickel means no chloride resistance, and the page notes a tendency to intergranular corrosion, so welded parts that sit in detergent belong to a 304 grade. 430F is the free-machining version for valve bodies and screw-machine parts in the same assembly: 552 MPa typical tensile, lower corrosion resistance, and a sulphur addition that makes it unsuitable for a weldment.

What goes on the purchase order

Grade, thickness with tolerance, coating mass per side, surface treatment, spangle, coil inside diameter. Tolerance is not a formality: the Z275 coil page holds the strip to 0.01 mm in thickness and 2 mm in width, and a coil at the top of that band needs a different die shim from one at the bottom. Surface treatment decides whether the press can run the strip at all: chromate and anti-fingerprint keep the suction cups and gloves off a surface that is already finished.

Grade is the harder line to write. DX54D is the drawing grade, and its page names the refrigerator, the washer and the switch cabinet among its uses; galvanized strip next to it quotes 200 to 550 N/mm² tensile across the family, true and useless as a single figure. It is bought on yield strength, elongation and the r-value on the certificate. Ask for tensile alone and you get a coil that meets the paper and splits in the draw. Slit and cut-to-length processing decides how much of it reaches the press without a weld in the middle of a door.