Edge quality and coating life: where the corrosion clock starts

An enclosure installed on a coastal site in 2019 came back for inspection with the face film intact at 78 µm, a gloss reading that still passed, and red rust running along every laser-cut edge. The coating had not failed. The edge had been under-coated from the day it shipped, and nobody measured it because the drawing said 80 µm and the gauge said 80 µm on the face.

Edge quality and coating life are the same subject seen at two moments: the geometry that decides how much film lands on the arris, and the years it takes for the shortfall to become rust. A part can pass every dispatch check and still fail at year two, because the check measures the face.

The rule is that the edge is a separate coating zone with its own film budget, and the budget is set by the drawing. A sharp arris gets 40–60% of the face build; an R0.5 gets 80–95%; an R2.0 gets 95–100%. Everything below is why that ratio decides the service life and how to specify it.

Edge quality and coating life: which defect ends it first?

The one that leaves bare steel at the arris, and the one that leaves a crevice beside it. Burr, dross and a sharp corner do the first; a burr against a mating face, a lap joint or a folded hem does the second.

Laser oxide is a different mechanism. It sits under the film, absorbs moisture and cracks the film from below, but that failure is a preparation problem and powder coating laser-cut edges covers the routes that remove it. This article is about what the edge does to film coverage and what that costs over the life of the part.

The arris is a coating shadow

Powder coating is line of sight and electrostatic. A 90° corner presents almost no area to the gun, so fewer particles land there, and the molten film then pulls away from the sharp arris under surface tension. The measured result is 40–60% of face build.

On an 80 µm face that is 32–48 µm at the edge. A C3 system typically assumes 60 µm minimum, so the edge is below specification before the part leaves the plant, and it is below it by more than the tolerance on the face reading.

A burr is a crevice with a wick

A burr of 0.1 mm under a bolted joint holds the two surfaces apart by 0.1 mm. Water enters that gap and then climbs, because a narrow gap is a capillary.

Capillary rise is h = 2γcosθ/(ρgd). With water at 0.072 N/m, a contact angle of zero on wet steel, and a 0.1 mm gap, h is 2 × 0.072/(1,000 × 9.81 × 0.0001), which is 0.147 m, or 147 mm.

That is not a small effect. A 0.1 mm burr at the bottom of a panel can pull water 147 mm up into a joint, which is why rust appears along a bolted flange well above the wet line. At a 0.5 mm gap the rise is 29 mm, and at 1 mm it is 15 mm. Deburring to 0.1 mm maximum does not close the gap; it makes the gap small enough to wick.

Holes are edges too

A panel has more cut edge than its outline suggests. Take a 500 × 500 mm door with 40 holes of 8 mm: the outline is 2,000 mm of edge, and the holes add 40 × π × 8, or 1,005 mm. A third of the coated edge on that part is inside a hole.

Holes are worse than the outline because the film has to wrap a closed curve with a gun that cannot enter it. Below 3 mm diameter the film bridges and closes the hole, and the edge that was supposed to be protected becomes a solid plug of powder with a bare ring behind it. Drill after coating, or open the hole to 3 mm and accept the larger fastener.

The same arithmetic applies to slots, louvre openings and cable entries. Count the edge length before deciding the edge specification, because a part with 40 small holes needs the radius and the edge film check more than a plain panel does.

What film arrives at each edge condition?

The Ganesh column is what the line can hold. A radius on a laser path is nearly free; a radius on a sheared edge is a second operation. A stripe coat is a manual pass and is quoted as one.

What does a bare edge cost in years?

ISO 12944-2 classifies the environment by first-year steel loss. C3, urban and industrial inland, runs 200–400 g/m² per year. C4, coastal and industrial with salt, runs 400–650. C5, very high salinity, runs 650–1,500.

Steel loses 1 µm of thickness for every 7.86 g/m², so C3 is 25–51 µm per year, C4 is 51–83 µm, and C5 is 83–191 µm per year, per exposed face.

A cut edge has two faces. On 1.6 mm sheet in C3, the pair removes 50–102 µm per year, so 10% section loss, 0.16 mm, arrives in 1.6 to 3.2 years. In C4 it arrives in under two years.

The flat face next to it, under a sound 80 µm film, is still intact at that point. The edge is not a small part of the corrosion problem; it is the clock.

Why the salt spray number is not the service life

A 480 h neutral salt spray test to ISO 9227 is the duration ISO 12944-6 sets for a C3 high-durability system, and 720 h for C4 high. Passing it means the system met a comparative test, not that the part will last 480 hours outdoors, which is obvious, or ten years, which is not.

What the test does give is a comparative number for the edge. Measure creep from the cut edge at the worst point, in millimetres, and compare candidates on that basis. A system that creeps 4 mm from the edge at 480 h and one that creeps 1 mm are not the same product at the same price, even when both say 480 h and no red rust.

The inspection report should carry the edge creep figure, the film thickness at the edge and the rust grade. A report with a pass and no numbers cannot be compared with the next batch.

Why bending after coating is a crack, not a stretch

Cured powder is a thermoset. Its elongation at break is far below that of the steel under it, which is why a coated part is normally bent first and coated second.

The strain at the outer fibre of a bend is t/(2R + t). At R2 on 2 mm, that is 2/(4 + 2), or 33%. At R0.5 on 2 mm it is 67%. At R4.5 on 2 mm it is 18%.

A cured polyester powder will not follow 33% strain. It cracks along the bend line, and the crack runs to the nearest cut edge, which is now a bare steel path into the joint. If a part must be formed after coating, the bend radius has to be large, the film thin, and the bend line masked, and the result still needs a visual check at the toe.

The practical rule: keep every bend at least 3 mm from a cut edge, so the film that survives the bend is not also the film that has to protect the arris.

What Ganesh actually runs

Cut edges come off a 12 kW fibre laser with a 40 ft × 16 ft bed, or off the 6 kW and 3 kW support machines, with a kerf and edge condition that ISO 9013 classifies by range. Where the drawing calls a radius, it is in the laser path. Where it does not, the standard edge is a sharp arris with a 0.1 mm burr limit, which is the cheapest edge and the shortest coating life.

Parts then go through the 16 ft × 16 ft blast chamber, seven-tank pretreatment, the 18 ft × 14 ft powder booth and the curing oven. Film is checked with a DFT gauge at nine points per part, and the edge is one of them where the drawing calls for it. Adhesion runs to ISO 2409, and salt spray to ISO 9227 with the edge creep recorded in millimetres rather than a pass mark.

Enclosures and canopies for Mahindra Powerol, Godrej & Boyce and TKE run on this route, and the shot blasting capability page sets out the chamber and the profile it holds.

What to put on the drawing

1. The edge radius, by edge: "all external cut edges R0.5" or a symbol with the radius, not a general deburr note.
2. The arris film minimum, for example "60 µm minimum at all cut edges", with the measurement point 2 mm in from the edge.
3. The burr limit and the permitted side, because a burr on the coated face is different from a burr on the inside.
4. The bend-to-edge distance, at least 3 mm on any part that is formed after cutting.
5. The corrosion class from ISO 12944-2 and the durability band, which together set the salt spray duration.
6. The edge creep acceptance in millimetres at the test duration, with the assessment method.
7. The stripe coat, where the class is C4 or above, and the zinc-rich product family.
8. A note that the part is coated after bending, if the sequence is not obvious.

The deburring and edge quality guide covers the cutting parameters and the methods that produce the edge, and the surface treatment hub indexes the rest of the coating guides, including why powder coating fails on outdoor enclosures.

When a radius is the wrong answer

A radius is wrong when the edge is a locating feature. A part that sits in a machined slot or registers against a shoulder needs a sharp edge, and rounding it moves the datum. In that case protect the edge with a stripe coat or a sealant, and keep the sharp geometry.

It is also wrong on a part that is welded along the edge, because the weld consumes the radius and the toe becomes the new coating problem. Prepare the toe instead, as the weld preparation route does.

And it is wrong when the failure is not at the edge at all. If rust appears in the middle of a panel, the cause is a film defect, a contamination spot or mechanical damage, and a radius on the edges will not change the outcome.

Send the drawing with the service environment, the expected life and the edges that matter to request a quote. We will confirm the radius, the film build at the arris and the test that proves the edge before the first part is coated.