Salt spray interpretation: what counts as failure

Two coated panels come out of the same chamber after 720 h. One is graded zero creep at the scribe. The other carries 4 mm of lifting either side. Both reports say pass, because the purchase order asked for a salt spray test and never said what failure looks like.

Salt spray interpretation is a measurement problem before it is a judgement problem. The chamber runs 5% sodium chloride at 35 °C, pH 6.5 to 7.2, with a fall-out rate of 1 to 2 mL per 80 cm² per hour, and it holds the panel wet for the whole duration. What the report then contains is a set of readings: creep either side of a scribe, a blister size and frequency code, a rust area percentage, and sometimes an adhesion grade. Each of those is defined by a standard. None of them is defined by the number 720.

This guide sets out what those readings mean on sheet metal parts, where the failures actually start, and the clause to add to a drawing so that a certificate can be accepted instead of filed.

What counts as failure in a salt spray test?

A panel fails when a measured defect exceeds the limit written for that part, and no standard supplies that limit on its own. ISO 9227 defines how the test runs: salt concentration, pH, chamber temperature, collector rate and scribing. It does not say what an acceptable result is, because 720 h is severe for one coating and routine for another.

The acceptance comes from somewhere else. ISO 12944-6 sets test durations and acceptance criteria for protective paint systems by corrosivity category and durability band, and it is the clause most industrial buyers should be quoting. For a C4 system rated high durability, 720 h of neutral salt spray is the reference exposure; for C5-M at the same durability, 1,440 h. If the drawing says only "salt spray as per ISO 9227", it has specified the apparatus and left the verdict open.

Three readings decide most disputes, and each has its own scale:

• Creep from the scribe, measured in millimetres from the edge of the cut to the first point where the coating has lifted from the substrate. Measured on one side, not across the scribe.
• Blistering, coded as size and frequency per ISO 4628-2: a number from 1 to 5 for the diameter of the worst blisters and a number from 1 to 4 for how densely they sit.
• Rust, rated per ISO 4628-3 as Ri 0 for none, Ri 1 up to 0.05% of the area, Ri 2 from 0.05% to 0.5%, Ri 3 from 0.5% to 1%, and Ri 4 from 1% to 8%.

A fourth reading matters on sheet metal and is usually missing: creep at a cut edge, a weld seam or a punched hole. Those are not scribed features, so the standard has nothing to say about them, but they are where an outdoor enclosure rusts first.

How scribe creep is measured

The scribe is a single straight cut through the coating to bare metal, 0.5 mm wide, made with a blade held in a guide. ISO 9227 refers to ISO 17872 for the scribing method, and the tool matters: a hand-held blade that wanders produces a 0.8 mm groove with a torn coating edge, and the tear reads as creep on the first inspection.

After exposure, the panel is rinsed, dried and examined within the hour. Creep is measured perpendicular to the scribe at five points along its length, from the edge of the cut to the boundary of the lifted film, and the worst reading is reported. The boundary is not always sharp. On a powder coating it is usually a clean line of undercut where the film has lost adhesion; on a wet paint it can be a soft zone of discoloration that is not yet detached.

The distinction matters because the number rejects a batch. A part with 1.8 mm of clean undercut and a part with 1.8 mm of soft halo are graded the same unless the report says which boundary was used.

What 2 mm of creep removes

Creep reads as a small number in millimetres, so it is worth converting to area. A 150 x 100 mm panel has 15,000 mm² of face. A 100 mm scribe with 2 mm of creep on both sides removes 2 x 2 mm x 100 mm, which is 400 mm², or 2.7% of the panel. At 4 mm of creep the same scribe removes 800 mm², or 5.3%.

The conversion explains why "2 mm allowed" is not generous: four percent of the coated face has failed on a panel that spent 30 days in a chamber, and that area becomes the starting point for under-film corrosion in service.

Salt spray limits worth writing into a specification

The values below are the ones that hold up on a powder-coated CRCA or galvanised enclosure. They are specification choices, not clauses of ISO 9227, and the test duration should match the corrosivity category the part will actually see.

The dry film thickness row is the one that decides most arguments. Salt spray hours and film thickness are not interchangeable, but on the same system a 50 µm film will fail earlier than an 80 µm film at the same exposure. If the report does not state the thickness measured on the tested coupon, the hours cannot be compared with any other supplier's hours.

Why the blistering code hides the area

ISO 4628-2 codes blisters twice: a size from 1 to 5 and a frequency from 1 to 4. Size 2 is 0.5 to 1 mm across, size 3 is 1 to 2 mm, size 4 is 2 to 5 mm. Frequency 2 is medium, 3 is medium dense, 4 is dense. The code reads like a grade, and a report that says 3(S2) looks mild next to 4(S4).

Convert it. A size 3 blister averages about 1.5 mm across, which is 1.77 mm². At frequency 2 there may be 25 to 100 of them on a 150 x 100 mm panel, so 44 mm² to 177 mm², or 0.3% to 1.2% of the face. At frequency 4 the count runs into the hundreds and the blisters merge, which is where the coating has effectively failed across 5% to 10% of the panel.

Two rules follow. Report the count with the code: a band can hide a factor of four. Second, photograph the panel at a fixed magnification with a scale in frame; a 1 mm blister photographed at arm's length with a phone is invisible in the report.

Blistering on a sheet metal enclosure is almost always a pretreatment problem rather than a resin problem. Trapped soluble salts under the film draw water by osmosis, and the blister forms where the salt was. That is why blisters cluster near a weld seam, a hem, or a corner where rinse water sat — the same places the seven-tank line has to reach, and the same places a badly designed jig shields. The pretreatment process explained covers the stage chemistry; the outdoor enclosure failures article covers what the resulting corrosion looks like after a season outside.

Where sheet metal parts fail first

On a flat, well-pretreated panel, salt spray mostly measures the resin. On a real part it measures the geometry, and the geometry is where the argument should be.

Cut edges

A sheared edge presents a vertical face with no coating and a burr with a rolled lip. Powder covers it at 20 to 30 µm against 70 µm on the face, and at the burr root the film is thinner again. That edge starts corroding within hours of exposure. On a laser cut edge the recast layer and oxide hold the film slightly better than a shear burr, but the edge is still the weak point.

On a 2.5 mm CRCA panel with 1,200 mm of cut perimeter, the exposed edge is 1,200 x 2.5 mm, or 3,000 mm². That is 20% of the coated face of a 150 x 100 mm panel, and none of it is in the salt spray report. A part can pass 720 h on the face and rust from the edge in the first wet season.

Weld seams

A MIG seam carries spatter, a heat-affected zone and often a silicate residue from an anti-spatter spray. Coating over it produces a film that is thinner over the spatter and poorly bonded on the scale. Blisters appear along the seam line first, then rust bleeds out of the seam and stains the face below it. The stain is what the customer sees, so a seam failure is reported as a panel failure.

Hems and returns

A hem traps pretreatment chemical when the rinse cannot reach inside it, and it traps water in service. Creep from a hem cannot be measured by the scribe method, and no rating covers it.

Punched holes and slots

A punched hole has a burnished wall and a rollover at the entry, and both hold a thin film. A 6.5 mm hole has 20 mm of wall on 1 mm sheet, so a 12-hole pattern carries 245 mm of edge, more than the perimeter of a small bracket.

The practical consequence is that salt spray panels should be cut from the real part, not from a flat test plate. A coupon with a hem, a weld and a punched hole tells you more at 480 h than a flat plate tells you at 1,440 h.

What Ganesh records, and what the coating line can hold

Coating at Ganesh runs seven-tank pretreatment into an 18 ft x 14 ft electrostatic powder booth, with a wet paint booth and a curing oven on the same route. Parts that need a mechanical key go through the 16 ft x 16 ft shot blast chamber first. Dry film thickness is measured with a gauge and adhesion is checked by cross-hatch, so every batch leaves with a thickness and an adhesion record rather than a visual opinion.

What the line does not do is run its own salt spray chamber, so a neutral salt spray result is a qualification test rather than a per-batch release check. In practice that means the coating system is approved once on coupons cut from the production route, and the ongoing evidence is the pretreatment stage record, the DFT reading and the cross-hatch grade. If a buyer needs salt spray on every lot, the test goes to an accredited laboratory and the sample plan has to be agreed at quotation, because the panels cost a batch slot and a week of chamber time.

The route covers CRCA, galvanised steel, aluminium and stainless, and the substrate changes what the salt spray number means. A galvanised part with a 70 µm powder film will show almost no rust at the scribe for 720 h because the zinc protects the exposed steel; a CRCA part with the same film shows creep from the first day. Comparing the two reports as if they were the same test is the most common misreading in a vendor comparison. The paint shop capability lists the stages and the substrates, and the plant and quality hub collects the inspection and pretreatment guides that sit behind these records.

What to put on the drawing

A salt spray requirement that can be accepted or rejected has eight fields. A note that says "salt spray 720 h" has one.

1. Method and standard: neutral salt spray to ISO 9227, with the edition.
2. Duration and corrosivity category: 480 h for C3, 720 h for C4, 1,440 h for C5-M, plus the durability band.
3. Substrate and pretreatment: CRCA with zinc phosphate, or galvanised with a chosen pre-treatment, because the same hours mean different things on each.
4. Film build: the dry film thickness range on the face and the minimum on a cut edge.
5. Scribe: to substrate, 0.5 mm wide, straight, cut before exposure.
6. Acceptance: creep in millimetres, blister code and count, rust rating, and the adhesion grade after exposure.
7. Sample: number of coupons, whether they carry a weld, a hem and a punched hole, and which batch they come from.
8. Record: the report format, the photographs at fixed magnification, and who holds the coupon.

If those eight fields are present, a report can be checked in ten minutes. If two of them are missing, the report is a claim.

When salt spray is the wrong test

Neutral salt spray is a ranking test for coating systems in a permanent wet chloride film. It is not a service life prediction, and it does not model the conditions that actually destroy an outdoor enclosure.

It misses UV entirely, so a coating that chalks and loses gloss in two summers can pass 1,440 h without difficulty. It misses the wet-dry cycle, which is what drives osmotic blistering on a real panel. It misses acid deposition and it misses mechanical damage from handling and installation. A C5-M coastal site with 200 to 1,500 mg/m²/day of chloride deposition, as classified in ISO 9223, is a different exposure from a 5% salt fog at 35 °C.

Where UV decides the outcome, the relevant test is a xenon-arc or fluorescent UV exposure to ISO 16474, or a cyclic test that alternates salt spray, drying and condensation. If the part sits indoors, in a data hall or a lift shaft, 480 h is already more than the service condition demands.

Where the risk sits at an edge or a seam, no chamber test replaces a coated sample sectioned and measured under a microscope. A 30 µm edge film read off a section is a harder fact than any hours figure.

Send the coating clause you intend to apply with the drawing and the quotation will state the duration, the substrate, the film build and the records that come with the batch. If the clause cannot be met on the route, it is better to find that out before the tooling is cut than after the first 720 h report — the quotation request is the place to settle it.