Film thickness is the one coating parameter everybody measures and almost nobody agrees on. The argument always arrives after the scaffolding is down and the money is spent, and it is nearly always about criteria that were never written into the specification.
Four Criteria, Not One
ISO 19840 supplements the ISO 12944 series with the measurement of, and acceptance criteria for, dry film thickness on rough surfaces. It is widely referred to as the eighty per cent rule, which understates it considerably. Acceptance of an inspection area requires four conditions to be satisfied together, and a job can fail on any one of them while passing the other three.
a) The mean meets nominal. The arithmetic mean of all individual dry film thicknesses must be equal to or greater than the nominal dry film thickness.
b) No reading below 80%. All individual dry film thicknesses must be equal to or above 80% of the NDFT. This is an absolute floor, not an average.
c) The 20% cap. Individual readings falling between 80% of the NDFT and the NDFT itself are acceptable only provided the number of such measurements is less than 20% of the total taken.
d) A maximum applies too. All individual dry film thicknesses must be less than or equal to the specified maximum dry film thickness. Where no maximum is specified, ISO 12944-5 is referred to.
Plus the sampling plan. The minimum number of measurements set out for the size of the inspection area must actually be taken, distributed randomly across the surface.
Criterion (c) is the one most often misread. The twenty per cent is not a tolerance on how far a reading may fall short. It is a limit on how many readings are allowed to sit in the band between eighty per cent and nominal. A job where a third of readings land just under nominal fails, even though every single reading is above the eighty per cent floor and the mean might still scrape through.
Eighty per cent is a floor no reading may cross. Twenty per cent is a cap on how many readings may cluster just beneath target. They are different rules and a specification needs both.
The Maximum Is Not a Formality
Criterion (d) exists because over-application is a defect in its own right, not a free margin of safety. Applying more than the system was designed for produces real failure modes.
Solvent entrapment is the common one in thick epoxy films. Where a coat is applied heavier than intended, solvent cannot escape the film before the surface cures, and it remains trapped beneath. The result is soft or under-cured coating, blistering, and poor intercoat adhesion when the next coat goes on. Excessive thickness also reduces flexibility, so a film that should accommodate thermal and mechanical movement instead cracks, and cracking exposes the substrate exactly where the coating is thickest.
Disbonding follows from both. A heavy film carries higher internal stress as it cures and shrinks, and that stress works against adhesion at the substrate interface. The practical consequence is that a specification without a stated maximum has given an applicator no upper boundary, and where a dispute arises the standard directs attention to ISO 12944-5 rather than to a figure the parties agreed.
The Correction Value That Makes It Conform
The measurement itself is where most of the technical trouble sits, because the standard deals specifically with rough surfaces.
A magnetic or eddy current gauge adjusted on smooth flat steel, then placed on a blast-cleaned surface, reads high. The probe sits on the profile peaks while the field penetrates toward the valleys, so the reading includes part of the profile as though it were coating. ISO 19840 addresses this directly: the thickness of the dry film above the peaks of the profile is defined as the instrument reading minus an appropriate correction value, and the dry film thickness is obtained by applying that correction to readings based on adjustment on a smooth flat steel surface.
The standard is blunt about what happens if this step is skipped. Where individual readings based on adjustment on smooth flat steel are used without correction values, that method does not conform with the standard. A DFT report produced without correction is not a non-compliant reading of a compliant method. It is outside the method.
Where the surface profile is unknown and no uncoated sample is available, a default correction value is applied. Where the profile deviates from the normal grades, or where the parties agree a specific approach, the standard sets out a procedure for determining a specific correction value against the actual substrate.
A gauge adjusted on smooth steel and used on blast profile is reading the profile as though it were paint. Subtracting the correction is not refinement, it is what makes the measurement the standard’s measurement.
Two scope limits are worth knowing. The standard applies where the nominal dry film thickness is 40 micrometres or greater, and it notes that where nominal thickness is less than the surface roughness of the substrate, measurement uncertainty increases. That is not a rare edge case: a medium blast profile can approach or exceed the nominal thickness of a thin primer coat, and in that situation high precision is simply not available, particularly in the field.
What Actually Settles a Claim Two Years Later
When a coating fails and the parties reach for records, thickness readings alone rarely decide anything. What decides it is the surrounding evidence, because the question is usually whether the conditions permitted application at all.
The dew point margin deserves a note, because it is routinely quoted as a universal figure and it is not one. The margin required is a manufacturer requirement for that system, and different products state different values. A specification that requires “the usual” margin has specified nothing enforceable, and a record showing air temperature but not steel temperature cannot demonstrate compliance with any margin at all.
The overcoating window matters at both ends for the same reason. Applied too soon, solvent from the previous coat is still leaving the film. Applied too late, the surface may have cured beyond the point where the next coat bonds chemically, and adhesion depends on a key that was never created.
Where Adhesion Testing Fits
Thickness measurement is non-destructive and tells you how much coating is present. It says nothing about whether it is stuck down. Pull-off adhesion testing to ISO 4624 answers the separate question, by bonding a dolly to the cured film and measuring the force required to detach it, together with the nature of the failure.
That second part carries most of the diagnostic value. A failure at the coating-to-substrate interface points at preparation, most often contamination or insufficient profile. A failure between coats points at overcoating interval or intercoat contamination. A failure within the body of a coat points at the coating itself, including solvent entrapment in an over-thick film. Recording only a number, without the failure mode, discards the information that would identify the cause.
Because the test is destructive, it belongs in the specification with its frequency and locations agreed in advance, along with the repair procedure for the test sites. Introduced after a dispute has started, it produces results both sides argue about rather than evidence either accepts.
None of this is complicated, and all of it is cheaper written into a specification than argued afterwards. Name the nominal thickness and the maximum. Require correction values and record the basis. Require the sampling plan to be followed. Log substrate temperature, humidity and dew point margin at each coat. Specify adhesion testing where the service warrants it. The parties then have a shared definition of what acceptable meant, agreed while the scaffolding was still up. The inspection file that comes with our coating work is built to exactly that definition.
Frequently Asked Questions
What are the ISO 19840 acceptance criteria?
Four, all of which must be satisfied. The arithmetic mean of all individual readings must equal or exceed the nominal dry film thickness. No individual reading may fall below 80% of NDFT. Readings between 80% of NDFT and NDFT are acceptable only if they number less than 20% of the total taken. And all readings must be at or below the specified maximum dry film thickness, with ISO 12944-5 referred to where no maximum is stated. The sampling plan for the inspection area must also be followed.
Does a maximum thickness apply as well as a minimum?
Yes. It is one of the four acceptance criteria. Over-application is a defect rather than a margin of safety: heavy films trap solvent, which leaves coating under-cured and prone to blistering and poor intercoat adhesion, and excessive thickness reduces flexibility so the film cracks under thermal or mechanical movement. Where the specification states no maximum, the standard refers to ISO 12944-5.
Why does the gauge reading need a correction value?
Because a gauge adjusted on smooth flat steel reads high on a blast-cleaned surface, since the probe rests on profile peaks while the field reaches toward the valleys. ISO 19840 defines film thickness above the peaks as the instrument reading minus an appropriate correction value. The standard states that readings taken without correction values do not conform to it, so an uncorrected report sits outside the method rather than merely failing it.
Which records actually matter in a warranty dispute?
Substrate temperature rather than air temperature, the dew point margin against the manufacturer’s stated requirement for that system, relative humidity, how the gauge was adjusted and what correction was applied, overcoating intervals against both the minimum and maximum, and batch traceability. Thickness readings alone rarely settle anything, because the question is usually whether conditions permitted application at all.
Sources: ISO 19840, Paints and varnishes, Corrosion protection of steel structures by protective paint systems, Measurement of, and acceptance criteria for, the thickness of dry films on rough surfaces (second edition, cancelling and replacing ISO 19840:2004), including the four acceptance criteria, the sampling plan, the correction value method and the scope limit at nominal dry film thickness of 40 micrometres or greater · ISO 12944-5, Protective paint systems, referred to where a maximum dry film thickness is not specified · ISO 8503 series, Surface roughness characteristics of blast-cleaned steel substrates · ISO 4624, Paints and varnishes, Pull-off test for adhesion
MarineCraft Journal
Get the next briefing in your inbox
Safety protocols, offshore training, regulation and market analysis, written for Malaysia’s maritime and offshore industry. No sales pitches.
We use your address only to send MarineCraft Journal. Unsubscribe any time. See our Privacy Policy.