Coating thickness measurement: DFT gauge use on curved and edged parts

A coating thickness measurement is a number produced by a magnetic field, and the field does not know what it is measuring. On a flat panel the gauge reads the film. On a 10 mm radius it reads the film plus the air gap under the probe, and the gap is larger than the coating. The instrument is in calibration, the certificate is current, and the number is still wrong by a factor that matters.

The geometry is easy to calculate and rarely calculated. A standard probe has a contact diameter of about 5 mm. On a 10 mm radius the contact sits on a chord with a sagitta of 318 µm, four times an 80 µm powder film. On a 50 mm radius the sagitta is 63 µm, still 78% of the same film.

What follows covers how the gauge works, what curvature and edges do to the reading, how the blast profile enters the number, and where to take readings so that the record means something.

How a coating thickness measurement turns a field into a number

Two sensor types cover most sheet metal work. A magnetic induction gauge measures the reluctance of the magnetic path between the probe and a ferrous substrate, which rises as the non-magnetic coating lifts the probe away. An eddy current gauge drives a coil and measures the change in impedance caused by the conductive substrate under the coating, and is used on aluminium.

ISO 2808 describes the methods for measuring film thickness, including the magnetic and eddy-current methods, and sets out the field of application and the precision of each. The important consequence is that both methods infer thickness from a field that extends a few millimetres into the substrate and the surrounding air. Anything that changes that field changes the reading: the substrate alloy, the curvature, an edge, the blast profile, and the probe seating.

A gauge is zeroed on the uncoated substrate, then checked on a foil of known thickness, then used. The zero is where most errors enter. Zeroing on a flat offcut and measuring a curved part is the same mistake as calibrating a calliper on a gauge block and measuring a hot part.

Curvature: the sagitta is the error

The chord geometry is the whole story. For a probe of contact diameter d on a convex radius r, the gap at the centre of the probe is:

s = r − √(r² − (d/2)²)

At d = 5 mm:

• On r = 10 mm, s = 0.318 mm, or 318 µm.
• On r = 25 mm, s = 0.126 mm, or 126 µm.
• On r = 50 mm, s = 0.063 mm, or 63 µm.
• On r = 100 mm, s = 0.031 mm, or 31 µm.

For the reading to be within 5% of an 80 µm film, the gap has to stay under 4 µm, which puts the minimum radius near 780 mm for a 5 mm probe. A 2 mm probe needs a radius of only about 125 mm for the same error, because the gap falls with the square of the diameter. That is why small probes exist, and why a small probe is the first thing to reach for on a curved flange.

There are three honest ways to measure a curved part. Use a probe small enough for the radius. Zero and check the gauge on a curved uncoated coupon of the same radius. Or coat a flat witness coupon of the same material in the same batch and measure that. The third is the most defensible in an audit, because the coupon can be kept with the batch record.

Edges and the field

Near an edge the field has less steel to close through, and the reading moves away from the true film. The practical standoff is 10–15 mm with a standard probe, and more on thin sheet where the field reaches the far side. A 1.2 mm panel gives a thinner magnetic path than a 6 mm plate, and the reading on the thin panel is more sensitive to the probe seating and to what is behind it.

Edges are where coatings fail first anyway, so the measurement that matters at an edge is coverage, not thickness. A coated witness coupon, a destructive cut and a microscope, or an eddy-current probe small enough to sit on the edge are the options. A standard probe held over a hem tells you about the hem, not the film.

The blast profile is part of the reading

A magnetic gauge measures the distance from the probe to the steel, and on a blasted surface the steel is a field of peaks and valleys. The gauge reads to the effective plane of the peaks, which means the reading includes some of the profile. ISO 19840 deals with this: it sets a procedure for verifying dry film thickness against the nominal on rough surfaces, defines inspection areas, sampling and acceptance, and provides correction values that are subtracted from readings taken on a smooth-surface zero. It applies where the nominal film is 40 µm or more.

The size of the correction depends on the profile. On a blast that leaves Rz in the 40–70 µm band, the correction is around 10 µm; on a coarser profile it is larger. On an 80 µm film, a 10 µm correction is 12% of the coating, and a plant that ignores it reports a film it does not have.

There is a second, physical effect. A film that follows a rough surface is thinner over the peaks than over the valleys even when the gauge reading is uniform. If the film is 60 µm over a profile with Rz 40 µm, the coverage over the peaks is closer to 40 µm, and that is where corrosion starts. This is why the coating specification has to pair a film build with a surface profile, and why blasting and coating are one process rather than two.

Where to take the readings

Two rows deserve emphasis. The substrate zero is the one that is easiest to get wrong and hardest to detect, because a gauge zeroed on mild steel and used on galvanised sheet reads through a zinc layer that is itself non-magnetic. The cure state is the other: powder before cure contains volatiles that leave during the cure, so a green reading is not the film that ships.

Film build and the coating route

Powder film build is set by the gun output, the distance, the line speed and the part geometry. A flat panel coats evenly; a box section coats thick at the edges and thin in the recess, because the powder wraps and because the faraday cage effect keeps the field out of the corner. That is a geometry effect, not a gauge effect, and it is why the reading locations have to include a recess and a flange, not just the middle of the largest face.

Wet paint behaves differently again: the film shrinks as solvent leaves, so the wet reading is not the dry reading, and the dry reading is the one the drawing specifies. Where a primer and a topcoat are both applied, each layer has its own nominal, and a gauge reading of the total does not prove either layer. Measure after each coat, or use a method that can separate the layers.

Curing is part of the thickness story. An undercured film is softer, and a probe pressed into it reads low and leaves a mark. An overcured film can be brittle and chip at the bend, which the cross-hatch and a bend test will show before the DFT gauge does.

What Ganesh runs

Coating runs through the seven-tank pretreatment line, the 16 ft × 16 ft shot blast chamber and the 18 ft × 14 ft electrostatic powder booth, with a wet paint booth and a curing oven for the finishes that need them. Film build is checked with a DFT gauge, and adhesion with a cross-hatch test, and both records sit with the batch.

The parts that reach the booth are cut on the 12 kW fibre laser with its 40 ft × 16 ft bed, formed on the 200 T, 5-axis brake with the 4,450 mm bed and welded on the CLOOS cells or at the manual stations. Thickness runs from 0.8 mm to 40 mm, with ±0.5 mm typical sheet metal tolerance, and the coating call is read against the drawing rather than against a house standard.

The system is certified to ISO 9001:2015 and ISO 14001:2015, and the effluent treatment plant is part of the same site, which matters where the buyer audits pretreatment chemistry and waste handling as well as the film.

What to put on the drawing

1. The nominal film build, with a minimum and a maximum, and the unit in µm.
2. The standard the measurement is made to, and whether the roughness correction applies.
3. The measurement locations: which faces, how far from edges, and whether a witness coupon is used.
4. The substrate and its grade, because the zero and the eddy-current response depend on it.
5. The surface preparation: blast grade and profile, or pretreatment type, since both change the reading.
6. The number of coats and the nominal for each, where a primer is used.
7. The adhesion test and its acceptance class, alongside the thickness call.
8. The colour, gloss and any cosmetic standard, kept separate from the film build.

When a DFT gauge is the wrong tool

It is the wrong tool below 40 µm nominal, where ISO 19840 does not apply and the uncertainty of the method becomes a large fraction of the film. It is the wrong tool on uncured powder, on plastic substrates, and on coatings over a non-conductive substrate where neither magnetic induction nor eddy current responds.

It is also the wrong tool for the question a buyer usually asks at the edge. A gauge reading near a hem or a cut edge does not tell you whether the edge is covered; a coupon, a section or a tape test does. And it says nothing about adhesion, colour, gloss or cure, all of which can fail while the film build is perfect.

The honest use of the gauge is as one instrument in a coating record: film build at stated points, adhesion by cross-hatch, cure confirmed by the process data, and the pretreatment chemistry logged with the batch. Send the drawing with the finish specification and the measurement points, and the engineering team will confirm what the coating route can hold before the price is fixed. Request a quote.

The plant and quality hub covers the rest of the coating records. The powder coating on laser-cut edges guide explains why the cut edge coats differently, the degreasing guide covers the step before the booth, and the roughness guide sets the profile that the film has to cover. Coated assemblies run through the paint shop capability.