Coating inspection to AMPP and SSPC: the hold points

A coating fails at year four and the argument begins. Was the profile in range? Was the steel above the dew point when the primer went on? Was the dry film thickness measured to SSPC-PA 2 or eyeballed with a gauge someone had in a truck? The applicator remembers one thing, the owner's inspector remembers another, and the daily report is a photocopy with a coffee ring. The hold point either has a record behind it or it does not. This paper walks the hold points an AMPP-certified coating inspector works to, what SSPC-PA 2 actually asks of DFT data, and what the record must carry so a release can be defended years later rather than remembered.

What are the hold points in a coating inspection?

A hold point is a stage where the applicator stops and waits for an inspector's release. The specification defines them, but for industrial steel they converge on a familiar sequence:

• Pre-job: specification review, coating data sheets, gauge calibration certificates, inspector qualifications on file.
• Before surface preparation: ambient conditions, solvent cleaning (SSPC-SP 1) verified, masking and protection confirmed.
• After surface preparation: cleanliness against the specified standard (SSPC-SP 10/NACE No. 2, SSPC-SP 5/NACE No. 1, SSPC-SP 6/NACE No. 3, or ISO 8501-1 grades), surface profile measured, dust and soluble salt checks where required.
• Before each coat: air temperature, steel temperature, relative humidity, dew point, and the recoat window from the previous coat.
• During application: wet film thickness against the target, stripe coats on edges and welds.
• After each coat: dry film thickness to SSPC-PA 2, visual for runs, sags, holidays, contamination and dry spray.
• Final: holiday testing on immersion service, adhesion where specified, cure verification, and the closeout report.

Each hold point produces data. The failure mode is not that the inspector skipped a step. It is that the step happened, the result was written on a clipboard, and the clipboard did not make it into the permanent record with a time, a signature and a gauge identity attached.

What does SSPC-PA 2 require for dry film thickness?

SSPC-PA 2 (now published under AMPP) is the procedure for determining conformance to dry coating thickness requirements on steel. Inspectors often quote it loosely, so the specifics matter.

• Gauge types. Type 1 gauges are magnetic pull-off. Type 2 are electronic. Both must be verified for accuracy before use, typically with certified shims or standards, and adjusted if the specification allows.
• Base metal reading. On blasted steel the profile influences the gauge. The standard describes how to account for the base metal reading so the DFT reported is coating, not coating plus profile.
• Spot measurement. The average of at least three gauge readings within a small circle, roughly 1.5 inches (4 cm) in diameter. Individual readings are expected to scatter. The spot is the unit of judgement.
• Area measurement. Five spot measurements over roughly 100 square feet (10 square meters). The area value is the average of the five spots.
• Sampling. The standard scales the number of areas to the size of the structure: every 100 square feet on small jobs, then a declining number of randomly selected areas as the structure grows.
• Restriction levels. The specification names a coating thickness restriction level (1 through 5). Each level sets how far spot averages and area averages may fall below or above the specified minimum and maximum. Level 1 allows no deviation; higher levels allow spot readings down to 80 percent of the minimum and progressively more overage.
• Nonconforming areas. When a spot or area falls outside the allowed range, the standard requires additional measurements to bound the nonconforming region, which is then repaired and remeasured.

The point for the record: a DFT report that lists a single number per component is not a PA 2 report. A PA 2 report carries the readings, the spots, the areas, the restriction level applied, the gauge, and the verification performed that day. Statistics (minimum, maximum, mean, standard deviation, count) are useful, but they are derived from the readings, and the readings must be there.

How should surface preparation and profile be recorded?

Surface preparation is where most premature coating failures originate and where records are thinnest. Three things need to be on the record, not in the inspector's memory:

• Cleanliness grade achieved, referenced to the specified standard, with photographs taken in the same light and at the same distance as the visual comparator. The photograph is the evidence; the checkbox is the summary.
• Surface profile, measured by the method the specification calls for (replica tape, depth micrometer, or a digital profile gauge to ASTM D4417). The record needs the method, the location, the readings and the mean, because profile that is too shallow reduces adhesion and profile that is too deep leaves peaks the primer cannot cover.
• Contamination checks: dust rating by tape test (ISO 8502-3) and soluble salt levels where the specification requires them (SSPC Guide 15 describes the methods). Salt readings in particular are frequently taken and rarely filed.

The reason this matters at the hold point is timing. Blasted steel starts to rust back within hours in humid conditions. The inspector releases the surface for priming based on cleanliness, profile and the ambient reading taken at that moment. If the primer goes on six hours later after a shift change, the release is stale. The record must show the release time and the coating start time on the same page so the gap is visible.

Why do ambient conditions decide the recoat window?

Most coating specifications require the steel temperature to be at least 3 degrees Celsius (5 degrees Fahrenheit) above the dew point before and during application, with relative humidity below the manufacturer's limit. The measurements are cheap. The discipline is in taking them at the right moments and tying each set to the coat it released.

• Before surface preparation, to confirm blasting will not drive moisture into the profile.
• Immediately before each coat, because a morning reading does not release an afternoon application.
• At intervals during long applications, since conditions move across a shift, especially near sunset and sunrise when the dew point closes on the steel temperature.
• At the start of the recoat window, to confirm the previous coat has cured to the point the data sheet requires and that the maximum recoat time has not passed.

The recoat window is the interval between minimum and maximum time to overcoat, and it depends on temperature. Missing the maximum means abrading or, for some products, removing the coat. The record needs the coat completion time, the ambient conditions, and the data sheet limits together, so the inspector's decision to release the next coat is a calculation anyone can check, not a judgement anyone can dispute.

Which readings should enter the record without retyping?

Every retyped reading is an opportunity for a transposition (0.25 mm becomes 0.52 mm) and an argument about whether the gauge said what the paper says. The gauges used at coating hold points can nearly all send data:

• DFT gauges such as the DeFelsko PosiTector 6000 report each reading with a timestamp.
• Surface profile gauges such as the Elcometer 224T record depth readings per location.
• Ultrasonic thickness gauges are used when the substrate itself is under inspection and the coating record shares the component.

A field app that accepts these readings over Bluetooth or through a keyboard wedge removes the transcription step. The reading lands on the record with the device identity, and the inspector's job becomes confirming location and count rather than copying digits. This also closes a common audit gap: the gauge verification against shims. If the verification readings are captured on the same device profile at the start of the shift, the auditor sees that the gauge read the 250 micron shim as 250 microns before it read the steel.

Photographs belong in the same category. A photo of the blasted surface, the wet film comb, the holiday detector sparking at a pinhole, each bound to the component and hold point at capture time, is worth more than a paragraph of description. Cameras roll photographs are sorted later, usually wrongly.

What does a coating exception report need to contain?

An exception report (some specifications call it a nonconformance report) is the document that carries weight in a dispute. It should exist for every hold point that was not released on the first attempt, and it needs:

1. The hold point and component, identified the same way the drawings identify them.
2. The measured condition with readings, not adjectives: profile mean of 40 microns against a specified 50 to 75, or an area DFT of 180 microns against a Level 3 minimum of 200.
3. The requirement quoted from the specification or data sheet.
4. The disposition: rework, repair, accept as-is with engineering concurrence, or reject.
5. The remeasurement after rework, with the same rigor as the original.
6. Signatures from the inspector, the applicator's representative and, where the disposition is accept as-is, the owner's engineer.

Exception reports are also the best leading indicator a coating program has. A quarter with many profile exceptions on one contractor's crews says something about their abrasive or their air pressure. A cluster of dew point exceptions in one month says the schedule is pushing night work in the wrong season. None of that is visible if the exceptions live in email attachments.

How intelliSPEC handles it

intelliPAINT is the coating and lining module in intelliSPEC. It carries DFT verification with statistics, coating exception reports, and passive fire protection appendices, all written to the same asset record the inspection and integrity modules use. The device profiles described above are shipped in the product: the DeFelsko PosiTector 6000 and UTG, the Elcometer 224T and MTG, and others connect over Bluetooth or keyboard wedge, and readings enter the record without retyping.

Hold points map to workflow transitions. The product ships transitions for submit, approve, reject, hold, resume and reopen, among thirteen in total, and tenants shape the hold-point sequence for a given specification through the form designer rather than through code. The signature widget records per-signer identity even when the inspector and the applicator's representative sign on one shared tablet.

Two properties matter most in a dispute. First, the field app works fully offline with no time limit on the local store, so a blast yard with no coverage produces the same record as a shop. Second, a completed record pins the field configuration and reference options it was executed with, so a 2026 hold point renders in 2032 exactly as it was signed. Every create, update and export event sits on a hash-chained audit trail with actor, time and device, and a verify routine recomputes the chain for auditors.

Fireproofing and coating often share a scope, and the same record structure serves both; see intelliFIRE for the passive fire protection side. Terms used in this paper are defined in the glossary.