Holiday Testing New Coatings: Getting the Voltage Right

What SP0188 is, and the boundary it draws around itself

SP0188 is the standard practice for discontinuity testing, universally called holiday testing, of new protective coatings applied to conductive substrates. A holiday is any void, pinhole, thin spot, inclusion or crack that leaves a conductive path from the surface of the coating down to the steel underneath. The practice describes how to find them, and it describes nothing else.

Three exclusions decide most disputes. The practice is written for new coatings, so aged coatings, coatings that have been in immersion service and coatings contaminated by process residues fall outside it. It applies to conductive substrates, so coatings on concrete, masonry or composites are not covered. And it establishes no acceptance criteria whatsoever, which is why a specification saying test per SP0188 without saying what a pass looks like has in fact specified nothing.

Its companions are ASTM D5162, which covers the same two techniques for non-conductive coatings on metallic substrates, and ASTM G62 for pipeline coatings. Referencing documents point at all three inconsistently. When a coating specification cites SP0188 and an ASTM practice together, the requirements need to be reconciled before mobilisation, because the voltage guidance and the surface condition expectations are not identical between them.

Wet sponge or spark: 500 microns decides

Dry film thickness selects the technique, and the dividing line is 500 microns, 20 mils. At or below that thickness the low-voltage wet sponge method is used. Above it, high-voltage spark testing applies. The reason is physical rather than administrative: a thin film cannot survive the voltage a thick film requires, and a low-voltage sponge cannot drive a detectable current through a thick one.

The wet sponge method uses a direct current source, typically in the range of 9 to 90 volts, a sponge dampened with water containing a wetting agent to lower surface tension, and a ground return connected to bare substrate. The sponge is passed over the coating slowly, on the order of 0.3 metres per second, and a holiday completes the circuit and triggers the alarm. The sponge must be damp, not dripping; free water running across the surface produces false alarms and can bridge into an adjacent area.

High-voltage testing uses a pulsed or continuous DC source and a conductive electrode, brush, spring or rolling, moved across the surface at a controlled rate with the substrate grounded. A discontinuity produces a visible and audible spark. Because the technique works by breaking down the air gap at the flaw, it will equally break down sound coating if the voltage exceeds what the film can withstand, which is why voltage selection is the entire discipline.

Setting the voltage from measured thickness

SP0188 derives test voltage from film thickness by formula. For coatings up to 1 millimetre, 40 mils, the voltage is 525 times the square root of the thickness in mils, or 3,294 times the square root of the thickness in millimetres. Above 1 millimetre the constants change to 1,250 and 7,843 respectively. Both forms give the same answer within rounding; the mils form is the one field crews actually use on the back of a tally sheet.

The step at 40 mils is worth understanding before it happens to you. At 40 mils the lower formula returns roughly 3,320 volts and the upper formula returns roughly 7,900 volts. Two inspectors testing the same coating, one treating it as the top of the thin range and one as the bottom of the thick range, will apply voltages differing by more than a factor of two. The specification has to say which band a nominal 40 mil system falls in, or the argument happens on site with the applicator watching.

The word measured matters more than the formula does. Thickness for voltage selection comes from gauge readings taken on the work in accordance with the applicable dry film thickness measurement practice, not from the product data sheet. A system specified at 20 mils and applied at 32 mils will be under-tested if 20 is used, and a system specified at 32 mils but applied at 18 mils will be burned through. Both failures are common and both are avoidable in the same five minutes.

Coatings you cannot test this way at all

Holiday testing depends entirely on the coating being an insulator. Coatings that conduct do not merely give inconsistent results; they give continuous indications across sound film, and crews respond by winding the voltage down until the alarm stops, at which point the test detects nothing at all but still gets recorded as performed and signed.

The usual offenders are inorganic and organic zinc-rich primers, which are deliberately conductive so they can protect galvanically, and linings pigmented with carbon black, graphite or metallic flake. A zinc primer cannot be spark tested on its own. It can be tested once the insulating intermediate and topcoat are applied, with the voltage then set from the total system thickness. Testing at primer stage and testing the finished system are different tests answering different questions, and only one of them is possible.

Very thin films are the other exclusion. A 50 micron shop primer will not survive any meaningful spark voltage and is a wet sponge subject if it is tested at all. Moisture-cured and solvent-retaining coatings that have not reached full cure are equally poor candidates, because residual solvent leaves the film conductive and the test reports holidays that will not exist a week later. Cure schedule before test schedule is the rule that prevents this.

The conditions that manufacture false indications

A surprising share of reported holidays are artefacts of the conditions rather than defects in the coating. Surface moisture is the single biggest cause: dew, condensation from a cold substrate, or a surface wiped down and not dried will carry current across the film and produce indications that move as the surface dries. High-voltage testing requires a dry surface, and a surface temperature comfortably above dew point is the practical control.

Conductive contamination does the same thing. Soluble salt residues, overspray of conductive materials, damp insulation debris and finger contamination on a tank lining all provide paths the tester will find. So does an inadequate ground: a clamp bitten onto painted steel, or onto a flange face with a gasket in between, leaves a high-resistance return that suppresses genuine indications rather than creating false ones, which is considerably harder to notice.

Then there is the operator. Travel speed that is too fast misses small holidays because the electrode does not dwell long enough to break down the gap. Travel speed that is too slow, or an electrode held stationary at a point, will initiate a discharge through sound film. Edges, weld crowns, bolt heads and internal corners concentrate field strength, so induced damage happens there first, which is also exactly where genuine holidays cluster because film thickness is hardest to hold there.

Acceptance criteria come from the owner, not the standard

This is the point that ends most specification arguments once it is understood. SP0188 tells you how to detect a discontinuity. It does not tell you how many are tolerable, whether an area is rejected, how the holiday is repaired, or whether the repair is retested. Every one of those decisions belongs to the coating specification written for the job, and if the specification is silent, the parties will fill the silence differently.

A workable specification states the coverage, commonly 100 percent for immersion service linings and tank floors and a defined percentage for atmospheric service; the technique to be used and the voltage or the formula used to derive it; the permitted holiday count; the repair procedure including surface preparation and feathering of the repair area; the cure period before retest; and the record to be produced. Written that way, the inspector and the applicator argue about the coating rather than about the standard.

Where the specification is silent, the practical default for immersion linings is 100 percent coverage and zero permitted holidays, because a single pinhole in a lining is a corrosion cell with the entire tank surface acting as its cathode. Atmospheric coatings are rarely holiday tested at all outside critical areas. Getting the wording right before bid is the cheapest intervention available on a coatings project, and it is a normal part of Level III technical authority support.

Instrument verification, safety and the records that survive

The tester's output voltage must be verified against a calibrated kilovoltmeter and the verification recorded, because output drifts with battery state, electrode wear and cable length. Many specifications require verification at the start of each shift. A tester reading 5,000 volts on its own display while delivering 8,000 is burning the coating it is meant to inspect, and there is no way to know that without measuring the output independently.

Safety is not a footnote on this test. High-voltage testing inside a tank or vessel takes place in a confined space that may hold solvent vapour from the coating just applied. Spark testing an incompletely cured lining in an unventilated tank is an ignition source in a flammable atmosphere. Lower explosive limit monitoring, forced ventilation and a hold on testing until vapour has cleared are the controls, and they belong in the inspection and test plan rather than in the incident report.

The record that survives an argument identifies the item and area tested, the measured dry film thickness range used to set the voltage, the technique, the instrument and its verification, the voltage applied, the number and location of holidays found, the repair reference and the retest result. Reports assembled that way pass an independent report review without a query, and they are far easier to retrieve years later when they sit in an inspection management system rather than in a site folder.

How the coating inspection team is deployed

Lining and coating campaigns run on the applicator's schedule, which means inspection has to be present continuously rather than by periodic visit. Atlantis mobilises inspection personnel from Houston and Hyderabad to the fabrication yard, the coating shop or the tank site, and once mobilised the team remains deployed for the duration of the contract rather than flying in for hold points and hoping the schedule held.

That continuity is what makes hold points work at all. The same inspector who verified the surface preparation and recorded the soluble salt result is the one who measures film thickness, sets the holiday test voltage from those readings, and signs off the repairs. Handing a lining job between rotating inspectors is how a coating ends up accepted against three different readings of the same specification.

Where the wider scope includes weld examination of the substrate before lining, personnel are qualified and certified under a written practice to SNT-TC-1A or ISO 9712, and NDT training and certification can be delivered against your own written practice rather than a generic syllabus. To scope a coating and lining inspection package, describe the asset and the governing specification.

Repair coatings and the DFT gradient that produces a false pass

A repair patch on an in-service coating is rarely applied to a uniform dry film thickness across its full area — DFT typically feathers down toward the patch edge where it overlaps the original coating, and a holiday detection reading taken only at the patch centre, where DFT is thickest and the coating most defect-free, can pass while the feathered edge, where the coating is thinnest and most likely to carry an undetected discontinuity, is never tested at the voltage appropriate to its actual measured thickness.

This produces a specific and recoverable error: the test voltage under SP0188 is calculated from the measured DFT, and a single voltage calculated from a centre-of-patch reading and applied uniformly across a feathered repair will be systematically too high for the thin edge zone — a condition that itself risks damaging a coating that is otherwise sound, or too low to properly test a discontinuity in a thicker zone measured elsewhere on the same patch.

A defensible repair holiday test takes DFT readings across the repair area including the feathered transition, not only at the patch centre, and sets the test voltage from the thinnest zone actually being tested in that pass — accepting that a single repair patch with a genuine DFT gradient may need more than one calculated voltage if it is tested in more than one zone. Recording only a single pass or fail result for the whole patch, without the DFT readings that justified the voltage used, leaves a later reviewer unable to tell whether the test was actually valid for the zone that eventually leaks.

Request a consultation