Holiday Testing a Tank Lining Before Cure
A 10,000 gallon carbon steel tank is lined with 20 mils of amine-cured epoxy. The specification requires a 100% holiday test after full cure (7 days at 75°F), high-voltage DC testing, plus DFT and cure verification. On day 2 the contractor's technician runs a high-voltage detector over the shell, finds 14 holidays concentrated near welds, and begins patching them. The inspector sees that the ground cable is clipped to a painted lug, the voltage setting is still the value used on the previous job (a 12 mil coating), and the film feels soft, giving a pencil hardness of HB. A wet-sponge test run the same day on the shell-to-floor weld finds two indications the high-voltage test missed. The owner's representative says a holiday is a holiday, so patch them and move on.
1. Are the 14 indications valid holidays? Justify your answer from the test conditions.
Read the worked answer
No, they are not decision-grade results. Several controlled variables were wrong:
- Ground path: clipped to a painted lug, the return circuit is unreliable. A high-resistance or open ground causes missed indications and erratic indications; the ground must be on clean bare metal.
- Voltage: a setting chosen for a 12 mil film applied to a 20 mil film is too low and will miss defects. Voltage must be selected for the actual thickness and the standard being used; too high is equally wrong, because it perforates the film and manufactures holidays.
- Cure state: a soft film at day 2 has not reached the 7-day full-cure condition the specification requires. Soft coating is easily penetrated by the probe, producing false indications, and an uncured film is more conductive, which affects both detection and the meaning of any indication found.
- Location pattern: a cluster near welds is plausible as real defects, which is exactly why the result must be confirmed under proper conditions rather than dismissed. Disposition: stop the testing, stop patching, and re-test after full cure under controlled conditions.
2. Why did the wet-sponge test find indications the high-voltage test missed, and what does that tell you about method selection?
Read the worked answer
Two different mechanisms were run outside their intended conditions, so the discrepancy tells you the tests were invalid, not that one method is superior.
- Wet-sponge testing is a low-voltage method suited to relatively thin films, broadly up to about 20 mils or 500 µm, and it is sensitive to pinholes that a low-voltage high-voltage test can miss. It also gives false positives over contamination, over conductive or uncured films, and at welds where the film is thin or irregular.
- High-voltage testing is intended for thicker films, but only at a correctly selected voltage with a proper ground; under those conditions it finds defects the sponge cannot.
- Because the film was uncured, the higher conductivity of a soft amine epoxy can produce wet-sponge indications that disappear once the film cures. The judgement: method and settings come from the specification and the coating thickness and cure state. You do not resolve a discrepancy by choosing the result you prefer; you restore the controlled conditions and repeat the test.
3. Set out the disposition, re-test plan and documentation you would require before sign-off.
Read the worked answer
Disposition
- Reject the day-2 test results in writing; they do not demonstrate conformance and cannot support acceptance or the patching already started.
- Halt patch repairs until the film is cured and re-tested. Patching an uncured, solvent-rich film risks trapped solvent, poor bond and a repair that becomes the next defect.
Re-test plan
- Confirm cure at the specified 7 days at 75°F, or per the data sheet, using a documented cure check such as pencil hardness and a solvent rub.
- Record DFT to confirm thickness, since the voltage selection depends on it.
- Verify the detector: correct ground on bare metal, voltage set for the measured thickness, function checked against a known-defect reference if available.
- Then run the 100% holiday test, marking and mapping every indication and re-testing repairs.
Documentation
- NCR with times, technician names, equipment identification, settings, ground location, cure data and photographs.
- The reason the earlier results were rejected, the re-test plan and the acceptance criteria, so the file shows a reasoned resolution.
- To the owner's representative: patching false holidays adds unnecessary thickness and risk, while missing real ones means a through-film failure in service. The answer to 'patch and move on' is a controlled re-test, then patch only confirmed defects.