Holiday Detection on a Freshly Coated Ballast Tank
A ballast tank has been coated with a two-coat epoxy system, nominal DFT 300 µm per coat, and is being inspected for holidays before flooding. The inspector uses a high-voltage holiday detector. The product data sheet lists a dielectric strength that corresponds to a test voltage of about 100 V per µm of DFT, but the inspector sets the unit to 3,000 V based on a generic 'tank coating' rule of thumb. During the sweep, several sparks jump to the steel at coating edges and around weld toes, and the inspector marks them as holidays. The applicator argues the sparks are false positives caused by the high voltage and that the coating is sound.
1. Evaluate the inspector's voltage selection and the applicator's 'false positive' argument.
Read the worked answer
The voltage selection is the problem. The data sheet gives roughly 100 V per µm of DFT. With a nominal two-coat system of 300 µm per coat (600 µm total), the appropriate test voltage is on the order of 60,000 V, not 3,000 V - but more importantly, the correct voltage must be set from the actual measured DFT of the coating, not a generic rule of thumb, and must stay within the detector and coating manufacturer's limits.
Wait - the numbers need care. A 3,000 V setting is far below what a 600 µm epoxy would require for a valid holiday test. At 3,000 V, the test is likely under-voltaged and would miss real holidays, not create false ones. So the applicator's 'false positive from high voltage' argument is backwards: 3,000 V is too low for this film thickness, not too high.
However, the sparks at edges and weld toes are a separate signal. Sparks concentrate at geometric discontinuities - sharp edges, weld spits, and thin spots - because the coating is thinner there. Those are not false positives; they are likely real holidays or thin-film defects at exactly the locations most prone to failure. The inspector should not dismiss them.
Correct approach: measure actual DFT at the spark locations, set test voltage per the data sheet and detector instructions, and re-test. Investigate each spark location as a potential thin spot or holiday.
2. The applicator says 'the coating passed visual, so the sparks must be instrument error.' How should the inspector respond and what is the disposition?
Read the worked answer
Visual inspection and holiday detection test different failure modes. Visual inspection confirms the film looks continuous and uniform; it cannot detect pinholes, thin spots, or discontinuities that are invisible to the eye. A coating can pass visual and still have holidays - especially at weld toes, edges, and around inclusions, which is exactly where the sparks appeared. So 'it passed visual' is not evidence against the holiday readings.
Disposition:
- Do not accept the coating as holiday-free based on visual alone, and do not accept the applicator's dismissal of the sparks.
- Verify the instrument: confirm calibration, correct voltage setting for the measured DFT, and correct probe/technique. If the voltage was wrong (3,000 V is likely too low for 600 µm epoxy), the test is invalid and must be repeated at the correct voltage.
- Re-test at correct voltage and re-examine each spark location. At edges and weld toes, measure DFT - thin film there is a common and real defect.
- Repair confirmed holidays: abrade, clean, and apply additional epoxy per the data sheet, then re-test.
- Document the voltage used, DFT at spark locations, and the repair/re-test results.
The inspector's job is to establish whether the sparks are real defects at the correct test voltage - not to accept or reject based on a visual pass or the applicator's assertion.