The Two-Mil Disagreement on a Tank Shell
You are the third-party coating inspector on the exterior of a 250,000-gallon aboveground storage tank. The specification requires a zinc-rich primer at 3.0–5.0 mils DFT and an amine-epoxy intermediate at 5.0–7.0 mils. After the intermediate coat, the contractor's QC technician reports a spot average of 6.0 mils using a magnetic pull-off (Type 1) gauge he calibrated last week with shims laid on a 1/4 in steel plate. Your electronic Type 2 gauge, checked on the same plate, reads 4.2, 4.4 and 4.1 mils at that same location. The shell is 3/8 in plate, the area is in direct afternoon sun, and the steel surface temperature is 128°F. The contractor wants your signature so the finish coat can be applied before the shift ends.
1. What are the most likely causes of the 2-mil discrepancy, and how do you resolve it before any acceptance decision?
Read the worked answer
Neither reading is trustworthy yet - the measurement system itself is in dispute, so a decision now is unsupported.
Likely causes:
- Substrate thickness bias: Type 1 magnetic pull-off gauges are sensitive to plate thickness. He calibrated on 1/4 in plate but is measuring 3/8 in shell - the calibration is not transferable.
- Calibration method: shims over bare plate do not reproduce a blasted, curved, hot surface. Both gauges should be verified on the actual substrate, same curvature, same profile, with certified foils.
- Hot surface: at 128°F the epoxy is soft; a pull-off magnet or probe foot sinks in and biases readings, and electronic probes can drift.
- Technique/geometry: probe not perpendicular, readings too close to welds, nozzles or stiffeners (curvature/edge effect), or drag point read inconsistently.
- Gauge condition: expired calibration certificate, damaged probe, wrong mode.
Resolution: have both parties zero and verify each gauge on the actual shell plate with certified shims, then re-measure the identical spot sequence side by side. If the gauges still diverge, use a third method on a removed coupon (destructive gauge or cross-section) and send both instruments out for calibration. Record everything, including serial numbers.
2. Assume the discrepancy is resolved in your favor. How do you apply the DFT acceptance criteria to a 5.0–7.0 mil specification, and what is your disposition of this area?
Read the worked answer
Under PA 2 logic:
- Each spot measurement is the average of three gauge readings.
- Each spot average must lie between 80% of the specified minimum and 120% of the specified maximum - for 5.0–7.0 mils that window is 4.0 to 8.4 mils.
- The average of all spot measurements in the reference area must fall inside the specified range, 5.0–7.0 mils.
Your spot averages of roughly 4.1–4.4 mils sit above the 4.0 floor, so no individual spot is automatically rejected. But if the reference area average is about 4.3 mils, it is below the 5.0 mil minimum, so the area is nonconforming as a whole, even though individual spots squeaked by.
Disposition: do not release the area. Confirm you have sampled at the frequency the standard requires for the reference area (a single spot is not a survey), record every reading by location, and require the contractor to apply additional epoxy to bring the area average into the 5.0–7.0 range - or obtain a written owner concession if 4.3 mils is judged adequate for the service. A 4.3 mil intermediate under a finish coat is a barrier and durability risk, not a cosmetic issue.
3. What documentation and communication actions do you take, and can the finish coat go on today?
Read the worked answer
No - the finish coat should not go on today, for two independent reasons:
- The DFT disposition is unresolved; topcoating would bury a possibly thin intermediate coat and turn a cheap fix into a full removal.
- A 128°F steel surface in direct sun very likely exceeds the product data sheet's maximum application temperature. Hot substrate causes rapid solvent/pot-life problems for epoxies and, on a 3/8 in plate in sun, outgassing pinholes and poor wetting.
Actions:
- Issue a written nonconformance report / hold to the contractor and owner, citing the specification clause, the disputed readings and the reason the finish coat is being stopped.
- Record gauge makes, serial numbers, calibration dates, operator names, time of day, ambient and surface temperatures, dew point and locations of every measurement.
- Recommend shading or water-cooling the shell and shifting application to cooler hours, then re-verify surface temperature at the moment of application.
- Keep the record factual and non-accusatory; your authority is to inspect, document and report - the owner decides concession, the contractor decides rework.