After 20-plus years of QA inspection on Florida wastewater concrete, Corrosion Probe, Inc. (CPI) knows what scattered UT DFT readings tell you — and what they don’t. SSPC-PA 9 is supposed to help. The problem is the spec that invokes it usually doesn’t know what it’s asking for, and neither does the engineer who wrote it.
If you’ve spent any time inspecting protective linings on Florida’s wastewater infrastructure — digesters, clarifiers, wet wells, lift stations, force main structures, manholes — you’ve been there. You’re standing inside a concrete structure that took a crew two weeks to prep and coat. The applicator did a solid job. You watched the work. Now you’re running your UT gauge across the cured lining, and the readings are all over the place. Ninety mils here. A hundred and sixty there. Eighty-two in the next spot. The spec calls for a 100-mil minimum, references SSPC-PA 9, and offers no additional guidance.
So, what do you do with that data?
If you don’t understand what’s driving that scatter, you’re going to make the wrong call — either you accept a job that has very thin areas, or you reject a perfectly good lining because you’re chasing numbers that were never going to be consistent on that substrate to begin with. CPI has seen both happen. Neither outcome serves the owner, which is the whole point of having an independent inspector on the job.
Here’s what the ultrasonic gauge manuals don’t emphasize enough: the gauge doesn’t care what you’re trying to measure. It emits an acoustic pulse, listens for the echo off the coating-substrate interface, and calculates thickness based on that travel time — using whatever acoustic velocity you calibrated it to. When the substrate is smooth, dense, and uniform, that works beautifully. When the concrete substrate is abrasive blast-cleaned or hydro-blasted to a CSP 3, 4, or 5 — which is what you’re looking at on virtually every wastewater immersion lining job in Florida — the substrate is anything but.
Every peak on that surface brings the substrate closer to the probe. Every valley pushes it away. The coating itself might be right on target, applied consistently by a competent crew using a properly calibrated plural-component rig, and your UT gauge will still bounce all over the place because the reference assumption — the concrete surface — is a landscape of highs and lows. That variability isn’t coating inconsistency. It’s topography. If you can’t tell the difference, your inspection report is going to cause problems.
Add to that the reality of Florida wastewater concrete specifically: older cast structures with carbonated surface layers, biogenic acid attack from H₂S that’s chewed up the paste matrix and left exposed aggregate, surfaces that remain damp even after you’ve followed the dew point protocol, and multi-coat systems where each product has a different acoustic velocity. Every one of those variables degrades what the gauge is telling you, and none of them has anything to do with how the contractor applied the coating.
Most Florida wastewater immersion lining specifications require a cementitious resurfacer prior to the primary epoxy system. Products like Tnemec N218 MortarClad fill bug holes and voids and bring the substrate nominally to plane. Good practice — CPI isn’t arguing against it. But here’s what the spec almost never acknowledges: that resurfacer typically goes on at about 1/8 inch. That’s 125 mils. You now have a substrate preparation layer roughly the same thickness as the epoxy body coat on top.
And it doesn’t go on at a uniform 125 mils. A trowel-applied cementitious product does exactly what it’s designed to do — it fills the lows and feathers on the highs. In a deep bug hole or form void, you might get 200, 250 mils of resurfacer. On an aggregate crown or a formed surface ridge where the concrete was close to plane, it feathers to near-zero. That variability is understood from a substrate prep standpoint. From a UT measurement standpoint, it’s a problem you’re not going to solve by taking more readings.
Your UT gauge, calibrated to the acoustic velocity of the epoxy body coat, can’t tell where the resurfacer ends and the epoxy begins. It reads a composite travel time through a stack of materials — each with a different acoustic velocity — and converts it all into mils using a single velocity. In a spot where the resurfacer filled a 200-mil void, your gauge reads high. Not because there’s extra epoxy on the wall, but because there’s 200 mils of cementitious material with a different acoustic signature beneath the epoxy, and the gauge counts it all.
Note: Advanced ultrasonic gauges with multiple-echo mode and velocity profiling can sometimes identify individual layer boundaries, but most standard field inspection gauges — including the PosiTector UTG in single-echo mode — can’t make this distinction. The practical field implication remains the same.
On a peak where the resurfacer feathered to 10 mils, your reading is much closer to the actual epoxy DFT. Here’s what confuses people: your highest gauge readings appear over your deepest bug holes, but that’s NOT because there’s more epoxy there. It’s because the resurfacer is thickest in those voids. A low reading over a high spot doesn’t mean thin epoxy — it means thin (or zero) resurfacer under relatively uniform epoxy. The gauge is reading a composite of both materials, so high composite readings don’t guarantee adequate protective coating. That’s the opposite of what the raw numbers imply: the gauge is tracking resurfacer depth, not epoxy film build.
The specification acceptance criterion of 100 mils minimum is being applied to composite numbers ranging from 115 to 310 mils — not because the epoxy film build varies by that margin, but because the resurfacer beneath it does. The gauge isn’t malfunctioning. The specification writer is requiring you to evaluate a number they don’t understand against a criterion you can’t meaningfully enforce with that instrument on that substrate.
SUBSTRATE CORRECTION — THE PRINCIPLE:
Where the prepared substrate is rough or irregular, the only way to isolate coating thickness is to establish a baseline before coating: take and record readings on the prepared substrate. Net dry film thickness is then the reading taken over the coated surface minus the substrate reading at that same location.
— the substrate correction approach, SSPC-PA 9: Measurement of Dry Coating Thickness Using Ultrasonic Gages
This is the step that most specifications referencing SSPC-PA 9 fail to require. Without documented substrate-corrected readings taken on the prepared concrete or cured resurfacer surface before the first coat is applied, the inspector is working from composite UT data with no defensible baseline, and no post-application reading can recover a baseline that was never taken.
There are two measurement paths in practice — and they matter enormously when a cementitious resurfacer is in the stack.
Option 1: Substrate Correction (Preferred for Resurfaced Concrete). Substrate correction directly addresses this problem. Before applying any coating, take and record UT measurements directly on the prepared concrete surface. Those readings capture the substrate’s acoustic behavior — the profile, porosity, and aggregate — and give you a location-specific baseline. Net DFT is then your post-application reading minus the substrate reading at that same point. When it’s done right, you’ve isolated the coating contribution from the substrate contribution. That’s meaningful data.
On a system with a resurfacer, best practice is to take substrate readings at two separate hold points: first on bare, prepared concrete, then on the cured resurfacer before the first epoxy coat. The second set is the relevant datum for the protective lining — the first set describes a surface that no longer exists after the resurfacer is applied. Mark or reference each measurement location — grid coordinates, photos, or physical marks — so you can return to the same spot before and after resurfacer curing.
Option 2: Direct Measurement. Direct measurement of the cured coating without substrate readings is the other path, provided the gauge is calibrated per the coating manufacturer’s specification for that product. This is a legitimate approach when the substrate is straightforward, and the coating system consists of a single material.
The problem: on a cementitious resurfaced substrate, the direct method inherits everything described above — a composite reading through two materials of unknown relative thickness, calibrated to the acoustic velocity of only one of them. The method is legitimate, but the number it produces can’t be cleanly compared to a 100-mil criterion written for the epoxy system alone.
That’s why the substrate correction method is the more defensible choice for resurfaced concrete: it gives the specification exactly what it’s asking for — the epoxy contribution — rather than a composite through a variable stack.
In CPI’s field work, the projects where substrate correction has been done completely can be counted on one hand — with pre-application readings documented at the correct hold points. The reason is straightforward: the spec doesn’t require it. The PA 9 reference in most Florida wastewater coating specifications is a single line — “measure dry film thickness in accordance with SSPC-PA 9” — with nothing after it. No substrate reading protocol. No hold point definition. No calibration standard for each product in a multi-coat system. No distinction between the bare concrete stage and the resurfaced surface.
When those substrate readings weren’t taken before the coating went on, you are in direct measurement territory — and the composite readings you have are real data with an ambiguous relationship to the epoxy system alone. PA 9 is still in play. The direct method is still valid. But without the substrate correction, you’re working from a composite number and applying it against a criterion that was written for something cleaner.
When a spec says the resurfacer shall bring the surface to plane, it does not mean the surface is now flat. It means the gross irregularities — the deep bug holes, the form tie voids — have been filled. The troweled surface of a cementitious resurfacer is still textured. It still has aggregate relief. It still varies.
Engineers writing specs sometimes assume that specifying a resurfacer resolves the measurement problem. It doesn’t. It changes the measurement problem — and in some ways makes it harder, because now you’ve got that additional acoustic layer of variable thickness in the stack. A clean concrete substrate, even a rough one, is at least one material. A resurfaced concrete substrate consists of two materials of unknown relative thickness. The specification that doesn’t account for that hasn’t solved anything.
When the PA 9 hold points aren’t observed — which is most of the time — CPI has found that material quantity reconciliation is a useful cross-check. It’s not the answer, but it tells you something the gauge alone cannot. The concept is straightforward: you know how much material was applied, you know the manufacturer’s theoretical coverage rate, you know the surface area, and you can apply a reasonable loss factor for the application method. From those inputs, you can back-calculate the average DFT applied to the structure for each coat, completely independent of the UT gauge and all its substrate complications.
The key is to do this per coat, per product — not as a composite. The resurfacer gets its own calculation using its own batch records and PDS coverage rate. The body coat gets its own calculation. The topcoat gets its own. When you run the numbers separately, you can see exactly what each layer contributed, independent of what the gauge reads across the stack.
MATERIAL QUANTITY RECONCILIATION — AVERAGE DFT PER COAT
Avg DFT = [Gal Applied × (1 − Loss Factor) × Theoretical Coverage (ft²/gal @ 1 mil)] ÷ Area (ft²)
Gallons Applied: From batch records and container inventory — tracked separately per product, per coat
Loss Factor: 0.10–0.20 for plural component airless spray/trowel application
Theoretical Coverage: From manufacturer PDS — pull it per product; 100%-solids materials like N218 MortarClad and G434 both rate at 1,604 ft²/gal @ 1 mil, but any product below 100% solids will not
Surface Area: From project drawings, field-confirmed; consistent across all coat calculations
Worked Example — 100-mil Body Coat Over 125-mil Resurfacer, Wastewater Digester
Resurfacer — Tnemec N218 MortarClad
Interior surface area (walls + floor, field-verified): 1,450 sq ft — used consistently across all coat calculations | Theoretical coverage (from PDS): 1,604 ft²/gal @ 1 mil | Gallons applied (batch records): 248.0 gallons | Loss factor (trowel/spray): 10% | Effective gallons on surface: 223.2 gallons | Calculated avg DFT — Resurfacer: 247 mils (nearly double the specified thickness)
Body Coat — Tnemec Series G434
Theoretical coverage (from PDS): 1,604 ft²/gal @ 1 mil | Gallons applied (batch records): 128.0 gallons | Loss factor (plural comp. airless): 15% | Effective gallons on surface: 108.8 gallons | Calculated avg DFT — Body coat: 120 mils
Raw UT composite readings on this digester — a different structure from the 115-to-310-mil example earlier — ranged from 190 to 380 mils. That range is driven by the variable resurfacer depth beneath the epoxy — not by inconsistent coating application.
Those raw UT readings — 190 to 380 mils — look alarming against a 100-mil spec. The reconciliation explains why: the resurfacer averaged around 247 mils, with real variability from near-zero on the peaks to 250-plus in the voids. The body coat reconciled at 120 mils — slightly under target, worth documenting and discussing, but not the out-of-control application the raw gauge numbers suggest. The material record tells a coherent story that the gauge can’t on its own.
Continuity testing per ASTM D4787 doesn’t go away because the reconciliation looks good. ASTM D4787 is the right standard for nonconductive linings on concrete; AMPP SP0188 (formerly NACE SP0188) is for conductive (metal) substrates. For a wastewater-immersion thick-film lining, high-voltage continuity testing is required; no quantity calculation can find the localized pinholes and thin spots that a high-voltage spark test will. All three legs of the stool —substrate-corrected thickness measurement, per-coat material reconciliation, and high-voltage continuity testing—matter.
What material reconciliation gives you is a separate volume-based estimate for each layer — independent of the UT gauge. It can flag problems the gauge alone might miss, but it can’t replace proper thickness measurement or tell you where a deficiency is located. It’s a cross-check, not a solution.
IF YOUR SPECIFICATION REFERENCES SSPC-PA 9 ON A RESURFACED WASTEWATER CONCRETE SUBSTRATE, IT SHOULD BE ABLE TO ANSWER ALL OF THESE:
If the engineer who wrote the spec has never stood inside a wet well or digester with a UT gauge in one hand and a column of readings that don’t make sense in the other, that’s worth understanding before you accept or reject work based on those numbers.
PA 9 is a good standard. It was written to cover concrete among other substrates. The concrete inside Florida’s wastewater infrastructure is some of the most demanding surface you’ll find in protective coatings work — rough, porous, H₂S-degraded, biologically attacked, and now overlaid with a resurfacer layer that complicates every acoustic measurement you take. It deserves a specification that takes the measurement challenge seriously at each stage of the system, not one that drops a standard citation into the coating section and moves on.
Running the per-coat material reconciliation in parallel with your UT measurements, from the first batch mixed, gives the owner a coherent and defensible record of what’s on their structure. That’s what they’re paying for — not just a column of numbers, but someone on site who can explain what those numbers mean and what’s driving them.
If your specification doesn’t require documenting substrate readings before the first coat, it isn’t a specification — it’s just a number generator.
