Moisture and Adhesion Testing Before Coating

Purpose

Turn "it looks dry" and "it seems solid" into numbers a second person can check, so the decision to coat is defensible months later. Two calls in this trade hang on this: whether a substrate is dry enough to take a film, and whether the coating already on it will hold the one you are about to add. Getting either wrong produces the same visit, a full strip of an elevation or a wall, but the record is what decides whether the shop pays for it or the customer does. A reading with no location, no basis and no instrument check is not evidence.

Scope

Pre-coating measurement on wood, gypsum board, plaster, concrete and previously coated surfaces in residential and small commercial work, from instrument verification through the go or hold call.

It does not cover the wash and drying that produces the condition being measured (exterior wash and prep day standard), the prep and deglossing that follows a pass (interior surface preparation standard), repair of what a failed reading reveals (wood repair and drywall repair standards), tracing an active leak to its source, which is a roofing or plumbing scope, or flooring moisture qualification, which runs different test methods entirely.

Roles and responsibilities

Procedure

1. Verify the instrument before you trust a number from it. Check the meter against its own reference standard or calibration block, at the value printed on that block, and confirm the battery is above the meter's low threshold. Acceptance: the meter reads its reference within the tolerance the manufacturer's manual states, and that manual is where the tolerance comes from rather than memory. What wrong looks like: a meter that reads low across a whole house, which releases wet substrates and is invisible until they fail. Stop rule: a meter outside its stated tolerance is tagged out of service and replaced; readings are not adjusted mentally to compensate. Hazard: none at this step, it is a bench check done before the ladder comes off the van, and it is first because everything downstream inherits it.

2. Pick the method to the substrate and name the basis of the number it gives. A pin meter driven into wood gives percent moisture content of that wood at the pin depth. A pinless meter reads a volume beneath the pad and gives a relative scale on non-wood materials, sometimes called a wood moisture equivalent, which is a comparison number rather than an absolute percentage. Acceptance: the record states which meter, which scale, and at what depth or setting each reading was taken. What wrong looks like: a pinless number written as a percentage and compared against a pin-meter gate, which is two different quantities wearing one symbol. Stop rule: if you cannot state the basis, the reading does not go in the record. Hazard: pins are driven into a wall cavity that may carry cables, so pin readings stay in solid members or at known depths, and no pin goes into a wall face near a device box or above a switch.

3. Take readings at the wettest points you can reach, and state how many and where. Read the low points within about 2 ft of grade, north-facing surfaces, ends and joints, under downspouts and around any stain, and take at least three points per elevation or per wall. Acceptance: every point recorded with its location, and the gate applied to each point individually. What wrong looks like: three readings averaged into one number, which is how a wet band at the bottom of a wall gets released. Stop rule: any single point over the gate holds that elevation or wall; nobody averages, and nobody re-reads until they find a passing spot. Hazard: reaching low points means kneeling next to grade where the ladder feet also sit, and reaching high points means the ladder, so readings are taken with both feet on a level surface and the meter in one hand, never with a meter and a light in the same hand on a ladder.

4. Apply the gate, corrected for what the meter's manual says to correct for. Use a shop default of 15 percent or below for exterior wood and 12 percent or below for interior wood and gypsum, with the coating manufacturer's data sheet governing where it states a different figure. Apply the species and temperature corrections in the meter's manual, and write the raw reading, the correction and the corrected value as three separate figures. Acceptance: three figures printed per corrected reading, and the corrected value compared against the gate. What wrong looks like: a corrected value in the record with no raw reading behind it, which nobody can re-check. Stop rule: a reading that clears only after a correction you cannot name in the manual does not clear. Hazard: none, this is arithmetic done off the readings, and it is named as its own step because a correction mentioned and never printed is a correction nobody applied.

5. Run the plastic sheet test on concrete and masonry rather than a pin. Tape an 18 in square of clear polyethylene tightly on all four edges to the surface, leave it in place at least 16 hours, and read it per ASTM D4263, in the edition your specification references. Acceptance: no condensation on the underside of the sheet and no darkening of the substrate beneath it when the sheet is lifted. What wrong looks like: a faint darkened square that dries within a minute of lifting, which is a positive result and gets read at the moment of lifting, not after. Stop rule: any condensation or darkening holds that area, and the test is repeated after the substrate has had more drying time; the sheet test is a screen, so a positive result is a hold rather than a diagnosis. Hazard: taped sheeting on a floor is a trip hazard on a customer's route, so it is marked and, where the area is a walkway, the test moves to an off-route area of the same slab.

6. Test adhesion where you are recoating an unknown, glossy or suspect film. Run a cross-hatch tape test per ASTM D3359, in the edition your specification references, Method B on films at or under 5 mils and the X-cut Method A on thicker films, and run at least one test per substrate type per elevation. Acceptance: a classification of 4B or better on the scale that edition prints, recorded with its location and the method used. What wrong looks like: squares lifting cleanly in sheets at the lattice, which is a bond failure in what is already there. Stop rule: below 4B, no topcoat is sold as warrantable on that substrate until the customer has chosen between full removal and a bonding-primer system, in writing. Hazard: the knife is drawn away from the body with a fresh blade from a stable stance, and on any pre-1978 painted surface this cut disturbs paint film, making the test itself a regulated renovation under 40 CFR 745 Subpart E, run by a certified renovator with that rule's containment and cleanup under the RRP standard, not bare on a porch step.

7. Read where the break happened, not just the number. Look at the tape and the substrate: coating on the tape with bare substrate underneath is an adhesive failure at the bond line, coating split within its own thickness is a cohesive failure inside the film, and old coating coming off with new tells you which layer is the weak one. Acceptance: the record names the failure plane, not only the classification. What wrong looks like: a 2B result recorded with no plane, which leaves the next person unable to tell whether a bonding primer would even help. Stop rule: if the failure is adhesive at the substrate, a bonding primer is not a fix and the conversation is removal; if it is cohesive inside an old chalky layer, that layer comes off regardless of what goes on top. Hazard: none beyond the knife already covered, and lifted flakes generated by the test are collected rather than left on the ground where a pre-1978 job would spread them.

8. Record the result and issue the go or the hold to the crew in writing. Enter every reading, its location, the method and basis, the corrections, the adhesion classification and plane, and issue a go or hold per elevation or per room. Acceptance: a written disposition for every area the crew is scheduled to coat that day, and any hold accompanied by a delay note sent to the customer the same day. What wrong looks like: a verbal go for the house while one wall is still holding. Stop rule: an area with no disposition is not coated, and a crew that arrives to no disposition calls the lead rather than deciding. Hazard: none, this is a desk step, and it is named because the whole SOP is worthless if the number never reaches the person with the sprayer.

The record this produces

Per area: meter make and the reference check result with a date; method and basis for each reading; every reading with its location, raw value, correction and corrected value; the gate applied and its source; plastic sheet test start and lift times with the result; adhesion classification, method and failure plane with a photo of the lattice; the go or hold and who issued it. The failure-plane photo is the field that decides a dispute, because a customer's own prior coating showing on the tape is the whole argument.

One run of this procedure, filled in

A 1980s single-story exterior repaint, three elevations plus a painted concrete porch slab, testing on the Wednesday after a Monday wash.

Step 1 failed. The pin meter read outside the tolerance in its manual against its own check block, twice. The stop rule fired: the meter was tagged out of service and the backup meter was checked and used, rather than the first meter's readings being written down with a mental adjustment. That is the whole point of running step 1 first; had the crew skipped it, every number in this record would be unusable and the failures they released would surface next spring with nothing to argue from.

Step 2 recorded pin meter, percent moisture content, pins at the shallow setting. Step 3 took three points per elevation: south 10, 11, 12; east 11, 12, 12; north 12, 13, 16 percent, the 16 taken 20 in above grade under a downspout. Step 4 applied the manual's species correction for the cedar trim, printed as raw, correction and corrected for each. Against the 15 percent exterior gate, north holds on one point and is not averaged; that mean would have been 13.7 and would have released the wall.

Step 5 taped an 18 in sheet on the porch slab at 7 a.m. Wednesday and lifted it at 8 a.m. Thursday, which is 25 hours and clears the 16 hour minimum. No condensation, no darkening at the moment of lifting, so the slab passed.

Step 6 ran one cross-hatch per substrate: cedar trim came back 4B, the fiber cement body came back 5B, and the porch slab's old floor coating came back 2B. Step 7 read that 2B as adhesive failure at the concrete, with bare slab showing under the lifted squares. Under the step 7 stop rule a bonding primer is not the answer there, so the porch went back to the customer as a removal decision rather than a recoat.

Step 8 issued a go on south and east, a hold on north until its low point cleared, and a hold on the porch pending the customer's removal decision. North re-read Friday at 11, 12 and 13 percent, all three under the gate, and was released.

References

• ASTM D3359, rating adhesion by tape test, in the edition your job specification references, for both Method A and Method B.
• ASTM D4263, indicating moisture in concrete by the plastic sheet method, in the edition your job specification references.
• Moisture meter manufacturer manual, the authority on the reference check tolerance, species and temperature corrections, and pin depth or pinless scan depth.
• 40 CFR 745 Subpart E, the EPA Renovation, Repair and Painting rule, which reaches the adhesion test itself on pre-1978 painted surfaces.
• See related: the exterior wash and prep day standard, the interior surface preparation standard.