A client in Mandarin picked a luxury vinyl plank she loved, in a colour she had been carrying a sample of for a year. We put relative humidity probes into her 1974 slab, waited the manufacturer's required period, and read between 88 and 91 percent across the main living area. The product's limit was 85.
That is the entire argument for testing. A floor that goes down over a slab reading above its limit does not fail immediately. It fails in eighteen months, as cupping, as adhesive release, or as a smell nobody can locate. By then the warranty conversation is over, because the manufacturer's first question is always what the slab read.
Where the moisture comes from
Concrete poured on grade sits on soil, and soil in northeast Florida holds water. If a vapour retarder was installed under the slab, most of that moisture is blocked. A great many slabs poured here before the 1990s have nothing under them at all, so ground moisture moves up through the concrete continuously, for the life of the building.
This is why the age of the house tells you nothing useful. A fifty-year-old slab is not dry because it is old. It is as wet as the ground beneath it, every day.
How the test works
The defensible method is in-situ relative humidity testing: small holes drilled into the slab to a depth set by the standard, sleeves inserted, left to equilibrate, then read with a calibrated probe. It measures what is happening inside the slab rather than at the surface.
Surface meters have their place as a screening tool. We use one on a first visit to decide whether a room is worth testing properly. They are not a compliance test and no manufacturer will accept a surface reading in a warranty claim.
What you do with a high reading
- Apply a moisture vapour barrier: a roll-on or trowel-applied membrane over the prepared slab, which raises the tolerable reading substantially. Real cost, real added days, and it works.
- Change the assembly: porcelain tile over an uncoupling membrane is far more tolerant than engineered wood, and sometimes the better answer on a wet slab.
- Do nothing and accept the risk, which we will not do under our name, and which voids the product warranty.
The second measurable problem: flatness
The other half of the preparation story is how flat the floor is. The common tolerance for long plank is about 3/16 of an inch over 10 feet. For large-format tile it is tighter, often 1/8 of an inch, because a 48 inch tile bridges dips and shows every one of them as lippage at the joint.
We survey flatness with a straightedge across the whole area and mark the floor. That map is what the levelling quantity is priced from. On one Ponte Vedra job it came to 71 bags of self-levelling underlayment across 2,700 square feet. That is not a pleasant number, but it was a known number before the tile was ordered rather than a surprise on install day.
The third cause: transitions nobody planned
The last common failure is not a failure of the floor at all, it is a lip at a doorway. Different products build up to different heights. Carpet with a pad can be 20 millimetres. Glue-down vinyl can be 4. If nobody works out the heights before demolition, somebody improvises a ramp at the end, and it looks improvised because it is.
We record every existing transition height at survey stage and draw the detail for each doorway before demolition. Where a change is over about half an inch we fix the substrate rather than trim over the problem, because a half inch lip is a genuine trip hazard for anyone using a cane or a walker.
What this costs you
Testing adds a few hundred dollars and a handful of days to the programme, and the days overlap with work happening elsewhere. Getting it wrong costs the floor, the labour, the furniture move and the argument. We have never regretted waiting for a reading.
Written for this demo site. Code references, tolerances and figures are illustrative and vary by jurisdiction, by product and by house. Always confirm with the authority having jurisdiction and a licensed design professional.
