Floor Screed Moisture Readings Before Laying: A Field Guide

Floor screed moisture readings before laying means measuring the relative humidity inside the slab, not on top of it, at a defined equivalent depth, and writing the result down in a signed protocol. Drill or place your probes at 40% of the slab thickness where the slab dries upwards only, or 20% where it dries from both faces. Seal the hole, let it equilibrate (24 hours under ASTM F2170, longer under the Swedish RBK method), then read it at a stable temperature. Record the RH value, the depth, the temperature, the hole position on a plan, the instrument and its calibration date, the time and the date, with photos. If the reading is above the limit set by the adhesive or covering manufacturer, you do not lay. And if you lay anyway on someone else's say-so, get that in writing, because a signed moisture protocol is the only document that saves you when the floor cups fifteen months later.

Why this got worse in 2026

Build programmes have tightened. Concrete has not started drying any faster. The screed goes down late, the building is closed in but not properly heated, someone needs the handover date, and you are asked whether you can start Monday. You say the slab is probably not dry. Someone says it will be fine. You lay it. A year later the boards cup and the letter arrives with your name on it. Without a reading log you are arguing about memory. With one you are arguing about a document.

Surface readings are not moisture readings

A pin meter or a handheld capacitance meter tells you about the top few millimetres. That is the part that dries first and tells you least. A slab can read comfortably dry on the surface and sit above 90% RH forty millimetres down. Seal a covering over it and that moisture redistributes upward until it finds your adhesive. Those tools are fine for a fast sweep to find wet spots worth investigating. They are not evidence, and no adhesive manufacturer accepts them as proof you followed their instructions. What counts is relative humidity in a sealed cavity inside the concrete, at depth, at equilibrium.

Taking floor screed moisture readings before laying: where to drill

A drying slab is a gradient: driest near the drying face, wettest at the far end. The number you care about is the RH the slab will settle to after you seal the top and the gradient evens out. Equivalent depth is where today's reading equals that sealed-in average. The Swedish and American methods land in the same place:

• Drying in one direction only (on ground, on a vapour barrier, on insulation, on steel deck): measure at 40% of the thickness from the drying face. A 200 mm slab on ground gives 80 mm.
• Drying from both faces (a suspended slab open top and bottom): measure at 20% of the thickness from the face you drill. A 250 mm two-sided slab gives 50 mm.
• Bonded screed over an older slab: take 40% of the screed thickness, plus a second, deeper reading in the base slab. A screed bonded to a wet base does not stay dry.
• Heating loops cast in: know where the pipes run before the drill goes in. Scan it. A moisture reading is cheaper than a heating engineer.

Do not average a dry reading and a wet one into an acceptable one. Each hole is its own verdict.

How many holes, and where

ASTM F2170 gives the clearest published baseline: three test locations for the first 100 m² (1,000 sq ft), and one more per additional 100 m². That is a floor, not a ceiling. Add holes where the risk is:

• The middle of any large pour, furthest from every edge
• Anywhere ponding, snow, rain or a burst pipe ever got in, however long ago
• Against external walls and under windows where a leak may have run
• Either side of a joint, or between two pour dates
• At any later patch or infill, since it has its own drying clock
• In cold or shaded corners the dehumidifier never reached

Letting the readings stabilise (this is where people cheat)

Drilling releases heat and exposes fresh concrete. The air in the cavity needs time to equilibrate with the pore moisture around it. Read too soon and you get a low number that flatters everyone and protects nobody. Two things stabilise, not one: the hole, then the probe inside it.

• Drill dry, and vacuum the hole out. No water. Brush and vacuum until no dust remains, or the dust becomes the thing you are measuring.
• Line and seal. Fit the sleeve or tube, seal it into the concrete and cap it. An unsealed hole reads the room, not the slab.
• Wait for the hole. ASTM F2170 requires 24 hours after the hole is prepared before the reading is taken. The Swedish RBK borehole method is stricter and uses longer equilibration windows; take the current figure from the RBK manual, not from a WhatsApp message.
• Then wait for the probe. If you insert the sensor at the end of the wait rather than at the start, it needs its own settling period. Follow the sensor maker's stated time.
• Temperature steady, slab at service conditions. ASTM F2170 wants the building at service temperature and humidity for 48 hours before testing. RH and temperature are coupled: roughly 1 °C of error costs you about 0.5 %RH. A slab read at 12 °C in an unheated shell and then handed a certificate is a fiction.
• Confirm, do not discard. If a reading looks surprisingly good, take it again the next day. Two readings a day apart that agree is a result. One reading is a rumour.

The limits: 75, 85, and why the numbers disagree

This is where installers get confused, because the internet mixes markets. The number is not universal. It depends on the standard, the method, and above all on what you are sticking down.

Sweden sits higher partly because the Swedish approach reasons explicitly about the critical moisture level of the actual build-up, and partly because equivalent-depth measurement already predicts the sealed-in condition rather than describing today's surface. That does not make 85 a licence. The binding number is always the lowest one: the standard, the adhesive data sheet, the covering data sheet, the specification. If the vinyl maker says 75%, then 75% is your number, whatever anyone on site says.

What a moisture protocol has to contain

A protocol is not a number in a text message. If it cannot be handed to a loss adjuster three years from now and understood without you in the room, it is not a protocol. Every entry needs:

• Site, room and a marked-up plan with each hole numbered. Hole 4 means nothing without a plan showing where hole 4 is.
• Slab thickness, construction and drying direction, plus how you established the thickness. This justifies the depth you chose.
• Measured depth per hole in millimetres, and the equivalent-depth calculation behind it.
• The RH value and the concrete temperature at the moment of reading, recorded together.
• Timestamps for both events: when the hole was drilled and sealed, and when it was read. The gap is your proof that you waited.
• Instrument make, model, serial and calibration date, plus any salt-solution reference check. An uncalibrated instrument produces an unusable protocol.
• Ambient conditions: room temperature and RH, whether heating was on, whether dehumidifiers were running.
• Photos of each sealed hole with its number visible, the probe in place, and the display showing the value. A photo carries its own date.
• The verdict, in words: approved for laying with this build-up, or not approved, and why.
• Names and signatures: who measured, who witnessed, who accepts the result. A site manager who signs cannot later say he never knew.

A digital template should force every one of those fields before it lets you close the job. That is what a self-inspection builder is for: OdinTask lets you build the moisture protocol once, attach photos from the phone in the field, capture the signature on site, and store the signed PDF against the job, so it is still findable in year three.

When the slab is not dry

Say it in writing the same day, not verbally in the site hut. Three honest options, one dishonest:

1. Dry it properly and wait. Heat plus air movement plus dehumidification, then re-measure. Not heat alone, which mostly drives moisture around.
2. Change the build-up. A surface damp-proof membrane or epoxy suppressant, warranted by its manufacturer for the RH you measured. Their number, their warranty, in writing.
3. Change the covering. A more vapour-open system tolerates what a fully sealed one will not.
4. Lay it anyway on a verbal instruction. This is the one that ends with your insurer.

If the client insists after being told, that is their right. Your job is the record: the reading, the manufacturer's limit, that you advised against laying, that they instructed you to proceed anyway. Signed, or at minimum emailed and acknowledged.

Why the protocol is the only thing that protects you

A cupped floor a year on has no clean answer. The moisture is gone by the time anyone looks. The evidence has evaporated, literally. What is left is documents.

With no protocol, the default assumption in almost every dispute is that the last trade on the floor failed to check. Checking was your competence. You cannot prove you checked. That is usually the whole case.

With a protocol (dated photos, an equivalent-depth calculation, a calibrated instrument, a signature) the conversation changes shape. Either it was under the limit, you laid it, and the cause lies elsewhere. Or it was over, you said so, and someone signed to overrule you. Both are survivable. Memory is not.

The protocol costs an hour of drilling and a couple of days of waiting. The claim costs a year and a floor. Start a free trial of OdinTask to keep moisture protocols, photos and signatures on the job card instead of in a phone gallery, or read more on the OdinTask blog.