London-wide leak specialists
ResidentialCommercial
Get in touch
Book an engineer

Pressure testing underfloor heating before the screed goes down

Fill and purge the loops, pressurise to the figure the pipe manufacturer specifies, hold it long enough to separate a real loss from temperature movement, and then leave the system under pressure while the screed is poured — a pipe damaged by a shovel shows on the gauge the same minute rather than the following winter.

Leak Fix LondonLast updated

Flooring and plaster being made good after a leak repair

The cheapest underfloor leak is the one found before the screed

Once screed covers the loops, everything changes. Finding a leak means thermal imaging, tracer gas and a saw. Fixing it means breaking out floor, cutting a section, joining it, and reinstating. Finding the same fault the hour it happens, with the pipe still visible, means a coupler and ten minutes.

If you are mid-build, the pressure test is not a formality your installer signs off before moving on. It is the last moment the floor is cheap to open.

What the test is actually for

People treat it as one job. It is three, and they answer different questions.

StageQuestion it answers
After the loops are laid, before screedAre the pipe and the manifold connections sound as installed?
During the pour, pressure still onHas anything been damaged in the last few hours?
After the screed has setDid the loops survive the pour and the curing?

Skip the middle one and you lose the only stage that tells you when a pipe was damaged. That matters later, when a leak appears and nobody can say whether it was the pipe, the installer or the screeder.

What a proper test involves

The sequence does not vary much between systems:

  1. Fill each loop individually from the manifold and purge it. Air left in a circuit will come out of solution over the following hours and drop the gauge, and you will spend a morning hunting a leak that was never there.
  2. Pressurise to the figure in the pipe manufacturer's instructions. That figure is set by the pipe and fitting system you have actually bought, and it is usually well above working pressure so that a marginal joint fails now rather than later. Do not take a number off a forum, and do not take it from this page — we are not going to invent one. It is printed in the installation manual, and the manual governs.
  3. Hold it for the period the manufacturer specifies before you read anything into the gauge. A short hold catches a spraying joint. It does not catch a slow weep, and it does not let temperature settle.
  4. Walk every loop. The gauge tells you the system is losing water. It does not tell you where. While the pipe is visible, your eyes are the best detection tool on site.
  5. Record the reading, the time and the temperature at the start and end of the hold. A reading with no temperature next to it is close to useless, for the reason below.

Some manufacturers permit an air test instead of water, usually where there is a frost risk before the heating is commissioned. Air behaves differently — it is far more sensitive to temperature and it will not show you a wet patch — so only do it if your own system's instructions allow it, and follow their method rather than adapting a water procedure.

Why the gauge moves when nothing is leaking

This is the single most common false alarm on site, and it costs people days.

Water expands as it warms and contracts as it cools. A sealed loop has nowhere to put that change, so the pressure moves with it. A slab that sat in the sun all afternoon and then cooled overnight can drop the gauge noticeably by morning. Nothing has escaped. The same volume of water is simply taking up less room.

Screed makes it worse in both directions. Cement curing is exothermic: the pour generates its own heat over the first hours, which pushes the reading up, sometimes enough to alarm someone who has just been told to watch for a drop. Then the slab cools over the following days and the reading falls back. A test read at 7am against a figure set at 3pm the day before is not comparing like with like.

Two habits fix this:

  • Log the temperature with every reading — slab or ambient, consistently, the same way each time.
  • Compare readings taken at the same time of day, once the slab has stopped moving thermally.

The signature of a real leak is different. Temperature movement is a reading that falls and then comes back. A leak falls and keeps falling, and it does not recover overnight. If the gauge has dropped and stayed down across a full warm-cool cycle, treat it as a loss and look for it.

Leave it under pressure through the pour

Of everything on this page, this is the part most often skipped and the part that pays for itself.

The pour is when the pipe is most likely to be hurt. Barrows, boots, rakes, mesh, a dropped tool, a screed rail pinned through a loop. Pipe under a set slab hides all of that. Pipe under pressure does not: puncture it and the gauge falls in front of the person who did it, while the pipe is still exposed and the screed is still wet enough to work.

So keep the pressure on. Put the manifold gauge where the screeding gang can see it, tell them what a falling needle means, and have someone glance at it through the pour. If it drops, stop and find the damage while it is reachable. A coupler now, or a floor breakout later. That is the choice.

Keep it on after the pour as well, through the set, with a further check once the screed has cured. It costs nothing and it closes the gap between the last person off site and the first person to blame.

What to photograph, and why it matters in five years

Disputes about underfloor heating surface a long time after the build, usually when a floor gets wet and three parties disagree about whose pipe it was. The people who win those arguments are the ones with a file.

Before the screed covers anything, take:

  • Overheads of every room with the loops laid out, taken from enough angles that pipe runs can be traced afterwards. A phone photo from a step ladder is fine.
  • Measurements written into the shots — a tape laid across the floor, or a marked-up plan photographed next to the real thing. This is what stops someone drilling into a loop when they fit a stair rail or a kitchen island three years from now.
  • Close-ups of every manifold connection, with the loop numbering legible.
  • The gauge itself: at the start of the hold, at the end, during the pour, and after the screed has set. Time-stamped, which phones do automatically.
  • The pipe manufacturer's name and the product label on the coil. Later, that identifies the exact system and the exact test regime it required.

Alongside that, keep the installer's test certificate, the pipe manufacturer's instructions, and a dated note of who was on site for the pour. Back it up somewhere off the phone. A dispute years later turns on records, and the record is either taken in that one afternoon or it does not exist.

The same file makes a later leak cheaper. Called to a floor with a documented loop layout, we already know where the pipe runs and roughly where it does not, which narrows the search before any equipment comes out of the van.

When the pressure really has dropped

If you have ruled out temperature and the loss is real, work outward from the easy end:

  1. Isolate loops one at a time at the manifold. This is the advantage you have before the floor closes and the reason manifold valves exist. If the system holds with one loop shut, you have narrowed the problem to a single circuit.
  2. Check the manifold itself. A large share of "underfloor leaks" are at the connections, not in the pipe, and the manifold is the one part that stays accessible.
  3. Look for the obvious wound along the suspect loop. Pre-screed damage is usually a puncture or a crush, not a mystery.

Once screed is down, none of that is visible and the work changes. That is a detection job: thermal imaging to follow the warm pipe run through the slab, and tracer gas on a drained circuit where thermal cannot separate the loops. We have written both up in detail in thermal imaging leak detection and finding a leak under a concrete floor.

Where we fit, and where we do not

To be straight about it: we are not the people who install your underfloor heating or pour your screed, and a pre-screed pressure test is normally your installer's job under their own manufacturer's instructions. If your loops are still exposed and the system holds, you do not need us at all.

Where we come in is the other case — a system that will not hold, or a floor already closed over a fault. We pinpoint it through screed and tile, cut small, repair, and our own handymen and builders put the floor back. That work is on the underfloor heating leak repair page, with the symptoms that usually bring people to it set out under underfloor heating leak.

The commercial terms are the same as everything else we do: £150 an hour, agreed before we attend, no call-out fee, and if we do not find the leak you do not pay for the time spent looking. The full breakdown is on our pricing page. Where a leak has damaged a new floor, the detection visit comes with a written report prepared to support a trace-and-access claim — what the policy then covers is the insurer's decision, not ours.

Frequently asked

How long should underfloor heating be left under pressure?
Follow the pipe manufacturer’s instructions for the hold time, and treat that as the minimum rather than the target. Beyond that, the useful answer is: leave it pressurised through the pour and through the curing of the screed, and check it again afterwards. There is no cost to leaving it on and a large cost to taking it off early.
The pressure dropped overnight. Is that a leak?
Not necessarily. Water contracts as it cools, so a slab that warms during the day and cools at night will move the gauge on its own. Compare readings taken at the same time of day, with the temperature logged. A drop that recovers as the slab warms is thermal; a drop that keeps going and never comes back is a loss.
Can I pour screed over loops that are not pressurised?
It is done, and it is a bad trade. Pressurised pipe reveals damage the moment it happens, while the pipe is still exposed and fixable with a coupler. Unpressurised pipe hides the same damage under a set slab, where finding it needs detection equipment and fixing it needs breaking out the floor.
What pressure should I test underfloor heating at?
Whatever the pipe and fitting manufacturer specifies for your system, which is normally well above working pressure and is printed in the installation manual supplied with the pipe. We deliberately do not publish a figure here, because it varies by pipe type and manufacturer and guessing it can either miss a marginal joint or damage a sound one.
What should I keep after the test, in case of a dispute later?
Time-stamped photographs of the loops laid out with measurable references, close-ups of the manifold connections, gauge readings before, during and after the pour, the pipe product label, and the installer’s test certificate. Store it somewhere other than the phone it was taken on. Underfloor heating arguments surface years later and are decided on records.

Sources

No find, no fee · £150 per hour. The rate is fixed before we attend and does not rise for a difficult property, an awkward access or an evening visit. The total depends on how long the job takes; the rate cannot change.