Summary
A tradesperson called out for "cold underfloor heating" needs to establish one thing before anything else: is the whole system not heating, or is most of the floor warm with one or two patches that stay cold? These are different fault families with almost no overlap in diagnosis. Total failure points to the demand chain, the boiler, the manifold, or (for electric systems) the supply and thermostat — all covered in the fault-finders linked above. A localised cold spot with the rest of the floor performing normally points instead to something specific to that section of floor: how far apart the pipe or cable runs are laid, how deep or even the covering above them is, whether that particular loop or zone is getting its fair share of flow, or whether something on top of the floor is blocking heat transfer.
This matters commercially as much as technically. A callout for "one corner near the patio doors never warms up" that gets treated as a manifold fault wastes a visit checking actuators and wiring centres that are working perfectly. The actual cause is very often a design or installation-stage issue baked into the floor structure — loop spacing widened to make the pipe run fit, a screed pour that came out thinner over one run, or perimeter insulation that wasn't tight against the wall. Some of these are cheap to fix (rebalancing a loop, moving a rug); others require lifting flooring or, in the worst cases, are simply an accepted limitation of the original design that can only be partly mitigated.
The most common misconception is that a cold spot means a broken pipe or cable. Genuine breaks are rare and almost always cause either a full loop/zone failure (no heat at all in that area) or, for electric cable, a tripping RCD from an earth fault — not a "slightly cooler" patch with everything else working. A patch that's cooler but not cold, with the rest of the floor performing as expected, is a heat output or heat transfer problem, not a broken circuit.
Key Facts
- Wet UFH loop centres — typically laid at 150mm, 200mm, 250mm or 300mm centres depending on room heat loss; wider centres (250–300mm) are used in lower-demand rooms to save pipe, but produce a visibly cooler floor surface between the pipe runs, most noticeable directly above the midpoint between two pipes
- Screed depth over pipe — for traditional sand:cement screed, typically 65–75mm cover over the pipe is specified (check the specific installer's design); for liquid/flowing screed, cover can be as little as 50mm — uneven screed depth means uneven thermal lag, and a high spot in the screed will read cooler at the same water temperature
- Perimeter/edge insulation gaps — a gap in the expansion strip or edge insulation lets heat migrate sideways into the wall/skirting rather than up into the room, producing a cold strip near that wall rather than in the room's centre
- Insulation board joints and thermal bridging — gaps or poorly-fitted joints in the insulation board beneath the screed allow heat to travel downward instead of up, reducing surface output directly above the gap
- Manifold balancing valves (wet systems) — each loop has a lockshield/balancing valve set at commissioning; if one loop's valve has drifted, been knocked, or was never correctly balanced, that loop runs starved relative to others, producing a cold patch matching that loop's floor coverage — see the balancing procedure in wet underfloor heating
- Air pockets confined to one loop or one section of pipe — unlike a whole-system air lock (which usually affects several loops and is caught at commissioning), a small trapped air pocket in a single loop's high point can produce a localised cold patch while the rest of the loop, and other loops, heat normally
- Furniture and floor coverings — sofas, beds, and especially thick rugs or high-tog underlay sitting directly over UFH pipework insulate the floor from the room, reducing the heat that actually reaches the air above; manufacturers typically specify a maximum combined floor covering tog rating (commonly around 1.5 tog for efficient operation, lower for heat-pump-fed systems)
- Retrofit floating-floor / low-profile systems — panel or joist-notched systems (common in retrofits over existing timber floors) rely on the pipe or cable sitting in a routed or notched channel with good thermal contact to a spreader plate; inconsistent notch depth means inconsistent contact, and a shallow or poorly-seated section transfers heat less effectively
- Electric mat/cable overlap and gaps — the single most common cause of an electric UFH cold patch: during installation, the cable run should maintain consistent spacing across the whole area (per the manufacturer's fixed centres, commonly 50–80mm); a gap where the installer ran out of cable run and had to skip a section, or where the cable was diverted around an obstruction without maintaining spacing, leaves a cold strip — the cable must never touch or cross itself, which rules out simply "doubling back" tightly to compensate
- Electric UFH loop resistance test — an electrician can measure the installed loop resistance against the manufacturer's datasheet figure; a reading within tolerance across the whole loop doesn't rule out a spacing/coverage problem, since the cable is still intact and drawing correct current — it's simply laid unevenly
- Zone vs whole-floor design — some rooms are deliberately laid with two zones at different output (e.g. a lower-output zone under a fixed kitchen island or fitted furniture footprint) — confirm the original design/commissioning record before assuming a fault, as a "cold" patch may be an intentionally reduced-output zone
Quick Reference Table
Diagnosed the problem? Create a repair quote in minutes with squote.
Try squote free →| Symptom pattern | System type | Most likely cause | First check |
|---|---|---|---|
| Cold strip along one external wall | Wet or electric | Perimeter insulation gap | Inspect skirting/edge detail if accessible, thermal camera |
| Cold band running in a regular stripe pattern | Wet | Loop centres too wide for that area | Check design drawing/loop spacing record |
| One whole loop's floor area cooler than rest | Wet | Balancing valve mis-set or partial blockage | Check manifold flow meter for that loop |
| Cold patch that moves slightly warmer after 24–48h running | Wet | Small trapped air pocket | Purge that loop individually at manifold |
| Cold patch matches exactly where a rug/sofa sits | Wet or electric | Floor covering insulating the surface | Remove/relocate covering, re-check |
| Cold strip with no obvious pattern, near a retrofit joist area | Wet or electric (retrofit) | Inconsistent notch depth / poor thermal contact | Check installation record, consider thermal camera |
| Cold strip of consistent width, straight line | Electric | Gap in mat/cable run during install | Thermal camera to trace cable path, compare to layout diagram |
| Cold patch near a fixed obstruction (island, WC pan) | Electric | Cable diverted without maintaining spacing | Compare as-built layout diagram to obstruction position |
| Screed surface visibly uneven (feel with a straightedge) | Wet | Screed depth variance | Check screed thickness records / commissioning report |
Detailed Guidance
Decision tree: is this a cold-spot fault or a total-failure fault first?
START: Complaint is "part of the floor is cold"
|
|-- Is MOST of the floor heating normally, with one or more
| discrete cooler patches?
| |-- NO, nothing is heating anywhere --> STOP. This is a
| | total-failure fault, not covered here. Go to
| | [underfloor heating manifold not calling heat](/knowledge/fault-finder/underfloor-heating-manifold-not-calling-heat)
| | (wet, nothing calling) or
| | [underfloor heating electric fault](/knowledge/fault-finder/electrical/underfloor-heating-electric-fault)
| | (electric, nothing heating)
| |-- YES, most of the floor is fine, patch(es) are cooler
| but not stone cold --> continue below
|
|-- Is this WET (water-fed loops) or ELECTRIC (mat/cable) UFH?
|
|===================== WET UFH =====================
|
|-- Does the cold patch correspond to ONE manifold loop's
| known floor coverage area?
| |-- YES --> Check that loop's balancing valve and flow
| | meter reading against neighbouring loops.
| | Low/no flow --> rebalance or investigate
| | partial blockage.
| |-- NO, patch is smaller than a whole loop area or crosses
| loop boundaries --> continue
|
|-- Does the patch sit along an external wall or cold bridge
| (bay window, patio door threshold)?
| |-- YES --> Suspect perimeter/edge insulation gap or
| | thermal bridging at that detail
| |-- NO --> continue
|
|-- Does the patch sit under furniture, a rug, or thick
| underlay?
| |-- YES --> Remove/relocate covering, re-test after
| | 24-48h before assuming a structural cause
| |-- NO --> continue
|
|-- Has floor covering or use of the room changed recently
| (new rug, new sofa position, re-carpeted)?
| |-- YES --> Covering change is the likely cause
| |-- NO --> Suspect loop spacing design, screed depth
| variance, or a localised air pocket in that
| loop - check design drawing, consider
| thermal camera survey before any invasive work
|
|===================== ELECTRIC UFH =====================
|
|-- Is the cold patch a consistent straight strip or band?
| |-- YES --> Compare to as-built cable layout diagram -
| | likely a gap left during installation where
| | spacing wasn't maintained
| |-- NO --> continue
|
|-- Does the patch sit near a fixed obstruction (island unit,
| WC pan, shower tray) that the cable had to route around?
| |-- YES --> Cable likely diverted without maintaining
| | manufacturer's fixed centres near the
| | obstruction - confirm with layout diagram
| |-- NO --> continue
|
|-- Does the patch sit under a rug, dense underlay, or fitted
| furniture footprint?
| |-- YES --> Floor covering insulating the surface -
| | check combined tog rating against manufacturer
| | maximum (commonly ~1.5 tog)
| |-- NO --> Consider a retrofit low-profile system with
| inconsistent joist-notch depth, or request
| thermal camera survey and loop resistance
| test before considering any floor lift
Diagnosing without lifting the floor: thermal imaging
For both wet and electric systems, a thermal imaging camera run over the affected area after the system has been running steadily for at least an hour is the single most useful non-invasive diagnostic step. It reveals the actual pattern of heat at the surface — a regular striped pattern points to loop/cable spacing, a single localised cold blob points to an air pocket or a covering effect, and a cold strip that tracks precisely with a known cable/pipe route confirms an installation-stage gap rather than a random fault. Compare the thermal image to the original installation layout drawing wherever one exists — this is the fastest way to separate "designed this way" from "installed incorrectly."
Wet UFH: loop spacing and screed depth as a design issue, not a fault
Where a thermal survey confirms a genuine stripe pattern matching wide loop centres, this is frequently not a fault at all but a design compromise made to fit the available pipe run within a room, or a genuine under-specification for that room's heat loss. Retrofitting tighter loop centres means lifting the floor — rarely justified for a mild cold-strip complaint. Practical mitigation without floor removal: increase flow temperature slightly at the blending valve (within the 35–45°C gas boiler / 30–35°C heat pump range — see wet underfloor heating for detail), or accept the design limitation and manage customer expectations. Where the loop centres are genuinely non-compliant with the manufacturer's design guidance for that room's calculated heat loss, this is a design fault that should be raised with whoever specified the system, not treated as a repair callout.
Uneven screed depth is harder to fix retrospectively. If a thermal survey and cross-reference with the original screed pour records confirm a high spot, the only reliable long-term fix is local screed remediation — cutting back and reinstating to correct depth — which is invasive and should only be proposed where the cold spot is a genuine functional problem, not a minor comfort complaint.
Wet UFH: rebalancing a starved loop
If a single loop is confirmed starved of flow relative to its neighbours, follow the manifold balancing procedure: open all actuators, run the pump, and compare flow meter readings across all loops on that manifold. Adjust the lockshield valve on the affected loop (anticlockwise to open, increasing flow) until it matches the design flow rate for its length — longer loops need proportionally more flow to deliver equivalent output per metre. A loop that reads zero or near-zero flow despite the actuator being open points to a partial blockage rather than a balancing issue; see wet underfloor heating for blockage diagnosis and powerflush guidance.
Electric UFH: confirming a spacing/coverage gap without lifting flooring
Where the layout diagram is unavailable (common in older retrofit jobs), a loop resistance test alone will not reveal a spacing gap — the cable is still intact and will read correctly on a multimeter even if it's unevenly spaced across the floor. Thermal imaging combined with careful measurement of the cold strip's position against known room features (door thresholds, fixed units, pipe boxing) is the only reliable non-invasive method. If a genuine gap is confirmed and the floor covering allows (e.g. a tiled floor with the mat under a flexible tile adhesive bed, rather than embedded in a permanent screed), localised remedial matting can sometimes be installed in the gap — this is a specialist repair and should not be attempted without confirming the exact cable route first, to avoid piercing the existing element.
Furniture, rugs, and floor covering effects
Any thick rug, high-tog underlay, or large piece of upholstered furniture placed directly over an active UFH zone will measurably reduce the surface temperature reaching the room above it, while the floor structure itself (and the heating element within it) may be running entirely normally. This is not a system fault and requires no repair — advise the customer on the manufacturer's maximum recommended combined tog rating for floor coverings over their specific system (commonly around 1.5 tog total for standard wet/electric UFH, lower again for heat-pump-fed systems running at reduced flow temperatures) and suggest relocating heavy coverings away from known cold-complaint areas where practical.
Retrofit floating-floor and joist-notched systems
Low-profile retrofit systems — panel systems laid over an existing subfloor, or pipe/cable notched directly into joists beneath a floating floor deck — depend heavily on consistent, tight thermal contact between the heat source and the floor covering above. Variable notch depth (too shallow loses contact with the spreader plate or floor deck; too deep can leave an air gap beneath the pipe) is a common cause of patchy performance in these retrofit installations, particularly where notching was done by hand across an uneven joist run. Confirming this as the cause typically requires either an installation record showing notch depths, or, where none exists, careful thermal imaging correlated with known joist spacing.
Frequently Asked Questions
Is a cold spot always a sign the underfloor heating is broken?
No. Most cold-spot complaints, in both wet and electric systems, are caused by design or installation-stage variance — loop spacing, screed depth, a mis-balanced loop, or a floor covering blocking output — rather than a broken pipe, cable, or component. Genuine breaks almost always cause a full loop or zone failure, not a patchy "slightly cooler" area with everything else working.
Can I fix a cold spot without lifting the floor?
Often yes, or at least partially. Rebalancing a starved wet loop, adjusting flow temperature, purging a localised air pocket, or moving furniture and rugs off the affected area all require no floor disturbance. Genuine spacing or screed depth issues baked into the original installation usually can't be fully corrected without invasive work, and the practical advice in many cases is to manage the limitation rather than propose an expensive floor lift for a comfort issue rather than a functional failure.
How do I tell the difference between a cold spot and a partial system failure?
Run the system for at least an hour and check the affected area against unaffected areas with a thermal camera or, at minimum, by hand. A cold spot fault typically shows a gradient — cooler but not stone cold, often in a regular pattern (stripes, a band along a wall, a patch matching furniture). A partial failure (a dead loop, a snapped cable section) typically shows a sharply defined area at or near room-ambient temperature with no gradient into the surrounding warm floor, and often corresponds exactly to one loop or one section of cable rather than a diffuse pattern.
Should I trust the original installer's design drawing if one exists?
Yes, always request it before diagnosis. A design or as-built drawing showing loop spacing, screed depth specification, or electric cable layout is the fastest way to distinguish "installed as designed, this is an inherent limitation" from "installed incorrectly, this is a genuine fault." Where no record exists, thermal imaging is the best available substitute.
Regulations & Standards
BS EN 1264 — Water-based surface embedded heating and cooling systems; covers wet UFH design output and installation tolerances relevant to loop spacing decisions
BS 7671:2018+A4:2026, Section 753 — Electric floor and ceiling heating systems; covers cable installation, spacing, and RCD protection requirements for electric UFH
BS 8204-1 — Screeds, bases and in-situ floorings — Part 1: Concrete bases and cementitious levelling screeds to receive floorcoverings — Code of practice; relevant to screed depth and uniformity over embedded pipework
The Building Regulations 2010, Approved Document L — Conservation of fuel and power; relevant to continuity of floor insulation and avoidance of thermal bridging at floor perimeters
The Building Regulations 2010, Part P — Electrical safety; any fault-finding or repair involving the fixed electrical circuit for electric UFH is notifiable work
BEAMA — Underfloor heating installation guidance — industry body guidance on electric UFH installation practice
Polypipe UFH Technical Support — wet UFH loop spacing and screed guidance
JG Underfloor Heating Technical Guide — manifold balancing and commissioning guidance
BPEC UFH Module Guidance — training and assessment guidance for UFH fault-finding
wet underfloor heating — general wet UFH fault-finding, including whole-system failure, air purging and balancing procedure
underfloor heating manifold not calling heat — total non-response fault, manifold never activates
underfloor heating electric fault — electric UFH not heating at all, sensor and element diagnosis
electric underfloor heating pricing guide — pricing for electric UFH installation and repair