Underfloor Heating Design: Heat Loss, Pipe Spacing and Loop Length
A tidy pipe layout does not prove that a floor can heat the room. Check its usable area and heat output first, then connect spacing, circuit length, flow and pressure loss.
Published by MEPFlow · Methods and limitations · Primary references
Use the method while you read
Follow the worked examples, check your inputs and use the related MEPFlow tool to explore your own design.
In this guide: jump to a step
Hydronic underfloor heating design has two separate questions: can the floor deliver the heat the room needs, and can the water circuit carry that heat at an acceptable pressure loss? Answering only one leaves an incomplete design. More pipe cannot overcome every floor-temperature limit, and a large circulator cannot correct inadequate emitter area.
This guide connects those questions through one original worked example. It covers water-based radiant floors, not electric mats, snow melting or radiant cooling. The example uses deliberately assumed thermal and hydraulic data to explain the calculations.
Connect the thermal check to a circuit layout in MEPFlow’s underfloor heating design workflow. Keep the selected floor assembly, room duty and hydraulic assumptions visible for review.
1. Start with room heat loss, not a universal watts-per-square-metre rule
Establish a room-by-room winter heating load using the project design temperatures, envelope construction and outdoor-air assumptions. A whole-building heat-loss total is not enough to allocate floor circuits: a highly glazed corner room and an internal room can need very different outputs per square metre. Use the building heat-loss worked guide to review that starting calculation.
Decide how much of each room's duty the floor must deliver. If another emitter intentionally serves part of the room, document that split rather than assigning the complete load to both systems. Keep the room's required upward heat delivery separate from heat leaving the pipe downward or through distribution routes.
ASHRAE's radiant heating design procedure treats panel area, construction, water temperature and flow as linked design choices. That is the right sequence: establish the duty and usable emitter surface before treating the pipe layout as final.
Define the thermal boundary once
A room heat-loss model may already contain a floor-loss component. Reconcile that model with the heated-floor assembly calculation before adding downward losses to water-side duty; do not count the same loss twice. Heat delivered to another conditioned room must also be allocated consistently.
2. Divide by the active heated area
Active area is the surface assigned to useful floor heating, not automatically the gross room area. Mark fixed exclusions, construction joints, restricted zones and the actual routing constraints. Confirm the treatment of built-in furniture and floor coverings with the selected system designer; do not quietly count an excluded area as a full-output emitter.
Our teaching room has 30 m² of gross floor area, with 6 m² deliberately excluded. Its remaining 24 m² must provide an upward duty of 1,440 W. The required output is therefore 60 W/m². Dividing by the gross area would give 48 W/m² and understate the duty on the floor that actually heats the room by 20%.
Scroll sideways to see both floor circuits.
Required upward output = room duty assigned to floor ÷ active heated area
For the example: 1,440 W ÷ (30 − 6) m² = 60 W/m². This is the required output, not evidence that the chosen floor can supply it.
| Input | Value | Meaning |
|---|---|---|
| Gross room floor | 30 m² | Architectural area |
| Excluded area | 6 m² | No useful output credited |
| Active heated area | 24 m² | Area used in the heat-output check |
| Upward room duty | 1,440 W | Duty allocated to this floor |
| Required upward output | 60 W/m² | 1,440 ÷ 24 |
Take the next step on your own floor plan
Create a MEPFlow account to work through loads, ductwork, domestic water or hydronic layouts in your browser.
Start free trial3. Check the floor assembly before choosing water temperature
Obtain output data for the actual assembly: pipe and spacing, embedding layer or heat-transfer plates, floor finish, insulation, room temperature and water conditions. Do not transfer a bare-slab rating to a timber-and-carpet assembly. Caleffi's heat-emitter guidance illustrates how spacing and floor-finish resistance change the water temperature needed for a given upward output.
Check the relevant comfort and material limits, including any lower limit set by the flooring supplier. A water temperature, a floor-surface temperature and a room-air temperature are different quantities. Nor is a published mean-water temperature necessarily the supply temperature: use arithmetic mean water temperature or logarithmic mean water-to-room temperature difference as specified by the manufacturer, together with its stated temperature drop.
For arithmetic only, assume a hypothetical assembly delivers 60 W/m² upward with 150 mm spacing, water at 40°C supply and 35°C return, and a 20°C room. Assume its surface and material limits have been checked. That gives 24 × 60 = 1,440 W upward, exactly matching the example duty. These are invented teaching inputs, not a claim that any real floor produces that output at those temperatures.
If the reviewed assembly cannot meet the duty within its limits, reduce the room load, revise the assembly or provide supplementary heat. Raising the water temperature without rechecking the floor is not a valid shortcut. Low-temperature performance matters particularly when matching a floor to a heat pump; see the heat-pump capacity guide.
Area sensitivity: what if only 20 m² can be heated?
Keeping the same assumed 60 W/m² output gives only 1,200 W. The room then needs 240 W of additional heat. The required output has risen to 72 W/m², but that requirement does not make the original floor rating increase.
4. Turn pipe spacing into a provisional circuit layout
For a first area-based estimate, active pipe length is approximately active area divided by pipe spacing in metres. This comes from parallel pipe runs covering a strip one spacing wide. It is a planning estimate: actual bends, borders, exclusions and connections must be measured from the routed layout.
At 150 mm spacing, the 24 m² example needs approximately 24 ÷ 0.15 = 160 m of active pipe. Split it provisionally into two equal areas: about 80 m active pipe per circuit. If each circuit has 8 m of combined supply and return tails back to the manifold, each full loop is about 88 m and the two circuits contain about 176 m of pipe in total.
The extra 8 m is per loop, not per manifold and not per tail. A loop is the complete continuous route out of the manifold and back. Check the final pipe length, bore, bend radius, joint policy and pressure loss against the selected system's instructions. There is no single maximum loop length that is valid for every pipe size, flow and floor system.
Spiral and serpentine layouts distribute the water's changing temperature differently. Choose the pattern with the thermal design and room geometry, then review doorway passages, feed congestion and access to the manifold. Two loops that look similar on a plan can still need different balancing settings.
Loop length ≈ active area per loop ÷ spacing + both tails
Example: 12 m² ÷ 0.15 m + 8 m = 88 m per loop. Replace this estimate with the actual routed length before hydraulic selection.
| Spacing | Active pipe for 24 m² | What remains to be checked |
|---|---|---|
| 100 mm | About 240 m | More pipe and a different circuit split |
| 150 mm | About 160 m | Example starting layout |
| 200 mm | About 120 m | Output and surface-temperature uniformity |
5. Calculate water flow from total circuit heat transfer
Water supplies both useful upward heat and any other heat released along its route. For this example only, assume the reconciled assembly calculation adds 160 W of downward heat at the same design state, and assume the tails have negligible heat loss. The total water-side duty is then 1,440 + 160 = 1,600 W. The 160 W is an independent teaching input, not a recommended loss allowance or a fixed percentage.
Use the heat balance Q = mass flow × specific heat × water temperature drop. With water approximated as 1 kg/L and specific heat 4.18 kJ/(kg·K), the total flow is 60 × 1.60 ÷ (4.18 × 5) = 4.59 L/min. Equal 800 W circuit duties give approximately 2.30 L/min per loop. Unequal areas or losses need their own circuit duties rather than an automatic equal split.
The 5 K difference is supply minus return, not water minus room temperature. Halving it to 2.5 K would double the calculated flow for the same duty, so it also changes the hydraulic calculation. It can change the thermal operating point too; do not alter flow independently and assume the same floor output without checking the coupled result.
For an IP-unit cross-check, use the free hydronic flow calculator and the load-to-flow guide. The usual water approximations are not interchangeable with glycol properties. A glycol design needs its actual concentration and operating-temperature data.
Water flow (L/min) ≈ 14.35 × heat transfer (kW) ÷ ΔT (K)
14.35 = 60 ÷ 4.18 using the stated water-density approximation. Here: 14.35 × 1.60 ÷ 5 ≈ 4.59 L/min total, or 2.30 L/min per equal loop.
6. Use the controlling circuit for pump head, not the sum of parallel loops
For each complete path, calculate pipe loss at that path's flow and add its fittings, manifold ports, valves and other devices. Add shared-component losses at the combined flow they actually carry. The circulator must deliver the simultaneous design flow against the controlling complete path; parallel circuit pressure drops are not added together.
To demonstrate the accounting, assume a reviewed pressure-loss rate of 110 Pa/m at 2.30 L/min for the example tube and water condition. An 88 m circuit then loses 9.68 kPa. Assume a further 4.00 kPa across its branch-specific manifold connections and valves, plus 8.00 kPa in all common supply/return components at the total 4.59 L/min. The complete path is 21.68 kPa, approximately 2.21 m of water head.
The 110 Pa/m, 4 kPa and 8 kPa values are placeholders, not pipe or valve performance data. Replace all three with actual calculations or manufacturer selections. Do not add the second identical parallel loop's 13.68 kPa to the first. Uponor's planning information explains manifold balancing against the highest-loss circuit and the separate common pressure requirements.
This simplified duty point is not a selected pump. Confirm the complete piping arrangement, available pump curve, valve authority and operation when zones close. The critical-circuit pump-head guide explains the wider path check, including why closed-loop circulation head is not simply the building height.
| Path element | Flow basis | Pressure loss |
|---|---|---|
| One 88 m pipe loop | 2.30 L/min | 88 × 110 = 9,680 Pa = 9.68 kPa |
| Branch manifold ports and valves | 2.30 L/min | 4.00 kPa assumed |
| Shared supply/return components | 4.59 L/min | 8.00 kPa assumed |
| Controlling complete path | Total pump flow 4.59 L/min | 21.68 kPa ≈ 2.21 m water |
7. Finish with controls, commissioning and a reviewable schedule
A design should explain what happens at part load, not just at the coldest condition. Coordinate room control, water-temperature control, floor-temperature protection and the heat source's minimum-flow and cycling requirements. A thick floor does not respond like a lightweight emitter; abrupt control changes can leave stored heat arriving after the room no longer needs it. Caleffi's key design concepts provides useful background on low-temperature emitters and thermal response.
Before issuing a layout, keep one schedule that ties the room duty to the selected floor assembly and every circuit. Record actual pipe routes and lengths, design flows, pressure losses, manifold assignments and any remaining supplementary heat. Follow the system supplier's pressure-testing, filling, purging, curing and initial heat-up instructions, together with local requirements; a calculation example is not a commissioning procedure.
Use MEPFlow's underfloor heating workflow to explore the relationship between room demand, heated areas, circuits and manifold results. Keep manufacturer data and independent checks alongside the model. Read the methods and engineering scope before relying on a preset; a result for one construction does not establish the performance of another.
- Thermal: room duty, active area, actual assembly, floor finish, water conditions and allowable surface temperature.
- Layout: pipe specification, pattern, spacing, exclusions, actual loop lengths and both tails.
- Hydraulic: circuit duties, flow, actual bore, pressure losses, balancing settings and complete pump path.
- Operation: supplementary heat, zoning sequence, floor limits, heat-source minimum flow and commissioning record.
Frequently asked questions
How do you calculate underfloor heating output per square metre?
Divide the room duty assigned to the floor by its active heated area to find the required upward output. Then compare that requirement with verified output for the actual assembly and operating conditions. Required output and available output are not the same quantity.
What pipe spacing should underfloor heating use?
There is no universal answer. Select spacing with the room duty, floor construction, floor finish, water conditions and surface-temperature limits. The 100, 150 and 200 mm values in this guide demonstrate pipe quantity only, not approved selections.
How long can an underfloor heating loop be?
Use the selected pipe and system manufacturer's limits together with a pressure-loss calculation at the design flow. Count the entire circuit, including both tails to the manifold. A memorized length limit cannot replace the pipe-size and flow checks.
Do I add the pressure drops of all manifold loops?
Not when the loops are parallel. Add their simultaneous flows, then calculate the controlling complete pressure path, including shared components at their combined flow. Balance the lower-loss branches appropriately.
Can underfloor heating be the only heat source in a room?
Only if its verified upward output meets the assigned room duty within surface, material and operating limits. A floor that falls short needs a revised design or a documented supplementary emitter; a larger pump alone does not solve the shortfall.
Is this worked example suitable for glycol or electric floor heating?
No. The flow and head example uses simplified water properties and hydronic circuits. Glycol needs mixture-specific thermal and hydraulic properties. Electric floor heating has different electrical, output and installation requirements.
Continue the engineering workflow
Put this guide into practice with MEPFlow
Move from the calculation method into a connected browser-based design workflow while keeping the assumptions visible for engineering review.
Start free trialPrimary references
Sources and further reading for the methods discussed in this guide.