CIBSE's New Domestic Heat Pump Design and Installation Guide: What Installers Need to Know

Published October 2026 · Industry news · 9 min read

CIBSE has published the Domestic Heat Pump Design and Installation Guide (2026), developed jointly by its Domestic Building Services Panel and the Heat Pump Association UK. It has been more than two years in the making and is written for the people doing the work: installation companies, engineers moving across from gas, and anyone who wants a single, manufacturer-independent reference for how a domestic heat pump system should be put together.

It is a companion to the CIBSE Domestic Heating Design Guide 2026 (version 2026-02, the 11th edition), not a replacement for it. The Design Guide holds the engineering method (heat loss, emitter sizing, pipe sizing) and is heat-generator-neutral. The new guide wraps that method in a set of heat pump-specific workflows running from the first survey visit through design, installation, commissioning and handover, and at each step it tells you which section of the Design Guide to use. CIBSE describes it as "presented at a lower technical level" than the Design Guide, and that is a fair description.

Here is what matters in it.

It is deliberately narrow, and says so

The guide covers one kind of system and is upfront about the boundaries. Its method assumes:

• • a monobloc air-to-water heat pump of up to 45 kW (split, hybrid, ground-source and air-to-air are not covered in this edition)
• • an owner-occupied, single-family house; flats, maisonettes and rented property are outside scope
• • a maximum space-heating flow temperature of 55°C and a 5°C design ΔT
• • one flow temperature for every emitter in the house, radiators and underfloor alike
• • a single circulator controlled by the heat pump, normally the one built into it or supplied with it, serving the whole system
• • no thermal store and no low-loss header
• • weather compensation, hot water priority and a constant heating strategy with a setback of no more than 3°C.

If a job falls outside any of those, the guide says so and points you to the full Design Guide, the manufacturer or a specialist. That is a strength. A lot of the guide's value is in telling you clearly when the simplified method stops applying.

It is worth being precise about what "outside scope" means here, because the guide is. On low-loss headers the instruction is: "For systems designed according to this guide: do not specify a low-loss header" (Note 740). On thermal storage: it "should not be specified in systems designed following this guide" (Note 873). Hybrids likewise (Note 811). None of that is CIBSE saying these components are wrong in themselves. The guide lists where they may be appropriate: a low-loss header where there are multiple circulators or circuits at different temperatures; thermal storage for time-of-use tariffs, solar PV or intermittent heating; a hybrid where the electrical supply or budget rules out a full heat pump. Those systems need a different design approach, and the guide sends you to the Design Guide, the manufacturer or a specialist for it.

But it does state a preference. For most standard systems, it says, thermal storage is unnecessary, and adding it "can add cost, complexity and space requirements without improving performance" (Note 873). Avoiding a low-loss header "can reduce installation complexity and system costs" (Note 740). Hybrids are "not recommended for designs following this guide unless necessary" (Note 811). If the integrated circulator can't manage the index circuit, the guide wants you to bring the resistance down (upsize a pipe run, swap a high-resistance fitting, Note 730) or ask the manufacturer for an alternative unit before you reach for hydraulic separation (Note 872). And existing underfloor circuits with mixing valves and their own pumps should have those removed, with the heat pump supplying the compensated flow temperature directly (Note 441).

A volumiser (a two-pipe buffer) to meet the manufacturer's minimum water volume is part of the method, and the guide is specific that it goes on the heating return only, never the common return (Note 810).

Heat loss comes first, and the guide wants two numbers from it

The heat-loss workflow (section 200) is the foundation. It doesn't contain a heat-loss method of its own; it hands you to section 2 of the Design Guide and its worksheets (A1, A2 and A3, free at cibse.org/dbsp-worksheets). What it does do is set out what has to be settled before you calculate anything:

• • Air permeability, ideally measured with a blower-door or pulse test. Default values are a fallback for retrofits where a test isn't possible (Note 210). Our guide to converting a q50 result into air changes covers what happens to that number next.
• • Outdoor design temperature, from the property's climate zone and a chosen cold-day percentile (99.6th, 99th, 98th or 95th), corrected for sites above 50 m. The guide notes that MCS or manufacturer rules may restrict which percentile you can use (Note 211).
• • Indoor design temperature: for homes built after 2006, 21°C for heated rooms and 22°C for bathrooms. Older homes use the Design Guide's room-by-room table (Note 214).
• • Measured areas and identified construction for every wall, floor, roof, window and door, with U-values from standard tables or calculated where needed (Notes 212–213). See finding the right U-value for an existing wall, floor or roof.

Two outputs come out the other end, and both go on the record. The design heat loss at the outdoor design temperature sets the peak the heat pump must deliver. The typical heat loss at the zone's reference (average) temperature is used to assess the minimum output the heat pump needs to be able to modulate down to (Notes 230–231). Whether a unit actually avoids short cycling also depends on its modulation range, the controls and the system water volume, but the typical heat loss is the number that starts that check, and it is easy to skip.

The guide accepts other methods and software where they produce "a fully compliant heat loss calculation in accordance with BS EN 12831-1" (Note 201). Over- and under-estimation get equal weight: too high and you pay for a bigger unit that short-cycles; too low and the house is cold on the days that matter.

The reused-radiator method is the most practical part

Section 400 is new ground. The Design Guide doesn't cover existing emitters, so this section stands alone, and it gives you a repeatable way to decide whether a radiator stays or goes.

Identify it. If you can find the make and model, use the catalogue output. If not, Note 421 lets you estimate: confirm it is a standard steel panel (type 11, 21, 22 or 33), measure width × height, and read an approximate ΔT50 catalogue output from Table 400-2 by frontal area and type. The guide's own worked example is a 1,200 × 700 mm type 22, which it treats as 0.84 m² and looks up at the next area band above. The estimate route only applies to a standard panel with its original finish and nothing shelved or boxed over it. Designer, column, vertical and flat-fronted radiators need manufacturer data.

Convert it. Table 400-4 turns catalogue output into actual output at the heat pump's flow temperature, for a 21°C room, a 5°C flow/return difference, bottom-opposite-end connections, original finish and no obstructions. The numbers are sobering. As the guide puts it, at 55°C "output may reduce to approximately 50% of the catalogue value", and it falls away faster below that: by 40°C a radiator is giving well under a quarter of its rating, and at 30°C only a few percent of it. The physics behind that curve, with a worked table, is in our guide to flow and return temperatures.

Compare it with the room's design heat loss. Equal or above: keep it, and fit a TRV. Below: replace it, add a second radiator sized only for the shortfall (Note 425), or reduce the room's heat loss. Pushing above 55°C or a 5°C ΔT to make an existing radiator work is outside the guide's method and, in its words, "may reduce heat pump efficiency" (Note 412).

Two related points from the survey section. Radiators that have been on an open-vented system and show even light corrosion are listed as unsuitable for reuse (Note 112). And where you have had to estimate a radiator rather than identify it, the customer should be told the system may struggle to hold temperature on the coldest days, and that should be written into the design record (Note 421).

Flow temperature is a whole-house decision

The guide's workflow starts every design at 30°C and only raises the flow temperature, in 5°C steps and for the whole system, when a room can't be made to work at the current setting. Because every emitter runs at the same temperature, one awkward room (a box bedroom with a small radiator and no wall for a bigger one) sets the flow temperature for the whole house.

That is exactly the trade-off the guide wants you to see. The options it offers are the real ones: a bigger or additional radiator in that room, fabric work on that room, or accepting a higher flow temperature and the efficiency hit that comes with it.

Fabric: not a prerequisite, but do it before you size the emitters

The guide's position on insulation is balanced. Its introduction states that high levels of insulation "are not a prerequisite for retrofitting a heat pump into an existing home" and that properly designed systems "can successfully heat the majority of UK housing stock" without extensive or uneconomic fabric upgrades.

It still asks surveyors to identify and discuss the cheap, high-impact measures (loft insulation, draught-proofing, sealing redundant chimneys) because they can mean a smaller heat pump, radiators that can stay, lower running costs and, in some cases, avoiding a three-phase supply (Note 113). And it is clear on sequencing: any fabric or airtightness improvement the client agrees to "should be completed before finalising and installing heat emitters" (Note 411). Survey, agree the fabric work, then do the heat loss and emitter sizing against the improved building.

The system criteria document

The thing most likely to change how you work day to day is the four-page system criteria document, carried over from the Design Guide and free to download. It records the design heat loss, air permeability, design temperatures, flow temperature and ΔT, every emitter's required and actual output, the index circuit, the expansion vessel and water treatment, and a plain-English "critical design parameters" page the customer signs.

Before installation begins, the guide says the document shall be discussed with the client, fully understood and agreed, and signed by both the client and the designer, "including confirmation of their competency" (Note 990). The commissioner initials it when checking the installed system against the design, and it is handed over at the end with the manuals and commissioning sheets. In England and Wales, CIBSE ties this explicitly to the Building Safety Act, and it notes that even when one person surveys, designs, installs and commissions a job, "each role carries distinct responsibilities".

How it relates to MCS

It isn't an MCS document. It references MCS 020(a) for noise and, at handover, MIS 3005-D and MIS 3005-I, and it notes that MCS may limit your design-temperature percentile, but the design method is CIBSE's own, built on the Design Guide. If you work to MCS you still need to satisfy MIS 3005-D; what this guide gives you is a clearer route to the numbers that go into it. The two calculation methods are compared in CIBSE vs MCS — choosing a calculation method.

Where Heatworx fits in

At Heatworx, our focus is the first half of this process: the survey, the room-by-room heat-loss calculation and the existing-radiator assessment.

Heatworx supports room-by-room heat-loss calculations using the CIBSE Domestic Heating Design Guide 2026 or the MCS methodology, and assesses what your existing radiators will deliver at the flow temperature you are designing to. That is the same comparison the new guide's section 400 asks you to make for every room.

The new guide reinforces why we've focused on getting the survey, building fabric, room-by-room heat loss and existing radiator assessment right. They're the starting point for a properly designed heat pump system.

Reference: CIBSE, Domestic Heat Pump Design and Installation Guide, 1st edition, October 2026 (ISBN 978-1-912034-61-1). Developed jointly by the CIBSE Domestic Building Services Panel and the Heat Pump Association UK. Written to complement the Domestic Heating Design Guide version 2026-02 (11th edition). The worksheets and the system criteria document are available at cibse.org/dbsp-worksheets. Note numbers above are the guide's own.

Related guides