The Flow-Temperature Test: What Really Decides If Your Old House Can Take a Heat Pump

The easy part is the box on the wall

When homeowners picture converting an old house to a heat pump, they picture the machine: a humming cabinet against the back wall, dropped in where the gas boiler used to be. In practice that swap is the least interesting decision in the whole project. A boiler and a heat pump both push the same hot water around the same pipes and the same radiators. What separates a warm, cheap-to-run installation from a cold, expensive one is a single number most people never hear named until it is too late: the flow temperature, the temperature of the water leaving the unit and entering the emitters.

A gas boiler runs that water hot, commonly 70 to 80C, and nobody minds, because burning gas at 75C costs almost the same as burning it at 45C. A heat pump is the opposite. It does not make heat, it moves heat, and the harder it has to lift the temperature the less of it moves per unit of electricity. The efficiency figure, the coefficient of performance, is highest at around 35C and falls steadily as the water gets hotter. Above roughly 55C the penalty turns brutal: installers reckon on losing about 2 to 2.5 percent of efficiency for every additional degree. Run an old system at boiler temperatures and a heat pump will still keep the house warm, while quietly costing as much to run as the gas it replaced. The whole point of the machine is to run cool.

Why radiators lie about their output

Here is the catch that traps most retrofits. The output printed on a radiator, 1.5kW or 2kW, is measured at a laboratory condition called delta T 50: a mean water temperature of 75C in a 20C room. That is a boiler-era benchmark. Ask the same radiator to work at heat-pump temperatures and its output collapses, because heat transfer depends on the gap between the metal and the air.

The relationship is not linear, and the maths is unforgiving. Output scales with the correction factor, delta T divided by 50, raised to the power of about 1.3. At delta T 30, a mean water temperature around 45C, a radiator gives only about half its rated figure. At delta T 20 it manages roughly a third. So a 2kW radiator that comfortably heated a room on a boiler may deliver barely 1kW once the heat pump is doing its efficient work. The radiator has not changed. The rules it is playing by have.

Reading the house, one room at a time

This is why a credible heat-pump retrofit does not begin with a quote for a unit. It begins with a room-by-room heat-loss survey. In the UK the MCS standard, MIS 3005-D, makes it mandatory: a calculation to BS EN 12831-1:2017 for every room before an installation can be certified. The survey establishes how much heat each room loses on a cold design day, and only then can anyone choose a heat pump size, a design flow temperature and, crucially, whether each existing emitter is big enough to meet that room's loss at the low temperature the system will actually run.

The output of this exercise is a map of mismatches. Some rooms pass untouched: older homes are often full of generously sized radiators, sometimes oversized by a previous over-cautious boiler fitter, and an oversized radiator is a gift to a heat pump. Other rooms, usually the coldest, or those with a small radiator squeezed under a window, fail, and need attention.

Three honest ways to close the gap

Where a room falls short, there are three moves, in rough order of disruption. The first is simply to fit a bigger or deeper radiator: a double- or triple-panel unit in the same footprint can more than double the surface area, and often the pipework already there will serve it. The second, for rooms where a larger radiator will not fit, is a fan-assisted emitter or a clip-on booster fan, which forces air across the fins and can lift effective output by 20 to 40 percent at low temperatures. The third, and the most transformative, is underfloor heating: a whole floor is an enormous, low-temperature emitter, perfectly matched to a heat pump running at 35C, and worth considering wherever floors are already coming up.

Fabric first, or emitters forever

There is a quieter route that changes the entire calculation: lower the heat loss instead of raising the emitter output. Every kilowatt of loss you insulate away is a kilowatt you no longer need a giant radiator to replace. This is the fabric-first logic behind deep-retrofit standards such as EnerPHit, the Passivhaus Institute's benchmark for existing buildings, which targets around 25 kWh per square metre a year for space heating, against 15 for a new-build Passivhaus, and delivers, in monitored projects, heating-demand reductions of 85 to 90 percent. At that point the flow temperature can drop into its most efficient band and the original radiators may suddenly be more than adequate. Fabric-first is not cheap, UK deep-retrofit figures for 2026 run to roughly 1,800 to 2,600 pounds per square metre, but it attacks the problem at its source rather than compensating for it forever.

A different kind of renovation

The lesson is that a heat-pump retrofit is not a plant swap, it is a heating-system redesign disguised as one. The visible unit matters far less than the invisible number it runs at, and that number is set by the fabric of the house and the size of its emitters, things a survey reveals and a spreadsheet decides long before anyone lifts a spanner. Ask any installer the flow temperature they are designing to, and demand a room-by-room calculation to back it up. A house that can run at 45C or below is a house that will be warm and cheap on a heat pump. A house quoted at 55C and above is a warning that the real retrofit, the radiators, the insulation, the detailing, has been skipped.