Different-ΔT primary–secondary heat pump systems

Heat pumps are going into both new buildings and existing ones at a really fast pace right now. This is especially true for older properties that were originally designed around boiler systems. Those older systems usually ran with a higher temperature difference than the 5 °C that most heat pumps prefer. In this article we will look at why we sometimes run different temperature differences on the primary and secondary sides. We will also see what that means in practice.

The thermodynamics of decoupling

The flow rate you need depends directly on the temperature difference you are using. To deliver a certain amount of heat to the building we use this basic relationship:

$Q = \dot{m} \cdot c_p \cdot \Delta T$

When we run one temperature difference on the heat pump side and a different one on the building side, the two flow rates naturally end up different as well. Here is a straightforward example:

• Primary $\Delta T$ = 5 °C

• Secondary $\Delta T$ = 10 °C

• Result: $\dot{m}_p = 2\dot{m}_s$

To keep the two circuits hydraulically independent, so what happens on one side does not directly affect the other, we normally fit a Low Loss Header (LLH). We size this header so the pressure drop across it is almost zero. That usually means the header pipe itself is larger than the primary and secondary pipes connected to it.

Here is how it works in practice. The primary pump pushes twice the volume of water. All of it leaves the heat pump at the target temperature of 55 °C. The secondary pump only draws off half that flow and sends it out to the building at the same 55 °C. The remaining hot water simply drops down through the low loss header and blends with the 45 °C return water coming back from the building. After mixing, the water heading back to the heat pump sits at exactly 50 °C. Both sides get exactly the flow they need. The two pumps never fight each other.

Flow, COP & condenser performance

You might assume that running a wider temperature difference on the primary side would help the heat pump’s efficiency by keeping the average water temperature lower. In reality it usually reduces the coefficient of performance (COP). The reason lies in what happens inside the brazed plate condenser.

If you try to run the heat pump itself at a 10 °C temperature difference, you get a very low mass flow rate. That slow flow turns laminar inside the narrow channels of the condenser. Heat transfer suffers. The compressor has to raise the refrigerant condensing temperature to push enough energy into the water. All that extra work lowers the overall COP.

With a low loss header in place, the primary pump can maintain a high flow rate. The water inside the condenser channels stays turbulent. The heat transfer coefficient improves. The compressor can settle at a lower condensing temperature. Efficiency goes up as a result.

Here is a clear comparison:

Metric	Direct 10 °C ΔT	Decoupled 5 °C ΔT
Mass flow	Low	High
Channel flow	Laminar	Turbulent
U-value	Low	High
Condensing temp	High	Low
Compressor lift	High	Low
COP	Reduced	Maximised

You could try to force turbulence by making the channels narrower. The pressure drop would rise sharply and you would need far more pump power. Those tiny channels would also foul quickly with normal system debris. Keeping generous channel sizes and maintaining good flow rates remains the practical engineering choice.

Flow protection for the compressor

Air source heat pumps are quite sensitive to sudden drops in water flow. In a directly coupled system, when zone valves or thermostatic radiator valves start closing as rooms reach temperature, the resistance on the secondary side rises quickly. Flow across the condenser can fall enough to cause sharp spikes in refrigerant pressure. This often triggers a high-pressure lockout.

A low loss header gives the source loop built-in protection. If the secondary flow drops because valves are closing, the primary pump simply keeps circulating its full volume. The excess water bypasses straight through the header. The heat pump sees the return temperature rising and smoothly reduces its output. It modulates down safely without ever hitting a flow fault.

Buffer vessel strategies

Air source heat pumps need enough thermal mass in the system to perform reverse-cycle defrosts cleanly. This melts frost off the outdoor coil without tripping low-temperature faults or blowing cold air into the building. Where you place that mass makes a big difference.

The series return buffer

The preferred arrangement is to place a buffer vessel in series on the primary return. It sits between the low loss header and the heat pump. Hot water from the heat pump flows straight through the low loss header and out to the building. The buffer sits on the return side. Its thermal lag does not delay the supply temperature reaching the emitters.

During initial warm-up the cold buffer actually encourages the heat pump to run at full output. This helps protect startup COP. We generally avoid fitting a mechanical bypass around this buffer. It can cause premature modulation and introduces another potential failure point.

The 4-pipe parallel buffer

Some designs use a 4-pipe buffer vessel that replaces the low loss header. It acts as both hydraulic separator and thermal store. The main drawback is internal volume. Even with top connections, the kinetic energy of the incoming water stirs the tank. The 55 °C primary flow mixes with cooler layers inside. The water leaving to the secondary circuit often drops a couple of degrees. For example it might leave at 53 °C instead of 55 °C.

A low loss header has almost no internal volume. It transfers the full 55 °C flow straight across without any mixing or temperature loss.

Buffer vessel sizing

During a defrost the heat pump reverses and effectively runs as a chiller. It pulls heat out of the system water to melt ice on the outdoor coil. To prevent the water temperature from dropping too far and causing faults, you need enough water volume to act as a thermal battery. Industry guidance usually suggests 12 to 15 litres per kW of heat output. This figure comes from the latent heat needed to melt the ice that typically forms. A defrost normally lasts 5 to 7 minutes at full capacity.

We can calculate the required volume with this relationship:

$V = \frac{Q_{\rm defrost} \times t}{\Delta T_{\rm allow} \times c_p \times \rho}$

For a 10 kW heat pump running a 5-minute (300 s) defrost and limiting the system temperature drop to 6 °C, the numbers work out like this. The energy extracted is roughly 3,000 kJ. Dividing by the allowable temperature drop and the properties of water gives about 120 litres. That matches the common 12 L/kW rule of thumb.

Having that volume turns what would be a sharp temperature plunge into a gentle, recoverable dip. It is also standard practice to subtract the water already present in high-volume secondary pipework and active radiators from the total buffer volume you need to add.

Sizing guidelines for common heat pump ratings:

• 6 kW: 72 to 90 litres

• 10 kW: 120 to 150 litres

• 16 kW: 192 to 240 litres

These figures sit inside the 12 to 15 L/kW band. In colder, frost-prone climates where defrost cycles happen more often, it makes sense to lean toward the upper end of the range.