Puzzling Heat Pump Efficiency Data (Daikin Altherma 3 HHT)

Applications
1

Hello All

I have been running a Daikin Altherma 3 HHT, model EPRA18DV3, since August 2025

Can anyone help explain some rather puzzling readings from my heat pump.

1) Heat Pump ‘COP’
The tiny user interface on the Daikin indoor unit has some information screens. These include data about Electricity Input and Produced Heat, I have been taking daily readings of totals since installation. This lets me calculate Electricity Input per day and Produced Heat per day. I ran the heat pump for a few months at its installed Leaving Water Temperature (LWT) of 55C. I wanted to see how it would perform with the installer setup.

I have taken my daily readings and divided the Produced Heat for each day by the Electricity Input. I am provisionally calling this Heat Pump ‘COP’, although I don’t know enough yet to be sure that’s what it is. The plot attached below shows that with LWT = 55 my Heat Pump ‘COP’ varies between about 3 and well above 5. (Delta T is the difference between internal and external temperature, measured in a rather labour intensive way using two outside Min / Max thermometers and ten inside)

Two or three weeks ago I reduced the LWT to 50. As expected the Heat Pump ‘COP’ is higher. There seems to be a kink in the plot. But it may simply be that I don’t have enough LWT = 50 data points yet

Heat Pump COP.pdf (9.8 KB)

Here is what is puzzling me.

2) Heat Pump System Electricity Consumption

My second plot below shows that with Delta T less than 11.5 C, the heat pump system as a whole uses more energy at LWT = 50 than at LWT = 55. In this plot the electricity is measured from the dedicated consumer unit that was installed to power the heat pump and all the related heating system equipment

Heat Pump_Consumer_Unit_Elec_vs_Delta_T.pdf (12.2 KB)

Reducing LWT has apparently increased ‘Heat Pump COP’ but made the system as a whole less efficient. Can anyone explain what is happening?

I should add a couple of other points.

  1. I believe the heat pump has been over-sized

  2. I don’t have specific DHW data, but I think our usage is fairly constant, apart from the weekly disinfection days.

Any thoughts will be much appreciated.

Chris Gordon-Smith

2

yes,you don’t have enough data, but imo you are overcomplicating the measurement .. if you are running it right,I ndoor temp should be fixed 20 or 21c (or your own target), +/- very little, so just measure outside temp average over the day (or ask openweathermap for history data) and use that for your x axis.

If you are letting the inside temp vary, then this is a bad way to run a,heatpump, and confuses heck out of the numbers. I estimate it take 10 to 20 kWh to raise the temp of 50 or 100 tons of house by 1c, which is a lot more than it takes to meet the steady state heat loss for an hour or two.

50c is awfully hot by the way, never mind 55c. Might need it when it is subzero, but when the temp is up around 5c or more,you could be running at 40 or 45c, and not having the compressor turning off every 10 mins. Weather compensation does that automagically, if you have the option.

1 Like
3

Thanks Dave

I just have one thermostat in one room. So without others I don’t have an average for the whole house. But I could no doubt have managed with fewer than 10!

Based on the measurements I have I’m pretty confident that the inside temperature varies vary little (by less than 0.5 C up or down)

55C may be high, but that is what the installer’s comprehensive design document set it to. The heat pump is capable of running at LWT up to 70C. Starting from the installer’s initial setup I am making small changes one by one to build up a picture of how the system operates. Once I have built up this picture I may well look at weather compensation.

Although I want to get more data, the data that I do have does seem to show that after reducing the LWT to 50C the overall system has become less efficient for Delta T lower than 11.5C. There are 15 magenta LWT = 50 points above the red LWT = 55C line, but only 3 below. It’s hard to explain that as a statistical blip.

Any explanation of what might be causing this would be very welcome.

Chris Gordon-Smith

4

Hi Chris,

Interesting - and as you noted, somewhat puzzling. I can appreciate why you want to understand what is going on before straying further from the installer’s settings.

A couple of observations:

  1. If I understand you correctly, you sourced the Electricity Input data underlying your first graph from the heat pump controller, whereas for your second graph that came from (presumably) some sort of electricity sub-meter installed in the dedicated consumer unit for the heat pump.

    • How do those two sets of Electricity Input readings compare with each other? Are there any clues in the differences between the readings from the two sources?
    • What sort of “related heating system equipment” do you have installed, that is included in the second graph but not the first?

  2. The underlying principle behind Weather Compensation is that the heat loss from a building is directly related to the Outdoor temperature - and should at least be consistent at any given Outdoor temperature (assuming the target Indoor temperature is the same).

    • Your second graph is showing quite a lot of variation in Electricity Input at a given temperature (i.e. quite a lot of ‘scatter’) - notably around 7 degrees of temperature difference.
      • The cluster of magenta data points are roughly double some of the red data points.
    • If this variation in Electricity Input is reflecting a variation in Space Heating Output, your property is accepting significantly different amounts of heat on different days with very similar Outdoor temperatures - and yet your Indoor temperature is remaining stable.
    • Passive solar gain can sometimes account for a reduction in space heating demand, and high winds can account for an increase - compared with the nominal expectation for a given Outdoor temperature.
      • Would you say that you sometimes get a lot of passive solar gain - i.e. do south-facing rooms get significantly warmer than north-facing rooms on cold-but-sunny days?

Two or three weeks is a relatively short time and some installations have been having issues with excessive defrosting due to ‘problematic’ weather conditions over the past few weeks. I wonder if you’ve just been unlucky with the weather since you swapped to LWT = 50.

David

1 Like
5

The installers ‘comprehensive design document’ sets the lwt needed for the worst case outside temp, and the rads you have. Unless you reduce the lwt when it is warmer outside than that worst case temp (which it will be 364 days a year, or more) the compressor’s only option is to turn off. Ideally a heatpump compressor never stops, the room thermostat never quite reaches target temp, and any trvs never close (unless, as mentioned, you get a lot of exra heat from sun, cooker, log burner etc).

Defrost will indeed mess things up, so will trying to heat dhw to more than ‘5c less than the lwt’ which will either take ages, or invoke immersion heaters (cop = 1, but not reported as such). What dhw temp are you Trying to achieve? (and what flow temp being used tilo get there)?

By the way, you might want to view the heat geek video on legionElla, and consider whether you even need that weekly money burning.

6

Hello David (davidMbrooke)
Thanks for your comments. Here are some responses.

1) … for your second graph that came from (presumably) some sort of electricity sub-meter installed in the dedicated consumer unit for the heat pump.

Yes, there is a dedicated consumer unit. It feeds the heat pump and all the other heating system equipment, and nothing else.

2) How do those two sets of Electricity Input readings compare with each other? Are there any clues in the differences between the readings from the two sources?

Here is a series of values from early January. They are rather puzzling

Date Dedicated Sensor (kWh)
Consumer Unit (kWh)
1 Jan 2026 58.4 66.5
2 Jan 2026 48.5 52.1
3 Jan 2026 55.8 61.3
4 Jan 2026 66.9 66.6
5 Jan 2026 70.2 62.0
6 Jan 2026 68.1 68.1
7 Jan 2026 51.7 54.0
8 Jan 2026 47.3 53.9
As you can see, the electricity input value based on the Daikin sensor is sometimes lower than the value based on the dedicated consumer unit readings. But it is also sometimes higher. I thought this would be impossible! Everything is fed from the dedicated consumer unit.

3) What sort of “related heating system equipment” do you have installed, that is included in the second graph but not the first?

Ah. By “related heating system equipment” I meant any part of the whole heating system other than the heat pump itself. I expect this includes the immersion heater and the water pump(s). But perhaps there are other items that are part of the system as a whole but are not being monitored by the sensor in the Daikin indoor unit.

We don’t use any other heaters (the heat pump is heating the house well!).

4) Your second graph is showing quite a lot of variation in Electricity Input at a given temperature (i.e. quite a lot of ‘scatter’) - notably around 7 degrees of temperature difference.

Yes, there is a lot of variation. As you say it is notable that a lot of it is round Delta T (temperature difference) = 7C. That’s a good spot, thanks. The variability seems to be less at lower values of Delta T, and also at higher values. I had thought that the scatter was just the kind of variability that could be expected when making such measurements, but perhaps there is more to it than that. Any thoughts on what it might be? Is there some system behaviour that could be kicking in at about Delta T = 7C or just before?

5) The cluster of magenta data points are roughly double some of the red data points.

Yes. It’s very odd. I have not filtered disinfection days out of the data, so that might increase the scatter. But it would affect both the red and the magenta points

6) Would you say that you sometimes get a lot of passive solar gain - i.e. do south-facing rooms get significantly warmer than north-facing rooms on cold-but-sunny days?

We notice this effect in the front bedroom. But mainly in Summer. I doubt whether it is a major factor in Winter. We do notice that that the kitchen, which is the lowest room, is colder than the rest of the house, which is running as a single zone. The thermostat is in the kitchen.

7) Two or three weeks is a relatively short time and some installations have been having issues with excessive defrosting due to ‘problematic’ weather conditions over the past few weeks. I wonder if you’ve just been unlucky with the weather since you swapped to LWT = 50.

That’s a possibility. But I haven’t noticed anything particularly unusual about recent weather conditions here. I’ll keep collecting data at LWT = 50 for a while before reducing to 45C

Any other thoughts you (or anyone else) might have would would be very welcome. In particular:

A) How it it possible that the Daikin Electricity Input sensor sometimes indicates more electricity use than the dedicated consumer unit, and sometimes less?

B) Why is my heating system using more electricity at LWT = 50 (for the same Delta T) than at LWT = 55 C.?That seems to fly in the face of the theory.

C) Why is the variability in electricity use higher at around Delta T = 7, and lower at Delta T =3 and Delta T =20

D) Why is the red line on my second plot almost linear above Delta T = 11 C. That is not what appears on a typical COP diagram focused on the heat pump itself.

Chris Gordon-Smith

7

I think the Daikin heatpumps, most them anyway, do not actually measure immersion heater power, they just measure how long they ask for it, and assume 3kW. I recall some folks seeing Daikin reported numbers of ridiculous usage, when their immersion was broken turned off, or the local thermostat/overheat was off. I should believe what the electricity meter says, not the heat pump.

As I said earlier, if you are asking for hotter water than the heat pump can easily achieve, in whatever time at whatever flow temp is allotted, it will try to use the immersion heater. Some of them will do it anyway, if they think it avoids defrost cycles.

8

Hello Dave (Storey)

Thanks for your posts.

1) Defrost will indeed mess things up, so will trying to heat dhw to more than ‘5c less than the lwt’ which will either take ages, or invoke immersion heaters (cop = 1, but not reported as such).

Very interesting. Can you say a bit more about “not reported as such”?

2) By the way, you might want to view the heat geek video on legionElla, and consider whether you even need that weekly money burning.

I’ll take a look.

3) I think the Daikin heatpumps, most them anyway, do not actually measure immersion heater power, they just measure how long they ask for it, and assume 3kW. I recall some folks seeing Daikin reported numbers of ridiculous usage, when their immersion was broken turned off, or the local thermostat/overheat was off. I should believe what the electricity meter says, not the heat pump.

That’s a bit worrying. Does that mean that my plots of “COP” are meaningless? They are based on what heat pump says about how much electricity it has used and how much heat it says it produced. I’m still puzzled about why the heat pump sometimes says it used more electricity than was recorded by the dedicated consumer unit. Our immersion heater is not broken / switched off.

Heat Pump COP.pdf (9.8 KB)

4) As I said earlier, if you are asking for hotter water than the heat pump can easily achieve, in whatever time at whatever flow temp is allotted, it will try to use the immersion heater. Some of them will do it anyway, if they think it avoids defrost cycles.

Wow. The hot water is set to 45C. Is there a document that describes all this logic? How can I know how to avoid these various pitfalls?

Chris Gordon-Smith

9

  1. the heatpump reports COP based on the electricity used by itself and the heat it thinks it produced (based on DT between flow and return, and flow rate). The former should be fairly accurate, the latter less so. However the former does not usually include any electricity used by immersion heater, and sometimes not by water pumps and controls, if they are not actually part of the heatpump. So the COP is probably ‘as good as you can get for free’, but will not include immersion use.

  2. Recommended. The number of victims of Legionella in UK from domestic water systems is currently a round number. Zero. You shower head is a more dangerous vector than your DHW tank. Unless you are absent / not using water for a week or two.

  3. See 1). The Daikin is reporting its COP best as it can, but not immersion etc use.

  4. 45c should be fine, but how long it takes to get there will depend on the water flow temp (and the tank coil size). The system may have a higher flow temp, or at least a different one, when doing DHW than when room heating, given room heating might only need a 35c flow (which aint EVER going to get the DHW to 45c, obviously).

The document to read is the installation / instruction manuals for your particular heatpump (and control system). If you give the model number someone can point you at an online copy, and maybe even tell you under what conditions it might try to use immersion heater instead of heatpump power (as I said, immersion COP = 1 .. 1kW in, 1kW out)

1 Like
10

Thanks Dave

11

I’ve been trying to understand why, in mild weather, lowering the leaving water temperature (LWT) sometimes increases daily electricity use, even though COP should improve.

After a lot of discussion and looking at my own data, it may be that it comes down to minimum heat-pump output and control behaviour, not thermodynamics. The idea is that when a heat pump goes into this behaviour, it switches off and on several times before reaching target room temperature because it can’t operate below it’s minimum power. As the LWT approaches target the heat pump tries to reduce power, but has to switch off to avoid going below its minimum power. But the room has not reached its target temperature, so the heat pump soon has to switch on again.

The problem gets worse as LWT reduces because it gets to its minimum power sooner.

Here is Chat GPT’s step-by-step explanation of the mechanism. Comments very welcome. I am keen to get to the bottom of this.

Typical cycling sequence in mild weather

1) Thermostat switches ON
The controlling room temperature (e.g. kitchen) drops slightly below setpoint, so a heat demand is issued.

2) Circulation starts
Primary pump runs and water circulates through the heat pump (and buffer / low-loss header, if present).

3) Compressor starts and ramps up
The compressor starts and heats water towards the target LWT.

4) Radiators extract heat
Initially, radiator output is high, return water temperature (RWT) is low, and
ΔT = LWT − RWT is healthy.
This is stable, efficient operation.

5) Indoor temperature approaches setpoint
The thermostat room warms towards its setpoint. Other rooms may already be warmer.

6) Heat demand falls
Radiators extract less heat, RWT rises, and ΔT across the heat pump shrinks.

7) Controller reduces compressor speed
The heat pump tries to match the falling load by slowing the compressor.

8) Minimum compressor output is reached
The compressor hits its minimum stable speed.
However, ΔT continues to shrink and LWT is near target, meaning output is still greater than load.

9) Internal control decides it can’t run stably
After a short persistence period with:

  • very small ΔT,

  • minimum compressor speed,

  • and little heat being absorbed,

the controller determines that stable operation is no longer possible.

10) Compressor switches OFF (cycling begins)
The compressor stops even though the thermostat is still calling for heat.
This is an internal protection / anti-hunting decision, not the thermostat switching off.

11) Water temperatures drift down
LWT and RWT fall, ΔT increases again.

12) Compressor restarts
Because:

  • the thermostat is still ON, and

  • internal conditions now allow stable operation.

Steps 3–12 may repeat several times.

13) Thermostat eventually switches OFF
Once the controlling room finally reaches setpoint, the heat demand ends and the system rests.

Key takeaway

The important point is that the compressor often switches off before the thermostat is satisfied, because it has reached its minimum stable output and the water circuit can no longer absorb heat fast enough.

This explains why, in mild weather, lowering LWT or using aggressive weather compensation can sometimes increase cycling and daily electricity use, despite improving instantaneous COP.

12

Hi Chris,

The implication is that one (or both) of the electricity metering devices is inaccurate - because as you say, the heat pump “Sensor” value should never be smaller than the “Dedicated Consumer unit” value (which is measuring the supply to the heat pump plus some other devices).

To be confident of the readings they need to be coming from a meter which is certified to an accuracy standard - typically the Measuring Instruments Directive regulations.

What sort of metering do you have on the dedicated consumer unit?

There are always limitations with a heat pump’s variable compressor speed control. It’s therefore almost universal to have some degree of ‘cycling’ when only a modest amount of heating is required. (It’s not that long since heat pumps only had fixed-speed compressors and had no option but to cycle, and most models cope with cycling fairly well.)

For the EPRA18DAV3, Daikin’s technical details summary shows:

  • Minimum heating output: 4.4 kW
  • Nominal heating output: 9.0 kW
  • Maximum heating output: 12.2 kW

Is that right? Thinking out loud: At a lower LWT, the radiators will lose less heat to the rooms (at a given room temperature) so ΔT = LWT - RWT will indeed be smaller and less heat will be transferred into the house, which increases the tendency to hit the lower limit of heat output. Running at a higher LWT will result in more heat transfer to the house (because LWT - Room Temperature is greater), so it will be easier to stay above the minimum heat output and avoid cycling.

It all depends how much heat the house actually needs though - and how good the radiators are at transferring that heat at a lower LWT. If you’re dumping 4.4 kW to the house (at LWT = 55) but it only wants 3 kW it will just ‘cycle’ on the room thermostat instead.

A lower LWT is certainly always better for (instantaneous) CoP.

So perhaps your Daikin unit is cycling more at the lower LWT (and cycling is reducing efficiency) - but why is that reduced efficiency not showing in the heat pump’s own CoP readings for LWT = 50 - and does that mean you can’t trust the Daikin’s reporting of electricity consumption?

David

1 Like
13

Hello David (davidMbrooke)

Thanks for your message

To be confident of the readings they need to be coming from a meter which is certified to an accuracy standard - typically the Measuring Instruments Directive regulations.

I don’t know what sort of metering I have on the Daikin. BUT…

The brochure for my heat pump identifies the following as an accessory:

K.ELECMETV: Electric meter for domestic RHI - Single-phase (Metering for performance compliant)
MID Class A electric meter to measure the electricity consumption of the Daikin Altherma heat pump

Since I didn’t pay for this accessory I’m assuming I didn’t get it. So I’m treating the current electricity readings from the heat pump with a large degree of scepticism!

What sort of metering do you have on the dedicated consumer unit?

It’s on the left of the consumer unit, labelled: FuseBox KWH1M45.

So perhaps your Daikin unit is cycling more at the lower LWT (and cycling is reducing efficiency) -

I think that is probably what is happening. I know that the heat pump is oversized. Much to my annoyance, the installer put in the model above the one that had been agreed. And without telling me! I didn’t find out about it until later. I already thought the unit was oversized, so the oversizing may be substantial.

For information, here is my latest plot:

Heat Pump_Consumer_Unit_Elec_vs_Delta_T.pdf (12.4 KB)

Chris Gordon-Smith

14

So based on the FuseBox website that meter is MID Approved and can be trusted, whereas I agree with your conclusion that the Daikin unit’s built-in metering won’t be as reliable.

[By the way, that exact same meter seems to get re-badged and sold under several different brand names; I’ve got a couple of Rayleigh D-175 models which appear identical to yours, apart from the ‘manufacturer’s’ labelling. The Rayleigh meters are available with an M-Bus interface which means I can read them automatically.]

That certainly is annoying and will lead to a bit more cycling in not-too-cold weather. There are some benefits of having a larger unit though - such as less-frequent defrosting and quicker heating of the hot water cylinder.

David

15

Thanks David

16

There is a definite minimum workable flow temp for your house/system, but 50c should not be it. You can work it out (but needs a spreadsheet), measure all the rads, lookup their DT50 kW output (that’s the one that assume room at 20c, average rad temp 70c, so 80c out 60c back) (Stelrad site is good), and then derate to sensible heatpump temps (50C flow is ROUGHLY DT 30, which iirc says the rads put out 51% of their DT50 output.

My 21kW of rads can get down to about 28c, at which point they put out ~2.5kW, which is the minimum for my heatpump.. Go any lower than 28c and it WILL cycle (but not very quickly).

However .. that assume your rads are all balanced and getting similar flow and dumping proper amount of energy/power. If there is a ‘short circuit’ through on nearby radiator, automatic bypass (yuk!) 4 port buffer or low loss header (double yuk) the amount of power you can throw into the house is way less. So hot water arrives back at the heatpump, which responds by quitting. You really need temp monitor on flow and return (especially return!) to catch this sort of misbehavior.

Your ‘why is it worse at 50 than 55’ might be easier to debug if you turned the immersion off during the experiments, but Daikin heatpump may throw an error under those conditions. That is the likeliest cause of ‘loads of electricity being used in inexplicable manner’. Even at 50 or 55c slow I doubt your heatpump is using more than 1.5kW, whereas the immersion will use (at least) twice that if/while it is on, wildly skewing the numbers.

1 Like
17

Thanks, that’s helpful.

One extra data point that might be relevant: before the heat pump installation, on gas heating I was seeing about 100 kWh/day total energy at ΔT ≈ 14 °C. I don’t recall the exact split, but assuming around 90% was gas space heating, that suggests an average delivered space-heating load of only ~3 kW at that ΔT (Inside - Outside), allowing roughly for boiler efficiency.

That may put a lot of mild-weather operation below the minimum steady output of an 18 kW unit, which would naturally push it into duty-cycling even with most TRVs open. (Do you have a rough idea of what the minimum output would be.)

Given that, I’m wondering whether the combination of minimum output plus the LLH (mine’s in the cellar) could be shortening run times at lower LWTs by driving the return temperature up faster, which might explain why 50 °C sometimes looks worse than 55 °C in daily kWh, even though the house stays warm.

18

Using these numbers, 77 kWh (factoring DHW and efficiency) over 24 hours at Δ10°K* gives a rough heat loss coefficient of around 320 W/K for the property. Scale this up to a design temperature of -3°C (Δ20°) comes to 6.4 kW heat loss for the property. Even with defrosts, a 18 kW unit is way oversized.

[*active heating is often not necessary when outdoor temperature is above ~17° as there’s sufficient heat coming from appliances, humans, sunshine, etc. Can see this on the first chart you posted. So calculate Δ from that temperature rather than actual indoor temperature. See heating degree days for more info on the concept.]

You can get more accurate figures by plotting daily heat against average Δ and fit a line to it. Compare both gas and heat pump data, and you should get similar results (adjusting for any improvements to the fabric).

4.4 kW, as mentioned in post #12.

As Dave notes above, the stable flow temperature is dictated by how much heat the radiators can emit. If your radiators are too small or too few or restricted in some way or there’s a bypass, then they won’t be able to shed the heat coming from the heat pump, and flow temperature rises. This could explain why consumption and COP is worse at 50° than at 55° - the system is cycling more, and being less efficient.

It would help to have a chart of flow and return temperatures throughout the day, as this will help us see the behaviour of the heat pump. The flow rate is also useful to know, as this can be used to calculate actual heating power.

Daikins are often accompanied by an automatic bypass valve to protect the system if the primary flow is restricted. This is unnecessary if you have a low loss header, so look to see if you have one, and make sure it’s not passing hot water.

There will be a 3-way valve that switches between central heating and domestic hot water. If this is broken, this can allow some of the hot water to bypass directly to the return, raising the flow temperature. This is precisely what is happening on the topic Mitsu Zubadan Flow Temperature Issue - check that the DHW side isn’t hot when not active.

The low loss header can impact efficiency a little, but not enough to cause the issues you’re seeing. Ideally, the secondary pump is matching the speed of the primary, and the temperatures on both sides of the LLH are equal. There is a process to balance these, but it’s quite involved. The topic linked above covers this too.

3 Likes
19

Hello Tim

Using these numbers, 77 kWh (factoring DHW and efficiency) over 24 hours at Δ10°K* gives a rough heat loss coefficient of around 320 W/K for the property. Scale this up to a design temperature of -3°C (Δ20°) comes to 6.4 kW heat loss for the property. Even with defrosts, a 18 kW unit is way oversized.

The numbers look about right to me. 6.4 kW is far less than the heat loss calculated by the installer - 10.63 kW with outside temp = -2.08 C and inside temp. = 20C (I’m not sure where the .08 came from)

I have attached the plot showing my pre heat pump data, as well at the post heat pump data. The green line shows the situation after we had completed insulation and double glazing, but before heat pump installation. It should be clear from the plot that heat loss would never be 10.63 kW. That would take the daily heat loss to 255 kW. I showed the plot to the designer. He went to look at it, but the answer that came back was that the estimate was accurate.

Energy per Day vs Delta T.pdf (23.7 KB)
(Note that the plot now has two points at LWT = 45C.)

I did eventually persuade the designer to reduce the estimate by asking him (for a second time) to look at the results of a blower test we’d had done a few years earlier (before the house was double glazed). That left the design with a 16kW model (EPRA16) rather than the originally proposed 18kW model). I decided to go ahead with 16kW, even though I thought it was probably oversized. There had already been many delays and I needed to get the heat pump installed. Finding an installer who would do the job in a way that left me confident it was compliant with the installation instructions had been difficult. I didn’t want to go back to square one.

Some weeks after the installation I found that the 18kW model had been installed anyway! I didn’t spot it at first because the model number on the unit was difficult to access, and I remember thinking it was unlikely that the wrong unit would be installed. But it was.

I think at the heart of this problem is that heat loss is estimated by inspecting the building material etc., rather than actually measuring it. At best estimating factors like air change rate and U values is uncertain, and there may be a tendency to err on the safe side to avoid a situation where a house is not heated properly on cold days.

(Do you have a rough idea of what the minimum output would be.)
4.4 kW, as mentioned in post #12.

Oops. Looks like I dropped the ball there. It was late at night!

It would help to have a chart of flow and return temperatures throughout the day, as this will help us see the behaviour of the heat pump. The flow rate is also useful to know, as this can be used to calculate actual heating power.

Absolutely! I would also like to be able to monitor electricity use at a much more granular level so that I can see the cycling. I’m hoping to get my hands on an emonPi

Daikins are often accompanied by an automatic bypass valve to protect the system if the primary flow is restricted. This is unnecessary if you have a low loss header, so look to see if you have one, and make sure it’s not passing hot water.

There is a ‘Buffer Tank / Low Loss Header’ in the cellar. A cylindrical unit two feet or so high. Two pipes in / out either side. I’ll see if I can check the point about passing hot water.

The low loss header can impact efficiency a little, but not enough to cause the issues you’re seeing. Ideally, the secondary pump is matching the speed of the primary, and the temperatures on both sides of the LLH are equal. There is a process to balance these, but it’s quite involved. The topic linked above covers this too.

Thanks

20

Following on from Dave’s post (#16) do you know the total rated output of your radiators?

The Radiator Output Formula is helpful for this sort of thing:

Heat_output = Rated_Heat_Output × (Delta_T ÷ Rated_Delta_T) ^ 1.3

50° flow through the LLH could have mean temperature at the radiators of 45°. Assuming a room temperature of 20°, that gives a delta of 25°. This means the radiators would put out only 41% of their rated output. Or, to put it another way, the radiators need to be rated at 2.5 times the power from the heat pump. So, at design temperature, 6.5 kW × 2.5 = 16.25 kW of rated output to keep up with the heat pump at that temperature. If any radiators are restricted, that will bring down the effective emitter surface.

If you want reduce mean flow temperature of 40° to get better performance, at the minimum power level, you’ll be needing 4.4 kW × 4.8 = 21 kW of radiator output.

If you already have this much rated output, then that suggests the high flow temperature is caused by something else, like the ones Dave and I mentioned before.

Here’s a topic about Correct setting for Automatic Bypass Valves? (ABV)