Performance of 9kW Daikin

I’ve got a 9kW Daikin system, installed by Octopus. With the help of members here, I’m monitoring it using ESPAltherma and a CT connected to a MyEnergi Eddi. The data is logged to Home Assistant and published to emoncms and heatpumpmonitor:

The house is 1960s detached, two storey and with reasonable insulation. The design flow temperature is 50C at -2C with a calculated heat loss by Octopus of approximately 7.5kW (according to the Radiator Schedule). We have radiators (no underfloor heating), a 250L Daikin hot water tank and a 20L volumiser. The heat pump has a pipe run of approximately 11m before it enters the house. The house has been extended and probably has a funny internal pipe run.

I’m running the following WD curve:
45C at -2
33C at 16C

Room thermostat control with a -1C thermostat offset to match the ambient temperature better.
I’ve configured 1C of modulation and 4C of overshoot.

Typical temperature targets are 19C from 6am to 10pm, 17C set-back overnight.
Hot water is schedule-only at 1pm, to heat to 45C. I use the Eddi to boost higher.

I’ve balanced the radiators so they all heat up satisfactorily with a decent temperature drop, although they aren’t all managing a delta of 5C. I’ve done this mostly to try and ensure there’s no unexpected short-return paths to the heat pump that might explain the high return temps it only seems to achieve (see below).

Some observations from my data:

  1. The difference between the R1T (before back-up heater) and R2T (after BUH) measurements is about 1.9C
  2. My system struggles to drive the LWT below about 33C. I’ve tried a lower weather curve and the system tries to get to a lower return temp, but can’t achieve it at minimum power and flow rate, and the flow-return dT doesn’t hit 5C and so the resulting COP is poor.
  3. It performs best (COP near 4.0) when it is ‘walking’ the flow temperature above the target temperature. It did this yesterday morning, but today the LWT has been sat mostly on or near to the target temp.
  4. The system seems to allow heating to continue running until about 20.5C which is pretty wild for a 19C target temp.

Any thoughts or suggestions most welcome :slight_smile:

Things I might try, after I’ve gathered a bit more data:
a) 4C delta-T, to see if it encourages the system to use a bit more power and higher flow rates
b) a slightly higher weather curve


Hi Jonathan,

I apologise in advance as I am going to be the bearer of doom and gloom, I have the same heat pump in a smaller house than yours and I am finding it a nightmare to be honest.

I have to ask, how big are your radiators?

Your radiator schedule would be good to see.

This heat pump has a minimum electrical input of about 900w, it doesn’t matter how low you go on the flow temperature, that is as low as it goes.

That is because it is really a 16kW heat pump.

Mine doesn’t like going below about 35c flow temperature and it often uses no less electricity running at 38c.

I have found that running as radiators works best for me, my COP is worse and my heat output lower if I try to run at dT 5c on fan coils.

I get to a dT of about 6.3c on radiators at a flow of 36c

i cannot get to the radiator dT of 8c as the flow rate is too high, it only goes as low as 10lpm.

We all see a difference between the tow temperature sensors R1T and R2T, there is a loss through the back up heater.

I am MID monitored and R2T is what comes into my house even though R1T is higher on my heat pump as well.

I would try radiators with the fixed dT of 8c, you probably won’t get 8c but it may be better.

I gave up on the Madoka fttb

I just run a fixed flow of 36c and either turn the heat pump off or open windows to control room temperature.

I find the initial heating phase to be drawn out and poor in efficiency terms caused by at least 30 minutes of high flow rate.

My one is difficult on occasions after DHW or periods of being off, sort of like a failed start, never really gets going and shuts down after an hour or so and has another go.

I find as it gets colder my heat pump cannot maintain 36c and the flow temperature slowly decays to around 34c, I have no idea why as I have an overshoot of 4c set. When it is warmer, no problem and the overshoot works as the house temperature rises.

I also find my heat output drops at a constant flow temperature when it gets colder outside, the opposite of what i would expect. I would have thought the radiator performance would improve as the heat loss increased, I see the opposite.

If I leave it running it always shuts down and restarts 6 hours later, you can set you watch by it. It must be a thing for these heat pumps. It is annoying as it sucks all the heat that I have paid for out of the house and starts again, for no apparent reason.

The 8kW version does not do this, it will run continuously.

I have tried different dT and none work for me, radiators has always been best in terms of efficiency.

My honest opinion is that this is a dog of a heat pump and it is all because it is really a 16kW heat pump with a lowest output commensurate with a home that needs a 16kw heat pump.

As you may be able to tell, I hate it!

1 Like

[Split into a new topic]

I have a very similar type of property, with a different brand of heat pump. [System #1]

Looking at the intantaneous COP, your Daikin seems to struggle to get up to the performance we would expect to see in mild weather (10°+)

We’ve seen this happening on other Daikins too, so I don’t know if it’s a common fault or bad design.

Other than that, your curve setting looks to be about right. The temps will continue to creep up as the radiators aren’t managing to emit the heat being produced on the minimum output. You’ll want your radiator valves as open as possible, and balancing the system may have limited emitter output.


Hi Tim,

I think I have worked out why this is.

I have looked at mine and a couple of 11kw Daikins and they all do the same, they run at a high pump speed initially. It seems to be timed. Mine is about 29 minutes, it does it without fail, and then it settles to my lowest of 10lpm.

This means for me I can never get much of a dT established for at least the first 30 minutes.

My radiators cannot handle it even though they are massive for the house, just still not enough for this heat pump.

Even though the flow temperature is established the pump keeps flowing at a high rate.

It is only when it slows that the efficiency arrives.

It takes 45 minutes and that is why I find it difficult to run this Daikin heat pump for short periods in my house.

1 Like

This is the bit I cannot really decipher or understand in the manual, overshoot vs modulation…

I have an overshoot of 1 (never changed) and have changed a modulation from 10 to 2 and now 4. I try to change one thing and see a difference

Hi Neil,

The overshoot is by how much the flow temperature can exceed the requested flow temperature.

It prevents short cycling by trying to maintain the dT between flow and return, when the return nears the flow it has to stop.

The modulation is related to the room temperature and the Madoka.

It’s been a while since I had my Madoka controlling anything so I cant remember exactly what it does, although I am not sure I really ever understood it either.

My 9kW Daikin is too big for my house, or at least it’s minimum output is too big for my house and I just couldn’t get on with the Madoka.

It didn’t really limit the room temperature and seemed to exert too much influence over the heat pump, but that is probably because my house and the heat pump are not well suited.

I found the Madoka made the heat pump work hard to get to the room temperature, then exceeded it and then shut down all at the worst possible efficiency.

There are no getting away from the facts with this heat pump.

If you want it’s best efficiency then you need to accept heat of 4,500w or more at a COP of 5.0 or better.

If you want or need less heat then you have to accept a lower COP.

If you want 3,000w of heat then the best COP you will ever get is 3.30.

If you want to turn it off periodically then it may well get worse in terms of COP.

I don’t know who decides such things but as far as I am concerned this heat pump is not fit for purpose and should not be sold.

For anybody at all interested in efficiency this heat pump has a range of operation of between 4.5kW and 6.3kW when the defrosting starts.

I cannot think of any house it is really suitable for.

It heats my house OK but everything I look at from other manufacturers seems to do a better job.

I know what they were thinking when they did this but I can’t believe they really thought this was a good idea.

I have yet to encounter an owner of a Daikin 9kW Altherma 3 who doesn’t have the same issues.

Performance during this DHW cycle looks reasonable though, about what I would expect. This one too. Both have about the right amount of time and kWh to heat 200 litres by 20°, which suggests the metering isn’t far off.

Could get better COP by targeting a lower temperature, say 45°, but this depends on household demand. Oh wait, you’re already doing this.

As a starting point it would be worth reviewing your actual heat loss (as told by octopus) and either check this yourself, using heatpunk, manual htc calculations (Estimating the heat capacity of my house | Protons for Breakfast) or getting an independent heat loss survey via a heat geek or similar.

This will give you a good starting point to understand what you actually need for your house. If it is less than 7kw then you may want to reach out to octopus to request the unit be swapped out. My 09 is being swapped out in 2 weeks for the EDLA08EV3 (my heat loss is 5.4kw and not the 7.3 they thought).

If you are able to it may be worth listing your radiator schedule and the updated heat loss you have calculated as from that you can work out the heat output at various temperatures (40, 45 and 50°c) and see if there is a way to tame the beast in your house.

One other point of concern would be the distance from the heat pump to the diverter valve and tank. If this is over 10 meters, as the EDLA09DA3V3 can easily output over 11kw and to reduce the resistance the pipework should be 35mm and not the usual 28mm that Octopus tend to use.

If you want a hand with any of the above or to discuss further let me know.


Thanks for looking at the graphs @Timbones.
Yes, the DHW target temperature is 45C. Is it normal to see flow temps considerably higher than the target temperature? I guess if you want to heat the DHW in a reasonable time you need to?

Here’s the data from Home Assistant sensors for the cycle on 29th December, which heated the water about 10C:

And here’s the cycle from 31st December, which heated the water about 13C:

I’ve included the Eddi temperature sensor as well.

@matt-drummer @KnightPhoenix here’s the radiator schedule:

The total room heat loss comes out to 6456W.

I’ve previously estimated the heat loss on Heat Punk with basic assumptions about wall / window / floor U-values and get around 6600W, so that’s pretty close to Octopus.

PS: Happy New Year everyone!


Have the installers fitted a short circuit between flow and return?

(auto bypass valve)

That is a bodge. It has no place in a great pump setup and will cause high returned if it opens due to insufficient fire through the radiator circuit.

The “inappropriate pump speed during warm up and performance only being delivered after the pump speed drops” is an interesting observation.

Could you force its hand so to speak with external dT control of pump speed I wonder? sorry heat pump, you’re too dumb, I’m taking a run signal from your compressor then coding my own circulation pump speed that you’ll have to work to"

1 Like

There is one thing to note about the radiator schedule from Octopus.

I found this out on my second heat loss survey.

It only includes the heat loss of heated areas so if you have any areas without a heat source, radiators in this case, then the heat loss from that area has to be included to get to the total they used.

In Jonathan’s case, they would not have fitted the 9kW heat pump unless the heat loss was over 7kW.

They told me that if they had calculated my heat loss at 7kW or less I would have got the 7kW at the time or the 8kW now.

My heat loss calculated by Octopus was 7.4kW in the end, 5.1kW for the house and 2.3kW for my 13m2 conservatory.

That is why I am struggling with it.

I had 9kW of radiators on my schedule (I didn’t actually because the surveyor put all my K1s as P-Plus).

I have since found out that I needed at least double the radiator sizes I had, I still don’t have enough really to get the best out of this heat pump.

You will never be able to run at a low flow temperature at anything like a decent COP with this heat pump and those radiators.

Hi Marko,

I have been thinking about this also.

The pump speed appears to be timed, it is not intelligent at all on the start up.

It is the source of most of my problem now with the heat pump.

If I could get it `come to life’ more quickly it would be easier to run in shorter bursts.

It’s impossible to fit big enough radiators in a house my size to get a dT of 5c when the pump is running at 20lpm.

The pump is capable of running at a lower speed.

My concern would be that other parts of what is really a 16kW heat pump require this minimum flow rate of 10lpm.

I don’t know.

Whilst the abv is undesirable I am not sure I would call it a bodge, they had to buy it and fit it, it would have been much easier and cheaper not to have done so.

My radiators all had trvs on them, what would have happened it they had all shut?

What they should have done was removed all the trvs as the last thing you need with this heat pump is less emitters working!

Hey Jonathan,

Thanks for sharing, just to echo what Matt has said, the 9kw pump is not likely to work great in your property (or any from what we have been seeing) as it needs to loose 4kw minimum to get a reasonable COP and prevent restarting (ideally 4.5kw or above). While it will do that at -3 and maybe as high as 5°c this is not where we spend most of the year.

It may be worth discussing this with Octopus especially if you find it is cycling.

Also the variance between radiators (48% to 134% coverage) is a little crazy. If it were me I would be re designing the system using either heatpunk or excel with a 45°c or 40°c flow temperature. This will mean changing radiators out however, this will allow for the existing heat pump to loose more heat from the system (and run a bit more efficiently) and if you go down the replacement route for the replacement to run much more efficiently as lower flow temperatures mean cheaper bills. If you do not want to replace everything as a minimum I would be upgrading the towel rail to a 1800 x 600 or something like the ‘Milano bow D’ as that will put out more heat and still act as a normal towel rail.

I get not everyone wants to rip out their radiators again and keep tinkering with their heating system but as you are here I am guessing you have a bit of a thing for efficiency and statistics (like a good few of us) so wanted to present the option.

Again if you need a hand contacting octopus or with the radiator stuff just shout (I have made a few waves with the tentacled beast so may know who to chat with).


I have Daikin coming to check my heat pump and installation on 12 January.

I hope they find something wrong with it because I cannot believe this is how it is supposed to work.

I can just about get it to perform to an average efficiency looking at all the other systems but it has taken a lot of effort to do that.

I have tried everything I can think of and everything that has been suggested by more knowledgeable people than me here.

At times it works really well but it isn’t really a sustainable way to operate a heating system.

I wanted to find out what the heat pump could do but I will have to change how I operate it at some point.

Anything I do now will drop the efficiency and it will be a heat pump that performs well below average.

I will be looking for answers from Daikin, whether I will get them or not I don’t know.

Everybody that I have encountered with this heat pump has the same problem and even the `bigger’ ones seem to suffer some of the issues that just aren’t prevalent with heat pumps from other manufacturers.

I really believe that there isn’t a home that they deem this heat pump suitable for where an alternative from any other manufacturer wouldn’t perform much better.

I really don’t see how they can explain that a heat pump designed to be fitted to modest, average size houses with heat losses of around 7.5kW has the same compressor, the same circulation pump and the same minimum operating level as a heat pump they make that will heat a house three or four times the size of ours.

I feel really strongly that those of us with this heat pump have been victims of misselling of a product that is not fit for purpose.

I expect a bit of a battle with them over this during the coming weeks and months.

1 Like

please keep us updated, one thing i would like to know is how to flatten the initial peak of energy consumption with a high flow rate, why not low and slow (less power and lower flow rate) like a good steak on the BBQ.

1 Like

Hi Neil,

I will keep everybody updated however boring it is! :slight_smile:

I would like to know this too!

One way to do it is manually walk the flow temperature up, @Stephen_Crown does this sometimes I think and I do at times.

But who has time for this all day every day?

The initial phase of heating is awful and for me, by the time it is done I need it to turn off.

I think I am going to try and run with the Madoka again soon, with my windows closed and see what the efficiency really is.

I will even close the windows in my conservatory!

Then I can compare the two extremes.

I have run like this before but not with the new radiators.

There is one thing I have noticed looking at the various Daikin models.

As we all know, the 9 to 16 all have the same compressor and circulation pump and the same minimum electrical input.

I expected the same from the 4 to 8 models as they have the same components (I think) but it does appear from what I can see that the 6kW goes as low as 300w input and the 8kW seems to only go as low as about 425w

Maybe the 9kw is a mistake, maybe it is possible for it to run at a lower level, maybe it is just the software that limits it?