Improve efficiency: Turn the heat pump off

Hi Chris, welcome aboard!

My EcoDan also runs 15/20 minutes cycles in mild weather, and I’ve not figured out how to prevent it either. I think it’s because the unit is slightly oversized for the property, and so can’t modulate its power low enough to keep a steady run without going over the target flow temps.

Increasing the flow temp kinda helps, but is a bit counter intuative. I’m flip-flopping between a minimum of 25 and 30 at the moment, but neither seems to work very well. I find that the COP is still pretty reasonable though as it’s warm outside, so I might not worry about it.

Increasing the volume of your heating system might help, by keeping all TRVs fully open and bleeding the radiators. This increases the heat output of the system, with larger delta between flow and return, allowing the HP to run for longer before reaching target temps.

Lowering the room thermostat (if you have one) will reduce the time the HP runs unnecessarily, but that’s a whole different optimisation that affects running costs vs. comfort.

Hope that helps…

1 Like

Yes I do this do this too for my UFH (boiler/TS based).

I’ve just installed an Ecowitt Weather Station which has a solar sensor on it, so (at some point) I’ll add that into the mix. I’m already using it to switch lamps on/off if it suddenly gets a bit too dull or at dawn/dusk.


I’ve since revised this further, having realised that defining my own thermostat in software means I can have more sophistcated control that a usual programmable timer provides. Now I have the target temperature increasing smoothly throughout the morning with the heat pump turning on as required.

(The dip at noon was when HP was busy heating water; at 5pm there’s a small boost)

The thermostat also feeds into weather compensation, with additional load compensation.
Still tweaking and experimenting, waiting to see if it’ll work when it’s proper cold out.

1 Like

Hi Tim,

Thanks for the reply.

I already have all the TRVs fully open but I’ll try bleeding them, thank you for the advice.

You can see in the graph below of my own system the COP drops offs noticeably once the cycling starts above 10°C outside. So think it is worth increasing the flow temp to stop this happening, the house will heat up quicker so it doesn’t need to be on as long.

I’ve noticed manually turn the thermostat down to prevent the heating coming on till the house is below 17.5°C it stops the cycling, but we’d prefer to keep the house warmer than this. Is there a way of increasing the hysteresis?
I assume this makes the delta T greater and stops the cycling. I’m only getting a delta T of 3°C between the flow and return, I’ve read it should be 5 to 7°C.
I’d seen something about reducing the pump speed to increase the DT, does the heat pump do this automatically or is there any way of altering it?

1 Like

Have you just written a script, or using something else?

The engineers who installed mine did reduce the pump speed using some setting deep in the FTC menu; I don’t know which one and wouldn’t want to guess. They were particularly concerned about keeping the flow rate above a certain value, otherwise the HP would shut down. So, may be best to talk to your installer about it. The “Installation Manual” may give some clues, but is very scant on detail.

I get dT between 3 and 5, yet still see the same cycling that you do, so I’m not sure that increasing the dT will necessarily fix the specific problem of cycling (though might improve COP).

Yes, I have a script that decides how warm the house should be, and turns the heat pump on/off accordingly, via MELCloud.

1 Like

@borpin although a few of us have written scripts, Chris doesn’t want to. This is part of the reason for open-sourcing our scripts and results, so the manufacturers can build the adaptations into their devices for regular users. I suspect the UK is a bit awkward because we have so much of the time when we’re around this 10 degrees phase which is tripping up the heat pumps control algorithms and causing cycling.

BTW @cre at least your cycling isn’t as bad as this:

(part of this bigger post)

Yes appreciate that, just couldn’t remember what Tim had done or if he’d moved to a more universal platform for the coding.

I won’t hold my breath :laughing:

I’ve only had my Arotherm a few weeks, but I think some of the “cycling” issue is just the floor/lowest output of many units.

For example, the datasheets for my 5kW Vaillant says a minimum heat output of 2.5kW.
And with it being 13C outside, you’re gonna get a COP of 5, so the unit it just running as low as it can.

In my case, with a heatloss of 5.5kW at -3, 2.5kW heat output is more than I need right now.
But there’s nothing I can do about that.

Vaillant do a 3.5kW Arotherm model. But when you look at the spec of that, it’s still got a floor of 2.5kW heat output. So that wouldn’t help.

So the cycling I’m seeing (and it’s not really cycling) is just the heat pump outputting its minimum. That being too much for the house right now. Heat pump turns off and tries again a bit later on.

An interesting observation looking at my Vaillant, @glyn.hudson Samsung and @TrystanLea Mitsi units, all 5kW is that the Samsumg seems to have a lower floor than the Vaillant and Mitsi. Which appears useful at this time of year.

But I think reality is, once we get into single figure outdoor temps, we’ll likely see the “cycling” go away and each run should start to have longer run times as the heat output starts to match the heat required within the house.


Well, there sort of is: when the house gets too cold, aim to run it at 2.5kW heat output until the house exceeds your too-warm value or a temperature from which it shouldn’t get too cold before electricity becomes cheaper or COP improves significantly.

By the way, does your Vaillant have a good way for it to be controlled by free software/DIY Home Automation? An API without connecting to the internet?

How isn’t that really cycling? And, sadly, for the Ecodans, “a bit later on” seems to be “about 5 minutes later” which is fairly silly and results in turning off again soon, really lowering COP.

Because I’m contrary, I’m slowly unwriting my script. I’ve split it into a MELCloud-MQTT bridge and a Domoticz event script. When I get time, I’ll tidy and publish those. If I get time while I’m home alone (so there’s no-one to inflict cold on if I get it wrong), I’ll try replacing my Domoticz script with widely-available plugins like the PID Controller, or the Smart Variable Thermostat which has some learning/adaptation feature.

The same MQTT bridge should work with Home Assistant (Domoticz and HA share an autodiscovery protocol) so if there’s a clever heating controller for HA, such as one which controls the flow temperature to do price compensation or carbon emission compensation as well as weather compensation, then I’m open to trying that.


MELCloud-MQTT bridge: GitHub - mjray/ Python script to translate MQTT command to MELCloud and MELCloud status to MQTT topics

The Domoticz event script switching the heat pump on and off is in a post at: Where do the json_attributes go? - Domoticz


The Vaillant looks woefully short on this i’m afraid.
Even their internet “app” is total trash.

There’s a Home Assistant integration, but warns of “depends which app you’re using”.

The best vaillant interface seems to be via ebus, but you need to buy a reader for that, which are in short supply.

Both @glyn.hudson and I have put our names down for one of these, but it says Q1 next year.


I sopke to the techs at the manufacturer for my new boiler and tried to convince them that it would be a good thing to sell a local access device to their boiler rather than the £250 ‘smart’ thermostat that talked to their cloud, but they wouldn’t go for it :frowning:

1 Like

What were your convincing arguments? Asking them to do more work and lose any future cloud subscription fees seems like a tough sell to me. I’m not sure how we can show that the extra sales from people wanting local control would outweigh the lost subscriptions.

I think I understand why the manufacturers want cloud, for commercial and diagnostic reasons, as well as probably firmware rollouts, even if I think it sucks and the drawbacks should be made clear to purchasers.

There aren’t many and they are not convincing (obviously). Main one is that there is a body of consumers that would be attracted to their boilers with this facility - would certainly have made a difference to me. Being able to get Real Time data from the boiler would be massive. Minimal cost to them - they have the design for the access - all it needs is a UART output to be useful. Yes other interfaces would be good but increases complexity and cost. Goodwill.

It will also still work when they decide that their cloud offering is too expensive or they get bought out. I doubt their system does firmware updates.

I never suggested it was a convincing argument.

I suggested before that perhaps OEM should do an eBUS reader :thinking:

Whilst I admit that I’m biased because my job is building, running and selling cloud software, I do have an example where I believe the cloud is useful when controlling devices.

I’ve built this complicated web of sensors and scripts which make our system work well. For example, my installer noted I’d used 8 units of electricity whereas similar customers had used 12.

However, when I had COVID all I was able to do was to flick the pre-existing switch in my code to enable COVID mode which permits heating more often because the windows are open. My head was way too fuzzy to work anything else out.

Taking this to it’s extreme, when I die my family have got no hope of keeping all this stuff running. They will rely on fairly dumb local control or some fancy-schmanzy cloud-optimizer.

Although we are all quite understandably sceptical of the cloud (they have our data, they go offline, they are motivated by their own needs) I can see why the mostly dis-interested user base would use it.

Now, without advertising my employer, I will say that my day job is trying to so out the “they have our data” problem and we are making some progress.

I think part of the issue is that these ‘cloud’ solutions don’t actually do very much, that local control can’t. I also think that most of them are extremely ‘simple’ and provide little benefit. I had (have as the backup parallel system) a HeatMiser system. I’ll swear it didn’t ‘learn’ anything - it is just sales hype that can never be disproved!

Yes but that isn’t limited to the heating!


I’ve been considering this, it’s not something we really want to do, I’m surprised there’s not already a commercial available adaptor eBus, has been around since the early 2000’s