Monitoring LWT in Homeassistant from Daikin EDLA04E2V3(4kw) pump … I read in many forums and experts in here that this pump “doesn’t like to operate at less than 30C LWT” [ setpoints ].
I queried this with Daikin and mentioned the installer was Octopus who provided a commissioning document (excerpt attached). I also mentioned alleged “rapid cycling of the pump” as stated by forum members.
Daikin said this was “fine” and “no issues”.
Why would the manufacturer state this pump can operate “fine” at LWT less than 30C, the installer sets the lowest temp in the weather curve at 25C - that is X1 in the table below and is the lowest LWT the pump should try to set … and experts on the forum contradict this.
Now please be constructive and objective, I am only asking the question because I am intrigued as to the conflict of opinions from “experts” from both domains. I am not questioning the wisdom of either - I just am interested in learning who is right and why the others may have or been providing - the wrong / unsuitable advice/opinions. I am not trying to evoke a war or flameout - I just genuinely want to understand rationales.
By the way, an “expert” (compared to moi) on here @chrisg runs an automation that I also adopted and he observes no cycling at these low temps - according to his ESP monitoring feedback , correct me if I’m wrong chris?
Here is the octopus commissioning document for my installation , note X1 value 25C
The problem with many pronouncements from manufacturers, installers, and knowledgeable amateurs is that they usually gloss over important “ifs and buts”.
For example, trying to heat a house to 23C with water at 25C will be much more problematic than heating a house to 17C with water at 25C. Heat transfer to the room is just lower so the likelihood that the heat produced by the HP cannot be transferred to your inside air increases, leading to stop-start behaviour. More or larger radiators can help you “get away with” using cooler water, as can using fancoil heaters.
Also, whether or not there is a buffer tank comes into play. These will always have some mixing between HP flow water and cool space heating return water (maybe a LOT of mixing). That could push up the lowest workable temperature by a couple of degrees.
And what the overall space heating water temperature drop is, influenced by radiator balancing and having a suitably sized and speed-controlled pump.
So: I’m sure it IS possible to design a system which can run at 25C but the issues come with how close to that each real system is. Retrofit brings compromises!
I do get some cycling of the heat pump below 30c LWT but it’s not rapid. Yesterday daytime, sunny spells and mild, it’s heating for about 15 minutes, sampling periodically for 45 minutes. It’d be better if I could have the 8kw unit though. Still waiting on Octopus reviewing my data.
Would it be better to run at min LWT 30c and have the heat pump heating for, say, 30 minutes then off completely for 2 hours? Maybe I should try that some day but a) the Madoka won’t do it in RT mode - the house will simply overheat and b) I want to reduce reliance on the cloud API as much as possible which means reducing the number of times on/off is requested from Home Assistant.
Really interesting to compare yesterday’s data - we’ve just had our 9kW replaced by the 8kW after agreeing to pay for the HP with Octopus covering the installation costs. Not sure if this helps, but here’s our system’s data during the same time period. About the same outdoor temp, room temp a little higher than yours but the system ran continually from 11:30 through 17:00. 2.48kWh consumed vs 2.36kWh yet 14kWh heat vs 6.4kWh.
Very few installs will cope well with super low flow temps under 30C without overshooting fairly quickly and cycling imo. Most installs will be at 30C when it’s fairly mild, so the heat pump will have high power out/COP from mild outdoor air, the rooms won’t be losing much heat, and there is a small dT between flow temp and room temp so emitters can’t emit much heat…
The daikins don’t deal particularly well with the cycling however, as on startup they tend to have quite a high power draw, and take a while to drop back to lower power, so they quickly overshoot again.
That’s not my experience, Ben. Take a look at my graph above. You can see that it never has high input power draw when restarting the compressor at low flow temps. After 45 minutes of sampling, the compressor starts up at its lowest power draw.
The only time I see a high input power draw on startup is if the requested flow temperature is much higher than the current flow temp - I’m seeing this more now the heat pump is off overnight due to the milder weather:
Interesting - your 9kW seems to be behaving differently certainly to what ours used to do. We’d always get a ramp up to higher output on startup. Perhaps there are differences in hardware generations? Ours was installed in August 2022 if that’s a comparison point.
That’s not my experience, either.
Hence my OP .. lots of voices aligned with your sentiments, but not much documentary measured evidence so far as I have seen, to support that view.
My target LWT is 25-28C, which obviously is too low for the rads to handle, so the HP gets too hot and cycles - great, that’s what I want
Your target LWT is around 40C! And it has some 2C steps in it - is that programmed by you - at 1.17-, 4.00+, and some +/- 19.00-21.00
My HP cuts off at 14C outside temp, as that’s how I’ve set it
It’s a Friday, so my legionella cycle starts at 2am, and includes a 3kW immersion session I also account for by adding 3kW to generated heat for that period
Something happens at 5am on yours - is that a defrost?? At 10C ambient?!
You’ve obviously turned the HP off 6.00-19.00
In-house mean temps: Tim 22.3C, John 21C
Heat output: Tim 49.6kWh, John 41.6kWh
Space heating COP: Tim 3.75, John 3.97
So yes, I can see that there’s quite a chunky startup effort at the beginning of each “cycle” compared to mine - is that what you refer to? But I’d argue your LWT is probably too high for that ambient, so I’m not surprised the HP felt (past tense!) it needed to try hard to reach that. Perhaps your heat loss is significantly less than mine, but then over the period you put more heat into the house, admittedly for a higher room temp. And yes, I’ve considered discussing the 8kW with Octopus, but it’s currently working.
Looking at your 2 Mar screenshot, which is at OAT average 4C, -0.6-9.5C, a very different situation where one wouldn’t normally see cycling, it looks like it’s responding to defrosts, which is what I’d expect in those circumstances, rather than cycling. You turned it off at 15.00? Here’s mine for that date (apologies for the data hole - my wifi obviously dropped out!):
Room temp mean 22.6; 21.3->24.5(!) - might have been sunny, and I use LWT only, no 'stat
At 20.00 it kicks in hard, but again that’s post-defrost
I’m not sure I’ve added much to the conversation here, just some observations and some questions about how you were/are using it. For the record, I don’t usually see a big kick when it is genuinely cycling, rather than recovering from defrosts. And none of this is intended as a criticism of anyone or anything, we’re all learning. Oh, and mine was installed in May '23.
Interesting comparison points.
LWT driven by weather curve setup + Madoka stat. Steps are whatever the system decides to apply as we don’t manage that.
With the 9kW we had a “boost” of +1 set to the target room temp for whenever our EV kicked in an out of hours charge on Octopus Intelligent so the 1900-2100 might have been linked to that. Not necessary for the 8kW as it’s happy to run steadily at much lower power consumption than anything the 9kW could achieve.
With the 9kW’s minimum power input stuck at 930-950w there was limited practical benefit to chasing lowest possible LWT temps for us as the unit would ramp up cycles & defrost cycles if LWT dropped much below 37ºC so most of the time it would be sat at 37-40ºC LWT rather than achieving lower temperatures.
5am - coming off a DHW cycle & boosting room temps by a degree.
We never turn the system off; house hit target temperature at 0600 and the heat loss is low enough that the HP didn’t kick back in until 1900.
Your unit’s going to be at least a year newer than our 9kW was. Install was August 2022, but no idea how long the hardware sat in a warehouse/distributor before shipping to us. Not impossible there are differences between them in behaviour or perhaps software.
