Two months of learning with Octopus Daikin

Hardware Heatpump
1

We’ve had our Altherma 3 M now for two months and I’ve learned an awful lot in this time.

The most recent of which has been that Octopus didn’t balance any of the radiators they installed (which was all of them). Ever since it was installed, the attic radiator barely got warm. It is the highest point so gets a little air and I’m still bleeding it to this day, probably every couple of days. The system pressure isn’t dropping, holding at 1.5bar so I think it’s just leftover air from the first fill. Initially it may have had an air lock. I could shift it by closing the first floor radiators TRVs all the way off and could hear the air moving. We’d have heat up there for a few days and then gradually it’d stop working again. More air to bleed and a slight improvement but nothing amazing. It’d barely register 20c on my infrared thermometer while the rest were 30-35c.

Another symptom I noticed was that the attic rad would get lukewarm during the initial higher pump speed phase and would then go cold when the pump dropped to 10lpm. It’s taken me this long to realise this is probably a balancing issue.

Sure enough, as I went round each rad, all the lockshields were fully open. This explains why the first couple of rads in the loop were always toasty while further away ones were cooler. I got dT on those first ones of 2-3c, same for all the bedrooms, the ground floor rads fared better at 3, 4 and 7. The attic was 0, no flow at all.

After shutting them all down and reopening 1/2 turn first, then 1/4 for some, I got dTs at all the rads of between 4 and 6. When I went up to the attic, which is fully open, it now was warm reading 31c on the flow side for the first time ever.

So… lesson number 1 (which I learned last of all so far), balance the radiators. Probably took a couple of hours with an infrared thermometer pointed at the bottom corners of the rad where the flow enters and exits (most reliable spot I could take measurements).

Lesson number 2 - the Madoka is basically useless (if you want fine control over your climate). On day 1 of operation back in mid October, the house warmed easily, steaming past the Madoka setpoint on an apparent road to oblivion. 19c setpoint, still heating at 20.5c. I quickly learned (with the help of some searches here as well) that the Madoka/Controller has a 2c swing between starting the heating and shutting it off.
It has other quirks too, like if you schedule a temperature change, or even just the same temp as the previous timeslot, it’ll re-evaluate the room and if it’s within +/- 0.3c or something, then it will shut off the heating or start heating, depending. I tried to work that to my advantage for a while but it was very hit and miss. I tried using Home Assistant to control when the heat came on and off but that was also hit and miss - if the cloud API was down, we’d be without heat or baking. I really wanted the system to be as self-sufficient as possible, without relying on the cloud. Which, again with the help of other forums posts here, brings me to…

Lesson number 3 - LWT is king, if you can get the weather curve right. As soon as I realised that you can still schedule with LWT (albeit on the MMI only), I knew that it could be used to schedule in an effective overnight setback (-2c seems about right).
But here is where Home Assistant can actually help. Once the WD curve was dialled in, optimised for gloomy days with no solar gain and little to no human activity generating heat, HA can use an external thermostat (three Switchbots on the ground floor in my case, the median is used) to adjust the flow offset down a couple of degrees if there is a warming trend. If my wife is baking, it goes down a couple more. So the heat pump is always running at some level. This keeps energy consumption more even, reducing the large warm up energy gobbles. So now the heat pump is pretty much self sufficient with a little guiding from HA very rarely. LWT mode also prevents the heat pump from shutting off completely. This also makes the Madoka useful again, because you can dial in manual offset adjustments if you feel the need and since this doesn’t cause the heat pump to completely stop, like if adjusting the setpoint in thermostat mode, it results in a more comfortable climate and keeps the heat pump active in some capacity.

Various other things I learned included, the water pump doesn’t need to be at 100% to work, at least for my installation, 60% (9-0D to 4) is much quieter. Indeed 100% didn’t resolve my attic issue at all partly because after 25 minutes or so, it drops to 10lpm whatever you do. Another useful (IMO) byproduct of going to LWT mode is you get sampling for free which means the water pump isn’t running constantly when heat isn’t required.

ESPAltherma is really useful if for no other reason that calculating COP on the fly. And it wasn’t nearly as scary or complex to setup as I imagined. Using the MMI ‘Produced Heat’ as a guide, I worked out that the ‘Leaving Water before BUH’ sensor was too generous and the ‘after BUH’ sensor too miserly. Adding 0.4c to the after sensor in the Home Assistant template brings the daily kWh produced calculation very close to that which MMI claims. The CoP numbers seem reasonable. Daily seems to be 3.2 which includes DHW to 50c every night as well as space heating.

I don’t need the legionella cycle, so I was able to use 4-03 value 4 and setup a Home Assistant toggle to trigger ‘powerful’ mode, which means we can initiate a manual hot water tank top up without using the resistive heater. We typically use most of the tank daily but rarely run out completely, so a fast reheat isn’t needed. Another lesson I learned early on was that schedule+reheat was overkill for us - robbing precious power from our storage batteries unnecessarily when it could wait until the overnight off-peak rate.

Also, when using LWT mode, it’s important to set the system shutoff external temperature 4-02 which by default is a mad 25c. I dropped that to 19c for now otherwise the thing will never stop running well into summer unless you turn it off manually. Which I guess we may end up doing anyway to stop it coming on on cooler summer nights.

And lastly, and I suppose the first lesson, was that Octopus didn’t bother to program a WD curve. Out of the box it was 50/25 at -8/22c. That probably explains why it overshot on day 1 but seems to be much too high at both ends. After lots of monitoring during a mix of freezing and mild weather, it seems to work well at 45/27, -8/13c (I suspect that 13c might nudge up a bit next spring) without too much offset fettling from Home Assistant.

It’s no wonder that people struggle with all this. I’m reasonably technically minded and been able to work a lot of it out by research, trial/error and some logic. But how is the average consumer meant to deal with all these variables? I know heat loss calculations aren’t an exact science, but if we’re to get mass adoption, some customers will need much more hand holding. Or they’ll simply think “this is a rubbish system and why did we ever ditch gas” etc.

In summary, to get the Altherma to provide a consistent temperature I had to:

  • get the installer reference guide and become familiar with installer mode
  • switch to LWT mode
  • tinker with the WD curve (standard, I know) - it wasn’t even close to the design 50/-3c
  • set 4-02 to 19c
  • set 9-0D to 4
  • balance the radiators << most important one!

Thank you to all who contribute here, your knowledge and experience really helped me get started with my first heat pump.

16 Likes
2

Thanks for sharing your learning and experience @chrisg !

1 Like
3

Well this gets even more interesting. A week ago, our system shut down with error C0-00, Flow Sensor. After the somewhat daft manual reset on the MMI, the unit started up again and was good for another two days when it shut down for the same reason again. Curiously a few minutes after the DHW schedule kicked in both times. Daikin attended today and replaced the sensor. They noted that the bypass valve next to the volumiser was wide open! I was wondering why it was struggling to hit dT 5 - it would at the lower flow rate but at the startup rate (even at 60%), it would struggle to reach 3. I can now hear a marked increase in flow through the radiators since Daikin made the change to the bypass valve. This appears to have reduced the initial flow rate down to 12lpm which I assume means the pump now has some resistance which it didn’t have before. I wonder if I’ll need to rebalance the radiators now. Thoughts?

2 Likes
4

Yes. In fact they explicitly told me they were all fully open, since that was best way to run the HP. But like you, I found two radiators were stealing all the heat. Once I turned those down, things were better, but I’ve never managed to get drop on them all the same - hall radiator possibly doesn’t get enough flow ?

Yes, I had initially tried using target room temperature for setback, but have now given up on that. But the Madoka does let you use modulation, which should be some form of feedback to reduce LWT if room gets too hot, or raise it if room is too cold. (But the steps are crude.)

Are you saying that the MMI can set up a timed offset for the LWT ? I’ve not seen that one. I know you can adjust it manually, but I’ve been using the API to adjust it programatically. Though lowering it inevitably causes the compressor to stop to allow the water to cool.

I’ve dropped max pump to 80%, but might take it down to 60%. Though it spends most of its time down at the minimum of 10l/min, and only goes up to 24 when it turns the compressor off and wants to cool the water (and then get back up to temperature).

5

Yes indeed! You can set it like you would a standard room temperature schedule (on the MMI only) but instead of defined temperatures, you dial in + or - offsets to LWT and the start time that offset is effective from.

It seems like Octopus installers have been told to ignore plumbing basics and just open all the valves - presumably hoping for the best and maybe in most cases it just works or the punter doesn’t notice any issues.

1 Like
6

To be able to schedule changes to LWT, you need to be in Weather Dependent mode rather than Room Thermostat.

One thing to note, if you switch to WD mode you are able to set a LWT schedule. If you then switch back to Room Thermostat the LWT schedule is still applied (though there is nowhere this is editable in either Onecta or via the MMI, unless you switch back to WD mode). I’m not sure if the same applies the other way around i.e. Does the room temperature schedule still apply if you create a schedule in room thermostat mode, then switch to WD and create a LWT schedule.

I would say that this is a bug/not by design. However, if you are aware of it, you can use the flexibility to your advantage.

1 Like
7

Just to be clear, you need to be in LWT mode to schedule an LWT offset.
WD is independent of LWT or RT modes.

That’s an interesting side effect @mortstar. Good to know!
Once switching from RT to LWT mode, the original temperature schedule is no longer active.

8

Could be useful for Octopus Cosy, set a negative offset at peaktime, and a positive offset in the cheap rate before peaktime, using room temperature to prevent overheating.

9

I’m finding that my system doesn’t seem very good at lowering LWT once it has reached some level. I had been trying lowering it (-1) in the evening, and then boosting it a little (+1) overnight during the cheap rate, and back to 0. But having elevated the LWT, it just tends to stay there, rather than slowly dropping back down.

Perhaps because it’s a 9kW unit, and probably oversized - it’s already at lowest input power. The pump tends to be running at lowest speed (10l/min), so it ought to be able to increase pump speed to reduce temperature that way? (Same power for higher volume of water.)

10

As far as I’ve been able to see, there is no way to adjust the lower flow rate, that it uses for the first 30 minutes or so, only the initial flow rate upper limit.

I suspect +/-1c isn’t enough to have enough of a difference. Are you monitoring flow temps?

11

That was an interesting couple of days. With a full house of 6 (usually we’re 3) and lots of Christmas cooking, even an offset of -4 wasn’t enough to stop the house overheating. Ended up turning the heating off for most of the day on Dec 25. This doesn’t discourage me from sticking with LWT mode - mainly because the Madoka 2c hysteresis is intolerable. But instead of using HA to toggle in offsets, I’m having it turn off the heating completely if the house goes over a threshold. This resulted in dramatically reduced running costs on Christmas Day, despite the cool weather outside and a comfortable home.

I expect that outside of the holiday periods, LWT mode will mostly keep the house at the base level of heat needed with occasional shutdowns. The higher restart power when the flow temp has dropped is more than offset by the long periods of being completely off. That said, I’ve not had a shoulder months period yet - so I think we may overheat then.

Does make me wonder if this thing is oversized. I mean it probably is, that seems to be a theme with heat pump installations. The design temp was 50c at -3c and that was much too much. At the last cold snap, we were comfortable at 40c/-3c.

Octopus fitted an 11kw unit and the survey listed radiator output of 9kw.
Given the actual design temp required is 40c, or about 80% of the design temp, does that mean that we’d be ok with a heat pump sized 80% smaller? I’m not sure how the calculation works on that front but something closer to 8kw was more what I expected before starting this journey (based off the A2A heat pumps we were using for a couple of years before).

I sorta think that the heat pump shouldn’t need this much 3rd party intevention to provide a comfortable climate. It really feels like we should have a smaller unit that can modulate down much lower than the EDLA11 can, thus chug along at a lower output when needed. This may also make the Madoka able to do its modulation near setpoint which it doesn’t seem to be able to do as it stands (if it’s already at minimum, it has nowhere further down to modulate to).

Should I push Octopus for a smaller unit?

3 Likes
13

8 posts were split to a new topic: Correct setting for Automatic Bypass Valves? (ABV)

15

More learnings…

The Madoka can modulate in both directions. At modulation 10, what I observe is:

  • at 0.5c over setpoint, the first modulation kicks in. Flow target drops by 3c.
  • at 1c over setpoint, the second modulation fires. Flow target drops by another 3c.
  • if the room temp then drops below 1c over setpoint, flow target raises by 3c.
  • interestingly, the first modulation is only cancelled when the room drops below the setpoint.
  • at 1.5c over setpoint, heat pump switches off completely.

It also modulates in reverse…

  • at 0.5c under setpoint, the flow target increases by 3c.
  • I’d imagine the same is true at 1c.
  • If you adjust the setpoint up, say from 19c to 20c and the current indoor ambient temperature is 19c, the flow target is immediately increased.
  • the same is true in reverse, but if you reduce the setpoint to, say 19c but the room is 19c or above, the heat pump will shut off until ambient drops to 0.5c below setpoint again.

I’ve been running in RT mode for the last week and I think using the Madoka is useful, if you want to avoid third party tweaking using Home Assistant and the Daikin API.
I think my issue with it in mild conditions is the oversized heat pump can’t modulate its output low enough so we overheat the house. I’m still not decided whether to use the LWT/HA method or RT. Perhaps it’ll become clearer if I have the right size heat pump.

After having further radiator troubles, I decided to try reopening all the lockshields fully, to see if we still had the problem of the attic radiator not warming after Daikin set the ABV correctly. I left the landing radiator at 1/4 since it is the first in the loop. All rads have worked since, although the attic rad is cooler than the other first floor rads when the circ pump drops to it’s minimum 11lpm but it is at least producing heat. If anything, upstairs is maybe 1c warmer than ideal, so I might yet throttle them all down to maybe 1 turn open or something, maybe.

[edit: I ended up throttling down the landing to 1/2 turn, bed 3 also 1/2 turn - it seemed to heat quite easily, downstairs loo and bathroom both 1 full turn open. Everything else fully open. Now we have heat everywhere, including the attic and good flow. I’ve also had to reduce the flow target a couple of degrees because the house was overall on a warming trend again - which suggests further that the flow was too restrictive across the whole system.

So it seems the original issue with the attic rad was the failure of someone to set the ABV correctly. Yay. So much time and effort wasted on that.

Also, one of my first posts was how noisy the circulation pump was at full chat. That also went away after the ABV was set correctly. I’ve kept it at 100% and it seems to be running much more stably now. Defrosts are still a mess and reducing the pump speed to 60 or 70% hasn’t changed that. I think the real issue was, despite the flow sensor reporting x lpm, the flow around the system was too restricted. DT during the cold snap peaked at 10-11c after defrosts, then eventually settled at 5c. After opening all the lockshields, this peak reduced to 8c but it settled at 5c much more quickly. Will need to wait for the next cold snap to know if the backup heaters would kick in again.

Annoyingly, I’d probably never have touched the lockshields if that attic rad was working, if the installer had set the ABV at commissioning it would have been. So many hours lost. Why didn’t I ask Octopus for help? I did. They promised to send someone but that someone didn’t turn up.

Oh yes, and the thing nobody tells you about heat pumps…
There is a lot of care taken to keep the heat pump at least 1m from the boundary and there are so many videos on youtube showing how quiet they are. But when its frosted up, it’s so very loud with high speed fan noise. It’s like a constant roar. Then, when it enters a defrost, you get a loud sound like a truck releasing it’s air brakes. If I had known, I’d never have agreed to site it on the patio below our bedroom window. And they are triple glazed. Maybe it’s something you get used to, I don’t know yet but for sure we can’t ever open the window again if we want a good nights sleep. Even the trickle vents have to be shut to keep the noise out.

An Octopus installer is coming out tomorrow.

2 Likes
16

This ^^ !
I have a long list of things which the average customer would have just suffered. The reputation of the technology is one thing, but the wasted energy is another. IMO Ofgem should require proper elec and heat monitoring and claw back the BUS grants from installers if they fail to deliver a system with satisfactory performance during the first winter. That would focus minds on quality of design (incl heat loss calcs) and installation.

6 Likes