@simonwig Thanks for the bump. Not sure why I didn't see the original message!

Ideal heating: It's on our roadmap but we don't have a concrete timeline yet. We have a few other brands before. If you make an account in the Havenwise app, you'll be placed on our wait list and receive a message when we launch support for Ideal Heating heat pumps.

Batteries: Havenwise has 2 modes for people with batteries. See more info here: https://help.havenwise.co.uk/en/articles/12417331-how-does-havenwise-work-if-i-have-batteries-and-or-solar-panels

@hcas Hello Henri, I know that we have had an exchange on WhatsApp which was not conclusive and I don't find those exchanges all that satisfactory, because they require instant response rather than a considered one.

So backtracking I wonder if you could comment on one aspect, i.e. the apparent minimal response to a target temperature of 24C (increased from 23C) and seemingly stuck at 20.5.

In the previous location for the PAR 60 (the coldest room in the house, our bedroom) we seemed to be getting some improving results both on response and on COP. but since moving the PAR 60 into the adjacent room on 24th Nov (about 4m distance but with more reliable WiFi), the response has been much reduced. It may be that the daytime temperatures have been higher but the night time have been around the same and I have gone back to 24 hr operation to try to maintain bedroom temperatures.

Could you explain why there is currently so little apparent response to a difference of 3.5degrees C between actual and target temperature?

(I did wonder if the PAR 60 was reading low but it is backed up with a Shelly H & T which is nowreading 20.8C

To be fair, since increasing the target temperature to 24C earlier this evening I do see a reaction as below, but it still looks lower than I would have expected. 

@davidalgarve Hi David! For communication with Havenwise, you can choose between email, WhatsApp, and the chat in the app. Whatever works best for you.

On the problem at hand, the root cause is that the machine learning models fail to calibrate well for your home. It comes down to the fact that the data doesn't show a clear pattern between room temperature, activity of the heat pump, and other data like outside temperature. If there are no patterns, Havenwise can't learn.

To make it very concrete: Since yesterday afternoon, the flow temperature has been continuously higher than 45C. And was actually 55C the whole night (benefiting from your cheap overnight tariff). Yet the room temperature is "stuck" at 20.5C. For similar outdoor temperatures yesterday, the heat pump ran less hard (flow temperatures 38-50C), yet the room temperature also stayed at 20.5C mostly. 

If the models can't calibrate well, they can't accurately track your setpoint.

Btw, as you can see in the above explanation, it's not that the heat pump just sits idle. It runs at 45-55C flow temperatures with a deltaT of 7C, while the outdoor temperature is around 10C. So it's definitely putting in a good amount of heat.

I think we have to recognise that Havenwise might not be a good fit for your house. And perhaps the best path forward is to run your system without Havenwise. 

I appreciate that's not optimal. It's also frustrating for us, and we haven't come across a situation like this. We've tried to solve it through various ways including manual calibration by our CTO, but I can't really see another option at this point.

@hcas Thanks Henri. I take your point, but feel that when it is running at the temperatures you describe, it is not just trying to heat the family room (where the Par60 is located) but around 170sq.m upstairs with 8 other radiators and 7 rooms, all in addition to the two rads downstairs. 

I wonder about one last try. At the moment the PAR 60 is in a large North facing family room downstairs. Earlier in November it was in a much smaller adjacent North facing room (possibly having worse wall insulation), but it seemed to work better.

The obvious place for the PAR60 would be in the living room, but this is upstairs and South facing so would be subject to more external temperature influences.

(I should consider that I have previously operated on weather compensation rather than adaptive and it may be that adaptive, with the PAR 60 is not going to work for me even without Havenwise)

Was the decision to locate the PAR 60 downstairs the correct one in an attemt to maintain reasonable downstairs temperatures or would this be worth a try? 

@hcas Hello Henri. It is disappointing, but I must presume that you were not persuaded with my suggestion of trying the PAR-60 in the upstairs living room so could you please advise how to disengage from Havenwise. I see that there is a "Delete account" option on the app, but I am only on a trial so I am not sure if this would be appropriate.  

It is very important to me to know, at that point, what my control situation will be on both MelCloud and the FTC 6, because I have only previously operated on weather compensation and without the PAR-60 having any control.

Thanks for your efforts.

@davidalgarve You can disable Havenwise optimisation in the Settings page. That will turn off any Havenwise influence. At that point, you can control the heat pump via the Mitsubishi controller / thermostat or Melcloud as you did before.

The root cause is that the data doesn't show a clear pattern between indoor temperature and the heat pump activity. That makes it impossible to calibrate the models for your home. As an example, since Sunday night the indoor temperature has barely moved from 21C. The reported room temperature by the Mitsubishi stat has been 21C for all hours but 1. But the outdoor temperatures have been between 10 and 20C, and the flow temperatures between 31 and 44C.
That's highly unusual, especially for rooms with radiators where we always see a pretty clear relationship between the different data points.   

@davidalgarve Potentially stupid question, but are you sure that the mobile thermostat device is working correctly, given that its reading barely changes, based on Henri's info?  Is there somewhere particularly cool or hot that you could move it to temporarily to see if it responds to this?

Unless your house is staying at a rock solid 21°C, and not moving from there, then it sounds to me like the unit isn't functioning correctly.

I am new to all this. We have a SamsungAE080RXYDEH that is not Internet connected. Is it possible to adapt it so that it is and can work with Haven wise? 

@john999 If your Samsung is Generation 7 it should have been supplied with a wifi kit. It its an earlier Gen, then a wifi kit may be available, and Samsung or an A.I search should enable you to find out. Best of luck.

@sheriff-fatman Sorry that I missed your comment earlier.

Yes the PAR 60 is working, but HW have revealed an aspect that I would have liked to have known earlier, i.e. that it has a granularity of    +- 0.5 C and that this is is not ideal for their control.

HW eventually suggested that I should use a Shelly H & T Gen 3 rather than the PAR 60, but having purchased and paid for the installation of the PAR 60, I wanted to persist with it. 

Our main living rooms are upstairs and South facing, whereas our master bedroom and family room/ office are North facing. This (and perhaps smaller than ideal radiators downstairs), dictates location of the PAR 60 in the downstairs bedroom.

I have subsequently bought 4 more Shelly H & T's just to monitor temperatures in upstairs and downstairs rooms and in recent weeks, the system is maintaining reasonable control with a temperature difference of 2C across 6 rooms, so perhaps the system is learning how to deal with the perceived deficiencies of the PAR 60.

However, I am quite undecided, because I don't recall having similar difficulties last year, when operating on weather compensation as opposed to managed adaptive.       

@broadsman Thanks. I think mine is pre-Gen 7. Installed May 2023

Posted by: @davidalgarve

↑

@sheriff-fatman Sorry that I missed your comment earlier.

Yes the PAR 60 is working, but HW have revealed an aspect that I would have liked to have known earlier, i.e. that it has a granularity of    +- 0.5 C and that this is is not ideal for their control.

HW eventually suggested that I should use a Shelly H & T Gen 3 rather than the PAR 60, but having purchased and paid for the installation of the PAR 60, I wanted to persist with it. 

Our main living rooms are upstairs and South facing, whereas our master bedroom and family room/ office are North facing. This (and perhaps smaller than ideal radiators downstairs), dictates location of the PAR 60 in the downstairs bedroom.

I have subsequently bought 4 more Shelly H & T's just to monitor temperatures in upstairs and downstairs rooms and in recent weeks, the system is maintaining reasonable control with a temperature difference of 2C across 6 rooms, so perhaps the system is learning how to deal with the perceived deficiencies of the PAR 60.

However, I am quite undecided, because I don't recall having similar difficulties last year, when operating on weather compensation as opposed to managed adaptive.