Thanks for the description @clockworks, that gives a better picture of your setup.

I’m not sure if I’m interpreting your graphs properly but it appears your return water temperature seems to always rise very quickly in line with the LWT rise. If the emitters are working properly your return water temp should stay more or less horizontal. Unless at the initial ramp up when there is a steady rise to the steady state temperature.

If correct This would suggest your emitters are not offloading the heat effectively.

there could be many reasons for this from emitter size, low system volume  to unbalanced LLH flow, secondary Pump speed.

have you got any picture of your site built LLH. 

Also can you post a MELCloud hourly graph of your flow and return. I can read them better than the one she ive seen. I find the vertical axis difficult to read.

here is one of ours showing 5c vertical increments you can perhaps see the RWT I see mostly horizontal when in steady state.

If you post Melcloud graphs its worth clicking off all the other series you don't need, like set tank temp, this allows the plots to rescale to the data of interest.

Melcloud graph for the last 24 hours. I have the heating scheduled to go off at 20:00 and back on at 23:30. This is because I'm on an EV tariff with cheap overnight rates. I effectively have my overnight setback in the evening. It ran a Legionella cycle last night too.

I'm running in weather compensation mode, so the lines are going up and down yesterday as the LWT tracked the outside temperature changes (between 7 and 10 degrees outside yesterday)

As you can see, the curves look an awful lot smoother than my other ones from Home Assistant. 

This is because the Melcloud data points are an hourly average, whereas the Home Assistant data points are 30 second (I think) real-time snapshots.

This is the same 2 traces, over the same time period, but captured in real-time by the Shelly sensors in the same pipe locations as the Mitsubishi sensors:

As you can see, the graph looks very different!

It's basically the same data though, just the first one is averaged out each hour, while the second one is close to real-time.

@clockworks if you could post the hourly temp plots rather than the last 24 hour ones they will have all the detail that should match your other plots

@gary Not sure I understand what you are asking for?

@clockworks under reports there is an hourly temp plot that you can then choose by hour, its starts with the current hour but if you click on the time at the top of the screen then it gives you a menu so you can choose any hour in the last 24 hours 

@gary Ah, OK. I find that way of doing it more confusing, as it doesn't show the actual time on the horizontal axis.

I get your point about matching the time period if showing graphs from 2 different sources though. In the last 2 graphs that I posted, I tried to match the Home assistant graph to the 24 hour period in the Melcloud graph.

In earlier posts, I used shorter time periods to show what was happening in greater detail.

I admit that it can get a bit confusing for others though.

Posted by: @clockworks

↑

Ah, OK. I find that way of doing it more confusing, as it doesn't show the actual time on the horizontal axis.

Hi @clockworks. I was just suggesting the one hour MELCloud graph so it’s more detailed. I think @gary could see what I was driving at....

the actual time, you are looking for  is visible since the hour is shown on the top heading eg (1pm) and the y axis is 60 minutes long. It would show exactly what’s happening minute by minute between flow and return temps. It’s handy since the scale is big and there should no background ‘noise’. It was just an idea for a unified scale.

Regarding your LLH - Is it a 4-port vertical cylinder with 22mm primaries and 22mm secondary pipes? 

Ah, I see. I hadn't seen that option. The title's a bit misleading- "hourly" meaning "for one hour" rather than what I assumed "hour by hour"

I've pisted above, and it's showing the same cycling as my Home Assistant graphs, but it looks a little less "spiky" because the sampling rate is lower.

I'm in weather compensation mode, and it's reaching the set temperature and shutting down because it's not losing heat to the rooms fast enough. Someone said earlier that the 6kW Ecodan can't modulate below 3kW, and that does seem to be my problem - I don't need 3kW of heat output unless it's a really cold day.

Here's the same time period, taken from Home Assistant, using the Shelly sensors. The flow and return sensors are positioned right next to the Mitsubishi sensors. The "radiator" sensor is fitted after the secondary pump, just before it goes up into the loft.

The radiator sensor tracks the flow sensor pretty closely, so I think the LLH is working OK.

Yes, the LLH is a length of 32mm pipe, about a metre, with 4x 22mm fittings.

This is the power consumption of the ASHP for the same one hour period.

You can see that it doesn't want to run for very long at much less than 850 watts, and seems to prefer to run at about 1kW. This is for the whole setup though, including the controller and pumps

Posted by: @clockworks

↑

Someone said earlier that the 6kW Ecodan can't modulate below 3kW, and that does seem to be my problem - I don't need 3kW of heat output unless it's a really cold day.

 

Here's the same time period, taken from Home Assistant, using the Shelly sensors. The flow and return sensors are positioned right next to the Mitsubishi sensors. The "radiator" sensor is fitted after the secondary pump, just before it goes up into the loft.

The radiator sensor tracks the flow sensor pretty closely, so I think the LLH is working OK.

 

 

Yes, the LLH is a length of 32mm pipe, about a metre, with 4x 22mm fittings.

Hi @clockworks

Yes I sent you the charts which showed 3kw minimum output. 

I was very impressed when you managed to create the same hourly chart as the MELCloud graph…. Very easy to interpret. And it does show the RWT rising quite steeply with no room to modulate 

And I agree that the radiator thermostat (secondary flow temp) is very close to the Primary flow temp… implying there isn’t a massive issue with the LLH.

Furthermore your description of the pipe sizes is very convincing.

So the 3 kWh minimum output may be the issue to overcome especially if lower flow temperatures are what you are searching for.

It's not that I'm chasing low flow temperatures forcthe sake of it,  more a case keeping the house at a comfortable temperature in the most economical way. If it costs me the same as my old oil boiler, I'll be happy.

I've got to balance the apparent economy of the way it's running now (with all the cycling) against possibly killing the compressor unit early and paying for repairs. 

I'll try fitting the Mitsubishi wireless remote, and using Auto Adaptation.

If that doesn't reduce the cycling, maybe I should look at scheduling the heating - get the house up to temperature by increasing the flow temp, then switching everything off for a couple of hours. Might be cheaper overall if I keep the Ecodan working "harder" for longer, but only for 50% of the time?

Or, leave it doing all the cycling, and just get a smaller unit when it dies?

@clockworks or get your installer back to fix it, the cycling you are getting isn't acceptable, the MCS guideline talks about volumisers to ensure there is enough system volume to prevent cycling.

What happened during the cold snap what flow temps were you getting then and was it still cycling?

What happened when it defrosted did it recover quickly, did the house temp drop during defrosts, if so, then this all points to low system volume in my opinion, that your installer should be open to rectifying, do they know what your system volume is?