AA uses your weather compensation curve as a starting point some info here.

Thanks for the page @gary

I'm re-posting it with higher resolution and colour correction to help with legibility  🙂 

If you want to check the context, I found the original diagram here in this Mitsubishi document.

@clockworks - it seems that AA was originally conceived as a concept to help switch intelligently between air-conditioning and heating modes for Mitsubishi units that do both.
In the UK, those air-con units aren't approved by the MCS, so we only have the heat-pump functionality.

We are therefore only experiencing a portion of the overall software strategy.
There may be better ways for us Brits to achieve greater efficiency than relying on AA, but it's a good starting point.

Thanks for the link and the chart. Says what the aims are, which is what I expected it to do.

I've decided to leave the Evohome relay connected for now, but set one of the TRVs to well above the temperature that it can ever reach. That'll keep the ASHP running just on the schedule, and the room temperature in the hallway. 

At least, that's how it seems to be working -

ASHP shuts right off if the Evohome is set to "heating off", or if all the TRVs are up to temperature. 

ASHP running and modulation (and hence flow temp) set by the wireless remote sensor  (actual room temp) and the room set temperature. Bigger gap between set and actual, compressor works harder, flow temp increases.

Now that it's in AA mode, it's still cycling, but so far not as much - twice an hour, rather than 3 times an hour.

Part of this is because it's spending longer with the compressor off during each cycle - 10 to 15 minutes, rather than 4 or 5 minutes in WC mode. This seems to because it's using the full 10 degree flow temp drop defined by the thermo diff setting, rather than switching back on as soon as the flow temp and return temp are the same (generally around a 6 or 7 degree drop).

I guess I'll leave it running like this for a couple of days and see how it goes...

Great sounds like you have it functioning correctly.

Graphs for midnight onwards this morning, cycling much reduced.

Running in Auto Adapt mode, with the Mitsubishi wireless thermostat set to 0.5 degrees above the temperature the hallway normally reaches with the radiator "open" when running in WC mode.

For info, I have the heating scheduled to go off at 20:00 (house is warm enough), then come back on at 23:30 when the cheap rate starts. DHW heating starts at midnight, tank temp set to 49 degrees. I could use a setback temp, rather than a hard switch off I suppose.

It does appear to run better in AA mode, but I can't be sure as it was a fair bit colder outside last night.

Presumably I can temporarily switch to WC mode, tweak the WC curve down a little, then switch back to AA mode, and it would then base the target temps off the new WC curve?

I think I'm going to leave the Evohome boiler relay connected, as it will switch off the ASHP if all the other rooms go over temperature while the hallway is still below target.

Things are slightly complicated because the pin is sticking on the hallway radiator TRV (sometimes doesn't open up), and the lockshield is on the inlet side. I've taken off the Evohome TRV heads, and I'm using the base handwheel to balance the rad, with the lockshield fully open.

A new radiators and valves will be going in at some point.

@clockworks Meanwhile … you may be able to exercise that sticking thermostat pin. Place a coin on the top of the pin and press down - release - press again and repeat a few more times, you may find this will ease the movement of the pin until replacement time arrives. I think that either the lubrication used in the assembly dries out or a little leakage has allowed some water through and this evaporates away allowing the additive in the system to form a ‘gunge’ over time. Regards, Toodles.

Jiggling the pin gets it going again, but it just jams again if I  (or the TRV) closes it right down.

I've left the Evohome TRV head off, and I've adjusted the wheel part to balance the flow. I've also deleted the hallway zone from the Evohome controller, leaving the Mitsubishi wireless remote "in charge".

The rest of the Evohome TRVs are just acting as limiters now.

2 days of running in AA mode, and we are back to the 20 minute cycling once the house has recovered from the 3.5 hour "setback" period.

Power consumption and flow/return temps from 08:00 this morning.

Outside temp varying between 8 and 9 degrees, which would give a target temp of 38 or 39 degrees according to the WC curve.

Mitsubishi wireless remote is showing 19 degrees, with a set temp of 19.5.

It looks like AA is doing its job, as the flow temp was a fair bit higher, with no cycling at all, while the house was warming up after the setback - about 2.5 hours.

The hallway temperature has been very stable, varying by less than a degree over the past 2 days. It never actually reaches the 19.5 degree setpoint though.

I'm going to try reducing the setpoint to 18.5 degrees, see if that'll make the heatpump switch off for a bit, then work harder without cycling when it switches back on.

Controlling the "on/off" using the Mitsubishi wireless remote isn't going to work in practice.

It does switch off, but only if I set the target temperature to 1.5 degrees below the actual room temperature.

It will only switch on again when the room temperature is 0.5 degrees below the target temperature.  A hysteresis of 2 degrees.

By the time the hallway has dropped 2 degrees, the rest of the house will be uncomfortably cold.

The way it's set now, the wireless remote can switch off the heatpump even if the Evohome controller is calling for heat.

The Evohome controller can switch off the heatpump even if the wireless remote is calling for heat.

So, both the wireless remote AND the Evohome controller need to be calling for heat for the heatpump to run.

I'm not really sure what the way forward is here, if I want to keep using it in AA mode?

I wonder if I can set it all up so that the Evohome does all the heat calls, but the wireless remote gives some performance feedback to utilise AA mode without being able to turn the system off?

With the wireless in the hallway, and the temperature there barely moving no matter what the rest of the house is doing, I doubt it's giving any useful feedback.

Maybe it would be better if it was in the kitchen/diner, but only if it couldn't switch the heatpump off....

Edit:

I might have found the answer.

The instructions for adding the Misubushi wireless remote say to change the sensor settings in the FTC to "Room RC1", with Room RC1 assigned to zone 1.

With settings, the remote (and its receiver) are assigned to zone 1 as the temperature sensor, and also a switching device.

I've changed one of those 2 settings, switching from "Room RC1" back to "Main RC" (how it was set before I added the wireless remote".

This seems to have stopped the wireless remote switching off the heatpump, while leaving it as a temperature sensor (the main display is still showing the room temperature) and AA mode is still active.

Hopefully this will leave the Evohome in charge of calling for heat, while the wireless remote can be used for adjusting the temperature offset, and providing performance feedback for AA mode

Looks like I got it wrong.

With it set to "Main RC", it seems to be using the temperature sensor in the wired controller, but the wired controller is displaying the temperature set by the wireless controller.

All very confusing. I've changed it back to how it was.

I don't think AA mode is going to work for me.

The outside temperature dropped sharply overnight, from 8 degrees to 3 degrees in less than an hour.

The hallway, where the wireless remote is placed, was at 19 degrees, with a target temperature of 20 degrees, and had been ramping up steadily after the setback. Flow temp was 37 degrees, pretty much where the WC curve said it should be. As the outside temperature started dropping, the heatpump ramped up the flow temp. On pure WC, it should've just increased the flow temp to 44 degrees.

When I got up just before 7am, it was flowing at 49 degrees!

I think it did this because the temperature in the hallway was responding very slowly - it just kept pushing the flow temp up, waiting for the room temperature to increase by a certain amount. When the room temp didn't increase enough to register, it just kept ramping up the flow temp..

Seems like it was caused by the stability of the temps in the hallway. As shown by a Tapo sensor (0.1 degrees increments), the hallway temp had a total swing of 0.9 degrees over the past 24 hours. The temperature in the hallway does go up and down with the heating, but not very much.

The Mitsubishi wireless only has a resolution of 0.5 degrees, so it can't really track what's happening in a room with such stable temperatures. 

Because the flow temps were so high, all the other rooms got up to a degree or two over the required temperature, and the TRVs started shutting them down.

I'm sure AA mode would work, if I could find somewhere to put the sensor where it was giving useful feedback.

Kitchen diner gets massive solar gain until a few hours after noon, and would shut down the rest of the house too quickly on a cold but sunny day.

Living room has the same issue in the afternoon.

The downstairs bedrooms are kept cooler.

Upstairs (2 small bedrooms in the dormer conversion) doesn't need to be heated outside of working hours.

I think I'll have to go back to WC mode.

Posted by: @clockworks

↑

When I got up just before 7am, it was flowing at 49 degrees!

I think it did this because the temperature in the hallway was responding very slowly - it just kept pushing the flow temp up, waiting for the room temperature to increase by a certain amount.

Hmmm.

Let me extrapolate further from your observation.

If your house had a concrete pad in that hallway, with UFH embedded in it, then the symptoms you describe would always occur.
UFH has a large thermal lag.

So that suggests the Samsung AA algorithm hasn't been designed to operate with UFH...
... which is a significant oversight.

What do you think @jamespa ?

Just to clarify, the hallway is heated by a 600x1100 T21 radiator.

It does leak a fair amount of heat up the stairs, and into the integral garage (older house, no fire door).

The heat leakage seems to be very consistent though, as there are no windows or external doors.

I guess the radiator input is very closely matched to the heatloss, so it's stable. No TRV fitted now.