@papahuhu I would definitely check the upstairs radiators. If they were locked down too much, that would effect the volume of the system and might account for frequent defrosts and swings by Homely. How long has homely been learning your house?

There are some low cost monitoring options for samsungs which might help you diagnose what is wrong

@bontwoody It’s been converted from 3 zone GCH to single open loop. All TRVs are fully open (on 5). I’ve bled it multiple times, there’s about 36l/min flow rate which although slightly less than nominal, should be fine. 
Homely has been running it since Sept, I was running on “smart” mode but it was causing short cycling (4 or 5 starts/hr), homely told me to switch to “standard”. It went from short cycling to not short cycling but to the “on 50% off 50%” cycle I have now.

Installer is offering to remove the homely and move the Samsung control interface into the habitable area from the garage. What’s the reasonable value monitoring solution please, without the homely it’s not Wi-Fi enabled?

@papahuhu OK. what about the lockshield valves, have you tried opening those with the TRVs set to max? Are the upstairs radiators a similar temperature to those on the ground floor? Have you got the Heat loss survey, if so does it flag the radaitors upstairs as being undersize for your flow temperature

Have a read of this:

https://community.openenergymonitor.org/t/monitoring-your-samsung-ashp-controller/27638

@old_scientist 

Thanks. How can it possibly detect frost if it’s not touching the coil please? It’s a good 5cm away from the fins.

The only way I can foresee it might work is if it’s measuring both temp and humidity and using the dew point and mass air flow and refrigerant temperature to calculate the theoretical rate of ice deposition. 

@papahuhu I think the rear sensor is probably a bit of a wild goose chase, unless its faulty.

@bontwoody OK. Just thought it looked different to others I’d seen, it was a straw to clutch at at. I used to have a heat pump decades ago on mainland Europe, it worked a charm.

@papahuhu I would concentrate on the upstairs radiators. If the downstairs rooms are warm then there is enough heat when the emitters are working properly and are appropriately sized

@bontwoody I have been. I forced a repeat heat loss survey. 
The surveyor when he was here told me they had made an error in the rooms that are cold, using the wrong U values. He took the data back and now the official feedback was that it’s all been done correctly. I’ve asked for the data and they have gone silent for the last few days.

But even if the rads have been undersized, it doesn’t explain why the heat pump only operates 50% of the time and doesn’t modulate.

@papahuhu That may be telling. Measure your radiators. Take a photo and measure the room its in. Let me know roughly its date of construction and we should be able to have a good go at if its the right size. At the end of the day if you need bigger radiators in a couple of rooms its not the end of the earth. 

Still answer my other questions too. Eg temperature of rads etc

@bontwoody 

Here’s one example where it runs 4 C below the reference temperature (the homely node is in the lounge underneath at design temp 21C). You can see they reduced the output by 50%! In this room there is a room in roof above it, half of the ceiling area has 300mm insulation and half has 95mm, I believe they have assumed it’s 300mm entirely. 

@papahuhu Im struggling to understand why they took a bigger radiator out? What type of construction is your house? when was it built? At 28 W/m2 Im guessing its quite a new construction?

The easy fix will be to top up the attic insulation

@bontwoody Me too, I formally questioned the decision 3 separate occasions prior to the install. Their answer was that the heat loss calculations indicate they have selected correctly. 
It’s a late 70s, brick and block, cavity was filled about 20 years ago. All double glazed, 15 years ago. Rads were original, when single glazed and no cavity fill.

Because there is an unheated room in roof above  that’s not so easy. We already went from 100mm to 300mm in the loft space, but the central portion can only fit 95mm rock wool between the rafters. It’s too fancy to call it a room in roof, the previous occupants created it to send their kids up there to play.