From what you show, 15:00-19:30 ish is standard defrost (every 40 mins) and you aren't hitting target LWT; then after that the cycles extend to around 1 hour, and you are hitting LWT target. It's whether that improvement is worth the cost of the COP=1 power of the LLH heater.

Can you add external air temp to your graph? Maybe it was just warmer after 19:30? The key temp for a Grant seems to be 3C and lower.

Do you know how much power the LLH is drawing? It's a 3kW element, but the Grant docs suggest it'll only pull 250W in "defrost cycle help" mode. I'm a little sceptical.

I will add another version of that chart with outside temperature later in the week - I'm away from my home assistant server right now so can't access the data unfortunately.  The temperature was mostly between 0 and 2deg during that day and did rise slightly into the evening but only to about 3deg and the unit was still defrosting fairly frequently into the night.  I would agree with you that the issue is mainly confined to around 3deg outdoor temperature and below.

I'm also a bit sceptical about the 250W figure - for starters, presumably they mean 250Wh rather than 250W.  It would also depend on how you set it up - you can define the flow temperature at which the immersion heater will switch off.  I set mine at 40degC and looking at the data the defrost cycle results in flow temperature being below that level for about 12 minutes.  I'm not sure at what point the immersion heater kicks in but 12 minutes out of every 37 minutes is about 1/3rd of the time so that would equate to about 1kWh of electricity consumption every hour - on a day like last Thursday that would have been quite an uplift on my total electricity consumption.  I don't currently have any means to check exactly how much is being used by the immersion.

I did give some further thought to my idea of switching off the central heating circulation pumps during the defrost cycle.  With the data I am collecting from the heat pump I have a sensor for "Defrost Temperature".  It appears that just before the Defros cycle begins the defrost temperature drops to about -7deg.  At the moment the defrost cycle starts the compressor switches off and the heat pump power consumption drops to 100W and stays there for 3 minutes before the defrost process begins.  So using the combination of the defrost temperature being below -6deg and the power consumption going to 100W I can trigger an automation which switches my three underfloor heatings zones off for 10 minutes.  This immediately switches off the underfloor heating pump an means that the colder water is not being pumped around my underfloor heating circuits.  The heating demand remains on from the other zones so the heat pump will go into recovery mode as soon as the defrost is finished.  This should provide two benefits - 1. the heated floors are not being actively cooled during the defrost - the water in the UFH pipes should stay warm and continue dissipating heat into the floor.  2. The remaining zones should be able to heat up more quickly as there is a smaller body of water to be heated by the heat pump during it's recovery from the defrost and by the time the UFH switches back on the heat pump is more or less back up to temperature. That's my theory anyway - but I'm probably displaying a complete lack of understanding of the laws of thermodynamics.  Now I just need some more cold weather to test it out properly - I did have it running overnight and the trigger was working correctly but I didn't quite have my signals to the ufh heating zones working correctly.  I also may need to fine tune the 10 minutes upwards a bit. 

  Here is a period where my OAT was around 2 deg, with no defrost cycles.  Heat pump is Samsung Gen6 16kW

Chart #1 - 4 hour period no defrost cycles

Chart #2 - Same heat pump, slightly cooler OAT, a couple of defrost cycles 11 hour period - defrost about once per 90 minutes

Chart #3 - Bit more difficult to get data out quickly further back, but here is some data showing about a 50 min gap between defrosts, this is when OAT was below zero.

From what you show, 15:00-19:30 ish is standard defrost (every 40 mins) and you aren't hitting target LWT; then after that the cycles extend to around 1 hour, and you are hitting LWT target. It's whether that improvement is worth the cost of the COP=1 power of the LLH heater.

Can you add external air temp to your graph? Maybe it was just warmer after 19:30? The key temp for a Grant seems to be 3C and lower.

Do you know how much power the LLH is drawing? It's a 3kW element, but the Grant docs suggest it'll only pull 250W in "defrost cycle help" mode. I'm a little sceptical.

I have the immersion heater connected up on my LLH and have experimented a bit with this. There are several settings linked to this which are a bit confusing and there are two different ways it can be used - or potentially a combination of both. 

1. It can be used as a supplemental heater so it switches on when outside temp is below a set point AND the temperature in the LLH is below a set point. E.g when it’s -2 outside and LLH is below 40 degrees. This one definitely had an impact but also has the potential to be quite inefficient depending what the set points are as it can end up being on for a prolonged period.  I didn’t try this for very long but it looked like it was causing the ASHP to throttle back a bit - presumably detecting that the delta between flow and return temperatures was falling so less heat was needed. it did help to bring up the temperature peaks between frequent defrost cycles though and also extended the time between defrosts because it was allowing the heat pump to work less hard.  

2. This option is supposed to support the defrost process by providing some extra heat into the LLH while the defrost is taking place but only for a limited time.  This is more subtle than option but did seem to help the recovery from defrost to get back to the target temperature.  However it didn’t seem to reduce the duration of the down time.

here’s a plot from yesterday showing normal defrosts for the first part and then the impact of the immersion heater in the second part.  I didn’t do this very scientifically unfortunately so I can’t remember exactly what settings I had applied in each case

I might try a lower target flow temp next time the conditions are right for a lot of defrosting and see if that helps. 

I do have a LLH so there’s a bit of a thermal store there to support the defrost if the heating circulation was to be stopped. I did have a go at manually stopping the central heating pumps today when I saw a defrost cycle starting but I don’t think I got the timing quite right and ended up disrupting the recovery from the defrost instead which prolonged the downtime - I guess the problem with this approach is that when the defrost finishes - if there is no heating demand then it doesn’t start heating up again.  So getting the timing right would be critical and probably quite difficult in the absence of any specific signals. I don’t think I have enough free time on my hands for that challenge. 

The cold and damp weather that we frequently experience in the UK is possible the worst operating conditions for the reliable and efficient operation of an ASHP.

This is the crux of it @derek-m. I've just come off a call with my colleagues in Oslo.

Right now it is -13degC and 74% RH, somewhat typical winter weather for them. Absolute moisture content: 0.91g/kg of air. Yesterday at home in the UK it was 2degC and 90% RH, absolute moisture content 3.94g/kg of air. That's over 4x the latent duty on the evaporator coil here in the damp and drizzly UK.

Scandinavian climes are often quoted as proof that ASHPs work in lower temperatures, and they do, but the milder damper temperatures we more frequently experience in the UK are more problematical for air source evaporators. There's less enthalpy at the lower air temperature, so the refrigerant process COP is lower, but the much lower frequency of defrost means a more consistent delivery of heat to the indoor space. Doesn't matter if it's a direct A2A refrigerant vapour condensing heat pump or an A2W indirect hydronic condensing heat pump - once the refrigerant evaporating temperature falls much below zero, evaporator coils prefer dry colder air over damp humid air.

I’m wondering what I can do to improve this situation?  Would a lower target flow temp reduce the frequency of defrosts and thus increase the total heat energy being delivered?

It seems a bit silly that during the defrost cycle the system continues to pump the rapidly cooling water around the house - I wonder if I could create an automation that recognises the start of the defrost cycle and tells my smart thermostat to stop calling for heat (which would in turn stop the secondary and ufh pumps). Would this allow the emitters to at least stay warm and thus improve the recovery time.

When operating in normal mode a ASHP absorbs thermal energy from the outside air and uses this thermal energy to heat the water being used to transfer this thermal energy into the home.

When operating in defrost mode the refrigerant gas flow is reversed, so thermal energy is being absorbed from the central heating water and being used to melt the ice that has built up on the evaporator. I therefore doubt that stopping the water flow will have any useful effect and may even make matters worse, particularly if there is no anti-freeze in the system.

The cold and damp weather that we frequently experience in the UK is possible the worst operating conditions for the reliable and efficient operation of an ASHP.

The following suggestions may possibly help to reduce the effect of defrost cycles.

1) If possible lower the required LWT setting. This should hopefully reduce the demand on the heat pump and reduce the rate of ice build-up. This may cause lower indoor temperatures, so would need to be a balanced approach.

2) If possible get a secondary heat source for use in damp and cold weather, thereby helping to reduce the demand on the heat pump. A log burner or fan heater may be a simple and cost effective solution.

3) Improve insulation and draft proofing to lower heating demand.

4) Larger heat emitters may help by reducing the LWT, but the heating demand would probably remain the same.