This sounds like what I noticed in our heat pump a few weeks ago.  The Honeywell room thermostat was set to have a 'cycle rate' of 5 times per hour.  I.e. if the temperature was too low it would come on, but only for 5 or 6 minutes, then off for the same, then on again until after much struggling the room temperature reached the desired one. 

So if you have the instructions for the room thermostat, what does it say about 'cycle rate' or similar?

its important to differentiate between, is it cycling because a thermostat has told it to turn off? Or is it cycling because the actual LWT is at or above the target LWT, and/or the RWT has got too high (so, delta T has got too low for the HP to be happy with it) and so the HP's internal control logic decides to "not add heat to the system" for a bit, to allow the circulating WT to drop? That second option is not the same as saying the house isn't calling for heat.

3rd party stats may cycle between on and off state too quickly for a HP. The samsung controller will not. @petra I deduce you are probably using the samsung controller?

2091 on the last option what I have always seen recommended (by people who genuinely do this for a living, I don't I'm an enthusiastic amateur) for a 3rd party stat install. 2093 is the exact equivalent if using the wired controller. That option allows the HP to do the back-off that I mentioned. the 3+7 min cycling is normal when the HP is backing off from adding heat.

However, what this can tell you about your system if its happening a lot / continuously, is that you may have an imbalance between emitter output and heat input, because the HP has to keep on stopping adding heat. If you have balance between input and output (which is what you are always aiming for), you will not see this backoff, or at least not see it much. In my system (full system performance link accessible in my sig) it can take an hour to stabilise (I am in my first winter so still tuning the WC numbers), here's this morning with some annotations:

so , if this is happening all the time may be worth working out what the heat output of your emitters is at the temperatures you are using, and does that match what the HP is putting in? Have the  emitters balanced (at the flow level, i.e. radiator lockshield valve adjustment), do they all get sufficient water flow so they all contribute? Do any get turned off by TRV's or similar, reducing available emitter output capacity ? also your system design can influence this: is it a single loop, or is there a buffer tank or low loss header splitting the system into primary and secondary? The latter two can make it harder to get the heat that the HP puts in to the buffer/LLH, back out again to the emitters : there has to be a flow-rate balance between the two sides of the system.

@petra how are you monitoring your system, you must have something that is telling you compressor is on or not? Do you have a way of collecting heat throughput data, a heat meter or similar?

@guthrie The room thermostat is integrated in Samsungs wired remote controller which is placed in my living room. There I cannot find any parameter which sounds like controlling the number of cycles per hour.

@iancalderbank Thank you for your answer. May I just ask at the beginning about the abbreviations you use? LWT and RWT?

Your first question: yes, I am using the Samsung wired remote for controlling the HP.
This 3+7 cycle that I also read of in the user guide isn't followed by the HP (although option 4 is selected in item 2093.

The delta T you mention, is it item 4052 (under "inverter pump")? And what exactly does it mean? And how does it influence the on/off-cylce?

What I observed until  now:  I've chosen the option of controlling temperature by the room thermostate. If the room temperature isn's reached, the HP turnes on. And it stays on until the water temp is reached according to the water-temp-law. So it doesn't turn off before water-law is reached, also if room temp is already reached. And on the other hand it turnes off with water-law also if room-temp is still to low.  Is it meant that way?

I turned on all radiators yesterday (also the one in the basement which is normally turned off), and it really got better after a while. I can show you my power usage curve (that's the only thing I can control over time ... my "normal" power need is about 300 W, so everything that is about 1000 W or higher is due to the HP turning ON. The peaks at about 8:30 AM and 15:30 PM is something else (washing machine), the peak at 11 AM is the hot water which I turn on every day at that time (hoping that my PV is working good at that time of day).

The plateau in the afternoon was very interesting to observe: the water temp increased VERY slowly, so the water-law temp was only reached after several hours. Normally it lasts only about 15-20 Minutes to rais the water temp from about 33 to 40 celsius. So I have no idea, what caused the WP to work that way but I was really happy about that 🙂
Unfortunatly after that, during the night, the "old" behaviour returned:

Is there a parameter in the Samsung controller that influences the number of cycles per hour? Or that influences the time, the water-law-temp is reached? I reduced the delta T parameter to 3 (from default 5) 2 days ago without really knowing, what it does and if it helps reducing cycles ...

Just do understand my initial problem: here's a graph a few days ago (HP was turned off during night at that time) ...

And regarding your further questions: no buffer tank, one single loop (I think), and "low loss header" I don't know what that means (and therefore don't know, if it was installed)
Monitoring the system I do via looking at my power usage via PV App - and by switching through status info screens in the Samsung wired remote. Unfortunatly I have no other monitoring tools available.

@petra LWT is Leaving Water Temperature, that is the temperature of the water leaving the heat pump and going to the radiators. RWT is Returning Water Temperature, that is the temperature of the water returning from the radiators back to the heat pump. The difference between the two temperatures is called Delta T. This usually varies between 3 and 6. The amount of heat energy that the heat pump produces is proportional to the flow rate and Delta T. However, you cannot control Delta T on the Samsung. Changing Setting 4052 doesn’t do anything in my experience. 

The heat pump runs continuously until either the room temperature limit that you have set on the controller is reached or the LWT rises above what the Water Law says it should be. For example if the Water Law setting suggests that the LWT should be 34, then the heat pump will try and maintain the LWT at 34. If it rises to say 36, then it switches off. The reason why the LWT exceeds the target of 34 is usually because the radiators are no longer getting rid of the heat energy produced by the heat pump. When there is an imbalance like that the RWT rises first. This means Delta T and the heat energy being produced fall, but LWT initially stays the same. There is a limit to how low the heat pump’s heat energy can go. When that point is reached the LWT ends up rising above its target temperature and the compressor stops. With some makes, it stays off for 20 minutes, but the Samsung seems to restart after 4-5 minutes. The 2093 setting suggests that after the compressor stops because of the high LWT , the pump will run for 3 minutes and then be off for 7 minutes. In practice, the pump runs for 3 minutes and switches off, but restarts 2-4 minutes later as the LWT has now fallen below the target of 34 and the compressor wants to start again. 

The best way to stop this behaviour is to make sure that all your radiators are open, particularly when the outside air temperature is higher and the heat requirement of the house is lower. When it gets really cold outside, you should find that the heat loss from your home is more than the minimum output from the heat pump. Therefore the target LWT is maintained and doesn’t rise and therefore doesn’t switch the heat pump off. The only thing that then switches it off is if the room temperature target is reached.

The other solution to prevent cycling when the outside temperature is higher, is to make the target LWT quite a bit higher than really necessary, so that your home heats up quite quickly and switches the heat pump off when the target room temperature is reached. Then the heat pump will remain off until the house slowly cools down below the target temperature. This means that the heat pump is still cycling on and off, but much less frequently. Doing it in this way may not be the most efficient or economic way of doing it but maybe worth trying.

Posted by: @petra

↑

The plateau in the afternoon was very interesting to observe: the water temp increased VERY slowly, so the water-law temp was only reached after several hours. Normally it lasts only about 15-20 Minutes to rais the water temp from about 33 to 40 celsius. So I have no idea, what caused the WP to work that way but I was really happy about that 🙂

Unfortunately after that, during the night, the "old" behaviour returned:

Is there a parameter in the Samsung controller that influences the number of cycles per hour? Or that influences the time, the water-law-temp is reached? I reduced the delta T parameter to 3 (from default 5) 2 days ago without really knowing, what it does and if it helps reducing cycles ...

Just do understand my initial problem: here's a graph a few days ago (HP was turned off during night at that time) ...

And regarding your further questions: no buffer tank, one single loop (I think), and "low loss header" I don't know what that means (and therefore don't know, if it was installed)
Monitoring the system I do via looking at my power usage via PV App - and by switching through status info screens in the Samsung wired remote. Unfortunatly I have no other monitoring tools available.

very likely what made it better was that you turned on that extra radiator. it is amazing to see the improvement in stability that comes from adding extra emitters / getting the emitters properly balanced so they all carry a fair share of the load. I am still working on this weekly in my own system. There's no parameter for cycles/hour. I suggest you don't focus on solving that directly  - focus on getting your system setup so that its stable, then the cycling will just sort itself out.

if you don't know if you have an LLH...  Whether you have one or not will actually make a really big difference to how your system is operating. I would suggest at this point you draw a logical diagram of your system. Lines for pipes, boxes for components (valves, pumps etc). There are some good examples on this forum from other members. then post it.

depending on how keen your are , and how much of a DIY'er / technical person you are (or not), there are various ways to setup more detailed monitoring, so that you know exactly what your heat pump is doing.  where do you sit on that spectrum?

@petra Ah,that sounds like a more sensible way to set things up than my system.