See here:

https://renewableheatinghub.co.uk/does-your-heat-pump-have-enough-water-volume-to-run-efficiently/

which suggests 25-27L per kW for very low flow temps, such as with UFH.

What is your heat loss, what is the output and minimum recommended volume for your heat pump, and what flow temperature are you targeting?

What is the volume of water in the slab?

Everything I've read indicates the manufacturers minimum recommended system volume is just that - an absolute minimum, and that substantially higher volumes will be beneficial.

I should note that the OEM will almost certainly not entertain any support if your system doesn't meet the OEM minimum recommended volume.

Further, there was a long discussion of the benefits of a volumiser to add system volume here:

https://community.openenergymonitor.org/t/system-volume/26959/51

For reference, I have a relatively large system volume (around 190L, of which 120L useful, for a heat loss of 4-5kW), and I can certainly notice a drop in temperature on the room thermostat when a defrost cycle happens. I'm glad I have a 50L volumiser.

Posted by: @old_scientist

↑

Everything I've read indicates the manufacturers minimum recommended system volume is just that - an absolute minimum, and that substantially higher volumes will be beneficial.

 

Agreed.  Manufacturers minimum is just enough in theory to avoid a defrost death loop.

Posted by: @old_scientist

↑

For reference, I have a relatively large system volume (around 190L, of which 120L useful, for a heat loss of 4-5kW), and I can certainly notice a drop in temperature on the room thermostat when a defrost cycle happens. I'm glad I have a 50L volumiser.

I cant understand the thermodynamics of that unless you have fancoils or a house with a very low heat capacity. 

The energy required for defrost (which of course does ultimately come from the house) is a tiny fraction of the heat capacity of the house in most normal cases.  With fancoils its plausible that cold air is blown out for a while locally cooling the air (but not the fabric), with anything else the question is - how are you losing enough heat to have a measurable effect on room temperature...unless one of the above is true or your system is doing something it shouldnt be doing during defrost.

What is more plausible is that your system does not recover completely from a series of defrosts and that the cumulative effect is significant.  This could be down to a poor control/defrost algorithm, insufficient capacity or some limiting factor.  Mine does suffer from this when its minus numbers overnight, but its because I enable quite mode which caps capacity well below that required to maintain equilibrium.  My choice.

Of course Im not doubting the observation, just trying to understand how the energy equation adds up!

@old_scientist Thanks for the reply.  I am looking at this question from the perspective of physics.  My system will be open loop with two large heat exchangers including a 1200 sq foot hydronic slab plus a 60,000 BTU water to air exchanger connected to the house ducting.  The system and circulating pump/fans are run always open/always on since the control is simply the ambient temperature curve.  This means circulating water is constantly able to exchange energy.  There is zero risk of a defrost death loop with this system since the house itself would need to run out of btus .. and if that happened the error would be in HP sizing or perhaps a high humidity weather anomaly.  In my case the system will be a hybrid using gas/wood as a peaker on days below -15C .. I am in Canada.  From a physics perspective, the circulating water will always pick up heat .. this means water volume should not matter at all.  My slab has as much energy per 1C as a 40 gallon buffer has per 50C delta!  The HP in theory will never know or care about system volume as long as heat is being exchanged.  I think I will try it on 1/4 the recommended volume just to see how it does .. physics usually works, it makes sense to me, I will save on cost (no volumiser) and less expense glycol (needed where I live) and I can always add a volume later if it doesn’t.

Posted by: @tony-stolz

↑

 I am looking at this question from the perspective of physic

I agree with your analysis from the physics standpoint.  With ufh in slab the slab is going to provide a larger reservoir than the water.   I guess the only other question is the dynamics ie does the hear transfer to the water quickly enough.  You know the rate of transfer from water to slab, at the same DT the reverse rate will be the same.  A back of an envelope calculation will hopefully answer the dynamic question.

Posted by: @jamespa

↑

Posted by: @old_scientist

↑

Everything I've read indicates the manufacturers minimum recommended system volume is just that - an absolute minimum, and that substantially higher volumes will be beneficial.

 

Agreed.  Manufacturers minimum is just enough in theory to avoid a defrost death loop.

 

Posted by: @old_scientist

↑

For reference, I have a relatively large system volume (around 190L, of which 120L useful, for a heat loss of 4-5kW), and I can certainly notice a drop in temperature on the room thermostat when a defrost cycle happens. I'm glad I have a 50L volumiser.

I cant understand the thermodynamics of that unless you have fancoils or a house with a very low heat capacity. 

The energy required for defrost (which of course does ultimately come from the house) is a tiny fraction of the heat capacity of the house in most normal cases.  With fancoils its plausible that cold air is blown out for a while locally cooling the air (but not the fabric), with anything else the question is - how are you losing enough heat to have a measurable effect on room temperature...unless one of the above is true or your system is doing something it shouldnt be doing during defrost.

What is more plausible is that your system does not recover completely from a series of defrosts and that the cumulative effect is significant.  This could be down to a poor control/defrost algorithm, insufficient capacity or some limiting factor.  Mine does suffer from this when its minus numbers overnight, but its because I enable quite mode which caps capacity well below that required to maintain equilibrium.  My choice.

Of course Im not doubting the observation, just trying to understand how the energy equation adds up!

We have large radiators as our emitters. Room temp is 20.0C, flow temp is around 33-35C at temperatures where we may see defrost cycles.

From memory (which is not what it used to be), the defrost cycle takes about 10mins to run during which time cold water enters the radiators (volumiser is on the primary return so doesn't help buffer this). It probably takes another 20mins to get that 190L system volume of water back up to temperature so the whole process lasts around 30mins, during which time there is a drop in recorded room temp (our room thermostat has a resolution of 0.1C).

From a physics pov, the room needs a steady 35C flow temp to maintain target temperature of 20.0C. If that steady 35C flow temperate is disrupted for ~30 mins then the result will be a fall in temperature. Maybe if the building has sufficient thermal mass then if will not be noticeable, but I can pretty much tell you when a defrost cycle occurs (or has recently occurred) by the drop in room temp displayed on the room thermostat (may only drop a few tenths of a degree, I really can't remember but I'd notice if it dropped to 19.8C when it's been sat at a constant 20.0C all evening)

Anyway, it's observable. But my system is not at all like @tony-stolz's so not sure if relevant to them.

The "thermal comfort" we feel is a mix of air temperature and infrared radiation received by the body. This radiation comes from every surface in the room. In a house with bad insulation, walls will be cold, so occupants receive less infrared radiation, so air temp has to be increased for the same feeling of comfort.  So, bad insulation increases losses by itself, but it also requires higher air temp, which also increases losses. That's also a reason why radiators are often placed below windows. It's not just a space saving trick: radiation from the rad compensates lack of radiation from the colder window.

If you are in direct view of a large radiator heated to 35°C, radiation will be an important part of your feeling of comfort. If the heat pump sucks the heat out of the radiator, it will take a while before room air and surface temperatures drop, but radiation from the radiator itself drops immediately as it cools, and this is noticeable. It should work the same for other radiant heating sources like radiant ceilings or floors, if they have low thermal mass.

That's one reason why a boiler with hot flow temp and on/off regulation is less comfortable than a "low and slow" heat pump: if the radiator swings from room temp to 60°C then there will be a large variation in radiation output from the rads, making occupants too hot and too cold on every cycle, even if air temp doesn't change much.

With a nice thick UFH slab it takes hours for the heat to percolate through so there's no chance it'll be noticeable.

@bobflux Yeah .. that is what I see as well.  In my case I’ll run the pump in parallel, always on, through the floor as well as through a whole house water to air exchanger with only mechanical valves to balance heat flow.  The floor will move at most 1C though a defrost cycle .. the air stream to the house will cool a little as well of course .. but that is a first world problem IMHO and one that I will happily sacrifice for the 25% increase in efficiency and lower capital costs vs buffered systems.  I’m planning on using lots of temperature and flow sensors to gather data on what I am doing and to share this data with my local supplier .. cheaper, more efficient systems means more systems for consumers.  Gas is sooo ridiculously cheap where I live the motivation to install a heat pump has more to do with how it can provide air conditioning in the summer as well as cheap heat in relatively warm conditions in winter (temps above -15C).  I have to come close to matching the cost of an air conditioner install for my market.

@old_scientist In my case I want to avoid any quick cycles on the HP, so in addition to using ambient temperature curves to control temperature, I plan to use an adjustable delta t thermostat for on/off limits.  I am fine with up to a 2C differential.  The human perception of temperature varies a lot anyway .. feel a bit cold? .. put on a sweater.  My wife and I have electric down “sweaters” .. and we use them inside as well as outside in the winter to fine tune how we feel .. they are absolutely awesome!

Posted by: @old_scientist

↑

From a physics pov, the room needs a steady 35C flow temp to maintain target temperature of 20.0C. If that steady 35C flow temperate is disrupted for ~30 mins then the result will be a fall in temperature. Maybe if the building has sufficient thermal mass then if will not be noticeable, but I can pretty much tell you when a defrost cycle occurs (or has recently occurred) by the drop in room temp displayed on the room thermostat (may only drop a few tenths of a degree, I really can't remember but I'd notice if it dropped to 19.8C when it's been sat at a constant 20.0C all evening)

OK but what I understood you were saying, and what I was struggling with because you have a large smoothing capacitor namely the house fabric, is that the cycling (as opposed to an overall drop in temperature) is detectable. 

I can see that not getting sufficient energy may cause it to cool overall (which could be compensated by increasing the WC curve a tad), its detection of the cycling due to defrost thats bothering me, given the large smoothing effect of the fabric of the house.

.

Posted by: @bobflux

↑

The "thermal comfort" we feel is a mix of air temperature and infrared radiation received by the body. This radiation comes from every surface in the room....

If you are in direct view of a large radiator heated to 35°C, radiation will be an important part of your feeling of comfort. If the heat pump sucks the heat out of the radiator, it will take a while before room air and surface temperatures drop, but radiation from the radiator itself drops immediately as it cools, and this is noticeable. It should work the same for other radiant heating sources like radiant ceilings or floors, if they have low thermal mass.

That's one reason why a boiler with hot flow temp and on/off regulation is less comfortable than a "low and slow" heat pump: if the radiator swings from room temp to 60°C then there will be a large variation in radiation output from the rads, making occupants too hot and too cold on every cycle, even if air temp doesn't change much.

Ah, thats an interesting observation.  I keep forgetting that, although radiators are primarily convectors, they are also radiators.  This explanation is consistent with the duration of defrost vs the thermal time constant of the house so perhaps is more plausible, although it wouldn't explain the change being detected by a thermostat (well maybe it could if the sensor faces the radiator).

This is an interesting topic.  My long term history is too granular to tell but when we get into defrost period in the forthcoming season Im going to do a stare and compare of defrost cycle vs measured room temperature.  

So far, mine hasn't defrosted yet, and the radiant ceiling was finished last month and isn't connected yet. I'll have to wait for next winter to see how defrosting feels... For the UFH loops I don't have any worries, the slabs are 10cm thick so there are tens of tons of concrete as thermal mass...