The debate over buffer tanks in heat pump systems has been a hot topic for years, but there’s another component that’s equally important yet often overlooked: the volumiser. While buffer tanks are designed to store and release heat energy, volumisers serve a different purpose: they store water volume to stabilise flow rates and prevent short-cycling. However, just like buffer tanks, it appears the placement of volumisers is a point of contention among installers and heat engineers.
Recently, I posed a question to my network of heat pump installers, engineers and experts: “Where do you place the volumiser: on the flow side return side?” The results were striking: 81% said the return side, 8% said the flow side, and 11% offered other suggestions, often with comments like “it’s needed at all if the system is correctly designed”.
While the majority opinion is clear, it raises an important question: Is this placement based on sound technical reasoning, or is it simply a case of following tradition?
What Is a Volumiser & What Does It Do?
A volumiser is a hydraulic component that stores water volume to stabilise flow rates in a heat pump system. Its primary role is to prevent short-cycling (frequent on/off cycling of the heat pump) and to ensure smooth operation during transient events, such as defrost cycles.
Unlike a buffer tank, which stores heat energy, a volumiser doesn’t store temperature. It stores volume. This distinction is crucial because it determines how the volumiser interacts with the system and where it should be placed for maximum effectiveness.
The Defrost Cycle: A Critical Consideration
To understand why volumiser placement matters, we need to look at what happens during a defrost cycle. When a heat pump defrosts, it temporarily reverses its operation to melt ice on the outdoor coil. This process introduces cold water into the heating circuit, which can disrupt comfort and system performance if not managed properly.
If the Volumiser is on the Return Side:
- Cold water from defrost flows directly into the return pipe, which is already at a lower temperature during heating mode.
- This cold water mixes with the cooler return water, further reducing the temperature entering the radiators.
- Result: Radiators can receive a sudden influx of colder water, causing potential cooling and discomfort.
If the Volumiser is on the Flow Side:
- Cold water from defrost enters the flow side, where it mixes with warmer water stored in the volumiser (heated by the heat pump during normal operation).
- The volumiser’s stored volume moderates the temperature of the cold influx before it reaches the radiators.
- Result: Radiators receive tempered water, minimising temperature shocks and maintaining comfort.
Thermodynamics and Energy Efficiency
The flow-side placement of the volumiser aligns with thermodynamic principles, particularly the First Law of Thermodynamics (conservation of energy). By moderating cold influxes during defrost, the flow-side volumiser reduces the energy required to reheat the system, improving overall efficiency.
On the return side, the volumiser’s stored energy does little to moderate the cold influx, as the return water is already cooler. This means the radiators receive colder water almost immediately, increasing the energy required to reheat the system and reducing efficiency.
Why the Majority Might Be Wrong
Given the technical advantages of flow-side placement, why do 81% of installers prefer the return side? The answer likely lies in tradition and training. Many installers are taught to place volumisers on the return side because it’s seen as the “standard” approach. But is that right?
What Homeowners Need to Know
As a homeowner, understanding the role and placement of volumisers can help you have more informed conversations with your installer. Here are a few key points to keep in mind:
- Ask About Volumiser Placement: If your installer plans to include a volumiser, ask where they intend to place it. If they say the return side, ask why and discuss the benefits of flow-side placement.
- Focus on Defrost Performance: A well-placed volumiser can improve comfort during defrost cycles by preventing cold shocks to the radiators.
- Consider System Efficiency: Flow-side placement can reduce energy consumption and improve overall system efficiency, saving you money in the long run.
Whilst I agree with the final conclusion (flow side placement appears to be best, perhaps with the exception that if the volumiser is outside the insulated envelope of the house there might be a loss argument for return side placement) I am not sure about the following:
Can you explain why that is the case please. The energy used for the actual defrost is (surely) the same either way and so its this amount of energy that must be replaced by the heat pump. So (surely) for the statement to be true it would need the heat pump, for some reason, to operate more efficiently during recovery in flow side placement than it would in return side placement. Why would that be?
@JamesPa the key point, from my own logical thought process (which could be flawed), is where and how the cold influx from defrost interacts with the system. When the volumiser is on the flow side, it contains warmer water from prior heating cycles. During defrost, this stored warmth moderates the cold influx, reducing the temperature drop in the water entering the radiators. This means the heat pump doesn’t have to work as hard to reheat the system after defrost, as the temperature difference it needs to overcome is smaller.
On the return side, the volumiser’s stored energy is less effective (in my opinion) because the return water is already cooler. Cold water from defrost mixes with this cooler water, creating a larger temperature drop inside the property. This forces the heat pump to work harder to reheat the system, consuming more energy in the process.
While the energy used for the actual defrost is indeed the same in both cases, the recovery phase is where I think the difference lies. By moderating the cold influx, the flow-side volumiser reduces the energy required to bring the system back to its target temperature, improving overall efficiency.
Not sure about this. The heat pump sees the water returning from the fins, which will be at a temperature determined by the defrost process not any temperature of water being sent to the radiators.
It certainly stores less energy. But its the same amount of energy that has to be made up (ie the energy needed for defrost).
agree!
Back to the first point, not sure!
probably not worth an extended argument since we agree on the conclusions!
@JamesPa Agreed. I think it’s a case of much over muchness, which is probably why the industry doesn’t pay too much attention to it. That said, when we eventually redo our piping to switch from a buffer to a volumiser, I’ll request it to be placed on the flow side. The reason I’ve posted this is because there have been questions from homeowners on this in the comments of our Youtube videos.
I’ve got a 7kW Arotherm+ with 50 L buffer tank (so no immediate slug of cold water going to the radiators) and see the recovery from defrosting as only taking a few minutes with the heat pump temporarily producing more heat than before the defrosting started. The output then drops again as the ice builds up. These temperatures were measured close to the buffer tank. The OAT was around -0.5C. The heat pump stopped producing heat for 6 minutes and the flow temperature was back to the defrosting temperature after a further 7 minutes. I’ve not been in the right place at the right time to observe if the circulating pump for the radiators is automatically stopped during defrosting but, at the normal heat pump flow rate the water through the heat pump during 6 minutes is 2.4 x the buffer tank volume.
The normal process is for the circulating pumps to continue running, as they need to circulate warm water from the radiators (or DHW) to the heat pump, where the heat pump operates ‘in reverse’, extracting heat from the warm water to warm itself and melt any ice that has formed. Once the heat has been extracted, this cooled/colder water is then continues it’s journey back into the property, hence the slug of cold water entering the radiators.
Playing devil’s advocate for a moment – it strikes me that everyone is looking at this from the viewpoint of protecting the heating system from that slug of cold water, not from the viewpoint of defrosting the heat pump which will require a certain volume of warm water. The purpose of the volumiser is to ensure that sufficient volume of warm water exists. If it is immediately diluted by the ingress of cold water, then the volume of warm water available for the defrost cycle is diminished. Or to put it another way – what is the primary purpose of the volumiser, or why did you install a volumiser? Was it to provide sufficient volume for effective defrost cycles, or was it to protect the heating system from the ingress of cold water following a defrost cycle? If it was the former, then that would suggest the return may be the better placement?
That’s a really good question. Ive been trying to find a clearer explanation of what specifically is the purpose or effect of increased volume in the heating circuit. For a while now I personally have seen that the amount of heat to defrost isn’t that great since it only takes around 4 minutes and this appears to be Bourne out by our system inbuilt temperature control which only tries to maintain 20c minimum water temperature through it’s freeze stat function. The recent discussions which you’ve been involved with have brought new other insights that it’s more about recovery time to get the flow and return back up to temperature to reinstate heat transfer through the emitters.
just for clarity, my post above shows our HP system which does not have a volumiser or a buffer tank which is why the recovery takes over 30 minutes. But the return temp only drops by 5c. But what I’m suggesting is that with repeated defrosts and a drop of 5c and a recovery of 30 minutes every hour this will have a disruptive effect on maintaining steady state heat transfer into the living space. This might be more pronounced in colder climates.
There may also be other ways of adding volume to help with defrosts which might be more effective than a volumiser, like perhaps simply adding more radiator volume in the first place. This would also have the added function of allowing systems to operate at lower flow temperatures making them more efficient.
Also it’s interesting to see how @JohnR and the buffer tank he has fitted reacts to a defrost.
It is indeed a good question.
The purpose is always said to be to ensure that there is sufficient water in the system to yield sufficient energy to effect the defrost. If there isnt sufficient energy available then the heat pump cant recover, and may either be damaged or (more hopefully) will shut itself down. Some heat pumps will activate a backup electric heater in this case, or indeed always for defrost.
There is a thread here where this behaviour is actually observed (and is a problem!) and the heat pump does indeed shut itself down, triggered by a return temperature of <15C. In this particular case the problem only occurs during DHW reheat following the drawing of a bath – which may be an argument for putting the volumiser in the flow before the diverter valve not afterwards return common return!
I am guessing that heat pump designers consider the something close to the worst case when recommending minimum system volume which would be a system operating at a particularly low flow temperature, eg a small UFH based system with dense UFH pipes. Obviously if the system is operating at a higher flow temperature, such as is likely with radiators, the energy stored in the water is correspondingly higher so less volume is required.
Obviously adding more radiator capacity is a ‘more’ efficient way to add more system volume, but radiator sizing is frequently limited by practicality!
Repeated defrosts surely do have an effect on average output, and some manufacturers (eg Mitsubishi) produce two tables of output vs OAT/FT, one including the effect of defrost and one excluding it. What assumptions they make about humidity and thus frequency of defrost, and other important parameters, is anyone’s guess.
@JamesPa
Hi James,
on this topic: you’ve given good advice to me in the past and I’m working through things with my heat pump set up with the help of my installer. There have been some design issues, but I’m slowly getting to grips with it.
Running my system totally open and constant has helped a lot. I’ve got a mix of UFH ground floor and rads upstairs. (Ecodan 11.2Kw). But the design temperatures were meant to be the same as we have no mixing valves, but they are not. I can run the UFH system with really quite low temps. when its around freezing an LWT of 33 keep those rooms to a lovely temp of 20. The bedrooms above are about 17, which I think I’ll just have to live with as I’m not putting more rads on the walls.
I was having big trouble with defrosts and recovery from them, which is a little better now I’ve stopped asking for LWT that my system can’t get to. But the recovery from defrosts is a struggle. My installer put in a bigger volume low loss header and zone values off them to try (I think) to stop the secondary circuits being open with defrost. But this has done nothing other than give me a small blending problem at the header than I didn’t have before. The secondary circuits keep pumping during defrosts. My heat loss is about 11k at -3. I think my system volume is about 150L. with my generally low flow temps, do you think there is a case for more system volume. On another forum, I’ve been advised that 300L of system volume may help defrost.
Or, defrost are just part an parcel of a heat pump, deal with it, and 150L is still plenty for an 11.2KW ecodan even at 30 degree flow temps.
Firstly thank you for the feedback, its always good to get.
The only way I can see you could balance the temperatures would be to throttle the UFH. This will push up the required LWT so will increase cost, but perhaps not by too much in return for the additional comfort. Do check that all your upstairs rads arent throttled (some may be but the majority probably dont need to be).
Defrost is part and parcel of a heat pump, so expect it. The issue arises if its taking too long and worse still if it doesnt complete, which will case the heat pump to lock up (In another thread locking up happens to someone when they have a bath and its cold outside). So if its not causing an insuperable problem, probably just leave it.
That said, what your installer has done is pretty much the exact opposite of what I would have done! I think the problem you have is that an LWT of 33 is quite low, and presumably its a bit lower still when the temp is above freezing. So the amount of energy available from the system to ‘fuel’ defrost is correspondingly low. I also suspect that much or part of it isnt reaching the heat pump during defrost because of the zone valves (please delete them or force them permanently open) and/or the LLH.
I can see two possible options if just forcing open the zone valves doesnt fix it:
1. Simplify the system – delete zone valves, secondary pump and LLH. This will ensure that the full system volume is available for defrost and also eliminate any mixing, improving COP
2. Add a volumiser on the heat pump side of the LLH to provide additional water for defrost
The first is obviously major surgery, but (unless there was a sound reason for the LLH in the first place) makes for a more efficient system where faults are easier to diagnose. Its just possible that, if you do this, you will still need a volumiser (repurpose the LLH for this!)
Hope that gives you some food for thought. There are some other options but they get more complex.
I should add that my system has two circulating pumps: One in the heat pump which circulates water to the buffer tank and DHW cylinder and one in the circuit to the radiators. Is there sufficient energy in a 10C drop in the 50 litre buffer tank temperature to melt the ice? If so, I would expect the pump in the radiator circuit to shut down until the buffer tank gets back up to temperature. I would need to buy two more of these temperature loggers to be able to monitor what’s happening in that circuit during the defrosting. However, it appears to me that the much-maligned buffer tank does have its uses.
@SUNandAIR my graphs look similar on my ecodan. And when you make it 2 defrost and hour…… you can see how they struggle
Further to the above I have just made some observations during a defrost cycle. My Arotherm+ kept the secondary (radiator) circuit pump running so the temperature in the flow to the radiators did drop by about half the drop in the temperature of the flow from the heat pump. The radiators, however, have sufficient thermal mass that they didn’t go cold.
There is absolutely nothing wrong with buffer tanks in principle. The problem lies with the execution which all too frequently, it seems, involves poorly balanced primary/secondary pumps resulting in excessive mixing and thus a reduction in overall system efficiency. They also make diagnosis of some problems more difficult.
A volumiser will do essentially the same job for defrost or reducing the frequency of cycling and cant suffer these problems. There are few if any domestic situations where the hydronic separation provided by a buffer tank (but not by a volumiser) is necessary. Hence the argument against buffer tanks in practice, given the wholly avoidable risk to the quality of the installation.
The irony is that those installers who are least likely to fit buffer tanks are most probably the very same installers who, if they fitted them at all, would fit them properly (and vice versa).
Can I ask a very simple question? The answer may be in this discussion somewhere so apologies if I have missed it…
When the heat pump goes into a defrost cycle, is the flow of water around the house actually reversed? Or is it just the heat pump refrigerant flow that is reversed?
The article states that “if the Volumiser is on the Return Side … Cold water from defrost flows directly into the return pipe". But is this the case? Only if the direction of flow of the entire system is reversed.
Surely the heat pump itself will send flow the same direction whether the volumiser is on the flow or the return? Or am I missing something?
The flow doesn’t get reversed but just keeps circulating without being heated at the heat pump. Consequently there’s a drop in flow temperature as the water passes through the heat pump (to melt the ice) instead of a rise which results in the flow going into the house being colder than the flow going out of the house.
As I thought. So isn’t the line “cold water from defrost flows directly into the return pipe" incorrect?
Not trying to be pedantic. Just trying to get my head around the whole subject. But this discussion on volumisers has confused me a little because it suggests that the heat pump will reverse the direction of flow of the whole system.
This whole “hot water from the primary is used in the defrost" is confusing me. Unless it is not general across all makes the Ecodan reverses the flow of hot refrigerant gas to defrost the evaporator after stopping the fan and primary flow pump.
Are you sure. Where does the defrost energy come from on that case?
Most stop the fan but leave the primary pump going thus extracting the defrost energy from the house. There are pumps that can start an inline electric heater and I think someone posted about one that has a built in water tank (not sure exactly how it’s used).
The energy must come from somewhere.
Well according to Mitsubishi’s technician in the video it is the hot refrigerant gas being reversed via a reversing valve and expansion chamber. No need to steal heat from the primary circuit or any volumiser. There may well be brands that do but that seems a very low tech solution compared to using the hot refrigerant gas in the system.
Sofaik all others do the same, ie they operate the refrigerant circuit in cooling mode to effect defrost. They thus cool the house and warm the heat exchanger using heat extracted from the house.
I think it’s highly likely mitsubishi don’t switch off the water pump because the water is the source of the energy.
@JamesPa
My mitshubishi does not turn off any pumps flows continue and takes some heat from the house. I’m fairly sure this is normal behaviour
“I think it’s highly likely mitsubishi don’t switch off the water pump because the water is the source of the energy." So the official Mitsubishi video is wrong? He clearly says that they use the already compressed hot refrigerant to defrost along with a tiny bit of heat from the HE at the end of the process, not the water from the primary as the heat source. Surely that is where the energy comes from.
On my Ecodan 14kW the primary pump stops for the 2/3 minutes it takes to clear the ice from the evaporator. The entire process from start to finish is over so quickly, and infrequently, that if it wasn’t for the vapour cloud and water running off you can easily miss it.
Now I am really confused!
@Abernyte
Post a picture of your flow and return, you will see they swap over for a short period during the defrost.
the video does say in takes a tiny bit of heat from the water.
the picture shows it stole a tiny bit of heat from the house to sort itself out.
Is that not just indicating that the primary pump has stopped momentarily while the unit defrosts using the hot refrigerant gas?
No, that wouldn’t account for the flow temp being lower than the return temp. This shows that the water pump continues to operate and as a result the water output from the heat pump is chilled by the fact that it has given up energy to the refrigerant and ultimately to the fins. Its quite clear from this plot that the Mitsubishi does defrost in exactly the same way as almost all other heat pumps.
@Abernyte no, it doesn’t turn the pumps off on mine.
My Mitsubishi does exactly the same as David’s above, there wouldn’t be enough volume of water/heat in the primary circuit alone to defrost the heat pump my run is very short.
Thank you gentlemen/people (can’t be too careful now) perfectly reasoned arguments that fit the observed evidence. The only fly in this ointment is why the manufacturer does not mention this in it’s otherwise clear and technical explanation of defrost. Taking a “tiny bit of heat from the HE at the end of the process" is the only mention. One would think if they were setting out to make such a video then the simplest route would be to say “we are briefly going to steal some heat from the house or DHW to defrost the HP" But they don’t.
Sigh….split jury, should never have studied law!
@abernyte – you really can’t be too careful…. alas there is possibly another complication in the evidence….. @davidnolan22 has a Low Loss Header fitted and he has been discussing poor recovery after defrosts (presumably the poor recovery is of room temps via the heating circuits and isn’t a failure of the heat pump)
A Low Loss header is entirely capable of masking the size of the cooldown during a defrost if the flow rates are not balanced on both sides of the LLH.
If the exit flow to his rad circuit and his 2 UFH circuits does not match the flow rate of his primary pump then it is entirely possible that the heat pump just sees a smaller cool down temperature than is actually happening. A small distortion might not be a problem But it mightn’t be worth investigating.
If the chilled water from the defrost was sent in to the secondary circuit (radiators and UFH then there could be SLIGHTLY hotter water sent straight back to the return thermistor to give a false reading.
This graph shows our heat pump during a defrost and you can see quite a high temperature drop compared to David’s graph
This graph is very clear and our system has no LLH or buffer or separate circuits. In other words there is no opportunity for distortion (or mixing)
here is the same graph where ive tried to sketch the DT as an overlay to show what the flow Andy return temps would have been like had ther not been a defrost. Hopefully you can better see the true drops in temps for both flow and return.
As was said earlier these defrosts could be repeat events every hour which could significantly upset the heating continuity of the home.
ive posted David’s graph below for comparison and I hope it can be seen the temp drop does not look that severe, when in reality it is likely to be just as large a temp drop as my defrosts.
So @davidnolan22 this was one thought I had and perhaps you’ve done it. Regarding recovery after multiple defrosts- Have you checked with thermostats what the temperatures are on all your connection ports on your Low Loss Header? This might give you an indication where the dominant circuit is or if the circuits are well balanced.
@SUNandAIR
Hi, ive read that a few times, I think I get what you mean
Yes, was definitely struggling to heat the house this winter at several points. But, I’m not sure if it was the defrosts that were the problem are the system design and how I was running it. My system was “designed" to run closed loop with fixed flow of 45 degrees. But my house would drop in temperature intolerably when it got close to freezing. When all the zones were calling for heat and the fixed flow was set to 45, it struggled like mad. The house dropped and I could never get it back.
But having worked through the “design" it seems that my emitter power was quite different in each zone. Overall I have 17Kw of emitter power at DT25 with an 11.2Kw heat pump. When all the emitters were open at the same and I was calling for 45 degrees, it could never get close, it sent the compressor to 120hz constantly and defrosted every 20 mins. And as I was running zones, the fabric of the house never got properly warm and the rooms were constantly going up and down in temp, and if the house got stuck too cold in minus weather, I could never get it back.
As for defrosting. On the old set up: all the systems flows kept running with defrost.
I then changed how I ran the system, all stats set to max and opened it all up all the time, only called got 28-34 degrees, but ran it constantly, and the house heated a lot better and held its heat. Ive only had a few minus days or periods since, but its been no problems with those temps with flows of 32 to 35, when its 3-5 degrees I’m running 26-28 degrees, no major cycling issues. With calling for lower flow temps it did defrost less and settled the whole system down, ran much calmer and less noisy and used less power.
However, the installer made some changes to try to help defrosts. but we, and I include the installer who designed everything, we’re still trying to get me to 45 degrees. The changes they made were to swap the initial 35mm piped header for a larger volume low loss header and zone valve off this to close the secondary flows if possible with defrosts. This was (I think) to increase the volume in the primary circuit so the pump had more energy to defrost and not need to use the secondary flows.
I spoke to Mitsubishi, who said that the secondary flows should close and the pump should be OK to defrost of the primary loop as it only needs 9 litres or so to do this in our climate.
Anyhow, after the changes nothing has changed other than a small blending problem that has irritated me more than ruined the system. But now I think what’s happening is that as I’m always running very cool water, the pump needs more energy than is currently in the primary circuit to defrost and has to keep the house flows open.
So in summary, I still don’t full know if the Mitsubishi heat pump is meant to defrost on the primary loop if its got sufficient energy or not. I need to wait until next winter to full tests my system to see if the deforesting issues I had re surface, or I can ride through them. If I can’t then I’m going to have to come up with a plan, a 3 piped buffer, or volumiser placed somewhere.
Do you think your drops look so much as your running at 40 in the picture and I’m running at 30. And, you retuns drop, but I’ve got 170l system volume and I get about on lap of the full circuit in a defrost so the return water does not drop until after the defrost finishes and its stole some from the UFH mainly. I think I’ve got this system pretty well balanced across the header now… Wow… long post. Time for a beer.
… Isn’t great advertising! Or is that too cynical?
Good call on the need for a beer.. that’s exactly what I did.
As you say, your system appears to have plenty of emitter capacity for your HP size, if you’ve got 17kwh capacity at DT25. 17kw capacity becomes 4.98kwh output capacity at 32c flow temperature which is more than the minimum operating output for the heat pump model you have. So even when your flow temp is set to 33c there shouldn’t be any significant cycling.
So the issue of recovery after/during several defrosts may need to wait for a cold spell before you are able to test. But the issue of volume may be a factor. The latest thinking on design sizing of volume seems to be requiring over 20 ltrs volume per 1 KWsize of the heat pump. So for 11.2 model that would be 11.2 X 20. That would make for a 224ltr design volume.
Under floor heating is perfect for slow release of heat but they are not very good at adding volume.
I noticed on the sketch you sent earlier you had a circuit volume of 150 litres does this include the volume of the new, BIGGER LLH? What exactly is the volume of the new LLH. Reason for asking is that LLHs tend to be fairly small in volume eg 2 to 15 litres. (If it’s not very big you wouldn’t really benefit by converting it to a volumiser)
So if you’re not thinking of extending the radiators any time and why should you, then you might benefit from adding 50 or more litre volumiser * if the recovery problem persists.* This would give you a surplus supply of hot flow-temperature water.
I also noticed on the sketch that there wasn’t a primary pump before the buffer I guess there is one since the ecodan doesn’t have an internal circulator like other makes.
regarding placement of the volumiser if you needed one… (and since we are on the thread talking about correct placement of a volumiser)
I still think it should be placed on the Flow side after the DHW valve. That would be in position C on the attached sketch.
Special note: And since the tank would be constantly carrying high water temp it very much should be placed within the insulated envelope of the home. (Even insulated tanks lose heat over time.)
reason being the chilled defrost water has a chance to mix with the flow temp water in the tank before entering the emitter circuits.
Please feel free to discuss this hotly debated topic further….
In all this discussion, sorry I forgot to ask do you still run it in fixed flow temp or are you now working in Weather Compensation mode. Do you ever bother with Auto Adapt?
One key point I think is quite important is that LLHs and buffer tanks rely on balancing the flows entering and exiting the tank but they introduce inefficiencies if there is a blending or mixing problem. The other issue is you have no knowledge of the flow rate through the various other zones since the flow meter is installed on the primary side of the circuits.
They are supposedly rarely needed but obviously serve an important role where they are needed.
A volumiser on the other hand actually relies on blending and mixing when the situation of chilled water needs to be warmed before it enters the under floor heating circuit. I believe that radiators cope with recovery better than high mass underfloor heating.
HI, yes, I forgot to add the primary pump! its on the return from LLH
I never bother with Auto adapt, I run the system all open, but run fixed flow where I control the flow temp, I’m looking to change this by next winter, but I just wanted to make sure my system was working. So, I’d look at the weather over night, set the flow manually and the change it in the day as it warms. This has given me a good idea about what temps do what at what weather. The weather comp on the R32 I struggle with, as my system tops out at 35-36 degrees, I can’t get it higher. And the difference between a flow of 32 and 34 is really quite a bit. The temp probe on the back of the unit gets colder during a cycle due to it cooling as the unit does and this drives the temp up 2-3 which on my system is a really big change.
Possibly 40c vs 30c flow has something to do with it…. Inevitably temp is what melts quicker. But what the graphs seem to show is both our Flow temps (LWT) appear to drop to similar degC fall 12c for yours and 15c for mine. The flow temps also both take about the same length of time to recover:- 16 minutes for your lwt recovery and 16 minutes for my lwt recovery. What is really different is that your RWT hardly drops more than 2c whereas mine drops 5c and takes 26 minutes to recover to its former 35c return temperature.
I have similar volume to yours but no UFH and no LLH. So yes perhaps your linear under floor heating means the feedback in temperature drop might be a lot longer.
And as you say, this might be robbing your heated slabs of some of the carefully built up heat your system had created. So is it not a fair question that this defrost-pause in heating followed by the “robbery” of heat with each cold influx from the defrost is what’s stopping your systems recovery to your target room temperature?just to be clear what now happens if you run your heating system at a LWT of 45c or maybe even 35c? Perhaps a WEATHER CURVE that is higher by only a few degrees might be all that you need to regain the room temps you’re after.
@SUNandAIR
I can’t get my system higher than 35 or 36 degrees, The only way to do that would be to start to close more of the secondary system off.
@davidnolan22 assume you have the flow rate set at 5, my system is the same I can’t get it above 42C when its freezing outside, if I close off some loops it can get there, its just under sized I probably should have an 11.2 rather than an 8.5, when I did heat loss on heat punk it comes out as nearly 10kw but was surveyed at 8kw.
@Gary
Its not the integral tank, so I have a separate pump.
My data from the second 1/2 of the winter is that 11.2 should be OK to cover this house.
I can not get my water above 36 degrees if its 10 degrees outside, let alone freezeing. The power gets lost in the 17kw of emitter capacity. The installer shoulder have spaced out the pipes a bit more in parts of the UFH.
@Gary @davidnolan22
I dont wish to go back over material you may already have covered, but its perhaps worth observing that another forum member (@cathodeRay) for several years couldn’t work out why his house wouldn’t warm even though all the indications were that the heat pump was oversized not undersized. Its not clear that its fully resolved but is looking a lot like it was a flow rate mismatch at the PHE used as system separation, resolved eventually by opening up some valves.
If you have any form of system separation its not inconceivable that you have a similar or related problem. Just saying but please accept my apologies if you have already been through all of this.
Hi James, we do have separation by the very small pipe I have placed pictures of before on the preplumbed cylinder that has no appreciable volume.
All I know is that if everything is open on the secondary side and its low single figures outside and my flow is set to 45C the heat pump can’t get there, compressor will be at 100Hz and the flow temp will get to 42C the return is at 32C and it just stays there till it defrosts so the heat is going into the house and the house temp is stable, hence, my conclusion that its just the heat pump that is undersized.
I could pay have it all replumbed to be completely open loop, but I think I saw someone get urban plumber do that for a Mitsubishi system and it made absolutely no difference to performance.
All noted.
it was @BobTSkutter that finally diagnosed the problem (which related to mismatch of pump speed causing throttling at the system separation point) with the @cathodeRay installation. Maybe if you post some flow rates, temperatures etc either side of the ‘very small pipe’ (a low loss header I suspect) he or someone else may be able to rule it out/in.
er, you cant have too much emitter capacity!
You can have an imbalance where some parts of the house heat more than others but never too much.
If the pipes are too dense (but not unbalanced) you can just operate at a lower flow temperature and save some money.
@JamesPa
yep, and this is what I’m doing
it’s the eveness of the spec. If your going to oversize, do it evenly. BUT, and it’s a big BUT, having spent many hours and weeks playing with things, I’ve think I’ve got this in a place that a). can live with and b) is working well. But in very cold weather, the bedrooms still run 2-3 degrees cooler than the main house. But for the 2-3 weeks of the year that’s its very cold, I think I can live with that.
Many people regard that as a requirement!
@JamesPa yeah, I agreed. My son’s bedroom is on the ground floor with his sisters room above. I need to keep his room all night to over 20 so the fabric of that part of the hose stays high and heats the rooms above.
but very first world issues. I’d like the option to run the main part of the house bit cooler, but I’d done with fiddling for another year. It’s working pretty good now. If I can hold it through heavy defrost conditions next winter, then I’ll stand down….
but, I’ve had work out the control mechanism and balance 5 pumps, 18 UFh loops and 6 rads myself. The 9 zone, 9 stats, closed loop system it was commissioned at did not work. There is no way a typical punter would have got here. If the industry is serious about mass roll out, I’m dubious.
I have yet to hear of a system that actually did work efficiently post installer commission. In fact I cant see how it can – no installer is going to spend the several days of elapsed time required to get WC curve and emitter balance right, and installations occur outside the heating season.
The best we can currently expect in practice is that the system heats and is tolerably efficient, but realistically nothing more. Unfortunately there appears to be a conspiracy of silence over this problem. However this is nothing new. Most, if not almost all, condensing boilers installed since they became mandatory appear to be set up so they dont condense or rarely condense. The fact the UK didnt mandate weather compensation in the UK and operate typically at ~FT75 pretty much guarantees this. So we have all been paying perhaps 10% more for our gas heating than we needed to. Plus ça change
Personally I don’t believe it is. Much easier to do a small number of high price jobs than a large number of low price ones. BG and Octopus may be different however.
@JamesPa
yeah, i agree. The time required to fine tune a system is not economical. I absolutely understand why heat pump companies oversize, stat control and get out of town. It breaks down, as mine did, when it did not heat the house. I think if I’d have had the 14kw unit, and I consolidated a few of the zones that I initially did, it would have powered through the problems and “worked”. So in a way, I think I’ll come out of this is a better place.
@BobTSkutter – you are welcome to rant. You are able to express yourself (and complex ideas) very well, and your posts are valuable contributions to the forum.
@BobTSkutter Hi Bob,
This is the ‘LLH’ on the preplumbed cylinder to separate the primary and secondary sides. (this isn’t my tank).
Flow rates on the primary side are 15L/min which is the recommended flow for the heat pump. On the secondary side there are 2 zones and associated secondary pumps for rads and UFH. No way of knowing the flow rates on the secondary have the 1,2,3 options on the pumps.
The issue is that the heat pump cannot get to the set flow temperature during the coldest periods similar to @davidnolan22 experience from what I have read.
Here is a plot from early Feb I have chosen a day when defrosts weren’t compounding the issue. Heating starts at 01.00 it took 2 hours for the heat pump to reach the set flow temp of 45C. The return temp is 10C lower than the flow. When there are defrosts hourly then it will never reach the set flow temp.
This isn’t because the house was cold the heating was running all day and only stopped at 23.30 to run the DHW for 90 mins prior to return to heating.
This is the reason I believe the heat pump is undersized.
If there are experiments I can do with the pump speeds happy to do that.
Looking at the plot it looks like the compressor frequency (yellow line) is zero from about 0L00 to 19:00. Am I misreading?
2 Hours to get up to temperature is not really a surprise (my house takes over 24hours to get up to temperature), particularly with UFH if the slab has been allowed to cool. Have you go a plot when defrost is compounding the issue.
@JamesPa
Your not misreading but that is after the event, if I posted the plot before midnight it would show its heating for 5 hours before hand
The house is up to temperature the floor won’t be cold it was running for hours before DHW at 35C, just trying to maintain it/batch heat during the coldest times overnight.
@davidnolan22 The UFH is retrofit so its only 20mm thick 150mm centres about 90sqm and then 8 rads upstairs.
Plot when it defrosts
@Gary how much UFH and how much rads are you? Whats your screed depth and pipe size?
Thanks for the additional information.
As I say 2 hours to get up to temperature after a long shut off is not totally surprising (albeit that the LLH may well be contributing – we may circle back to that later). However I would initially focus on performance when defrost is occurring which, as I understand it, is when your house is not warming sufficiently (please correct me if I have misunderstood your problem).
Looking at the plot most recently posted, it seems that the system keeps up initially (I can be certain about this because I dont know the target flow temperature, but it stabilises so I presume this is at or around the target), but then there are again extended periods when the compressor frequency is zero and the flow temperature thereafter is a bit all over the place. Why are there extended periods when the FT is zero and what is happening to the room temperature (which I think is the problem) in this period and the OAT.
Im basically trying to find the evidence in the plots of the problem you refer to. If you hadn’t said there was a problem the plots appear fairly normal except for the unexplained periods when the compressor is off for an extended time.
Hi,
I’m saying this from my experience, not as a heating expert:
Your compressor is turned off for too long periods. You may think the UFH is at temp, but low down its cooling. and when its OAT is cold, needs constant flow going through it. I would turn down the flow temp and run it constantly. DT 10 the unit straining like mad to heat the screed. I too though my unit was undersized, but after I turned to flow down and ran it constant, I could keep the house topped up with heat for 8kw an hour produced at freezing (heat loss 11) and the pump had lots more in the tank.
My graphs looked just lie yours when was aiming for a temp that the screed could not get to.
In mild weather, I tend to batch heat a bit more until I get a weather comp that I’m happy with
at DT10, your little unit is putting in over 10-11kw of energy, then is idle or just ticking over for large parts in the day. You need to even this out more
Can I just say that this is what I think too, but I was waiting for @Gary to answer the questions before being certain about the conclusion.
Heat pumps need to be on 24*7 during the height of the heating season, assuming that they have been correctly sized. The FT needs to be turned down as low as possible (to be done with thermostats, TRVs etc set well above desired temperature). In the shoulder season, or with a well oversized heat pump, batch heating may be possible and may work out cheaper, although there is no guarantee.
Operated this way its reasonable to expect the running cost to be similar to gas or oil, or a little cheaper. If you qualify for OvOs 15p tariff then definitely much cheaper, particularly if you have UFH and so can operate at a max FT of 35C or below.
As a concrete example I am operating with radiators and a max temp of 42C, the heat pump is working out ~12% cheaper than gas. Im buying electricity for it at an average of 19p by using EON Next Drive which gives 7hrs cheap rate at night in return for a 10% daytime price hike. I operate 24*7 without setback or set forward, so I’m not exploiting the overnight cheap rate to any great extent.
I did think about trying that this year just need to convince my financial voice to try it, as it will be significantly more expensive, it will pull 25% less energy but at 4 times the price
@Gary
I get where you’re coming from. But I don’t think you’re undersized . You’re trying to cram in all the heat your house needs into certain periods of the day. This is a perfectly reasonable thing to do, but with the nature of heat pumps, you’ll defrost a lot.
Hello, I confused about why this is a problem for you. Is your house not getting warm enough?
Thanks Bob
Correct it’s designed to 45 at -2C and can’t get to 45C house isn’t cold but 20 rather than 21C
@Gary mine was “designed” to 45, bit system can’t get above 36. That bit was nonsense.
from looking at your graphs of one day: you’ve got loads of scope to get more energy into your house, if you want/need to, by changing your control strategy.
Yes the examples I gave weren’t when it was at its coldest that week when it didn’t get above zero and dropped to -4 overnight I don’t have data for then but it was running 24/7
@Gary some the trends you’ve posted show the LWT is at the setpoint. They also show the setpoint is changing and they show the compressor frequency at zero (not running). This makes me think you’re running with a control scheme that’s changing the LWT to meet a room temperature demand and then switching off when the demand is met. Do you have an EcoDan installed with room temperature auto adapt?
Bob
No not auto adapt running hotter at cheaper times cooler at expensive times. Off is when demand is met on milder days. System won’t run below 32c without cycling