Performance figures after removing buffer tank?

CH_18 · 8 February 2026 00:18

Hi all,

First post here. I’ve got a 7kw Vaillant heat pump with a 100l 4 port buffer. Like many on the forum, reading about “max output capacity of the 7kw units” I find that for 95% of the time the unit performs well, albeit at a lower than expected COP @ 2.9 average. On the days where the outside temp drops below 2 degrees, the unit struggles to meet the target temperatures due to defrost cycling. My question is, how much (if any) improvement will I see if I repipe my buffer to a volumiser or remove it completely? I have enough system volume to remove. Also does anyone have any performance figures before and after removing the buffer? Thanks in advance

Timbones · 8 February 2026 00:38

Hello, and welcome to the forum.

I don’t have any direct experience myself, but here’s a few links that might give you some useful insights:

It may be easier to balance the two pumps instead…

CH_18 · 8 February 2026 00:46

Hi @Timbones , thanks for the reply, I’ll have a read through them. That’s another issue I have that could be negatively affecting efficiency. I have 1 x pump post buffer and then another at the UFH manifold. I did read that some guy removed his pump set and mixer at the manifold and saw an immediate improvement

Timbones · 8 February 2026 08:44

Can you tell us more about the configuration of your system? Is it all UFH, or are there some radiators too? Single zone, or multiple zones controlled by thermostats and actuators? What flow temperature does it typically run at?

When diagnosing heat pump issues, it really helps to chart the flow and return temperatures - do you have any monitoring in place, or only what Vaillant provide? It’s common for their temperature sensors to be a little off, and give inaccurate COP information.

Measuring the temperatures at each of the 4 ports of the buffer can help determine if the two pumps are adequately matched. Can use cheap digital thermometers like Mark did here using John’s spreadsheet to calculate how it affects power delivery. Could the do the same across the manifold.

dMb · 8 February 2026 12:26

Having a mixing valve at the UFH manifold - as is typical with UFH set-ups designed for operation with boilers - is a bigger issue than having a supplementary circulation pump at the manifold. Other people have reported that even when they’re ‘closed’ these mixing valves still allow some recirculation of the return from the UFH, which is never* desirable with a heat pump.

Having said that, if you already have a circulation pump downstream of your buffer tank then having a further one at the UFH manifold might also be excessive; it depends whether you also have radiators etc.

* unless you have a ‘High Temperature’ heat pump running at >50C flow temp, which would be hotter than some floor coverings over UFH would be happy with

arc12 · 9 February 2026 14:42

That COP looks a bit poor, so I would start by investigating the HP on/off behaviour, the flow temperatures at the 4 buffer pipes, check the various “delta T’s”, that the rads are all balanced, that the DHW isn’t somehow using the immersion heater etc… and not expect buffer elimination to be a quick fix.

The usual issue with buffers is “blending” - ie mixing hot water with cold return water before circulating in the emitters - but this is unlikely to cause you to be using a LWT more than a few degrees higher than if it were perfect. That will probably give a 10%-15% (ballpark guess) loss of COP.

greentangerine · 9 February 2026 19:42

I converted my four pipe buffer to three in order to mostly eliminate the blending issue. That worked in my situation.

CH_18 · 9 February 2026 21:21

Thanks all for the info

So I bought a infrared thermometer and have done a bit of a temperature survey and I think I may have bigger issues, I think the unit is massively undersized.

I have the Vaillant integration setup on home assistant so I’ve looked at the history in cold spells (less than 3 degrees) and noticed that my flow temperature has never exceeded 30 degrees. FYI my design flow temperature at -1.79 degrees should be 46 degrees. I’ve never got anywhere near that figure.

Fast forward to today, I have 2 circuits, UFH downstairs, Rads upstairs. I set the control in Man on both circuits and set the target room temp on the sensocomfort to 30 degrees to max the performance of the pump. For reference the outdoor temp was 10 degrees so not cold at all. As expected the target temp asked for 44 degrees and the compressor wound up to 100% over an hour of a half of testing.

From 16:08 to 17:25 the actual flow temp increased from 30.7 to 40 degrees, it never actually reached 44 degrees in over an hour time and went through 1 defrost in this time? Is this a normal timescale?

My DT across the heat pump (through the buffer) maintained approx 7 degrees most of the way during the test. Relatively normal?

Also my temperatures across the 4 ports of the buffer didn’t seem crazy inefficent, for example at 39.2 flow temp:

Buffer inlet from HP = 38.2
Buffer outlet to circuits = 36.8
Buffer inlet from circuits = 30
Buffer outlet return to HP = 30.5

I have stripped most of the insulation off atm to allow testing but don’t think this would make much difference?

I ran out of time for testing so will revisit tomorrow but does this sound like an undersized unit? The unit pulled a constant 3.2kw during the testing but looking back, when its cold it pulls 2kw max and then hits a defrost cycle, never getting above 30 degrees.

Do you think this could be a balancing issue/undersized unit or a software/setting limit that is in place?

Sorry for all the info and numbers they probably don’t make much sense, but any help of info would be great.

Timbones · 9 February 2026 21:49

What is the estimated heat loss of your property by your installer?
And what is your daily delivered heat on the coldest day?
Is your house warm enough, day to day?

Can you chart your flow and return temperatures over a typical 24 hour period?
Knowing the flow rate and heating power can also be useful, if you have it.

Did the room temperature increase during this time? The flow temperature is determined by the total radiator + underfloor area, and lower is generally more desirable. Not managing to reach 44 degrees is not necessarily a problem, so long as the required heat power is being delivered in the house.

The defrost is unexpected though, for a relatively mild outdoor temperature.

Yeah, not too bad. Might be costing you a few percent in performance, probably not worth messing with, IMHO.

Hard to say. What really matters is how much heat is being put into the house, and if that matches the heat loss at whatever the outdoor temperature is. Usual configuration is to set the weather compensation curve to increase/decrease the flow temperature as the weather gets colder/warmer - what is yours set to?

MBC · 11 February 2026 04:41

I tried to use an infra red thermometer on my pipes and found it to be too inconsistent - factors such as distance and angle to pipe affected the reading and so in the end I got a digital pipe thermometer with 2 clamps so it can show each pipe and display the delta T too.

glyn.hudson · 11 February 2026 14:17

This is where the inefficacy from a buffer tank comes from. You’re loosing 1.4C due to distortion (mixing) in the buffer tank. This means the heat pump needs to deliver 1.4C higher flow temperature compared to if it was directly connected to the system. 1.4C is not terrible, here’s an example of a recent remedial job that @johncantor was involved in:

The first graph shows the temperatures when the buffer was connected in the usual 4-pip arrangement, there’s a 4.5C drop in flow temperature due to distortion. After re-piping the buffer as a 3-pipe arrangement the distortion dropped to 0C, this allowed the heat pump to run 4.5C cooler flow temperature while delivering the same MWT into the emitters. John estimated this should improve efficiency by about 10%

johncantor · 12 February 2026 16:59

Sorry, not read much of the threads here, but i did convert a system with ‘bad mixing’ buffer at bottom of DHW cylinder.. managed to reduce the flow temp by 4deg. . I chise the 3rd option, and have a before and after emoncms graph

johncantor · 12 February 2026 17:01

CH_18 · 12 February 2026 22:10

Hi all, thanks for the replies.

Today I spent 5 hours testing the system and I still think the unit is under sized.

So, outside temp was around 9.5 degrees +/- 1 degree. I set up 6 temperature monitors on the system mainly around the buffer and asked for 30 degree on the sensocomfort. Target temp went to 44 degrees and compressor continued to ramp up throughout the test.

After approx 40mins the unit went through its first defrost. I’ll attach the excel spreadsheet to show figures, I would try and chart them but not exactly up to speed with excel.

From the start of the test to the defrost time was 40 mins and I lost approx 5 degrees temperature from the buffer during the defrost, which doesnt seem terrible?

The unit continued to ramp up targeting the 44 degrees over the next hour where increased from 28 degrees to 36 degrees. Seems slow?

The delta T between the buffer inlet and outlet (points 1 & 2 on the hot side) was negligible all the way up to 36 degrees, averaging less than 0.5 degree. Although I did see an increase in temperature across the buffer on the cold return side back to the heat pump, it was picking up approx 1 degree, which again doesn’t seem too bad, maybe a sensor out of calibration slightly. At this point I had the buffer pump on minimum and the UFH pump on the mid setting.

As the flow temp got closer to the target temp the compressor started to modulate down, at this point I increased the weather curve to MAX 4.0 from 0.4 and increased my max flow temp to 55. The unit then targeted the 55 degrees.

At this point I also did a temp survey across the UFH manifold (results attached) and found that the average delta T was 7.4 degrees (6.2 min, 9 max) which I (maybe incorrectly) thought was too high and the ideal was 5 degrees? So I sped up both the buffer pump and UFH manifold pump to max (to confirm all of the flow valves are already fully open.) After the speed ups the delta T across the UFH circuits decreased to 6.6 degree average (5 min, 7.9 max)

After I increased the pump speeds the Delta T across the circuits got better but caused the delta T across the buffer to worsen. On the original pump speeds I losing less than 0.5 degree across the buffer but this widened to 1.5 degrees.

I also did a survey across the rads which i’ll attach.

Sorry to throw a load of figures at everyone but I think it boils down to this, my design flow temp is 46 degree’s @ -1.79 outside temp. After testing the highest temp I reached was 40.5 degrees @ 9.5 outside temp, that seems well short. The flow temperature seemed to plateaux after 5 hours so I dont think it would ever reach 45 degrees @ 9.5 never mind -1.79!

Day to day the house is fine, it never feels cold. During cold spells (below 2 degree’s) the house loses temperature but again has never felt cold. My weather curve is 0.4.

I have attached a few spreadsheets/drawings to show my findings if anyone is interested and can send any advice my way. I really appreciate your help so far.

A few questions since doing the test:

Shall I stick with a better delta T across the UFH circuits and a worse delta T across the heat pump or revert the pump speeds back to before?
I discovered the average delta T across my rads was 2.7 degrees (min 1.0, max 5.3) what should I be aiming for on the rads? They are aluminium if that makes any difference.
Looking at one of the coldest days of the year so far on the vaillant app (outside temp 2 degrees) my heat generated for that day was 2.5kWh per hour. My calculated heat loss by the designer is 5.5kw, does that mean my heat pump is less than half the size of what I need on a cold day? To compare looking at todays heat generated I averaged 9.2 kwh per hour, but that is at 10 degrees outside temp.
On my sensocomfort I have 3 circuits. Circuit 1 UFH, Circuit 2 Rads, but I have no idea what circuit 3 is? Is my sensocomfort set up wrong?
The vaillant temperature sensors for UFH and rad circuits have a discrepancy of approx 3 degrees, they should be exactly the same temperature. Could this cause issues? Miscalculations?
What is my next step? Does it sound like an undersized unit? Are there any mods I could retrofit to help my situation? (i.e. back up heater) Or do I go down the route of getting the supplier to replace as it doesnt meet the criteria/figures they said it would?

Sorry about the essay, I’ve had a guts full of typing “delta” thats for sure

Any more help/info is greatly appreciated, thank you all

UFH_Delta_T_Comparison.pdf (47.3 KB)

UFH manifold temperature survey - Sheet1.pdf (30.7 KB)

Rads survey - Sheet1.pdf (33.3 KB)

ASHP testing.xlsx (193.3 KB)

CH_18 · 12 February 2026 22:16

Hi Mark, yeah I found the same, not great on copper pipes. I taped some black matte tape to the pipes and got a decent reading that way

I ended up buying 8 x digital probe thermometers for the testing just to be sure

Timbones · 13 February 2026 00:03

This is the key point for me. If the house can maintain temperature, then the heat pump must be sufficiently sized. It is the power generated that is important, rather than the temperature it can achieve. There’s a good chance that the installer’s design temperature is incorrect.

Heating power is delta T × flow rate × specific heat capacity of water, so the actual flow temperature doesn’t not matter.

If we guess the heat pump is running at maximum output of 7 kW, it will produce a delta T of 7° at a flow rate of 14.5 l/min. I don’t have Vaillant’s data sheet to hand, but that seems pretty normal. As the compressor modulates down to a lower power, the resulting delta T or flow rate will drop according to that formula.

A defrost with an outdoor temperature of 9.5°? That is unusual. One would only expect to see defrosts around 2°, depending on local humidity. Either the outside is frosting up from condensation, or the Vaillant thinks it is.

No, this means your house needs less than half of the calculated heat. Heat pump may actually be oversized.

The calculated heat loss by the designer is likely to be overestimated, so take that with a pinch of salt. Your heat pump has a rated output of 7 kW (with caveats), so on paper should be more than big enough. If you are struggling to get that heat into the house, then there may an issue with the system design, or potentially a fault with the unit.

Are you sure? 9.2 kW seems way too much for 10°.

CH_18 · 13 February 2026 00:24

Top man @Timbones , thanks again for the reply.

I’ve attached a screenshot of my generated heat for the test period. Discounting the “warming up” and “cooling down” periods, the middle 3 hours are 8.9kwh, 9.3kwh and 9.4kwh per hour, unless I’m reading that wrong. Outside temp did step up to 11 degrees throughout the test aswell

As you can see, it was putting the heat into the house fairly efficiently over this period.

Timbones · 13 February 2026 00:46

Oh, right - your peak heat output was 9.4 kW, with a significant rise in internal temperature. That seems as expected, and demonstrates plentiful heat delivery in mild weather.

CH_18 · 13 February 2026 00:57

@Timbones I’ve just had a look at my notes again and noticed that what I thought was a defrost cycle may have been something else. At the time it displayed “heating overrun” on the control unit, followed by “compressor blocked” then “pre run” and then “compressor active”.

I can’t remember seeing a defrost message?

I’ve just had a quick Google and it says “heating overrun” occurs when the heat demand has been met and stops the compressor, but at no point during the test was any of the target temperatures met? Any ideas?

Begall · 13 February 2026 12:20

So:

7kW heat pump installed, on a 5.5kW calculated heat loss
Manages to peak at 9.5kW output in mild conditions
Kept the house warm while averaging 2.5kW output on a day that averaged 2c

I don’t think there’s an issue here? You have a massive headroom to keep warm in freezing conditions, it’s really unlikely you need to get anywhere near a constant 7kW output at any point.