As I said in the introduction thread my partner and I are planning to extend our Victorian semi. We are wondering whether it would also be a good time to replace our 14-year-old combi boiler with a heat pump system. We have planning permission but yet to complete detail drawings or engage builders, so we hope the right time to be thinking about making sure the building could work with low temperature heating. I hope this post hasn't got too long but there is so much to get straight. Any advice or guidance is welcome!
The house is currently about 120 square metres with the extension planned to take us up to around 160. Typical layout with a wing extending out the back of the main house. The original house had solid walls. The kitchen wing has been previously extended, in the early 90s so the kitchen and the bathroom above have a mix of solid walls and insulated cavity. The extension will fill in the internal corner between the original house and the kitchen and add a two storey extension on the gable end, so will cover some of the solid walls, but some will remain and several chimneys, so I think the heat load will be quite high. I need to make the plunge and get some installers in to do calculations on this: hopefully working from plans will not be an obstacle.
Looking at the particular features of the house and our plans leaves me with a number of questions.
I am wondering what to do with the suspended wooden floors downstairs and whether underfloor heating is a sensible choice for these rooms.
The back room, which will end up being part of the open plan kitchen-diner and having no external walls has a suspended floor over a shallow void. The boards are, I think, original and we would like to retain. Could I lift the boards, insulate, fit UFH and re-fit? What other options would I have?
The front room and hall are over a cellar. Front room currently insulated with kingspan boards, the hallway just has a lath and plaster ceiling in the cellar. I've seen some UFH spreader plate systems that are said to be suitable for fitting from below. Might they be suitable for these two rooms? Is there anybody on the forums that has experience with these sorts of systems, good or bad?
The kitchen has a direct-to-earth floor. We will be extending this room significantly, so I think we want to go for a screeded under floor system over insulation here. Anything in particular to be aware of with the existing areas? I guess we would ask our builders to excavate out to get the depth for insulation and the screed with the UFH pipes.
For new rooms upstairs, what are the relative merits of underfloor heating against fitting radiators?
In terms of plumbing layout, I am thinking that the plant room could be in our new utility room with the heat pump itself standing out the back of this. It would be below a window: does that lead to any particular considerations? We will be too close to the neighbours to put the heat pump at the side of the house. We have quite a large bathroom so I think I would site the hot water cylinder in there. That would be upstairs from the plant room but with reasonably straightforward pipe runs. Cylinders for heat pump systems look huge: are structural calculations required for the joists that will be supporting the cylinder?
When you are renovating is always a good time, as it makes any installation much easier and you can plan properly. In terms of the detail I can only comment on two aspects of this:
UFH vs Radiators: UFH takes up no space and can generally be run at a lower temperature than radiators (typically 30-35 rather than 45 or sometimes even more). This significantly improves the efficiency of an ASHP. With radiators run at a design temperature of 45 you should see about the same running cost as gas, maybe a bit lower with UFH run at a design temperature of 30-35 it is 20-30% cheaper
Heat pump cylinders: because we typically store water at 48C rather than 60C heat pump cylinders do tend to be a little larger than cylinders for boiler based systems, however its also the case that they tend to be designed for a more luxurious lifestyle (just because time and expectations have moved on). 200l is quite common, 300l not uncommon, it depends obviously on your usage (you get people who want 6 off 10 minute showers at 20l/min in quick succession!). I haven't heard of structural calculations being done but I suppose it may be necessary in some cases and certainly if the cylinder is going mid span. Domestic flooring is I believed designed for ~150kgf per sq m, a 200l tank weighs 200kg, so would exceed the loading. I have a 200l cylinder and no calcs were done, but it replaced an existing cylinder the joists underneath are 200mm and the span at this point is only about 1m, so it was obviously going to be OK.
I dont know anything about the detail of alternative UFH arrangements, hopefully someone else will.
There is a good guide to ASHPs here which you may want to read. Note particularly the parts that say to avoid external controls, in particular 'smart' controls which are generally very bad for ASHPs. With 120sq m you are looking at one zone (max 2 - one up and one down) directly connected to the ASHP run, almost certainly, fully open loop on weather compensation.
Hoe that helps, feel free to ask as many questions as you want.
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
I've got a couple of leads courtesy of Mars, and requested design consultation via the Heat Geek web portal, so I'll see What some experts say about underfloor heating options and surely generate some more questions.
It has been a long time since I've posted. In the mean time a long grind of gathering quotations. I think we are ready to get under contract with a builder and we are pretty close on the heat pump too.
So, to a question. We have two heat pump quotes received. One from our local HeatGeek via the HeatGeek service, one from a trusted installer that @editor recommended. Both recommending 12 kW heat pumps. That's big but, looking at smart meter returns from the coldest day this year year, I can see my current gas system topped out at about 10.5 kW, of gas consumption rather than heating power, presumably limited by the radiators, and ran nearly all day non-stop. So, although difficult to correct for all the changes we are planning, I can see why this is a reasonable rating. Also backed up by the SAP assessment we had done as part of the building regulations approval pack. Both similar in cost, advice given, estimated SCOP.
The tricky bit. The local HeatGeek recommended a Vaillant heat pump, the other installer a Samsung one. Any experience on here to offer? Vaillant seems to get the better write up - in particular I see a couple of threads on here that are quite critical of Samsung. On that basis, I'd probably go with the Vaillant solution. The rub is very different shape factors. The Vaillant is twin fan, tall and narrow. I can't see where we could put it without blocking a window. The Samsung is a single fan unit, short in height but wide and deep. Aesthetically I can see the Samsung pump fitting into the overall scheme a lot better. Of course, that can only be a consideration if it is likely to perform well for us.
I have a he 7kW Vaillant which is good. My comment would be there is, sofaik (based admitedly on what others say and the specs) nothing particularly wrong with the Samsung that a better UI wouldn't fix. So get homely or havewise as well.
Incidentally the 11.2kW Mitsubishi has a similar height to the Samsung and is narrower, worth considering.
If either installer is proposing a 3 or 4 port buffer tank, plate heat exchanger or low loss header between hp and emitters tell them to remove or look elsewhere. I doubt the heat geek one will but Joule, who seem to be agents for Samsung, frequently do.
I'm a little concerned about your sizing however, 12kW is quite big for 160sq m semi. When you say 10.5kW all day do you mean 22- 24 hours If not then you are likely oversizing. What your annual consumption and what fabric improvements need to be factored in? Can you get a seasons worth of smart meter data so you can do a more rigorous analysis?
This post was modified 2 weeks ago 8 times by JamesPa
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
Interesting. I wouldn't be worrying about form factors if a 7 kW unit was proposed. They are a lot smaller.
An example cold day would be this (I hope that you'll be able to view that bigger than I can see in the message editor). About 10 kW for about 11 hours, then a bit of a break before it fired up again. The current system has a bit of a primitive programmable room thermostat with a 0.6 C hysteresis. I guess you are right that this is a lot less energy than averaging 10 kW for 24 hours. I think it's fairly typical of peak winter.
I've got usage data back to 2015, when we moved in. Peak in this graph is somewhere around 120 kWh per day (so ~ 5kW), but calculated from monthly meter readings, which will tend to smooth things out a bit. We've been on Octopus for about 17 months so higher fidelity data available from their API and would cover at least a year.
In the extension, we are adding a lot of floor area but that does come with some improvements to fabric that I hope will be significant. Insulating the floor of the back room that is currently uninsulated suspended timber, and that in the kitchen (direct-to-earth concrete), and to some extent wrapping the old solid-brick-walled building with a new one with modern insulated cavities.
It's a bit daunting to get an expert in to size your system and then tell them they've done it wrong and to knock 40% off the unit rating but I've seen from discussions here and elsewhere how important right-sizing is. It would probably be worth going back with some decent data to see what they say.
Peak in this graph is somewhere around 120 kWh per day (so ~ 5kW), but calculated from monthly meter readings, which will tend to smooth things out a bit
This smooths things out a lot, basically its useless.
I guess you are right that this is a lot less energy than averaging 10 kW for 24 hours. I think it's fairly typical of peak winter.
Thats nowhere near an average 10kW
It looks like you switch heating off at night which of course does reduce consumptio0n if the house cools substantially. So you may not get an accurate read out from gas consumption and/or may have to make some assumptions. How much does the house cool?
It's a bit daunting to get an expert in to size your system and then tell them they've done it wrong and to knock 40% off the unit rating but I've seen from discussions here and elsewhere how important right-sizing is. It would probably be worth going back with some decent data to see what they say.
Unless they have done a full survey they haven't yet sized the system, just done a rough estimate for quoting purposes. Even if they have, if they are wrong they are wrong, the task is to find out if they are wrong or not.
Attached is what I did with 2 seasons worth of half hourly readings (most of the 'experts' including the two who did a full 3 hour survey. were telling me I needed 16kW). I was running the boiler as close to heat pump mode as I could. This clinched it and I found some better experts to fit the system I actually needed!
I am not saying your experts are wrong, but it makes a big difference in your case as it did in mine, so it's worth going the extra mile to be as sure as possible.
This post was modified 2 weeks ago 2 times by JamesPa
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
Heat Geeks did a full survey (although as the survey is of the house before the extension, it's a bit hard to see how it properly accounts for the work to be done: I was a bit disappointed that the much-vaunted heat geek toolset couldn't cope with this better). The other installer assessed from architect's plans with a follow up visit to inspect existing radiator sizes. I was pretty happy with thoroughness.
Heating typically doesn't run at night but it is a 2.5 C set back rather than fully off. Thus we might guess the house doesn't cool down more than that. We have to note the apology that the thermostat is in the hallway, and so its reading isn't very representative of the temperature in the living spaces.
I had a go at my own heating degree days regression analysis.
I downloaded gas usage for my account from Octopus. Full data range available would be 14th Feb 2024 (when I changed supplier to them) to today, although I used 1st April 2024 to 31st March 2025.
I used https://www.degreedays.net/ to source historical weather data. Their free version is limited to max 366 data points, hence the date restriction above. I might work around to pull more data points into my chart, or even consider paying a subscription.
15 C baseline temperature for heating degree days is just the number suggested by the degreedays.net regression analysis but it does look like a pretty good fit for the data.
Regression line is suggestion 0.335 kW / degree day with a very small zero intercept which I am going to neglect (cooking and hot water, I guess). Design temperature -3.2 C (lifted from one of the installers). That is 18.2 degree days on my 15 C baseline. Suggests power of 6.1 kW would be sufficient at design temperature.
Drawbacks:
Analysis is for the house as it is now, not as it is to be. Unfortunately I can't collect the data for the house as it is to be until after I make a heat pump decision!
Current heating set up isn't great and frankly it's a bit cold in the winter: thus we might run the house a bit warmer with the heat pump and so use more heating power. We can already see that in our 10 year data. There is a bit of a bump in the 12 month average line in 2021, which I associate with the covid lockdowns. We were at home during the day so we ran the heating until the house was comfortable.
I didn't run the boiler in a heat-pump-like way over the sample period.
I don't know the gas boiler efficiency (but I guess using gas usage rather than heat output is just giving extra margin from the point of view of choosing a heat pump size).
Anyway, I think I need to ask my prospective installers some questions.
Anyway, I think I need to ask my prospective installers some questions.
Agree, but you may have a hard time convincing them, as an overestimated survey is their ultimate protection. If you can get their calculations (and if you paid for the surveys you should) then it would help.
Analysis is for the house as it is now, not as it is to be. Unfortunately I can't collect the data for the house as it is to be until after I make a heat pump decision!
If you can get the survey calculations you could do a 'difference' correction. Even without you could have a go. It would be good to work out for sure whether the changes were + or -
Current heating set up isn't great and frankly it's a bit cold in the winter: thus we might run the house a bit warmer with the heat pump and so use more heating power. We can already see that in our 10 year data. There is a bit of a bump in the 12 month average line in 2021, which I associate with the covid lockdowns. We were at home during the day so we ran the heating until the house was comfortable.
You could also do an estimate for this, but its a bit vaguer. Still an estimate is better than no estimate
I didn't run the boiler in a heat-pump-like way over the sample period.
What flow temp were you operating at.
If it was the typical 75C then at best boiler efficiency will be 95% because it wont be condensing much of the time - which you can also deduce if you have thick plumes of vapour coming out of the flue.
If it was 55C or less then you might get 110% (because the recovered latent heat of condensation isn't included in the declared calorific value of the gas)
Unfortunately our hapless heating industry insists on setting up condensing boilers so they don't condense, adding about 10% to heating bills, and if you change it yourself even resetting it when serviced!
I don't know the gas boiler efficiency (but I guess using gas usage rather than heat output is just giving extra margin from the point of view of choosing a heat pump size).
Yes and see comment above
Based on my experience and to an extent what others report here the most likely 'errors' in the installer calculations are
Ignoring fabric upgrades which are invisible (and thus they cant verify)
Counting room to room losses but not room to room gains (effectively therefor double counting
Overestimating ACH. The default is 2-3, actual may be more like 0.5-1 it seems from various sources
In my case the first two accounted for the difference between 16kW (surveyed and calculated by 'experts') and 10.5kW (calculated by me using MCS assumptions) and the second the difference between 10.5kW and 7kW (measured).
I would say its definitely worth taking this to the next step and doing some of the estimates above (in a way that allows you to adjust ACH), particularly as form factor matters. FWIW a neighbour of mine went through something similar and likewise had a form factor constraint. He ended up fitting the 7kW with the Vaillant 3kW supplementary electric heater. The latter is never, or almost never, used but kept the installer happy. Its a matter of debate whether this arrangement meets the BUS rules!
FYI not much point in getting the 10kW Vaillant, its just the 12kW one software limited. Similarly with the 3kW which is the 5kW software limited. Unfortunately many manufacturers play this game which is very annoying. Daikin, for example, do about 8 sticker variants but only 2 actual physical models.
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
