Concrete Screed for wet UFH - advice please

[JohnW]

I will be installing wet UFH and I have been advised to lay 100mm depth of 35N concrete screed (sand/cement/10mm stone) laid as a wet mix with a brush finish. The thinking behind it is, that the stone increases the thermal mass of the floor (more so than just sand/cement). This doesn't appear to be a common screed. From reading various threads on here, it appears that most people are using either liquid screeds or sand/cement.

 

Have any of you used such a concrete screed or can you advise?

[jack]

Anyone with an MBC slab (including me, @JSHarris, @TerryE and several others) have just this arrangement, although ours are reinforced and sitting on 300mm EPS.  Works very well with decent insulation under it.

 

What's your floor buildup more generally?  

[JohnW]

Thanks @jack ground floor build-up is... (from top down)

• 100mm concrete screed (with wet UFH pipes embedded)
• 125mm foil backed polyurethane (2 layers 75mm and 50mm with staggered joints)
• DPC (1000 gauge polythene)
• sub floor

[Dudda]

What's your final floor finish? Are you polishing the concrete or putting something else over it like carpet, tile, timber, etc?

 

I'd consider trying to get more insulation say 150mm in the floor unless someone has worked out the u-values in detail and can justify 125mm. You can get 150mm thick boards so no need for multiple layers with staggered joints. Quicker to put in. Also it's advisable to put a slip layer or an additional layer of polythene on top of the insulation to stop the concrete seeping between the joints. You could also tape the joints instead of the additional layer of polythene.

[Jeremy Harris]

As above, more insulation, ideally.  UFH will increase the floor slab temperature, so it increases the level of insulation needed in order to keep the heat losses down into the ground at a sensible level. 

[JohnW]

Thanks @Dudda, we plan to finish the floors with tiles.

 

We have been advised that using the 125mm will achieve a u-value of 0.15, however I accept your point about increasing it to 150mm. One of the reasons for the 2 layers of insulation is to be able to accommodate the first fix plumbing, e.g. areas that have pipes will be built-up as 50mm below pipes, 50mm beside pipes and 25mm above pipes, giving a total of 125mm to match areas that do not have pipes (50mm + 75mm). Thanks for the tip regarding the tape/polythene for the joints.

 

 

 

 

[JohnW]

Thanks @JSHarris, what depth of insulation would you recommend?

[Jeremy Harris]

27 minutes ago, JohnW said:

Thanks @JSHarris, what depth of insulation would you recommend?

 

It depends on the insulation material; we have 300mm of EPS, but if that were PIR it would reduce to about 200mm and still have a similar U value.  I don't know what your heating requirements are, in terms of heat loss, but if, say, you needed to run the UFH at around 24 deg C in order to heat the house OK, with a typical ground temperature of around 8 deg C (reasonable for most of the UK, except particularly cold regions) then the heat loss into the ground with a U value of 0.15 W/m.K is going to be around 2.7W/m² and the heat input to the room (with a room temperature of 21 deg C) is going to be about 27 W/m², so about 10% of the heat input will be wasted.  That's the maximum allowable normally under building regs, as it happens. 

 

For comparison, our house has a lot lower heating requirement; the UFH floor temp is usually around 22 to 22.5 deg C, and our under slab insulation has a U value of just under 0.1 W/m².K. For a 21 deg C room temperature and 22.5 deg C floor temperature, our UFH heat loss into the ground is about 1.45 W/m² and our heat output into the room is about 14 W/m², so a bit over 10%, even with 300mm of insulation. 

 

One of the paradoxes of improving the thermal efficiency of a house is that you pretty much always increase the UFH percentage heat loss, because of the ratio between heat delivered to the room versus heat delivered into the ground under the house.

 

If you want to calculate the heat delivered into the room from a UFH system, then the formula is P (in watts/m²) = 8.92 x (floor surface temp - room temp) ^1.1

Edited February 27, 2017 by JSHarris

[JohnW]

Thanks again @JSHarris, that's very useful information. I'll spend the evening trying to calculate the sweet spot of insulation thickness for my build.

 

As you very clearly know your way around the thermal calculations would you mind if I re-position my original question?...

 

The original question was regarding the concrete screed, and I believe the thinking behind the inclusion of 10mm stones is to increase the thermal mass and therefore ability to retain heat within, allowing it to slowly release over a longer time. My builder has suggested using 50mm of liquid screed (Cemfloor Liquid Screed C20/F4) instead. I can see and understand the benefits of the liquid screed during installation, however what I can't see is whether it will be as efficient and effective as the concrete in heating the house. Any thoughts/calculations would be appreciated.

[Nickfromwales]

Allow me 

 

[Onoff]

Tiles

100mm concrete

A142 mesh 

Polypipe tray with UFH

100+50mm PIR

DPM

25mm EPS 

Sharp sand blind

Compacted sub base

 

Btw I have no idea what I'm doing. Standing on the shoulders of giants etc.

[Jeremy Harris]

As someone's already mentioned, "thermal mass" isn't a measurable parameter, has no units associated with it, and no known way of calculating what it actually is for any material.

 

Concrete and screed are very similar in terms of heat capacity (defined as the amount of heat energy required per unit mass, or volume, per unit change in temperature) and thermal conductivity (defined as the rate at which heat energy flows through a given thickness of material for a particular temperature differential).

 

There won't be a significant, or even probably detectable, performance difference between laying UFH pipes in the slab and laying them in a screed on top of a slab, in practice, particularly if the overall heating requirement is low, so the UFH temperature is also low.  If the heating requirement is significant, say more than about 40 to 50 W/m², then there may be a very slight advantage in terms of initial heat up time to laying the UFH in a screed, the question is really whether it's worth all the extra cost and time involved.

 

My own view is that, for a well-insulated house, with a decent level of airtightness, and MVHR to significantly reduce the ventilation losses, then the heating requirement is so low as to make it not worth considering all the additional cost and time involved in laying UFH in  a screed on top of a slab, especially as it increases the overall floor height build up and the screed adds nothing structurally.

 

We have a 100mm thick slab, with the UFH pipes cable tied to the top of the steel reinforcement fabric on 200mm centres, so the centre of the UFH pipes is around 42mm below the surface of the floor.  Our slab was power-floated dead flat, so the floor finishes (12mm thick travertine stone and 12mm thick bonded down bamboo) are laid directly on to the slab.  This was a pretty cheap and quick way to get a heated finished floor.

Edited February 27, 2017 by JSHarris

[JohnW]

46 minutes ago, Nickfromwales said:

Allow me 

 

Thanks @Nickfromwales & @JSHarris...I used to consider myself a reasonably intelligent person, but having read that thread I realise I am walking among giants! I think my floor slab will have fully heated by the time I absorb all that information.

[JohnW]

11 minutes ago, JSHarris said:

As someone's already mentioned, "thermal mass" isn't a measurable parameter, has no units associated with it, and no known way of calculating what it actually is for any material.

 

Concrete and screed are very similar in terms of heat capacity (defined as the amount of heat energy required per unit mass, or volume, per unit change in temperature) and thermal conductivity (defined as the rate at which heat energy flows through a given thickness of material for a particular temperature differential).

 

There won't be a significant, or even probably detectable, performance difference between laying UFH pipes in the slab and laying them in a screed on top of a slab, in practice, particularly if the overall heating requirement is low, so the UFH temperature is also low.  If the heating requirement is significant, say more than about 40 to 50 W/m², then there may be a very slight advantage in terms of initial heat up time to laying the UFH in a screed, the question is really whether it's worth all the extra cost and time involved.

 

My own view is that, for a well-insulated house, with a decent level of airtightness, and MVHR to significantly reduce the ventilation losses, then the heating requirement is so low as to make it not worth considering all the additional cost and time involved in laying UFH in  a screed on top of a slab, especially as it increases the overall floor height build up and the screed adds nothing structurally.

 

We have a 100mm thick slab, with the UFH pipes cable tied to the top of the steel reinforcement fabric on 200mm centres, so the centre of the UFH pipes is around 42mm below the surface of the floor.  Our slab was power-floated dead flat, so the floor finishes (12mm thick travertine stone and 12mm thick bonded down bamboo) are laid directly on to the slab.  This was a pretty cheap and quick way to get a heated finished floor.

 

You mentioned that you have a 100mm slab with the UFH pipes within, would you mind sharing the full floor build-up please?

[Nickfromwales]

12 minutes ago, JohnW said:

I think my floor slab will have fully heated by the time I absorb all that information

PMSL 

That made me laugh out loud. Good man! 

[Jeremy Harris]

18 minutes ago, JohnW said:

 

You mentioned that you have a 100mm slab with the UFH pipes within, would you mind sharing the full floor build-up please?

 

 

We have a pretty standard "passive slab", which is also the foundation for the house.  The under-insulation build up varies a bit with the type of soil, but in our case we excavated to a depth of 150mm over the whole area of the house, plus about 300mm all around.  This was filled with type 3 stone (18mm to 35mm, no fines) compacted in layers until it was 150mm deep (so the same as the ground level.  On top of this was laid a 25 to 50mm thick layer of coarse blinding grit, very carefully levelled.  The EPS insulation was laid directly on to this, with L shaped sections all around the edges and the interior filled with over-lapping 100mm thick EPS sheets.  The DPM was laid under the last 100mm thick insulation sheet.  The steel reinforcement fabric was then laid in place, including a reinforced ring beam around the edge and for two internal load bearing walls.  The UFH pipes were fitted, then the concrete was poured and power floated smooth.  The whole process of laying the foundation to finished and level slab, with UFH pipes in, took 4 days.

 

So, the build up, from the bottom, is 150mm compacted type 3, plus grit blinding, 200mm of EPS, DPM, 100mm of EPS then 100mm of reinforced concrete on top, with the 12mm floor finishes (travertine and bamboo) on top of that.

 

It's documented in our build blog, starting at about this entry, with photos: http://www.mayfly.eu/2013/10/part-sixteen-fun-and-games-in-the-mud/

Edited February 27, 2017 by JSHarris

[JohnW]

@JSHarris, once again a very comprehensive and helpful response. Thank you.

[Ferdinand]

@JohnW

 

I am thinking that perhaps you need to be questioning that concrete screed.

 

One question to ask might be by how much it increases the "thermal mass" over concrete. The not very meaningful answer is likely to be to do with the increased density - and I suspect the increase is not very much.

 

The cost per cubic meter difference and the pfaffing involved in 2 materials means you may be able to fund a West End theatre weekend (or a posh meal out) to help you unwind from the difference while it starts to set. You can never have too many unwinding weekends.

 

Edited February 27, 2017 by Ferdinand

[Jeremy Harris]

1 hour ago, Ferdinand said:

@JohnW

 

I am thinking that perhaps you need to be questioning that concrete screed.

 

One question to ask might be by how much it increases the "thermal mass" over concrete. The not very meaningful answer is likely to be to do with the increased density - and I suspect the increase is not very much.

 

The cost per cubic meter difference and the pfaffing involved in 2 materials means you may be able to fund a West End theatre weekend (or a posh meal out) to help you unwind from the difference while it starts to set. You can never have too many unwinding weekends.

 

 

It's a good question to ask. 

 

Generally, some screeds have a very slightly better thermal conductivity than concrete, but a similar heat capacity.  There are some screeds that have a higher heat capacity, but they are the ones that are very, very slow to harden, and generally a bit of a pain when it comes to trying to bond things like tiles or anything else to.

 

Unfortunately there is a fair bit of mythology surrounding a lot of building-related stuff, often not deliberate, just a consequence of the natural conservatism of the building trade in general, and a reluctance to change older ways of working.

 

If you want to store a lot of heat in something like a slab, and release it slowly over a long period of time (eee Terry's useful thread here:

) then it's hard to beat plain concrete as a multi-function material.  Not only does it have the strength and stiffness needed to provide a sound foundation system, but it also has a reasonable heat capacity for a solid material (about 880 J/kg°C , compared with about  4181 J/kg°C for water, so about 20% of the heat capacity per unit mass of water which is pretty good for a building material) and a reasonable thermal conductivity, about the same as water for the sort of medium density concrete used in floor slabs (water's about 0.58 W/m.K, medium density concrete varies from around 0.4 to 0.7 W/m.K).

 

It's worth noting that as a general rule of thumb, you can ignore heat stored in any part of the structure deeper than around 100mm, as the time taken for heat from that depth to flow to the surface and contribute to the heat input to the building is usually so long as to be insignificant in terms of thermal modelling the internal thermal time constant.  The latter is important for comfort, and it's worth looking at the above ground structure to determine how that will respond.  A good example of something often overlooked is the decrement delay of insulation.  There's a good explanation of that in this link: http://www.greenspec.co.uk/building-design/decrement-delay/ that's worth a read.  I can vouch for the fact that having a relatively high decrement delay insulation system, in our low overall mass new house, has a very marked impact on temperature stability, particularly helping to keep it nice and cool in hot weather.

 

Edited February 27, 2017 by JSHarris
typo

[PeterW]

@JohnW you have a similar floor build up to me but with a couple of exceptions.I have a 125/130 PIR going down with a 25mm EPS all round to absorb the lumps instead of sand a la @Onoff. My top course is a flowing concrete from a local supplier with fibre reinforcement in it - a polyester fibre - and a number of plasticisers that make it self levelling and self compacting. Target for us is 0.14, but that is a choice as we have a conversion element too and it won't allow us without a significant amount of work to get to a very low uValue on the walls so we've gone for a very warm "hat" on the build ...!

 

Concrete flooring is probably the easiest part to get wrong but I've basically gone with it as it will make life easier. Is it more "dense" than normal..? Who knows..! Unless concrete is aerated, the mix of aggregate and sand to cement and water will make a very minor difference to the uValue overall.

 

 

[Declan52]

How do you intend to  build the floor of  house assuming the ground conditions should be good where you are.

Are you considering as you are going block build doing the founds then pouring a 100mm subfloor then continue to build and put the insulation and ufh bits plus screed after first fix. This would be by far the most common method here in NI.

Or do you only intend to do one floor which will contain steel mesh with the ufh in this concrete. Probably a bit more risky as the floor will have a lot of weight on it when it's loaded out with a few loads of blocks.

 

[joe90]

Due to difficult ground conditions we have gone with a different foundation, copying the Golcar foundation design. http://www.greenbuildingstore.co.uk/golcar-passivhaus-ground-floor-foundations/. But we have had a 100mm oversite concrete slab with steel mesh then the 200 poly insulation then the UFH in a 60 mm screed ( this was done partially because of illness timing on my part). All topped with either engineered timber or tile floors.

[JohnW]

Ok...having digested The Great Thermal Mass Myth................ thread I will attempt to summarise my understanding, so please correct me if I've got it wrong.

 

• "Thermal mass" is a term used frequently which has no meaningful measurement and therefore should be ignored
• The 2 measurable characteristics I should look to compare are; Heat Capacity (J/deg K) and Thermal Conductivity (W/m K) for convenience I will refer to them as HC and TC

Ok, armed with this information it appears that a concrete screed (with 10mm stone) probably isn't very different in terms of both HC & TC to a standard sand/cement screed. For arguments sake we'll say concrete screed has the following values HC(880) and TC(0.8) so to compare it to liquid screeds I need to know their HC & TC...

 

The 2 liquid screeds I'm considering are;

• RTU Ultraflo with a HC(?) and TC(1.66-1.88) and
• Cemfloor C20/F4 with a HC(?) and TC(2.9) 

I haven't been able to find out the HC of these liquid screeds, however the higher TC values suggests that they will conduct heat quicker than concrete and will therefore "respond" quicker, I.e. gain heat quicker and cool quicker. So the question is, what impact is this likely to have on my heating strategy? I am assuming not very much in a well insulated, reasonably airtight house or will the unknown HC values for the liquid screed scupper my master plan???

Edited February 27, 2017 by JohnW

[JohnW]

2 hours ago, Declan52 said:

How do you intend to  build the floor of  house assuming the ground conditions should be good where you are.

Are you considering as you are going block build doing the founds then pouring a 100mm subfloor then continue to build and put the insulation and ufh bits plus screed after first fix. This would be by far the most common method here in NI.

Or do you only intend to do one floor which will contain steel mesh with the ufh in this concrete. Probably a bit more risky as the floor will have a lot of weight on it when it's loaded out with a few loads of blocks.

 

We've yet to break ground but we are very hopeful that the ground is good enough for strip foundations, concrete sub-floor (not 100% sure what depth), then as you say insulation, UFH and screed after 1st fix.

[Frogeye]

I'd be a bit concerned that the 10mm gravel may damage the ufh pipe as it is being crunched around on whilst being laid