Concrete Screed for wet UFH - advice please

[Declan52]

15 minutes ago, JohnW said:

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.

Any particular reason for 100mm screed and not 75mm depth with the extra 25mm going to improve your insulation depth. One sheet of 50mm to cover the first fix pipe work then overlapping with a 100mm sheet.

Or build it to suit 200mm insulation if the budget can stretch that far.

[JohnW]

42 minutes ago, Declan52 said:

Any particular reason for 100mm screed and not 75mm depth with the extra 25mm going to improve your insulation depth. One sheet of 50mm to cover the first fix pipe work then overlapping with a 100mm sheet.

Or build it to suit 200mm insulation if the budget can stretch that far.

No reason that I am aware of. If the budget allows we will consider more insulation and less screed. In fact it is one of the reasons I'm considering the liquid screed, it's only 50mm so would potentially give us 50mm to play with.

[PeterW]

3 minutes ago, JohnW said:

No reason that I am aware of. If the budget allows we will consider more insulation and less screed. In fact it is one of the reasons I'm considering the liquid screed, it's only 50mm so would potentially give us 50mm to play with.

 

Be careful - most liquid screeds need 50mm above the heating pipes so you need 65mm min

 

Why not go for a ground bearing slab and just put insulation then concrete with UFH in it ..? One less pour and type 1 is much cheaper than concrete ...!

[JohnW]

25 minutes ago, PeterW said:

 

Be careful - most liquid screeds need 50mm above the heating pipes so you need 65mm min

 

Why not go for a ground bearing slab and just put insulation then concrete with UFH in it ..? One less pour and type 1 is much cheaper than concrete ...!

This approach sounds the same as JSHarris http://www.mayfly.eu/2013/10/part-sixteen-fun-and-games-in-the-mud/

I hadn't considered it before now will certainly give it some thought - thanks.

 

Is this approach more expense?

Edited February 27, 2017 by JohnW

[oranjeboom]

@JohnW: Not sure whether you are retrofitting or not....I don't think you are, but I am considering using a product called "Agilia", which is described as self-levelling/self placement, so no power floating needed, less labour needed and works fine with UFH. Good for retrofit as in my case where access for power floating is not an option. But, I think you are new build, so may as well go a la JSHarris with C25(?) concrete.

 

[JohnW]

16 minutes ago, oranjeboom said:

@JohnW: Not sure whether you are retrofitting or not....I don't think you are, but I am considering using a product called "Agilia", which is described as self-levelling/self placement, so no power floating needed, less labour needed and works fine with UFH. Good for retrofit as in my case where access for power floating is not an option. But, I think you are new build, so may as well go a la JSHarris with C25(?) concrete.

 

Yeah, it's a new build.

[PeterW]

7 hours ago, JohnW said:

This approach sounds the same as JSHarris http://www.mayfly.eu/2013/10/part-sixteen-fun-and-games-in-the-mud/

I hadn't considered it before now will certainly give it some thought - thanks.

 

Is this approach more expense?

 

@JSHarris and others have a Kore or other manufacturer EPS fully insulated slab. With a standard strip or trench fill construction you can do the same, building up your external walls until you get to DPC and then backfilling with 100-150mm MoT1 and 100-150mm PIR insulation before casting the slab. PIR needs approx 1/2 the thickness of EPS for the same uValue but in certain situations it makes more economic sense if you have a lot to lose on levels. 

 

I'll dig out our sections later and show what the build up is 

[Jeremy Harris]

4 minutes ago, PeterW said:

PIR needs approx 1/2 the thickness of EPS for the same uValue but in certain situations it makes more economic sense if you have a lot to lose on levels. 

 

 

Worth mentioning that Kore EPS for the floor system we have has a lower lambda than some EPS, the info is on the Kore website here: https://www.kore-system.com/kore-products/floor-insulation/kore-floor/technical  PIR has a lambda of around 0.021 W/m.K to 0.022 W/m.K, and  Kore under floor EPS has a lambda of between 0.031 W/m.K and 0.036 W/m.K (the 0.031 W/m.K figure being for EPS70, the higher lambda for EPS100).

 

So the ratio between EPS and PIR is not as bad as 2:1, more like 1.6:1 for a structural floor, or around 1.5:1 for a non-structural floor.  For example, to get the same sort of U value as we have, for a structural slab, but using PIR instead of EPS 100, would (simplistically, and ignoring perimeter loss)  reduce the insulation thickness from 300mm to about 190mm.

 

 

[PeterW]

Thanks @JSHarris - more a rule of thumb than technically accurate as I quoted ..! I've had some fun with ours as the PA values vary depending on which section of the building it is so I've calculated them as 3 separate values and had to combine them.

 

Also doesn't help that the integral garage also has an EPS layer under it to try and reduce some heat loss from within there ...!! 

[Jeremy Harris]

I think the main advantages of EPS are that it's usually a fair bit cheaper than PIR and has decades of proven use deep underground, as it's been used for things like basement insulation and boggy ground railway track foundation support for a long time now. 

 

PIR should be as durable, but there isn't as much long term usage data as there is with EPS.  IIRC, they started using EPS under basement foundations in Germany around 40 years ago, which gives a fair bit of reassurance that it has the long term durability needed.  PIR should be similar, as accelerated testing shows it to be OK, it just hasn't been used as much, or for so long, as EPS in this application.

 

For a non-structural floor in a conventional build, like the one shown on that Kore floor insulation link above, then 200mm of EPS70 looks like it should give a reasonable performance, without too much build up in level and I'd guess it'd be a fair bit cheaper than PIR.

[Temp]

You mentioned tiling so beware as some screed is hard to file over...

 

http://www.bal-adhesives.com/tiling-onto-calcium-sulfate-anhydrite-screeds/

 

It may seem obvious but with UFH you also want to avoid aerated or light weight screed.

[JohnW]

5 hours ago, Temp said:

You mentioned tiling so beware as some screed is hard to file over...

 

http://www.bal-adhesives.com/tiling-onto-calcium-sulfate-anhydrite-screeds/

 

It may seem obvious but with UFH you also want to avoid aerated or light weight screed.

Thanks @Temp

[JohnW]

7 hours ago, PeterW said:

 

@JSHarris and others have a Kore or other manufacturer EPS fully insulated slab. With a standard strip or trench fill construction you can do the same, building up your external walls until you get to DPC and then backfilling with 100-150mm MoT1 and 100-150mm PIR insulation before casting the slab. PIR needs approx 1/2 the thickness of EPS for the same uValue but in certain situations it makes more economic sense if you have a lot to lose on levels. 

 

I'll dig out our sections later and show what the build up is 

Thanks @PeterW, I would appreciate a look at your sections

[PeterW]

This isnt the latest but you get the idea

[JohnW]

Thanks @PeterW, it's not dissimilar to what we have been advised to use. (See below - I should also point out that this image  and all technical & design information contained within are the copyright of Reinco Insulation & Renewables consultant, Eric Davidson)

Edited February 28, 2017 by JohnW
Added copyright information for the image

[Jeremy Harris]

@JohnW, your build up looks better, as you're partially mitigating the thermal bridge at the internal wall/floor junction with the Quinnlite blocks, but I'd question the need for the separate, cold, sub-floor.  I think you could useful change this detail to the one in the Kore link I gave earlier, leaving your foundation and wall detail as-is.  If you have room, then I'd suggest laying the insulation either direct on to a blinded, packed type1 sub-base (cheap!) , or with the blinding covered with the DPM, then the insulation, and making the finished floor the structural layer, with the UFH pipes closer to the top than the bottom, to further reduce losses.

 

If I get time later I'll try and do a sketch to show exactly what I mean, but I'm pleased to see that some effort has been taken to reduce thermal bridging in the internal wall - this is a detail that is often overlooked, it seems.

Edited February 28, 2017 by JSHarris
typo

[joe90]

32 minutes ago, JSHarris said:

@JohnW, your build up looks better, as you're partially mitigating the thermal bridge at the internal wall/floor junction with the Quinnlite blocks, but I'd question the need for the separate, cold, sub-floor.  I think you could useful change this detail to the one in the Kore link I gave earlier, leaving your foundation and wall detail as-is.  If you have room, then I'd suggest laying the insulation either direct on to a blinded, packed type1 sub-base (cheap!) , or with the blinding covered with the DPM, then the insulation, and making the finished floor the structural layer, with the UFH pipes closer to the top than the bottom, to further reduce losses.

 

If I get time later I'll try and do a sketch to show exactly what I mean, but I'm pleased to see that some effort has been taken to reduce thermal bridging in the internal wall - this is a detail that is often overlooked, it seems.

That makes it similar to the golcar passive house I linked to earlier.

[JohnW]

2 hours ago, joe90 said:

That makes it similar to the golcar passive house I linked to earlier.

Thanks @joe90, @JSHarris it's good to hear we appear to be planning correctly...just have to build it correctly now

[Jeremy Harris]

It's really largely about getting the detailing right, and a bit of research before work starts can reap a useful benefit at little or no additional cost, or may even give you a cost saving.

[joe90]

29 minutes ago, JSHarris said:

It's really largely about getting the detailing right, and a bit of research before work starts can reap a useful benefit at little or no additional cost, or may even give you a cost saving.

This forum has given me soooo much important information to allow me to plan my build exactly as I want it and to the best standards available.

[JohnW]

3 hours ago, JSHarris said:

@JohnW, your build up looks better, as you're partially mitigating the thermal bridge at the internal wall/floor junction with the Quinnlite blocks, but I'd question the need for the separate, cold, sub-floor.  I think you could useful change this detail to the one in the Kore link I gave earlier, leaving your foundation and wall detail as-is.  If you have room, then I'd suggest laying the insulation either direct on to a blinded, packed type1 sub-base (cheap!) , or with the blinding covered with the DPM, then the insulation, and making the finished floor the structural layer, with the UFH pipes closer to the top than the bottom, to further reduce losses.

 

If I get time later I'll try and do a sketch to show exactly what I mean, but I'm pleased to see that some effort has been taken to reduce thermal bridging in the internal wall - this is a detail that is often overlooked, it seems.

In the scenario you paint above with the UFH pipes close to the top of the structural layer, would the structural layer not be deeper than 100mm and therefore a large percentage of the heat from the UFH pipes would transfer downwards and never make it to the surface where we need it?

[Jeremy Harris]

9 hours ago, JohnW said:

In the scenario you paint above with the UFH pipes close to the top of the structural layer, would the structural layer not be deeper than 100mm and therefore a large percentage of the heat from the UFH pipes would transfer downwards and never make it to the surface where we need it?

 

It's the way our slab's laid, and works well.  We have a layer of reinforcing steel fabric on 50mm chairs, so at the mid-point in the 100mm thick slab.  The UFH pipes are on top of this, so are closer to the top than the bottom (the top of the pipes is around 34mm from the surface of the slab), which improves the response time slightly and also reduces the losses slightly.  The thermal conductivity of concrete is pretty close to that of water, so pretty high, and the heat capacity of concrete is a lot less than water, so in practice heat travels from the water in the pipes to the surface of the slab quite quickly.  There will still be a temperature gradient both ways when the systems running, with the centre of the pipe being the highest temperature and the temperature dropping the further away from that you go in the slab.

 

If the pipes were at the bottom, then the losses would increase slightly, as the hottest part would be further away from where you want the heat delivered, and closer to the place you don't want the heat to go, the ground underneath.  No insulation is perfect, there will always be heat lost through it whenever there is a temperature difference from one side to the other, and the key thing here is that the higher the pipes are in the slab the lower the temperature of the slab will be at the bottom, which is a good thing overall.

Edited March 1, 2017 by JSHarris

[Dudda]

On 28/02/2017 at 07:34, JSHarris said:

 

Worth mentioning that Kore EPS for the floor system we have has a lower lambda than some EPS, the info is on the Kore website here: https://www.kore-system.com/kore-products/floor-insulation/kore-floor/technical  PIR has a lambda of around 0.021 W/m.K to 0.022 W/m.K, and  Kore under floor EPS has a lambda of between 0.031 W/m.K and 0.036 W/m.K (the 0.031 W/m.K figure being for EPS70, the higher lambda for EPS100).

PIR can be as low as 0.018W/m.K eg Kingspan K103 floor insulation. I've had to use it in our refurb where I'm keeping the sub floor in areas and still want to retain the existing sub floor and not increase floor levels.

 

In addition if you're renovating you obviously can't install aerated or Quinnlite blocks in a wall so another trick to reduce the thermal bridge is to put vertical perimeter insulation where you're removing the sub floor and hardcore. It increases the distance for the cold to transfer thus reducing the thermal bridge.

 

Edited March 1, 2017 by Dudda

[DeeJunFan]

15 hours ago, JSHarris said:

@JohnW, your build up looks better, as you're partially mitigating the thermal bridge at the internal wall/floor junction with the Quinnlite blocks, but I'd question the need for the separate, cold, sub-floor.  I think you could useful change this detail to the one in the Kore link I gave earlier, leaving your foundation and wall detail as-is.  If you have room, then I'd suggest laying the insulation either direct on to a blinded, packed type1 sub-base (cheap!) , or with the blinding covered with the DPM, then the insulation, and making the finished floor the structural layer, with the UFH pipes closer to the top than the bottom, to further reduce losses.

 

If I get time later I'll try and do a sketch to show exactly what I mean, but I'm pleased to see that some effort has been taken to reduce thermal bridging in the internal wall - this is a detail that is often overlooked, it seems.

 

In NI this is the "Way" it would be very difficult to find a builder locally who would even consider doing anything differently.

 

This is the same build up as i have used.  We have 150mm concrete sub-floor, quinn lite blocks at the slab level and then we are currently debating to go either 200mm PIR with Sand/Cement screed or 250mm PIR and 50mm CemFloor screed.

 

 

[Declan52]

I wouldn't want my finished floor put in at the founds stage if I was going to be building blocks. The floor will be ruined!!! 

Plus all that weight on your floor with the ufh piped in it is taking a chance. Depending on how access is you might have to stack blocks up 5 straps high which is 5t sitting on 1m2 which is a lot.