UFH location in concrete slab with mesh

[SuperPav]

So we're digging up the existing uninsulated slab (50sqm) to the same level as the oversite under the suspended timber floors (which are also coming up) approx 400mm below FFL.

 

With some hardcore/blinding, we're looking at a build up of approx 225mm of insulation (100mm PIR with rest in EPS), on top of which we'll have about 100-125mm of SLC with a A193 or similar mesh in it.

 

We're getting someone in to pump and lay the SLC, but will be doing the UFH and mesh ourselves. Plan was to lay the mesh on soldiers and then simply tie the UFH pipes to the mesh.

 

Question is, could we staple the UFH down to the PIR, and then set the mesh above it on soldiers? Saw this on an episode of something recently (maybe an old Grand Designs or similar), and it occurred to me that it might be a better way to prevent any damage to the pipes from walking over, plus might make laying the pipe easier if I'm not having to navigate the mesh to clip to. Is there something obvious I'm missing here (like extra heat loss etc.) or is this an option?

 

Bay window and internal floor dig out.pdf

[Conor]

The depth of the mesh is dependent on its purpose. I'm assuming it's only for anti-cracking reasons? In which case it would be set 40-50mm below the surface... So perfect to tie your UFH pipes to.

 

If you have the pipes deeper, the response time wil be slower. You'll have a bit more heatloss but not sure how significant it would be, of at all.

[Chanmenie]

I’m doing the same, A193 mesh and 120mm concrete, so will cable tie the pipes to the mesh then lift in onto the 40/50mm rebar chairs.

[Simplysimon]

i'd be looking at 100mm concrete on top of 200mm PIR and fill the rest with hardcore/ EPS

[TerryE]

Your profile is similar to those of us that have MBC warm slabs.  The rebar mesh was laid on ~40mm soldiers and the UFH pipe tied directly to the mesh.  Have a look at my blog; there are pics of this.  And others have similar. Absolutely no probs.  Remember to tape the UFH pipe ends before the pour, to prevent dirt ingress.  Our slab was laid in late Oct, so we didn't fill and pressure test until after the frame and windows were up and the structure airtight.

 

BTW, I note that you have a cold bridge between the slap and the inner wall leaf.  You typically put a small EPS upstand here to prevent this.  Plastering out + skirting will hide this.

[Nickfromwales]

8 minutes ago, TerryE said:

BTW, I note that you have a cold bridge between the slap and the inner wall leaf.  You typically put a small EPS upstand here to prevent this.  Plastering out + skirting will hide this.

+1. I've just specified 40mm PIR thermal break / upstand for a client where the walls will battened for a service cavity, PB + skirting. There will also be a 10mm foam expansion skirting at the edge of the slab to allow ( the miniscule amount of ) expansion. Pennies to add the foam skirting, so better to be belt and braces. Not necessary where you have a passive raft, but very necessary where you are pouring slabs within 4 boxed in sides and the slab may wish to push out in all 4 directions.  

[joth]

On 24/05/2022 at 12:47, SuperPav said:

So we're digging up the existing uninsulated slab (50sqm) to the same level as the oversite under the suspended timber floors (which are also coming up) approx 400mm below FFL.

 

With some hardcore/blinding, we're looking at a build up of approx 225mm of insulation (100mm PIR with rest in EPS), on top of which we'll have about 100-125mm of SLC with a A193 or similar mesh in it.

 

Orthogonal to your question, but interested if you've had a structural engineer or building control approve this design yet?

I was just mentioning on another thread we had trouble getting an insulated slab approved on a retrofit, so interested to hear others can do this. The fact you had a suspended floor to start with maybe of relevance - depends what the existing concrete slab is actually doing I guess.

 

 

 

[SuperPav]

Thanks everyone

 

@jothThe existing floors do nothing - there are spine internal masonry walls which support everything above, the slab will only be for the ground floor and a few (non-structural) stud partitions. The plan is to give BC the proposed build up and then if they decide they want a SE assessment I'll send it to my SE (who is very good) and he can confirm. I don't think the proposed is a particularly unconventional construction for a slab, so don't anticipate too many issues. Worst case is we'll need to use chunkier mesh..

 

With regards to upstands, yes that's not on the detail as that section was quickly knocked up for the builders to know where to dig down to. We'll be laying all the insulation, and we'll have a 50mm PIR upstand all around the external perimeter (solid walls so we'll be using 52.5mm insulated PB for the walls, so can easily lose 50mm PIR thickness).

 

For those that have tied to the mesh, what mesh did you use? I've used A393 in raft slabs before, and that's relatively sturdy, but as I'm hoping to get away with thinner 6mm or 7mm stuff here, I take it there's no real difference if the pipes are laid before the mesh is raised onto soldiers?

 

With a 120mm slab, do heat up times really matter (in case we stapled the UFH to the PIR)? I half-expected this slab to have such a slow response time to almost make it irrelevant as it'll need to be timed to batch "charge" or just be on all the time at low flow temps anyway?

 

[TerryE]

@SuperPav, Mesh selection is really a choice for your SE or whoever did your slab design and it is driven by that.  Using the mesh as a placement framework for the UFH piping is really a secondary bonus and shouldn't impact on mesh selection.

 

As to slab heating, there are broadly two strategies at the extreme: 

1. Dump heat into the slab at a steady rate so that the slab remains at a slight Δt above room temp.  Here it will radiate ~ 7 ×Δt × A W into the interior.  If this matches the net heat loss of the house, then your house will stay at the set temperature.
2. Calculate your total kWh house losses for the coming day, and for a given input heating rate this give a total heating time. Just dump this into the slab in one or more "chunks", and accept that this will result in a slight ripple (say under 1°C) on your internal temperature.

We have a passive house and have adopted the latter as this was easier to implement, in terms of kit required, ease of control, and simplicity leading to maintenance risks and reduction. 

 

The thermal dynamics here is a different Q entirely and one where I feel more qualified to answer on.  I did my slab design back in 2015, and I was one of the first ones on this forum to go with a UFH solution which went against the prevailing wisdom of including a buffer tank.  I decided to use a 3 kW Willis heater to heat the slab directly and use it as the thermal store for the heat.  I needed to do some modelling of the slab for me to be comfortable with this solution and to ensure that it's thermal performance during the heating cycle was well within sensible limits, before finalising on this implementation.  Luckily my professional background give me some experience of doing this type of modelling, but the main challenge was one of producing a model of the correct simplicity yet enough detail for the exercise to be meaningful.  I have previously documented this in my blog posts and and specifically in my thread, Modelling the "Chunk" Heating of a Passive Slab written in late 2016.

 

For the modelling, I didn't want to get into the complexities of CUDA compute engines and CFD libraries, so to keep the math and computation simple enough to be computable in Fortran on a single core (which is what I used back in the 1980s when I did this sort modelling for a living), I approximated the slab heated by one UFH loop as a (radially symmetric) concrete pipe some 0.120m in diameter and 100m long with a 15mm heating pipe running down its centre and with water circulating through it and heated by a 1 kW element before return.  This allowed me to use a simple radial approximation for the Fourier heat equation and to solve over time for the (r, l, t) coordinates over time and the length of the pipe.  (See this post for more details.)  OK, it's a model and an approximate one at that because the actual heat flow through the pipe is not radially symmetric as the area around the heating pipe is insulated below, radiating to the air above, and adjacent to another.  Even so, this did allow me to investigate the overall varying heating characteristics over time and both along and across the concrete.  This was good enough to confirm that a "Willis direct into slab" approach would work fine for me.  I  subsequently instrumented the actual implementation with lots of DS18B20 digital thermometers which have been continuously logging now for ~5 years, and the slab behaves as predicted.  This can be summarised by more simple and intuitive approximation.

• The slab surrounding a single pipe heating loop can be thought of a (folded) long box of concrete that is insulated below, radiating to the air above and with similar boxes on either side (since these are being heated by a return run at a similar flow temperature).
• The UFH pipe flow is dumps ~1kW (in my case) heat along the centre of this long box, and so the water cools as it flows along the pipe.  The overall  Δt is largely dictated by the water flow rate which in turn depends on the pump head, flow resistance, etc.  In my case this was the biggest difference between the model and actual implementation, as I went with a slower flow rate than my modelled 1 m/s, in order to keep the circulation noise to a minimum. (The UFH is in a services cupboard off or G/F toilet and I don't like to hear the pump noise when I am taking a dump).
• So as the long "box" heats you get a standard radial heat flow curve where the circulating water in the middle is maybe 5°C hotter than the surface during heating period, with maybe a 2-3°C drop along the 100m run length.
• As the heat is dumped into the slab, it slowly but steadily heats up over the heating period, say by 5°C or so over a 7hr heating window.  Certainly if you walk over the slab at the end of a heating cycle in bare feet (as we do), you can notice the 1-2 °C variation across the floor between the flow and return legs of the UFH runs and any gaps in UFH coverage.
• Pretty much as soon as the heating stops, the heat from the warm spots spreads so the feel becomes more uniform (as the heat only needs to flow ~50mm or so through the concrete).
• We heat our slab overnight, so by midnight the slab might be ½-1 °C below room temperature. In the morning after heating it might be 4°C above room temp, and it is now radiating ~28 W/m² into the environment.  This rate will fall during the day as the slab cools, leading (in our case) to a ~1°C ripple on the overall air temperature.  The integrated heat dump from the slab must match the overall house losses, so the colder it is the more we need to heat the slab; and the warmer, the less.

I hope this makes sense.

Edited May 28, 2022 by TerryE

[SuperPav]

Thanks Terry, our house is far from PH standards, although the insulation under the floor will be decent, therefore I suspect our flow temps will need to be much higher than the levels you're talking, but the 120mm concrete should still act as a decent buffer. In my head also the heat loss difference between sticking the pipes on the bottom of the slab rather than the middle are minimal, but putting them at the bottom of the slab should result in a (marginally) more uniform slab surface temperature with regards to flow and return legs.

[TerryE]

We have a heating strategy of keeping our house at the same temperature 24 × 7 (accepting the ±½°C ripple that I mentioned above), though one side effect of having a super-insulated house is that even in winter the house only cools by around 1°C / day if the heating is off.  But the same thermal design process needs to be applied even if you live in a conventional low-insulation leaking house: If you are using UFH as your primary heating source then if your house looses X kWh / per day heat externally, then your UFH heating strategy must supply the same X kWh to keep overall heat balance.  

 

That  (~7 × Δtemp × Area) W is a good rough estimate of how much heat the slab can throw into the room environment, so the more heat you need to input then the higher the Δtemp needs to be, though once this gets up to 10°C then the floor starts to get pretty uncomfortable, and you really need to worry about potential hot spots under in-contact furniture, etc.  Higher rates of temp rise can cause heat stress in the slab which isn't a good thing.  This is why IMO you want to insulate where practical and cost-effective and try to remove external air leaks and use MVHR as this will dramatically reduce your overall heat requirements.

 

A more standard higher O/P slab will benefit from a TMV mixer and buffer tank design, though because the water circulating in the slab is a lot cooler than that in a radiator installation, UFH installations are a far better fit to ASHP installations.

 

Last thing to remember if you are going for a non-constant Time-of-day profile, is the issue of time lag. With a 100mm+ slab (especially if the UFH runs are placed at the base of this) then the heat has to propagate through the say 80mm of concrete before the surface is at temperature and the floor starts to radiate heat into the room.  This means for example if you want the house to be warm when you come home at 6pm, say, then you might need to turn the UFH on at 2 or 3pm and likewise turn it off at 9pm, say.

Edited May 28, 2022 by TerryE

[SuperPav]

Rather than start a new thread, I thought I'd post back here as it's still relevant to the same install

So I've now taken up all the timber floors down to the original concrete oversite. I will through down some sand for a very thin layer of blinding, just to even out any bumps etc. but my question is:

 

Is it OK to have steps due to a small variation of depth between different areas (e.g. 15mm or less)?

 

Some rooms' oversite (where the dividing wall has also been removed) are at slightly different levels to each other (they clearly originally built the footings up to DPC and then filled the oversite to "approx" a consistent level throughout). I really can't be arsed blinding more sand unnecessarily to lift a 3rd of the footprint up just because one ex-closet area has a slightly higher oversite.  (it's a faff barrowing it due to the site layout/restrictions)

 

So in essence the insulation would be the same throughout, but my concrete slab on top would be 5-15mm deeper in some areas than others. I'll most likely be clipping the pipes to the mesh so the step shouldn't affect any UFH laying etc, and even if I was stapling to the celotex, I can't see how 10mm or so would make any real difference... 

 

Thoughts?

[crispy_wafer]

if the higher level is in a particular part of the property you could use a slightly thinner layer over this to mask the base difference, if its all over then beware as thicker PIR sheets are quite rigid and like to rock over undulations .  I think others may get round this by using thinner PIR in multiple layers.

[SuperPav]

Sorry I might not have been clear - individual rooms will be levelled - no undulations etc.

It's where one large room meets another large room, the oversite levels are slightly off, is it OK to have a step in the insulation (the sheets will be cut at that point so will be solid on their respective bases)?

 

Where there are a very localised protrusions (a few concrete remains), I will just channel/reduce thickness of the PIR or leave off the 25mm.

 

Build up is 25mm EPS, 100mm EPS, 100mm PIR

[garrymartin]

Apologies for updating a couple of older topics but if you're like me and have been diligently combing through old threads for nuggets of wisdom, the following might be helpful.

 

Short version, out of three scenarios (pipes at 19mm centres below surface of slab, at 50mm centres below surface of slab, and pinned to the insulation);

• the pipes at 19mm give the best performance, but obviously very close to the surface (~79.2 W/m²)
• the pipes at 50mm showed a small decrease in performance against those at 19mm (~75.1 W/m²)
• the pipes pinned to the insulation showed a much larger performance hit (~56.2 W/m²)

So when considering in-slab UFH and choosing between tied to rebar or pinned to the insulation, tied to the rebar will win from a performance perspective every time.

 

Longer version, in Imperial measurements (US-based author) https://www.pmengineer.com/ext/resources/PME/2019/July/014-017-pme-0719_siggy.pdf

[JohnMo]

7 minutes ago, garrymartin said:

when considering in-slab UFH and choosing between tied to rebar or pinned to the insulation, tied to the rebar will win from a performance perspective every time

Not sure what to state is real in reality. Unless your heat losses are huge of course - but then you wouldn't be installing UFH unless you were mad with money to burn.

 

If I needed 70+W/m2 the house would require 13kW to heat, reality is I need 3kW at -9. So why would I design the UFH to give out that much heat. Mines at the bottom of a 100mm concrete pinned to insulation and at 300mm centres. Slow to respond yes, does it make any difference - no, as you keep the slab (60 Tonnes) up to temperature, by low and slow heating or batch heating on E7.

[garrymartin]

9 minutes ago, JohnMo said:

If I needed 70+W/m2 the house would require 13kW to heat...

Not an expert in this area, and I may have failed to use the correct conversion factors from the BTU/h/ft² used in the article. Regardless, it seems clear that tied to rebar is better for in-slab UFH than pinned to the insulation from a performance perspective.

[JohnMo]

Please see the other response on the other thread. The article uses 25mm of polystyrene insulation under the slab so downward heat losses are huge and make the whole article a bit of a joke and at best misinformation.

 

The lowest output they mentioned is a mean flow temp of 35 degC for 47W/m2. Highest is 95W/m3 at 56 degC mean flow temp. All pretty huge, maybe ok for energy rich USA.

[Nic]

On 15/05/2024 at 19:46, JohnMo said:

Not sure what to state is real in reality. Unless your heat losses are huge of course - but then you wouldn't be installing UFH unless you were mad with money to burn.

 

If I needed 70+W/m2 the house would require 13kW to heat, reality is I need 3kW at -9. So why would I design the UFH to give out that much heat. Mines at the bottom of a 100mm concrete pinned to insulation and at 300mm centres. Slow to respond yes, does it make any difference - no, as you keep the slab (60 Tonnes) up to temperature, by low and slow heating or batch heating on E7.

 

forgive me for such a basic question. 

when adding, say 100mm onto of your UFH , does it matter if it’s screed or just concrete? 

[JohnMo]

6 minutes ago, Nic said:

forgive me for such a basic question. 

when adding, say 100mm onto of your UFH , does it matter if it’s screed or just concrete? 

We did concrete, but did it before the walls went up.  Not sure all screeds can be done in a 100mm, so you have to do some investigation. Hopefully someone will be along to answer in more detail.