First Floor UFH - Role of Subfloor Insulation

[SanMan]

I have upstairs bedroom with UFH and it is struggling in this weather. Max room temp 18 degrees. The set up is:

- suspended timber floor 

- ProFloor routered boards with spreader plates

- Kahrs Tuplex underlay (recommended for UFH), then Kahrs engineered oak

- subfloor void is insulated with loose fibreglass, avoiding downlights for room below, so somewhat gappy 

- room below is heated kitchen

Yesterday, I removed a downlight and put my hand in the void. It felt very toasty. I could touch the underside of the ProFloor boards and these felt quite warm. 

So my suspicion is that the insulation is insufficient and I am losing a lot of heat downwards into the kitchen ceiling rather than into the room. 

Other info:

- room has insulated cavity walls and new uPVC windows

- bedroom ceiling void is insulated with 120 mm Celotex (cold loft above)

 

The system settings are:

- mixing valve set to 50 degrees

- flow rate 1.4l/min

- pump speed 3

- only circuit on manifold (70 m length pipe around the floor area of the room) 

- manifold only 3 m away from room, approximately 8 m between manifold and condensing gas boiler 

- all pipes to/from manifold lagged 

 

I was told by Building Control that insulation in internal floors between ground and first is not necessary, but as we are talking underfloor heating, I did it anyway. Well, I don’t think it is enough and I think it should have had 100mm rocksilk slabs or similar pushed right up against the underside of the ProFloor boards, to ensure heat travels upwards into the room and not downwards into the floor void. Yes, heat rises, but I think here it goes both ways.
Thankfully, the ceiling below is due to come down and be reboarded/skimmed anyway as the downlights are all in the wrong place - so I can add the insulation from underneath. 
Can I have your views on this before commencing these works? Thank you.

[Olf]

10 hours ago, SanMan said:

ProFloor routered boards with spreader plates

This system seem to rely on good contact of spreader plate with the top of the pipe, if air gap is present you'll have only half of the pipe surface transferring heat, and that into the chipboard - not the best heat conductor. Anyway, it is what it is now.

 

10 hours ago, SanMan said:

so I can add the insulation from underneath

Definitely! The boards seem to receive heat from the pipes, by adding extra insulation you limit the flow downwards - so the only way is going to be up, where you want it.

As a bonus improvement of acoustic properties.

Of course that will reduce what the kitchen receives, will that be a problem?

What is the temperature limit for the surface, are you able to increase flow temperature in that room?

[Nickfromwales]

This will be down to a lack of physical contact between the diffuser and the upper floor. I see this christ knows how many times.

Did you witness this being installed, and do you know if the deck boards are sitting snugly against the UFH system? If there's even a couple of mm gap the results ( unless it was a passive style dwelling ) would be dramatically reduced.

[SanMan]

Thank you for your reply @Olf. 

 

The kitchen is quite a warm room with its own heating - it definitely doesn’t need help from above 😇 (except on the energy price front maybe! 😂). 

 

Floor surface temperature limit is 27. I can increase the mixing valve to 55 but I find a hotter flow might just trigger switch off by the floor limit function sooner - or am I wrong there? 

 

Can you help with this:

 

As the joists are a not dead straight, I would like to use Rocksilk slabs rather than Celotex as the slabs are more forgiving if cut slightly too big. Is it worth going for foil faced slabs? 

And if so, should the foil side be nearest to the floor deck? 

And finally (sorry) - Do I need an air gap between the underside of the floor deck and the slabs or is a close contact fit best? The total joist height in the floor void is 200 mm, so I was thinking add 100 mm slabs, secure from underneath and leave 100 mm space between slab and kitchen ceiling plasterboard to ventilate the floor void. 

 

Thank you so much. 

[SanMan]

Thank you @Nickfromwales. Yes, I saw it installed. The floor is Kahrs engineered floating on the spreader plates. Did not look like lack of contact - the fitter even screwed all the spreader plates down as they would otherwise have been a bit springy and that would have reduced contact. ProFloor don’t ask for this, but another well known manufacturer of the routered board/spreader plate system does.

 

How do you overcome poor contact? ProFloor is sold as suitable for taking a floating engineered wood floor. 

 

Thank you. 

[SanMan]

17 minutes ago, Nickfromwales said:

This will be down to a lack of physical contact between the diffuser and the upper floor. I see this christ knows how many times.

Did you witness this being installed, and do you know if the deck boards are sitting snugly against the UFH system? If there's even a couple of mm gap the results ( unless it was a passive style dwelling ) would be dramatically reduced.

Have replied above. Thanks. 

Edited December 11, 2022 by SanMan

[SanMan]

23 minutes ago, Olf said:

This system seem to rely on good contact of spreader plate with the top of the pipe, if air gap is present you'll have only half of the pipe surface transferring heat, and that into the chipboard - not the best heat conductor. Anyway, it is what it is now.

 

Definitely! The boards seem to receive heat from the pipes, by adding extra insulation you limit the flow downwards - so the only way is going to be up, where you want it.

As a bonus improvement of acoustic properties.

Of course that will reduce what the kitchen receives, will that be a problem?

What is the temperature limit for the surface, are you able to increase flow temperature in that room?

Have replied above. Thanks. 

Edited December 11, 2022 by SanMan

[Nickfromwales]

17 minutes ago, SanMan said:

Thank you @Nickfromwales. Yes, I saw it installed. The floor is Kahrs engineered floating on the spreader plates. Did not look like lack of contact - the fitter even screwed all the spreader plates down as they would otherwise have been a bit springy and that would have reduced contact. ProFloor don’t ask for this, but another well known manufacturer of the routered board/spreader plate system does.

 

How do you overcome poor contact? ProFloor is sold as suitable for taking a floating engineered wood floor. 

 

Thank you. 

Much higher flow temps. Problem being, is you’ll need to raise this incrementally in 2°C lifts until you’re at or around the max permitted temp stated by the flooring manufacturer. That’s usually 27°C. 
Beyond that, it’s failed by poor design / implementation. 

Edited December 11, 2022 by Nickfromwales

[SanMan]

21 minutes ago, Nickfromwales said:

Much higher flow temps. Problem being, is you’ll need to raise this incrementally in 2°C lifts until you’re at or around the max permitted temp stated by the flooring manufacturer. That’s usually 27°C. 
Beyond that, it’s failed by poor design / implementation. 

Thanks. Flow temperature as in temperature selection at mixing valve? Doesn’t this just cause the limiting function to kick in sooner? Limit is set at 27 degrees. The mixing valve is presently set to 50 degrees, but I understand it can be increased to up to 55 for suspended timber floors, provided that you always have a working floor temp probe to protect the floor.

 

I must admit it sounds like this spreader plate contact problem is an inherent drawback of this type of system because truly, I cannot see what could have been done differently. 

 

Thank you so much. 

[Adsibob]

15 hours ago, SanMan said:

I have upstairs bedroom with UFH and it is struggling in this weather. Max room temp 18 degrees.

I don’t see what the problem is. Why would you want a bedroom warmer than this?

[SanMan]

6 hours ago, Adsibob said:

I don’t see what the problem is. Why would you want a bedroom warmer than this?

Thank you for your reply. 

It is my daughter’s room, so in the school holidays especially she spends long periods in there. It is then used more like a living room and as you know they are set slightly warmer than bedrooms. Also, 18 is the point it gets to and then the floor protection probe switches the heating off until the boards have cooled enough. It switches back once the temperature has dropped to nearer 17.5 - so in practice the room is rarely at 18 for all that long, but just under. 

My question is really: is my concern of heat loss into the ceiling void below valid? Or is it not worth upgrading the insulation and directing the heat upward as the room below is heated? Thanks. 

[Adsibob]

1 hour ago, SanMan said:

Thank you for your reply. 

It is my daughter’s room, so in the school holidays especially she spends long periods in there. It is then used more like a living room and as you know they are set slightly warmer than bedrooms. Also, 18 is the point it gets to and then the floor protection probe switches the heating off until the boards have cooled enough. It switches back once the temperature has dropped to nearer 17.5 - so in practice the room is rarely at 18 for all that long, but just under. 

My question is really: is my concern of heat loss into the ceiling void below valid? Or is it not worth upgrading the insulation and directing the heat upward as the room below is heated? Thanks. 

Ah, ok, that makes sense.

Well I guess that if you are feeling warmth underneath the ufh system, then the insulation is likely insufficient. Could you temporarily affix some PIR to an area of the floor/ceiling void and see if it makes any difference, before doing the whole thing?

[Onoff]

13 hours ago, Nickfromwales said:

This will be down to a lack of physical contact between the diffuser and the upper floor. I see this christ knows how many times.

Did you witness this being installed, and do you know if the deck boards are sitting snugly against the UFH system? If there's even a couple of mm gap the results ( unless it was a passive style dwelling ) would be dramatically reduced.

 

So.....in the sketch below, if the black is the aluminium spreader, that has to be butt tight against the underside of the routed chipboard (where the arrow is)?

 

 

 

[Marvin]

Good picture to see detail:

 

https://www.first-traceheating.co.uk/blog/what-floor-type-is-best-for-underfloor-heating

[SanMan]

5 minutes ago, Onoff said:

 

So.....in the sketch below, if the black is the aluminium spreader, that has to be butt tight against the underside of the routed chipboard (where the arrow is)?

 

 

 

Thank you for the reply. 

 

The layers are actually like this (starting from the bottom up - some insulation, routered chipboard, spreader plates, thin Tuplex underlay, Kahrs oak engineered flooring). In other words, the spreader plates are on top of/bedded into the routered chipboard, not the other way round. I attach a photo. The thing I am missing from that photo is a slab of insulation.

 

My question is will adding insulation make much difference, given there is a heated room below. The floor joists are 200 mm tall, so the void is quite spacious. There is some insulation there, but only loose full fibreglass. 

 

And if I add insulation slabs, should they be pushed right up against the chipboard. And should they be foil face up? 

 

Thank you. 😊

[SanMan]

5 minutes ago, Marvin said:

Good picture to see detail:

 

https://www.first-traceheating.co.uk/blog/what-floor-type-is-best-for-underfloor-heating

Thank you. Yes, that is a good picture, but my set up is like this….

[SanMan]

33 minutes ago, Adsibob said:

Ah, ok, that makes sense.

Well I guess that if you are feeling warmth underneath the ufh system, then the insulation is likely insufficient. Could you temporarily affix some PIR to an area of the floor/ceiling void and see if it makes any difference, before doing the whole thing?

Good idea. The ceiling underneath is a mess and has to come down anyway. I wonder if I might just take one sheet down and insert an insulation slab, to see what that does to the surface temp feel of the floor. PIR is a nightmare between our joists because of the age of the house. No two joist bays are the same and as the timber has expanded and contracted over time, the joists are a tiny bit “wriggly” (not 100% flat/straight) - so I was going to use 100 mm Rocksilk slabs as they are more accommodating if cut slightly too large. You know what fun it is to try to cut PIR exactly right! Thank you. 😊

[Marvin]

Yes heat travelling down through chipboard or up through chipboard, underlay and floor covering...

 

If the thermal resistance of these layers is more that the insulation below, and the underside of the routered chipboard is not touching all over the insulation below, then I think lots of heat will go down.

 

Yes Celotex or similar underneath (touching the routered chipboard) is best.

 

Another thing that can cause problems is if the joist ends are not sealed into cavity walls properly and the insulation below has a gap between that and the routered chipboard allowing air to flow...

[Marvin]

10 minutes ago, SanMan said:

Thank you. Yes, that is a good picture, but my set up is like this….

Notice in the diagram the Insulation is fixed up tight to the underside of the floor makeup

 

[SanMan]

4 minutes ago, Marvin said:

Yes heat travelling down through chipboard or up through chipboard, underlay and floor covering...

 

If the thermal resistance of these layers is more that the insulation below, and the underside of the routered chipboard is not touching all over the insulation below, then I think lots of heat will go down.

 

Yes Celotex or similar underneath (touching the routered chipboard) is best.

 

Another thing that can cause problems is if the joist ends are not sealed into cavity walls properly and the insulation below has a gap between that and the routered chipboard allowing air to flow...

That is really helpful. Thank you for your advice. 😊

[SanMan]

3 minutes ago, Marvin said:

Notice in the diagram the Insulation is fixed up tight to the underside of the floor makeup

 

Yes. I was wondering if that is ok. No need for air gap? I guess that would defeat the object, having read your other reply. Thank you. 

[Onoff]

It's very easy to install Celotex wrong. How you cut, i.e how square can have a major impact in how tight it fits against the joist. One way of near guaranteeing at least one side is tight to the joist is to cut 5mm under width. Hold in place with some nails then gun foam the hell out of the gap.

 

 

 

 

[SanMan]

2 minutes ago, SanMan said:

Yes. I was wondering if that is ok. No need for air gap? I guess that would defeat the object, having read your other reply. Thank you. 

And how do you feel about 100 mm Rocksilk? Much easier to fit than Celotex. And would it need to be foil faced like the Celotex? Thank you. 

[SanMan]

Just now, Onoff said:

It's very easy to install Celotex wrong. How you cut, i.e how square can have a major impact in how tight it fits against the joist. One way of near guaranteeing at least one side is tight to the joist is to cut 5mm under width. Hold in place with some nails then gun foam the hell out of the gap.

 

 

 

 

Yes. I read one article once where they said cut it 5 mm LARGER to ensure a good fit. 🤔 The problem is the joists are tapered/uneven in themselves and using foam when working in a ceiling against gravity is no fun. So I thought 100 mm Rocksilk will solve s multitude of fitting issues and that can easily be oversized and will still stay in. 

[SanMan]

Just now, SanMan said:

Yes. I read one article once where they said cut it 5 mm LARGER to ensure a good fit. 🤔 The problem is the joists are slightly tapered/uneven due to their age in both directions and using foam when working in a ceiling against gravity is no fun. So I thought 100 mm Rocksilk will solve s multitude of fitting issues and that can easily be oversized and will still stay in.