Suspended floor insulation

[jacko555]

Sorry for flogging a dead horse, however...

I have a suspended wood ground floor and will shortly have parquet laid on ply, around 50m2

Underfloor is 4x2 with about 4 inch to the subfloor, so could fit a max of 4inch insulation.

I had the idea to insulate and discussed it here after reading an article saying "netting and loft insulation is the wrong way to do it" https://www.diynot.com/diy/threads/floor-insulation-article.601960/

Using a few u value calculators, e.g. (https://www.rockwool.com/uk/resources-and-tools/tools/u-value-calculator/), and a p/m ratio of 0.36, 4inch of rockwool will get me a u value of around 0.03 Watt/mK

There's also the pain of pulling up the floor, compromising on the insulation thickness around the power cables etc.

A different thread https://www.diynot.com/diy/threads/10-or-12mm-pir-insulation-board.601978/ got me thinking.

10mm STS tile backer board https://www.sts-uk.com/tiling-preparation/professional-tile-backing-board is designed for flooring and has a thermal conductivity 0.034 Watt/mK.

Similar boards with the same u value are used for underfloor heating.

So if I am happy with gaining 10mm of floor height, laying these, then the ply will give the same u value as 100mm of rockwool.

Did I work that out correctly?

Is this a valid method for insulating the floor?

What am I missing

 

[John D v2.0]

Conductivity (r value) is the inverse of the u value.
The R value is used for individual components as you can add them up easily based on the thickness. Then you calculate the u value for the whole buildup because that's the useful number.
To match 100mm of mineral wool with 10mm thickness you're talking some kind of space age insulation! Normal celotex would be 50mm roughly.

 

[jacko555]

Needed a sanity check, thank you.
Rereading the specs, while the foam has that value, the board is a u value of 2.62.
Kinda useless then.

Nuts. Looks like the floor is coming up

 

[omph]

10mm STS has an R-Value of 0.29 and a u-value of 3.4.

You could check out foil based insulation. It relies on reflection of heat back into the house and needs a reasonably sized air gap making it very well suited for suspended floors, and if done properly can bring the u-value all the way down to 0.12. See https://www.superquilt-insulation.co.uk/solutions/suspended-timber-floor-ybs-superquilt/

 

[oldbutnotdead]

Needed a sanity check, thank you.
Rereading the specs, while the foam has that value, the board is a u value of 2.62.
Kinda useless then.

Nuts. Looks like the floor is coming up

Yeah. Rockwool isn't as efficient as correctly installed pir but it's very difficult to correctly install PIR in old woodwork- twisted joists, uneven centres, nothing square makes it hard to achieve airtight fully filled compartments. Rockwool on the other hand is very easy to install correctly and is much more cost efficient (far less wastage, cost for the volume required is between half and a third of cost of pir to achieve the same u value).

 

[jeds]

I'm very suspicious of foil insulation. They are fond of quoting things like; 'a u-value of 0.12 can be achieved' etc. But to achieve that the P/A ratio is unrealistic - 50/500 for example. A more realistic P/A is 1 - which would give you a u-value more like 0.27. And I'm still suspicious of that.

 

[jacko555]

Also unconvinced. Rockwool it is

 

[jeds]

As a basic guide; PIR is closely approximate to being double the value of standard mineral wool. i.e. 50mm of mineral wool is equivalent to 25mm of PIR. 100mm of mineral wool is about the same as 50mm PIR. I say 'standard' mineral; there are different values. Standard is about 0.039 - the best might be 0.032, which alters that ratio slightly.

 

[jeds]

PS, whatever you do, make sure you don't cut off the ventilation to the void.

 

[omph]

I'm very suspicious of foil insulation. They are fond of quoting things like; 'a u-value of 0.12 can be achieved' etc. But to achieve that the P/A ratio is unrealistic - 50/500 for example. A more realistic P/A is 1 - which would give you a u-value more like 0.27. And I'm still suspicious of that.

A P/A of 1 is still very good., and the same principle is also used on glazed windows - Low E coats do practically the same thing and reflect about 20-30% of the heat back into the house which plays a major role in getting lower u-values for double glazing. The physics does work, and the same principle is used in space to retain heat and deflect external radiant energy.

 

[jeds]

A P/A of 1 is still very good.

P/A is neither good nor bad, it is just the ratio of perimeter to area and is what it is. The point is that a u-value of 0.12 for foil is only applicable to a P/A of 0.1, which is unrealistic. You won't very often, if ever, come across a P/A of 0.1, especially in housing. Therefore foil manufacturers are presenting a false scenario to make the product look better than it is, and if they need to do that to sell the product it tells you something.

The problem with foil is not with the physics, the physics of radiation and emissivity is well understood. The problem is that the physics rarely relates to actual use. In order for a foil to reduce surface radiation it must not be touching anything - and the gap is critical to the effectiveness. With several layers of foil the inner foils are obviously touching each other, and therefore are working only to reduce conduction, not radiation. Also, the foil must be attached in some way, and every attachment eliminates the radiation process. Therefore, to be effective, the foil must be suspended in air, with an accurate and minimum free gap and minimally attached to anything. This won't and doesn't happen in 90% of real cases.

 

[omph]

P/A is neither good nor bad, it is just the ratio of perimeter to area and is what it is. The point is that a u-value of 0.12 for foil is only applicable to a P/A of 0.1, which is unrealistic. You won't very often, if ever, come across a P/A of 0.1, especially in housing. Therefore foil manufacturers are presenting a false scenario to make the product look better than it is, and if they need to do that to sell the product it tells you something.

The problem with foil is not with the physics, the physics of radiation and emissivity is well understood. The problem is that the physics rarely relates to actual use. In order for a foil to reduce surface radiation it must not be touching anything - and the gap is critical to the effectiveness. With several layers of foil the inner foils are obviously touching each other, and therefore are working only to reduce conduction, not radiation. Also, the foil must be attached in some way, and every attachment eliminates the radiation process. Therefore, to be effective, the foil must be suspended in air, with an accurate and minimum free gap and minimally attached to anything. This won't and doesn't happen in 90% of real cases.

All valid points, but I think you are still underplaying it. Conduction is also a part of physics so when engineering solutions you have to factor in all possibe issues, not just focus on radiation or conduction.
Low-E coats suffer from the same conductivity issues but can still reduce heat loss by 20-30% and are pivotal in getting the u values down - The real world performance for such a thin coat is exceptionally good. Conduction was also an issue with glazing, so warm spacers have aided in reducing the effects of that problem.

Also, you would need to calculate the surface area of the joists vs the surface area of the air suspended foil if you want to calculate roughly how much is conducted vs radiated.

If we assume a room of 3m x 3m with an area of 9m^2. 10 Joists of 0.05m x 3m at 0.4m apart = 1.5m^2 - ~83% air.
A 2x4 joist with 20mm ply and 5mm floorboards, and a wood k-value of 0.1154 will give us a u-value of 0.9232. If we assume a temp DT of 23c this will result in a heat loss of about 37.37 watts (127.43BTU) an hour by conduction if the temp of the joists is -3c. (which hardly ever happens in a heated house). That isn't too bad, but to bring it down further a 10mm PIR board laid over the ply with a k-value of 0.022 will bring the u-value down to 0.65 and heat loss of about 22.4w per hour @ -3 by conduction.
We can go out on a limb and add another 20% to the joists area but the loss will still be low, and can easily be mitigated in other ways.

We could tweak our setup by decreasing the thickness of the joist and compensate in length, or space the joists further apart (410-450mm instead of 400) or use thicker PIR.

Now I'm not suggesting we all use foil but many of the issues you describe can be improved and it all depends on the circumstances. In some cases PIR is better, in others its rockwool, and in jobs where removing and insulating is too costly, then foil with thinner insulation overlay may be a more viable solution.