Does anyone know the regulatory requirement for insulating the stud partition
walls of an attic room?

I'd sorta assume the same for roof spaces - 300mm rockwool equivalent? But
rockwool would be difficult to fix for vertical walls - meaning c.150mm PIR?
That'd be expensive, as well as resulting in uneven insulation where it meets
the pitched section, with only 50mm PIR.

Anyhoo, about to board out a loft space so any advice welcome.
--
Cheers, Rob, Sheffield UK

RJH <patchmoney@gmx.com> wrote:
> Does anyone know the regulatory requirement for insulating the stud
> partition walls of an attic room?
>
> I'd sorta assume the same for roof spaces - 300mm rockwool equivalent?
> But rockwool would be difficult to fix for vertical walls - meaning
> c.150mm PIR? That'd be expensive, as well as resulting in uneven
> insulation where it meets the pitched section, with only 50mm PIR.
>
> Anyhoo, about to board out a loft space so any advice welcome.

https://assets.publishing.service.gov.uk/media/61d727d18fa8f50594b59305/retrofit-room-in-roof-insulation-best-practice.pdf
is worth a read. It says:

"58. Currently, in England, Wales and Northern Ireland a vaulted or flat roof should achieve
an ‘improved’ thermal transmittance (U-value) of 0.18 W/m 2K, and a roof insulated at
ceiling level should achieve an ‘improved’ thermal transmittance (U-value) of 0.16
W/m 2K. Where this is not technically or functionally feasible, then the roofs should be
upgraded to the best U-value possible. In all circumstances, the U-value must be
calculated in accordance with the conventions in the current version of BR443
conventions for calculating U-values.
59. In Scotland a roof should be upgraded to meet a U-value of no worse than 0.35 W/m 2K,
and the area weighted average for all roof elements should be no worse than 0.25
W/m 2K"

It doesn't say anything for room-in-roof walls as such. Meanwhile:
https://insulation4less.co.uk/blogs/guides-and-news/building-regs-u-values
(a rehash of Building Regulations Part L) says:
"For new elements in existing dwellings, the U values will be:

Roofs: 0.15 W/m²K
Walls: 0.18 W/m²K
Floors: 0.18 W/m²K

For existing elements in existing dwellings, the limiting U values will be:

Roofs: 0.16 W/m²K
Walls: 0.30 W/m²K
Floors: 0.25 W/m²K"

so the Part L U-value for new (exterior) walls is roughly the same as a
roof, with more leeway for existing walls. I think 0.18 sounds like a
reasonable target, although there's wiggle room if not achieveable.

I like this site for doing U-value calculations (use the demo version):
https://www.ubakus.de/en/r-value-calculator/

If I make a stackup with:
inside (20C)
12.5mm gypsum board (plasterboard)
80mm spruce studs on 400mm centres
filled with 80mm PIR ('Rigid Foam, PUR, aluminium laminate')
Another layer of PIR
unheated roof space (-5C)

I need 80+80mm of PIR to get to U=0.18. If I switch PIR for 'Rockwool,
stone wool' I need 80+150mm to get to U=0.18.

Theo

Theo <theom+news@chiark.greenend.org.uk> wrote:
> I need 80+80mm of PIR to get to U=0.18. If I switch PIR for 'Rockwool,
> stone wool' I need 80+150mm to get to U=0.18.

Also, about fixing Rockwool to vertical surfaces: in my loft conversion
there are fibreglass batts between studs, they're held in with string which
is stapled to the studs, a bit like a wire fence. I suppose you could do a
second layer if you had something to attach a new set of wires to.

For what I've got I'm not super happy about the arrangement, because the
existing batts are cut too small and there's a gap between the batt and the
stud, ie a chimney where air can convect. Also, I'd like to have a foil
barrier to reflect heat. However the current thinking is to add foiled PIR
between the rafters to make the loft a conditioned space, making it more
useful for storage etc.

Theo

On 3 Dec 2023 at 13:13:48 GMT, Theo wrote:

> RJH <patchmoney@gmx.com> wrote:
>> Does anyone know the regulatory requirement for insulating the stud
>> partition walls of an attic room?
>>
>> I'd sorta assume the same for roof spaces - 300mm rockwool equivalent?
>> But rockwool would be difficult to fix for vertical walls - meaning
>> c.150mm PIR? That'd be expensive, as well as resulting in uneven
>> insulation where it meets the pitched section, with only 50mm PIR.
>>
>> Anyhoo, about to board out a loft space so any advice welcome.
>
> https://assets.publishing.service.gov.uk/media/61d727d18fa8f50594b59305/retrofit-room-in-roof-insulation-best-practice.pdf
> is worth a read. It says:
>
> "58. Currently, in England, Wales and Northern Ireland a vaulted or flat roof
> should achieve
> an ‘improved’ thermal transmittance (U-value) of 0.18 W/m 2K, and a roof
> insulated at
> ceiling level should achieve an ‘improved’ thermal transmittance (U-value) of
> 0.16
> W/m 2K. Where this is not technically or functionally feasible, then the roofs
> should be
> upgraded to the best U-value possible. In all circumstances, the U-value must
> be
> calculated in accordance with the conventions in the current version of BR443
> conventions for calculating U-values.
> 59. In Scotland a roof should be upgraded to meet a U-value of no worse than
> 0.35 W/m 2K,
> and the area weighted average for all roof elements should be no worse than
> 0.25
> W/m 2K"
>
> It doesn't say anything for room-in-roof walls as such. Meanwhile:
> https://insulation4less.co.uk/blogs/guides-and-news/building-regs-u-values
> (a rehash of Building Regulations Part L) says:
> "For new elements in existing dwellings, the U values will be:
>
> Roofs: 0.15 W/m²K
> Walls: 0.18 W/m²K
> Floors: 0.18 W/m²K
>
> For existing elements in existing dwellings, the limiting U values will be:
>
> Roofs: 0.16 W/m²K
> Walls: 0.30 W/m²K
> Floors: 0.25 W/m²K"
>
> so the Part L U-value for new (exterior) walls is roughly the same as a
> roof, with more leeway for existing walls. I think 0.18 sounds like a
> reasonable target, although there's wiggle room if not achieveable.
>
>
> I like this site for doing U-value calculations (use the demo version):
> https://www.ubakus.de/en/r-value-calculator/
>
> If I make a stackup with:
> inside (20C)
> 12.5mm gypsum board (plasterboard)
> 80mm spruce studs on 400mm centres
> filled with 80mm PIR ('Rigid Foam, PUR, aluminium laminate')
> Another layer of PIR
> unheated roof space (-5C)
>
> I need 80+80mm of PIR to get to U=0.18. If I switch PIR for 'Rockwool,
> stone wool' I need 80+150mm to get to U=0.18.
>
> Theo

Many thanks for that, appreciated, plenty to go on.

At the moment, at least on the bits I can see, there's 75mm rockwool between
the studs, held in place with bits of plasterboard.

--
Cheers, Rob, Sheffield UK

On 3 Dec 2023 at 14:57:14 GMT, Theo wrote:

> Theo <theom+news@chiark.greenend.org.uk> wrote:
>> I need 80+80mm of PIR to get to U=0.18. If I switch PIR for 'Rockwool,
>> stone wool' I need 80+150mm to get to U=0.18.
>
> Also, about fixing Rockwool to vertical surfaces: in my loft conversion
> there are fibreglass batts between studs, they're held in with string which
> is stapled to the studs, a bit like a wire fence. I suppose you could do a
> second layer if you had something to attach a new set of wires to.
>

Yes, it could be done, but not as neat as PIR. I've just ordered a load of PIR
seconds - I'll see what that's like when it arrives before making a decision.

> For what I've got I'm not super happy about the arrangement, because the
> existing batts are cut too small and there's a gap between the batt and the
> stud, ie a chimney where air can convect. Also, I'd like to have a foil
> barrier to reflect heat. However the current thinking is to add foiled PIR
> between the rafters to make the loft a conditioned space, making it more
> useful for storage etc.
>

I've seen the PIR between rafters in the roof void on a few re-roofs I've
watched. But I really don't get the point - surely heat will be lost around
the soffits etc? Added to the fact roofers seem to use a rough interference
fit, which must leave gaps.

--
Cheers, Rob, Sheffield UK

On 03/12/2023 11:32, RJH wrote:
> Does anyone know the regulatory requirement for insulating the stud partition
> walls of an attic room?

I take it you mean exterior stud walls rather than partition?

If so 0.18 W/m^2K

> I'd sorta assume the same for roof spaces - 300mm rockwool equivalent? But
> rockwool would be difficult to fix for vertical walls - meaning c.150mm PIR?
> That'd be expensive, as well as resulting in uneven insulation where it meets
> the pitched section, with only 50mm PIR.

PIR in the space between studs and then a thinner layer over the lot.

Here is how I did it...

https://wiki.diyfaq.org.uk/index.php/DIY_Loft_conversion#Insulation

(note that was 20 years ago - u value requirements are a little tighter now)

--
Cheers,

John.

/=================================================================\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\=================================================================/

On 3 Dec 2023 at 19:52:46 GMT, John Rumm wrote:

> On 03/12/2023 11:32, RJH wrote:
>> Does anyone know the regulatory requirement for insulating the stud partition
>> walls of an attic room?
>
> I take it you mean exterior stud walls rather than partition?
>

Yep.

> If so 0.18 W/m^2K

Thanks.
>
>> I'd sorta assume the same for roof spaces - 300mm rockwool equivalent? But
>> rockwool would be difficult to fix for vertical walls - meaning c.150mm PIR?
>> That'd be expensive, as well as resulting in uneven insulation where it meets
>> the pitched section, with only 50mm PIR.
>
> PIR in the space between studs and then a thinner layer over the lot.
>
> Here is how I did it...
>
> https://wiki.diyfaq.org.uk/index.php/DIY_Loft_conversion#Insulation
>
> (note that was 20 years ago - u value requirements are a little tighter now)

Yes, exactly that. PIR has pretty much doubled in price since. I got some
grade B at £10 per 25mm sheet locally. Not sure yet what that entails, but the
whole thing is nuts. The payback at retail prices would be decades. For such a
simple retrofit fix it should, at the very least, be VAT-free.
--
Cheers, Rob, Sheffield UK

On 12/3/2023 5:38 PM, RJH wrote:
> On 3 Dec 2023 at 19:52:46 GMT, John Rumm wrote:
>
>> On 03/12/2023 11:32, RJH wrote:
>>> Does anyone know the regulatory requirement for insulating the stud partition
>>> walls of an attic room?
>>
>> I take it you mean exterior stud walls rather than partition?
>>
>
> Yep.
>
>> If so 0.18 W/m^2K
>
> Thanks.
>>
>>> I'd sorta assume the same for roof spaces - 300mm rockwool equivalent? But
>>> rockwool would be difficult to fix for vertical walls - meaning c.150mm PIR?
>>> That'd be expensive, as well as resulting in uneven insulation where it meets
>>> the pitched section, with only 50mm PIR.
>>
>> PIR in the space between studs and then a thinner layer over the lot.
>>
>> Here is how I did it...
>>
>> https://wiki.diyfaq.org.uk/index.php/DIY_Loft_conversion#Insulation
>>
>> (note that was 20 years ago - u value requirements are a little tighter now)
>
> Yes, exactly that. PIR has pretty much doubled in price since. I got some
> grade B at £10 per 25mm sheet locally. Not sure yet what that entails, but the
> whole thing is nuts. The payback at retail prices would be decades. For such a
> simple retrofit fix it should, at the very least, be VAT-free.
>

https://insulationgo.co.uk/blog/insulation-cost-price-expensive/

"WHY IS PIR INSULATION EXPENSIVE?

Some may wonder why PIR boards are expensive. The answer is quite simple.
The extraction of basic raw materials for PIR board production is costly,
resulting in a higher price for the final product."

"the best time to buy PIR boards appears to be in November and December." <=== Then you check what month,
the web page was updated :-)

But of course, material cost is not the only factor. If a market isn't
competitive, you don't get the best price either. There are lots of
ways for suppliers to fix prices. They fixed the price of bread here,
having previously succeeded at it, one hundred years ago. Price fixing
is nothing new.

Imagine for example, if BEV cars came without a battery pack, and you
shopped the open market to get one. Your car would be "damn cheap" then.
But in the long run, if you arranged things that way, you would be
very very sorry. The illusion of success would be transitory (one company
would end up making all the batteries, and it would be a monopoly just
as the battery industry has achieved on previous occasions).

https://en.wikipedia.org/wiki/Polyisocyanurate

PIR typically has an MDI/polyol ratio higher than 180.

https://en.wikipedia.org/wiki/Methylene_diphenyl_diisocyanate

As of 2019, the largest producer was Wanhua Chemical Group.[5]
Other major producers are Covestro,[6] BASF, Dow, Huntsman, Tosoh,
Kumho Mitsui Chemicals. All major producers of MDI are members of
the International Isocyanate Institute,

You would look for monkey business then, among that set
of companies. Did someone cut off supply, restrict exports ? Etc.
It's a chemical, and I don't think you "mine it".

Paul

Paul <nospam@needed.invalid> wrote:
>
> https://insulationgo.co.uk/blog/insulation-cost-price-expensive/
>
> "WHY IS PIR INSULATION EXPENSIVE?
>
> Some may wonder why PIR boards are expensive. The answer is quite simple.
> The extraction of basic raw materials for PIR board production is costly,
> resulting in a higher price for the final product."
>
> "the best time to buy PIR boards appears to be in November and December." <=== Then you check what month,
> the web page was updated :-)

I'd have thought the raw materials come from oil, so the extraction isn't
that costly. I don't know how difficult the chemical processes used to make
them are, but the boards are lightweight so there isn't a lot of material in
each one.

> But of course, material cost is not the only factor. If a market isn't
> competitive, you don't get the best price either. There are lots of
> ways for suppliers to fix prices. They fixed the price of bread here,
> having previously succeeded at it, one hundred years ago. Price fixing
> is nothing new.

Also transport costs. You don't get too many boards to a truck and they
don't exactly pack down like Ikea flatpacks. There are only a few plants so
there's a lot of transporting involved from them.

An alternative option would be spray foam, where you make the insulation
foam on site from the liquids. But that has quality control issues, and the
fact that rigid insulation boards are nicer to work with than the sticky
spray foam.

Theo

RJH <patchmoney@gmx.com> wrote:
> On 3 Dec 2023 at 14:57:14 GMT, Theo wrote:
>
> > Theo <theom+news@chiark.greenend.org.uk> wrote:
> >> I need 80+80mm of PIR to get to U=0.18. If I switch PIR for 'Rockwool,
> >> stone wool' I need 80+150mm to get to U=0.18.
> >
> > Also, about fixing Rockwool to vertical surfaces: in my loft conversion
> > there are fibreglass batts between studs, they're held in with string which
> > is stapled to the studs, a bit like a wire fence. I suppose you could do a
> > second layer if you had something to attach a new set of wires to.
> >
>
> Yes, it could be done, but not as neat as PIR. I've just ordered a load of PIR
> seconds - I'll see what that's like when it arrives before making a decision.

Depends... rockwool is easier to shove into gaps, OTOH PIR gaps can be
foamed to hold it in. In the US they insulate stud walls with rockwool all
the time.

Another option would be rockwool between the studs (easier to fill out the
gaps) and then PIR boarded over the top. The challenge there is avoid gaps
that form chimneys behind the PIR.

> > For what I've got I'm not super happy about the arrangement, because the
> > existing batts are cut too small and there's a gap between the batt and the
> > stud, ie a chimney where air can convect. Also, I'd like to have a foil
> > barrier to reflect heat. However the current thinking is to add foiled PIR
> > between the rafters to make the loft a conditioned space, making it more
> > useful for storage etc.
> >
>
> I've seen the PIR between rafters in the roof void on a few re-roofs I've
> watched. But I really don't get the point - surely heat will be lost around
> the soffits etc? Added to the fact roofers seem to use a rough interference
> fit, which must leave gaps.

You'd insulate the soffits too. To allow the rafters to breathe you'd need
to leave an air gap between the PIR and the tiles, and that gap is vented
out the soffits and ideally ridge ventilation too.

The problem with an interference fit is that I doubt the gaps between joists
or rafters are square or even, so cutting the foam to interference fit is
awkward. The alternative is leaving a gap and expanding foaming it, but
that makes it harder to remove if you ever need access afterwards.

Theo

On 03/12/2023 22:38, RJH wrote:
> On 3 Dec 2023 at 19:52:46 GMT, John Rumm wrote:
>
>> On 03/12/2023 11:32, RJH wrote:
>>> Does anyone know the regulatory requirement for insulating the stud partition
>>> walls of an attic room?
>>
>> I take it you mean exterior stud walls rather than partition?
>>
>
> Yep.
>
>> If so 0.18 W/m^2K
>
> Thanks.
>>
>>> I'd sorta assume the same for roof spaces - 300mm rockwool equivalent? But
>>> rockwool would be difficult to fix for vertical walls - meaning c.150mm PIR?
>>> That'd be expensive, as well as resulting in uneven insulation where it meets
>>> the pitched section, with only 50mm PIR.
>>
>> PIR in the space between studs and then a thinner layer over the lot.
>>
>> Here is how I did it...
>>
>> https://wiki.diyfaq.org.uk/index.php/DIY_Loft_conversion#Insulation
>>
>> (note that was 20 years ago - u value requirements are a little tighter now)
>
> Yes, exactly that. PIR has pretty much doubled in price since.

OOI I looked at what I paid in 2003, it was £670 (today that would be
£1,171).

(note that I got mine at a significant discount from the builder's
merchant price from a local supplier of seconds - the BM wanted ~£2,200
- I expect there is still a similar variation depending on what you get
and form where)

> I got some
> grade B at £10 per 25mm sheet locally. Not sure yet what that entails, but the
> whole thing is nuts. The payback at retail prices would be decades. For such a
> simple retrofit fix it should, at the very least, be VAT-free.

Payback will depend a lot on the starting point. With any insulation
each doubling in thickness (and hence cost) only gives half as much
again in savings.

--
Cheers,

John.

/=================================================================\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\=================================================================/

On 4 Dec 2023 at 11:32:00 GMT, Theo wrote:

> RJH <patchmoney@gmx.com> wrote:
>> On 3 Dec 2023 at 14:57:14 GMT, Theo wrote:
>>
>>> Theo <theom+news@chiark.greenend.org.uk> wrote:
>>>> I need 80+80mm of PIR to get to U=0.18. If I switch PIR for 'Rockwool,
>>>> stone wool' I need 80+150mm to get to U=0.18.
>>>
>>> Also, about fixing Rockwool to vertical surfaces: in my loft conversion
>>> there are fibreglass batts between studs, they're held in with string which
>>> is stapled to the studs, a bit like a wire fence. I suppose you could do a
>>> second layer if you had something to attach a new set of wires to.
>>>
>>
>> Yes, it could be done, but not as neat as PIR. I've just ordered a load of PIR
>> seconds - I'll see what that's like when it arrives before making a decision.
>
> Depends... rockwool is easier to shove into gaps, OTOH PIR gaps can be
> foamed to hold it in. In the US they insulate stud walls with rockwool all
> the time.
>
> Another option would be rockwool between the studs (easier to fill out the
> gaps) and then PIR boarded over the top. The challenge there is avoid gaps
> that form chimneys behind the PIR.
>

Yes, I'd do that - if only to reuse the rockwool.

>>> For what I've got I'm not super happy about the arrangement, because the
>>> existing batts are cut too small and there's a gap between the batt and the
>>> stud, ie a chimney where air can convect. Also, I'd like to have a foil
>>> barrier to reflect heat. However the current thinking is to add foiled PIR
>>> between the rafters to make the loft a conditioned space, making it more
>>> useful for storage etc.
>>>
>>
>> I've seen the PIR between rafters in the roof void on a few re-roofs I've
>> watched. But I really don't get the point - surely heat will be lost around
>> the soffits etc? Added to the fact roofers seem to use a rough interference
>> fit, which must leave gaps.
>
> You'd insulate the soffits too. To allow the rafters to breathe you'd need
> to leave an air gap between the PIR and the tiles, and that gap is vented
> out the soffits and ideally ridge ventilation too.
>

Still don't get it! The rafters I understand. But what's the point of
insulating an under-roof area open to the outside? The under-roof void,
beneath the rafters, is always going to need fresh air passing through. In the
properties I'm talking about (Victorian terraces) there aren't soffits as I
understand the term - it's just a wall plate; the rafters rest on that and
ventilation works by entering the roof void between the roof and the wall
plate (onto which the gutter is fixed).

And another thing(!) - what's the point of the permeable membrane the roofers
use over the PIR? There's no way I can see that moist air is going to get
through that. Anyhoo. I'm sure I'm missing something as I've seen this method
used on 3 roofs.

> The problem with an interference fit is that I doubt the gaps between joists
> or rafters are square or even, so cutting the foam to interference fit is
> awkward. The alternative is leaving a gap and expanding foaming it, but
> that makes it harder to remove if you ever need access afterwards.
>

Yes, been there, with some very irregular joists/rafters - so plenty of foam
filling. The roofers I watched seemed to simply chuck in the PIR - and the
ineffectiveness of it all was demonstrated every winter as their new roofs
were the first to thaw snow - albeit not knowing the internal temperatures.

--
Cheers, Rob, Sheffield UK

On 05/12/2023 00:29, RJH wrote:
<snip>
> Still don't get it! The rafters I understand. But what's the point of
> insulating an under-roof area open to the outside? The under-roof void,
> beneath the rafters, is always going to need fresh air passing through. In the
> properties I'm talking about (Victorian terraces) there aren't soffits as I
> understand the term - it's just a wall plate; the rafters rest on that and
> ventilation works by entering the roof void between the roof and the wall
> plate (onto which the gutter is fixed).

Sounds like a "warm roof" with insulation above the rafters and no air
flow below them.

> And another thing(!) - what's the point of the permeable membrane the roofers
> use over the PIR? There's no way I can see that moist air is going to get
> through that. Anyhoo. I'm sure I'm missing something as I've seen this method
> used on 3 roofs.

A breathable membrane allows any moisture on the PIR (including
condensation) to escape out rather than in.

>> The problem with an interference fit is that I doubt the gaps between joists
>> or rafters are square or even, so cutting the foam to interference fit is
>> awkward. The alternative is leaving a gap and expanding foaming it, but
>> that makes it harder to remove if you ever need access afterwards.
>>
>
> Yes, been there, with some very irregular joists/rafters - so plenty of foam
> filling. The roofers I watched seemed to simply chuck in the PIR - and the
> ineffectiveness of it all was demonstrated every winter as their new roofs
> were the first to thaw snow - albeit not knowing the internal temperatures.
>

--
Robin
reply-to address is (intended to be) valid

On 04/12/2023 11:32, Theo wrote:
> RJH <patchmoney@gmx.com> wrote:
>> On 3 Dec 2023 at 14:57:14 GMT, Theo wrote:
>>
>>> Theo <theom+news@chiark.greenend.org.uk> wrote:
>>>> I need 80+80mm of PIR to get to U=0.18. If I switch PIR for 'Rockwool,
>>>> stone wool' I need 80+150mm to get to U=0.18.
>>>
>>> Also, about fixing Rockwool to vertical surfaces: in my loft conversion
>>> there are fibreglass batts between studs, they're held in with string which
>>> is stapled to the studs, a bit like a wire fence. I suppose you could do a
>>> second layer if you had something to attach a new set of wires to.
>>>
>>
>> Yes, it could be done, but not as neat as PIR. I've just ordered a load of PIR
>> seconds - I'll see what that's like when it arrives before making a decision.
>
> Depends... rockwool is easier to shove into gaps, OTOH PIR gaps can be
> foamed to hold it in. In the US they insulate stud walls with rockwool all
> the time.
>
> Another option would be rockwool between the studs (easier to fill out the
> gaps) and then PIR boarded over the top. The challenge there is avoid gaps
> that form chimneys behind the PIR.
>
>>> For what I've got I'm not super happy about the arrangement, because the
>>> existing batts are cut too small and there's a gap between the batt and the
>>> stud, ie a chimney where air can convect. Also, I'd like to have a foil
>>> barrier to reflect heat. However the current thinking is to add foiled PIR
>>> between the rafters to make the loft a conditioned space, making it more
>>> useful for storage etc.
>>>
>>
>> I've seen the PIR between rafters in the roof void on a few re-roofs I've
>> watched. But I really don't get the point - surely heat will be lost around
>> the soffits etc? Added to the fact roofers seem to use a rough interference
>> fit, which must leave gaps.
>
> You'd insulate the soffits too. To allow the rafters to breathe you'd need
> to leave an air gap between the PIR and the tiles, and that gap is vented
> out the soffits and ideally ridge ventilation too.

There are a number of factors at play here. If the tiled roof has a
breathable sarking, then you can full fill the depth of the rafters with
insulation. If it is an older impermeable sarking then you normally
leave a ventilation gap between the insulation the the tiles. Then have
vents at the ridge and the soffit. So you have an envelope of insulation
that isolates the inside of the building, but there is an air passage on
the potentially damp side of it.

> The problem with an interference fit is that I doubt the gaps between joists
> or rafters are square or even, so cutting the foam to interference fit is
> awkward. The alternative is leaving a gap and expanding foaming it, but
> that makes it harder to remove if you ever need access afterwards.

The modern solution are products like gappotape. They fit to the edges
of the boards and provide enough "wiggle room" to make a proper full
depth air tight seal even of the board cutting accuracy is less than
perfect. They also mean in many cases you can precut all the boards
without needing to adjust the fit to cope with slight variations in
rafter gaps etc, since they can take up 20mm of variation.

--
Cheers,

John.

/=================================================================\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\=================================================================/

RJH <patchmoney@gmx.com> wrote:
> On 4 Dec 2023 at 11:32:00 GMT, Theo wrote:
> > You'd insulate the soffits too. To allow the rafters to breathe you'd need
> > to leave an air gap between the PIR and the tiles, and that gap is vented
> > out the soffits and ideally ridge ventilation too.
> >
>
> Still don't get it! The rafters I understand. But what's the point of
> insulating an under-roof area open to the outside? The under-roof void,
> beneath the rafters, is always going to need fresh air passing through. In the
> properties I'm talking about (Victorian terraces) there aren't soffits as I
> understand the term - it's just a wall plate; the rafters rest on that and
> ventilation works by entering the roof void between the roof and the wall
> plate (onto which the gutter is fixed).

Here's a picture:
https://www.diychatroom.com/attachments/vent-gif.541089/

Your setup is the left side. You insulate the 'kneewalls' and between the
floor joists, so the loft is at outside air temperature and ventilated.

My proposal is the right side. You insulate between the rafters so the part
behind the kneewalls becomes conditioned (heated) space. That means you can
store stuff in there without it getting damp/rusty/mildewy/etc.

To do so, you create a ventilation path behind the insulation so the rafters
are exposed to outside air and moisture/condensation in the timber can
escape. For example a 50mm gap is traditional. This gap is ventilated top
(ridge ventilation) and bottom (soffit vent). You join the rafter
insulation with the wall insulation so you have a continuous blanket.

How to do the joint is a question that depends on your wall construction,
but if it's a cavity wall then presumably you want to join your rafter
insulation with your cavity fill so there's no air gap. There is a
complication that the cavity needs some way to breathe to allow moisture
out, which may affect the choice of insulation in that spot (rockwool,
sheeps wool and wood fibre insulation are porous to some degree, PIR isn't).

> And another thing(!) - what's the point of the permeable membrane the roofers
> use over the PIR? There's no way I can see that moist air is going to get
> through that. Anyhoo. I'm sure I'm missing something as I've seen this method
> used on 3 roofs.

A permeable membrane allows the roof timbers to breathe while preventing
wind and driven rain water getting at them.

In short, a building enclosure design needs 4 layers:
1. Water control layer
2. Air control layer
3. Vapour control layer
4. Thermal control layer

That's in order of precedence - if water can get in, don't bother trying to
control anything else.

In such a roof we have:
Tiles - rain control
Permeable membrane - air control
PIR - thermal control
Foil backing - vapour control

For more info on the theory see:
https://buildingscience.com/documents/insights/bsi-001-the-perfect-wall

> > The problem with an interference fit is that I doubt the gaps between joists
> > or rafters are square or even, so cutting the foam to interference fit is
> > awkward. The alternative is leaving a gap and expanding foaming it, but
> > that makes it harder to remove if you ever need access afterwards.
> >
>
> Yes, been there, with some very irregular joists/rafters - so plenty of foam
> filling. The roofers I watched seemed to simply chuck in the PIR - and the
> ineffectiveness of it all was demonstrated every winter as their new roofs
> were the first to thaw snow - albeit not knowing the internal temperatures.

If the roof is a habitable space, the roof surface is say 150mm of stuff
away from the heated space. If the roof isn't habitable, most of the
structure is several metres of ventilated air away from the heated room
below - the heat will be lost out of the vents, not conducted through the
tiles. So it isn't necessarily a good guide to look at the snow melt - it's
only if you're comparing equivalently constructed roofs.

Theo

John Rumm <see.my.signature@nowhere.null> wrote:
> There are a number of factors at play here. If the tiled roof has a
> breathable sarking, then you can full fill the depth of the rafters with
> insulation. If it is an older impermeable sarking then you normally
> leave a ventilation gap between the insulation the the tiles. Then have
> vents at the ridge and the soffit. So you have an envelope of insulation
> that isolates the inside of the building, but there is an air passage on
> the potentially damp side of it.

Indeed, the key here is the ventilation is for the roof timbers. If there
is condensation it needs to be able to get out, otherwise the timbers will
stay damp and rot. Traditionally the roof space was ventilated so the
timbers were exposed to outside air and rot wasn't a problem (unless it
leaked, when you had bigger problems). If you start insulating you need to
make sure that moisture can still get out. Wood is porous, so as long as
some part of the timber has a breatheable contact with the outside air then
that's enough.

> > The problem with an interference fit is that I doubt the gaps between joists
> > or rafters are square or even, so cutting the foam to interference fit is
> > awkward. The alternative is leaving a gap and expanding foaming it, but
> > that makes it harder to remove if you ever need access afterwards.
>
> The modern solution are products like gappotape. They fit to the edges
> of the boards and provide enough "wiggle room" to make a proper full
> depth air tight seal even of the board cutting accuracy is less than
> perfect. They also mean in many cases you can precut all the boards
> without needing to adjust the fit to cope with slight variations in
> rafter gaps etc, since they can take up 20mm of variation.

Gapotape helps to an extent, but more of a problem is that older hand-built
roofs may be much more uneven than that - eg you could have 350-450mm
joist/rafter spacing, which could vary from one side to the other (joists
not square to each other). This makes it a PITA to cut insulation because
you can't just cut a stack of 400mm, you have to measure every single gap
and cut accordingly.

Theo

On 05/12/2023 13:13, Theo wrote:
> John Rumm <see.my.signature@nowhere.null> wrote:
>> There are a number of factors at play here. If the tiled roof has a
>> breathable sarking, then you can full fill the depth of the rafters with
>> insulation. If it is an older impermeable sarking then you normally
>> leave a ventilation gap between the insulation the the tiles. Then have
>> vents at the ridge and the soffit. So you have an envelope of insulation
>> that isolates the inside of the building, but there is an air passage on
>> the potentially damp side of it.
>
> Indeed, the key here is the ventilation is for the roof timbers. If there
> is condensation it needs to be able to get out, otherwise the timbers will
> stay damp and rot. Traditionally the roof space was ventilated so the
> timbers were exposed to outside air and rot wasn't a problem (unless it
> leaked, when you had bigger problems). If you start insulating you need to
> make sure that moisture can still get out. Wood is porous, so as long as
> some part of the timber has a breatheable contact with the outside air then
> that's enough.
>
>>> The problem with an interference fit is that I doubt the gaps between joists
>>> or rafters are square or even, so cutting the foam to interference fit is
>>> awkward. The alternative is leaving a gap and expanding foaming it, but
>>> that makes it harder to remove if you ever need access afterwards.
>>
>> The modern solution are products like gappotape. They fit to the edges
>> of the boards and provide enough "wiggle room" to make a proper full
>> depth air tight seal even of the board cutting accuracy is less than
>> perfect. They also mean in many cases you can precut all the boards
>> without needing to adjust the fit to cope with slight variations in
>> rafter gaps etc, since they can take up 20mm of variation.
>
> Gapotape helps to an extent, but more of a problem is that older hand-built
> roofs may be much more uneven than that - eg you could have 350-450mm
> joist/rafter spacing, which could vary from one side to the other (joists
> not square to each other). This makes it a PITA to cut insulation because
> you can't just cut a stack of 400mm, you have to measure every single gap
> and cut accordingly.

Yup it can be a PITA job. If you have the head room, then less
insulation between the studs/rafters, and a thicker layer on the warm
side is easier to fit.

--
Cheers,

John.

/=================================================================\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\=================================================================/

On 5 Dec 2023 at 12:59:14 GMT, Theo wrote:

> RJH <patchmoney@gmx.com> wrote:
>> On 4 Dec 2023 at 11:32:00 GMT, Theo wrote:
>>> You'd insulate the soffits too. To allow the rafters to breathe you'd need
>>> to leave an air gap between the PIR and the tiles, and that gap is vented
>>> out the soffits and ideally ridge ventilation too.
>>>
>>
>> Still don't get it! The rafters I understand. But what's the point of
>> insulating an under-roof area open to the outside? The under-roof void,
>> beneath the rafters, is always going to need fresh air passing through. In the
>> properties I'm talking about (Victorian terraces) there aren't soffits as I
>> understand the term - it's just a wall plate; the rafters rest on that and
>> ventilation works by entering the roof void between the roof and the wall
>> plate (onto which the gutter is fixed).
>
> Here's a picture:
> https://www.diychatroom.com/attachments/vent-gif.541089/
>
> Your setup is the left side. You insulate the 'kneewalls' and between the
> floor joists, so the loft is at outside air temperature and ventilated.
>

Ah yes - thanks - and for the remaining explanation snipped here. I'm actually
talking about directly below the apex - where Gable Vent is shown. That
roofspace was membraned/PIRd.
--
Cheers, Rob, Sheffield UK

RJH <patchmoney@gmx.com> wrote:
> On 5 Dec 2023 at 12:59:14 GMT, Theo wrote:
>
> > Here's a picture:
> > https://www.diychatroom.com/attachments/vent-gif.541089/
> >
> > Your setup is the left side. You insulate the 'kneewalls' and between the
> > floor joists, so the loft is at outside air temperature and ventilated.
> >
>
> Ah yes - thanks - and for the remaining explanation snipped here. I'm actually
> talking about directly below the apex - where Gable Vent is shown. That
> roofspace was membraned/PIRd.

For the top, they have it with an open space that's ventilated at the gable,
ie outside air can flow from the gable through to all the truss bays. That
obviously only works if you have a gable, so is no good in a terrace.

The alternative is not having airflow between bays, but having ridge vents
instead. Then you get airflow from soffit through to ridge.

As also mentioned, if you have a permeable sarking membrane then the timbers
can breathe through the membrane, so you don't need a separate air space.

Theo