Welcome to the forum :)

A vapour barrier near the outside is generally considered a complete no-no, and some don't even like them in their usual place on the inside of a wall. What led you to such a decision? What does condensation analysis show you?

Personally I wouldn't use PIR anywhere, because of the pollutants it can produce in a fire, but each to their own.

Thank you for the feedback :)

I must admit, the PIR route stemmed purely from past experiences and the analytical view of U value performance. I think I came across a detail somewhere that illustrated this build up, did the basic maths and moved on. I assumed so long as the warm room air didn’t enter the wall there was no risk of condensation.

Having played around with a WUFI calculator it is fascinating how different materials have such different behaviours , how moisture moves through the wall structure and how you need to think of the complete construction not just the U values. Also the change in behaviour with vapour barrier and where they’re located.

In theory the current wall does hold up, but is very dependent on that interior vapour layer to allow no interior air through. I do much prefer the idea of a moisture permeable wall, that allows movement and has no risk due to lack of sealing or puncturing.

Funnily, the best performance for drying and moisture movement, comes from moving the OSB internally, using rockwool (or similar) between studs and then wooden batts directly to the outside of the studs and render. You do have to increase the thickness however to achieve the same results. I’d have to do a cost comparison also.

Thank you for the suggestions Tom, I’ve seen you suggest similar elsewhere, and the calculations do stack up for condensation. I’ve a preference for fibreglass over warmcell as I can DIY, and less mess. Presumably you glue and mechanically fix the EPS?

EPS does seem to be the way to go, I like the idea of wood batts, but they’d have to be thicker, and nervous of using it only 150mm from ground level.

Whoo, there's a thought

Only a few points from me in response to others. I started a response yesterday but it fell off the cliff...

Of all the mainstream insulants PIR gives close to the best 'headline' figure but it's only as good as the fit and the 'cloaking' of any thermal bridges. As FT highlights, anything with no thermal bridge mitigation is less than ideal, so anything that's 'all between studs' is, in my humble view, worth avoiding. For the OP's original construction, full (140mm) fill with Frametherm 32 (lambda 0.032W/mK) comes v close to the U value you'd achieve with PIR, and the fit will be better, because of the 'squidge'. *But you still want to cloak the bridge of the studs (see FT's suggestion).

''I like the idea of wood batts, but they’d have to be thicker, and nervous of using it only 150mm from ground level.''

I used a lot of WF, but XPS (or, for my preference but not necessarily the comfort of system providers, EPS) below ground and above ground to at least the height of 'local splash-up' (the industry suggests rain splashes up 150mm. I have a 300-ish 'tide mark' on my shed...). In my view (and contrary to 'conventional wisdom') a set-back (plinth layer not as deep as main EWI layer) is not essential.

Try that for size and I'll come back later if any 'supplementaries' req'd.

How high rain splashes depends on what the surrounding surface is like. We have a gravel topped French drain all round our house for that reason. The downpipes from the roof gutters discharge onto the gravel and we haven't seen any problems.

Posted By: fostertomwhat software is that screenshot? Doesn't use the obsolete, flawed Glaser method, does it?

It does - it's Ubakus, so not very helpful.

Posted By: Oj0001EPS does seem to be the way to go, I like the idea of wood batts, but they’d have to be thicker, and nervous of using it only 150mm from ground level.

A big advantage of wood (or hemp) batts is that they significantly reduce the decrement delay - effectively the time taken for heat to reach the inside from the outside; once heat reaches the inside then the building starts to heat up, and you want to delay that in a heatwave. Wood beats hemp for decrement delay, hemp beats wood for its moisture handling properties, both beat non-biological insulations (EPS, mineral wool, etc). You could use both though - EPS on the outer face, wood / hemp between studs.

For various reasons, look to achieve good levels of airtightness. OSB3 could form part of an airtightness layer if the joints are sealed, but not all OSB3 is airtight, so check the manufacturer's data for the brand concerned.

Posted By: fostertomIf the stud above is 95 wide, I don't see why not 100 block base wall - significant economy of construction, smaller footprint. But I've had BldgInsps and Engineers insist it must be min 190 thick (meaning 215 or 2x100 cavity wall) but they can't say why.

I'd guess that's due to Building Regs structural stability requirements (Part A, 2C6) Solid walls constructed of coursed brickwork or blockwork should be at least as thick as 1/16 of the storey height - although most of wall is not in brickwork or blockwork, the part that is would be subject to similar forces.

Posted By: fostertomOj, what software is that screenshot? Doesn't use the obsolete, flawed Glaser method, does it? You've previously used WUFI?

In your cross section drawing at top, have you met any resistance to base wall 'only' 140 thick? If the stud above is 95 wide, I don't see why not 100 block base wall - significant economy of construction, smaller footprint. But I've had BldgInsps and Engineers insist it must be min 190 thick (meaning 215 or 2x100 cavity wall) but they can't say why. Anyone?

It is Ubakus, apologies that's my ignorance, I hadn't fully appreciate the differences between the two methods. Although I don't fully understand why this method is considered less accurate from thermal and vapour gradients standpoint? I presume your actual WUFI build up of the above suggestion didn't show condensation risk?

Posted By: Mike1
For various reasons, look to achieve good levels of airtightness. OSB3 could form part of an airtightness layer if the joints are sealed, but not all OSB3 is airtight, so check the manufacturer's data for the brand concerned.

Is there a specific difference between air tightness (I presume to reduce thermal losses due to air passage) and Vapour diffusion through the wall? Since vapour barriers are designed to prevent air movement, but still allow vapour transport, why is there such an objection to them?

Posted By: Oj0001140mm block ... accepted as a full plans submission

Good news - encouraging. Here's a previous attempt to get it past BldgRegs:

-----------------------------

1.4.3 215 thick rule See drawing 316Ja-60.

Just convention? Or is it deemed-to-satisfy A1/2 2C?
If so, Table 3 requires 190mm, or A1/2 2C.12 requires 90mm for an ‘annexe’ not exceeding 60m2 (our new stud walled areas RED total 56.94m2).

But does A1/2 2C apply at all to a masonry found wall under a studwork upper wall?
A1/2 2C20 applies it to masonry walls (ours is mainly stud);
A1/2 2C2 says the wall must extend to the full storey height (the masonry part of ours does not).
The only specific (drawn) reference to a found wall is in A1/2 2C Diagram 8, showing the found wall only included because it’s part of a (multi-)storey-height masonry wall (ours is not). Note, our floor slab does not provide lateral support.

Even if A1/2 2C applies, 2C3.d. allows a minor departure on the basis of judgement and experience.
For a 100 thick found wall up to approx GL or FL, supported on both sides by compact material (as made significant in A1/2 2C Diagram 8) as in variants A,B,C and D:
My judgement is that this low 100 thick found wall has multifold greater stability than a 215 thick wall of (multi-)storey height; and
My judgement is that carrying stud down to the main found wall as A and C, compared with to top of 2c of Marmox laterally unrestrained as B and C, provides better lateral stability.
My experience is that variants A and B enable the easiest, most reliable continuity seal between DPM and DPC in accordance with C 5.5.a., and airtight seal between sole plate and DPM in accordance with L1(a) 4.23.b. Once the sole plate is bolted down, all that sealing is protected from further works; and
My experience, on previous residential buildings having similar found wall/floor/upper wall construction, and unquestioned by Bldg Insp, is that it has been trouble-free.

As allowed by A1/2 2C3.d., the design should be released from deemed-to-satisfy A1/2 Table 3 (190 thick)

---------------------------------

On that occasion, passed without comment, after previous objection - more due to oversight than agreement, I fear - I'd hope to get specific acceptance hence a precedent, not least to be no longer over-ruling the Engineer's spec.

Posted By: fostertom

Posted By: Mike1Wood beats hemp for decrement delay, hemp beats wood for its moisture handling properties, both beat non-biological insulations (EPS, mineral wool, etc)

Compare also to Warmcel blown in to quite high density?

Good question. My French hemp batt manufacturer also does a Warmcel equivalent, Jetfib'Ouate. So, putting those through the calculator at https://www.kenzai.fr/calculateur/dephasage.html, at 145mm thick:

decrement delay Jetfib'Ouate = 4 hours (Warmcell)
decrement delay Biofib'Chanvre = 4.5 hours (hemp batts)

So hemp batts have the edge by 30 minutes.

Raw data:

WUFI1 EPS

WUFI5 EPS

WUFI6 EPS

The climate file input into WUFI was Brussels - because it comes free, unlike any UK file, which cost £100s. Does that approximation sound reasonable/fatal?

I think that re-drying potential is the explanation for 'performs best with no VCL at all'. Without VCL, re-drying is both inward and outward. With internal VCL, it's outward only. With variable VCL, a bit of inward drying is allowed.

Of course, the sandwich has to be designed as suitable for 'fully breatheable' function - it's not a matter of leaving out VCLs for just any design. I arrived at this sandwich design after trying many variants, not just to work with WUFI, but also for buildability, fool-proofness, robustness as well.

It's a tribute to the hypnotic power of profitable membrane and sticky tape manufs' adverts, that this VCL-less buildup apparently hasn't been tested and adopted widely.

Posted By: fostertomThe climate file input into WUFI was Brussels - because it comes free, unlike any UK file, which cost £100s. Does that approximation sound reasonable/fatal?

Seems reasonable to me - see, for example, https://weatherspark.com/compare/y/41841~51121/Comparison-of-the-Average-Weather-in-Coventry-and-Brussels Though global heating is going to rip up the external environment data - maybe model it on Grenoble for an insight into the future.

However, from my experimentation, the internal climate used is likely to be the biggest source of deviation from reality, if the external is within the right ballpark.