Here is the Ubakus calculation for the membrane in-between the PIR layers
(only one attachment per post)

Hi djh,
thanks for spotting a typo -- indeed it's PIR board that I was thinking of (the question is corrected now).

The U-value is 0.16 -- I thought this is consistent with building regulations from 2022 of 0.16 (it's a retrofit) -- has there been an update that I missed?

Clearly 0.16 is not Passivhaus, but probably still quite "green" compared to just roof tiles and 10 cm of aged rockwool on the fllor ;)
Which solution would you recommend for the roof insulation?

Many thanks!

You could easily go 100mm PIR under the rafters there? Would get it a bit better. That’s what we did. Also we just foil taped the joints and didn’t bother with a VCL

Also just looking at the photos I see that we taped both layers (belt & braces) and used tile battens (on the inside) to fix through the 100mm PIR into the rafters, creating a service void, then plasterboard finish.

Is it right that there's no underslating felt? If so, these must be original tiles/battens from 1927 - impressive durability. So the 25mm gap on top of the proposed insulation must be well ventilated via the plain tile joints, unless they're the flat (not crosswise-cambered) type, laid unusually close-fitting, whose gaps may have been well gummed up. Assuming the gap is indeed well ventilated:

This is close to my standard preferred design, for both sloping-ceiling roofs and for external stud walls, except that I leave out any vapour barrier at all. I have a question for the forum, which might help answer blubb's original question. Well, to be precise my preferred buildup is:

Rafters (or studwork), plasterboard, skim and vapour-breatheable paint internally, which can be punctured at will for electrics etc; 12 ply sheathing (or 11 OSB3) gapfilling (polyurethane) glued and screwed/nailed externally, as air barrier, blown full of Warmcel insulation.

EPS (or wood fibre) insulation; if 95 rafters then 150 EPS; if 145 rafters then 100 EPS, glued external to the sheathing EWI-style, with highly gappy glue application, to allow unimpeded water vapour passage. Board-edge joints filled with expanding foam, by long nozzle pushed in full depth, so filled full depth from inside outward. Bead-board EPS is by far least environmentally offensive, compared to any foam-board insulation such as PIR, and is sufficiently vapour-breatheable.

Breather underslating felt, downslope battens, crosswise tiling battens, tiling.

This design, fully breatheable without any vapour barrier (VCL), results from my one opportunity, years ago, to play with WUFI. I tried many use-cases, and for this type of construction, found it performed much better without any VCL. Note, there is a (vapour permeable) air barrier layer and a (vapour permeable) external wind barrier layer. I put this superior performance down to its maximisation of re-drying potential, unimpeded both outward and inward. With or without VCL, there is, at some times of the year, interstitial condensation in the outermost centimetre(s) of the EPS, but this is harmless as there's no embedded timber in that zone. Neither EPS or woodfibre mind such intermittent wetness, as long as fully re-dried soon enough, whether diunal (winter sunny day) or seasonal (after a long cold winter).

For years, Building Inspectors have accepted this construction, but now there's a change - post-Grenfell they avoid all possible blame, in all topics, by requiring certification for every detail. I've found no certification to quote for this VCL-less construction, not even scrappy results from my own WUFI exploration of years ago. No consultancy seems willing to provide same for a fee, only to use WUFI in their own projects. I'd intended to join TGR's WUFI training (happened last week) once again, to get the 6wks free access to WUFI - but missed the boat (next in the autumn). Not even sure a Building Inspector would accept my semi-amateur WUFI evidence as certification.

So, dear GBF members, I've asked before, but can anyone suggest any certification for this VCL-less construction?

As far as blubb's original question, no direct answer. but might gain some insight from what I've said?

PS AFAIK, plywood does in fact have some effect as an 'intelligent' (but weak) VCL - higher vapour resistance in humid conditions than dry. Any comment?

Posted By: fostertomPS AFAIK, plywood does in fact have some effect as an 'intelligent' (but weak) VCL - higher vapour resistance in humid conditions than dry. Any comment?

Err, I think Intello is more vapour open in more humid conditions? Exactly opposite to what you said.

I'm not sure I am reading your build-up correctly from internal to external Tom?

Plasterboards & skim
Fully-filled rafters (Warmcell)
12mm plywood
100-150mm EPS
Weatherproof breather membrane
Counter battens / battens / Tiles


Plywood on the outside of the rafters would create a vapour-closed construction - or at least a vapour-restricted one. I'm told the correct ratio of insulation that can be on the warm side of the VCL is 1:5.

OSB3 (depending upon the make and thickness) can be used as a moisture vapour-variable layer - fully taped and sealed we use it inside our timber frame wall systems as combined racking/VVCL/airtightness with woodfibre between studs and direct rendered woodfibre sheathing board outside.

As much as we try to remove the need for plastic membranes in construction, we would still always ask for a VCL under the roof. A service void will prevent accidental penetrations. However, with some woodfibre boards, external breather membrane is not required thanks to Hydrophobic surface treatments.


Our preferred roof build-up, would be:

Plasterboards & skim
25 - 50mm service void (can be filled with WF)
VCL membrane (fully taped and sealed as draft proofing)
Fully-filled rafters (Warmcell / zell/ WF)
Woodfibre sarking boards with hydrophobic surface treatment, (weatherproof sarking and insulation combined)
Counter battens, battens and tiles

Hi sgt_woulds

It's always great to hear the build-up 'sandwich' that practitioners and TF manufs have settled on. For instance Seamus O'Loughlin at Viking House of this forum was standardised on pretty much what I've described, tho relying for OSB airtighness on tapes instead of glue-and-screw, and I think may have modified a bit since.

Yes, that's my build-up.

Posted By: sgt_wouldsPlywood on the outside of the rafters would create a vapour-closed construction - or at least a vapour-restricted one. I'm told the correct ratio of insulation that can be on the warm side of the VCL is 1:5

Interesting - that 1:5 (some say 1:3) was exactly what I first set out to test in my aforementioned WUFI-test period (and found it made no systematic difference). This was widely quoted, and was really the only guideline, I guess till 15yrs ago, before WUFI became generally available. I really searched, back then, for its original source and authority but found nothing. Warmcel used to quote it prominently. I asked them where they got it but they couldn't say. They don't say it now. I'm pretty sure it was the cause of the catastrophic failure of UK's first PH school, in Dartington, Devon, which ruined its PH-pioneering architect. The truth never came out, mired in litigation, described as 'leaks' showing after a few years, but I'm pretty sure it was year-on-year accumulating intertsitital condensation. Not sure exactly how they got it wrong, tho.

As I say, applying 1:5 or 1:3 in WUFI, as opposed to no such gradienting, correlated with no systematic difference. Nor did the difference between super-permeable materials like Panelvent, vs ones like OSB3 or EPS, which despite being only moderately permeable, are still a world away from VCLs. From this I concluded that as long as materials are at least moderately permeable, their combination and order in the sandwich doesn't matter much. This would mean that a "vapour-closed construction - or at least a vapour-restricted one" (at least as you've described it) barely exists or hardly matters. The big surprise was how much better some of my sandwiches performed with no VCL at all.

I wonder if you've evolved the build-up sandwich that you use, in WUFI? If so, what data you used for OSB3, and EPS. Figures quoted vary widely, to either support or deny their suitability and reliability - OSB3 airtightness and EPS permeability are in much doubt (or were). I took much comfort from hearing Peter Warm say "I've never seen a PH fail on airtighness because of using OSB3" (this was before Smartply's plastic coated ProPassiv).

And I wonder if you tested any sandwiches without any VCL. VCLs are so engrained that I suspect no one tests that in WUFI. The earliest housebuilder TFs in UK, back in the 70s, were built without VCL (because no one knew better), turning into a national scandal of rot and condensation. The industry response was internal VCLs, and all has been well since (??). The theory developed to explain this - that moisture in buildings was internally generated (clothes washing, breathing) and should be shut out of the sandwich by VCL. Now we know that internally generated moisture is trivial compared to external moisture which flows inward and outward seasonally and diurnally; and that a VCL that isn't absolutely perfect (never is) is no protection against a long term trickle of vapour permeation, which sees a pinhole, or aging/fatiguing sticky tape, as an open door. What an imperfect VCL does do, is impede short-term rapid re-drying when the sun comes out.

Since my early WUFI adventure, it's become known that bulk-air leakage is at least as important, if not greater, in intersitial condensdation, than vapour permeation. Bulk air carries water vapour with it, to the cold zone where it condenses, just as permeating water vapour does. Strangely, it was found that a major bulk-air leak is less dangerous that a small one, because a major leak can warm up the air passge, where a slow leak leaves it as a cold condensing surface. My guess is that those 70s VCLs were succesful, not as highly imperfect vapour barriers, but as effective air barriers. So nowadays, if a sandwich has good airtightness, does it need an imperfect VCL as well? I found 'No'.

People do seem to realise this, but instead of leaving the VCL out, they specify expensive VVCLs - 'intelligent' or variable VCLs which, as corrected

Posted By: djhErr, I think Intello is more vapour open in more humid conditions? Exactly opposite to what you said.

Some of my old WUFI conclusions do seem a bit surprising - for instance if combination and order of layers is so non-critical, what exactly did cause the Dartington catastrophe? Unfortunately I've lost that WUFI data and couldn't read it anyway, not having a WUFI copy handy. I do resolve to join the autumn TGR course, or maybe even buy WUFI, to re-check what I've been relying on, and even rashly advising on, for so long.

V interested to hear more. Anyone?

BTW, what's zell?

Hi Tom,

lots of interesting points that I'd like to go over with our technical guys.

My knowledge is based on what I've learned from discussions with our technical team and my own games with UBAKUS, so I'll defer to your WUFI knowledge regarding the placement of layers within the build-up.

Our own build-ups have been extensively tested in WUFI and in continuously monitored real-life construction studies in Germany. We aim to design-out VCL and airtightness membranes wherever possible; this not only speeds the build and reduces costs, but allows for the typically low-skilled workers engaged in airtightness/vcl work. It is considerably easier to cut / install / seal OSB3 internal to the frame, and much harder for them to damage it as other work progresses.

This works well for our wall constructions, but roofs still require VCL due to - as explained to me - 'the higher internal/external moisture loads expected in the roof insulation'.
Yes, this could just be belt-and-braces thinking - 'Rectum Protectum' is always the first order of any business!

However - apart from weight - there is no reason not to use OSB3 on the roof as well and ditch the plastic VCL. As with Plywood, it will vary with the glues and quality control applied.

It is important to note that not all OSB3 is the same. I'm told a lot of what we have in the UK would not pass muster in Germany.

I believe the issue with early TF/VCL construction was placing the racking boards external to the frames which caused a restriction in breathability. Adding VCLs, (applied with typical bob the builder cack-handedness) only exacerbated the issue - allowing moisture into the insulation zone, but restricting its movement out when conditions allowed. The old maxim of creating a build-up that is progressively more breathable from inside to outside holds true. From everything I've learned over the years - and reinforced working here - I agree with your sentiments; air tightness is more important than vapour control in a breathable construction, but both are vital in typical non-breathable construction.

BTW, zell is a blown wood fibre - similar in concept to flock insulations.

Posted By: djhstatement at https://www.passivhaustrust.org.uk/passivhaus_awards/Passivhaus%20Educational%20Buildings/ about the first PH schools

Ah, so not PH, just Zero-whatsit and eco - a v brave even earlier lead.

Posted By: djhThere's a bit at https://www.architectsjournal.co.uk/news/seven-year-old-zero-carbon-school-demolished-due-to-leaks ... where the account of the failure mechanism seems plausible to me.

Maybe

Posted By: djh... why you try to drag PH into the mud?

Easy up Dave, I'm not your enemy, nor PH's. Why would I?

Posted By: sgt_wouldsnot all OSB3 is the same

True - the only British Isles alternatives AFAIK are Scottish Sterlingboard and Irish Smartply - I prefer the latter. None seem to quality-control for airtightness - must vary a lot from batch to batch. At a WUFI training in Dublin, I met a senior boffin from Smartply, told him that what we need is a board of quality-controlled high air resistance and low vapour resistance - and why. He took it as a lightbulb moment, invited me to join a panel to help develop same - never happened, instead they produced plastic-coated PassivPlus OSB, which misses the point entirely. And leaving 3mm gaps between board edges, then taped over, guarantees that the tape will fatigue and fail, as a movement joint, at some point in its life. Airtightness expert Paul Jennings says that's one year.

I also corresponded with a Fraunhoffer boffin, who provide WUFI data testing on materials, asked why they don't test for airtightness. Answer - airtightness is negligibly affected by the board's characteristics, entirely by the joints. Not so if it's OSB3 or ply gapfilling glued and screwed. At that time, having discovered the moisture-transport importance of airtightness, to handle it WUFI just introduced three grades for airtightness of joints - good fair and poor, that would be entered intuitively by the designer! After all the scientific rigour of the permeabilty modelling, this for the even more important airtighness. It may have all changed by now.

Posted By: sgt_wouldsI am also slightly doubtful of the lifespan of tapes. As wet plastering seems to be very effective for airtightness, perhaps we should change to specifying a lime-plastered wood wool board internally.

Are your doubts based on anything more than a feeling? For myself I trust the tests done by Siga and by Pro Clima. There may be other trustworthy firms. Application technique is very important. Lime plaster is certainly much less prone to cracking than gypsum and as long as it is thick enough seems to work well though the stiffness and relative movement of the substrate is also important. There's still the question of the junctions - floors, ceilings/joists, internal corners etc. Other approaches include liquid airtightness products. And membranes. And Tom's glued OSB.

What are your thoughts on the longevity of using aluminium foil as airtightness tape

I would never consider using aluminium tape for airtightness purposes. Are there any with good long-term experience?

"Are your doubts based on anything more than a feeling?"

50/50 Gut feeling and experience. Having worked on buildings from many eras I've seen first-hand which materials stand the test of time and exposure. Even the best synthetic materials seem to have a limited lifespan. The tests that the manufacturers have done are based on guesstimates from accelerated testing which can never replicate true decades-long conditions.

Plastic degrades, glues and chemicals lose their covalent bonds.

I have seen flash-band that has lasted for over 30 years as internal draft sealing and is still strong and supple, but conversely, only lasted a bare few months when used on a highly exposed roof. I have often thought that used internally, it would probably be cheaper and longer lasting than some of the specialist tapes, and easy to mould and seal against awkward junctions.

The aluminium tape often gets used for sealing foil-backed PIR as VCL and draft proofing - again, being internal and not under extremes of weathering or movement I would expect it to fare quite well, but I've never seen any studies to prove this.

Back when I was an industrial abseilor, I used to crawl around in a lot of old ductwork that was sealed with alutape and apart from where it had suffered mechanical damage it seemed to last quite well against extreme heat and moisture.

I also worked as a crewman on one of the Goodyear blimps and that thing was mostly repaired and held together with (aviation) duct tape, so it's fair to say I have a pragmatic regard for what I trust in terms of tapes.

Posted By: sgt_wouldsAny recommendations for OSB/plywood/woodwool

Have you looked at Elka Strong Board? It is a type of OSB racking board that claims to be "up to 5 times more breathable (vapour open) than OSB3 and has greater tensile strength".

Recommended for natural insulation, vapour resistance factor is stated as 40.