What is considered best practice at the moment for air tightness on cavity walls that are being externally insulated?

I know about stopping convection currents in the cavity by filling and sealing (although maybe this plays into the overall solution).

However I'm talking about the air tightness of the entire wall itself. Is a membrane mechanically fixed externally before insulation? Parge coat the exterior?

** Edit ** I'm interested in _retrofit_ EWI on existing cavity walls.

I would have thought that you'd want the airtightness barrier on the inside of the wall. That's what you need for integrity of the thermal performance of the house. Further outboard of that, you're simply trying to eliminate 'wind washing' and convection currents in the insulation as I understand it, so there's no need for rigourous airtightness. Just a good attempt to fill voids to prevent convection, and a mostly sealed barrier to prevent the wind blowing in.

''I would have thought that you'd want the airtightness barrier on the inside of the wall. That's what you need for integrity of the thermal performance of the house. Further outboard of that, you're simply trying to eliminate 'wind washing' and convection currents in the insulation as I understand it, so there's no need for rigourous airtightness. Just a good attempt to fill voids to prevent convection, and a mostly sealed barrier to prevent the wind blowing in.''

You may be right, depending on how bad the 'wind-wash' is. My take in terms of my post above was that cold air might be entering the cavity through gaps in the masonry. Therefore parging could help. I realise that the air-tightness layer needs to be internal, but one could have a fantastically air-tight house with no insulation, so reducing (to as close to nil as possible) the chance of wind-wash has to be high up on the shopping-list too.

Posted By: MarkyPmy extension is having EPS EWI over new medium density blockwork cavity walls with full fill mineral wool insulation. I think the combination of the adhesive on the EPS and then the render on the EWI will make it air tight externally.

If the EPS is applied with dot and dab then you can (will) get convection air flow between the wall and the EPS. Either use full-coverage adhesive or (what I do is) put a solid line of adhesive (as a minimum) on the top and bottom edge or better on all 4 edges and then a couple of dollops in the middle. I have found that a line around the edge is better than full cover if the wall is uneven.

Posted By: gravelldSurely it's more complicated? And therefor more liable to error? It means going around every joist etc

No it's generally less complicated. Change the design to avoid problems. For example use a ledger board inside the airtightness barrier to support joists instead of running them into the blockwork or studwork. Outside there are lots of penetrations for things that all would need care, and the external environment is a lot harder on components and junctions. For example how do you ensure drainage of any water that does somehow find its way into a wall? How do you do the wall-floor detail and the wall-roof details?

Disruption - if EWI is sold as low disruption, and we want to perform proper holistic retrofit, how are we going to do air tightness? Plus I don't fancy ripping out every internal wall covering in my own house.

I would've thought, conceptually, an external (although internal to insulation) air tightness layer is easier

I suppose it is less clear cut for a retrofit. The main thing is to have a complete and detailed plan before you start and to think through exactly how to achieve the planned result. Have an airtightness test at the earliest possible opportunity and then another one at the end, and maybe a third in between. The cases of failure that I know of were largely due to incomplete and/or impractical plans, IMHO.

It is also possible to have part of the barrier inside and part outside, though that makes things more complicated to design.

If the existing cavity is uninsulated would it be possible to fill it with something which provides airtightness? Not necessarily an insulant, it could be concrete since you are also installing EWI. Is this a bad idea? If so what are the pitfalls? My first thought would be rising moisture at the footing level since there would be no DPC.

Expanding foam might damage the structure so don't take a chance, polystyrene beads are easy,

Posted By: djhIcynene is specifically marketed for cavity fill insulation

Didn't know that - probably ideal - expanding to very full fill but permanently rubbery/flexible, unlike other expanding foams. And breatheable. And largely made from organic source, not petrochemical. And not relying on exotic gas fill, just air, so only medium insulativeness but no long-term degradation of same on that account.

Or just remove the first row of roof tiles and fill the top of the cavity with expanding foam along with all vents etc. Air movement in the cavity should not be an issue provided the air can't mix with the outside air, and all of the wall is covers with EWI.

no, the air could set up convection currents, effectively wicking heat away from the inner/warmer leaf and to the colder outer leaf. Better to prevent air movement where ever possible.

Posted By: DarylPno, the air could set up convection currents, effectively wicking heat away from the inner/warmer leaf and to the colder outer leaf. Better to prevent air movement where ever possible.

What "colder outer leaf" as it will be covered with EWI......

About sealing up already partial filled cavities, I emailed the UK Icynene people about this but they haven't had a reply. I got the Retrofoam people interested for a while but they didn't call back.

What about membranes, can they have a place, fixed on the existing external leaf before EPS applied?

Ok, just wanted to put it out there.

So it looks like the answer is boiling down to either stopping air leakage in the cavity and/or using render externally or plaster internally to act as a second air tightness layer, to which (for example) joinery can be taped, or Tom's OSB tea cosy on the roof structure.

Anyone think I've misunderstood that?

Well the JUB site says that their EPS adhesive is elastic. It doesn't say it is airtight though and I haven't found any such statement, so I hope it's a good idea and I hope they have more test certs than I've seen.

Depends on price as well as specification, of course. It needs to be a fairly cheap adhesive if it's going to be applied over 100% of the surface instead of just dabs, or dab plus perimeter. It also seems like it would be difficult to guarantee/repair. By the time you have all the adhesive on so you are able to do an airtightness test, it is all covered up by insulation and difficult to fix any defects. So it might be good if you have 100% trust in the workmanship but less so with some random contractor.