Forming an airtight wall-ceiling junction

We have started to create the airtightness layer at various points in the building. In this post, I will describe the sequencing for one of our ceiling-wall junctions. The airtightness layer in this part of the build is formed by plaster on the wall and by Intello vapour barrier on the underside of the roof I-beams.

Stage 1) The Intello was stapled to the underside of the I-beams. Tescon tape was applied to join up each row of Intello. We also taped over the staples, in case the Intello is stretched and a small gap appears around the staples, although apparently this is not strictly necessary to achieve the Passivhaus 0.6 air changes per hour (ach) target.

Stage 2) In order to join the plaster to the Intello, we used Contega PV tape - half adhesive backed, half mesh. We had to use Tescon tape to stick the Contega to the Intello, as the Contega adhesive would not stick properly to the Intello. I had a very helpful conversation with Niall Crosson from Ecological Building Systems where he told me that if the Intello is exposed to humidity, it can affect its ability to stick to the Contega. I guess this must be because of the property of the Intello to vary its vapour permeability according to the ambient humidity.

Stage 3) After the first photo below was taken, an initial parge layer was applied under the Contega. We took care to put a kink along the length of the Tescon tape by pushing it up slightly, to allow for any possible future movement between wall and ceiling.

 

 Stage 4) Once it had dried, a second layer was applied to encase the mesh and the Contega tape above it. Then battens were screwed through the Intello to the I-beams. The Intello grips the screws and forms an airtight seal around them, which is only compromised if the screw is removed: it is vital therefore not to remove a screw if it has been incorrectly positioned - better to leave it in place. The space above the Intello (i.e. between the I-beams) is to be filled with Warmcel cellulose, so the battens will help the Intello support the weight of the Warmcel and will provide a 25mm service void to the ceiling plasterboard. 25mm is normally too narrow but I didn't want headroom to be reduced any more than is necessary in this attic space.

The remainder of the wall will be plastered later, along with the ceiling.

Airtightness around existing floor joists (and along the top of internal walls)

The last few weeks has seen tremendous progress on the build and one issue that I wanted to focus on in this post is how we managed airtightness around the ends of the existing floor joists.

The picture below shows how the floor joists looked before. Normally, the block work would only be plastered above and below the joists, where the finished wall would be visible. In a Passivhaus refurbishment, the plastering needs to run continuously around the joists.

We had originally intended to keep the plasterwork below, if it was sound enough, but mostly it just fell off with minimal if any encouragement. As discussed in a previous post, water penetration in the cavity, due to poor detailing at the original roof-wall junction, together with some poor choices with the original render and paintwork seemed to have been the cause.

The issue of how to make airtight joist ends has been covered in this AECB YouTube video. One of the key points they mention is that the approach needs to take account of whether the wall is going to be insulated internally or externally. If a wall is to be internally insulated, there will be a significant temperature drop at the end of the joist, which brings a risk of condensation because the joist end sits within the cold wall. In the film, to avoid this, they have mounted the joists on joist hangers on an internally insulated wall. Steps then need to be taken to achieve airtightness around the joist hangers.

In our project, the existing house is being insulated on the outside, so the inner leaf of the existing cavity wall will be the same temperature as the inside - i.e. no condensation risk. This has been our approach to achieving airtightness around the joist ends:

Stage One

Carlite Bonding parging around the joists onto the original concrete block wall or sometimes new built sections of Thermalite block wall:

The same principle has been applied to a steel girder fabricated with an additional L-shaped piece, welded with the bottom of the L welded to the web of the girder. This means that the profile of the end of the girder is a solid rectangle, like a joist end, rather than an "I" profile, simplifying making it airtight.

Stage Two

Priming around the joist ends with Pro Clima's Tescon Primer adhesive. This helps to make sure the Tescon tape sticks to the plaster.

Stage Three

Taping around the joist and girder ends using Pro Clima's Tescon Profil tape.

 

Stage Four

Dotting the corners with Pro Clima's Orcon F adhesive. This is a "belt-and-braces" step. We were not absolutely sure that our tape corner junctions would stay completely airtight. We applied dots of Orcon F to be sure! 

Stage Five

Apply Unibond to the Tescon tape to provide a key for the second parge layer (next stage).

Stage Six

Apply a second parge layer to seal in the Tescon tape. In the picture below, the joist on the left was originally a double joist. We cut short the thinner of the two rather than try to treat the double joist along its length. Obviously, if the double joint were needed for structural reasons, we would not have been able to cut one joist back.

 

Dealing with the edge joists
The joists run north-south in our build. There was no edge joist on the east side due to the run of the original roof rafters, however the edge joist on the west wall was mounted close up against the wall and there wasn't enough space around the joist ends to parge and tape around them. One option was to wrap the entire length of the joist in Intello Plus membrane, then tape it down above and below the joist. We decided it was more practical to cut out the existing joist, reduce its length a few cm and re-attach it with coach bolts along its length of the west wall. We used resin anchor gel to secure the bolts in the wall but also applied a generous ring of Orcon F around each a penetration to provide a flexible airtight seal (picture below).

YouTube video of project

I've uploaded an interview about the project given in December about the project.

View it here: http://www.youtube.com/watch?v=EjNbNJ3YgfQ

Photos from weeks 9 and 10

I've uploaded photos of the build - weeks nine and ten - on flickr. Our progress was slowed down by the snow, but the team has been trying to catch up since the new year.

Xmas and New Year update

It's been a while since I posted. The week-long installation of the scaffolding and 'tin hat' (photo below), the freezing weather and the Xmas/New Year break have meant that there is less to report. I feel very pleased that I didn't decide to skimp on getting a full 'tin hat' - I think it would have been very hard to make the progress we did during December without it. The project reached a turning point in December when we stopped removing bits of the existing building and started adding to it. December also saw a large outflow of money as I tied down as much of the spend on materials as possible before VAT goes up to 20% on 4th January. The rise in VAT further shifts the economic balance away from refurbishments in favour of new builds. Given that we have to address our existing housing stock if many of us are not going to faced with "heat or eat" choices in the coming decades, this policy choice of successive Lab-ConDem governments shows that, in practice, they are more concerned with the welfare of corporate big builders' interests than the citizens who elect them.

Week 4 ends

Windows finally ordered on Wednesday. I managed to get the window spec selectively trimmed using PHPP to test what effect this would have on the building's performance. The PHPP paid for itself many times over just in this one week of 'optimising' the spec.

Getting to grips with Therm (also after a long struggle). The model of the external floor-wall junction I created, has produced a provisional psi value of 0.15W/mK, lower than the 0.2W/mK estimate I had previously. Putting that more accurate figure has trimmed another 0.5kWh/m2.a off our Annual Heat Demand. Helpful, as it allowed me to trim a bit more off the window spec.

I have a sense of having achieved quite a bit this week.

Meanwhile, the scaffolders have been busy by this time next week, the whole house will be shrouded in a tarp and tin weather shield that will protect the existing structure from the rain and provide a working environment for the site operatives more conducive to accurate and quality work. It also much reduces the risk of weather-related project delays. We have been lucky in that we have missed most of the bad weather affecting the rest of the country, although night time temps have dropped to as low as -10C. Even though we have not had much snow, it has not been nice weather to be working outdoors.

Week 4 begins ... the latest on windows and on thermal bridges

Windows

Week 4 starts and we Still Haven't Placed Our Window Order. The process of placing the order is taking a lot longer than I'd anticipated. I am optimistic that we will be in a position tomorrow, finally, to press the Go button.

This has all come as surprise to me, as I'd thought naively that we could spend time fine tuning our window spec, then present our carefully worked out window schedule to our helpful, local Internorm distributor who would then be able to process it quickly and simply ... no probs, job done.

I realise now that we should have sat down and gone through our schedule with the distributor and talked prices and options at least two months ago, despite not having planning permission or completely finalised window dimensions at that stage. The order process is slow because, while it is easy to spec out the windows for their required energy performance, there are so many other variables to consider and requirements to fulfil. One is the thickness of each pane of glass. If the glazed unit is over a certain area, 4mm glass has to become 6mm. For a given width of bead, this means 4mm less space between the panes, which has a significant impact on the U-value of the glazing. None of these are problems in themselves but overcoming them adds to the cost of the windows unnecessarily.

Better to use slightly less exacting window performance figures in the PHPP. That way, you can order standard products and get a much more cost effective solution. It seems obvious to me now that I am writing this and, in fact, I have used been using conservative figures throughout the PHPP but I think I got carried away by all the exciting "Leading Edge" or best case figures that the different window manufacturers banded about. Assuming your chosen window provider is in the business of manufacturing windows with near Passivhaus performance, the conversation needs to be about what their standard spec is on those windows. It also needs to be about whether window sizes or other variables will make it harder to achieve the energy performance you are planning for in the PHPP.

Here are my rules of thumb:

	My "Leading Edge" Spec Assumptions this time	More conservative PHPP assumptions I will use next time
Glazing "g"-value	0.6 or 60%	0.5 or 50%
Glazing U-value	0.5W/m2K	0.6 or even 0.7W/m2K
Frame U-value	0.94W/m2K	0.94W/m2K
Spacer psi-value	0.038W/mK	0.05 or even 0.1W/mK

I got it right with the frame U-value, because I knew that the "leading edge" spec was a lot more expensive than their standard Passivhaus spec.

These figures aren't set in stone and I'm sure that each year what is considered standard "Passivhaus suitable" spec will improve.

Thermal bridging

On Thursday, I attended a one day course on how to use Therm, together with a very helpful Excel spreadsheet developed by Peter Warm, to calculate the psi-values of thermally bridged building junctions.

Therm is not at all intuitive but does have the advantage of being the only free software that can be used to derive a psi value for many types of thermal bridge. Therm can only model in two dimensions and more complex (and very costly) software needed to model certain types of thermal bridge junctions.

I have been grappling with this for some months but am now finally biting the bullet and getting to grips with Therm, so that we can replace the conservative (I hope) thermal bridge psi-values (0.2W/mK) we have used in the PHPP to quantify the additional heat loss through the floor-wall junctions of the existing house: this type of thermal bridge is unavoidable in a refurb but can be designed out in a new build.

Likewise, I want to model the junction between the existing house walls and the window jambs (sides), heads (top) and cills.

Meanwhile, the first insulation is being installed around the base of the walls of the existing building and in the base of the new build side extension.


We had to buy double quantities of the insulation in half thicknesses because, even though the thicker sizes do exist, the suppliers will only sell them in very large quantities, unsuitable for a project of our scale. I asked our builders to stagger them slightly to reduce possible thermal bridging in the inevitable, if tiny, gaps between each piece. So instead of this...

As we got around the corner, we started adding a 50mm overlap, like this...

It means a little bit of extra labour but it is essentially a free way to get the best from the insulation you are using. The images above are of the insulation around the base of the existing building, 2 x 60mm thick, which we are fitting from DPC level down about 400mm.

Installing dual layer of Foam Glas, under the toe of the concrete slab, where the weight of the walls of the new build will bear down - Foam Glas can take much heavier loads than other insulation  - we are also taking the opportunity of staggering the two layers to minimise unnecessary thermal bridging.

Later this week, the scaffolders will be here to start setting up the shroud that will cover the building for the next three months while the external wall insulation, the windows and the new roof are put in place.