To conduct the test the building is sealed and a calibrated fan is mounted in a doorway. It sucks air out of the building to depressurise it. Then they can test the pressure differences between outside and inside and watch what happens as they test at different pressures. A number of different data points are collected and graphed, creating a final averaged figure. We have tested twice now and will have one final test when the building is finished. So the first two tests have been used as aides to discover where in the building the airtightness needs to be improved. As the inside is depressurised it pulls in air from outside (actually air from outside pushes in and tries to fill the space) and you feel with the back of your hand where air is moving through the building envelope.
|Searching for leaks|
Building Regulations state a maximum figure of 10 air changes per hour, although 3 is the minimum standard for AECB (Association Environmentally Conscious Buildings) Silver Standard for environmental building. Passive Houses have to achieve 0.6. Getting a score of 10 down to 5 is relatively easy, however from 1 to 0.6 is very challenging...
Paul Jennings, the airtightness guru, arrived with all his gear and got to work. We achieved a result of 4 in the first test, we were all disappointed! Leaks were searched for around junctions of the walls and windows and roof and sealed with more tape. This got the figure down to 3.6. Not the start we were hoping for.
Unusually for a passive-style house, we have used traditional concrete blockwork. Blockwork can vary a lot in air permeability and the more environmentally friendly blocks composed of fuel ash can make the blocks even more porous and very leaky. So the blockwork could not be relied upon to give us the airtightness. The redeeming feature of blockwork is that it is a great thermal store, absorbing heat and very slowly releasing it. It will take a lot of energy to change the temperature of the mass. At this point in construction we had used blackjack, a paintable bitumen, as a seal when perforating the blocks, so we guess that this had not been as effective as we had hoped either.
Once our internal render had been completed we conducted our 2nd test. This time the first score was 0.9 - hooray! We were under 1, which was the figure we were aiming for when we first began this journey. A huge accomplishment. We always said that the aim was to do the best we could without throwing huge amounts of money at it. The costs can increase exponentially as you get down to the final 0.1s. Just goes to show what a difference a layer of render can do.
But Andrew and Charlie were really hoping they could make it to 0.6. It's boys- they see a target and they want to hit it, goal driven! With more rushing around with smoke pens and backs of hands against junctions they found some more air movement and used an airtightness sealant and mastic to fill the voids. Apparently this stuff is the business and doesn't fully harden and leak at a later date. A lot of our issues were around the electrical switch boxes and the wire perforations through the render.
We managed to get a result of 0.72! I think that is amazingly good! The photo below shows a result if 0.62, however each result has to be plotted on a graph and it is the average which determines the final score.
I am so delighted. 0.6 would be great but to be honest the difference of 0.12 is nothing that we will notice, our house is going to be super low energy and that is what we wanted. It remains to be seen whether we'll improve the score by the 3rd test. Andrew is planning to use some of the mastic inside the switch boxes in an attempt to improve the score. We've also completed the plaster layer now which may have an impact. However it has been known for scores to increase in the final test, as during the final finish more perforations are made in the airtight skin when you hang shelves and so on.
(I have to confess that I failed to write down the scores at the time, believing that they would be indelibly etched into my memory, however that has turned out not to be so! Paul may correct me if I've got any of these numbers wrong.)