A39 House

A39 House – Performance and Monitoring

Air Sealing Testing

Introduction and Rationale

“Build tight and ventilate right” is the mantra for passive houses in cold climates, such as “Passivhaus”. The idea is to heavily reduce air entering the building through cracks, windows and the like. This reduces the cooling and heating loads, enables no or little air-conditioning. It’s debatable whether this is worth pursuing in a warm climate such as Brisbane. I think is it, because

  • The Winter climate design temperature is 6°C, so we have a temperature difference of about 14K. Sealing the building means you can omit heating with a good solar design.
  • The shoulder season is very mild but cool at night. Building sealing keeps the heat in during the night.
  • The Summer season is hot and humid (32°C dry-bulb), so building sealing is not a big benefit, since the temperature difference is only about 6K.
  • In Summer, if you’re air-conditioning, it’s the humidity that increases load. Building sealing is a big benefit here.
  • Builk insulation is air-permeable. In most houses it performs poorly, as air just convects heat through it. THis is why batts are dirty – they are filtering air. Building sealing allows insulation to work properly, provides it is external to the insulation.

Unit Mayhem

Units for measuing building air leakage are numerous and confusing. Here is the summary:

Units Name Units Desc Note
ACH50 Air change rate (under pressue) ACH Air changes per hour under pressure ACH measures do not account for volume:surface area relationship
PERM50 Air Permeability (under pressure) m³/hr/m² Leakage per unit façade  under pressure Façade includes walls and floors and ceiling
ACH Natural Air change rate (normal conditions) ACH Air changes per hour in normal conditions Empirical formula is ACH50/20

The two main ones are ACH50 and PERM50. They measure similiar things: building leakages at 50Pa. One uses a volumetric unit and the other an area unit. The PERM50 is generally “better” because it accounts for different size buildings better. ACH50 penalises small buildings. These units are not directly interchangable. Based on AIRAH published data, we can see the difference:

Small single floored office building built ~2004                          2,960                          7,965 6.8 5.69 0.28 1.2 2.7
Converted and extended house used for office space – 1950s                              403                          1,140 16.9 16.19 0.81 1.0 2.8
Small office – part floor of offices around an atrium                          1,084                          3,902 9.1 6.23 0.31 1.5 3.6
Ground floor – new office building                          1,233                          4,661 11.7 7.92 0.4 1.5 3.8
First floor new office building                          3,223                        11,539 7.7 5.23 0.26 1.5 3.6
Free standing office/library building, large  vaulted ceiling ~1984                              925                          4,315 17.7 13.9 0.7 1.3 4.7
ACT Houses 22.8 17.96 0.9
Scott Road 91 281 3.73 3.9 0.19 1.0 3.1

Note the AIRPERM:ACH50 ratio does vary, but not by that much considering we are looking at houses, offices and libraries.

Building Tight

To add more info!

Running the Test


Show Me the Money: Results

We’ll use Air Permeability (AIRPERM) in m³/hr/m², or “air flow per unit facade area” to compare. The final results is:

The house results were, tested on a still day to EN 13829:2001

Name Value Unit
Air Permeability @ 50Pa, PERM50 3.7 m³/hr/m²
Air Change Rate @ 50Pa, ACH50 3.9 ACH
Air flow @ 50Pa 303 l/s

This is “Best Practice” for new-builds in the USA and three times tighter than the UK standards. On the other hand it is six times leakier than the Passihaus standard. Result summary: good but not great.

This was a depressurisation test. The HX dampers, ceiling fans and kitchen exhaust hood were all sealed. The HX has automatic dampers. Another test (30l/s higher) showed that the HX dampers leak 30l/s@50Pa when open. The kitchen and bathroom fans have “exhaust hats” which appeared to work well in depressurisation mode.

Results and Improvements

Where did it leak?

  • Windows – main source of leaks. Alumiunium. Leaks via glass-to-frame and frame to frame.
  • Louvres – similiar to windows, a bit worse. Leak via handles’ entries.
  • Sliding doors – seem pretty well sealed

Comparison to an Old Queenslander

On the same site is an old Queenslander (around 60 years old) with a raised tongue and groove floor. It looks like this:

I tested this as a reference:

Scenario Notes Air flow ACH50 PERM50
Units l/s @ 50Pa ACH50/hr m³/hr/m²
As-is 1950 26 23
Normalise kitchen Tape shut open vent and terrible door in kitchen. Leaks via other doors and windows, etc. 1800 24 21
Normalise bathroom Tape shut – loosely – bathroom door.Was leaking via exhaust and louvres. 1520 20 18
Remove B1 Tape well door to B1 and sleepout, removing lourves. 710 11 10

The original ‘as-is’ house is very leaky (and feels it in Winter!). It is the same as ACT houses in the referenced study. For the test, I used the internal door, as the external door is a sieve. Therefore the volume of the entry is excluded. Values are approximate as I guessed dimensions.

The old casement windows were actually pretty good. In contrast to the aluminium windows in the new build, they only leak around the edges. They could be weatherstripped. THe aluminium ones cannot be improved.

The floor gaps were also not terrible, as the tongue and groove mostly sealed the floor.

The key problems were obvious, and isolating them in the test above significant improved results:

  • Kitchen: open vent to outside, terrible french doors (2 ACH50 improvement)
  • Bathroom: old style wood lourves, exhaust fan (3 ACH50 improvement
  • Bedroom: 5m² of ‘plantation shutters’ in aluminium with terrible seals (8 ACH improvement!)

The final result of 10 ACH50 would be achieved if basic “fill the big holes” improvements were made. It would not require sealing the floor or weatherstripping. This gets to UK minimum standard. This is surprising, considering how leaky Queensland appear to be.

Leave a Reply

Your email address will not be published. Required fields are marked *