Making an Airtight House: Finding the Biggest Air Leaks First

airtight membrane attachment failure

Up to this point, we had performed our preliminary air tightness efforts; had our airtight membrane installed; and blew a preliminary blower door number of 847cfm equating to about 0.85 ACH50.  Not bad for the first go 'round.

The next step was to perform a thorough blower door test with the help of an expert in energy auditing.  We opted, as we did in all of our subsequent blower door tests, to also use a high-end (and very expensive) Fluke thermal imaging device that could also take digital pictures with precise thermal spreads to the 10th of a degree.

This way we would see, in real-time, exactly where and how we were failing in establishing an airtight interior conditioned space.  We were pleasantly surprised to find areas of air-sealing that we had missed with Floris, such as a couple of open plumbing lines that weren't sealed and a couple of areas where the membrane hadn't been attached well.  Having identified and remediated these issues, the real work (and fun) began...

Ferreting out failures in air sealing gave me a healthy reality check as to why so many passive houses are of simple and minimalistic shape and configuration.   It's a helluva lot easier to build 'em that way!

airtight membrane attachment failure in a corner

Just alone from an air tightness standpoint, lots of wall intersections, cathedral ceilings, trey ceilings, bumped-out bay windows, and the many dreaded corners all required great pains to seal completely.

complex corner in bay window resulting in air-leakage

The purpose of our first blower door test was not to determine what the ACH50 number was, its purpose was to tell us where and how we were failing in creating an airtight envelope; it served as our baseline for all subsequent air tightening efforts.

Towards the conclusion of this initial blower door test we managed to attain an air-leakage rate of 687cfm at 50 Pa or 0.69 ACH50 about 15% short of the Passivhaus airtightness standard.  This was our baseline.

Just to satisfy my curiosity we cranked the blower door up to 75 Pa.

An interesting thing happened, which was totally unexpected.  The air-leakage rates remained about the same.  We then cranked the blower door slowly up towards 100 Pa, being mindful of the potential damage we could do to our otherwise unprotected membrane or other interior things that wouldn't take too kindly to too much negative pressure.

Again we were surprised to see that the air-leakage numbers didn't increase, that is right up to the time that the blower door sucked itself out of the door frame!  How's that for a blower door test; a blown door.

All joking aside, I discovered a few empirical facts, about air leakage and air flow rates relative to air pressure, that I will share as I think they are important.

First, increasing positive or negative air pressure, doesn't necessarily mean that air exfiltration or infiltration rates will rise accordingly.  The reason, we concluded was simple: fluid dynamics.

There is a limit to how much fluid (in this instance air) can flow through any given opening--that limit is its maximum volumetric flow rate.  If the flow-rates (ie; air-leaks) are already at their maximum at any given pressure, simply increasing pressure will not increase the air flow (or air leak).

Second, and as a consequence of our first observation, it would be a mistake to assume an air-leakage rating obtained from a test according to ASTM E283 (at 75 Pa) would result in an improved rating if performed merely at 50 Pa.  The reason is that the air-leakage flow rate may already be at its maximum at 50 Pa, or even 20 Pa.

This realization proved very valuable to me going forward from that time. Now, every time I read a particular air-leakage rating performed at 75 Pa, I use that number in anticipating its performance in conditions of only 50 Pa.  Prior to this, I had been mistakenly assuming those performance ratings would improve by one-third.  No longer.

Third, as openings get smaller, air leakage (air flow) velocities go up (until they reach their limit).  This is a double-edged sword when it comes to identifying progressively smaller and smaller air leaks.  On one hand it makes small leaks easier to find with the hand, but it can also make it harder because as one goes through the iterative tasks of plugging leaks, the overall air leakage rate, as measured by a blower door, may (and oftentimes does) not appreciably change for the better! 

And that my friends is the real challenge.  Plugging the leaks in a manner sufficient to drop the aggregate air leakage rate.  And for that to happen, the remaining largest leaks, must already be at their maximum flow rates, for if they are not, plugging other leaks will only cause their flow rates to increase.

This is why, I believe there are no short cuts to  making an airtight house.  It has to be accomplished with an iterative process of diminishing returns; you have to walk that asymptotic line.