Window Testing – Getting it Right with a Mock-Up

Institutional building owners require a higher standard of building performance than other sectors. Imagine a new hospital having to shut down a wing for six months while it repairs a faulty window installation. It can’t happen, so they test for performance during the construction stage.

p1070179

Testing windows for air or water penetration usually consists of temporarily sealing off the inside face of the window frame to make a “chamber.” The chamber is depressurised to what ever specification the builidng envelope consultant specified. For measuring air leakage, the pressure across the window is usually 75Pascal. For inducing water penetration, the pressure can go as high as 720Pascals.

 

Just like the anxiety felt before an exam; it’s no different for curtain wall glaziers or window installers undergoing a field test for water penetration and air leakage. On commercial or institutional job sites, there’s a lot at stake pending the results of window tests.So it’s not uncommon for the project manager, the architect, the building envelope consultant, the glazier and the window manufacturer to be on hand to witness tests.

There are two kinds of windows;

those that leak and those that will leak.

                                                                                  Dr. Joe Lstiburek

It’s a strong statement. But after field testing all kinds of fenestration products, we’ve come to appreciate his point; it is only a matter of time before a bulb gasket or caulked joint fails and that’s just for the window, not the the installation!
For new construction, window products have to meet a minimum standard of performance and are lab tested to confirm they meet the standard. Our field testing of an installed product merely seeks to confirm that installed window will perform as it did in the lab. It’s important to note that almost all field window tests don’t require any performance testing on the transition joint between the fenestration product and the rough opening, so if you’re writing specifications insisting that a random, representative sample of windows get tested, consider testing both the window and the installation. Typically, the performance specifications are more demanding for commercial and institutional windows than they are for residential, even still, window failures are common on institutional projects.

I’m glad to report that when glaziers/installers are on site witnessing their installation being tested, they think differently. We often help them dissect a failed unit looking for water’s point of entry and they get it. They see how the window assembly failed and I can see that we’re all learning from the experience. Once a professional glazier experiences a test, they think differently. They carry the insights gleaned to all other installations and will (hopefully) never repeat the same assembly error or will get back to the manufacturer and ask serious questions about the product and its components.
Because BlueGreen Group does a good deal of high performance building envelope testing in the residential market, we’ve begun to pay more attention to windows. What we’re finding, with surprising regularity, are systemic air leaks through window assemblies, even those made overseas by reputable European companies. Lesson learned: no testing means unknown performance.  Typically, the building owner will designate a window at grade as a “mock-up” where the full installation process is carried out to specification and the window is tested. If the window passes both air and water testing, they crews proceed. If it fails either test, a diagnosis of the problem is completed and repairs are made along with changes to the written specification if applicable.
While no one likes to have their work tested and it takes a bit of time and money to test, the result is always a better quality installation, with increased performance and it gives the window manufacturing industry the impetus to improve their product and installation techniques.

Tags:

%d bloggers like this: